What is international relations (IR) theory? Is it an arcane academic enterprise or relevant to the real world? What do we mean by theory? What kinds of knowledge do theories seek or represent? How do they stipulate it should be found? How should we evaluate any resulting knowledge claims? What do answers to these questions tell us about the theory project in IR, and more generally in the social sciences? I address all of these questions in this book and set of companion videos. The videos can be accessed free of charge at www.nedlebow.com.
I cover epistemological ground with which you may be familiar, but perhaps not. I lay bare and query the assumptions of positivist and interpretivist approaches to knowledge in general and to IR in particular. I identify and explore the most important inner tensions of these epistemologies and question the assumptions they make about the world. I identify problems that need to be addressed, suggest some possible strategies or reformulations toward this end, and discuss how adherents of the two epistemologies might learn from one other. Above all, I am interested in the extent to which IR theories can tell us something useful about the world.
This book is about theory in IR, not about theories of IR. This distinction is an important one. It focuses on theory, what it is, what it purports to be, and how it organizes our understanding of the discipline and the world. It engages some IR theories and research programs, but only as illustrations or for the insights they offer into the questions I have posed. It differs significantly from most other books on IR theory. They are organized around paradigms, theories, or substantive areas of inquiry.1 Some books that claim to be about theory are really methods texts that insist on shared standards and make an uncritical pitch for those rooted in neopositivim (defined in due course).2 Methods are metaphorical tips of icebergs. They rest on a much larger structure that may as well be submerged as it attracts so little attention, as least in the neopositivist literature. I look beyond methods and paradigms to “epistemology,” which derives from the Greek word epistēmē, meaning knowledge. I identify the different kinds of knowledge people seek, how they believe it can be discovered, when and why they believe they have found it, and what they think it is useful for. We need to articulate and interrogate the epistemological foundations of all relevant research traditions so we can proceed in an informed way to theories, methods, and empirical research.
We can evaluate theory internally and externally. Is the theory logically consistent, are its terms well defined, is its scope carefully specified, and is it subject to empirical evaluation? If it meets these criteria, we can go on to ask if the theory or propositions derived from it offer good explanations or predictions or, alternatively, say something else valuable about the world? External criteria for validation are highly controversial both within and across positivist and interpretivist epistemologies. Positivists tend to believe in objective criteria for establishing truth claims. Interpretivists deny this possibility. Many, perhaps most, acknowledge the importance of protocols for conducting and assessing research and the possibility of achieving an consensus about both kinds of protocols among the community of researchers. Exploring differences between and within these epistemologies offers insights into the nature of and impediments to disciplined inquiry.
Theories of all kinds rely on reason in the form of logic. Reason is generally treated as universal and unproblematic, especially by positivists. It is a meta-concept, as it is foundational to all other concepts. Following Max Weber, I contend that reason is historically and culturally specific in understanding and use. It is closely related to values because people use different kinds of reason depending on the ends they seek. Some theories assume a high degree of instrumental rationality – one kind of reason – on the part of actors whose behavior they study. We confront an empirical question here, not one to which an answer can merely be assumed. To construct good theories, we need to know more about the kinds of reason actors employ, in what circumstances, and the degree to which they use these formulations in a consistent manner. Some social scientists insist that theories are free to incorporate any assumptions they want, including that of instrumental rationality. I devote Chapter 12 to this controversy.
The other concept central to theory is cause. It is understood in different ways by different research programs. They also put different values on its utility. Some assume cause to be the cement of the universe and the goal of inquiry. For others, it is a human artifact that we use with varying degrees of success to make sense of the world. Some research programs finesse cause or dispense with it altogether. I devote Chapters 13 and 14 to these different understandings of cause, their implications for knowledge, and the problems to which they give rise.
A century of IR theory might be regarded as an experiment from which we might learn something about the nature, appeal, and life cycle of research programs and the theories they have spawned. It also has the potential to teach us something about the relationship between theory and praxis. Theories come and go. Epistemologies also rise and fall in appeal, but over a longer time period. Both kinds of turnover reflect changing conceptions about what theory is or should aspire to be. These conceptions are closely connected to the ends we want theory to serve. It makes little sense to examine the rise and fall of paradigms and theories in isolation, as so much of the literature by IR scholars does with their fascination with so-called first, second, and third debates unless they are linked to more fundamental assumptions about knowledge and its purposes.3
Although we cannot agree about what theory is or should be, there is no escaping the fact that it has become more central to our field. It might be said to help constitute it because it is what separates us from foreign policy analysis and history. Foreign policy analysts and historians may use theories but IR creates them. There is nevertheless a certain irony in our reliance on theory to demarcate and legitimate IR because so many of our theories are borrowed from other fields and disciplines and so much IR is deeply involved with the analysis of foreign or national security policy. To the extent that IR theory is imbricated with theory in other fields, it offers an opportunity to address the problem of theory more generally.
Questions, Tensions, Controversies
International relations theory is often described as a twentieth-century phenomenon. The conventional date for its birth is the immediate aftermath of World War I when the first chair in the subject was created at the University of Wales in Aberystwyth. In reality, people had thought and written about relations between political units since at least the time of Thucydides (460–400 BCE). Kautilya, Niccoló Machiavelli, Jean Bodin, Thomas Hobbes, John Locke, Immanuel Kant, Jean-Jacques Rousseau, and Max Weber all stand out in this regard.4 What changed in the twentieth century was the creation of a discipline devoted to the study of IR. In its effort to establish itself, it claimed these thinkers and others as de facto IR theorists and representatives of a long-standing tradition – but in truth, a largely imagined one. Beginning with realism and liberalism, every IR paradigm would create a lineage for itself and co-opt distinguished thinkers from the past as its founders. By doing so, advocates of these paradigms hoped to add legitimacy and luster to their projects.5
Another distinguishing feature of IR theory is its policy focus. This too is a function of its origins. David Davies endowed the first chair in Aberystwyth as a memorial to the students killed and wounded in World War I. He named it after Woodrow Wilson because he embraced the American president’s vision of restoring and maintaining peace through the League of Nations.6 After World War II, IR scholars were equally committed to finding ways of preventing war and addressed what they thought were its immediate and underlying causes. This commitment is evident in the research and writings of, among others, Hans Morgenthau, John Herz, Ernest Haas, Ernst-Otto Czempiel, Karl Deutsch, Alexander George, Ole Holsti, Stanley Hoffman, Hedley Bull, and the English School. They produced landmark studies of war, war prevention, crisis management, nationalism, public opinion, and regional integration. Many of their research programs still find contemporary resonance.
