Review of International Studies page 1 of 21 doi:10.

1017/S0260210511000052 

2010 British International Studies Association

Why and how should we go for a multicausal analysis in the study of foreign policy? (Meta-)theoretical rationales and methodological rules
YONG-SOO EUN*

Abstract. This article argues that International Relations (IR) researchers concerned with why-questions about the state’s external behaviour ought to employ a multicausal approach attentive to the interrelated relationship between external structures and internal agents, presenting the (meta-)theoretical rationales underlying its argument. Here the author suggests ‘a rich/bold ontology’ regarding foreign policy behaviour. Then the article elaborates on detailed and explicit guidelines on how to traverse the bridge that connects the insights of that rich ontology to the empirical research necessary to make claims about the real world of any one moment. In a related vein, the article claims that a multicausal approach should be established using what the author calls ‘loose-knit deductive reasoning’ through which epistemological and methodological openness can be preserved in a manageable way. More importantly, this article discusses the role of theory for IR scholarship and the standards for judging theoretical contributions and progress in the field of IR. Ultimately, the author argues that a complex and flexible approach – both as a useful mode of explanation and as a progressive model of theory construction – can make important contributions to a better understanding of foreign policy and world politics, not only because it enables researchers to become keenly sensitive to the complex reality underlying a nation’s foreign policy and to the interrelated relationship between structures and agents in international relations, but also because it can serve to provide a secure base for the progressive accumulation of the evidence closely associated with multiple causation on which any adequate explanation about complex foreign policy behaviour must surely be founded and without which general theory cannot really flourish. Yong-soo Eun is broadly interested in International Relations theories, Foreign Policy Analysis, and the international relations of East Asia. His current research, in the Department of Politics and International Studies at the University of Warwick, addresses questions of why states behave as they do in world politics (from a theoretical perspective) and how one can more fully and accurately develop causal explanations of complex foreign actions of the state (from a methodological perspective), with a focus on the Asia-Pacific region (from an empirical perspective). Previously, he worked at the Institute for Far Eastern Studies as a Research Fellow and was offered a Visiting Fellowship from the Lowy Institute for International Policy. He is presently completing two journal articles which discuss modes of construction of IR Theory and the security dynamics of East Asia, respectively. Young-soo can be contacted at: {Y-S.Eun@warwick.ac.uk}.

* I would like to thank Valerie Hudson, Peter Ferdinand, Colin Hay, Christopher Hill, and the RIS’s anonymous reviewers for their helpful comments on earlier drafts of this article.

1

2

Yong-soo Eun

Introduction Why do states behave the way they do on the global scene? This is a crucial question in the field of International Relations (IR). Just as in any field of scholarly inquiry in social science, most, if not all, IR scholars utilise theories – or, at least some sorts of analytical tools – in order to answer such a challenging causal question. And every theory in every field of study in social science has its own initial assumptions about the nature of the social and political reality to be investigated – assumptions about what exists, what it looks like, and what units make it up – and claims about the conditions of acquiring knowledge of that which exists and the relation of our knowledge to that reality. To use philosophical terminology, all theoretical positions are, even if implicitly, dependent upon particular assumptions about ontology (the philosophy of being) and epistemology (the philosophy of knowledge), which directly affect the theory of method – shaping the approaches that researchers take to their research. In Marsh and Furlong’s analogy, ontology and epistemology are like ‘a skin, not a sweater that can be put on when we are addressing philosophical issues and taken off when we are doing political research’.1 In short, ontology, epistemology, and methodology are closely related. Yet this by no means suggests that all theories in the same field of inquiry have to retain the same ontological or epistemological assumptions on which they are premised. Indeed, in social science there exist distinctive and even exclusive assumptions and theories that attempt to explain the same phenomena. So it is with the discipline of IR. Clearly, different theoretical schools of thought compete against each other to explain the same features of world politics within the discipline. Stated differently, there have been divisive ontological and epistemological debates, or so-called ‘great debates’, in the history of the discipline.2 Not surprisingly, it follows that divergent theories – perhaps too many – in the study of International Relations and foreign policy have been formulated.3 However, this does not simply mean that all theoretical schools of thought have been valued equally in this field. In effect, the vast majority of IR scholars have employed realist and rationalist perspectives when it comes to the explanations and predictions of interactions among states in the international system. It appears that it was commonplace by the 1980s to speak of the three approaches – realism, liberalism, and Marxism – as constituting an ‘interparadigm debate’ in IR scholarship, yet what actually happened, as Steve Smith has commented, was that ‘realism dominated the discipline of IR’; it is ‘only in the last twenty years that it has become common to discuss more than rationalist accounts’.4 And, as will be
1

2

3

4

David Marsh and Paul Furlong, ‘A Skin not a Sweater: Ontology and Epistemology in Political Science’, in David Marsh and Gerry Stoker (eds), Theory and Methods in Political Science (New York: Palgrave McMillan, 2002), p. 21. Ole Wæver, ‘The Rise and Fall of the Interparadigm Debate’, in Steve Smith, Ken Booth and Marysia Zalewski (eds), International Theory; Positivism and Beyond (Cambridge: Cambridge University Press, 1996). Stephen Walt, ‘International relations: One world, many theories’, Foreign Policy, 110:1 (1998), pp. 29–46. Steve Smith, ‘Diversity and Disciplinarity in International Relations Theory’, in Tim Dunne, Milja Kurki, Steve Smith (eds), International Relations Theories: Discipline and Diversity (Oxford University Press 2007), pp. 4–9. Note, however, that in Europe, when comparing with the US IR

