Liberalism

The following appears as “Liberal International Relations Theory: A Scientific Assessment,” in Colin Elman and Miriam Fendius Elman eds., Progress in International Relations Theory: Appraising the Field (Cambridge: MIT Press, 2003), 159-204. To appear as Chapter 5 in: Colin Elman and Miriam Fendius Elman, eds. PROGRESS IN INTERNATIONAL RELATIONS THEORY (Cambridge, MA: MIT Press, forthcoming 2002). proofs 7/24/02 5 Liberal International Relations Theory A Scientific Assessment Andrew Moravcsik T his paper advances three arguments. First, there exists a distinct liberal research program in international relations. Section 1 of this chapter proposes three “hard core” assumptions shared by all work within the liberal “scientific research program” in international relations (IR) and introduces three variants of liberal theory — ideational, commercial, and republican liberalism — that 1 share those assumptions. Each type of liberal theory explains interstate politics by tracing the influence of variation in pressure from domestic and transnational societal actors on underlying state preferences. Second, this liberal research program is “progressive.” Section 2 of this chapter assesses the liberal research program using three criteria for novel “excess content” derived from Imre Lakatos’s philosophy of science. Judged by these criteria, liberalism has been and continues to be a “progressive” research program. It appears progressive, moreover, no matter which received interpretation of empirical fruitfulness we employ, although the most useful, I argue, is “background theory 1. A scientific research program, the essential unit of analysis for a Lakatosian analysis of scientific progress, contains a hard core of inviolable assumptions, a positive heuristic, and a resulting “protective belt” of “auxiliary hypotheses.” I am grateful to Michael Barnett, Philip Cerny, David Dessler, Colin Elman, Miriam Elman, Stefano Guzzini, Stephen Holmes, Robert Jervis, Peter Katzenstein, Robert Keohane, Stephen Krasner, Jeffrey Legro, Lisa Martin, Robert Paarlberg, Randall Schweller, and Anne-Marie Slaughter for helpful suggestions and comments, and to Marius Hentea for swift and sure research assistance. An earlier version of this paper appeared as Andrew Moravcsik, “Liberal International Relations Theory: A Social Scientific Assessment,” Weatherhead Center for International Affairs Working Paper Series No. 01-02 (Cambridge, Mass.: Harvard University, 2002). 2 progress in international relations theory novelty.” (This interpretation of Lakatos turns out to be a more rigorous criterion than some, including Elman and Elman in Chapter 2 of this volume, believe it to be.) The progressive nature of the liberal paradigm is particularly evident in comparison with alternative paradigms, notably realism, which has recently tended to “degenerate” (in the strict Lakatosian sense) by borrowing arguments from competing liberal and non-realist paradigms that are incompatible with any plausible realist “hard core.” Third, we should be skeptical about Lakatosian criteria. The tendency of recent realists, including the editors of this volume, to overlook creeping incoherence in the realist paradigm might well lead us to question whether the invocation of Lakatosian philosophy of science provides sufficient incentive to impose logical consistency on IR theories. In Section 3 of this chapter, I highlight — very reluctantly, given how well my favored theory has performed — some limitations of any application of Lakatosian criteria to IR theory. To be sure, Lakatosian philosophy of science usefully highlights the need for consistent assumptions, and the “background theory novelty” criterion for measuring the empirical fruitfulness of assumptions offers a more useful standard for doing so than many believe. Yet the Lakatosian view of theoretical disputes as “fights to the finish” among a few monocausal theories, decided ultimately in favor of the theory that has the greater empirical scope, may impose too constraining a criterion to encourage creative, empirically fruitful social science. There is no reason to believe, at this stage in the development of IR theory, that only the theory with the widest scope is useful. Such a view forgoes two potential benefits of a less conflictual interaction among theories: the delineation of relative explanatory domains and the construction of creative multicausal syntheses. These, I submit, offer more fruitful roads forward for contemporary IR theory than gladiatorial combat among monocausal claims. We should adopt a healthy skepticism towards the doctrinaire application of Lakatosian philosophy of science, narrowly understood, to IR theory — a conclusion broadly consistent with most other contributions to this volume. Overall, this draft -- do not cite or quote without author’s permission moravcsik: chapter 5 3 conclusion may well be more consistent with Laudanian than a 2 Lakatosian philosophy of science. The Liberal Scientific Research Program This section frames liberal IR theory as a Lakatosian scientific research program, delineating the “hard core” assumptions and “protective belt” of auxiliary propositions. the hard core: three common assumptions The liberal scientific research program in IR places state-society relations at the center of world politics. It is based on the fundamental premise that a critical causal factor influencing a state’s behavior is the relationship between the state and the domestic and transnational society in which it is embedded. This basic insight can be restated in terms of the three hard core assumptions shared by all liberal theories, which specify the nature of societal actors, of the state, and of the 3 international system. These three assumptions distinguish liberal IR theory from realist, institutionalist, and epistemic (or constructivist) paradigms. The Nature of the Actors in International Politics. The first assumption is that the fundamental actors in international politics are rational individuals and private groups, who organize and exchange to promote their interests. Liberal theory rests on a “bottom-up” view of politics, in which the demands of individuals and societal groups are treated as exogenous causes of the interests underlying state behavior. Socially differentiated individuals define underlying material and 2. Larry Laudan, Beyond Positivism and Relativism: Theory, Method and Evidence (Boulder, Colo.: Westview Press, 1996). 3. A more detailed and fully cited version of some arguments in Sections 1 and 2 can be found in Andrew Moravcsik, “Taking Preferences Seriously: A Liberal Theory of International Politics,” International Organization, Vol. 51, No. 4. (Autumn 1997), pp. 513–553. Some material for Section 2 has been drawn from Andrew Moravcsik, “Liberalism and International Relations Theory,” Center for International Affairs Working Paper Series 92-6 (Cambridge, Mass.: Harvard University, 1992/93). draft -- do not cite or quote without author’s permission 4 progress in international relations theory ideational tastes and preferences concerning future “states of the world,” and advance them through political exchange and collective action. The central intuition is that we cannot understand the exercise of interstate power or promotion of interstate collective action unless we first understand what fundamental social purposes each state 4 seeks. Liberal theory thereby rejects the utopian notion of an automatic harmony of interest among individuals and groups in society. Rather, scarcity and differentiation render some competition inevitable. Patterns of political order and conflict result from the variations in the underlying pattern of interaction in pursuit of these preferences for material and ideal welfare. As an empirical matter, societal demands so conflictual that social actors are likely to consider coercion as an acceptable means to promote them tend to be associated with three factors: divergent fundamental beliefs, scarcity of material goods, and inequalities in political power. These three potential motivations define 4. This assumption should not be controversial. This is tantamount only to saying that relevant domestic groups have some consistent preferences concerning the ultimate goals of foreign policy, based on underlying interests and ideals, and that they are translated into political preferences through individual and group action. Neither the assumption that individuals pursue their preferences instrumentally (shared by many “constructivists”), nor the assumption that the formation of such preferences is exogenous to interstate politics (in any given round of interaction), implies that individual preferences are atomistic. Cultural or sociological arguments that privilege collective social beliefs, either domestic or transnational, as sources of such social preferences, are not excluded. Some metatheoretical discussions between “constructivists” and “rationalists” obscure this potential complementarity between rationalist and cultural explanations, but more recent discussions tend instead to acknowledge it. See, for example, 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; Martha Finnemore and Kathryn Sikkink, “International Norm Dynamics and Political Change,” International Organization, Vol. 52, No. 4 (Autumn 1998), pp. 887–917. Thomas Risse completes the conceptual convergence with his notion of “liberal constructivism.” Thomas Risse-Kappen, “Collective Identity in a Democratic Community: The Case of NATO,” in Peter J. Katzenstein, ed., The Culture of draft -- do not cite or quote without author’s permission moravcsik: chapter 5 5 three strands of liberalism — “ideational,” “commercial,” and “republican” liberalism — described in more detail below. The Nature of the State. The second assumption of liberal theory is that states (or other political institutions) represent some subset of domestic society, whose weighted preferences constitute the underlying goals (“state preferences”) that rational state officials pursue via foreign policy. Representative institutions thereby constitute a critical “transmission belt” by which the preferences and social power of individuals and groups in civil society enter the political 5 realm and are eventually translated into state policy. In the liberal conception of domestic politics, the state is not an actor but a representative institution, constantly subject to capture and recapture, construction and reconstruction, by coalitions of social actors. This pluralist premise assumes neither that all individuals and groups have equal influence on state policy, nor that the structure of state institutions is irrelevant. To the contrary, every government represents some individuals and groups more fully than others — from the idealtype of a single tyrannical individual, a Pol Pot or Josef Stalin, to broad democratic participation — and thus political institutions can be of 6 decisive importance. Variation in the precise nature of representative National Security: Norms and Identity in World Politics (New York: Columbia University Press: 1996), pp. 357–399. 5. This assumption does not privilege the nation-state absolutely. Institutions and practices of political representation result from prior contracts, which can generally be taken for granted in explaining foreign policy. This currently privileges existing nation-states, yet where the primary interests and allegiances of individuals and private groups are transferred to a sub-national or supranational institution sufficiently empowered to represent them effectively — as may be true in, say, some aspects of politics in the European Union — a liberal analysis would naturally shift its focus to these levels. 6. Representation, in the liberal view, is not simply a formal attribute of state institutions, but may include other stable characteristics of the political process, formal or informal, that privilege particular societal interests, including informal ties, the form of individual and group rights, the nature of opportunities for exit, or an inegalitarian distribution of property, risk, information or organizational capabilities that establish socioeconomic monopoly power that can be translated into political influence. See Charles draft -- do not cite or quote without author’s permission 6 progress in international relations theory institutions and practices helps define which groups influence the “national interest.” The Nature of the International System. The third core assumption of liberal theory is that the configuration of state preferences shapes state behavior in the international system. States require a “purpose” — a perceived underlying stake in the matter at hand — in order to provoke conflict, inaugurate cooperation, or take any other significant 7 foreign policy action. The precise nature of the stakes shapes policy. In a pure liberal explanation, the distribution of capabilities, central to realism, and the distribution of information, central to institutionalism, are thus treated as either fixed constraints or as endogenous to state 8 preferences — or both. This is not to assert, of course, that each state simply pursues its ideal policy, oblivious of others. Instead, each state seeks to realize its Edward Lindblom, Politics and Markets: The World’s Political Economic Systems (New York: Basic Books, 1977). 