More than other social sciences, IR is engaged with practice, in analytical and hands-on ways. Theorists aspire to develop insights into the causes of international conflict and cooperation in security, economics, immigration, and other substantive domains. Most hope their research will not only advance our knowledge but also inform the policy of leaders, states, and regional and international organizations. Ethics also enters the picture. Many IR scholars are committed to developing norms for foreign policy, drawn for the most part from theories of ethics. I, for one, have mustered empirical evidence to demonstrate that ethical policies are more likely to succeed than unethical ones, and vice versa.7
Students of IR are nevertheless of two minds about policy engagement. I think it fair to say that for most it is what makes our field relevant and attractive. However, some prominent theorists, including Quincy Wright, Morton Kaplan, and Kenneth Waltz, questioned the coupling of theory and policy. They might be considered purists, committed to theory as an end in itself and to maintaining a firewall between international relations and foreign policy. For Waltz, international relations is a system-level phenomenon, and the goal of IR theory is to capture the character of different systems and their consequences for war and peace. By his own account, his theory has no practical utility.8
Even those scholars who want their research to inform foreign policy are generally keen to distinguish IR theory from foreign policy analysis, area studies, and journalism. PhD programs have tried hard to distance themselves from these enterprises, and many professors, especially in the United States, have been wary of them, if not downright hostile. At the same time, many IR theorists, including such luminaries as E. H. Carr, Hans Morgenthau, and Raymond Aron, had active lives as journalists or governmental advisors. Noted scholars like Harold Lasswell, Quincy Wright, John Herz, and Joseph Nye Jr. held government office. Not a few contemporary IR scholars are active bloggers with sizeable followings. The New York–based Council on Foreign Relations has for many decades offered fellowships to faculty to spend a year in Washington, DC, working in the government. I was the recipient of one of these fellowships in 1974–75.
A related controversy concerns the extent to which the profession of IR should engage in politics. Hans Morgenthau frequently maintained that the goal of IR scholarship was to “speak truth to power.” He was an early and outspoken opponent of the Vietnam War, debated National Security Advisor McGeorge Bundy on television, and spoke at antiwar rallies.9 He was punished by pro-war conservatives who lobbied hard to deny him the presidency of the American Political Science Association (APSA) and pass a resolution that the APSA was a professional association that should avoid taking any political position.10 The conservative hold on APSA led to the creation of a breakaway organization, and some years later to a reconciliation.11 In the decades since, IR theorists have not been shy about speaking out on political issues. Many signed petitions against the 2003 invasion of Iraq or spoke out publicly against it. Some of their British colleagues did the same. IR scholars in many countries have tried in this century to have their voices heard on a range of political questions.
Morgenthau’s plea for independence and courage has been taken to heart by many IR scholars. A few, as he feared, have been co-opted by governments or national security establishments and have become partisans of national or ideological causes. Still others act this way because they are embedded in their nation’s political culture and socialized into accepting a particular view of the world and their nation’s role in it. Simon Reich and I contend that this is true of many American realists and liberals who maintain that the United States is a hegemonic power and deserves special privileges because its leadership is essential to preserve global security and economic intercourse.12 This kind of tension and co-option is inescapable in a field so closely related to real-world developments and where it is possible for scholars to become practitioners, as many have. It is more likely still in great powers like the United States, China, and Russia, who all too often use their power in ways not sanctioned by their neighbors or the global community.
The difficulty of separating IR from politics has always touched a raw nerve in the American academy.13 Behavioralism become increasingly dominant in the early postwar decades Steeped in the positivistic tradition, its advocates insisted that social science was “objective,” “fact based,” and “value free.” They denounced other approaches as unscientific. The pendulum began to swing in the other direction by the end of the century with poststructuralists contending that social science is little more than ideological window dressing for existing power structures and their hierarchies. I follow the great fin-de-siècle German sociologist Max Weber in staking out a middle position. Our values influence, often determine, the conceptual categories we use to organize the world, define problems of interest, construct theories used to study them, and determine what constitutes relevant evidence and how it should be evaluated. Nevertheless, protocols for conducting good research distinguish good scholarship from mere advocacy. These protocols are constantly challenged and revised over the course of time. However, there is no escaping the fact that what we do reflects our position in society, life experiences, beliefs, and commitments. Again pace Weber, I contend that the first step in doing ethical research is to recognize and acknowledge as far as possible our parochial perspectives and reflect upon the ways in which they shape our research agenda and even our findings.14 We will see just how the ultimately unresolvable tension between the social nature of our enterprise and our desire to do science creates serious problems for positivist and interpretivist epistemologies alike.