Keohane and Joseph S. 27–44. Robert Gilpin. ‘Structural Realism after the Cold War’. Crawford. pp. European Journal of International Relations. 87:3 (1993). p. Waltz. ‘Normative and Structural Causes of the Democratic Peace’. Charles W.(Meta-)theoretical rationales 3 discussed in detail below. Kegley (ed. it is now openly stated that successive rounds of the ‘interparadigm debate’ have drawn the two approaches ever closer together such that it is ‘difficult to position clearly once prominent neo-realists or neo-liberals’. See. Kenneth N. 1975). Power and Interdependence: World Politics in Transition (Boston: Little. 24:2 (1999). and War: Rethinking the Democratic Peace Debate (Boulder. assuming that states are rational actors. Liberalism. regime types) at the expense of the individual decision-maker. Theory of International Politics (New York: Random House.10 Scholars who work within the approach of constructivism also tend to accept a structure-weighted form of IR theorising: they draw attention to the scholarship.9 In this sense. ‘A New Grand Strategy’. Introduction (New York: Palgrave. constructivism. 2001). pp. pp. Kenneth N. Colin Hay. pp. 35:1 (2009). Robert O. International Political Economy flourished to a much greater extent than realism in the 1980s and 1990s. 5 6 7 8 9 10 . Waltz. much of the analytical attention of neo-liberalism focuses on the social and political context. 1979). as that of neo-(structural) realism does: neo-liberals place explanatory weight upon such contextual or structural factors as international institutions7 or interdependence8 in examining international relations. English School. realism and rationalism maintain structuralist tendencies which privilege structural or contextual factors and thus marginalise the autonomy of the state’s personnel (human agents) in world political analysis. Democracy. for instance. ‘The Passion of World Politics Propositions on Emotion and Emotional Relationships’. Tarak Barkawi and Mark Laffey (eds). 289:1 (2002). 1:3 (1995). the recent challenges to realism from other theoretical approaches in IR are also – to a greater or lesser degree – grounded in structuralism. ‘The Stability of a Unipolar World’. Oran R. John Mearsheimer. Marc A. 19. see Neta C. 624–38. International Security. On this point. pp. rationally or collectively respond – in relation to external stimuli and the environments around them. 267–331. pp. William C. 116–56. Levy. the literature on the democratic peace theory including Zeev Maoz and Bruce Russett. though there are clear differences in emphasis between neo-realists and neo-liberals. ‘The study of international regimes’. ‘American power preponderance and the nuclear revolution’. It should therefore come as no surprise that throughout much of the twentieth century a whole generation of IR scholars became preoccupied with the question of how the structure of the international system affects the courses of action of the state. Bruce Russett. Review of International Studies. International Security. 1977). 1993). which has led to the identification of a ‘neo-neo synthesis’.6 Furthermore. Grasping the Democratic Peace: Principles for a Post-Cold War World (Princeton: Princeton University Press. 25:1 (2000). Martin Press. 2002). a great deal of attention has recently been focused on the role that domestic politics play in shaping the state’s external action. Norton. 1995). US Power and the Multinational Corporation: The Political Economy of Foreign Direct Investment (New York: Basic Books. 5–41. 5–41. for a useful critique of the democratic peace thesis. Brown and Company. International Security. Nye. 2001). Wohlforth.5 To be certain. For example. W.). Christopher Layne. 35:1 (2009). Controversies in International Relations Theory (New York: St. See. pp. Christopher Layne. See. Young and Michael Zürn. assuming that decision-makers react in an undifferentiated manner – in other words. Craig Campbell. American Political Science Review. The Atlantic Monthly. but much of the work has tended to emphasise the importance of domestic structures or conditions (for example. Review of International Studies. See also the 2009 special issue (no. Political Analysis: A Critical. 1) of World Politics in which nine authors explore the impact of the unipolar distribution of capabilities in the contemporary international system on the behaviour of the dominant state and on the reactions of other states. ‘America’s Middle East grand strategy after Iraq’. 24:4 (2000). 5–25. for example. for example. CO: Lynne Rienner. 36–42. for example. pp. See. The Tragedy of Great Power Politics (New York: W. See also.

Ibid.15 The realist perspective is. . 1:1 (2005). averred that ‘to search for the clue to foreign policy in the motives of statesmen is both futile and deceptive’. Theory of International Politics. then. ‘Foreign Policy Analysis: Actor-Specific Theory and the Ground of International Relations’.16 Instead. Morgenthau.] in the wider society’ to explain agents’ social actions.11 In Wendt’s own words. Foreign Policy Analysis.. The state’s external behaviour is explained as a function of its position vis-à-vis structures like geography or a distribution of material power. Since states are assumed to be unified actors that have the same operating motivation – acquiring and securing the national interest – in international relations. 1948). 92. Alexander Wendt. Instead. Social Theory of International Politics (Cambridge: Cambridge University Press. Waltz. because all leaders act in ways consistent with the ‘national interest’ of the state. Politics Among Nations: The Struggle for Power and Peace (New York: Alfred A. Hans Morgenthau. classical realists claim that analysts can ‘consider all decision-makers to be alike’. for instance. tended to accept a set of explanatory assumptions based on structuralism. Agents. p. In Morgenthau’s words: 11 12 13 14 15 16 Alexander Wendt. even if implicitly. actor-general in its orientation: only if people are reacting regularly to external constraints does it make sense to see external constraints as explaining their actions. 9. 117. . To be sure. Before proceeding further.14 In short. in short. . but it is important to remember that the individual per se is not the subject of this perspective. classical realism considers individuals. .] what is required is more [.] accurate structural accounts’. for instance. The image of the role of state leaders in conducting foreign policy comes close to being a mechanical image in which their choices are shaped by the international ‘structural constraints’ that they face. it is important that we are first clear about the underlying assumptions of the mainstream realist and rationalist approaches and their tendencies towards structuralism in relation to causal explanations. 113. As Colin Wight rightly observes. we need to ‘invoke the structure of shared understandings existing [. say the individual decision-maker. . 6. . is unimportant. Hans J.4 Yong-soo Eun intersubjective nature of structure. ‘for Wendt [. they agree that it is fruitful to conceptualise a state as a unitary actor behaving in a rational manner in pursuit of maximising power or survival in the international system. 2006). Hudson. p. Valerie M. p. Structure-oriented IR theories Although realists do not constitute a homogeneous school. uses Kenneth Waltz’s structural realism as a counterpoint to his own preferred ‘structuration theory’. what is within the state. Structures and International Relations (Cambridge: Cambridge University Press.12 Put otherwise.13 In this sense. 1999). Knopf. the central analytical focus is on the structure of the international system. the majority of IR scholars have. p. Valerie Hudson has commented that IR theorists ‘currently provide much more insight into structure than agency’. p. Wendt does not reject structural theorising. Colin Wight. .

11:4 (2009). pp.18 their emphasis is on the positions that policymakers occupy rather than on individual characteristics of the policymakers who hold the positions. and World War I’. To put it another way. 927–40. As such. Hence. . ignored. 45. and Foreign Policy (Cambridge: Cambridge University Press.] provides for rational discipline in [foreign policy] action [. 2009). . American Political Science Review.’20 In short. there is. World Politics. p. for example. . ‘The Power of Human Cognition in the study of World Politics’. 84:4 (1990).(Meta-)theoretical rationales 5 We assume that statesmen think and act in terms of [national] interests defined as power [. whatever decisionmaking unit is involved. and the balance of power. Papayoanou. present or future – has taken or will take on the political scene [. no need to delve into individual decision-makers or closely examine decision-making processes – these can be ‘minimized. In this vein. with the concept of the ‘national interest’ and the notion of the ‘rationality’ of the leader or excessively focus on the common role/position that the pertinent agent. Paul A. IR scholars who work according to realist and rationalist approaches can be called methodological structuralists. domestic institutions. Germany. . states are alike in the way that they are propelled by external forces. ‘Taking Stock of Neoclassical Realism’.] [T]hat assumption allows us to retrace and anticipate [. changes in leadership have little consequence. and intellectual and moral qualities of successive statesmen. p. since they attempt to account for the causal effect of structure on state behaviour by appeal to a systemic logic that operates as a whole. and thus policymakers as such have ‘been mostly missing’ from neoclassical realist scholars’ actual accounts of foreign policy19 while still taking ‘the state’s relative power their [the neo-classical realists’] chief independent variable. pp. And. ‘Neoclassical Realism and Theories of Foreign Policy’. .22 17 18 19 20 21 22 Morgenthau. ‘Concrete Theory: An Emerging Political Method’. Jerel A.] the steps a statesman – past. seeking for regularities in observed patterns of the behaviour of states in world politics. be it a human being or a group of humans. Explorations of foreign policy and world politics from this position are thus sought through emphasis on the structure at the system level rather than on a state’s internal variables at the unit level. and the Balance of Power: Britain. . p. the realist perspectives generally conflate leaders and countries. International Studies Review.] regardless of the different motives. since all black boxes (countries/regimes/leaders) work the same way. See. 20:2 (1996).21 Like billiard balls. . 51:1 (1998). from this perspective. Much the same can be said about the works of a Realist-Liberal synthesis. International Security. Gideon Rose. Lane Ruth. or assumed away’. p. . 42–76. Neoclassical Realism. 5. the realist and rationalist vision of International Relations is often referred to as the ‘billiard ball model’ or ‘black-box’ concept of state interactions. preferences. Since from the realist perspectives the national interest does not change and all national leaders have to be concerned with the effects of foreign policies on the country’s security and economies. Politics Among Nations. Lobell and his colleagues’ edited book. such as the president or the prime minister. International Studies Review.] The concept of national interest defined as power [.17 Also although neo-classical realist researchers unanimously emphasise the role of policymaking executives. . 151. . Papayoanou’s 1996 articles on economic interdependence. the State. realist scholars turn their analytical attention to the structure of the state and the international system. 2:3 (2000). See. for example. that unit can be modelled as a unitary rational actor and therefore be made equivalent to the state. 802. Shiping Tang. Rosati. Institutions. ‘Interdependence. . takes.

