Abstract

Both Gideon Rose's neoclassical realism and Andrew Moravcsik's liberalism attempt to solve the problem of how to incorporate domestic factors into international relations theory. They do so in very different ways, however. Liberalism is a “bottom-up” perspective that accords analytic priority to societal preferences; neoclassical realism is a “top-down” perspective that accords analytic priority to systemic pressures and treats domestic factors as intervening variables. These two approaches are not equivalent, and the choice between them has high stakes. Although it has gained rapidly in popularity, neoclassical realism is fundamentally flawed. Its intellectual justification is weak; it is logically incoherent; and it induces the commission of methodological errors. Realism can incorporate certain domestic factors without losing its theoretical integrity, but it does not need and should not use neoclassical realism to do so.

Introduction

The late 1990s was a fertile time for new thinking about how to incorporate domestic politics into international relations theory. First, in “Taking Preferences Seriously,” Andrew Moravcsik formulated a new paradigm to integrate diverse strands of existing scholarship on the topic.1 The result, which he called “liberalism,” made societal preferences the analytic foundation of state behavior. Then, in “Neoclassical Realism and Theories of Foreign Policy,” Gideon Rose argued that several recently published books had successfully synthesized structural realism with fine-grained insights about the domestic factors that mediate systemic pressures.2 Rose dubbed this research program “neoclassical realism” and championed it as an alternative to established theories both within and outside the realist paradigm. In short, both Moravcsik and Rose sought to draw new conceptual boundaries for the study of international relations.

Thus far, the field's response to these two works has been quite different. Most scholars have accepted Moravcsik's characterization of liberalism as a paradigmatic alternative to realism, but few have described their own work as liberal, even when it qualifies as such. Meanwhile, neoclassical realism has generated considerable enthusiasm. Rose has not developed it further, but others have written theoretical essays that expand on its logic. Even more important, numerous individuals have identified their empirical research with it. As of 2017, more than forty published works have used the term “neoclassical realism” or “neoclassical realist” in their title, and countless others have aligned themselves with it. If paradigm debates are popularity contests, neoclassical realism has already won.

What is at stake here? Each perspective represents an attempt to “solve” the problem of combining levels of analysis in international relations theory. Liberalism is a bottom-up, domestic society–based theory that nevertheless incorporates external variables; neoclassical realism is a top-down, international system–based theory that nevertheless incorporates internal variables. If their strengths and weaknesses were parallel, the choice between them might simply be a matter of personal preference. Those who believe that systemic pressures matter more could declare themselves to be neoclassical realist, whereas those who believe that domestic politics matter more could declare themselves liberal, and it would make little difference for the progress of the field.

In this article, I argue that the choice is not so easy. There are right and wrong ways to incorporate domestic variables into international relations theories, and bad choices lead to flawed inferences. Liberalism has the advantage of being designed specifically to focus attention on domestic politics, and it can generate powerful insights into what moves states. Thus far, however, the case for liberalism has not been well made: its proponents have hurt their cause, and lowered their own standards, with the claim that all research on the domestic sources of state behavior is liberal. Liberalism is defined by its analytic method; it does not own the domestic level of analysis. Neoclassical realism, on the other hand, is fundamentally flawed. As originally conceived, it contained ambiguities that all but ensured that it would become increasingly incoherent once taken up by others. Those tensions have become outright contradictions in recent attempts to develop its logic. Realism can incorporate certain domestic factors in a way that is consistent with its core assumptions, but doing so requires close attention to the boundaries of the paradigm. Neoclassical realism transgresses those boundaries and, as a result, leads to error.

I proceed to this conclusion in several steps. First, I lay out the core assumptions of realism and liberalism, focusing on the ways in which realism can incorporate internal variables and liberalism can incorporate external variables. Second, I examine the founding statement of neoclassical realism by Rose. He argues that realist theories can engage with domestic politics, but he does not set boundaries on this engagement. Consequently, the first two decades of neoclassical realist research saw a wide range of viewpoints about the appropriate role of domestic politics in realist theory. In the third section, I examine four of the most important of such works and evaluate their proposed solutions. Fourth, I review the latest and most ambitious restatement of neoclassical realism, Neoclassical Realist Theory of International Politics, by Norrin Ripsman, Jeffrey Taliaferro, and Steven Lobell.3 These authors attempt to unite the different strands of the literature, but it is an impossible task. I point out the internal contradictions in neoclassical realism, explain why they matter, and argue that the paradigm should be abandoned. In the conclusion, I call on realists to recognize the pitfalls of research that crosses levels of analysis but fails to “take preferences seriously.”

Paradigms

Before Kenneth Waltz's Theory of International Politics, all realism was classical realism.4 What makes classical realism “classical” is not its attention to human nature, ideas, domestic politics, or any other causal variable, but rather its lack of self-conscious scientism. That is, it is not formulated as a nomothetic social science. Such an enterprise did not exist at the time of Thucydides, Machiavelli, and Hobbes, at least not by current standards.5 Even in the twentieth century, classical realists explicitly denounced attempts to treat international relations as a social science.6 They sought to distill the wisdom of the ages, not to advance an academic research program. As Joseph Parent and Joshua Baron write, “Because they tended to be historians, like Meinecke, or were strongly influenced by historians, like Morgenthau, or were trying to gain the ear of policymakers unimpressed by social science, like Kennan, their works contain a multiplicity of causes for at least as many outcomes on every level of analysis.”7 They felt no obligation to state all of their assumptions, derive hypotheses from those assumptions, and maintain a strict separation between endogenous and exogenous theoretical factors. This practice did not prevent them from developing sharp insights into international politics, but it did limit the ability of future scholars to develop those insights into a coherent body of theory, and it excludes classical realism from consideration as a scientific paradigm. It must instead be categorized at a higher level of conceptual aggregation, that of a “worldview”8 or “tradition of thought.”9

Waltz broke from this classical approach. He was the first to formulate realism as social science: he carefully laid out his concepts of system and structure, developed a logic of interaction based on microeconomic theory, and generated hypotheses from that logic. What makes contemporary realism “neo,” therefore, is its self-conscious attempt to constitute a scientific paradigm.10 A similar process occurred with liberalism. “Classical” liberalism is a hodgepodge of theories about the causes of peace and prosperity; it is distinguished by its optimism rather than its internal coherence. Not until Moravcsik organized preference-based theories into a systematic alternative to realism could liberalism have been considered a scientific paradigm.11 Finally, constructivism, the most recent addition to the “big three,” was reconceived as a paradigm by Alexander Wendt barely a decade after its conceptual birth.12

Waltz's status as exemplar of the modern realist paradigm is complicated by the fact that one of the central elements of his theory sharply limits its analytic utility: the distinction between international politics and foreign policy.13 International politics is about systemic outcomes (e.g., “balances will form against powerful states”); foreign policy is about state behavior (e.g., “state A will balance against state B”). Waltz argues that his theory explains only the former, not the latter.14 In his view, the logic of anarchy is so strong that it creates selection pressures on state preferences. A state might prioritize some goal other than survival, but it will suffer for doing so, given the ever-present threat of war. If it does not learn from its mistake, or become “socialized,” it will eventually be eliminated. Thus, at any given time, the states populating the international system are likely to be those that prioritize survival. The qualifier “likely” is key: Waltz claims only that “structure selects” for states to prioritize survival, not that it forces them to do so.15 Given this uncertainty, realism cannot make claims about foreign policy.

Had the field accepted Waltz's injunction, much of the phenomena that realists seek to explain would have remained off limits. Waltz himself offers only three hypotheses: balances form; successful innovations are emulated; and bipolarity is more stable than multipolarity. These claims are so vague as to be almost unfalsifiable, and, to the extent that they are falsifiable, they are often wrong.16 Only through the study of foreign policy can realism constitute a scientific research program. Waltz's bare-bones theory had to be fleshed out, and a new generation of scholars rose to the challenge.17 As a result, realism came to be defined not by an evolutionary logic, which cannot be used as the basis of theories of foreign policy, but by the assumption that states are rational actors motivated by survival, which can.18

Over the past two decades, scholars have constructed various realist “subparadigms” to explain how survival-motivated states respond to the pressures of anarchy. Each employs the rationality assumption, and each serves as the basis of a theory of foreign policy. The most prominent are defensive, offensive, and hegemonic realism. Defensive realists argue that states balance effectively against threats.19 A state that acts overly aggressively will find itself encircled by adversaries, a dangerous situation. To avoid this outcome, a rational state will act moderately, or “defensively.” Offensive realists, in contrast, assert that collective action dilemmas impede cooperation between states.20 States try to “pass the buck” to each other, so aggression often pays. Under such conditions, “offensive” behavior is imperative for survival. If one state does not take advantages of opportunities to increase its relative power, then some other state will. Finally, hegemonic realism derives from Robert Gilpin.21 He argues that states compete to structure in their favor the distribution of gains and losses from international cooperation and conflict. The state with the most “prestige,” defined as the reputation for power, becomes the hegemon. It attempts to maintain its status for as long as possible, but eventually it faces challenges from states whose prestige is not yet commensurate with their rising power.

The full set of assumptions of the realist paradigm, shared by each of these subparadigms, is laid out systematically by Patrick James.22 Drawing from Imre Lakatos, he describes them as the “hard core” to which all realist theories must conform.23 He finds six “axioms with parametric status,” which I relabel (with R for realism, instead of P for parameter) and condense slightly.

R1 The most important actors in world politics are territorially organized entities …

R2 State behavior is rational.

R3 States seek security and calculate their interests in terms of relative standing within the international system.

R4 Anarchy is the ordering principle of international relations.

R5 States … are undifferentiated by function.

R6 Structure is defined by the distribution of capabilities among states.24

James derives these axioms from Waltz's Theory of International Politics, with the notable exception of the rationality assumption (R2). That one is a correction: it resolves the problems that arise from Waltz's evolutionary theory and distinction between international politics and foreign policy, as discussed above.

James's six assumptions accurately represent the ontology of the modern realist paradigm. That is, they define the essential concepts of the realist world. Two additional points are needed, however. First, realism has an epistemology: it is nomothetic, in that it is designed to reveal lawlike regularities in state behavior. Second, and more important for this article, it has a method: it requires that systemic variables take priority in any analysis of state behavior. In other words, realism is a top-down paradigm. Every realist theory must start with a specification of systemic imperatives; only then can it address other factors.

This methodological rule is essential to understanding not only realism itself but also the boundary between realism and liberalism. Liberalism, as developed into its paradigmatic form by Moravcsik, adopts the opposite position: it “rests on a ‘bottom-up’ view of politics in which the demands of individuals and societal groups are treated as analytically prior to politics.”25 Every liberal theory must start with a specification of societal actors and their preferences; only then can it address other factors. Given this difference, the line between the two paradigms is not fuzzy, as some scholars claim.26 It is definitive, controlling, and has critical implications for understanding the sources of state behavior.27

Of course, liberalism also differs from realism on ontological grounds. Moravcsik states his assumptions explicitly, which I abridge and quote with the addition of identifying labels:

L1The fundamental actors in international politics are individuals and private groups

L2States … represent some subset of domestic society, on the basis of whose interests state officials define state preferences and act purposively in world politics.

