But, Mr. Voltaire, declared lover of truth, tell me in good faith, have you found it? You combat and destroy all errors; but what do you place in their stead?
– Mme. du Deffand, letter to Voltaire, December 28, 1765Footnote 1
Introduction
International Relations (IR) is uneasy about its status as a rational enterprise capable of producing knowledge about the world of international politics.Footnote 2 This uneasiness is reflected in periodic bouts of skepticism, disagreement, and debate on the field’s standards of inquiry and prospects for scientific knowledge and progress. In short, IR is insecure about its foundations as a social ‘science’.
In search of a foundation to ground the discipline, IR scholars routinely turn to the Philosophy of Science (PoS) for guidance. At first blush, the turn to the PoS appears eminently reasonable for a field in search of scientific credentials. After all, philosophers of science spend their time constructing frameworks to illuminate the dividing line between science and pseudo-science and identifying logics of inquiry that make scientific practice intelligible and that account for its achievements. And, at least prima facie, much is at stake in IR’s foundational debate. Without the right foundations to guide inquiry and confirm the merits of particular questions, theories, and methods, how can IR scholars be sure their findings are valid? Without a robust yardstick to separate good from bad (i.e., non-scientific) work, on what basis should graduate students be trained, journals select articles to publish, and universities hire and tenure faculty? Finally, without establishing their enterprise as ‘scientific’, how can IR scholars expect policymakers to take the intellectual fruits of the field seriously?
Unfortunately for IR, there is no consensus among philosophers of science on the right answers to these questions. Indeed, the PoS is itself engaged in an ongoing debate and remains deeply divided over the nature, possibility, and desirability of foundations for science and knowledge. The multiplicity of reasonable foundational positions, supported by prominent philosophers of science, virtually guarantees that each contending perspective in IR’s ‘science’ debate will find a well-supported philosophical position to justify its particular approach. Indeed, the most prominent foundational positions in the PoS – Instrumentalism, Social Constructivism, and Scientific Realism – have become the most commonly endorsed foundations for IR, mapping on to the distinction between ‘positivism’, ‘anti-positivism’, and ‘post-positivism’.Footnote 3
Advocates of positivism, like Kenneth Waltz, endorse Instrumentalism in an attempt to make IR more scientific. Instrumentalism entails the view that objective knowledge is limited to what can be verified through observation. Anti-positivists, like David Campbell, build on Social Constructivism, the position that truth is a function of social and political processes. They deny the possibility of objectivity and therefore science.Footnote 4 Post-positivists, defined by their desire to find a PoS foundation acceptable to both sides of the science debate in IR, are more varied in their choice of foundational positions. To date, the most influential post-positivist position is Scientific Realism, which argues that skepticism about the possibilities for knowledge is unwarranted. Prominent advocates for Scientific Realism in IR include Alexander Wendt and Colin Wight.
Yet, despite post-positivists’ efforts to reconcile the foundationally divided discipline, the quest for philosophical foundations in IR continues, with appeals to philosophical authority playing a signal role. Commitments to Instrumentalism, Social Constructivism, and Scientific Realism now anchor a three-cornered fight between positivists, anti-positivists, and post-positivists, with each side wielding its favored philosophical position to settle the science debate once and for all. It should therefore come as no surprise that the turn to philosophical authority in IR has failed to produce a consensus on the proper foundations for the field. Indeed, the quest for a single foundation to define IR’s relationship to science has had the opposite effect. By escalating substantive and methodological disagreements to the foundational level, such imperial projects have contributed to the fragmentation of the discipline along meta-theoretical lines.
In this article, we challenge the usefulness of the foundational project in IR. Instead, we argue, a different approach to the field’s relationship with the PoS is not only warranted but necessary. As we show below, the ongoing division among positivists, anti-positivists, and post-positivists is the inevitable result of each side’s claim to represent the right position in the PoS. We argue that IR must come to terms with the fact that there is no Archimedian point capable of settling, once and for all, the debate among foundational positions. We appreciate that, as in any other social science, questions of meaning and interpretation loom large in IR. Choosing appropriate questions, theories, methods, and standards of evidence remain important challenges for IR scholars – challenges worthy of reflection and debate. And we believe the PoS has much to offer in terms of concepts, arguments, and historical analysis that help illuminate the range of meta-theoretical possibilities. What the PoS is unable to do, however, is settle the question, ‘What should IR look like in order to qualify as a scientific enterprise?’ To claim otherwise amounts to meta-theoretical sleight-of-hand.