Models of bounded rationality (Cambridge. Explaining Foreign Policy (NJ: Prentice-Hall.6 Yong-soo Eun Let us. MA: MIT Press. 1982). see Lloyd Jensen. Crawford. 12 (2008). . pp. World Politics. 116. rid ourselves of the troublesome abstraction. the concepts of ‘bounded rationality’23 and ‘multiple rationalities’24 epitomise the problem of the implausible assumptions about the rational capacity of individuals. Steven Lukes and Martin Hollis. p. ‘The Constructivist Turn in International Relations Theory’. is it fruitful to follow the line of the structuralist arguments and their actor-general theoretical position? I posit that we must answer this question in the negative. Robert Jervis. The New Foreign Policy: US and Comparative Foreign Policy in the 21st Century (Rowman and Littlefield. decisions cannot be made in terms of comprehensive rational calculation. constructivist scholars will find themselves ‘unable to explain where their powerful social structures come from in the first place. without such attention. not only realist researchers but constructivist scholars are also often criticised for not giving sufficient and sustained attention to internal factors and especially to human agents. pp. Perhaps the most serious weakness of rationalist and realist theory is that it assumes rational calculation on the part of decision-makers. 10–12. Neta Crawford notes that ‘deterrence theory may be fundamentally flawed [. to human agents – they are to run into difficulties in explaining and/or projecting significant changes in the international system. For more on this. because. In this vein. and thus speak of them in a vague way. first of all.’25 What is more. because structure-oriented realism and rationalist approaches have many limitations and weaknesses in explaining as well as predicting the foreign behaviour of the state. falling short of answering the questions that are of central interest to many IR researchers and are crucial in their further research. 3.27 Robert Jervis notes that ‘if social constructivism is to develop seriously. 1982). 50:1 (1998). 1982). ‘state’! Yet. owing to a lack of proper attention to what is within the state – in particular. Jeffrey Checkel.28 23 24 25 26 27 28 Herbert Simon. Rationality and Relativism (Oxford: Blackwell.26 In short. liberal theories of cooperation under anarchy and the formation of security communities that stress actors’ rational calculation of the benefits of communication and coordination are deficient to the extent that they do not include careful consideration of emotion and emotional relationships. And. Theory Talks. p. which is an idealised situation but one that is seldom realised. where variations can be identified. it has to be grounded in psychology’. ‘Nuclear Weapons. In this sense. The variations in behaviour during the time when a certain system is maintained cannot be explained by reference to the system structure because the structure has not changed. 339. 2003). p. Because of the great uncertainty and lack of complete information in international affairs. as well as the many public and private actors who are concerned with foreign policy issues. structure-oriented approaches explain outcomes in international affairs only in the most general manner. . Explanations of such variations must be found at lower levels of analysis. Laura Neack. . and why and how they change over time’. 8–15. Explaining the non-Realist Politics of the Bush Administration and US Military Presence in Europe’.] Similarly. ‘The Passion of World Politics Propositions on Emotion and Emotional Relationships’. structure-oriented approaches stand accused of a failure to appropriately acknowledge the role of domestic politics in and the influence of human agents upon the course of world events.

our accounts need to factor in the pertinent policymakers’ beliefs about the world and their images of other political actors. In other words. precisely because the core political beliefs held by the key policymakers involved in the foreign policy issue in question can wield a significant influence on their behaviours (decisions and choices). If we would like to lay out a compelling explanation of the state’s external action. Human decisionmakers are not interchangeable with generic rational ‘utility maximizers’. 63–96. individual decision-makers should not (and cannot) be conceptualised as being equivalent to the state and assumed to be unified rational or social actors who have the same operating motivation – acquiring and securing the national interest defined in material. As discussed. . we should neither conflate national leaders and their countries nor consider idiosyncratic human policymakers exogenously. I argue. in order for them to have a sound basis for closely examining human agents. 42:1 (1998). NJ: Princeton University Press. societal or normative terms – in pursuing foreign policy. Cognitive psychology and cognate theories29 demonstrably tell us that humans perceive and simplify reality based on their individual beliefs. Plausible explanations of the external action of the state must offer causal mechanisms that pass through the pertinent individuals. in doing so. Political Psychology. discard the idea of the state as a metaphysical abstraction while instead embracing human policymakers as a realistic concept of an agent in international relations. policymakers) as an analytical category in its own right and that treats the specific factor of human agents as an end in itself. there are individually tailored perceptual filters and cognitive constraints that work in practice. human beings view the external condition differently. Deborah Welch Larson. Put differently. ‘The Role of Belief Systems and Schemas in Foreign Policy Decision-Making’. Mershon International Studies Review. the conflation of states and individual policymakers’ interests is a fundamental impediment to advancing our understanding of world politics. schema theory and cognitive mapping. the analysts who attempt to answer why-questions on the state’s foreign behaviour correspondingly need to give a considerable amount of attention to the policymakers (for example. Because real agents in international relations are human policymakers who have a capacity to consciously act and. See also. to attempt to realise their intentions. they must. 15:1 (1994). The Ecological Perspective on Human Affairs (Princeton.(Meta-)theoretical rationales 7 It seems thus clear why IR researchers need an approach that posits an individual human agent (that is. but rather that they burn down the forest and then model only the charred stumps that remain. ‘Is There Method in Our Madness? Ways of Assessing Cognition in International Relations’. My criticism of structuralist or quasi-structuralist types of analysis is not so much that they are looking at the forest not the trees. For a useful exploration of subtle differences among them. 17–33. their beliefs about politics or perceptions of the reality which are often ignored or assumed away by structure-oriented approaches). see Michael D. Harold Sprout and Margaret Sprout. the IR researchers concerned with why-questions about the state’s external behaviour need to begin their inquiry by animating an abstract concept of a state through replacing it with those human agents who consciously act in the name of the state. As Daniel Little has 29 30 They include cognitive consistency theory.30 Put simply. In this vein. pp. pp. Rather. Therefore. especially their perceptions and beliefs. 1965). Nor are their motivations always to be crammed into socially constructed identities. Young and Mark Schafer. and tend to operate within their own ‘psychological environment’.