The close connection between the field’s agendas and concepts and what it studies has another pronounced effect. Research agendas, theories, even paradigms rise and fall in importance in response to changes in the political world. IR’s first hundred years witnessed phenomenal upheavals, among them the Bolshevik coup and Russian civil war; the Versailles settlement and its profound consequences for the map and politics of Europe, the Middle East, and Africa; the Great Depression, the rise of the dictators, World War II, communist victory in China’s civil war; the Cold War; nationalist struggles and decolonization in the Caribbean, Africa, and Asia; divided nations and partitioned countries; the European project, the advent and spread of nuclear weapons, globalization and related economic crises; mass movements of peoples, many of them refugees; the end of the Cold War and implosion of the Soviet Union; the information revolution and the advent of cyber warfare; the phenomenal rise of China as an economic, political, and military superpower; the self-mutilation of the United States; the resurgence of the nationalist right in Europe and the United States; and, most recently, a global pandemic. Many of these events generated intense interest, and some were catalysts for paradigm shifts. More importantly, they influenced or redefined our conceptions of what global, international relations meant and even what it is for something to be characterized as political.
Fluidity in substance, theory, and methods gives further urgency to the question of just what IR is. There has never been any consensus as to the scope of the field. In its early days, IR was almost entirely focused on war prevention and conflict management among the great powers, and from a political-historical, legal, or geographic perspective. Gradually IR scholars became interested in nationalism, modernization, deterrence, crisis management, human rights, refugees and migration, women, and the reconstruction of civil society after civil wars. International political economy developed as a subfield in the 1970s and came to rival security as a focus of interest. More recently, international political theory emerged as a subfield, making efforts to link IR theory to political theory and using the latter to formulate new questions for empirical research.15 IR is not unique in this regard. In my student and early teaching days, comparative politics was largely synonymous with the governments of the great powers. To the extent that political processes came into the picture, they were largely limited to elections and interest groups. Comparative politics has arguably evolved in its interests even more than IR.
Any definition of our field is also closely tied to paradigmatic and institutional commitments. In the United Kingdom and most of Europe, IR is considered a discipline and universities have freestanding IR departments that augment and sometimes compete with departments of politics or political science. In North America, IR is one of several fields within the discipline of political science. To justify IR as a separate discipline, or even as a separate field, it was necessary to distinguish it from departments or fields to which it had previously been attached. The initial move in this direction was to break away from history and international law. This began in the 1920s and gathered steam in the 1930s. Realists in international relations, among them Carr and Morgenthau, invoked the Kantian distinction between “is” and “ought.” They argued that IR should describe the world as it was and not as it ought to be and denied that law was useful in resolving questions that were fundamentally political in nature. In making this move, realists propagated an unfair depiction of international lawyers, whom they branded as misguided and even dangerous “idealists.” In reality, these scholars were among the first to recognize the limits of law and to describe and publicize the threat posed by Hitler.16
New arguments were coined in the aftermath of World War II. Once again Morgenthau was in the forefront and was joined by such theorists as Frederick Schumann, William T. R. Fox, and John Herz. International politics was distinguished from its domestic counterparts by the lack of any central authority.17 In contrast with the domestic politics of well-established states, there was no Leviathan to enforce laws. In keeping with this distinction, realists mobilized Weber’s characterization of a state as a territory over which a government exercised a monopoly on the use of force. This definition also suited them because it stressed power, not community, and authority, not society, as the essence of politics at every level of aggregation. In 1979, Kenneth Waltz turned this distinction into a sharp binary in his Theory of International Politics. He described the international “system” as one of utter anarchy, which sharply set it off from domestic politics and made security the first concern of states.18
These claims about the unique character of international relations were the conventional wisdom for decades. They strengthened the hold of realism, which, for most of the postwar era, was the dominant paradigm in the United States. Its success, I believe, had little to do with the intellectual power of these questionable conceptual moves. Rather, it was attributable to the Cold War and concern – even fear – about the possibility of another world war. Power-centered theories that emphasized military might and offered jaundiced views of law, treaties, and international cooperation seemed appropriate to those who regarded the Soviet Union as a serious threat. Government officials and two generations of college students lapped up realism. Denuded of the sophistication found in the writings of Morgenthau and Herz, it rapidly became indistinguishable from Realpolitik: power politics with no concern for ethics. It provided the justification for Washington’s support of right-wing dictators, coups against popularly elected left-leaning governments, assassinations of their leaders and other politicians thought to be friendly to the Soviet Union, and military interventions culminating like that in Indochina. One of the academy’s leading realists, Henry Kissinger, became a principal architect of these policies.19
The relationship between theory and policy in the Cold War is not without irony. Realism was developed and propagated by scholars like Morgenthau in the hope that it would encourage policymakers to reframe their conflict with the Soviet Union. Officials and the informed public were encouraged to see it less as an ideological confrontation between good and evil and more as a power struggle between two superpowers anxious to avoid war. Rather than a fight to the death, it was a conflict that might be managed by the balance of power, diplomacy, and self-restraint. Deterrence was mobilized in a similar way. It was intended to keep war at bay by restraining the Soviet Union. However, as practiced by both sides, deterrence became a source of tension in its own right and the principal cause of the Cuban missile crisis, the most acute confrontation between the superpowers.20 Subsequent arms deployments ratcheted up tensions and prolonged the Cold War.21 Sophocles’s Oedipus Tyrannus dramatizes the possibility that our actions can produce results diametrically opposed to those we intend. A compelling argument can be made that deterrence theory has had this effect.
Waltz’s success, which occurred late in the Cold War, was also very much a matter of timing. The behavioral revolution conquered American social sciences in the postwar era and economics was increasingly regarded as a science. Its academic practitioners won Nobel Prizes and were recipients of high status and pay within the academy. Political scientists sought to emulate them. Waltz’s theory was a particularly crude attempt to colonize IR with concepts drawn from the theory of the firm. He sought to make IR more scientific by defining theory in a way that all but excluded history, context, agency, and ethics. I read his Theory of International Politics when it first appeared in 1979 and could not believe anybody would take it seriously. How wrong I was! In Europe it had little impact, but in the United States it all but defined the field the next decade.