In our view, IR should adopt an open-minded and prudent ‘attitude toward’ – rather than a ‘position in’ – the unsettled PoS debate. This implies recognizing the limits of the PoS as a legitimating discourse. Indeed, perhaps the most important lesson to draw from the PoS as a discipline is that being open-minded about the boundaries of science is the only way to go. Adopting a prudent attitude toward the foundational debate necessarily forecloses imperial foundational projects. That means eschewing a priori judgments based on foundational commitments on what constitutes legitimate work in IR, thus judging work on its substantive contribution, not on the degree to which the chosen approach conforms to a particular conception of science.
To be sure, we are not the first to try to move the discipline beyond its foundational insecurities – others have gone to great lengths to ground IR on a foundation that makes room for problem-driven pluralism. What differentiates our project is that we do not seek to put forth the philosophical flavor du jour as the right foundation for IR. Our goal is to make room for problem-driven research and pluralism in approaches to substantive questions in IR by exposing the contradictions in, and inherent limitations of, imperial foundational projects designed to ground IR on a single vision of what science is. We are therefore emphatically not engaging in yet another example of meta-theoretical one-upmanship. Rather, we aim to put foundational arguments in perspective, recognizing their strengths and limitations, what they can and, above all, cannot do.
Recognizing the limits of foundational arguments is not the same as rejecting their usefulness in toto. The attitude we endorse is not anti-foundational, but post-foundational.Footnote 5 It tries to make the best use of PoS arguments on par with substantive and methodological arguments, while doing away with the imperial foundational projects that seek to impose a single foundation on IR as a whole. We call this attitude ‘foundational prudence’. Adopting it would lead IR scholars to move beyond foundational debates. In our view, IR can do well – indeed better – without worrying about securing its philosophical foundations. Philosophers have been unable to agree on what constitutes an unshakable foundation, and we have little reason to expect IR scholars to fare any better. The quest for foundations is best left to the philosophers.
The remainder of this article proceeds as follows. In the second section, we lay out the process through which the PoS became important in IR debates and summarize the problems this creates for IR scholarship. Then, in the third section, we turn to the contemporary PoS, summarizing the main foundational positions, showing how they relate to IR, and highlighting why none of them can serve as an unshakable foundation. The fourth section presents our attempt to move from a ‘position in’ to an ‘attitude towards’ the PoS debate, and lays out our vision of IR freed from foundational debates. The fifth section addresses likely objections to a post-foundational IR and implications for stakeholders in the ongoing ‘science’ debate. We offer a brief conclusion in the sixth section. By putting the PoS in perspective, we hope to lay the groundwork for future philosophically informed discussions about IR scholarship, while defusing some of the more counterproductive divisions in the field.
The ‘science’ debate in IR
The ‘science’ debate in IR has divided the field on the very possibility of a science of international politics. A recent survey of the profession conducted by Maliniak et al. (Reference Maliniak, Oakes, Peterson and Tierney2007) demonstrates the centrality of foundational divisions in IR. When asked to identify the principal disciplinary divide, 22% of IR scholars pointed to philosophical foundations (specifically, epistemology) as the major divide. Furthermore, the only other two issues that gathered more than 10% of the votes as the most important disciplinary divide – methodology (36%) and paradigms (27%) – implicitly index foundations. In other words, a full 85% of respondents name as the most important divide in IR one that is tied to the unresolved state of the foundational debate (Maliniak et al., Reference Maliniak, Oakes, Peterson and Tierney2007: 28). Foundational positions have become part-and-parcel of the way IR scholars think about themselves and their work. Indeed, seven out of ten American IR scholars consider themselves positivists, with the remainder upholding either non-positivist or post-positivist foundational commitments (Maliniak et al., Reference Maliniak, Oakes, Peterson and Tierney2007: 37). Foundations in IR are hardly buried; they are at the forefront of disciplinary debates.