what we ought to do in the first place of causal inquiry into the state’s external behaviour is to rid ourselves of the troublesome abstraction ‘state’ while instead giving human policymakers the status of a real agent in international relations. if we neglect the external structure and its constraints – this may result in analyses that depict policymakers as free agents with an almost unrestricted menu of choices. not to assert that the agents’ psychological predispositions are the only (or the most powerful) causal source.8 Yong-soo Eun noted – and at a common-sense level – I believe that we should accept what he calls ‘trivial’ individualism. we need an approach that posits a human agent as an analytical category in its own right. which in turn logically allows us to delve into their individual intentions as well as the common role of the policymakers in a position of authority to commit the resources of the state. Human agents do not make decisions in a vacuum but in a certain structural or contextual setting. that my discussion above by no means implies that the whyquestion as to the state’s external action can be explained solely in terms of the actions and intentions of individual human agents. that only individuals ‘act’ in the literal meaning of the term. Varieties of Social Explanation (Boulder: Westview Press. . countries do not act. Stated differently. the researchers must also examine the manifold contextual or structural factors and conditions associated with a supra-individual level in order to comprehend the dynamics and in particular the limits of the policymakers’ perspectives and choices. ‘to say “the US intervened” is part of our everyday language. (Meta-)theoretical rationales for a multicausal analysis attentive to both agential and structural sources of the state’s foreign behaviour Note. then the causal mechanisms we need to provide to build a valid explanation of the actions must ultimately pass through specific individuals. limited only by the scope of their own ambitions. Hence. p. p. our causal enquiry into the state’s external action must be begun by treating idiosyncratic human policymakers as real agents in world politics. If we take what is perceived or believed by policymakers as the only causal/explanatory factor of the state’s foreign behaviour – put otherwise. and may thus constrain and/or facilitate (limit and/or enable) the decisions and 31 32 Daniel Little. people act. To this end. As such. ‘The Power of Human Cognition’. 47. but the enquiry should not end there. A real agent in international relations should be conceived as human beings working in the name of the state. and thereby offer fuller and more satisfying explanations of the complex reality underlying foreign policy and international relations. in order for us to establish a plausible basis for taking the intentions or beliefs of the pertinent human agents as one of the important causal factors. [Yet] in reality. 1991). Rosati. the structure of the international system or the societal context provides frameworks for interpretation and conditions for action.’32 If the state’s actions operate through individuals. 183. I consider that although it is absolutely crucial for foreign policy researchers to discard the idea of the state as a metaphysical abstraction looking below the state level of analysis to the actual policymakers involved.31 To borrow Jerel Rosati’s analogy. however.

1968). Agents.]’. Structures and International Relations. Thinking about the sources of foreign policy behaviour in the ontological sense The structured nature of international reality is the product of the interaction of agents. how supportive the spectators are of the game) and. Every social act or event is. 13. This is much like the relative indifference of a corporation as to just who its employees are. as long as there are a sufficient number of persons with sufficient qualifications in the proper relationships to each other and to their tools and instruments.33 Hence either agential sources or structural sources are alone not enough to account for social causality. the properties of structures are greater than the mere aggregations of individual wants. . In the simplest terms. what the weather is like on the day the game takes place. In this respect. we should not assume that the structure is simply reducible to individual agents. Max Weber. In Max Weber’s parlance. above all. equally important. Speaking more fundamentally. Causal mechanisms for social actions must thus pass through individuals but may not be reduced to them. indeed. 44:4 (2000). the structure is seen to have a set of properties that cannot be defined solely in terms of the properties of individual agents. ‘no agents. The Systems View of the World: A Holistic Vision for Our Time (Hampton Press 1996). beliefs. it is impossible to analyse a football game in purely individual terms: there are rules (for example. 98. ‘only possible insofar as the conditions for action exist as well as the agents who act’. p. p. and which 33 34 35 36 37 Heikki Patomäki and Colin Wight. Hay. the agential elements (for example. Specifically. beliefs or intentions of individuals) are not the only existing real sources linked to causal mechanisms for the actions of individuals and (by extension) states. p. conditions (for example. . then from where? Ex nihilo?’ Put simply. Although the structured nature of world reality derives from the interaction of the agent. Economy and Society (New York: Bed Minster Press. p. players cannot use their hands while they play a game). Wight. pp. the properties of agents and of structures are both relevant to explanations of individuals’ social behaviour. and actions. Political Analysis. ‘After Post-Positivism? The Promises of Critical Realism’. Ervin Laszlo. . If we think in the ontological sense. this line of argument becomes clearer. 231. settings (that is. ‘if not from the interactions of agents. any professional football game cannot take place without a football pitch) within all of which individual players find themselves.36 and once in place.37 In short. which is irreducible to their individual parts. structures exhibit a certain uniqueness of characteristics as wholes. one might ask. but an ‘organic’ one which displays constant interaction and coalescence. International Studies Quarterly. Such entities possess their own characteristics as wholes which individual agents are conscious of. 31–7.35 But.(Meta-)theoretical rationales 9 choices of individual agents. the former cannot be understood solely in terms of the latter – since the relationship between agents and their environment in the social world is not a mechanistic one. then no structures’. To give a simple analogy. social structures are ‘the resultant and modes of organizations of the specific acts of individual men . the properties of the structure equal the properties of its parts plus the relations of the parts within the structure.34 Indeed. 125.

rather straightforward: causes of the state’s behaviour in world politics can be both structural and agential. personal traits. 17:4 (1991). then. A Realist Theory of Science (Hassocks: Harvester Press. See also. the fact that both individual and supra-individual sources work in tandem with each other qua causes influencing world political activities renders a multicausal approach to the explanation of the state’s external behaviour plausible and necessary. beliefs. This point urges us to open up wide avenues in our approaches to the study of the state’s external behaviour suggesting the certain utility of seeking to combine the analysis of structure and agent and of being sensitised to the complex interactions between the two in any given situation. Put differently. The key point is that in the social world both agents and structures are necessary for any act to be possible. the state’s foreign policy outcomes. It is. human agents who play a football game or make history: whatever the context. Thus it is essential to consider human decision-makers’ perceptions. Martin Hollis and Steve Smith. pp. pp. rather than deterministic forces. but not reducible to each other. There must exist something – external and contextual factors and/or conditions – that can be perceived by human decision-makers in order to make foreign policy (actually. Review of International Studies. ‘Beware of Gurus: Structure and Action in International Relations’. 1979). Central Problems in Social Theory: Action. and the like in the study of foreign policy and world politics.39 Ontologically speaking. so the concept of ‘influence’ does have a non-mechanistic meaning. then B’ manner. the causation in world politics appears far beyond one-dimensional or unidirectional features. there would be no foreign policy or international relations. 20:3 (1994). This means that a variety of structural and agential (material and ideational) things or conditions can (and should) be recognised as causes of. Bhaskar.10 Yong-soo Eun thereby constrain how they play and thus affect the outcome of the particular game that they play. after all. states’ external actions occur due to existence and gatherings of human policymakers and the structural conditions with which their nations are confronted. 1978). the outcome is not determined by the structure of the situation itself. after all. What this puts forward is. they share the common ontological core: structures and agents in the social world are internally related.40 while instead 38 39 40 Here I draw upon the insights of the ‘critical realist philosophy’ of Roy Bhaskar (1978) and the ‘structuration theory’ of Anthony Giddens (1979). See Giddens. or at least as possessing causal capacities in relation to. In this regard. 393–410. to interact with other actors at all). Although they draw different implications from the relationship between structure and agent. This is problematic especially in the causal study of the state’s behaviour in global politics given the fact that the state’s foreign behaviour invariably derives from multiple sources operating in . Structure and Contradiction in Social Analysis (London: Macmillan. 241–51. Review of International Studies. in a sense. since they are ontologically real objects (causes for actions) and interrelated. the concept of cause in IR needs to be understood as ‘influences’ which do not have deterministic connotations: it is quite common for us to talk about someone exerting their influence but to no avail. if there were no external and situational sources.38 Implications of the discussion thus far for international relations are. that IR researchers in quest of valid and fuller explanations of why-questions as to foreign policy behaviour reject a dichotomised view of the relationship between structure and agent. I am against the position put forward by Martin Hollis and Steve Smith in ‘Two Stories about Structure and Agency’. In this vein. However. Their proposition that ‘there are always two stories to tell about the agency-structure debate’ is an epistemological stance which is trapped into the positivists’ narrow and mechanistic understanding of cause in which causes are treated as deterministic (‘pushing and pulling’) forces working only in a ‘when A.