L3The configuration of interdependent state preferences determines state behavior…. [E]ach state seeks to realize its distinctive preferences under varying constraints imposed by the preferences of other states.28

Finally, like realism, liberalism has a nomothetic epistemology, in that it is intended to generate “arguments, explanations, and predictions.”29

Realism and Domestic Variables

The two paradigms differ in their analytic method, top-down versus bottom-up. Neither, however, forbids engagement with the other's primary level of analysis. None of the core assumptions of realism explicitly excludes the use of domestic variables; none of the core assumptions of liberalism explicitly excludes the use of systemic variables. How far, then, can each be pushed? Some scholars suggest that anything goes: an infinite number of intervening variables can be layered into a theory without affecting its paradigmatic status. In contrast, this article argues that international relations paradigms become logically incoherent and methodologically flawed without close attention to their conceptual boundaries.

Realism's engagement with domestic variables is constrained by three of its ontological assumptions: R3, “States seek security and calculate their interests in terms of relative standing within the international system”; R4, “Anarchy is the ordering principle of international relations”; and R5, “States … are undifferentiated by function.” These three assumptions are tightly linked. If anarchy were not the ordering principle, states would not have to prioritize their security; their behavior would not be determined by their relative power; and they could become differentiated. In Waltz's terminology, they would no longer be “like units.”30 Thus, realist theories must not use domestic variables in a way that would make states functionally unlike each other. In a realist world, states can vary in their power; they cannot vary in their preferences.

The need for “like units” is especially important for the realist definition of “security” (in R3). “Security” is a slippery term, one that can easily be stretched in an ad hoc manner.31 Is it defined narrowly, in terms of homeland borders, or should it include overseas colonies, foreign commerce, and diplomatic commitments? Does it allow for concessions in autonomy and sovereignty? Realists are not obliged to agree with one another as to the meaning of security, but they are obliged to use it consistently within any given theory. That is, all states must define security the same way. Otherwise, they would be functionally differentiated, which the paradigm disallows.

For what purpose, then, can a realist theory incorporate domestic variables? As James explains, the central concepts of any paradigm are “subject to revision (as needed) at the level of measurement.” In the case of realism, “the failure of a hypothesis produces questions about the empirical meaning of capability, security, and other concepts within a particular analysis rather than the hard core of axioms itself.”32 One target for conceptual refinement is “rationality” (in R2). Humans are not perfect calculators of information; rather, they are subject to a range of cognitive biases. Because the makers of foreign policy are human, foreign policy should be affected by those biases. It does little violence to the realist paradigm to suggest, for example, that states become risk-averse when experiencing unexpected gains and risk-acceptant when experiencing unexpected losses.33 Cognitive biases do not change states' goals; they only make the pursuit of such goals less efficient. Thus, their incorporation does not undermine the coherence of the realist paradigm.

More than for any other concept in realism, the measurement of “capabilities” (in R6) demands attention to domestic variables. Every realist agrees that capabilities consist of internal attributes such as economic output, technological expertise, natural resources, and population size. The state must transform these raw materials into instruments of foreign policy, such as economic aid and military power. Furthermore, no realist could deny that domestic institutions mediate this transformation. Private wealth must be taxed; resources must be mobilized; and soldiers must be conscripted. If institutions vary in their effectiveness at performing these tasks, they should be considered a part of state capability.34 For example, suppose that democracies are better than authoritarian states at raising revenue when faced with a sharp increase in military threat. If so, “regime type” enters into the measurement of capabilities and thus can be incorporated into realist theory. One could go even further into the complexities of domestic politics: suppose that governments that represent lower-class interests are better at the same task than those that represent upper-class interests.35 If so, then “class interest” enters into the measurement of capabilities as well. Properly conceived, and made explicit rather than smuggled in ad hoc,36 these adjustments can be made consistent with realism.

Of course, realists may not employ “regime type” and “class interest” willy-nilly. The assumption that states are “undifferentiated by function” (R5) creates a bright-line boundary between realism and liberalism. Domestic factors must be integrated into realism as components of structure, like capabilities. They cannot shape actor preferences. Thus, it is not permissible within the realist paradigm to incorporate some variant of the Hobson-Lenin theories of imperialism, which claim that either the “regime type” or “class interest” of a government affects its predilection for territorial expansion. Such theories are premised on the idea that what states want from foreign policy varies predictably according to their domestic attributes. In terms of international relations paradigms, they are liberal.37

Furthermore, realist theories must treat the domestic components of capabilities as exogenous. To explain why a state has a particular regime type or class interest would require an extended foray into state formation and social choice, domains in which realism is wholly out of its depth. Even if one were to assume that societal actors, just like states, seek above all to ensure their own survival, it would not follow that systemic pressures can explain domestic politics. The problem is that societal actors have different requirements for survival than states. Foreign invasion is one concern, but no less of a threat to them are crime, poverty, and political oppression. Thus, they do not necessarily defer to the “national interest.” Domestic politics follows a different logic from that of international politics, even when systemic pressures are extreme. This is why states at war experience strikes, riots, revolution, and secession. The more they squeeze society for desperately needed resources, the more likely society is to push back. Given this threat from below, states rarely mobilize to the fullest extent possible. Instead, they limp along with their hands tied.38 Realism informs bellecist theories of state formation, but the complexities of domestic political change are beyond its remit. Realists should note the effects of such change on capabilities, but they must not imagine that it can be theorized endogenously within realism.

Early Controversies

Over the past few decades, the influence of domestic politics on foreign policy has been a subject of constant controversy among realists. One of the most polarizing works is Jack Snyder's Myths of Empire.39 Snyder accords analytic priority to systemic pressures; thus, he starts with realism. Specifically, he adopts a baseline of defensive realism, which assumes that states balance effectively against threats. In this view, aggressive behavior is counterproductive, so rational states act defensively. The focus of Snyder's research, however, is not rational behavior. Instead, it is to explain deviations from the norm. To this end, defensive realism serves as an instrument of case selection: it indicates which states “overexpanded” and therefore do not fit defensive realist theory. To account for their behavior, Snyder turns to societal preferences. He argues that certain historical paths of economic development create political “pathologies” that empower domestic groups with a material interest in promoting nationalist “myths of empire.”

What should we make of this? In the second stage of the theory, state leaders do not “seek security and calculate their interests in terms of relative standing within the international system” (R3). Rather, they manipulate foreign policy for their personal benefit. This argument violates the core assumptions of the realist paradigm. Indeed, it is quintessentially liberal. The overall theory cannot be considered realist; only its first stage can.40 For advocates of “analytic eclecticism,” such a two-stage, two-paradigm approach may be perfectly appropriate.41 For those who prize axiomatic consistency, however, it raises red flags. As Fareed Zakaria explains, “Defensive realism begins with a minimalist systemic assumption that explains very little state behavior, generating anomalies instead. It then uses auxiliary domestic politics theories to explain away those inconvenient cases—which comprise much of modern diplomatic history. Lakatos terms such projects theoretically ‘degenerative.’”42

Zakaria's line of criticism has been extended by two of the founders of the modern liberal paradigm, Jeffrey Legro and Moravcsik.43 They argue that the only true realist theories are those that “assume the existence of high conflict among underlying state preferences”;44 generate an imperative for balancing against power; and exclude beliefs, information, and any unit-level characteristics other than material capabilities. They then review recent works by self-proclaimed realists and find that few conform to their standard. Such works typically assert that the distribution of power in the system is a necessary, but not sufficient, cause of state behavior. Thus, they need to incorporate domestic variables to explain foreign policy outcomes. According to Legro and Moravcsik, this makes them not realist, but rather liberal, and signals the degeneration of realism as a scientific paradigm.

Legro and Moravcsik make some valid points about the conceptual “sloppiness” and methodological errors in the works they review.45 As Gunther Hellmann notes, however, their broader thesis founders on their inconsistent use of philosophy of science.46 On the one hand, Legro and Moravcsik set up their argument in that field's terms: they identify realism as a paradigm; they specify “coherence” and “distinctiveness” as criteria for evaluating it; they delineate realism's boundaries vis-à-vis three other paradigms (one of which, the “epistemic paradigm,” is their own neologism); and they adopt the concept of paradigmatic degeneration from Lakatos.47 On the other hand, their own definition of realism has no apparent justification from philosophy of science: it is not based on a Kuhnian “exemplar” like Waltz's Theory of International Politics;48 it ignores the break between classical realism and neorealism; and it does not reflect contemporary practice in the scholarly community of self-described realists. As their critics point out, their formulation of realism is their own, and no one else's.49 In short, it is a straw man.

Some of Legro and Moravcsik's realist critics push back even further against their use of philosophy of science, dismissing talk of “abstruse Lakatosian matters”50 and denying that realism can be considered a paradigm at all.51 To do so, however, is a mistake. First, it requires a certain amount of special pleading for a scholar to identify as a social scientist whose research is realist but to deny that realism is a research paradigm or that philosophy of science has any relevance to it. Only classical realists, who are not in the business of nomothetic inquiry, could plausibly object to being subject to the scrutiny of philosophy of science. Even this position may not be defensible, given the importance of methodological rigor to the production of idiographic knowledge.52 Second, paradigms have their benefits.53 It is difficult to participate in a scholarly community devoted to the accumulation and systematization of knowledge without accepting the role they play in it. Third, and most important, philosophy of science, properly applied, helps reveal the boundaries within which realism can take domestic factors into account without becoming internally inconsistent and methodologically flawed. The trend in recent years, however, has increasingly been to disregard those boundaries.

Founding Flaws

The incorporation of domestic factors into realist theories has drawn as much praise as criticism. For many realists, it does not mark the degeneration of the paradigm but rather is a sign of disciplinary progress. In an influential review of several such works, Rose argues that they retain the essential insights of realism while providing greater explanatory leverage over state behavior.54 Viewed collectively, they constitute a coherent school of thought, which Rose dubs “neoclassical realism.” He argues that it provides the basis for a “general theory of foreign policy,” on par with offensive and defensive realism.55 In the language of this article, he designates it as a subparadigm.

Rose argues that neoclassical realism fills a unique need in international relations theory, one unmet by existing realist or liberal perspectives. This assertion rests on two claims. The first is about the limitations of structural realism. According to Waltz, structural realism is a theory of international politics, not foreign policy: it explains systemic outcomes, not state leaders' decisions. To move from the former to the latter requires the development of a new (sub)paradigm.56 Second, Rose argues that liberalism is unsuited for this task. He describes it as a paradigm of Innenpolitik, composed of “pure unit-level explanations.”57 Because it focuses exclusively on the domestic sources of foreign policy, it ignores systemic constraints. Realism, in contrast, is a systemic paradigm. It has a built-in advantage: it recognizes that systemic constraints always loom in the background of foreign policymaking, and consequently it can explain how strategic necessity comes to triumph over domestic preferences.