Waltz and his brand of neorealism dropped like a stone in water at the end of the Cold War. The peaceful resolution of this conflict and the breakup of the Soviet Union stood in sharp contrast to the expectations of his theory. A new set of theoretical and policy concerns emerged in response to the resurgence of identity politics, nationalism, terrorism, globalization, and regional integration. Neorealism was regarded as irrelevant to all of them. A notable exception was John Mearsheimer, whose The Tragedy of the Great Powers, published in 2001, tried to adapt neorealism to foreign policy.22 He stipulated that great powers have two strategic goals: to acquire as much power as possible and to prevent the hegemony of other powers. He insists that great powers have always been willing to go to war for either end.
The fate of neorealism is more evidence of the sensitivity of IR theory to real world developments. The rise and fall of paradigms and theories are hardly ever due to critiques, research, and internal debates. Major shifts are invariably responses to outside events, as indeed was the origin of international relations itself. This linkage has important implications. Like all social theories, IR theories are meant to shape our view of the world and help us cope more effectively with it. The history of IR suggests that theories do not do a good job of explaining the world, in part because they are the products of preexisting understandings of it. Waltz’s neorealism and John Gaddis’s best-selling book on the so-called long peace found receptive audiences because so many people in the 1950s and 1960s – especially Americans – thought war between the superpowers was highly likely or even inevitable.23 This belief rested on the assumption that Stalin and Khrushchev like Hitler before them were intent on world conquest. The Cold War declined in intensity, although it did heat up again in the aftermath of the Soviet invasion of Afghanistan in 1979. The unexpected absence of superpower war required an extraordinary explanation and Waltz and Gaddis provided it. If policymakers and academics had supposed that both superpowers were aggressive and hostile but terrified of war and never so dissatisfied with the status quo that they were willing to risk one – let alone start one deliberately – then the Cold War’s failure to turn hot would not have appeared an anomaly. This latter assumption, we now know, would also have been much closer to the truth.
A mirror phenomenon occurred in international political economy. The 1970s witnessed a relative economic decline in the United States relative to the much faster growing economies of Germany and Japan. Hegemonic stability theory would expect these newly ascendant countries to push for radical changes in institutions governing the international economy from which they would benefit.24 Instead, they remained committed to defending the institutions, procedures, and norms that Washington had put in place following World War II. Robert Keohane wrote a well-received book based on the premise that American hegemony had eroded and invoked the concept of regimes to explain why cooperation had nevertheless persisted.25 Once again, such an elaborate explanation would not have been needed if people had reasoned – correctly, as it turned out – that Germany and Japan benefited from the existing order and risked losing a lot from the uncertainty and tensions any efforts to restructure this system would have generated. In effect, those in charge of Germany and Japan were basically satisfied and risk-averse.
In both instances, questionable political assumptions prompted expectations that never materialized. These failures gave rise to the need for new theories to account for their absence. Scholars – and the public – were more willing to buy into these ex post facto accounts than to question the initial underlying assumptions of theorists. The story does not stop here. These misleading assumptions were so ingrained because they were the outgrowths of other assumptions and theories that had become deeply entrenched in the postwar psyche. Something similar happened with deterrence. It failed quite dramatically in the lead-up to the Cuban missile crisis; Kennedy put Khrushchev on notice that he would not tolerate missiles in Cuba and Khrushchev nevertheless attempted to deploy them. American pundits and scholars never considered the possibility that deterrence as practiced by both sides had exacerbated mutual insecurities and promoted more aggressive policies for fear of being seen weak or timid by their adversary. In retrospect, they sought to salvage deterrence strategy and theory by arguing that President Kennedy had failed to practice it forcefully enough.26
The relative appeal of theories is very much a product of the Zeitgeist. There is a complex interplay between what happens in the world, how we theorize it, and the lessons we learn from practical applications of our theories. In formulating and evaluating our theories, we need to recognize how embedded they are, but this is difficult because we are also products of our time and society. So are the people we try to influence. Following the publication of We All Lost the Cold War in 1994, I tried to place a short piece in the New York Times about how deterrence had helped provoke the Cuban missile crisis. The then-editor of the Op Ed page rejected it, writing to me that “surely I could not expect anybody to believe such a claim.”
International relations is international in a double sense. The founding generation of IR scholars trained in Europe or the United States as lawyers, historians, political scientists, economists, geographers, and regional specialists. A few had been journalists or diplomats. All had international experience by virtue of their studies, research, employment, or immigration. The profession’s geographic basis narrowed considerably in the 1930s when its most prominent continental representatives sought refuge in the United Kingdom or the United States. International relations theory was a largely Anglo-American enterprise in the early postwar decades. Realism and liberalism, the two dominant paradigms of the era, had British and German origins but developed in ways that made them distinctively American. Institutionally, the profession also remained Anglo-American for much of the postwar era. European, Latin American, Asian, and African students came in large numbers to Britain and the United States to do their PhDs. Many went home to train their students in Anglo-American approaches. Until the founding of the European Journal of International Relations in 1995, all the leading journals in the field were published in the United States or the United Kingdom.
The field only become international again in the current century. It is now possible to obtain a first-rate postgraduate education outside of North America and the UK. Funding for postgraduate study and postdoctoral fellowships is readily available in Europe, Turkey and Israel, the Pacific Rim, and Australia. The top presses in the field are still in the United States or the United Kingdom and publish in English. But they offer outlets to scholars from many countries and research traditions. Language is not the kind of barrier it used to be since English became the lingua franca. Books and articles in languages other than English receive correspondingly less attention. Research published in French, German, or Chinese has little impact beyond audiences familiar with these languages unless it is translated into English.