have structuralist tendencies which privilege external conditions in relation to causal explanations of foreign policy actions. Charles F. 2000). in the end. International Organization. There are complexity and variety intrinsically involved in world political causality. Disarmed Democracies: Domestic Institutions and the Use of Force (Ann Arbor: University of Michigan Press. in Barry Farrell (ed.41 As discussed earlier. even those studies which illuminate internal and external factors often fail to posit a human agent as an analytical category in its own right – rather they concentrate on the role/position that human agents (for example. Fearon. Yet these studies are more in the nature of exceptional examples rather than representative cases in a generally accepted research method in the field of IR. pp. Haney (eds). ‘How Do Leaders Make Decisions?: A Poliheuristic Perspective’. James N. Rosenau. as already mentioned. and tend to see structural and agential sources as hierarchical or discrete. American Political Science Review. See. NJ: Prentice Hall. 1995). what structure theorists argue for is that ‘an adequate international relation theory [in a methodological sense] must be more structure based than agent orientated’. David Auerswald. See. For this reason. favouring agent-based explanations over structural ones.K. pp. 39). Institutions. (neo-) liberalism. namely (neo-)realism. Michael Brecher. the mainstream approaches employed by IR scholars. Although we may have different hunches about which causal factors are most important. but also necessary. Unfortunately. pp. IR scholarship still lacks multicausal approaches and accounts finely attuned to complex and interrelated relationships between human policymakers and the international environment. 41:3 (1987). causes are understood here as ‘influences’ which do not have deterministic connotations. Patrick J. 427–60.44 combination: recall the ontologically interrelated relationship between the agent and the structure. ‘Diplomacy and Domestic Politics: The Logic of Two-Level Games’. and constructivism. I discuss in detail in the next section. 247. Hermann.(Meta-)theoretical rationales 11 considering both the structural constraints on or opportunities for individual policymakers in realising their intentions and also the policymakers’ intentions and actions which give rise to such structural consequences. as already mentioned (see fn. Images. Robert D. To be sure. for example.43 Agent-centred theories reverse this prioritisation. p. in Laura Neack. or the material and the ideational as two discrete or exclusive entities. 88:3 (1994). 41 42 43 44 . Certainly. ‘The Agent-Structure Problem in International Relations Theory’. The Journal of Conflict Resolution. ‘Pre-theories and Theories of Foreign Policy’. this is not to say that the mainstream IR theories completely disregard domestic politics and the internal factors: even Waltzian structural realism acknowledges ‘a modest role for domestic factors’. Alexander Wendt. 577–92. ‘Domestic Political Audiences and the Escalation of International Disputes’. Bruce Bueno De Mesquita and David Lalman. however. James D. Furthermore. Interests. The Foreign Policy System of Israel: Setting. 3–13. and Information (Princeton: Princeton University Press. and the studies employing them are not only useful. 340. Approaches in Comparative and International Politics (Evanston: Northwestern University Press. national leaders) occupy – or retain polarised views of international relations that express the agential and the structural. or as ‘game tables’ (in Robert Putnam’s words). International Organization.42 Yet. for studies of this kind. the domestic and the international. In this sense. 1992). Jeanne A. 1997). Hey. 1972). ‘Epilogue: Reflections on Foreign Policy Theory Building’. As regards how to conduct an empirical analysis with the nonmechanistic and broad conception of cause. War and Reason (New Haven: Yale University Press. 42:3 (1988). there have been integrated studies on foreign policy and attempts to develop integrated theoretical frameworks. Helen Milner. 1966). Alex Mintz. and Process (New Haven: Yale University Press. 48:1 (2004). we should give (potential) causal status to both of them on a pari passu basis. p. Putnam.). Foreign Policy Analysis: Continuity and Change in Its Second Generation (Englewood Cliffs. theories or analytical schemes that can accommodate various types of causes associated with both agential elements and structural conditions.

Christopher M. although these scholars often agree that multileveled. Explaining Foreign Policy Ole R. What is more. Eugene Wittkopf and his colleagues comment that a single-cause or singledimension explanation is still ‘the widespread impulse’ in the field of IR because of its ‘simplicity that most of us intuitively find satisfying’. Put simply. Laura Neack. than it should do. I believe that fruitful integrative theories of foreign policy behaviour can only begin to flourish when a nuanced and complex view of world politics and of the relationship between structure and agent in states’ external behaviour – which might be referred to as ‘a rich/bold ontology’ regarding foreign policy behaviour – is matched by the development of detailed and explicit guidelines on how to traverse the bridge that connects the insights of that rich ontology to the empirical research necessary to make claims about the real world of any one moment. Martin Press. p. for example. methodological and epistemological issues) may hinder our efforts to move forward. . 18–9. Kegley. Hudson. The New Foreign Policy: US and Comparative Foreign Policy in the 21st Century (Rowman and Littlefield 2003). see Ian Hall. For a similar critique. they tend to focus more on why we should go for a multicausal analysis rather than on how to go for it. pp. they do not go beyond this statement of preference to indicate how one might (and should) conduct analyses that would lead to confirmation of this. respectively. however. 2008). Jensen.48 In other words. more one-dimensional. ‘Theories of International Relations and Foreign Policy: Realism and Its Challenges’. multicausal explanations of international relations are preferable. Milja Kurki. 11:4 (2009). Controversies in International Relations Theory (New York: St. pp. 629–30. 1995). our moves towards multicausal analysis have not yet gone far enough in that the currently dominant approaches to causation in IR are not finely attuned to the complex interplay between human agents and international structures. Pattern and Process (Thomson Wadsworth 2008)..47 do not explicitly articulate how to formulate and utilise such a complex and broadened multicausal approach. Wittkopf. for example. Equally. Jones. Wittkopf et al. in Charles W. See.46 as well as scholars who attempt to broaden the concept of cause in an ontological sense as a solution to the structure-agent problem. American Foreign Policy. Kegley (ed. 189–90. ‘Foreign Policy Analysis’. Developing the methodological and epistemological positions and rules commensurate with a multicausal analysis It needs to be emphasised at the outset of this section that my advocacy of the multicausal analysis underpinned by the rich ontology by no means indicates that 45 46 47 48 Eugene R. American Foreign Policy.12 Yong-soo Eun In short. pp. 19. See. Charles W. those who admit and even emphasise the importance of a multicausal analysis that accommodates both (human) agential and structural factors. Holsti. ‘What Causes What: The Ontologies of Critical Realism’.45 Structuralist theories and individualist theories tend to be sensitive to external factors and internal factors. I feel that this lack of sufficient attention to the latter (that is. they tend not to be sensitive to the fact that the reverse line of causation also holds. Causation in International Relations (Cambridge: Cambridge University Press. The consequence of this is to flatten out the rich and diverse textures of world reality: the world appears as far less interesting.). International Studies Review.