7. Here it is essential to avoid conceptual confusion, given the inconsistency of common usage, by keeping state “preferences” distinct from national “strategies,” “tactics,” and “policies,” that is, the particular transient bargaining positions, negotiating demands, or policy goals that constitute the everyday currency of international politics. States’ preferences, as the concept is employed here, comprise a set of fundamental interests defined across “states of the world.” They are by definition causally independent of and prior to specific interstate strategic interactions, such as external threats, incentives, withholding of information, or other interstate bargaining tactics. The phrase “Country A changed its preferences in response to an action by Country B” would be an abuse of the term as defined here, implying less than consistently rational behavior. By contrast, strategies and tactics — although they are sometimes termed “preferences” in game-theoretical analyses — are policy options defined across intermediate political aims, as when governments declare an “interest” in maintaining the balance of power, containing or appeasing an adversary, or exercising global leadership. Liberal theory focuses on the consequences for state behavior (and state strategies) of shifts in fundamental preferences, not shifts in the strategic circumstances under which states pursue them. This definition of preferences restricts liberal theory, distinguishing it from a loose intuition that “state interests matter.” 8. Liberals also set aside variations in psychology and instrumental beliefs, which lie at the core of epistemic and some constructivist theories. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 7 distinct preferences under constraints imposed by the preferences of other states. In this regard, liberalism is not, in any greater sense than realism or institutionalism, a “domestic” or “second image” theory. All are “systemic” theories, in the strict Waltzian sense, the difference being only that liberals view the distribution of preferences, rather than capabilities (realism) or information (institutionalism), as the systemic characteristic that decisively shapes those strategies. For example, where interstate interaction generates an outcome like trade protection, widely viewed as Pareto sub-optimal, liberals turn first for an explanation to countervailing social preferences and unresolved domestic and transnational distributional conflicts, whereas institutionalists look to the mismanagement of information due to the absence of an appropriate institution, and realists to countervailing considerations arising from the need to manage security competition within the prevailing configuration of political power. In assuming that state preferences vary exogenously, liberal theory thereby sets aside both the (realist) assumption that state preferences must be treated as if they are naturally conflictual, and the (institutionalist) assumption that they should be treated as if they are conditionally convergent. In their place, liberals assume that the critical theoretical link between varying state preferences, on the one hand, and varying interstate behavior, on the other, is provided by the concept of policy interdependence. Policy interdependence can be described as the set of costs and benefits for dominant social groups in foreign societies (the pattern of transnational externalities) that arise when dominant social groups in a given society seek to realize their own preferences internationally. Liberal theory assumes that this pattern of interdependence among state preferences — “asymmetrical 9 interdependence” — imposes a binding constraint on state behavior. Following conventional analyses of international strategic behavior, fundamental patterns of policy interdependence can be divided into at least three broad categories, corresponding to the strategic situation 9. Robert Keohane and Joseph Nye, Power and Interdependence: World Politics in Transition (Boston: Little, Brown, 1977). draft -- do not cite or quote without author’s permission 8 progress in international relations theory 10 (the pattern of policy externalities) that results from unilateral action. First, underlying state preferences may be “zero-sum” or “deadlocked”; that is, an attempt by dominant social groups in one country to realize their preferences through state action necessarily imposes costs (negative externalities) on dominant social groups in other countries. In this case, governments face a bargaining game with few mutual gains and a high potential for interstate tension and conflict. The decisive precondition for costly attempts at coercion, for example, is not a particular configuration of power, as realists assert, or uncertainty, as institutionalists maintain, but configurations of preferences conflictual enough to motivate willingness to accept high cost and risk. In other words, intense conflict presupposes that an “aggressor” or “revisionist” state advance demands to which other 11 states are unwilling to submit. Preferences need not be conflictual, however. A second category arises where preferences are naturally compatible or “harmonious.” Where the externalities of unilateral policies are optimal for others (or insignificant), there are strong incentives for coexistence with low conflict and simple forms of interstate coordination. Still a third category arises where motives are mixed, as when states have an incentive to negotiate institutionalized policy coordination because a shift in expectations, precommitments, or greater information can 10. See Lisa Martin, “Interests, Power, and Multilateralism,” International Organization, Vol. 46, No. 4 (Autumn 1992), pp. 765–792. See also Andreas Hasenclever, Peter Mayer, and Volker Rittberger, Theories of International Regimes (Cambridge, UK: Cambridge University Press, 1997). 11. Revisionist preferences — underlying, socially grounded interests in revising the status quo — are distinct from revisionist “strategies,” that is, a need to alter the status quo to protect enduring interests under new strategic circumstances. Liberals focus on the former, realists and institutionalists on the latter. Hence while realists and liberals might predict security conflict, they expect it to arise under different circumstances. For example, increased military spending in response to the emergence of a large adversary is a capability-induced change in strategy (with preferences fixed) consistent with realism, whereas increased spending initiated by a new ruling elite ideologically committed to territorial aggrandizement is a preference-induced change in strategy consistent with liberalism. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 9 improve the welfare of both parties relative to unilateral policy adjustment. Further differentiation is possible. As Kenneth Oye, Duncan Snidal, Lisa Martin, and others have argued, games such as Coordination, Assurance, Prisoner’s Dilemma, and Suasion have distinctive dynamics, as well as imposing precise costs, benefits, and 12 risks on the parties. Across and within each of the qualitative categories above, the form, substance, and depth of conflict and cooperation vary according to the precise nature and intensity of preferences. By focusing on this structural element of world politics, liberal theory explores a distinct dimension of the international “system.” auxiliary propositions and the protective belt: three variants of liberal ir theory These three “hard core” liberal assumptions, like those of institutionalism, realism, or any other Lakatosian scientific research program, are relatively “thin” or content-free. While they exclude most existing realist, institutionalist, and epistemic theories, as well as many domestic explanations not based on pluralist and rationalist assumptions, they do not, taken by themselves, define a single unambiguous model or set of theories or hypotheses. This ambiguity is, of course, precisely what the Lakatosian understanding of a “paradigm” leads us to expect. Core assumptions define a paradigm, but auxiliary propositions are required to specify it. While the core assumptions of liberal theory may appear almost limitless, the empirically and theoretically viable variants of liberal theory are in fact few and focused. There are three such variants of liberal theory: ideational, commercial, and republican liberalism. At the core of each lies a distinct view concerning the sources of the preferences of powerful domestic social groups, the causal mechanisms whereby they are transformed into state preferences, and the resulting 12. Kenneth Oye, ed., Cooperation under Anarchy (Princeton, N.J.: Princeton University Press, 1983), Martin, “Interests, Power, and Multilateralism”; and Duncan Snidal, “The Game Theory of International Politics,” World Politics, Vol. 38, No. 1 (January 1985), pp. 25–57. draft -- do not cite or quote without author’s permission 10 progress in international relations theory patterns of national preferences in world politics. Let us consider each 13 in turn. Ideational Liberalism: Identity and Legitimate Social Orders. Ideational liberalism views the configuration of domestic social identities and values as a basic determinant of state preferences and thus of interstate conflict and cooperation. Drawing on a liberal tradition of political philosophy dating back to John Stuart Mill, Giuseppe Mazzini, and Woodrow Wilson, it defines “social identity” as the set of preferences shared by individuals concerning the proper scope and nature of public goods provision; this in turn specifies the nature of legitimate domestic order by stipulating which social actors belong to the polity 14 and what is owed to them. Three essential elements of domestic public order often shaped by social identities are geographical borders, political decision-making processes, and socioeconomic regulation. Each can be thought of as a public or “club” good insofar as its provision typically requires that it be legislated universally across a jurisdiction. Recall that for liberals, even the defense of (or, less obvious but no less common, the willing compromise of) territorial integrity, political sovereignty, or national 13. For a more detailed discussion, see Moravcsik, “Taking Preferences Seriously.” 14. The concept of preferences across public goods employed here is similar to but deliberately more precise than Ruggie’s “legitimate social purpose” and Katzenstein’s “collective identity.” John Gerard Ruggie, "International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order," International Organization, Vol. 36, No. 2 (Spring 1982), pp. 195–231; and Katzenstein, The Culture of National Security. Here is a point of intersection between traditional liberal arguments and more recent constructivist works, which tend to stress the social rather than interstate origins of socialization to particular preferences. Risse-Kappen, “Collective Identity.” Liberals take no distinct position on the ultimate origins of social identities, which may stem from historical accretion or be constructed through conscious collective or state action, nor on the question of whether they “ultimately” reflect ideational or material factors — just as long as they are not conceived as endogenous to short-term interstate interaction. The ultimate origin of preferences (“all the way down”) is an issue on which IR theorists, the speculations of constructivists notwithstanding, have little comparative advantage. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 11 security is not an end in itself, but a means of realizing underlying preferences defined by the demands of societal groups (Assumption 1). Social actors provide support to the government in exchange for institutions that accord with their identity-based preferences and are therefore deemed “legitimate” (Assumption 2). Foreign policy will thus be motivated in part by an effort to realize social views about legitimate borders, political institutions, and modes of socioeconomic regulation. The ultimate consequences of identity-based preferences for IR depend on the resulting patterns of policy interdependence — in other words, on the transnational externalities necessarily created by attempts to realize those preferences (Assumption 3). Hence liberal theory predicts that where national conceptions of legitimate borders, political institutions, and socioeconomic equality are compatible, generating positive or negligible externalities, harmony is likely. Where social identities are incompatible and create significant negative externalities, tension and zero-sum conflict is more likely. Where national claims can be made more compatible by reciprocal policy adjustment, cooperation is likely. Parallel predictions about international politics follow from each of the three “ideational liberal” sources of societal preferences: national, political, and socioeconomic identity. The first basic type of social identity concerns the scope of the “nation”: specifically, the legitimate location of national borders and the allocation of citizenship rights. Where borders coincide with underlying patterns of identity, coexistence and even mutual recognition are more likely, but where there are inconsistencies between borders and underlying patterns of identity, greater potential for interstate conflict exists. This novel prediction of liberal theory is broadly confirmed. Over the last century and a half, from midnineteenth century nationalist uprisings to late twentieth-century national liberation struggles, the desire for national autonomy constitutes the most common issue over which wars have been fought and great power intervention has taken place. The Balkan conflicts draft -- do not cite or quote without author’s permission 12 progress in international relations theory preceding World War I and after the Cold War are among the most 15 notorious examples. The second basic type of social identity comprises the commitments of individuals and groups to particular political institutions. Where the realization of legitimate domestic political order in one jurisdiction threatens its realization in others (a situation of negative externalities), conflict is more likely. This differs from realist theory, which accords theoretical weight to domestic regime type only insofar as it influences the distribution of capabilities, and from institutionalist theory, which accords such influence only insofar as it contributes to the certainty of coordination and commitment. Recent trends in Cold War historiography, as well as political science analysis of the United States and the Soviet Union — both based on Soviet documents heretofore inaccessible to Western scholars — lend weight to liberal predictions 16 about the power of ideology, even in a central area of realist concern. The third basic type of social identity is the nature of legitimate socioeconomic regulation and redistribution. Modern liberal theory (in contrast to the laissez faire libertarianism sometimes labeled as quintessentially “liberal”) has long recognized that societal preferences concerning the appropriate nature and level of regulation impose 15. Even those such as James Fearon who stress the absence of domestic credible commitment mechanisms or the interaction between ideational and socioeconomic variables in explaining patterns of nationalist conflicts concede the importance of underlying identities. See David Laitin and James Fearon, “Violence and the Social Construction of Ethnic Identity,” International Organization, Vol. 54, No. 4 (Autumn 2000), pp. 845–877. Remaining dissidents include John Mearsheimer, who bravely asserts that nationalism is a “secondorder force in international politics,” with a “largely … international” cause, namely multipolarity. John Mearsheimer, "Back to the Future: Instability in Europe after the Cold War," International Security, Vol. 15, No. 1 (Summer 1990), pp. 5–56. This disagreement lends itself to empirical resolution: is violent nationalism more of an international problem in Central and Eastern Europe than in Western Europe, as liberalism predicts, or an equal problem in both areas, as realism predicts? The last decade tends to confirm liberal theory. 