Both of these foundational claims are mistaken. First, as discussed above, Waltz's distinction between international politics and foreign policy does not define the modern realist paradigm. It has not impeded the flourishing of the subparadigms of offensive, defensive, and hegemonic realism; thus, it does not necessitate the creation of an alternative to them. Second, Rose's characterization of liberalism, in which the international system is said to be “unimportant,” is wrong.58 No liberal international relations theorist, not even Immanuel Kant, would agree. In fact, as formulated by both Legro and Moravcsik, liberalism explicitly incorporates international factors.59 The first step of every liberal theory is to specify domestic actors' preferences and how they are aggregated politically; the second step is to predict how policymakers will carry out the goals of their constituency given the constraints of the international environment. Consider, for example, a state leader who represents a domestic interest group that derives income from an overseas empire, and who therefore places a high priority on the defense of the empire. The leader will puzzle long and hard over what diplomatic maneuvers and military commitments will best serve that end, knowing that other states' pursuit of their own preferences may either facilitate or impede it.60 This second step is integral to liberal analyses (per L3). Thus, one cannot argue that neoclassical realism is needed to compensate for the inherent limitations of liberalism.61

That liberalism already bridges domestic politics and the international system does not mean that realism cannot do the same. As explained above, it can and does. This creates an additional problem for the justification of neoclassical realism, however. If the core assumptions of the realist paradigm already permit the use of some aspects of domestic politics, why create a new paradigm to license it? Realists in every subparadigm have the option of adjusting their theories in the manner discussed above. Neoclassical realists have no special advantage over the rest, at least not if they wish to remain within the constraints of the modern realist paradigm. If they do not wish that, then there would be nothing “neo” about them: they would be classical realists, both unbound from and unbenefited by paradigmatic strictures. Neoclassical realism is not needed to fill any holes in the paradigmatic landscape; it is superfluous.

Superfluity, however, is the least of neoclassical realism's founding flaws. Another is its basis in two logically unrelated claims. As Rose defines it, neoclassical realism is an unwieldy hybrid that invites confusion and contestation over its meaning. On the one hand, it focuses attention on “how systemic pressures are translated through unit-level intervening variables.” Threats and opportunities are hard to measure and evaluate; thus, domestic factors matter. On the other hand, Rose makes a sweeping claim about state motives. He asserts that the “central empirical prediction of neoclassical realism” is that the “resources countries possess will shape the magnitude and ambition … of their foreign policies: as their relative power rises states will seek more influence abroad, and as it falls their actions and ambitions will be scaled back accordingly.”62 In essence, he substitutes “influence” for what Gilpin calls “prestige.”

There is no necessary, or even affinitive, connection between these two claims. One could favor the use of unit-level variables in realism without assuming that states are influence maximizers, and one could assume that states are influence maximizers without favoring the use of unit-level variables in realism. Even Rose himself does not assert a relationship between them: unit-level intervening variables are not what make states greedy for influence, and states' greed for influence is not what necessitates the study of unit-level intervening variables. In short, the two distinctive features of neoclassical realism do not interrelate or interact.

Must a neoclassical realist accept both elements of the theory?63 In practice, most works self-identified as neoclassical realist have only nodded to the motivational component. Instead, they have focused on neoclassical realism's use of domestic politics. Some reject its motivational component altogether. Taliaferro, for example, argues that neoclassical realists need to incorporate a theory of motives from other subparadigms.64 Thus, he divides neoclassical realism into two schools of thought: defensive neoclassical realism and offensive neoclassical realism. In Rose, defensive and offensive realism are alternatives to neoclassical realism; in Taliaferro, they are subcomponents of it.

One of the main purposes of a paradigm is to provide conceptual clarity to a collective enterprise. It allows individual scholars, who might live on different continents, have no communication, and know nothing of each other's work in progress, nevertheless to contribute to the same project for the accumulation of knowledge. Thus, disagreement among self-identified neoclassical realists over the first principles of neoclassical realism is problematic.65 Given its superfluity, ambiguities, and disconnections, neoclassical realism is bound to generate endless debates over its meaning.

Building on a Flawed Foundation

Rose seeks to liberate realism from a narrow focus on structure, but he does not formulate neoclassical realism as a cry for ontological anarchy. He addresses only two kinds of unit-level intervening variable: perceptions of relative power dynamics and the ability of the state to mobilize resources. In terms of the essential concepts of realism discussed above, he seeks to refine “rationality” and “capabilities.” Furthermore, he speaks of a single neoclassical realist theory, not a disparate set of theories under a broad neoclassical umbrella. It has a “central empirical prediction” and makes specific claims about individual states' behaviors. It promises not analytic freedom but rather “satisfying comprehensive explanations of foreign policy” while retaining “significant abstraction and parsimony.”66

At no point, however, does Rose make explicit the boundaries of neoclassical realism's engagement with domestic politics. Nor does he draw a clear line between it and liberalism. To the contrary, he anticipates convergence: “If neoclassical realists continue to incorporate unit-level intervening variables into their basic power-oriented argument, ironically, they might find themselves bumping into chastened Innenpolitikers coming from the other direction.”67 Rose does not elaborate on this claim; thus, neoclassical realists would be justified in seeing it as a green light for the incorporation of any sort of domestic factor whatsoever into realism, without constraint.

Such has been the case. Numerous scholars have taken up the banner of neoclassical realism, but few have adhered faithfully to Rose's vision. Instead, they have adapted it to their own perspectives and purposes, some of which are mutually incompatible. There are far too many works in the literature to address in this space, so I focus on a handful of monographs by well-known scholars who explicitly identify their theories as neoclassical realist.68 My approach is critical, but I do not attempt to provide a comprehensive review of their strengths and weaknesses. Rather, my goal is to demonstrate the analytic and methodological problems that arise from their use of neoclassical realism.

Among these books, Mark Brawley's Political Economy and Grand Strategy hews most closely to the core assumptions of the realist paradigm.69 Brawley seeks to explain how states, particularly great powers, respond to threats. He argues that three variables are critical: the immediacy of the threat, the availability of allies, and the rate at which the state under threat can convert wealth into military power. The most novel claim in this argument is about the relationship between the first and last variables. In the short term, a state's response to threat is constrained by its capacity to extract resources from society. In the long term, however, society will recognize that security is a public good and acquiesce to the demands that the state places on it. Thus, systemic pressures trump domestic politics.

The problem with this argument is not that it contradicts realism, but rather that it takes realism too far. As explained above, realism can incorporate an array of domestic variables into the concept of “capability,” but it must treat those variables as exogenously given. Brawley's attempt to graft onto realism a “second-image reversed” theory of domestic political change simply does not work.70 At the start of the book, Brawley boldly claims that states “override domestic cleavages formed on the basis of other interests” when responding to external threats. In the case studies, however, the state struggles against public opinion, confronts outright opposition from parties out of government, and loses control of its reforms.71 By the end of the book, every theoretical claim comes with an escape clause: the state is “relatively autonomous”; it is “often quite capable of exercising” authority; “[s]ecurity from external threat may trump preferences on other policies”; and “security from foreign threats tends to be a collective good.”72 So qualified, the theory is unfalsifiable.73

One of the purposes of paradigms and subparadigms is to signify that scholars in the same school of thought agree over fundamentals. This has not occurred, however, within neoclassical realism. One of its earliest and most vocal proponents, Randall Schweller, takes on the same question as Brawley and reaches nearly the opposite conclusion. In Unanswered Threats, he writes that “whether states choose to balance against threats … is not primarily determined by structural-systemic factors but rather, like all decisions that concern national defense, by the domestic political process.”74 His argument proceeds in two stages. He first specifies a baseline theory of how states “ought” to behave in the absence of internal constraints, then he explains how such constraints alter state behavior.75 Three variables are key: elite cohesion, social cohesion, and government/regime stability. When all three are high, elites reach consensus on foreign policy, and the state responds appropriately to systemic pressures. If one or more declines, such that elites and society become fragmented and the government/regime becomes vulnerable, then elites divide over foreign policy and the state experiences policy paralysis. Most modern states experience such divisions to some degree; thus, “underbalancing” is common.

The second half of Schweller's theory, which analyzes the domestic determinants of mobilization capacity, is consistent with the realist paradigm. It treats domestic politics not as the wellspring of state behavior but rather as a constraint on its capacity for action. Regardless of what states want to do, they cannot effectively balance against threats or take advantage of opportunities for expansion if they lack the political ability to do so. Thus, Schweller refines the realist concept of capability (in R6) “at the level of measurement.” His three domestic variables—elite cohesion, social cohesion, and government/regime stability—are simply components of capability, and he treats them as exogenously given.

The problematic part of Schweller's argument is its analytic starting point, his baseline theory. It initially resembles the hegemonic realism derived from Gilpin and designated by Rose as the “central empirical prediction” of neoclassical realism. In this view, states seek influence in proportion to their power. Those that were defeated in the last major war, and therefore disadvantaged in the peace settlement, are particularly keen to challenge the status quo distribution of power. So far, so good: states “calculate their interests in terms of relative standing within the international system” (R3). Where Schweller departs from realism is in his inclusion of a distinction between “limited-aims revisionists” and “unlimited-aims revisionists.” The former type of state seeks to improve its position within the existing international order; the latter is on “a quest for global domination and ideological supremacy.” With “revolutionary states” such as Revolutionary France and Nazi Germany, there can be no accommodation; their wars are “not a clash of interests but of ideologies.”76 In other words, revolutionary and non-revolutionary states are functionally differentiated, a violation of a core assumption of the realist paradigm (R5).

If the revolutionary ideology of unlimited-aims revisionist states could be explained by the dynamics of mobilization and had no further consequences for state behavior, it might be possible to reconcile it to realism. This would be a difficult case to make, however, and Schweller does not attempt it. To the contrary, he argues that mobilization capacity “alone tells us nothing about how a state chooses to use its enhanced capabilities.”77 Ideology, not systemic pressures, is what motivates revolutionary states. It also determines the objectives of states that few would consider revolutionary. U.S. territorial expansion in the 1840s, for example, “was rooted more in idealism than realpolitik.”78 Schweller attributes variation in state behavior to domestic preferences, but he does not theorize the origin of those preferences. This is neither realism nor liberalism; rather, it is ad hoc.