p. Sociological Reasoning (London: Palgrave Macmillan. 1996). Robert Jervis comments that even scholars who use large N clusters of data ‘often look at many cases to see if a proposed generalisation fits the data. 83. to establish an integrated multicausal theory and answer the question of why states behave the way they do in world politics. Kurki. what the researcher concerned with why-questions about foreign policy behaviour needs to attempt first.(Meta-)theoretical rationales 13 one must (and can) observe all the possible variables and relevant evidence. Those who are at the deductive end of the spectrum are apt to see only theory-laden factors and confine their findings to post hoc rationalisation: the purely deductive logic of inference places its emphasis upon the process of logical (theoretical) deduction. any non-observation-based statements could be rejected as ‘meaningless’50 or ‘mere fictions’. ‘After Post-Positivism?’. and the other that attempts to make prescriptive generalisations deduced from simple assumptions about the complex world. See Stones. she or he must combine the ontological boldness that the multicausal approach possesses with caution and openness in relation to epistemology and methodology. Instead. our causal analysis of foreign policy needs to commence by making richer (in other words. this position has significant methodological implications. if one prefers to work with a rich and complex multicausal approach. p. Rejection of the absolutism of deductive and inductive logic First of all. but this implied knowledge is not matched by the levels of such confidence. then – while bearing in mind that the causal relationship between explanan and explanandum in world political phenomena is multicausal – is to find potentially important factors that might impinge 49 50 51 Rob Stones notes that those engaged in a strict and formal form of deductive logic. As the reader may notice.49 On the other hand. [But] this is a form of confirmation.51 These strict forms of logic are problematic in terms of epistemic range and nuances. Let me clarify these points further. those who are at the opposite end of the spectrum tend to concentrate on what can be easily measured and readily observed phenomena rather than what might be theoretically important and the more subtle and deeper social and political forces: from the point of view of the purely inductive logic favoured by strict empiricists. Patomäki and Wight. 217. we should avoid both these extremes of logic – one side that values the (supposedly) neutral assessment of all the cases and empirical data without any theoretical assumptions. less parsimonious) theoretical or explanatory assumptions with a perceptive emphasis on the role of the empirical investigation in testing the consequences deduced by the assumptions. ‘Pluralistic Rigor’. 146. Thus. If accepted. I argue that any analyst should reject the absolutism of both deductive and inductive logic. These extremists make similar but opposite types of mistakes when they analyse what they observe. p. The real virtue of the multicausal approach has nothing to do with how many factors it examines. not the discovery of new facts’. p. . See Jervis. somewhat ‘omnipotent’ manner about events. Causation in International Relations. The multicausal approach put forward here should not be equated with an approach of ‘everything matters’. Going a step further. considering all the cases. while giving insufficient attention to the means by which theoretical propositions are tested empirically. Furthermore. known as ‘deductive-nomological’ modelling tend to write in an all-knowing. 45.

More specifically. . p. which in turn entails serious weaknesses in explanations of the dynamic relationship between agents and the external environment. as analysts cannot survey impartially all the possible variables. Employing a multicausal approach in a tentative and flexible manner The logic here requires. but also flexibility in an epistemological and methodological sense. it is crucial not to see structural and agential factors as oppositional or discrete. causes of states’ behaviour in the international system can be structural and agential in both material and ideational senses. let us recall the fact that the nature of the ontological object has an important role in defining which ways of knowing are appropriate to it. 52 Walter Carlsnaes. because the idea that explanations that emphasise different causal factors should be conceived as being in competition with each other leads us to settle on a truncated and impoverished view of the rich and complex textures of world reality or to impose unnecessary restrictions on the range of possible causes of social phenomena. IR researchers must be open pari passu to both quantitative and qualitative methods using hard data – from government statistics. 256. they must deductively come up with and select out some major causal factors on the basis of their own explanatory assumptions. In this vein. ‘The Agency-Structure Problem in Foreign Policy Analysis’. As discussed. As such. Yet. before and in the process of selecting from among the many. Again. no one in ‘today’s postbehaviouralist world’ would cavil unduly about this way of reasoning. and there are crucial factors that cannot be observed through ‘sense experience’. given the complex reality underlying a nation’s foreign policy and the ontologically interwoven relationship between agents and structures in international relations. in my view. since the foreign behaviour of the state almost invariably results from multiple sources.14 Yong-soo Eun upon and affect the observed phenomena on the basis of having rich theoretical assumptions and explanatory expectations. The important point here is that such investigation should be made on the basis of both quantitative and qualitative forms of empirical analysis. the analyst must think about and analyse observed foreign policies or world events in multicausal terms. for example – and soft data – from interviews or speeches – so that they can examine the causal power of the material and ideological factors associated with the complex interplay between structures and agents. the researcher should employ a tentative and open-ended approach within which not only richness in an ontological sense. And such a mental effort leads us to raise one critical question: how would one know if the chosen factors were really explanatory variables or not? This consequently leads the researcher to discern the explanatory strength of each of the chosen factors and trace their causal links.52 But in this process there are two very important – or ‘unorthodox’. International Studies Quarterly. First. need to be fully accepted and implemented by analysts. then. that in analysing and theorising foreign policy. can be preserved. the way states behave as they do in world politics mirrors the complex and dynamic relationships between the (intentions and perceptions of) human decision-makers and the international environment. Surely. 36:3 (1992). if you like – points that.

such as causal ‘influences’ or ‘capacities’ are preferred here – in lieu of a term. the relative causal powers and the causal links of the concerned factors are not predetermined. the idea that causal status ought to be given to both structural and agential factors does not imply that the causal capacities of the factors will necessarily be drawn upon or activated. ‘Interdependence. Loose-knit deductive reasoning in the study of foreign policy What my arguments thus far suggest is. attempt to search for causation beyond correlation – must perforce examine if the causal capacities of the factors to which the potential causal status is conferred have been activated through quantitative and qualitative 53 54 55 See. ‘the interactive and contingent nature of global politics limits the extent to which complete and deterministic theories are possible’. As Jervis observes. International relations and foreign policy are surely dynamic and full of contingencies. NJ: Princeton University Press). operate in combination to a varying extent. for a fuller exposition of this point. they. 105–11. at the first step of causal inquiry into the state’s external behaviour. . As they exist in ‘open [social] systems’ where the constant interactions with each other occur. This implies that we do not know if the factors chosen are actual explanatory variables of an observed phenomenon until after we have carried out empirical investigations into the explanatory strength of each of the several factors chosen and trace their causal links. pp. The key point here is that structural and agential sources of foreign policy should neither be deemed exclusive or separate nor be granted explanatory priority a priori. Method in Social Science: A Realist Approach (London: Hutchinson.55 Therefore. States. Institutions. Robert Jervis. for this reason. that in order to fully answer why-questions as to foreign policy actions. the intended consequence may not come out. Furthermore. they possess causal capacities as potentials. p. there is a clear need to employ a multicausal and open-ended approach in which while the causal status of both agential and structural (material and ideational) factors are accepted. may behave extremely variedly because ‘states can have varied preferences in similar strategic situations’. 1984). pp.(Meta-)theoretical rationales 15 A second – in my view more important – point is that at this stage of picking and choosing. and material and ideational factors – constantly vary from situation to situation. even when they are in the same situation. Bhaskar. 162–6.54 This calls for the analyst to have a flexible and pluralistic epistemological and methodological position. IR researchers who seek to obtain a plausible answer to the question of what really happened causally in the observed phenomenon – in other words. Consequently. the terms. rightly in my view. 44. Andrew Sayer. and the Balance of Power’. after all. Papayoanou. indeed. There can be no doubt that the actual causal influences and causal directions of variables – agential and structural. p. System Effects: Complexity in Political and Social Life (Princeton. 295. there is no reason a priori to reward extra credit to any particular factor. even when a causal capacity that objects or agents possess is exercised. it is satisfying to have only the ingredients of an analytical approach – which will assist us to undertake certain mental processes in deciding what it is that should be observed.53 Put simply. such as causal ‘forces’ which has deterministic connotations. although analytically separable. A Realist Theory of Science.