16. John Lewis Gaddis, We Know Now: Rethinking Cold War History (Oxford: Oxford University Press, 1997); William C. Wohlforth, Witnesses to the End of the Cold War (Baltimore: Johns Hopkins University Press, 1996). draft -- do not cite or quote without author’s permission moravcsik: chapter 5 13 legitimate limits on transnational markets. In a Polanyian vein, John Ruggie reminds us that domestic and international markets are embedded in local social compromises concerning the provision of 17 regulatory public goods. Such compromises underlie variation in national policies toward immigration, social welfare, taxation, religious freedom, families, health and safety, environmental and consumer protection, cultural promotion, and many other public goods that have increasingly been the subjects of international economic negotiations. Recent work has confirmed the novel predictions of this model — in particular, the emergence of so-called “Baptist-bootlegger” coalitions 18 around recent regulatory issues. Commercial Liberalism: Economic Assets and Cross-Border Transactions. Commercial liberal theories seek to explain the individual and collective behavior of states based on the patterns of market incentives facing domestic and transnational economic actors. At its most general, the commercial liberal argument is broadly functionalist: changes in the structure of the domestic and global economy alter the costs and benefits of transnational economic exchange, creating pressure on domestic governments to facilitate or block such exchanges through appropriate foreign economic and security policies. Commercial liberal theory does not predict that economic incentives automatically generate universal free trade and peace — a utopian position often wrongly attributed to it by critics who treat liberalism as an ideology — but instead stresses the interaction between aggregate incentives for certain policies and the obstacles posed by domestic and transnational distributional conflict. Liberal IR theory thereby employs market structure as a variable to explain both openness and closure. The greater the economic benefits for powerful private actors, the greater their incentive, ceteris paribus, to press governments to facilitate such 17. Ruggie, “International Regimes, Transactions, and Change.” 18. David Vogel, Trading Up (Cambridge, Mass.: Harvard University Press, 1995); John Gerard Ruggie, At Home Abroad, Abroad at Home: International Liberalization and Domestic Stability in the New World Economy (Fiesole, Italy: The Robert Schuman Centre at the European University Institute, Jean Monnet Chair Papers, 1995). draft -- do not cite or quote without author’s permission 14 progress in international relations theory transactions; the more costly the adjustment imposed by the proposed economic exchanges, the more opposition is likely to arise. The resulting commercial liberal explanation of “relative gains–seeking” in foreign economic policy is quite distinct from that of realism, which emphasizes security externalities and relative (hegemonic) power, or that of institutionalism, which stresses informational and institutional 19 constraints on optimal interstate collective action. One source of pressure for protection, liberals predict, lies in uncompetitive, monopolistic, or undiversified sectors or factors. These tend to have the most to lose from free trade and thus have a strong incentive to oppose it. Such pressure induces a systematic divergence from laissez faire policies — a tendency recognized by Adam Smith, who complained that “the contrivers of [mercantilism are]…the producers [merchants and manufacturers], whose interest has been so carefully attended to,” and echoed by countless liberals since. Recent research supports the view that free trade is most likely where strong competitiveness, extensive intra-industry trade or trade in intermediate goods, large foreign investments, and low assetspecificity internalize the net benefits of free trade to powerful actors, thus reducing the influence of net losers from liberalization. Novel predictions about cross-sectoral and cross-national variation in support 20 for protection have been confirmed. Commercial liberalism has important implications for security affairs as well. Trade is generally a less costly means of accumulating wealth than war, sanctions, or other coercive means, not least due to 19. This body of literature on “endogenous” foreign economic policy theory is exceptionally deep. For a review and discussion of the relationship between commercial and republican liberal theories, see Robert O. Keohane and Helen V. Milner, Internationalization and Domestic Politics (Cambridge, UK: Cambridge University Press, 1996). 20. Keohane and Milner, Internationalization and Domestic Politics; James Alt and Michael Gilligan, “The Political Economy of Trading States: Factor Specificity, Collective Action Problems, and Domestic Political Institutions,” Journal of Political Philosophy, Vol. 2, No. 2 (1994), pp. 165–192; Helen Milner, “Trading Places: Industries for Free Trade,” World Politics, Vol. 40, No. 3 (April 1988), pp. 350–376. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 15 the minimization of collateral damage. Yet governments sometimes have an incentive to employ coercive means to create and control international markets. To explain this variation, domestic distributional issues and the structure of global markets are critical. Stephen Van Evera argues that the more diversified and complex the existing transnational commercial ties and production structures, the less cost21 effective coercion is likely to be. Cost-effective coercion was most profitable in an era where the main sources of economic profit, such as farmland, slave labor, raw materials, or formal monopoly, could be easily controlled in conquered or colonial economies. Economic development, this line of theory predicts, tends to increase the material stake of social actors in existing investments, thereby reducing their willingness to assume the cost and risk of costly coercion through war or sanctions. Again, substantial empirical evidence supports this 22 view. Republican Liberalism: Representation and Rent-Seeking. Where ideational and commercial liberal theory stress, respectively, particular patterns of underlying societal identities and economic interests, republican liberal theory emphasizes the ways in which domestic institutions and practices aggregate such interests, transforming them into state policy. The key variable in republican liberalism is the nature of domestic political representation, which determines whose social preferences dominate policy. While many liberal arguments are concerned with the “capture” of state institutions by administrators (rulers, armies, or bureaucracies), a parallel argument applies to societal groups that capture the state or simply act independently of 21. Stephen Van Evera, “Primed for Peace: Europe after the Cold War,” International Security, Vol. 15, No. 3 (Winter 1990/1991), pp. 7–57. 22. Realist theory, with its assumptions of a unitary state and fixed preferences, simply presumes that the greater the wealth and power of a state, the less the marginal cost of deploying it. Power is thus reduced to capabilities. liberal theory suggests different predictions, and the competing empirical implications are testable. draft -- do not cite or quote without author’s permission 16 progress in international relations theory 23 it. When institutions of political representation are biased in favor of particular groups, they tend to employ government institutions for their ends alone, systematically passing on cost and risk to others. The simplest prediction of this pluralist view is that policy is biased in favor of the governing coalition or powerful domestic groups, but more sophisticated extensions are numerous. One focuses on rentseeking. When particular groups are able to formulate policy without necessarily providing gains for society as a whole, the result is likely to be inefficient, sub-optimal policies from the aggregate perspective, of which costly international conflict may be an example. If, following the first assumption, most individuals and groups in society, while acquisitive, tend also to be risk-averse (at least where they have something to lose), the more unbiased the range of domestic groups represented, the less likely it is that they will support indiscriminate use of policy instruments, like war, that impose enormous net costs or risks on a broad range of social actors. Aggressive behavior — the voluntary recourse to costly or risky foreign policy — is most likely in undemocratic or inegalitarian polities where privileged individuals can 24 offload its costs. Like other strands of liberal theory, republican liberalism is potentially quite complex, yet nonetheless it generates powerful and parsimonious predictions about international conflict in practice. With respect to extreme but historically common policies such as war, famine, and radical autarky, for example, broad and fair representation 23. Both possibilities are consistent with Assumption Two, whereby the state represents some weighted subset of societal actors; whether that subset comprises those who direct the state, or those who influence those who direct the state, is secondary. 24. This does not, of course, imply that broad domestic representation necessarily always means international political or economic cooperation, for two reasons. First, in specific cases, elite preferences in multiple states may be more convergent than popular ones. Second, the extent of bias in representation, not democracy per se, is the theoretically critical point. There exist predictable conditions under which specific governing elites may have an incentive to represent long-term social preferences in a way that is less biased draft -- do not cite or quote without author’s permission moravcsik: chapter 5 17 appears to inhibit international conflict. Republican liberal theory thus helps to explain phenomena as diverse as the “democratic peace,” modern imperialism, and international trade and monetary cooperation. Given the plausibility of the assumption that major war imposes net costs on society as a whole, it is hardly surprising that the most prominent republican liberal argument concerns the “democratic peace,” which one scholar has termed “as close as anything we have to an empirical law in international relations” — one that applies to tribal 25 societies as well as modern states. This line of argument, as James Lee Ray notes in Chapter 6 of this volume, has generated many novel predictions. Often overlooked is the theoretical obverse of “democratic peace” theory: a republican liberal theory of war that stresses abnormally riskacceptant leaders and rent-seeking coalitions. There is substantial historical evidence that the aggressors who have provoked modern great power wars tend either to be extremely risk-acceptant individuals, or individuals well able to insulate themselves from the costs of war, or both. Jack Snyder, for example, has deepened Hobson’s classic rent-seeking analysis of imperialism — in which the military, uncompetitive foreign investors and traders, jingoistic political elites, and others who benefit from imperialism are particularly well-placed to influence policy — by linking unrepresentative and extreme 26 outcomes to log-rolling coalitions. Consistent with this analysis, the than would broad public and elite opinion. This explains the existence of insulated trade policy-making institutions such as “fast track” provisions. 25. Jack S. Levy, “The Causes of War,” in Philip E. Tetlock, et al., eds., Behavior, Society and Nuclear War, Vol. 1 (New York: Oxford University Press, 1989), p. 270. 26. Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.: Cornell University Press, 1991). It is indicative of the conceptual confusion that besets metatheoretical labeling in international relations that this argument has been advanced by those often termed “neoclassical realists,” including Stephen Van Evera, Stephen Walt, Randall Schweller, and Jack Snyder. For an early critique along these lines by a scholar who subsequently fell into the same trap, see Fareed Zakaria, “Realism and Domestic Politics,” International Security, Vol. 17, No. 1 (Summer 1992), pp. 177–198. For a comprehensive critique of the mislabeling and incoherence of attempts to draft -- do not cite or quote without author’s permission 18 progress in international relations theory highly unrepresentative consequences of partial democratization, combined with the disruption of rapid industrialization and incomplete political socialization, suggest that democratizing states, if 27 subject to these influences, may be particularly war-prone. While such findings challenge what is sometimes referred to as liberal ideology, they are predicted by liberal theory. Precise analogs to the “democratic peace” exist in the area of political economy as well. As we saw in the preceding section, perhaps the most widespread explanation for the persistence of illiberal commercial policies, such as protection, monetary instability, and sectoral subsidization that may manifestly undermine the general welfare of the population, is pressure from powerful domestic groups. The power of such groups may ultimately result from the inherent power of certain business interests in civil society, as argued by pure commercial liberal theory, but might also reflect biases within representative institutions, as republican liberals theory suggests. Where the latter sort of biases exist — and it is seen in most contemporary representative institutions — rent-seeking groups are likely to gain protection through tariffs, subsidies, favorable regulation, or competitive devaluation. Where policy makers are insulated from such pressures, which may involve less democratic but more representative institutions, or where free trade interests dominate policy, open policies are more viable. Recent studies of commercial 28 policy have evolved in this direction. specify a realist paradigm, see Jeffrey Legro and Andrew Moravcsik, “Is Anybody Still a Realist?” International Security, Vol. 24, No. 2 (Fall 1999), pp. 5– 55. 27. Jack Snyder and Edward Mansfield, “Democratization and the Danger of War,” International Security, Vol. 20, No. 1 (Summer 1995), pp. 5–38; Jack Snyder, From Voting to Violence: Democratization and Nationalist Conflict (New York: Norton, 2000). 28. From this insight follows an entire line of literature about the role of national executives in foreign economic policy. For example, see Stephan Haggard, “The Institutional Foundations of Hegemony: Explaining the Reciprocal Trade Agreements Act of 1934,” International Organization, Vol. 42, No. 1 (Winter 1988), pp. 91–120. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 19 Is Liberal IR Theory a Progressive Research Program? Assessing whether any given scientific research program — such as the liberal program set forth above — is progressive demands that we ask whether it generates “excess content” in the form of “novel” predicted facts. Chapter 2 provides a helpful discussion of four possible Lakatosian criteria for judging the novelty of facts, of which I will consider three: “strict temporal novelty” (Lakatos1) and “the heuristic definition of novelty” (Lakatos3), considered in the first section below, and “background theory novelty” (Lakatos4), considered in the second 29 section below. The first two I assess with reference to the intellectual history of the liberal scientific research program. The latter I assess with reference to the current research findings of liberal theory and its competitors. No matter which criterion is used, the conclusion is unambiguous, namely, that liberal IR theory is progressive in a Lakatosian sense. In drawing this conclusion, we learn something about the practical utility of Lakatosian criteria. In contrast to Elman and Elman, I find that the most compelling criterion is “background theory novelty.” This is because in practice it proves quite difficult — contrary to what Lakatos and the Elmans both assume — to subsume new empirical results through auxiliary assumptions within the constraints of fixed hard-core assumptions. Recent modifications in realism, for example, which have adopted the propositions and assumptions of liberal theories to explain anomalies, demonstrate the difficulty of modifying realism itself. Nonetheless, all three Lakatosian criteria offer some unique insight, and their joint application permits us to draw a consistent and convincing conclusion that the liberal scientific research program in IR is progressive. 29. I set aside one of these criteria, namely “new interpretation novelty” (Lakatos2) on the ground, reported by the Elmans, that it has little support in the secondary literature. draft -- do not cite or quote without author’s permission 20 progress in international relations theory temporal and heuristic novelty: the intellectual history of liberal ir theory The intellectual origins of a scientific research program are directly relevant to judging consistency with two Lakatosian criteria: temporal and heuristic novelty. Both embody “the simple rule that one can’t use the same fact twice: once in the construction of a theory and then again 30 in its support,” as John Worrall puts it. “Strict temporal novelty” (Lakatos1) asks whether the scientific research program successfully predicts facts unknown, “improbable, or even impossible in the light of previous knowledge,” while “the heuristic definition of novelty” (Lakatos3) asks whether the scientific research program successfully predicts facts that did not “play some heuristic role in that theory’s 31 construction.” Elman and Elman voice suspicion about these criteria. Strict temporal novelty seems too restrictive, because it treats as “not novel” any “fact that is known to anyone at any time before the theoretical modification” — a criterion they believe is so strict as to “exclude almost any social behavior from ever being counted as a novel fact.” While unlikely to code degenerating scientific research programs as progressive, it may overlook some progressive scientific research programs. Elman and Elman side with the second, heuristic novelty, but note that it is difficult to employ, since “the determination of novelty depends on private, inaccessible biographical knowledge 32 about the scientist.” I submit that, at least at the broadest level, the liberal scientific research program meets the strict temporal and heuristic criteria — and does so in a way that belies some of the Elmans’ methodological and pragmatic misgivings about them. The most fundamental hypotheses of modern liberal IR theory were initially advanced by political philosophers and publicists of the eighteenth and nineteenth centuries, who wrote before the independent variables underlying 30. Quoted in Elman and Elman, Chapter 2 in this volume, p. [20 of 69]. 31. The first implies the second, of course. 32. Elman and Elman, Chapter 2, p. [same as n. 30 or plus 1]. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 21 liberal theory (democratization, industrialization, and secular belief systems) were widespread enough (if they existed at all) to generate any consistent record. The critical insights of liberal IR theory in the eighteenth and nineteenth centuries can be found in the writings of the three most prominent philosophers and publicists in this tradition: Immanuel Kant, Adam Smith, and John Stuart Mill. Each was a visionary who predicted the implications for international relations of a social trend that had only just begun when they wrote. To be sure, phenomena of the obverse were visible — wars waged by autocrats, pre-industrial mercantilism for control over fixed resources, and religious fundamentalism. Yet liberal philosophers advanced predictions about the potential for change on the basis of only a small spectrum of historical or geographical variation. Temporal and heuristic novelty are evident in each variant of liberal theory. Kant advanced a theory about the pacific implications of republican governance for foreign policy at a time when there were no 33 more than a handful of republics in the world. Numerous subsequent thinkers, from Woodrow Wilson to George Kennan to Francis Fukuyama, further developed this view. Adam Smith advanced a firmly grounded theory about socioeconomic and regulatory pressures for free trade and protectionism in a world still governed by great power mercantilism. Subsequent thinkers in this vein included Richard Cobden and John Maynard Keynes. John Stuart Mill advanced systematic conjectures about the implications of collective cultural phenomena — national identity, education, and cosmopolitan values — in an era in which these were only beginning to emerge as a dominant locus of political organization. Subsequent liberal thinkers in this vein included Giuseppe Mazzini and Wilson. 33. Kant is often misunderstood in this regard as a global federalist. Yet his movement from the world republic envisioned in “Theory and Practice” of 1793 to the structured relations among republics envisioned in “Toward Perpetual Peace” in 1795 is unambiguous. In the latter, Kant’s definitive statement, the internal sovereignty of nations is a constitutive principle of global order. James Bohman and Matthias Lutz-Bachmann, “Introduction,” in Bohman and Lutz-Bachmann, eds., Perpetual Peace: Essays on Kant’s Cosmopolitan Ideal (Cambridge, Mass.: MIT Press, 1997), pp. 6–7. draft -- do not cite or quote without author’s permission 22 progress in international relations theory An assessment of temporal and heuristic novelty requires that we investigate origins, but — it might be objected — perhaps Kant, Smith, and Mill should not be treated as early social scientists, but instead as idealistic visionaries whose predictions just happen in retrospect to have been correct? Since Machiavelli advanced his celebrated distinction between “the effective truth of things” and the “imaginary republics and monarchies that have never been seen or have been known to exist,” this has been the attitude of realists. Liberal arguments have been ridiculed as based on idealized notions of enlightened, benevolent individuals inhabiting a state of nature — notions drawn from very limited experience of world politics, if not pure philosophical utopianism. Liberals assume the existence of a perfect harmony of interests, between individuals as between nations, which the spread of education and cosmopolitan values will progressively make known to all. Thus Martin Wight calls the Kantian tradition a “revolutionary” and “utopian” project; Michael Howard criticizes liberals for their naïveté in demanding a Gandhian sense of individual self-sacrifice; and Hans Morgenthau contrasts liberal views with realism’s “theoretical concern with human nature as it actually is, and with the historical processes as they actually take place.” Arnold Wolfers and Laurence Martin treat it as a narrow doctrine bred of the insularity and unique domestic political legacy of the Anglo-American tradition. Even social scientists sympathetic to the liberal scientific research program have been quick to grant that liberal theories are more philosophy than social science; they cannot meet the standards of rigor set by realism — a remarkable claim in itself! — precisely because their underlying philosophical assertion of the moral worth and independence of the individual introduces, Robert Keohane argues, an 34 ineluctable source of “indeterminacy.” I submit, however, that liberal IR theory, as developed by such philosophers and essayists as Kant, Smith, and Mill, was grounded not in utopian philosophy in what we would term today a distinctive 34. Robert O. Keohane, “International Liberalism Reconsidered,” in John Dunn, ed., The Economic Limits to Modern Politics (Cambridge, UK: Cambridge University Press, 1990), pp. 192–194. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 23 social-scientific analysis of world politics — and thus it should count in favor of temporal novelty. Eighteenth and nineteenth-century liberals did not offer simply a particular ideal of global harmony, but sought to account for variation in international cooperation and conflict. As explained above, distinctive liberal theories of peace beget corresponding liberal theories of war, liberal theories of free trade and cooperation beget liberal theories of protectionism and mercantilism, and liberal theories of ideological conflict beget liberal theories of ideologically-induced consensus. By the time of Smith in Britain, Kant in Germany, and Benjamin Constant, if not Montesquieu, in France, such utopian notions — even if they occasionally reappeared later — had been definitively supplanted by efforts to ground liberal political philosophy in sociological theory. At the risk of gross oversimplification, it could be said that the essential move of modern liberal political philosophy was to place a richly varied society of individuals making choices at the basis of theorizing about political order. Thus the normative claims of subsequent liberal philosophers generally rest on a set of sophisticated claims about the variety of possible relationships between the state and society, of which their ideal prescriptions are simply a limiting case. Modern attempts to assert a normative liberal position must begin by accepting what John Hall has termed a sociological “wager on reason,” namely, the assumption that civil society precedes the state and that certain conditions will impel rational individuals in civil society to act 35 politically in predictable ways. This was as true for classical philosophers as for modern theorists. It is doubtful that even early liberals subscribed to such idealistic views as that their doctrines could be deduced from a mythical state of nature, that societies would harmoniously tend toward progress, or that human beings, once persuaded by liberal arguments, could be trusted to regenerate themselves morally. Sheldon Wolin has observed that: 35. John A. Hall, Liberalism: Politics, Ideology and the Market (Chapel Hill: University of North Carolina Press, 1987). draft -- do not cite or quote without author’s permission 24 progress in international relations theory Liberalism has repeatedly been characterized as “optimistic” to the point of naïveté; arrogant in its conviction that human reason ought to stand as the sole authority for knowledge and action; bewitched by a vision of history as an escalator endlessly moving upwards towards greater progress; and blasphemous in endowing the human mind and will with a godlike power of refashioning man and society in entirety. For the most part, these criticisms have little or no support in the writings of the 36 liberals. Kant constructed a plan for movement toward world peace that he 37 asserted would be effective “even in a world of devils.” Of Benjamin Constant, Stephen Holmes observed that: Once again following Montesquieu and other eighteenth-century (particularly Scottish) examples, [Constant] deliberately supplanted the contract myth with the theory of social change. The liberal state is desirable not because it mirrors human nature or respects eternal human rights, but because it is the political arrangement most adequate to solving the problems of European society in its current state of economic, scientific 38 and moral development. [Similarly,] Smith made the intellectual journey from a notion that commercial activity could tame or at least successfully oppose the 36. Sheldon S. Wolin, Politics and Vision: Continuity and Innovation in Western Political Thought (Boston: Little, Brown, 1960), p. 305, also pp. 286–294, 305–309. See also Don Herzog, Without Foundations: Justification in Political Theory (Ithaca: Cornell University Press, 1985), pp. 204–207; John Dunn, Rethinking Modern Political Theory: Essays 1979–83 (Cambridge, Mass.: Cambridge University Press, 1985), pp. 154–163; John Gray, Liberalism (Minneapolis: University of Minnesota Press, 1986), pp. 45–56. 37. Immanuel Kant, “Perpetual Peace: A Philosophical Sketch,” in Hans Reiss, ed., Kant’s Political Writings, trans. H.B. Nisbet (Cambridge, UK: Cambridge University Press, 1970), pp. 93–130. 38. Stephen Holmes, Benjamin Constant and the Making of Modern Liberalism (New Haven: Yale University Press, 1984), p. 32. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 25 more militant passions, to the notion of a self-regulating society largely 39 independent of the nature of individual norms. The remarkable prescience of early liberal IR theorists, and the resulting ability of liberal theory to meet the criteria of temporal novelty, stands in striking contrast to its realist and institutionalist counterparts. The realist scientific research program emerged from the inductive analyses of Thucydides, Niccolo Machiavelli, Thomas Hobbes, Friedrich Meinecke, and Hans Morgenthau. Here there is little temporal novelty. Each observed, in his era, characteristic realist pathologies of anarchy — an overriding concern for security, the formation of balances of power, the dynamics of deterrence and preventive war — then developed a theory to explain them. (To be sure, much subsequent history confirmed the balance-of-power theory, yet new cases of balancing were far less novel than the emergence and spread of modern republican government.) Similarly, it might be argued that the modern institutionalist scientific research program, which emerged in the 1970s, was developed to explain the success of post–World War II international organizations, which appeared anomalous from a realist perspective, as Robert Keohane and Lisa Martin note in Chapter 3. This is not to say that these research programs have not explained some temporally novel facts, only that new facts and major developments in world politics appear to have preceded major theoretical innovations in realist and institutionalist scientific research programs to a greater extent than was the case with the liberal scientific research program. While this speaks well for the liberal scientific research program, I remain unconvinced that temporal or heuristic novelty is an essential criterion for judging scientific research programs. Whether the behavioral regularities that a theory can convincingly explain are known before or after the development of the theory is an entirely secondary consideration. It seems to me that the fact that realist theory was distilled from widespread observation of world politics does not 39. Albert O. Hirschman, The Passions and the Interests: Political Arguments for Capitalism before its Triumph (Princeton, N.J.: Princeton University Press, 1977), pp. 100–112, 120. draft -- do not cite or quote without author’s permission 26 progress in international relations theory make the evidence of its importance, particularly before the modern period, any less compelling. Despite the Elmans’ skepticism, it is in fact far from trivial to develop a coherent theory to unify an extensive set of facts, even if they are fully understood in advance. Thus whereas I do maintain that liberal theory meets many criteria for theoretical fruitfulness and, accordingly, is unjustly neglected in current theoretical debates, I do not reach this conclusion primarily because liberal theory was derived deductively rather than inductively. More important than novelty, in my view, is performance — confirmed predictions minus confirmed anomalies — as compared to competing scientific research programs. If a particular theory provides a better fit to a specified pattern of facts without generating a greater number of offsetting anomalies, it should considered more plausible. I turn now to a Lakatosian criterion — “background theory novelty” (Lakatos4) — more consistent with this view. background theory novelty: liberalism and its competitors in current research “Background theory novelty” (Lakatos4), an interpretation of Lakatos proposed by Musgrave, instructs us to assess the excess content of novel facts explained by research programs over time by asking whether the liberal scientific research program “predicts something 40 which is not also predicted by its background theory.” This criterion I find more powerful than the alternatives, and according to it, liberal theory is an even more progressive program. By social-scientific standards, as we have seen, there exists remarkably strong support for key liberal predictions across the board, such as those concerning the democratic peace in the republican liberal tradition, endogenous international trade and monetary policy in the commercial liberal tradition, and the role of societal preferences across public goods in a range of phenomena from nationalist conflict to regulatory harmonization in the ideational liberal tradition. We have 40. Elman and Elman, Chapter 2 in this volume, p. [18], quoting Alan Musgrave. Also see Imre Lakatos and Alan Musgrave, eds., Criticism and the Growth of Knowledge (Cambridge, UK: Cambridge University Press, 1970). draft -- do not cite or quote without author’s permission moravcsik: chapter 5 27 seen that liberal theory has generated fruitful new lines of theory in security studies, international organization, and international political economy. Certainly the liberal scientific research program advances a wide range of distinct confirmed predictions not successfully predicted — or in any way derivable from — realist or institutionalist theory. Perhaps most important, given the Lakatosian tendency to view inter-paradigmatic conflict as a “three-cornered fight” between two theories and the data, is another point. Recent empirical and theoretical debates demonstrate that non-liberal scientific research programs have a very limited capacity to generate plausible — internally coherent and empirically confirmed — explanations for certain important regularities predicted by liberal theory. Contrary to what Elman and Elman suggest in Chapter 2 about “background theory novelty,” it seems in fact quite difficult to generate plausible auxiliary explanations for many phenomena uncovered by competing IR scientific research programs. A comparison of specific areas in which realist and liberal theories have been applied not only generates numerous anomalies where realists have tried and failed to generate satisfactory explanations for confirmed liberal predictions, but also numerous cases in which realists, even in the absence of a direct liberal challenge, have advanced formulations of realism that overtly degenerate toward liberalism, even when we judge “degeneration” according to core definitions that realists themselves have advanced. The next section summarizes the more detailed evidence for this charge. Realist Anomalies, Novel Facts, and Liberal Theory. We turn first to areas where realism has failed to propose any detailed explanation for salient phenomena that are well-explained within liberal theory, or where realist explanations for confirmed liberal predictions have, on closer inspection, proved unconvincing. Consider some examples. Realism provides no explanation for differences in the substantive nature of formally similar orders. What accounts, for example, for differences between Anglo-American, Nazi, and Soviet plans for the post–World War II world? What accounts for the substantial differences between the compromise of “embedded liberalism” underlying Bretton Woods and arrangements under the Gold Standard? divergences between economic cooperation under the draft -- do not cite or quote without author’s permission 28 progress in international relations theory European Community and ComEcon? the greater protectionism of agricultural policy of the Organization for Economic Cooperation and Development (OECD), compared to its industrial trade policy? These are realist anomalies. Yet, as John Ruggie and others have shown, there are plausible, parsimonious, and empirically confirmed liberal explanations grounded in the variation in national socioeconomic 41 preferences for each of these novel phenomena. Another example, the “democratic peace” proposition, remains a robust and significant anomaly for realism. Attempts by Joanne Gowa, David Spiro, Randall Schweller, and others to debunk the “democratic peace” hypothesis advanced by Michael Doyle, Bruce Russett, and others have not succeeded in reversing the strong presumption in its 42 favor. More broadly, realists provide no explanation for the consistent tendency of perceived threats to vary independently of the relative power of the threatener. Why do states tend to provoke war with large states and, more often than not, lose the subsequent war? What explains why U.S. concern about a few North Korean, Iraqi, or Chinese nuclear weapons is greater than that for the larger arsenals held by Great Britain, Israel, and France? The democratic peace hypothesis, as well as theories of ethnic attachment, offer plausible explanations for what are striking realist anomalies. A third example is the distinct nature of politics among advanced industrial democracies, grounded in reliable expectations of peaceful change, domestic rule of law, stable international institutions, and 41. Ruggie, At Home Abroad; John Gerard Ruggie, “Embedded Liberalism Revisited: Institutions and Progress in International Economic Relations,” in Emanuel Adler and Beverly Crawford, eds., Progress in Postwar International Relations (New York: Columbia University Press, 1991), pp. 201–234. 42. Joanne S. Gowa, Ballots and Bullets: The Elusive Democratic Peace (Princeton, N.J.: Princeton University Press, 1999). The only consistent result of such studies is that under certain extreme specifications — limited periods of time and limited numbers of countries — the relationship between democracy and peace can be reduced to statistical insignificance. Others assert that the democratic peace may not hold in the future. No critique consistently reverses the direction of the causal effect (i.e., democracies go to war more) or proposes a consistently powerful opposing theory to explain the patterns we observe. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 29 intensive societal interaction. This is the condition Karl Deutsch terms a “pluralistic security community” and Keohane and Nye term 43 “complex interdependence.” Whereas realists (and, as discussed below, Constructivists) offer no general explanation for the emergence of this distinctive mode of international politics, liberal theory argues that the emergence of a large and expanding bloc of democratic, interdependent, nationally satisfied states has been a precondition for such politics. Consider, for example, Western Europe since 1989. Unlike realism, liberal theory predicts and explains the absence of competitive alliance formation among West European powers. The lack of serious conflict in the rest of Europe over Yugoslavia — avoiding the “World War I scenario” — reflects in large part a shared perception that the geopolitical stakes among democratic governments are low. liberalism similarly makes more sense of the sudden reversal of East-West relations, a shift made possible by the widespread view among Russian officials (so interview data reveal) that Germany is ethnically 44 satisfied, politically democratic, and commercially inclined. These facts are novel by both the temporal and the background criteria. By contrast, John Mearsheimer’s realist alternative to democratic peace theory’s predictions of peace in post–Cold War has yet to find confirmation. Mearsheimer offers a heroic argument that external threats under multipolarity have triggered nationalist reactions in Yugoslavia. Yet this auxiliary claim fails to explain perhaps the most salient fact about post–Cold War European politics, namely, the disparity between East and West. We observe total peace among the established democracies of Western Europe, yet conflict (and threat of conflict), if sporadic, among the transitional democracies and nondemocracies of Central and Eastern Europe. In an effort to account for 43. Karl Wolfgang Deutsch, et al., Political Community and the North Atlantic Area: International Organization in the Light of Historical Experience (Princeton, N.J.: Princeton University Press, 1957); and Keohane and Nye, Power and Interdependence. 