Although IR theory is international, it is also national. There are striking differences between American IR theory and the rest of the world. In the United States, as noted, IR is a subfield of political science rather than a stand-alone discipline. Political science was in the forefront of the behavioral revolution, and IR is accordingly largely positivist in its orientation. It is heavily colonized by those who do quantitative research, rational choice, or rationalist modeling. Non-positivist approaches, most notably constructivism, critical theory, and poststructuralism, have nevertheless blossomed in recent years. However, in the United States, in contrast with Europe, so-called thin constructivism – ushered in by Alex Wendt – is prominent and more positivist than constructivist in orientation. There are positivists in the UK, on the continent, and in Australia, but they are a distinct minority. In these locales, research is very much more in the interpretivist tradition.
The same divide is evident with respect to substantive paradigms. Realism in its many forms is practiced everywhere, but neorealism, offensive realism, and so-called neoclassical realism, are almost uniquely American phenomena. Neorealism does not exist in the UK, prompting John Mearsheimer, its leading American advocate, to accuse his British colleagues of discriminating against it, a charge that aroused disbelief and anger.27 Liberalism is British in origin, but the liberal paradigm in international relations is an American product and largely limited to American scholars, and neoliberalism even more so.
The conventional understanding of science, proclaimed initially by Isaac Newton, is that knowledge is cumulative and that today’s scientists stand on the shoulders of their predecessors. For many positivists, this model is equally applicable to the social sciences. Quantitative researchers in particular believe that they are producing a body of knowledge that will one day result in well-established theories of universal import. An alternative view, which I espouse, defines intellectual evolution in terms of the questions we ask, not the answers we find. These questions change, not primarily due to earlier research, but because of the changing nature of perceived international challenges and problems. To the extent that we can speak of progress, it comes in the form of interrogating and sometimes refuting theories. Such evaluation gives rise to new questions and theories.
My understanding of social science further assumes that in contrast to the hard sciences, there is no such thing as a general law. Social knowledge is always limited in nature because the understandings and practices that underlie behavior and give it meaning are culturally and historically local. These underlying conditions and understandings evolve in the course of time undermining what appeared to be appropriate rules of behavior or explanations. Unlike molecules, human beings learn from experience – although not necessarily the right lessons. As Max Weber observed, the half-life of any law will be brief because once people understand it they will take it into account, thus altering the environment that made it possible. To be fair, I suspect there are also instances of regularities being strengthened, ceteris paribus, once they become known.
For some social scientists, the problematic nature of laws is a cause of concern and frustration. For others, it makes their projects more challenging and interesting. Albert Einstein is reputed to have said that “God gave physics all the easy problems,” and he may have been right.28 Social science, and IR theory in particular, must embrace a different conception of knowledge than physics does, and also different ways of making knowledge claims. American pragmatist philosophers – Charles Sanders Peirce, William James, George Herbert Mead, John Dewey, and their German contemporary Max Weber – thought social science findings should be evaluated in terms of their ability to help us address problems.
These contrasting approaches to knowledge are a core cleavage among IR theorists. There is no way of proving any of these epistemologies right or wrong. They involve fundamental choices about the nature of our enterprise from which everything else derives. It is not unlike choosing a partner. To make this choice, you must follow your heart. In epistemologies, as in relationships, there is as much room for self-delusion as for rewards.
Epistemologies cut across most paradigms of IR theory. There are positivist and interpretivist realists, liberals, and constructivists. Philosophical realism, another approach to knowledge, also finds adherents in these and other paradigms. Marxism, critical theory, poststructuralism, feminism, and post-colonialism have their own epistemologies or can be described as variants of positivism and interpretivism. I organize my book around epistemologies. I elaborate, explore, and critique their understandings of knowledge, their assumptions about the world, and how they guide research. When I address paradigms, I do so to show how diverse they are with respect to their epistemologies, research programs, and favored methods. At most they share a certain core of assumptions.
An analysis of this kind encourages readers to see connections as well as differences across the deeply dug trenches of the IR theory battleground. I am inclined toward interpretivism but take a position, so to speak, at least in part in no man’s land. This is a dangerous place to be as combatants on all sides are just as hostile to those who hesitate to support them fully as to those who join the enemy. If you – and here I speak to students – become interested enough in IR theory to pursue the subject further, you will sooner rather than later make choices that put you in some camp. Such decisions are impossible to avoid once you begin to do theory. Many people start out with a preference for an epistemology, paradigm, and method, often because they have studied with professors who have strong commitments and present them as the only defensible choices. Some expect their students to replicate and help justify their commitments. The better alternative is to explore different epistemologies and research programs with an open mind. Learn about their respective assumptions, their relative strengths and weaknesses, what divides them and what they share in common. In the longer term, this will make you better, more open-minded, and more tolerant people.
Structure of the Book
Chapter 2 raises the most fundamental of epistemological questions: what is knowledge? Drawing on the ancient Greeks, I describe four kinds of knowledge and associate each with distinctive projects. What we generally refer to as positivist IR aspires to what Aristotle calls epistēmē: scientific knowledge that can be expressed mathematically. Interpretivist approaches aspire to phronēsis, which is best understood as practical knowledge that helps us cope with the world. This distinction cuts across paradigms and is a far more fundamental division.
The kind of knowledge we seek also determines the extent to which we frame our theory project as anchored more in the sciences or the humanities. They are profoundly different enterprises in today’s understanding of knowledge, although from ancient Greek to modern times they were bundled together. I suggest that the distinction between the sciences and the humanities can be overdrawn. Interpretivist approaches can be scientific in their protocols for research and evaluation, just as positivist approaches can accept the culturally local nature of knowledge and investigate domains normally thought of as humanistic.
Chapter 3 provides an overview of positivism and interpretivism and the different ways in which these epistemologies seek knowledge. I describe briefly their different historical roots and core assumptions shared by most researchers who identify with one or the other of these traditions. The two epistemologies offer opposing judgments about the similarities and differences between the physical and social worlds and the extent to which knowledge is universal or local or takes the form of regularities and laws or constitutive and causal narratives. They differ greatly in their belief about the extent to which fact and value can be separated. For positivists, this distinction is the sine qua non of science. For interpretivists, values are implicated in every stage of research. Interestingly, advocates of these epistemologies defend their respective positions by offering different readings of Max Weber.