‘The Causal Nexus between Cognitive Beliefs and Decision-Making Behavior: The Operational Code Belief System’. standing on a rich ontological platform formulated prior to application of the approach. speeches.58 And if the test confirms the explanatory power. and then causally reconstruct the world of the observed phenomenon with the factors that receive empirical support concerning the explanatory power. 95–124. it can be referred to as ‘loose-knit deductive reasoning’. Deborah Welch Larson. what we should do is to examine the causal influence of the chosen factor – the subject’s political beliefs. pp.16 Yong-soo Eun forms of empirical investigations into the explanatory weight of the factors. George. 15:1 (1994). Falkoswki (ed. for instance states’ material interests. Bush) as one of the causal factors56 in searching for an answer to this why-question. This may sound too complicated. see Alexander L. for example. 13:2 (1969). in L. Pre-theories. Political Psychology. I describe the way of reasoning put forward here as a logical process in which one finds multiple causes using a flexible epistemological and methodological approach. ‘The Role of Belief Systems and Schemas in Foreign Policy Decision-Making’. as a potential explanatory factor. this way of reasoning enables the observer to discern the real causes of an observed phenomenon and reconstruct its causal processes in a systemic and clear manner. suppose that we are puzzled by why a particular nation (for example. does not yield significant explanatory weight. 190–222. pp. 1979). 17–33. then the chosen factor (Bush’s political beliefs) becomes an indispensible part of the multiple causation of the observed phenomenon (that is. Brecher. in Alexander George’s terminology. the US) made a certain external action (for example. The Foreign Policy System of Israel). set aside in conceptual brackets are considered. a more accurate and satisfying causal explanation (of why the US decided to go to war against Iraq) can be developed. but a multicausal and open-ended approach is not only preferable to a monocausal and complete approach given the reality of the sheer complexities and dynamics that pervade world politics. . then it is passed over and other potential factors. my intention in this article is to elaborate on how to utilise a multicausal approach and to make clear what ontological and epistemological assumptions and premises should be made before using it. we need to first infer the subject’s (Bush’s) political beliefs through a close investigation of his public and/or private statements – for example. There is no need to recount them here. While isolating or setting aside the other potential causal factors for the moment.S. Consequently. George W. Rosenau. In order to do so. George. many possible causal factors have already been described in previous works (see. 56 57 58 Here no attempt is made to provide a comprehensive list of all possible causal variables of states’ external behaviour. Alexander L. interviews. International Studies Quarterly. we can determine the explanatory power of the subject’s political beliefs – by seeing whether the final decision made by him was consistent with his political beliefs: this is. but is also manageable if a method of isolation and exclusion is invoked in conducting a multicausal analysis. we will be able to discern if the causal capacity of Bush’s political beliefs was activated – that is to say. Or. ‘The “Operational Code”: A Neglected Approach to the Study of Political Leaders and Decision-making’. the ‘congruence procedure’. See also. stated more succinctly. For classic illustrations of this method. and that we choose the political beliefs of key policymakers (for example. Psychological Models in International Politics (Boulder: Westview Press.57 Having inferred his political beliefs. But if the investigation indicates that the political belief. press conferences – which display his views on the nature of political and social life. In turn. As noted.). pp. the US war against Iraq). For example. go to war against Iraq).

Milja Kurki. there are shortcomings to this open-ended multicausal approach: for instance. p. In short. see Waltz. Fla. incidents and time frames pursuing parsimony in such a way as ‘to make them usable in the present as guides to the future’. 195. p. we should not reward extra.] a theory is false or not [. Hugh Ward. . emphasis in original. In this vein. 8 (2005). it has many loose ends as compared with conventional deductive approaches. the work of Gary King. and predictions relating to the phenomena under investigation.59 In effect.’61 59 60 61 For example. and as such it hardly provides any prediction or generalisation about the future courses of foreign policy behaviour. agential and structural (material and ideational) factors operate in combination to a varying extent: agential and structural factors are interacting to produce foreign policy outcomes. ‘Rational Choice’. Through a theory. these limitations should not obscure the underlying contributions of the multicausal and open-ended approach to our understanding of world politics. To repeat. we can present descriptions. ‘The Relationship between Theory and Policy in International Relations’.] than how much of the world the theory can help us explain. Keohane. .(Meta-)theoretical rationales 17 The crucial point here is that while we employ a multicausal approach. Going a step further I argue that such a complex and tentative approach – both as a useful mode of explanation and as a progressive model of theory construction – is well worth being given favourable attention in IR scholarship. stressing that ‘the question is less whether [. Theory and Methods in Political Science (New York: Palgrave McMillan. Structuralist and rationalist theorists claim that a good theory must generalise across cases. ‘Causes of a divided discipline: rethinking the concept of cause in International Relations theory’. this view constitutes today’s mainstream standards for judging theoretical contributions and progress in the study of international politics. . . credit to any particular variable. 32:2 (2006). 101. Gary King. which emphasises the presence of several causal variables. a multicausal approach to the study of the state’s external behaviour needs to be premised on the starting assumption that none of the explanatory factors chosen are more deterministic than others. 1994). it appears to be an ‘inelegant’ alternative as compared with the more parsimonious models. The Annual Review of Political Science. what we need to search for is a compound explanation. explanations. events. and Sidney Verba. Review of International Studies. I believe that the limitations are not as serious as the mainstream approaches in IR might consider. Indeed. Robert O. in David Marsh and Gerry Stoker (eds). which must begin with the important premise that none of the chosen factors are privileged a priori. However. . 2002). pp. The Logic of Comparative Social Inquiry (Malabar. Stephen Walt. Robert Keohane and Sidney Verba – which has ‘become very influential as a guiding light of causal analysis in political science and IR’60 – makes clear that generality is the single most important measure of social science progress. not one particular factor is determinate in the last instance. Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton: Princeton University.: Krieger 1982). let alone fundamental. 26–8. in world politics. For example. Rethinking the role of theory for IR scholars and the mode of construction of IR theory Admittedly.