44. Interview data reported in personal communication from Professor Celeste Wallander, Harvard University. draft -- do not cite or quote without author’s permission 30 progress in international relations theory this, even Mearsheimer is led to invoke the autonomous importance of 45 underlying patterns of national identities in the former Yugoslavia. This evolution, which emerged after predictions were on the table in 1990, confirms the novelty of the liberal theory. Similarly, under the rubrics of hegemonic stability theory and relative gains–seeking, Stephen Krasner, Joseph Grieco, David Lake, and others have posed realist challenges to liberal theories of economic integration and commercial liberalization advanced by Helen Milner, Jeffry Frieden, Ronald Rogowski, John Ruggie, myself, and many others within the now massive literature on endogenous tariff theory. Yet a series of disconfirmations have all but removed hegemonic stability theory from the academic scene. At best, it does not appear 46 robust beyond a single case, that of U.S. policy after World War II. Grieco has offered no convincing answer to criticisms that relative gains–seeking fails to demonstrate a link between security and trade, 47 as well as omitting direct tests with liberal hypotheses. The most that can be said empirically for this line of recent realist work is that some scholars have succeeded in demonstrating the existence of a modest 48 correlation between alliances and trade. One final example: liberal theory offers a plausible explanation for long-term historical change in the international system, whereas the static quality of both realist and institutionalist theory — their lack of an explanation for fundamental long-term change in the nature of international politics — is a recognized weakness. Global economic 45. Mearsheimer, “Back to the Future.” 46. For the best effort at a revival, see David A. Lake, “Leadership, Hegemony, and the International Economy: Naked Emperor or Tattered Monarch with Potential?" International Studies Quarterly, Vol. 37, No. 4 (December 1993), pp. 459–489. 47. For criticisms of Grieco, see Robert Powell, “Anarchy in International Relations Theory: The Neorealist-Neoliberal Debate,” International Organization, Vol. 48, No. 2 (Spring 1994), pp. 313–344. 48. Joanne S. Gowa, “Bipolarity, Multipolarity, and Free Trade,” American Political Science Review, Vol. 83, No. 4 (December 1989), pp. 1245–1256; and Edward Mansfield, “The Concentration of Capabilities and International Trade,” International Organization, Vol. 46, No. 3 (Summer 1992), pp. 731–764. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 31 development over the past 500 years has been closely related to greater per-capita wealth, democratization, education systems that reinforce new collective identities, and greater incentives for transborder economic transactions. Realist theory accords these changes no theoretical importance. Theorists such as Kenneth Waltz, Robert Gilpin, and Paul Kennedy limit realism to the analysis of unchanging patterns of state behavior or the cyclical rise and decline of great 49 powers and their success in making war. Liberal theory, by contrast, forges a direct causal link between economic, political, and social change and state behavior in world politics. Hence, over the modern period, the principles of international order have been decreasingly linked to dynastic legitimacy and increasingly tied to factors drawn directly from the three variants of liberal theory: national selfdetermination and social citizenship, the increasing complexity of 50 economic integration, and democratic governance. This is a novel fact — so much so that Michael Howard, a leading realist, was forced to reverse course and concede the limitations of realism in the second edition his classic critique of liberal IR theory, War and the Liberal 51 Conscience. These examples, each of them involving a significant issue of modern world politics, suggest that liberal theory has, at least in some matters, broader scope than realist theory, and that the latter is accumulating anomalies that are especially visible from the liberal perspective. Realist Degeneration in the Direction of Liberal Theory. Even more striking than the ability of liberal theory to explain realist anomalies is the increasing tendency of self-styled realists to explain core security 49. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: AddisonWesley, 1979); Paul Kennedy, The Rise and Fall of the Great Powers: Economic Change and Military Conflict from 1500 to 2000 (New York: Random House, 1987); Robert Gilpin, War and Change in International Politics (Princeton, N.J.: Princeton University Press, 1981). 50. Kalevi J. Holsti, Peace and War: Armed Conflicts and International Order, 1648– 1989 (Cambridge, UK: Cambridge University Press, 1991). 51. Michael Howard, War and the Liberal Conscience, 2nd ed. (New Brunswick, N.J.: Rutgers University Press, 1991). draft -- do not cite or quote without author’s permission 32 progress in international relations theory relations — patterns of war, alliance formation, arms control, and imperialism — by invoking core assumptions and causal processes drawn from liberal and institutionalist theory, including exogenous variation in societal preferences and transnational information flows 52 through international institutions. A closer examination of this tendency demonstrates not only the power of liberal IR theory, but also the difficulty of explaining anomalies through viable auxiliary assumptions, while retaining the integrity of hard-core assumptions. This confirms the utility of Musgrave’s conception of “background theory novelty” (Lakatos4), contra Lakatos and the Elmans, who assume it is trivially easy to explain away anomalies in this way. Jeffrey Legro and I have recently demonstrated that leading selfdeclared realists — among them Stephen Van Evera, Jack Snyder, Stephen Walt, Charles Glazer, Fareed Zakaria, Randall Schweller, Gideon Rose, William Wohlforth, and Joseph Grieco — have advanced as “realist” theories that water down the hard core of realism to generic assumptions of rationality and anarchy shared by nearly all major IR 53 theories. These self-styled “neoclassical” and “defensive” realists, who dominate modern realist theory, seek to explain the tendency of states to make war and alliance decisions. Some such efforts, to be sure, explain anomalies in a way consistent with a realist “hard core” focused on the resolution of interstate conflict over scarce resources through the application of relative power capabilities — while holding preferences and perceptions constant. Examples of such “progressive” realist shifts include “auxiliary hypotheses” that stress the role of geographical proximity and of offensive or defensive military 54 technology. Yet most “neoclassical” or “defensive” realists emphasize factors derived from liberal, institutionalist, or sometimes even constructivist 52. For a detailed summary, see Legro and Moravcsik, “Is Anybody Still a Realist?” 53. Legro and Moravcsik, “Is Anybody Still a Realist?” 54. This applies to the first three of the four elements of Stephen Walt’s revised formulation of realism. See Stephen Walt, The Origins of Alliances (Ithaca, N.Y.: Cornell University Press, 1987). draft -- do not cite or quote without author’s permission moravcsik: chapter 5 33 core assumptions. Their explanations invoke variations in the transaction-cost-reducing influence of international institutions, misperceptions and belief systems, and (most relevant for an assessment of liberal theory) state preferences — each traditionally seen as fundamentally opposed to realism. The influence of these factors often reverses the empirical predictions of traditional realists. Judged by its core assumptions, rather than its label, recent “realist” literature has done much to strengthen the liberal, institutionalist, and epistemic paradigms. Legro and I argue that realists have failed to advance a set of distinctive hard core assumptions that subsume these “realist” writings without expanding the “realist” category to include nearly all rationalist theories and causal processes in world politics. This is necessarily so: once realists permit preferences and perception, as well as power, to vary exogenously and influence state behavior, they can invoke as “realist” almost any rational decision-making process. Legro and I argue that the broadest “hard core” that could plausibly be thought of as distinct to realism is one that assumes rational unitary states, fixed conflictual preferences (the element that distinguishes realism from liberalism), and strategic interaction based on relative control over material resources (the element that distinguishes realism 55 from institutionalism). This would, it appears, exclude “neoclassical” realist theories, which would be more properly (i.e., in accord with their core assumptions) categorized as liberal, institutionalist, or epistemic/constructivist. Most realists who seek to set forth core realist assumptions (surprisingly few do so explicitly) propose instead what Legro and I 55. Returning to Elman and Elman’s definition of neorealism, this analysis implies that the seven assumptions they set forth — states are rational, egotistical, and strategic, possess limited resources, seek security, and act in anarchy — are insufficient even to define neorealism. These assumptions are, at least at the level of generality are stated, entirely consistent with the “democratic peace,” theories of interdependence and war, the importance of “security regimes,” and many other ostensibly non-realist bodies of theory. Again, either neorealism becomes another word for all rationalist IR theory or it is underspecified. draft -- do not cite or quote without author’s permission 34 progress in international relations theory term a “minimal realist” definition. In this view, realists need only assume that states are rational, unitary, self-interested actors, act in an 56 anarchic setting, and are concerned about security. As a “hard core,” this is manifestly inadequate. The only state behaviors it excludes are outright self-abnegating altruism and delegation of power to a world 57 state. Hardly any IR theorist today — certainly no thoughtful regime theorist or liberal theorist — maintains that states are altruistic, irrational, unstrategic, inward-looking, omnipotent, or oblivious of security matters. Nor do many maintain that the international system, even if influenced by international regimes, is anything but an 58 anarchy. Finally, while some liberal theories stress national goals other than security, most liberals see states as placing a preeminent value on 56. For examples, see Legro and Moravcsik, “Is Anybody Still a Realist?” Schweller’s contribution to this volume (Chapter 9) is an example of a realist analysis that crisply and insightfully sets forth the problem, namely that realism cannot progress without sacrificing its essential assumptions of constant underlying conflict of preferences resolved by the applications of material power resources. I find Schweller’s diagnosis of the crisis in realist thought, and what needs to be done about it, clearer than that of any other scholar in security studies. Yet he unaccountably concludes that the move from neorealism to theories that do not hold these assumptions about conflict and power — namely “neoclassical realist” theories — constitute a progressive shift. Schweller never resolves the obvious tensions by presenting a Lakatosian “hard core,” or some other measure of theoretical coherence, that subsumes both neorealism and neoclassical realism, nor explains why neoclassical realists should not be viewed, as Legro and I argue, as grafting on non-realist arguments. At the very least, Schweller’s account leaves us unclear, from a Lakatosian perspective, what any of the theoretical labels mean. 57. This is made very clear in Charles L. Glaser, “Realists as Optimists: Cooperation as Self-Help,” International Security, Vol. 19, No. 3 (Winter 1995/1996), pp. 50–90. 