Another book that attempts to incorporate ideas into realism is Colin Dueck's Reluctant Crusaders.79 In his “neoclassical realist model of strategic adjustment,” Dueck echoes Rose's assumptions about the implications of anarchy: states seek influence in proportion to their power, and the international system is the “most important long-term cause of changes in any nation's foreign policy behavior.” At the same time, systemic pressures are often indeterminate, which means that there is “a range of strategic options that might plausibly serve … [the] national interest.”80 Dueck's contribution to this framework is his assertion that culture fills the gap. He examines three critical junctures in the history of the United States and finds that, in each case, its behavior was constrained by its liberal national identity. That identity ruled out some strategic options entirely. When more than one option remained, the choice among them was determined by contestation among “strategic subcultures”: internationalist, nationalist, progressive, and realist. International conditions affect which subculture is most persuasive, but this is not the only relevant factor. Also important are the level of conceptual fit between the subculture and the liberal national identity of the United States, electoral politics, and individual politicians' leadership.81

Is this realism? The answer depends on the relationship of strategic culture to established realist concepts. It is not an element of capability (in R6), nor is it a refinement of rationality (in R2). Instead, it affects the meaning of security (in R3). If states have different “national identities,” and those identities shape their strategic behavior, then those states define their security in different terms. This makes states functionally differentiated, contradicting one of the paradigmatic assumptions of realism (R5). Thus, Dueck's theory is not realist. His strongest connection to realism is his claim that the international system is the most important long-run determinant of state behavior. This assertion is not the focus of his book, however. At no point does Dueck operationalize the meaning of “most important,” nor does he test it against other explanations. It is a commonplace among neoclassical realists, but it signifies little, as discussed later in this article.

In his attempt to layer strategic culture into realism, Dueck imposes sharp limits on the causal role of ideas. He does not argue that they determine state motives; rather, he says that they “narrow down the range of acceptable policy options.”82 This limitation does not suffice to keep the theory within the paradigmatic boundaries of realism. It does, however, create an artificial barrier to the analysis of the impact of ideas on state behavior. If the national identity of the United States is so powerful that it can persist through 250 years of history, affecting the decisions of policymakers at every step of the way, how is it merely a “permissive cause, or filter,”83 rather than an underlying source of preferences? To assert a role for strategic culture in the making of foreign policy is to open a Pandora's box of confounders for realism.

Consider the case of the United States at the end of World War I. Dueck argues that policymakers had three options: commitment to the League of Nations, an alliance with France, or disengagement from Europe.84 In his theory, those options should derive from realism. Yet what realist would argue that systemic pressures, even indeterminate ones, create a “national interest” in collective security?85 According to Dueck, “From a realist perspective, a peacetime Western alliance through a nominal league commitment was an entirely feasible alternative—in fact, it was the optimal one.”86 This claim is, to put it mildly, debatable. Even if realists accept it, however, there is a further problem: Woodrow Wilson's commitment to the League was not “nominal.” Rather, it was “breathtakingly ambitious.” When Senate Republicans sought to “destroy his vision of a League of Nations by gutting the crucial Article 10 commitment to a collective security system,” he refused to compromise.87 If realism sets the menu of choice, Wilson's actual strategy would never have been considered, much less pursued. The final strategic option, disengagement, is also problematic. It cannot be derived from Dueck's own theory of systemic pressures, which claims that states pursue influence in proportion to their power. In his words, “Given America's immense material power by 1918, international conditions could not have predicted the return to disengagement.”88 Thus, strategic culture is no mere filter. Rather, it dictates choices far removed from realism.89

Christopher Layne's The Peace of Illusions also examines the sources of U.S. foreign policy, asserts the importance of ideas, and identifies with neoclassical realism.90 It has little else in common with Dueck's book, however. First, Layne does not adopt the same theory of systemic pressures. His realism does not assume that states should expand their influence in proportion with their power. Quite the contrary, he argues for a narrow view of security, consistent with offensive and defensive realism.91 Within the Western Hemisphere, the United States enjoys regional hegemony and a highly defensible geographic position. Its only concern should be the rise of a peer competitor, so its systemic imperative is a strategy of “offshore balancing,” designed to prevent any other great power from establishing regional hegemony over Europe or East Asia. Under most circumstances, other states will balance against threats in their own neighborhood, so the United States need not establish a permanent presence in these distant regions. Rather, it is better off husbanding its resources and avoiding commitments overseas until all other alternatives are exhausted.

Why, then, has the United States done the opposite of this since World War II? Layne argues that the international environment was uniquely permissive in 1945: Western Europe was a vacuum of power, and the United States was by far the most powerful country in the world. Systemic conditions therefore allowed the United States to pursue a strategy of “extraregional hegemony.” Why it chose to do so, however, is another question. According to Layne, the United States possesses a distinctive pathology: “Open Door” economic interests and ideology. Drawing from the Marxist historians William Appleman Williams and Walter LaFeber, the latter of whom provides a back-cover endorsement of The Peace of Illusions, Layne argues that politicians and policymakers, under the influence of corporate lobbyists, believe that the freedom and prosperity of the United States depends on the maintenance of a liberal international economic order. To this end, the United States requires a global network of military bases and diplomatic commitments. Although it could have withdrawn from Western Europe as early as the 1950s, and its economy groans under the ever-increasing burdens of geopolitical overextension, the United States will follow this strategy “as long as the current foreign policy elite remains in power.”92

In a diagram of his theory, Layne labels the “motors of expansion,” Open Door interests, and ideology, as intervening variables, while conferring the status of independent variable on the “opportunity for expansion,” the international configuration of power in 1945.93 This modeling decision makes little sense. In realism, constraints are analytically prior to motives because motives are endogenous to constraints. In liberalism, motives are analytically prior to constraints because motives are exogenous to constraints. One cannot split the difference, as Layne does, and claim both that motives are exogenous to constraints and that constraints are analytically prior to motives. Neoclassical realism cannot serve as a via media between realism and liberalism, because the two paradigms rest on logically irreconcilable assumptions. Instead, it is a rationale for incoherence.

In response to this criticism, Layne's best defense might be to assert that the United States is exceptional. In his discussion of realist subparadigms, he accepts the offensive realist claim that “great powers that live in very dangerous neighborhoods … face almost irresistible structural pressures to break out of the omnipresent insecurity condition.”94 Thus, he seems to discount Snyder's argument that World Wars I and II were caused by domestic pathologies. Germany and Japan were responding to systemic imperatives; only the United States violates the rule. For someone so committed to the idea that domestic interest groups warp the U.S. “national interest,” and who draws so heavily on Marxist historiography, this would be an odd position to hold. Might not other countries be afflicted by the same malady? Again, neoclassical realism seems to constitute an artificial barrier to the analysis of the broader implications of the “intervening” variable under consideration. For Dueck, it limits inquiry into the full impact of strategic culture within his cases; for Layne, it limits inquiry into the full impact of economic interests and ideas on other cases. As a tool for the expansion of knowledge about the role of domestic politics in foreign policy, it leaves much to be desired. It fails the test of realism, yet it is a poor substitute for liberalism.

Each of the books discussed in this section make interesting and important claims about the domestic sources of foreign policy. The point of this section is not to prove them wrong; far from it. Rather, it is to interrogate the meaning of neoclassical realism among those who explicitly identify their own research with it. Several conclusions follow from this exercise. First, none of the four authors holds the same view of neoclassical realism. Brawley and Dueck accept Rose's baseline theory; Schweller and Layne do not. Brawley adopts theoretical constructs that are consistent with the realist paradigm; Schweller, Dueck, and Layne do not. Neoclassical realism emerges not as a coherent subparadigm but rather a Rorschach blot that can take the shape of whatever one imagines it to be. Second, neoclassical realism does not effectively facilitate the production of knowledge. It encourages scholars to move just far enough out of the paradigmatic orbit of realism that they lose the logical consistency of its core assumptions, yet it discourages them from confronting the full set of challenges and implications of their engagement with domestic politics. Halfway between realism and liberalism, it is an unhappy medium.

Third, these works collectively encompass a wide range of perspectives on social behavior. What is domestic politics about: contestation between subcultures, control by dominant elites and economic interests, or the authority of a relatively autonomous state? To imagine that these works are all contributing to the same intellectual project would require an impressive degree of ecumenicism, a willingness to accept that everything matters and that each perspective captures a significant portion of the truth. Suppose that neoclassical realists were to take this position. As each new piece of neoclassical research was published, the amount of variance in state behavior explained by domestic politics would expand. How long, then, would neoclassical realists be able to claim that systemic pressures matter more than domestic factors? In the long run, therefore, neoclassical realism is self-defeating. The more it reveals that domestic factors matter, the less able it will be to sustain its justifying assumption that systemic pressures deserve analytic priority.

The Newest Neoclassical Realism

After nearly two decades of intellectual ferment, neoclassical realism has reached its logical endpoint. Neoclassical Realist Theory of International Politics, by Ripsman, Taliaferro, and Lobell, is an ambitious attempt at integrating various strands of research identified as neoclassical realism into a deductive theoretical structure.95 Much like Waltz, Moravcsik, and Wendt, the authors self-consciously present their approach “in paradigmatic terms” and refer to it as a “research program” that “presents a common set of assumptions.”96 Furthermore, they characterize it as a “direct descendant of structural realism” and sharply differentiate it from classical realism. Neoclassical realism aspires “to the positivistic scientific rigor that structural realism introduced to realism”; classical realism does not.97 Thus, their vision of neoclassical realism falls squarely within the analytic boundaries of this article, whether it constitutes a subparadigm of realism or a fully fledged paradigmatic rival to realism (as discussed below).

The neoclassical realism of Ripsman, Taliaferro, and Lobell is distinctive in several ways. First, departing from Rose, and following Taliaferro, the authors avoid making definitive claims about state motives.98 Instead, they propose that every work of neoclassical realist research must adopt a baseline theory that defines the standard for rational state behavior. Options include offensive realism, defensive realism, balance of threat theory, hegemonic theory, or any other realist subparadigm. Second, Ripsman, Taliaferro, and Lobell identify a wide range of intervening variables that mediate between systemic pressures and state policy. Four categories of variable affect the perceptions and policy options of policymakers: cognition (“leader images”), strategic culture, state-society relations, and domestic institutions. Finally, the authors deduce the conditions under which certain variables should matter more than others. They offer two considerations: the level of information that policymakers have about the international environment (“systemic clarity”) and the level of threat that policymakers face (“nature of strategic environment”).

As noted above, Rose incorporates domestic factors into just two of realism's core concepts: rationality and capabilities. With a limited view of which aspects of domestic politics matter to realism, he is able to develop these concepts in a way that is consistent with the realist paradigm. In contrast, Ripsman, Taliaferro, and Lobell open the floodgates. No variable is excluded from their neoclassical realism, and there are no effective constraints on how such variables are used.99 Consider their discussion of strategic culture.