foreign policy does not occur in a static manner. To borrow Otto Neurath’s analogy. Neurath (Boston: Reidel. and increased seaworthiness is itself learned during the voyage on the open sea. Rather. Thus. that is focusing more on the development of an accurate explanation of limited scope rather than on the development of law-like generalisations in which local differences and contingencies are intentionally or unintentionally swallowed up or ignored.’62 And I believe that continual rebuilding is necessary to increase the seaworthiness of our ship (theory). diverse ideas should be welcomed and 62 Otto Neurath. Yet the task of identifying formidable explanatory variables and scope conditions is unfinished. Philosophical Papers 1913–1946. precisely because there are inherent complexities and dynamics involved in states’ foreign actions in world politics. there is no escaping the fact that causes in international relations do not run in one direction: they vary across time and space. Parsimony itself – as well as generalisations – results from constant testing and refinement of diverse factors and manifold conditions. p. What this suggests to us is. and thus of explanations based on the probabilistic symmetry between explanation and prediction. an epistemologically cautious position. the inclusive and tentative approach has more immediate profit – profit which in the end would provide the foundations for a general theory. trans. This is not to say that the explanation and theory of the state’s foreign behaviour ought not to be pushed to a higher level of generalisation. ‘We are like sailors who have to rebuild our ship on the open sea. because empirical data and logical conceptualisation are as yet inadequate to support such a general theory. there are many unobservable but important causal factors and conditions at work in world politics. The role of theory for IR scholars should be as an evolutionary guide to the empirical exploration of the open social world and as an emergent means toward the accurate explanation of the complex reality underlying world politics rather than as the simplification of it as a condition for the generation of predictive hypotheses. All of this brings into question the notion of prediction and generalisation in the study of International Relations. International relations are full of complexity and contingency. we should be more modest in our aspirations for a theory’s universal applicability. S. then. . 92. The task of identifying crucial explanatory variables which can be applied to a wide range of situations is unfinished.18 Yong-soo Eun However. 1983). Put otherwise. without ever being able to dismantle [and reconstruct] it in drydock. Put simply. Further. after all. variation also exists across different kinds of states and policymakers. as far as the present inquiry is concerned. IR theory ought to be. what this indicates is that a multicausal and open-ended approach makes more sense now. Until we have them – the components of such a general theory – it is probably best to speak of a theory qua a mere starting-point from which the researcher can manipulate a number of variables or of a provisional analytical framework in which several variables are loosely knit. and eds R. Cohen and M. doubtless the generality of social and political phenomena is an open question. No one doubts the number and complexity of factors that influence national action in the international area. able to sensitise the analyst to a wide range of ontological features of world politics and to multiple causation in which more than one cause is involved in the production of an effect.

Rather. as such. Further. ‘analytical’ (not ‘statistical’) generalisation – aimed to modify. analysis of one or of a few cases. generates narrower observable implications. . MA: MIT Press. I argue. but it is not only preferable – if our goal is to arrive at fuller explanations of social and political phenomena – but is also manageable. to say ‘IR researchers need to go for a multicausal analysis’ is not a suggestion that they must examine all possible variables of what they observe. 2005). but this point should not be exaggerated. that is small-N work. until the evidence that shows causation more fully and clearly is sufficiently accumulated. single-case analysis can provide a relatively complete account of the particular phenomenon in question because it enables researchers to focus on a single event or policy issue and study it in depth: it generates a wealth of data. Put otherwise. as illustrating general patterns. True. as discussed. expand and 63 Alexander L. In this way we may cut ourselves off from ontological richness and methodological openness that are the most fruitful way to study complex international relations. and the outcomes produced by such single-case research may not be seen.63 but also because making generalisable knowledge about causation must be supplemented by – indeed. it has to be acknowledged that it is not easy to use such an inclusive approach in the analysis of a series of events. my emphasis is placed on the following idea: we should not give an a priori preference to one or the other type of source of the state’s foreign behaviour – at least in the early stage of the enquiry – because doing so may induce us to suppress uncomfortable pieces of evidence in favour of those more suitable for confirming causal strength of the factors we chose a priori. The real virtue of the multicausal and open-ended approach instead centres on its capability to assist us to avoid becoming preoccupied with a single particular factor or dimension and especially from seeing structural and agential (material and ideational) factors as hierarchical or discrete. George and Andrew Bennett. In effect. there can be no real flourishing of a theory which attempts to explain complex causation. To this end. a theoretical approach within which richness and flexibility can be preserved is necessary. if a method of isolation and exclusion is invoked. The multicausal approach I advocate should not be equated with an approach of ‘everything matters’. needs to be begun from – the detailed causation-tracing evidence to be established through in-depth small-N research. Yet. Having said that. and thus allows the researchers to argue convincingly about the relationships between variables and present a plausible and satisfactory causal explanation of that phenomenon. More importantly. a complex and tentative approach is taxing as against a simple and complete approach. not only because even the most particular claims generate at least some cross-case implications and important inferences can often be drawn from causation-tracing evidence pertaining to one or to a few cases without broad cross-case regression. at first. The status of the multicausal approach put forward here has nothing to do with how many factors it has. This in turn leads to the identification of the causal mechanism that connects to the phenomenon through which really valid generalisations about similar phenomena can begin to be made. analysts preferring multicausal approaches may focus on a single important event or phenomenon.(Meta-)theoretical rationales 19 various factors must be tested. Case Studies and Theory Development in the Social Sciences (Cambridge.

. ‘International Studies in the Twentieth Century and Beyond: Flawed Dichotomies. because a multicausal and open-ended approach leads the observer to ask about and study multiple causes simultaneously. . Synthesis. Moreover. such accumulated evidence and knowledge provides a secure base on which more specific causal patterns can be identified or established. Thus. a multicausal and open-ended approach can be used as a critical building block for making more realistic generalisations and more precise predictions. These specific patterns would initially consist of generalisations of quite limited scope. As a good example here.64 I am not claiming that the ontological position. International Studies Quarterly. Michael Brecher – in his Presidential Address to the International Studies Association in 1999 – commented that ‘intolerance of competing paradigms and models [. and an incremental and integrated path to theory. . 2008). Case Study Research: Design and Methods (London: Sage Publications. Correspondingly. consider Graham Allison’s original study of a single case. 43:2 (1999). not to make inferences about a ‘population’ (or universe) on the basis of data collected about a ‘sample’ from that universe – can proceed through even a single-case study as long as the study has theoretical propositions at the outset of an inquiry.65 I do believe that much the same can be said about IR scholarship of the twenty-first century. Yin. We still need a flexible mentality as an approach to knowledge. And the multiple causes and causal processes elucidated can be fed back into the construction of a body of theoretical knowledge necessary to make valid claims about the real world in which complexity. the epistemological rules and the methodological suggestions put forward here are beyond criticism – only that they are well worth criticism and development.] a closed-mind mentality [. intricacy. because they allow us to add fresh and complementary theoretical thought and complexity-sensitive empirical evidence to the existing IR literature. emphasis added.] the low-value placed by most scholars on accumulation of knowledge’ are the shortcomings in IR scholarship of the twentieth century. . and thus establishing evident causation of an observed phenomenon or action. 15–38. In short. ISA Presidential Address’. pp. it is a better tool. See Graham Allison. Accumulation. 213–5. to move one step closer towards resolutions of fundamental puzzles about world politics.20 Yong-soo Eun generalise theories. Brown. it is the first and most essential step for the progressive accumulation of our knowledge about why states behave as they do on the world 64 65 Robert K. as compared with a monocausal and complete approach. However. Essence of Decision: Explaining the Cuban Missile Crisis (Boston: Little. . pp. Again. Consequently. I am not saying that a multicausal and open-ended approach is the only way in this endeavour. in terms of discerning how the causal links are formed and reifying causal processes. Such middle-range theory would initially consist of more general theory. 1971). it is able to assist us to recognise and explore previously unrecognised sources of the state’s external behaviour. giving rise to the identification of previously unidentified or yet-to-be discovered sources and conditions. the empirical evidence that shows causation of the state’s foreign action more fully and deeply begins to be accumulated. to succeed partially is not to fail completely. the 1962 Cuban Missile crisis which forcefully demonstrates how a single-case study can be the basis for significant explanations and generalisations. and in turn. Michael Brecher. Instead. and variety are pervasive. Nor do I assert that it will solve at once all conceptual or empirical puzzles in the field of IR. I believe that by virtue of the richness and the flexibility that the approach itself possesses.

we are able to better comprehend.(Meta-)theoretical rationales 21 scene. then. . With such progress. A multicausal and open-ended approach – both as a useful mode of explanation and as a progressive model of theory construction – is. a challenge worth taking up. explain. and thereby prepare for various eventualities in the international arena. predict.