Chapter 4 is the first of six chapters on how approaches rooted in positivist and interpretivist epistemologies seek knowledge. Positivism makes a set of substantive assumptions about the world. Most important for many researchers is the existence of observable regularities, discovery of which will allow explanation and prediction. Most positivists, but by no means all, assume that regularities are the equivalent of causes or can substitute for them as a source of knowledge. Regularities are discovered through large “n” (many case) quantitative studies that compare the frequency and timing of “things” called “variables.” The search for regularities makes use of independent and dependent variables that must be defined clearly and carefully. The cases that comprise any data set must be comparable and independent. Critics contend that these conditions are rarely met in international relations and, more fundamentally, that these categories are often arbitrary and invariably historically and culturally sensitive, rendering questionable comparisons across large numbers of “cases.” They further maintain that the search for regularities is confounded by the open-ended nature of the political world and nonlinear interactions. Many events – wars, for example – require catalysts, and they are often independent of underlying causes and not easily addressed by current quantitative research.
Chapter 5 examines experiments, another positivist research strategy. They test propositions or theories under controlled conditions, or occasionally by observing natural experiments. Experiments are designed to generate data that can be analyzed statistically. Experimenters are concerned with internal and external validity and have developed elaborate means for obtaining them. Internal validity is achieved when experimenters can rule out other explanations for their findings, usually in the form of systematic error or bias. External validity is the extent to which results can be generalized. Internal validity accordingly describes the level of confidence researchers have about the findings in the data set from which they were derived, and external validity pertains to their confidence that they can be applied to other data sets. The most important assumption underlying experimental research is that findings rigorously derived in one data set will be applicable to another or, put another way, that contexts are either irrelevant or can be controlled for. I examine this and other foundational assumptions of experimentation.
Chapter 6 explores rationalism. It is a paradigm that is largely in the positivistic tradition. I say largely because some modelers and gamers have little interest in the real world and are focused instead on the internal logic, patterns, and equilibria or solutions to their models or games. Most practitioners of this art in our field are drawn to it because they believe it has something to say about international relations. It can offer a template against which to assess real-world behavior or a vehicle for actually explaining and predicting behavior. This last purpose is where the overlap with positivism comes. Modeling and gaming or case studies based on rational actor assumptions assume that politics can be studied as a science. Rationalists attribute to actors the capability of “choosing the best means to gain a determined set or end.”29
Most rationalist modelers build on realist assumptions and predict foreign policy choices based on environmental opportunities and constraints. Most expected utility and game theoretic models make additional demands. They require actors to have transitive preferences, engage in Bayesian updating, and resort to sophisticated forms of signaling to convey their preferences, commitments and resolve and to learn about those of other actors. Leaders are assumed to be free of any ideological commitments, political constraints, or psychological pressures and limitations that might seriously interfere with the application of reason to foreign policy decision-making or their accurate reading of signals from other actors. They must also be focused on the problem at hand and not formulate their policy in response to some other problem they consider more important. These conditions are rarely encountered in the social world, and arguably less so in foreign policy decision-making.
Chapter 7 turns to interpretivism and its focus on the reasons people have for acting as they do. This bottom-up approach has the advantage of attempting to reconstruct the world through the eyes of actors, thus taking into account their culturally and historically specific understandings of social interactions as well as the importance of shared understandings for goal-oriented behavior. Good bottom-up analysis must consider the diverse goals of actors and the kind of rationality they utilize. These are difficult to determine. Analysts accordingly risk reading their own understandings onto actors, even in evidence-rich environments. Moreover, the quest for knowledge should not stop here. We want to know the consequences of actor behavior and how they came about. To do this, we need to study aggregation, the process by which the behavior of multiple actors leads to outcomes. Some of these outcomes may never have been desired or even envisaged by actors. To date, interpretivists have not effectively engaged this problem.
Chapter 8 examines practice theory, a rapidly growing paradigm within the interpretivist tradition. It seeks to explain the relationship between human action by reasoning that most behavior is socially determined and best studied through practices and the institutions they represent and instantiate. Practice theory conceives of behavior as patterned deeds in socially organized contexts. I provide an overview of the practice turn literature. I discuss different understandings of practice; the mechanism of habit, which is closely associated with them; and whether practice theory can be considered causal. I distinguish between works – by far the majority – that rely heavily on Bourdieu from those more indebted to Wittgenstein and the linguistic turn. I conclude by asking the extent to which performance theory has the potential to overcome or circumvent the agency-structure problem.
Chapter 9 engages counterfactuals, commonly known as “what-if” arguments. They change some feature of the past to bring about a different world in the present and connect the two by means of a causal chain consistent with empirical evidence. Counterfactuals are central to positivist and interpretivist claims to knowledge. They are used to evaluate propositions, theories, and causal narratives and also to imagine them. In different ways, the failure to consider counterfactuals or deploy counterfactual arguments weakens claims to knowledge in both research traditions. Beginning in the late 1990s international relations scholars began exploring the problems and prospects of counterfactual. In this chapter, I offer a synopsis of these developments and some thoughts about how counterfactuals can be used more effectively in both epistemologies.