Through socialization and institutionalization (in rules and norms), these collective assumptions and expectations become deeply entrenched and constrain a state's behavior and freedom of action by defining what are acceptable and unacceptable strategic choices…. For instance, democratic transnational norms and culture are the causal mechanisms of one of the central explanations for the democratic peace or special peace among liberal democratic states since 1815…. We would also include dominant ideologies, which can affect the state's attitudes toward international affairs and willingness to use force, and degrees of nationalism as important components of state culture.100

In this discussion, they cite one of the scions of liberalism, Bruce Russett, and several prominent constructivists, including Alexander Wendt. Thus, neoclassical realism incorporates some of the most foundational, distinctive, and novel claims of rival paradigms. Moreover, it does so without any attempt to reconcile these factors with realism's core concepts. Culture, norms, and ideologies are preferences: they define what states want. They cannot be derived from the survival motive, and they vary from state to state.

Ripsman, Taliaferro, and Lobell's discussion of state-society relations is equally expansive. The authors argue that the “foreign policy executive,” if autonomous and insulated from societal forces, will respond to the “demands of the external environment.”101 When exposed to the demands of societal actors, however, it may reverse its priorities.

To the extent that a particular socio-economic interest group, economic sector, or coalition of interests captures the state, it [the state] may be unable to enact policies that diverge from the preferences of that underlying coalition. This can occur either because the leaders are drawn from that political coalition and therefore view international affairs through the prism of their parochial interests, or because they recognize that they can maintain their power positions only by satisfying their support base's demands. In this regard, scholars who take a political economy approach to the state assume that states captured by inward-oriented nationalist coalitions will pursue politics of protectionism and military competition, whereas those whose dominant coalition is comprised of outward-oriented internationalists will pursue grand strategies of freer trade and international cooperation. Thus, unless the state possesses sufficient institutional autonomy to shield it from domestic pressures, the composition of the dominant coalition and its relationship with the state can affect the state's policy preferences.102

In other words, there is no guarantee as to what motivates state leaders in neoclassical realism. If they are not insulated from domestic politics, they may be unwilling, even “unable,” to respond effectively to systemic pressures, and societal preferences will be key to explaining their behavior.

An implicit assumption of the foregoing passage is that some set of institutional characteristics differentiates autonomous states from captured states. As described by Ripsman, Taliaferro, and Lobell, however, capture is indistinguishable from everyday politics. When, in either authoritarian or democratic states, are leaders not drawn from and responsive to a political coalition? In Plato's Republic, perhaps; nowhere else. All state leaders are biased both by their coalition's parochial interests and their own interest in remaining in office; thus, “the composition of the dominant coalition and its relationship with the state” are always an underlying cause of state behavior. The authors do not endorse this conclusion, but it follows logically from their premises.

Ripsman, Taliaferro, and Lobell assert that the influence of societal actors is circumscribed not only by the state's institutional autonomy but also by the level of external threat. Only in a “permissive” strategic environment will state-society relations affect foreign policy; a “restrictive” strategic environment will “compel societal actors to the sidelines in the national interest.”103 This claim is no less problematic than the previous one. First, as discussed above and demonstrated by Schweller, states often fail to overcome internal barriers to the mobilization of resources against existential threats. This is one of the central findings of the research of one of the most vocal advocates of neoclassical realism; it cannot simply be assumed away. Second, the strategic environment of a state at time t is shaped by its behavior at t – 1. Suppose, for example, that policymakers are drawn from a coalition that has a parochial interest in imperialism. When the strategic environment is permissive, the state expands its empire. When the strategic environment becomes restrictive, will it abandon the territories it acquired? If the same coalition remains in power, the answer is almost certainly not: policymakers will label the empire a vital “national interest” and devote considerable resources to its defense.104 An increase in threat does not allow for a strategic reset that zeroes out the effect of societal preferences in earlier times; history matters.

The underlying problem with neoclassical realism is that it attempts to operationalize an irretrievably amorphous intuition: that domestic politics matters, but not too much. Theoretical tensions are therefore inevitable. At one point, Ripsman, Taliaferro, and Lobell seem to step back from all of the claims quoted above: “Since neoclassical realism is, like all realist theory, a state-centric approach, it has comparatively little to say about the behavior of non-state actors, be they corporations, societal interest groups…. To explain the behavior of these actors, we would need theories of the motives, interests, and constraints they face, which neoclassical realism does not attempt to do.”105 At another point, the authors cite as an example of the most evolved, theoretically sophisticated form of neoclassical realism a book that takes the “political economy approach” to state-society relations.106 Then, in a later chapter, that same book is cited as a “liberal approach.”107 In short, confusion reigns.

Ripsman, Taliaferro, and Lobell insist that societal actors' interests are merely intervening variables, and they assure their readers that neoclassical realists consider systemic pressures to be more important than domestic politics. This position amounts to little more than hand-waving, however. If state behavior is a function of both systemic pressures and societal preferences, what does it mean to call the former variable “independent” and the latter “intervening”? And by what criteria can one measure which variable is more important? The two interact; both are necessary and neither is sufficient. As Benjamin Fordham writes, “It makes no sense to talk about the demands of the international system in the abstract, since these are as much a function of the domestic faction in question as of the international environment.”108 It is a matter of faith, not social science, to claim ontological priority for one or the other. The more relevant question is which variable deserves methodological priority, as I explain later in this article.

Another problem with the claim that systemic pressures are more important than domestic politics is that Ripsman, Taliaferro, and Lobell's neoclassical realism has no definite view of the systemic pressures that it deems to be more important. As noted above, the authors propose that every neoclassical realist analysis begin with a baseline theory, chosen from one of the realist subparadigms. That baseline provides an objective standard against which state behavior can be judged. If behavior diverges from expectations, one can infer that a domestic variable intervened on the policymaking process. The problem is that different baselines will produce markedly different expectations and, therefore, divergent inferences about the importance of domestic politics. Offensive and defensive realists, looking at U.S. foreign policy in the Middle East under President George W. Bush, might find marked deviation from their subparadigmatic standard of rationality and therefore see a major role for domestic politics. Hegemonic realists, looking at the same case, might find only a few mistakes in the way in which U.S. foreign policy was executed and therefore see a minor role for domestic politics. Without a theory of motives, neoclassical realism cannot claim that systemic pressures are more important in any given case.109 It can claim only that systemic pressures matter more in some metaphysical, unfalsifiable way. This is not a solid foundation on which to build theory.

The neoclassical realism of Ripsman, Taliaferro, and Lobell is caught in an inescapable dilemma. On the one hand, if it regards baseline theories as relatively unimportant, then it must concede that structure is relatively unimportant, and it cannot claim to be realist. On the other hand, if it claims that baseline theories account for most of the variation in state behavior, and much depends on the choice of baseline, then the defining feature of any given neoclassical realist theory is not its neoclassical realism, but rather its baseline. Thus, it should be identified as a work of that baseline subparadigm, not of neoclassical realism. In that case, neoclassical realism would have to be seen as something less than a paradigm or subparadigm: a shared intuition, perhaps, but not a coherent school of thought.

Paradigm, Subparadigm, or Neither

What is neoclassical realism: a paradigm, subparadigm, or neither? Ripsman, Taliaferro, and Lobell are not entirely clear. On the one hand, they state that neoclassical realism is a “direct descendant” of structural realism and that it “fully unleashes the explanatory power of structural realism.”110 On the other hand, they assert that it is “not subsidiary to structural realism” and is a “better overall theory.”111 Given that neoclassical realism departs so radically from the core assumptions of modern realism, it is probably safest to conclude that it should be treated as a competing paradigm, one that is a direct descendent not of Waltz's Theory of International Politics but rather of the classical realist worldview—what Ripsman, Taliaferro, and Lobell call “the broader Realpolitik tradition.”112 This categorization leaves the authors free to violate the modern realist paradigm without being subject to the charge of causing its degeneration.113 When they call themselves “realist,” they do so under a less strict standard than that accepted by most other realists. In particular, they abandon the assumption that states are functionally undifferentiated (R5).

Absent this assumption, what makes neoclassical realism authentically realist? To their credit, Ripsman, Taliaferro, and Lobell address this question explicitly: “Theories properly designated as neoclassical realist in accordance with the principles set out in this book all share the common assumptions that attest to their structural realist origins. Specifically, they assume (1) that the international system is anarchic and, consequently, that states must rely on themselves to ensure their survival; (2) that survival is the most important national interest in an anarchic realm; and (3) that anarchy makes cooperation difficult, as it leads states to prefer relative over absolute gains.”114 The first claim, that the international system is anarchic, is purely factual and not a major cause of inter-paradigm controversy. Moreover, the third point should be set aside: the notoriously unproductive debate over relative gains has been solved and no longer serves as an important marker of paradigmatic boundaries.115 Tellingly, it plays almost no part in Ripsman, Taliaferro, and Lobell's exposition.

The key point, therefore, is the second: anarchy forces states to place their survival before other interests. Yet, as explained above, Ripsman, Taliaferro, and Lobell do not adhere to it. Whenever states are “captured” by societal groups, they choose policies that satisfy those groups' interests, “even at the expense of international ones.” Similarly, strategic culture may lead a state to “self-defeating behavior … that jeopardize[s] its primary security interests.”116 Thus, neoclassical realism violates not only the core assumptions of modern realism but also its own. This is not a paradigm; it is a jumble of contradictions.

Paradigmatic Imperialism

One of the founding flaws in Rose's formulation of neoclassical realism, as noted above, was its mischaracterization of liberalism. Contrary to Rose, liberalism does not regard the international system as unimportant to state behavior. Rather, liberalism asserts that every policymaker must wrestle with the question of how to pursue societal preferences given systemic constraints. Survival is almost always one such preference,117 but it does not preclude the pursuit of other preferences, nor is there a fundamental tension between them. Indeed, survival is necessary to achieve most domestic actors' international goals. Few groups would benefit from military defeat and occupation; thus, most can be expected to support policies needed to ensure their state's survival when faced with a threat from the external environment.

What liberalism lacks is not attention to external factors but rather a parsimonious view of system dynamics. A liberal theory would not suggest that the international system has inherent tendencies toward balancing, buck-passing, or hegemony; rather, it would assert that the relationship between states is determined by the interaction of their preferences (per L2). In a system of multiple states with multiple preferences, complexity may be too high to generate reliable predictions about the likelihood and nature of conflict.118 In such a situation, baseline theories are counterproductive: they make oversimplified assumptions about what states want and consequently produce abundant anomalies. Thus, liberals should be reluctant to use them.

Rose's mischaracterization of liberalism has spread to other neoclassical realists. Ripsman, Taliaferro, and Lobell similarly contrast neoclassical realism to “Innenpolitik theories that understand foreign policy solely as the product of domestic political coalitions or leader preferences.”119 If neoclassical realists were to take liberalism seriously, rather than treating it as a useful straw man, they would have to pay more attention to paradigmatic boundaries and, consequently, recognize some of the contradictions within neoclassical realism itself. As things stand, confusion is inevitable. For example, Ripsman, Taliaferro, and Lobell write that “Innenpolitik approaches incorrectly assume that the external environment does not impose preferred strategies on states,” whereas in neoclassical realism, the policy effects of “external threats will be moderated by the unit-level variables we identify.”120 If one strikes the word “incorrectly,” then what, exactly, is the difference here?