58. It is true that the liberal hard core assumes that contestation among subnational actors influences national preferences, but this is employed only to explain variation in preferences. Few liberals deny that states are the major instrumental actors in world politics. Even those who stress the role of nongovernmental organizations increasingly focus on their ability to influence states to act in a particular instrumental manner. See, e.g., Margaret Keck and Kathryn Sikkink, Activists beyond Borders: Transnational Advocacy Networks in International Politics (Ithaca, N.Y.: Cornell University Press, 1999). draft -- do not cite or quote without author’s permission moravcsik: chapter 5 35 security. The democratic peace predicted by liberals, for example, and the formation of arms control regimes predicted under certain conditions by institutionalists, are held together precisely by the high 59 value placed by participating governments on security. Despite their commitment to Lakatos’s concept of a “hard core,” Elman and Elman seem to perpetuate this degenerative tendency, as is reflected in the Elmans’ own seven-point proposal for a realist hard core (“illustrative specification of the neorealist research program”). They suggest that the neorealist “hard core” might consist of seven assumptions, summarized as that egotistical, rational, strategic states employ limited resources to assure security in an anarchic international system. They conclude with the assertion that “work by structural realists” — by which they clearly mean to encompass far more than Waltzian neorealism — ”would share these central and unchanging 60 elements.” Yet, having accorded this definition Lakatosian status, Elman and Elman immediately undermine it. They point out that Walt, Van Evera, Snyder, Zakaria, Schweller, Grieco, and Glazer do not in fact accept all of these assumptions. Elman and Elman concede that these theorists explain outcomes by invoking exogenous variation in national preferences — what Elman and Elman somewhat misleadingly term “internal factors” — and international institutions. This the Elmans term a paradigm shift from “neorealism” to “neoclassical” (or “neotraditional”) realism. Yet they never answer the essential Lakatosian question, left open by their own apparent abandonment of their seven-part definition, namely: to what core realist propositions do neoclassical realists adhere? Does “neoclassical” 59. Indeed, as Schweller and Van Evera have argued, a realist world seems to assume the existence of revisionist aggressors, that is, states that seek far more than security. See Randall L. Schweller, Deadly Imbalances: Tripolarity and Hitler’s Strategy of World Conquest (New York: Columbia University Press, 1998); and Stephen Van Evera, Causes of War: Power and the Roots of Conflict (Ithaca, N.Y.: Cornell University Press, 1999). 60. Colin Elman and Miriam Fendius Elman, “Lakatos and Neo-Realism: A Reply to Vasquez,” American Political Science Review, Vol. 91, No. 4 (December 1997), pp. 923–926. draft -- do not cite or quote without author’s permission 36 progress in international relations theory realism have a distinctive hard core? These questions must be answered if the Lakatosian approach is to have any meaning at all. In lieu of setting forth such competing paradigms, Elman and Elman shift to the threadbare “level of analysis” distinction, now nearly a half century old, whereby realist theories — neorealist or not — are said to share a focus on the “external” environment of the state, as opposed to its “internal” environment. This formulation of the levelof-analysis distinction is untenable, precisely because it cannot be reduced to distinct core assumptions. (Indeed, all level of analysis distinctions are probably incoherent.) No rational calculation in world politics focuses entirely on the “external” environment. All such calculations compare instead the attributes of one country to the attributes of others; it is the relative position of a country that matters. In this sense, realists, liberals, and institutionalists all assume that states strategize in response to “systemic” imperatives; that is, they make policy by comparing their own internal characteristics with those of foreign states. In this view, the primary difference between realism, liberalism, and institutionalism lies not in the tendency of some to focus instead on “domestic” or “second-image” variables, but in the particular characteristics of states that they choose to compare in the formulation 61 of national strategy. For realists, it is material power resources. For 61. For this reason, I disagree with the suggestion of Stephen Krasner and Robert Jervis at the Progress in International Relations Theory conference (Scottsdale, Arizona, January 1999) that liberal theories simply subsume what were traditionally called “second image” theories. The core liberal claim is not that “domestic politics” is dominant. For liberals, two other conceptual distinctions are fundamental: the first stresses the fundamental sources of differences among states, the second the way in which those differences translate into political behavior. The first is a distinction between the international political system, on the one hand, and civil society (domestic and transnational) on the other. Liberals, in contrast to realists and institutionalists, stress the importance of state-society relations and the ultimate primacy of the societal context. In other words, underlying interdependence among societies, which drives interdependence among policies, is the fundamental force underlying state behavior. The second distinction, entirely at the interstate level, is between different characteristics of states that might drive policy: the draft -- do not cite or quote without author’s permission moravcsik: chapter 5 37 institutionalists, it is information. For liberals, it is underlying preferences. Once we set aside the misleading “level of analysis” language and focus on assumptions — as Lakatos invites us to do — we find that the realist emperor has no core. The unwillingness of realists, including Elman and Elman, to confront this issue raises serious concerns both about the integrity of the modern realist paradigm and about the power of Lakatosian language to police the integrity of paradigms in general. The lack of distinctiveness of realist theory is a flaw so fundamental that it transcends debates about the relative virtues of specific philosophies of science proposed by Lakatos, Laudan, and others. If a set of core assumptions is so broad as to be shared by a paradigm and nearly all its recognized competitors, what use is it? In addressing this problem, realists face a difficult choice. They may either define realism narrowly, and thereby admit the existence of increasing numbers of empirical anomalies, or they may water down the hard core to a “minimal realist” foundation, thereby permitting realist theory to degenerate into a loose and generic rationalism consistent with nearly every claim about world politics advanced in the past generation. The Lakatosian framework has the not inconsiderable virtue of making this choice explicit. If it fails to force a choice, as Elman and Elman appear to believe, then we must surely question whether Lakatos’s philosophy has any utility whatsoever in social science. distribution of preferences, resources, information, and beliefs. Liberal analysis stresses the distribution of preferences, and hence all major liberal variables are “systemic,” at least insofar as the influence of commercial incentives, national ideals, and regime type on the foreign policy of a given country cannot be assessed in isolation from the corresponding characteristics of other countries. More broadly, this suggests that the level-of-analysis distinction is a hindrance to understanding. The real debate in IR theory is not between second-image and third-image theories, but between different conceptions of the structure of the international system. Is that structure best understood in terms of the distribution of preferences, of resources, or of information? This is consistent with the framework proposed by David A. Lake and Robert Powell, eds., Strategic Choice and International Relations (Princeton, N.J.: Princeton University Press, 1999). draft -- do not cite or quote without author’s permission 38 progress in international relations theory International Relations: The Limitations of Lakatosian Assessment In this essay, I have offered a reconstruction of the liberal hard core and the theories that follow from it, and a demonstration that the resulting research program meets the three most important criteria for excess explanatory content: temporal novelty (Lakatos1), heuristic novelty (Lakatos3), and background theory novelty (Lakatos4). An admittedly crude comparison suggests that the liberal scientific research program has generated results at least as progressive as other major scientific research programs within broad domains of state behavior. Most strikingly, recent research has consistently led to the degeneration of other theories into liberal theory, not the reverse. Let me conclude, however, by turning away from these substantive conclusions about liberal theory to three considerations concerning the topic of this volume, namely the application of Lakatosian philosophy of science to IR theory. Despite the seemingly unambiguous positive result (for a theory I happen to favor), and the clear virtues of forcing social scientists to focus on the explanatory power of distinct core assumptions, my argument suggests some limitations as well as strengths of Lakatosian philosophy as a tool to assess IR theory. Overall, a more pragmatic “problem-solving” approach based on Larry Laudan’s philosophy of science seems more appropriate than one 62 based on strict Lakatosian criteria. the rigor and utility of the “background theory novelty” criterion Let us first concede the virtues of the Lakatosian approach. The analysis above suggests the utility of the “background theory novelty” criterion (Lakatos4). Elman and Elman, we have seen, follow Lakatos’s own tendency and reject background theory novelty because they believe that, in the face of anomalies, it remains trivially easy to develop auxiliary propositions that successfully protect the hard core. One can always add appropriate auxiliary propositions to account for anomalies, without thereby creating additional anomalies. 62. Laudan, Beyond Positivism. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 39 I see little evidence that this is the case, except in a trivial sense, and the recent failures of the realist research program demonstrate why. It is in fact difficult to explain new facts within a consistent set of hardcore propositions without generating overt contradictions. The result, in the case of recent realist writings, has been a transparent appropriation of propositions based on assumptions that — as a matter of intellectual history as well as modern paradigmatic reformulation — are anything but realist. The failure of realism to progress, it is critical to note, is not simply an outside judgment reached by liberals (such as myself) or institutionalists and epistemic theorists defending arbitrarily 63 chosen terrain. Instead, recent realists have taken a position that is internally contradictory. Realists find it impossible to match the distinctive and confirmed empirical claims of other paradigms without either violating the traditional realist hard core or loosening it to the point where it no longer has any theoretical power. These conclusions suggest — in the spirit of essays in this volume by Andrew Bennett (Chapter 14) and David Dessler (Chapter 11) — that “background theory novelty” is a more useful criterion than the Elmans’ introductory chapter suggests, as well as one that casts the liberal scientific research program in a favorable scientific light. theory synthesis and the liabilities of lakatos Now the limitations: although the liberal scientific research program appears to be vindicated by the analysis in this chapter, and “background theory novelty” a more useful criterion than Elman and Elman concede, I maintain that we must nonetheless acknowledge significant problems inherent in any application (even metaphoric) of Lakatosian philosophy of science to IR theory. Lakatosian theory is designed to explain the resolution of conflict among a small number of fundamental theories within a uniform field 63. This is a charge made by our critics. See Legro and Moravcsik, “Is Anybody Still a Realist? The Authors Reply,” in “Correspondence: Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?),” International Security, Vol. 25, No. 1 (Summer 2000), pp. 184–193 (critiques by Peter Feaver, draft -- do not cite or quote without author’s permission 40 progress in international relations theory of scientific inquiry. Lakatos and those who have sought to elaborate his approach tend to view theories as claiming plausibly to explain an entire scientific domain. The image is one of a series of discrete conflicts among such theories with ever-expanding empirical scope. While there can be extended failures to agree upon a single paradigm, these tend to be the transitional consequences of the need to assemble and analyze a large body of ambiguous data, rather than fundamental uncertainty about the nature of the microfoundations of the phenomena in question. Under such circumstances, Lakatos expects that conflict among theories will eventually result (or, hypothetically, could ideally result) in the vindication of one, which will subsume the loser by explaining all of its content. This image implies heroic confidence in the universal applicability of some single set of microfoundational assumptions — confidence that has been vindicated in 64 some areas of the natural sciences. The study of world politics, by contrast, often manifestly fails to meet these criteria — at least at its current state of development. Even broad scientific research programs such as realism, liberalism, and institutionalism (let alone specific theories such as work on the “operational code” or the democratic peace) do not make any plausible claim to universality, even within a circumscribed domain. It is next to impossible to find any reputable scholar willing to advance such universal claims for liberal, realist, or institutionalist theory. More importantly, there is no a priori reason to believe that such a universal claim would be valid. By contrast to the claims advanced by Newton, Einstein, Darwin, and other scientific revolutionaries, which rested on what was arguably a unique and exclusive conceptual foundation, there is little fundamental theoretical reason to assume that war is the result of, say, the non-democratic governance and underlying social conflict cited by liberals, rather than the perturbations in the balance of Gunther Hellmann, Randall Schweller, Jeffrey Taliaferro, and William Wohlforth). 64. Anything less would reduce Lakatos’s criterion to a pragmatic admonition to seek evidence for competing claims, thus ridding it of almost all distinctive content. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 41 power cited by realists or underdeveloped international organization cited by institutionalists. It is not difficult to conceive of sociological and psychological microfoundations (say, a “rationalist” framework of analysis) that encompass all of these. In this context, the tendency of Lakatosian analysis to focus attention on zero-sum conflict among all-encompassing theories is a liability, most obviously because it poses a manifestly unrealistic 65 standard. No one expects any of these theories, including liberal theory, to supplant or “knock out” its competitors, even within a 66 limited realm. International relations theory without realism or institutionalism strikes me as absurd on its face. The fundamental problem is that Lakatosian philosophy, even when employed as a heuristic, inhibits full recognition that international relations is ineluctably multi-paradigmatic. Lakatosian philosophy of science tends to block other trajectories of theoretical and disciplinary development. Two are of particular importance for IR theory. First, Lakatosian thinking inhibits appreciation of the possibility that liberal and other IR theories may be differentially applicable across different specific empirical domains of world politics. Each may have areas of relative power and relative weakness. Keohane and Nye theorized some years ago, for example, that the world of anarchic competition and the world of “complex interdependence” required 67 different theories. In other words, Lakatosian emphasis on maximal claims about the scope of an explanation may blind us to narrower, subtler, and more nuanced conclusions about the conditions under which particular theories have explanatory power. Such a world of 65. This tendency is related to what Keohane and Martin (Chapter 3 in this volume) term the “endogeneity problem.” 66. No one except, curiously, defensive realists, who incorrectly attribute this view to Legro and me. See Randall Schweller and Jeffrey Taliaferro in “Brother, Can You Spare a Paradigm? (Or Was Anybody Ever a Realist?)” International Security, Vol. 25, No. 1 (Summer 2000), pp. 174–178, 178–182. 67. Keohane and Nye, Power and Interdependence. draft -- do not cite or quote without author’s permission 42 progress in international relations theory accurate mid-range theories seems closer to our grasp than one with a single dominant theoretical paradigm. Second and more fundamental, Lakatosian thinking inhibits appreciation of the possibility that paradigms such as realism, institutionalism, and liberalism can usefully be deployed as complements rather than substitutes. From this perspective, the central challenge facing IR today is not selecting the correct philosophy of science most likely to help us develop a universal theory of IR, but selecting frameworks that permit us to engage in rigorous theory synthesis. The central issue here is how analysts should combine major theories into testable explanations of classes of phenomena in world politics, without permitting the resulting empirical analysis to degenerate into a mono-causal approach, on the one hand, or an indeterminate “everything matters” approach, on the other. Each would be deployed to explain different aspects of the same interstate interactions. The potential complementarity of basic IR theories follows from precisely the aspect that bedevils efforts by “neoclassical realists” to specify a distinct “hard core,” namely their shared rationalist assumptions. Within a rationalist world — and most IR theories are predominantly rationalist — there is little fundamental reason to believe that any single theory of the scope of liberalism, realism, or institutionalism could or should triumph. To see why, one need only consider a basic form of rationalist analysis, such as bargaining theory 68 or negotiation analysis as practiced by its leading analysts. In such analyses, it is possible, indeed conventional, to combine preferences (liberalism), coercive resources (realism), and information and norms (institutionalism or constructivism), as well as other factors, into synthetic explanations of bargaining outcomes. Indeed, coherent “bargaining theory” without variation in all these factors seems nonsensical. For the purposes of empirical analysis, separating the 68. For an overview, see Howard Raiffa, The Art and Science of Negotiation: How to Resolve Conflicts and Get the Best out of Bargaining (Cambridge, Mass.: Harvard University Press, 1982). draft -- do not cite or quote without author’s permission moravcsik: chapter 5 43 problem into competing theories may often be counterproductive. A structured synthesis would be far more illuminating. An example of structured synthesis, taken from recent empirical research on European integration, places major theories in sequence. In my analysis of major negotiations to create, develop, and amend the treaty structure of the European Union, liberal theory is employed to account for national preferences, rationalist bargaining theory (which could be seen as a non-coercive variant of realism) to account for the efficiency and distributional outcomes of negotiations, and 69 institutionalist theory to account for subsequent delegation. This is only one — although arguably the most general — of many competing generalizable models for synthesizing theories, including qualitative 70 frameworks, multivariate equations, and formal models. ontologies, paradigms, theories: the proper scope of research programs This leads us to a final consideration, namely the proper scope of a paradigm. Some might concede that Lakatosian criteria are inappropriate for theories such as liberalism or realism, yet nonetheless maintain that Lakatosian concepts can nonetheless usefully be employed to evaluate smaller or larger theoretical aggregations: narrower theories or broader “ontologies.” Many of the writers in this volume maintain that Lakatosian criteria are appropriately applied to narrower theories, such as democratic 69. For a explication and empirical application of this method, see Andrew Moravcsik, The Choice for Europe: Social Purpose and State Power from Messina to Maastricht (Ithaca, N.Y.: Cornell University Press, 1998). This is consistent with the implicit model set forth in Lake and Powell, Strategic Choice and International Relations. 70. Even if liberal IR theory, to take one element of this proposed synthesis, could be shown to be currently underutilized, of greater power or scope than the alternatives, or analytically prior to other theories (in the sense that the variation in interstate preferences it explains determine the conditions under which these other theories are valid), this would not constitute a valid reason to reject the realist or institutionalist paradigms entirely. draft -- do not cite or quote without author’s permission 44 progress in international relations theory 71 peace theory or theories of international regimes. I do not respond to the full argument of these chapters, but my analysis does suggest caution in accepting such claims. If, as I argue, realism, liberalism, and institutionalism are often complements rather than substitutes, would this not be even more true of narrower hypotheses within these traditions? It is hard to see, for example, why democratic peace theory should plausibly constitute an exclusive theory of war, and thus it is difficult to see what is gained by evaluating its progress and promise within a Lakatosian framework. Insofar as any theoretical constructions in IR could plausibly advance the type of exclusive claim to explanatory power within a given domain favored by Lakatosian philosophy of science, it must therefore be a theoretical paradigm at a broader level, such as what Alexander Wendt terms the “ontological” level of “rationalism” or “sociological theory.” An ontology can plausibly make a universal claim across a broad domain, and many believe that such claims are mutually exclusive. One might more reasonably speak of a rationalist research program in IR, with realist, institutionalist, and liberal 72 “paradigms” as leading elements. There is firm grounding in fundamental social theory for advancing 73 such a claim. Rationalist theories of social interaction, regardless of their substantive scope, tend to isolate three or four basic categories of fundamental causal factors — normally resources, preferences, beliefs, and perhaps information. Hence within a rationalist paradigm, which might perhaps be properly judged using Lakatosian criteria, we should find theories that give causal priority to the international distribution 71. This position tends to be held by those who are uncomfortable with the breadth of the liberal paradigm as formulated here. It is important to reiterate that while liberal theory is broad in theory, it tends to be narrow in practice. There are relatively few specifications of each variant — in Lakatosian language, relatively few sets of auxiliary propositions — that can survive empirical testing. The resulting research has therefore been quite focused. 72. Many variants of so-called “liberal constructivism” would be included. Risse-Kappen, “Collective Identity.” 73. James S. Coleman, Foundations of Social Theory (Cambridge, Mass.: Harvard University Press, 1990); Lake and Powell, Strategic Choice. draft -- do not cite or quote without author’s permission moravcsik: chapter 5 45 of resources (realism), preferences (liberalism), information 74 (institutionalism), and beliefs (epistemic or constructivist theory). Future research might profitably assess such meta-paradigms, given how many scholars today seek to reconceptualize international relations theory in terms of a dichotomy between “rationalist” and 75 “sociological” (or “constructivist”) theory. Yet even at this very broad level of abstraction, there remains space for skepticism. The same criticism of Lakatosian analysis advanced above in regard to realism, liberalism, and institutionalism applies equally here. There is no reason to believe that the psychological underpinnings of rationalist or sociological explanation are, in the real world, mutually exclusive. Complex combinations are possible. Few if any serious scholars are willing to assert that only “rational choice” or only “socialization” exists. Recent constructivist efforts to reformulate IR theory as debates between “rationalist” and “sociological” theory 76 are being abandoned by more sophisticated proponents. The constructivist challenge is now focused primarily on the need to forge a structured synthesis between rationalist and sociological theory, rather than demonstrating the dominance of one or the other. Under such circumstances, it is unclear what is to be gained by structuring academic discourse as a battle among mono-causal claims. Second, as Alexander Wendt, Iain Johnston, and others concede, there is only a very loose connection, if any at all, between ontology, at the level of 77 rationalism and constructivism, and concrete testable theory. Many predictions — including realist ones, as Johnston has shown, and liberal ones, as Wendt has demonstrated — are equally consistent with 74. Moravcsik, “Taking Preferences Seriously.” 75. Peter J. Katzenstein, Robert O. Keohane and Stephen D. Krasner, “International Organization and the Study of World Politics,” International Organization, Vol. 52, No. 4 (Autumn 1998), pp. 645–685. 76. Finnemore and Sikkink, “International Norm Dynamics”; and RisseKappen, “Collective Identity.” 77. Alexander Wendt, Social Theory of International Politics (Cambridge, UK: Cambridge University Press, 1998); Alastair Iain Johnston, Cultural Realism: draft -- do not cite or quote without author’s permission 46 progress in international relations theory constructivist and with rationalist ontology. Once the connection between ontology and concrete hypotheses has been broken, it becomes unclear exactly how Lakatosian criteria could be employed or what meaning they would have. This suggests not only that a paradigm can be too narrow for Lakatosian assessment, but also that it can be too broad. In conclusion, the discipline imposed on theory construction and development by the Lakatosian approach — at least in retrospect — is surely a useful reminder of the need for consistent assumptions, rigor, comparative theory testing, and the need to explain patterns in empirical data efficiently. Yet Lakatos’s focus on the scope of theories might encourage scholars to advance “universal” and mono-causal claims when it is inappropriate to do so. More appropriate may be a clear specification of proper empirical limits or more subtle theoretical syntheses. Whatever benefits the Lakatosian metaphor may offer, the debates among IR “isms” framed in universal and mono-causal terms that it helps perpetuate can hardly be considered a spur to scientific progress. Overall, the viability of the “background theory novelty” criterion and the more pragmatic “problem-solving” approach adopted here suggests that criteria proposed by Larry Laudan are 78 more appropriate than those of Imre Lakatos. Lakatosian standards — and, for the same reasons, any rigid definitions of paradigms as building blocks for theory development — should be imposed in 79 international relations only with the utmost caution and modesty. Strategic Culture and Grand Strategy in Chinese History (Princeton, N.J.: Princeton University Press, 1995). 78. Laudan, Beyond Positivism. 79. Overall, however, this finding is consistent with existing work on IR paradigms that deliberately employs more straightforward criteria, such as distinctiveness and coherence, rather than explicit philosophy of science. See Legro and Moravcsik, “Is Anybody Still a Realist?”; and Moravcsik, “Taking Preferences Seriously.” draft -- do not cite or quote without author’s permission