Chapter 10 is the first of two chapters that address knowledge claims. Many positivists embrace Karl Popper’s concept of falsification. Popper argued that theories and propositions must be formulated in a way to allow their testing, and that the best test is an effort to disprove them. Repeated, unsuccessful attempts to falsify a proposition or theory “corroborate it,” that is, give it the status of a provisional truth. Falsification was part and parcel of the project, initiated by the so-called Vienna Circle, to provide a rational basis for truth claims in the sciences. Scientists have never responded favorably to such efforts and remain more impressed by seeming proofs than they are by apparent falsifications. Critics of the early Popper and Vienna Circle regard science as a practice, regulated by agreed-upon protocols of what constitutes good science. These protocols arise from ways in which science is conducted and reflection about it. They evolve over time, sometimes dramatically, in response to criticism and new ways of conducting research. I explore some of the implications of this model of science for IR theory.
Chapter 11 turns to interpretivism and its take on truth claims. Interpretivism is rooted in the concept of verstehen (to understand) in contrast to positivism’s emphasis on erklären (to explain). Explanation for positivists consists of theories and they in turn encode factual descriptions of the world. Verstehen can take diverse forms, although all of them ultimately seek to describe the reasons people behave as they do, either as individuals or collectively. Interpretivists differ in their emphasis on cause. Contra positivism, all interpretivists are relativists. They reject outright the idea of there being any kind of objective test or standard for evaluating research findings. They treat social science as a practice with protocols for research and its evaluation specific to research traditions. This diversity is advantageous and problematic in equal measure for reasons I elaborate.
Chapter 12 is the first of four chapters on reason and cause. It explores the concept of reason. My starting point is Max Weber’s insistence that all theory must be rational, by which he meant logical and consistent in its definitions of terms and the relationships it stipulates among them. It must also specify the scope of any propositions and the conditions in which they are expected to operate. These several conditions avoid internal contradictions and allow external efforts of evaluation. Weber conceived of rational behavior on the part of actors as an ideal type but hoped that people might be rational enough to allow instrumental rationality to be used as a benchmark for analysis. Early game theorists followed Weber and regarded their models as ideal types that were useful for establishing guidelines for policy and identifying departures from reason in practice.30 More recent theorists of all kinds in the positivist tradition, and some interpretivists, have ignored Weber’s caveat. They also mistakenly treated reason as objective and universal in nature when in practice; as Weber recognized, there are different kinds of reason appropriate to different kinds of goals and visions of life.
Chapter 13 takes up the problem of cause. From the ancient Greeks to the present day, philosophers differ on what cause is, how it should be defined, and how central it should be to our efforts to understand the world. There is controversy about whether cause is a feature of the world or an artifact of the human mind. If the former, how do we establish cause? If the latter, are all assertions of cause nothing more than rhetorical? And if cause is a product of our imagination, could we dispense with it? Perhaps, as some scientists assert, it stands in the way of our search for knowledge. These various choices find expression in IR theories but are often left untheorized. I explore these several controversies and make the case for cause as a useful concept but one that does map neatly on to the world, for “thick” over “thin” conceptions of cause, and attention to the mechanisms and processes that link theorized causes to their alleged effects.
Chapter 14 explores what I call the “causal paradox.” People think causally when doing so appears to confer advantages in coping with the world. Causal inference accordingly becomes more important in the very conditions that make it more difficult. When people feel less confident about discovering causes, they feel less able to control or manage their environment and correspondingly are more at risk. They become insecure and vulnerable, and all the more so to the degree that their livelihood, status, or security depends on causal inference. People have strong instrumental and psychological motives to find ways around this dilemma. I identify and evaluate four generic strategies for coping with the causal dilemma in international relations theories. I conclude with some thoughts about why modernity has made cause more important than ever before and how this development has made us more responsive to some strategies than to others.
Chapter 15 takes up thick understandings of cause. These accounts engage in process tracing. They direct their attention to the mechanisms that link causes to their alleged effects and in doing so explain why some event or phenomenon occurs. Researchers can begin with an outcome, theorize a cause or causes, and look for evidence in support of links between them and the outcome. Or they can start with a theorized cause and attempt to discover its diverse effects. I identify different kinds of mechanisms and the functions they are expected to perform for different epistemologies. Positivists are attracted to thin mechanisms, as they are to thin conceptions of cause. They often take the form of searching for regularities to explain other regularities. Interpretivists gravitate to thick mechanisms that often entail unobservables. I suggest that we must theorize our mechanisms in a convincing manner before using them and stipulate relevant criteria toward this end.
A concluding chapter offers generalizations about IR theory, its evolution, and its prospects. I pose three sets of questions about each epistemology in this book: how it defines knowledge, how it seeks it, and how its adherents claim to have found it. I have stressed the differences between positivism and interpretivism, but there is also considerable variation within them, variation that should encourage dialogue not only within but also across epistemologies. I discuss some of the forms this dialogue might take and the ways in which researchers in each tradition can learn from the other.
Terms and Their Use
I do my best to provide useful definitions of the various concepts I use. They include ontology, epistemology, paradigm, theory, and method. This is a challenging task because they are all used in different ways, and some of them, like epistemology and theory, are hotly contested. Different definitions may also stress different features of what is being defined. In this connection, Peter Halfpenny identifies at least twelve versions of positivism.31 Different definitions support different understandings of knowledge, making it all but impossible to be neutral and precise about any of these terms. I choose definitions appropriate to my project but recognize their subjectivity. On occasion, I offer and contrast multiple definitions, starting with that of knowledge.