A common symptom of inattention to boundaries is “paradigmatic imperialism,” the claim to encompass viewpoints that originated in opposition to one's own. As initially formulated by Moravcsik, liberalism was clearly defined. When Legro and Moravcsik asked, “Is Anybody Still a Realist?” however, they contrasted liberalism with a caricature of realism. The result was unfair to realists and a setback to liberalism itself, suggesting that it had less rigorous standards for the treatment of societal preferences than originally laid out. Having misrepresented Innenpolitik approaches, Ripsman, Taliaferro, and Lobell commit a similar error. As explained above, they appropriate the central findings of both liberalism and constructivism: democratic peace, economic interdependence, normative socialization, and strategic culture. Their neoclassical realism is a kind of “theory of everything.”121 In attempting to explain nearly everything, it excludes nearly nothing. In the hands of neoclassical realists, realism becomes trivial.

Methodological Bets

As Jennifer Sterling-Folker notes, a common marker of the success of a paradigm is not its logical coherence, its methodological rigor, or even its ability to generate new insights into social behavior. Rather, it is the ability of that paradigm to find self-identifying adherents and generate research under its banner.122 By that standard, neoclassical realism has been a smashing success. Suppose it continues to flourish: What will its output look like? Its trajectory thus far has been clear. At first, Rose called on realists to address nuances of policymaking and pay close attention to diplomatic history, a research program to which only a die-hard behavioralist could object. Yet he failed to set boundaries on this endeavor, and neoclassical realism became a kind of catch-all paradigm for anyone who admired the insights of classical realists and disliked the austere parsimony of Waltz's Theory of International Politics. It brought together all sorts of scholars who seek to contribute to a broad research program based on the vague premise that systemic variables matter most but domestic variables matter too. Ripsman, Taliaferro, and Lobell's Neoclassical Realist Theory of International Politics is the logical endpoint of this trajectory. In pushing Rose's argument to its limit, however, it has deepened its contradictions. The result is an ostensibly realist paradigm that, with its menu of baselines, makes no definite claim about how the international system functions, and that, with its claims about strategic culture and state-society relations, allows any state to possess preferences that directly undermine its security.

What is at stake here? One might argue that, so long as the theories produced by neoclassical realists are persuasive, concerns about intellectual coherence are mere quibble. If neoclassical realism is methodologically flawed, however, it is less likely to produce persuasive theories. Moravcsik, in his original statement of the liberal paradigm, explains why.

Waltz, Keohane, and many others recommend that we synthesize theories by employing realism first (with preferences assumed to be invariant) and then introducing competing theories of domestic politics, state-society relations, and preference change as needed to explain residual variance. Yet this conventional procedure lacks any coherent methodological or theoretical justification. Methodologically, the procedure introduces omitted variable bias by arbitrarily privileging realist explanations of any phenomena that might be explained by both realist and liberal theories, without ever testing the latter explanation…. Preferences determine the nature and intensity of the game that states are playing and thus are a primary determinant of which systemic theory is appropriate and how it should be specified…. In short, liberal theory explains when and why the assumptions about state preferences underlying realism or institutionalism hold, whereas the reverse is not the case.123

The choice of a paradigm is not just an ontological bet; it is also a methodological one. If a paradigm's method is flawed, it will generate incomplete or mistaken claims about the sources of state behavior. In trying to reconcile a top-down method in which security ostensibly matters most (per realism) with the possibility of varying preferences (per liberalism), neoclassical realism sets up its adherents for error.

The methodology of neoclassical realism is particularly susceptible to what Jeffry Frieden calls “sins of omission.” It “often leads analysts to ascribe outcomes to strategic interaction itself, without paying careful attention to actors' underlying preferences. Disagreements among states might be owing to the difficulty of sustaining cooperation in an uncertain and information-poor world that lacks third-party enforcement; or they might, in fact, occur because of inherently antagonistic goals of two or more states. Trade protection could be the result of a breakdown of cooperation or of a simple desire on the part of both countries to reduce their imports—for whatever domestically derived reason.”124 Frieden's example is apt. In a chapter titled “Resolving Key Theoretical Debates with Neoclassical Realism,” Ripsman, Taliaferro, and Lobell propose to explain Britain's turn toward protectionism as early as 1931–32 as a response to “scarce security.”125 What about the Great Depression, the consequent collapse of the British economy, its effect on domestic interest groups, and the representation of their interests by the government? These go unmentioned.126 In neoclassical realism, preferences matter, but they are not taken seriously.

Conclusion

Neoclassical realism began its journey at a fork in the road. As conceived by Gideon Rose, it made two distinct claims: first, that the realist concepts of “rationality” and “capabilities” should include certain domestic variables, and second, that the realist view of state motives should emphasize the pursuit of influence. These two ideas were neither logically connected nor clearly delimited. Thus, scholars could make what they wanted of neoclassical realism. This was part of its appeal: its ambiguities invited scholars to appropriate and adapt neoclassical realism to their own research agendas. It rapidly gained adherents, but few hewed faithfully to Rose's original vision. Most downplayed its motivational component or abandoned it altogether. Instead, they focused on domestic politics. In doing so, however, they went far beyond Rose. Instead of merely refining “rationality” and “capabilities,” they abandoned the assumption that states are “like units.” Their so-called realism asserts the analytic priority of systemic pressures, but it makes few claims about the nature of those pressures, and it opens the door for societal preferences to undermine the survival motive. Lacking internal coherence, neoclassical realism has become less a scientific paradigm than an invitation to methodological error.

What is to be done? Neoclassical realism is beyond saving; it should be abandoned. Scholars who believe that the international system is important enough to take analytic priority in explaining state behavior should identify their research by their baseline assumptions about state motives and system dynamics: offensive, defensive, or hegemonic realism. This list is not exhaustive, and it will almost certainly expand as scholarship evolves. To qualify as realist, however, a theory must assume (1) that all states have the same motive, (2) that this motive is derived from a desire for survival, and (3) that changes in state behavior are determined by changes in the international environment, not a change in underlying preferences. A theory that departs from any one of these rules is not part of the realist paradigm. It might be influenced by the classical realist worldview, but it belongs in a different paradigm—“realist constructivism,”127 for example—or none at all, per “analytic eclecticism.”128

Contrary to Jeffrey Legro and Andrew Moravcsik, realism does not preclude the use of domestic variables. As discussed above, there are ways that realists can refine their core concepts without overstepping their paradigmatic boundaries. Most of the works originally reviewed by Rose do, in fact, meet the requirement that states are motivated by survival and are functionally undifferentiated. Scholars whose own work meets this standard should make explicit exactly how they do so. Furthermore, to avoid confusion on the part of others who might build on their research, they should identify the conditions under which the use of their domestic variables would “cross the line” and violate the realist paradigm. Attention to paradigmatic boundaries has the further benefit of mitigating tendencies toward misrepresenting rival paradigms' central claims and engaging in paradigmatic imperialism. With boundaries clearly specified, debate between paradigms should focus on substantive points of historical interpretation, where it is most likely to generate knowledge about important phenomena.

Although realism does not forbid engagement with domestic politics, it does impose significant limitations on what can be said about it. Consider, again, Randall Schweller's Unanswered Threats. Schweller argues that elite cohesion, social cohesion, and government/regime stability should be treated as components of a state's capability. That is, he focuses on a concept that is central to realism and refines it at the level of measurement. As explained above, this part of his theory is consistent with the realist paradigm. Whether it is intellectually satisfying, however, is another question. If elite cohesion, social cohesion, and government/regime stability affect states' capacity for action, should they not also affect states' motives for action? It is improbable that political conflicts among societal actors could be so severe as to cripple the state's ability to respond to foreign threats yet not affect the goals that the state pursues in the international system. Thus, societal dissensus over the means of foreign policy implies societal dissensus over the ends of foreign policy. Within the realist paradigm, however, only capability (means) can vary across states, not motives (ends). No matter how closely connected they may be in reality, realism forbids consideration of the latter.

Liberalism, in contrast, assumes that states have different motives and provides scholars with the analytic tools for understanding them. Such a task is theoretically and methodologically tricky. It requires adherence to certain standards of consistency and deductive logic, as laid out expertly in Jeffry Frieden's “Actors and Preferences in International Relations”—an essay that should be required reading for any scholar whose research crosses levels of analysis. Liberal theory construction is not simply a matter of layering in intervening variables to see how systemic pressures are translated into foreign policy. Rather, it requires close attention to the pathways through which social preferences are formed, aggregated, and represented by political coalitions. Scholars who look for evidence of domestic politics in foreign policy only in government archives and personal papers will miss out on the deeper structure of politics,129 which creates a close connection between constituents' economic interests and party leaders' ideologies.130 Liberalism demands attention to these dynamics, and consequently it has far more to say about the role of domestic politics in foreign policy than realism.

Given the advantages of liberalism, scholars who wish to explore the impact of domestic politics on state behavior should consider adopting it. For longtime realists, this would require not a Damascene conversion, but rather intellectual adaptability. Realism is an effective tool for the advancement of knowledge about international relations, but it is not a theory of everything, nor should anyone imagine that it can be made into one. In subjects of inquiry outside of its proper domain, it becomes the wrong tool for the task. We can tinker with its concepts, but if we push it too hard, we will break it. Of course, realists who believe that domestic politics is unimportant have no reason to change their minds. Those who both believe in its importance and seek to understand it, however, should think hard about the benefits of stepping completely outside the confines of their paradigm and, instead, taking preferences seriously.

Acknowledgments

The author is grateful to Arman Grigoryan, John Schuessler, and the anonymous reviewers for their comments and advice.

Notes

1. 

Andrew Moravcsik, “Taking Preferences Seriously: A Liberal Theory of International Politics,” International Organization, Vol. 51, No. 4 (Autumn 1997), pp. 513–553, doi:10.1162/002081897550447.

2. 

Gideon Rose, “Neoclassical Realism and Theories of Foreign Policy,” World Politics, Vol. 51, No. 1 (October 1998), pp. 144–172, doi:10.1017/S0043887100007814.

3. 

Norrin M. Ripsman, Jeffrey W. Taliaferro, and Steven E. Lobell, Neoclassical Realist Theory of International Politics (Oxford: Oxford University Press, 2016).

4. 

Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979).

5. 

Cf. Alan Ryan, “Hobbes's Political Philosophy,” in Tom Sorell, ed., The Cambridge Companion to Hobbes (Cambridge: Cambridge University Press, 1996), pp. 212–215.

6. 

Stefano Guzzini, Realism in International Relations and International Political Economy: The Continuing Story of a Death Foretold (London: Routledge, 1998), pp. 23–24, 37–38; and J. Samuel Barkin, Realist Constructivism: Rethinking International Relations Theory (Cambridge: Cambridge University Press, 2010), pp. 122–124. Cf. Colin Wight, “Philosophy of Social Science and International Relations,” in Walter Carlsnaes, Thomas Risse, and Beth A. Simmons, eds., Handbook of International Relations, 2nd ed. (London: Sage, 2012), pp. 32–34.

7. 

Joseph M. Parent and Joshua M. Baron, “Elder Abuse: How the Moderns Mistreat Classical Realism,” International Studies Review, Vol. 13, No. 2 (June 2011), p. 203, doi:10.1111/j.1468-2486.2011.01021.x. Cf. Martin Hollis and Steve Smith, Explaining and Understanding International Relations (Oxford: Oxford University Press, 1990), pp. 23, 25, 27; and Wight, “Philosophy of Social Science and International Relations,” p. 33.

8. 

Patrick James, International Relations and Scientific Progress: Structural Realism Reconsidered (Columbus: Ohio State University Press, 2002), pp. 70–71; and Randall L. Schweller, “The Progressiveness of Neoclassical Realism,” in Colin Elman and Miriam Fendius Elman, ed., Progress in International Relations Theory: Appraising the Field (Cambridge, Mass.: MIT Press, 2003), p. 323.

9. 

Timothy Dunne, “Mythology of Methodology? Traditions in International Relations Theory,” Review of International Studies, Vol. 19, No. 3 (July 1993), pp. 305–318, doi:10.1017/S0260210500117450; and Jennifer Sterling-Folker, “Forward Is as Forward Does: Assessing Neoclassical Realism from a Traditions Perspective,” in Annette Freyberg-Inan, Ewan Harrison, and Patrick James, eds., Rethinking Realism in International Relations: Between Tradition and Innovation (Baltimore, Md.: Johns Hopkins University Press, 2009), pp. 191–218.

10. 

Guzzini, Realism in International Relations and International Political Economy, pp. 127, 129; James, International Relations and Scientific Progress, p. 70; and Ole Wæver, “Waltz's Theory of Theory,” International Relations, Vol. 23, No. 2 (June 2009), p. 216, doi: 10.1177/0047117809104635.

11. 

Liberalism is often incorrectly identified with Robert Keohane's functionalist theory of international regimes. See Moravcsik, “Taking Preferences Seriously,” pp. 536–538. In fact, Keohane initially identified his approach as “modified structural realism,” because it was based on the core assumptions of realism. The label “neoliberal institutionalism” came from one of Keohane's critics, Joseph M. Grieco. See Grieco, “Anarchy and the Limits of the Cooperation: A Realist Critique of the Newest Liberal Institutionalism,” International Organization, Vol. 42, No. 3 (Summer 1988), pp. 485–507, doi:10.1017/S0020818300027715.

12. 

Alexander Wendt, Social Theory of International Politics (Cambridge: Cambridge University Press, 1999). Cf. Nicholas Onuf, “World of Our Making: The Strange Career of Constructivism in International Relations,” in Donald J. Puchala, ed., Visions of International Relations: Assessing an Academic Field (Columbia: University of South Carolina Press, 2002), pp. 119–141.

13. 

Colin Elman, “Horses for Courses: Why Not Neorealist Theories of Foreign Policy?” Security Studies, Vol. 6, No. 1 (Autumn 1996), pp. 7–53, doi:10.1080/09636419608429297.

14. 

Waltz, Theory of International Politics, pp. 121–122; and Kenneth N. Waltz, “International Politics Is Not Foreign Policy,” Security Studies, Vol. 6, No. 1 (Autumn 1996), pp. 54–57, doi:10.1080/09636419608429298.

15. 

Waltz, Theory of International Politics, p. 92.

16. 

See, for example, Stuart J. Kaufman, Richard Little, and William C. Wohlforth, eds., The Balance of Power in World History (Basingstoke, U.K.: Palgrave, 2007).

17. 

Elman, “Horses for Courses,” p. 10 n. 9; and John J. Mearsheimer, The Tragedy of Great Power Politics, 2nd ed. (New York: W.W. Norton, 2014), p. 432 n. 60.

18. 

James, International Relations and Scientific Progress, p. 122.

19. 

Jack L. Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.: Cornell University Press, 1991), pp. 11–12.

20. 

Mearsheimer, The Tragedy of Great Power Politics, pp. 19–22.

21. 

Robert Gilpin, War and Change in World Politics (Cambridge: Cambridge University Press, 1981), pp. 10–39.

22. 

James, International Relations and Scientific Progress, p. 121. Cf. Wæver, “Waltz's Theory of Theory,” p. 210.

23. 

Imre Lakatos, “Falsification and the Methodology of Scientific Research Programmes,” in Lakatos and Alan Musgrave, eds., Criticism and the Growth of Knowledge (London: Cambridge University Press, 1970), pp. 91–196.

24. 

James, International Relations and Scientific Progress, p. 121.

25. 

Moravcsik, “Taking Preferences Seriously,” p. 517.

26. 

Patrick Thaddeus Jackson and Daniel H. Nexon, “Paradigmatic Faults in International-Relations Theory,” International Studies Quarterly, Vol. 53, No. 4 (December 2009), pp. 907–930, doi: 10.1111/j.1468-2478.2009.00562.x.

27. 

Moravcsik, “Taking Preferences Seriously,” pp. 538–544.

28. 

Ibid., pp. 516, 518, 520 (emphasis in the original).

29. 

Ibid., p. 514.

30. 

Waltz, Theory of International Politics, pp. 93, 96–97.

31. 

Arnold Wolfers, “‘National Security’ as an Ambiguous Symbol,” World Politics, Vol. 67, No. 4 (December 1952), pp. 481–502, doi:10.2307/2145138; Jack Donnelly, Realism and International Relations (Cambridge: Cambridge University Press, 2000), pp. 54–56; and Burak Kadercan, “Making Sense of Survival: Refining the Treatment of State Preferences in Neorealist Theory,” Review of International Studies, Vol. 39, No. 4 (October 2013), pp. 1015–1037, doi:10.1017/S0260210512000538.

32. 

James, International Relations and Scientific Progress, p. 124.

33. 

Jeffrey W. Taliaferro, Balancing Risks: Great Power Intervention in the Periphery (Ithaca, N.Y.: Cornell University Press, 2004).

34. 

Fareed Zakaria, From Wealth to Power: The Unusual Origins of America's World Role (Princeton, N.J.: Princeton University Press, 1998); and Jeffrey W. Taliaferro, “State Building for Future Wars: Neoclassical Realism and the Resource-Extractive State,” Security Studies, Vol. 15, No. 3 (September 2006), pp. 464–495, doi:10.1080/09636410601028370. See also Brian Rathbun, “A Rose by Any Other Name: Neoclassical Realism as the Logical and Necessary Extension of Structural Realism,” Security Studies, Vol. 17, No. 2 (2008), pp. 301–303, doi:10.1080/09636410802098917.

35. 

Kevin Narizny, “Both Guns and Butter, or Neither: Class Interests in the Political Economy of Rearmament,” American Political Science Review, Vol. 97, No. 2 (May 2003), pp. 203–220, doi: 10.1017/S0003055403000625.

36. 

Annette Freyberg-Inan, What Moves Man: The Realist Theory of International Relations and Its Judgment of Human Nature (Albany: State University of New York Press, 2004), p. 150.

37. 

Some might object to calling a Marxist theory “liberal” and to other incongruities that arise from Moravcsik's choice of name for the paradigm. See Deborah Boucoyannis, “The International Wanderings of a Liberal Idea, or Why Liberals Can Learn to Stop Worrying and Love the Balance of Power,” Perspectives on Politics, Vol. 5, No. 4 (December 2007), pp. 703–727, doi:10.1017/S1537592707072180. A better choice would have been “neopluralism.” See Andrew S. McFarland, Neopluralism: The Evolution of Political Process Theory (Lawrence: University Press of Kansas, 2004).

38. 

Michael N. Barnett, Confronting the Costs of War: Military Power, State, and Society in Egypt and Israel (Princeton, N.J.: Princeton University Press, 1992); Victoria Tin-bor Hui, War and State Formation in Ancient China and Early Modern Europe (Cambridge: Cambridge University Press, 2005); and Randall L. Schweller, Unanswered Threats: Political Constraints on the Balance of Power (Princeton, N.J.: Princeton University Press, 2006).

39. 

Snyder, Myths of Empire.

40. 

Wæver, “Waltz's Theory of Theory,” pp. 212–214.

41. 

Rudra Sil and Peter J. Katzenstein, Beyond Paradigms: Analytic Eclecticism in the Study of World Politics (Basingstoke, U.K.: Palgrave Macmillan, 2011).

42. 

Fareed Zakaria, “Realism and Domestic Politics: A Review Essay,” International Security, Vol. 17, No. 1 (Summer 1992), pp. 192–193, doi:10.2307/2539162. See also Randall L. Schweller, “Neorealism's Status-Quo Bias: What Security Dilemma?” Security Studies, Vol. 5, No. 1 (Spring 1996), pp. 90–121, doi:10.1080/09636419608429277.

43. 

Jeffrey W. Legro and Andrew Moravcsik, “Is Anybody Still a Realist?” International Security, Vol. 24, No. 2 (Fall 1999), pp. 5–55, doi:10.1162/016228899560130.

44. 

Ibid., p. 14.

45. 

Peter D. Feaver, “Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?),” International Security, Vol. 25, No. 1 (Summer 2000), p. 166, doi:10.1162/016228800560426.

46. 

Gunther Hellmann, “Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?),” International Security, Vol. 25, No. 1 (Summer 2000), p. 171, doi:10.1162/016228800560426.

47. 

Legro and Moravcsik, “Is Anybody Still a Realist?” pp. 8–12.

48. 

Thomas S. Kuhn, The Structure of Scientific Revolutions, 4th ed. (Chicago: University of Chicago Press, 2012), pp. 10–11.

49. 

Hellman, “Brother, Can You Spare a Paradigm?” p. 170; Randall L. Schweller, “Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?),” International Security, Vol. 25, No. 1 (Summer 2000), pp. 175–176, doi:10.1162/016228800560426; William C. Wohlforth, “Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?),” International Security, Vol. 25, No. 1 (Summer 2000), p. 183, doi:10.1162/016228800560426; and Rathbun, “A Rose by Any Other Name,” pp. 298–301.

50. 

Schweller, “Brother, Can You Spare a Paradigm?” p. 176. See also Jeffrey W. Taliaferro, “Security Seeking under Anarchy: Defensive Realism Revisited,” International Security, Vol. 25, No. 3 (Winter 2000/01), p. 158, doi:10.1162/016228800560543.