Ontology is closely associated with Aristotle’s question of being: about what entities are real and what they share in common. The word “ontology” comes from the ancient Greek words ōn and ont, referring to “being,” plus the suffix “logy,” meaning “the study of.” Ontology is often described as the science of being and belongs to the branch of philosophy known as metaphysics. Ontology refers to the nature of the social world and of human beings, and the relationship between the two. Many ontologists are interested in classifying entities and determining the relationships among them. Commonly used categories are events, agents, substances, properties, relations, and states of affairs. These entities are assumed to possess concreteness, particularity, universality, and sometimes necessity. There is no consensus among philosophers about what entities exist and their relevant importance, utility, or connection to other entities.32
In social science, ontology and its entities refer primarily to units of analysis. For many theorists, these are people; they are described as irreducible units and the agents responsible for all social and political behavior. As we shall see, the emphasis on human agents as the proper focus of study spans epistemologies. Ontology is sometimes also used to describe other key assumptions that theorists make about the world. Among the most important is the extent to which some things and processes are real entities or artifacts of the human imagination. A related controversy is the degree to which the properties of things are independent or dependent on human understandings of them. Positivists assume many things to be real and their properties independent of what we think about them. Interpretivists, by contrast, start from the assumption that the social world is composed of intersubjective meanings that find expression in norms and practices. For positivists, successful behavior requires understanding of the external world, or at least behavior in conformity with its opportunity and constraints. For interpretivists, people can only come to know their worlds and develop the ability to act in them through shared meanings.
I address both ontology and epistemology in this book but focus more on the latter. All concepts have uncertain and porous borders. Depending on how epistemology is framed, it bleeds into ontology on one end and method on another. This is equally evident when discussing individual epistemologies. Positivism can and has been described as an ontology and as an epistemology and is associated with certain kinds of methods. The Stanford Encyclopedia of Philosophy has multiple headings for positivism, each of them referring to a different understanding of it in philosophy, law, and social science. For the Vienna Circle – central to my treatment of positivism – it was a project to demonstrate that metaphysics was not only false but also cognitively empty and meaningless. Their understanding of positivism qualifies as a philosophical system that holds that every rationally justifiable assertion can be scientifically verified or is capable of logical or mathematical proof.33 Positivism is strictly speaking a philosophy of science. However, in social science, it is a term used to describe an empiricist epistemology.
According to the Stanford Encyclopedia of Philosophy, epistemology can be translated as “knowledge.” However, epistemology really only takes on meaning with respect to particular philosophical projects. Plato’s epistemology was an attempt to understand what it was to know, and how knowledge benefits the knower. Locke’s epistemology was directed at understanding the operations of human understanding, Kant’s sought to grasp the conditions that made human understanding possible, and Russell’s to understand how modern science could be justified by appeal to sensory experience. Much recent work in formal epistemology is an attempt to understand how our degrees of confidence are rationally constrained by our evidence. In poststructuralist and feminist epistemology, the emphasis is on understanding how interests affect our evidence and rational constraints more generally.34 In the chapters that follow, I distinguish positivist from interpretivist epistemology and describe the different kinds of understandings they seek and some of the variation within each of these research traditions.
The other problem worth addressing arises from my binary of positivism and interpretivism. One of the manuscript’s reviewers rightly noted that by subsuming all the non-positivist approaches under the interpretivist banner, I am likely to annoy representatives of these approaches for not addressing the specifics of their positions. This reviewer further contends that “Only to a neopositivist would ‘interpretivism’ appear to be anything like a coherent category, since the only thing unifying the motley crew of scholars and scholarship that Lebow places under that heading is that they are all not neopositivists!” I respectfully disagree. Interpretivism is a meaningful epistemological category, and one in many ways opposite to positivism. In the succeeding chapters, I describe its foundational assumptions with regard to the nature of knowledge, how it is acquired, and how it is claimed. These assumptions are shared across a range of approaches that otherwise differ. The reviewer is right in thinking that my bundling these approaches together may cause some wonderment among some of their adherents. Perhaps this is a good thing. When focused on research programs and methods, IR scholars are bound to emphasize their differences with one another. It is useful for them to recognize what they share in common.
It is undeniably true that any effort to categorize anything is both a useful and a frustrating enterprise. It is useful because it highlights what seemingly different things share in common. Biologists have long used the categories of kingdom, phylum, class, order, family, genus, species to organize relationships among living things. They are the first to admit that classification of this kind involves some degree of shoehorning and does not neatly encompass all kinds of life – especially some microbial forms. Social categories are more arbitrary still because social kinds are not real in the same sense and not often the products of common ancestors. The test of any typology – and in this instance, of its associated binary – is its utility. Biologists start with dividing plants from animals, knowing that this excludes some life forms like fungi and ignores the many ways in which animals and plants – and many species in both categories – are codependent. Their taxonomic rankings nevertheless tell them many things of value. The proof of this assertion is that a scheme that began with Linnaeus in the late eighteenth century and periodically added to and revised has endured to this day and is a starting point for all students of biology.
So it is with my binary of positivism and interpretivism. It is one commonly made and widely used in the social sciences because, like that of plants and animals, it captures what is perhaps the most fundamental distinction in our search for social knowledge. The categories of plants and animals are subdivided into multiple categories down to species and subspecies. Those who use these last categories are interested in slotting individuals into them and distinguishing between and among them. This is an important task but one that, of necessity, ignores what they all share in common. Zoologists might become annoyed if a colleague came along who studied phyla or classes and told them how all but indistinguishable the the living things they studied were. This is, in effect, what I am doing to my IR colleagues, reviewers included, who work – quite properly – at the social equivalents of species and subspecies. For reasons I make abundantly clear in the narrative, moving up a level or two in the taxonomy to epistemology is a useful, even necessary enterprise. We should be equally aware of the similarities and differences in what we study and in social science of how we study.
The takeaway here is that interpretivism encompasses or can be used to describe a wide range of research programs. It is more diverse than positivism, but that epistemology is still reasonably variegated. I am not imposing my binary on research programs and associated methods, but rather building it bottom-up on the basis of what these programs share in common. There are still many nuances, as research programs that share core assumptions generally emphasize certain epistemological features over others and may interpret them somewhat differently. I highlight some of these differences but feel justified in focussing primarily on what these programs and approaches have in common and which set of conceptual and substantive assumptions divide positivists from interpretivists. The positivist-interpretivist binary accordingly serves my purpose of interrogating the theory project in IR in the most fundamental way.