51. 

Hellman, “Brother, Can You Spare a Paradigm?” pp. 172–174; and Jeffrey W. Taliaferro, “Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?)” International Security, Vol. 25, No. 1 (Summer 2000), p. 179, doi:10.1162/016228800560426.

52. 

Aviezer Tucker, Our Knowledge of the Past: A Philosophy of Historiography (Cambridge: Cambridge University Press, 2004); and Richard C. Carrier, Proving History: Bayes's Theorem and the Quest for the Historical Jesus (Amherst, N.Y.: Prometheus, 2012).

53. 

Larry Laudan, Progress and Its Problems: Towards a Theory of Scientific Growth (Berkeley: University of California Press, 1977), pp. 86–89, 92–93; and James, International Relations and Scientific Progress, pp. 77, 108.

54. 

Rose, “Neoclassical Realism and Theories of Foreign Policy.” See also Schweller, “The Progressiveness of Neoclassical Realism.”

55. 

Rose, “Neoclassical Realism and Theories of Foreign Policy,” p. 145.

56. 

See Taliaferro, “Security Seeking under Anarchy,” p. 133.

57. 

Rose, “Neoclassical Realism and Theories of Foreign Policy,” p. 148, including n. 8.

58. 

Ibid., p. 154, table 1.

59. 

Jeffrey W. Legro, “Culture and Preferences in the International Cooperation Two-Step,” American Political Science Review, Vol. 90, No. 1 (March 1996), pp. 118–137, doi:10.2307/2082802; and Moravcsik, “Taking Preferences Seriously.”

60. 

Kevin Narizny, “The Political Economy of Alignment: Great Britain's Commitments to Europe, 1905–39,” International Security, Vol. 27, No. 4 (Spring 2003), pp. 184–219, doi:10.1162/016228803321951135; and Kevin Narizny, The Political Economy of Grand Strategy (Ithaca, N.Y.: Cornell University Press, 2007).

61. 

There is a certain hypocrisy to this: realism, a systemic paradigm, is allowed to use domestic factors, but liberalism, a domestic paradigm, is not allowed to use systemic factors.

62. 

Rose, “Neoclassical Realism and Theories of Foreign Policy,” p. 152. Rose claims that this motive is not “security,” but he never suggests any other reason why states might be influence maximizers. Out of reluctance to over-interpret a single sentence clause, I do not pursue the point.

63. 

Asle Toje, Barbara Kunz, and Michael Agner, “Introduction: Neoclassical Realism in Europe,” in Toje and Kunz, eds., Neoclassical Realism in European Politics: Bringing Power Back In (Manchester, U.K.: Manchester University Press, 2012), pp. 4–5, 8–10.

64. 

Taliaferro, “Security Seeking under Anarchy,” p. 132 n. 14.

65. 

Toje, Kunz, and Agner, “Introduction,” pp. 8–10, 12.

66. 

Rose, “Neoclassical Realism and Theories of Foreign Policy,” pp. 152, 167, 168.

67. 

Ibid., p. 170.

68. 

For a sample of article-length works on neoclassical realism, see Steven E. Lobell, Norrin M. Ripsman, and Jeffrey W. Taliaferro, eds., Neoclassical Realism, the State, and Foreign Policy (Cambridge: Cambridge University Press, 2009); and Toje and Kunz, Neoclassical Realism in European Politics.

69. 

Mark R. Brawley, Political Economy and Grand Strategy: A Neoclassical Realist View (Abingdon, U.K.: Routledge, 2010).

70. 

Peter Gourevitch, “The Second Image Reversed: The International Sources of Domestic Politics,” International Organization, Vol. 32, No. 4 (Autumn 1978), pp. 881–912, doi:10.1017/S002081830003201X.

71. 

Brawley, Political Economy and Grand Strategy, pp. 91, 85, 122–124.

72. 

Ibid., pp. 140–141.

73. 

Other aspects of the book are also problematic—most notably, its caricature of liberal counterarguments. See ibid., pp. 19, 76–82, 91, 114–115, 121–122. Brawley repeatedly criticizes those counterarguments for their supposed implication that changes in government must result in changes in policy. In fact, they make explicit claims about the conditions under which the interests of coalitions converge, resulting in policy continuity despite turnover in office. See Benjamin O. Fordham, Building the Cold War Consensus: The Political Economy of U.S. National Security Policy, 1949–51 (Ann Arbor: University of Michigan Press, 1998); Peter Trubowitz, Defining the National Interest: Conflict and Change in American Foreign Policy (Chicago: University of Chicago Press, 1998), chap. 3; and Narizny, The Political Economy of Grand Strategy, pp. 22, 55–57, 63–64, 175–186.

74. 

Schweller, Unanswered Threats, p. 9.

75. 

Ibid., p. 23.

76. 

Ibid., pp. 29, 30.

77. 

Ibid., p. 105.

78. 

Ibid., p. 110.

79. 

Colin Dueck, Reluctant Crusaders: Power, Culture, and Change in American Grand Strategy (Princeton, N.J.: Princeton University Press, 2006).

80. 

Ibid., pp. 18, 19.

81. 

Ibid., pp. 33–36.

82. 

Ibid., p. 25.

83. 

Ibid., p. 166.

84. 

Ibid., p. 44.

85. 

See Kevin Narizny, “Rational Idealism: The Political Economy of Internationalism in the United States and Great Britain, 1870–1945,” Security Studies, Vol. 12, No. 3 (Spring 2003), pp. 1–39, doi:10.1080/09636410390442856.

86. 

Dueck, Reluctant Crusaders, pp. 52–54, at p. 54.

87. 

Ibid., pp. 61, 69.

88. 

Ibid., p. 164. See also pp. 51–52.

89. 

Another problem with Dueck's argument is that he does not test it against any other source of preferences, despite recognizing that domestic economic interests may underpin cultural differences. See ibid., pp. 169, 174 n. 21. Research into this question suggests that ideational conflict over U.S. strategic adjustment is, in fact, largely endogenous to economic interests. See Kevin Narizny, “The New Debate: International Relations Theory and American Strategic Adjustment in the 1890s,” Security Studies, Vol. 11, No. 1 (Autumn 2001), pp. 151–170, doi:10.1080/714005317.

90. 

Christopher Layne, The Peace of Illusions: American Grand Strategy from 1940 to the Present (Ithaca, N.Y.: Cornell University Press, 2006).

91. 

Ibid., pp. 19–22.

92. 

Ibid., p. 201.

93. 

Ibid., p. 31.

94. 

Ibid., p. 22.

95. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics.

96. 

Ibid., pp. 7, 177.

97. 

Ibid., p. 168.

98. 

Rose, “Neoclassical Realism and Theories of Foreign Policy”; and Taliaferro, “Security Seeking under Anarchy.”

99. 

See also Thomas Juneau, Squandered Opportunity: Neoclassical Realism and Iranian Foreign Policy (Stanford, Calif.: Stanford University Press, 2015).

100. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, pp. 67–68, 69.

101. 

Ibid., p. 71.

102. 

Ibid., pp. 72–73.

103. 

Ibid., p. 95.

104. 

Snyder, Myths of Empire.

105. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, p. 97.

106. 

Ibid., p. 104. The book in question is Steven E. Lobell, The Challenge of Hegemony: Grand Strategy, Trade, and Domestic Politics (Ann Arbor: University of Michigan Press, 2003).

107. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, p. 153 n. 44.

108. 

Benjamin O. Fordham, “The Limits of Neoclassical Realism: Additive and Interactive Approaches to Explaining Foreign Policy Preferences,” in Lobell, Ripsman, and Taliaferro, Neoclassical Realism, the State, and Foreign Policy, p. 257. See also Jeffry A. Frieden, “Actors and Preferences in International Relations,” in David A. Lake and Robert Powell, eds., Strategic Choice and International Relations (Princeton, N.J.: Princeton University Press, 1999), pp. 39–76.

109. 

Toje, Kunz, and Agner's vision of neoclassical realism, in which “what states actually want should be of central importance,” does not solve the problem, because they also assert that “states can want many things.” This is liberalism. See Toje, Kunz, and Agner, “Introduction,” pp. 9, 10.

110. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, pp. 180, 88.

111. 

Ibid., pp. 181, 88.

112. 

Ibid., p. 181.

113. 

See John A. Vasquez and Colin Elman, eds., Realism and the Balancing of Power: A New Debate (Upper Saddle River, N.J.: Prentice Hall, 2003).

114. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, p. 179.

115. 

Robert Powell, “Absolute and Relative Gains in International Relations Theory,” American Political Science Review, Vol. 85, No. 4 (December 1991), pp. 1303–1320, doi:10.2307/1963947; and Duncan Snidal, “International Cooperation among Relative Gains Maximizers,” International Studies Quarterly, Vol. 35, No. 4 (December 1991), pp. 387–402, doi:10.2307/2600947.

116. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, pp. 71, 69–70.

117. 

In some circumstances, such as the fulfillment of a unification project among sovereign states (Italy and Germany in the mid-nineteenth century), the voluntary dissolution of a federal state (the Soviet Union in 1991), or the peaceful secession of a substate unit (Czechoslovakia in 1992), the survival motive may be rejected entirely. Although unusual, these cases demonstrate that survival is not the raison d'ětre of states, which is a problem for realism.

118. 

Cf. Quddus Z. Snyder, “Integrating Rising Powers: Liberal Systemic Theory and the Mechanism of Competition,” Review of International Studies, Vol. 39, No. 1 (January 2013), pp. 209–231, doi:10.1017/S0260210511000593.

119. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, p. 96. See also ibid., p. 104; and Michiel Foulon, “Neoclassical Realism: Challengers and Bridging Identities,” International Studies Review, Vol. 17, No. 4 (December 2015), p. 642, doi:10.1111/misr.12255.

120. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, p. 146.

121. 

See Toje, Kunz, and Agner, “Introduction,” p. 7.

122. 

Sterling-Folker, “Forward Is as Forward Does,” pp. 213–214.

123. 

Moravcsik, “Taking Preferences Seriously,” pp. 542–543 (emphasis in the original).

124. 

Frieden, “Actors and Preferences in International Relations,” pp. 52–53.

125. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, pp. 154–155.

126. 

Ripsman, Taliaferro, and Lobell are on firmer ground when they note that the British government imposed economic regulations in 1937–38 to facilitate rearmament. They overlook, however, the serious inadequacies of these measures. See Narizny, “Both Guns and Butter, or Neither,” p. 210.

127. 

Barkin, Realist Constructivism.

128. 

Sil and Katzenstein, Beyond Paradigms.

129. 

Ripsman, Taliaferro, and Lobell, Neoclassical Realist Theory of International Politics, pp. 131–137.

130. 

Narizny, The Political Economy of Grand Strategy, pp. 25–30.