The quest for philosophical foundations in IR
How did we get here? Three events stand out as particularly important in explaining the roots of the foundationally divided discipline.Footnote 6
The first is the behavioral revolt of the 1950s and 1960s that launched IR’s science debate. The behavioral revolt in IR is significant because it represents the first time that the PoS became central to the field’s self-definition as a science. Prior to the behavioral revolution, as Wight notes, there was little if any concern within IR about the scientific status of the discipline: ‘science simply was traditional forms of inquiry’ (2002: 28, Wight’s emphasis). And because the behavioral revolt was built on a position in the PoS, subsequent attempts to re-orient the field inevitably turned to the PoS as well.
Behaviorists sought to make IR more ‘scientific’ by importing methods and standards from the natural sciences. In making their case, they turned to Logical Positivism, the dominant PoS at the time, which defined science as the search for natural laws through rigorous empirical observation and systematic data collection, while according little value to theory and explanation. Behaviorists argued that only by systematizing the study of international relations on the model of Logical Positivism could IR truly be made into a science. Buoyed by this philosophical warrant, behaviorists sought to marginalize traditional approaches to IR that were based on historical analysis and therefore did not conform to the logical positivist ideal of science. The resulting clash is memorialized as IR’s ‘second great debate’ (Hollis and Smith, Reference Hollis and Smith1990: 28–32; Wight, Reference Wight2002: 26–32).Footnote 7
Unfortunately for behaviorists, philosophers soon abandoned Logical Positivism. Behaviorism was now vulnerable to attack on its own foundational terms. Initial challenges to behaviorism sought to update rather than replace the discipline’s now outmoded logical positivist foundation. Taking into account the major shifts in the PoS, this foundational makeover turned to Instrumentalism to bring theory back in, freeing IR from the sway of behaviorism’s strict empiricism.Footnote 8 The systematic collection of data would no longer be an end in itself. Instead, data was to be used to adjudicate among competing explanations for behavior and outcomes in international politics. This updated (and attenuated) empiricism would come to define IR’s positivist orthodoxy as the taken-for-granted foundation for scientific IR.
The demise of Logical Positivism, however, did not result in a single dominant position in the PoS. In IR, when the positivist mainstream failed to predict the end of the Cold War, a space was opened for a small but vocal coalition of critical theorists, feminists, post-structuralists, and post-modernists, marginalized by the discipline’s entrenched positivist orthodoxy, to challenge the mainstream’s scientific credentials. Just as the positivists had done a decade or so before, these anti-positivists too turned to the PoS. They drew on Social Constructivism – a foundational position competing with Instrumentalism – to justify their claim that the objectivity of science was undermined by the ineluctably social nature of knowledge.Footnote 9 For anti-positivists, IR must abandon any pretense of ‘science’ and instead focus on interpretation and critique: identifying taken-for-granted truths and showing these to be constructed, not ‘real’ in an objective sense. Thus began IR’s second foundational debate (its third ‘great’ debate).Footnote 10
In an attempt to overcome the often acrimonious debate between positivists and anti-positivists, IR scholars again turned to the PoS to find a position capable of reconciling these seemingly irreconcilable poles. The most prominent post-positivist position endorsed in IR is Scientific Realism, which promises a common foundational ground for the manifold existing approaches to the study of international politics.Footnote 11
As a via media, Scientific Realism claims to shield both positivist and anti-positivist approaches against subsequent foundational attempts to exclude them. It thus tries to open a space for a sort of inter-faith dialogue in the discipline, making possible a truly question- or problem-driven – as opposed to foundationally driven – IR.Footnote 12
Would the post-positivist intervention have worked, IR would finally have been able to move beyond the PoS foundational debate. But, alas, it did not. Nor, for philosophical reasons we highlight below, could it. The introduction of Scientific Realism has in fact been the object of much PoS flak and new foundational positions continue to be advanced. Additional post-positivist foundational contenders for IR include Quasi-Duhemian Conventionalism (Chernoff, Reference Chernoff2002) and Deweyan Pragmatism (Cochran, Reference Cochran2002; Festenstein, Reference Festenstein2002). In spite of such efforts to unite the discipline on one PoS foundation, there appears to be little hope of finding a common ground for IR.
In our view, the advocates of each PoS position are engaging in what could be called philosophical curve-fitting. Dissatisfied with the state of IR, they turn to the PoS, picking and choosing foundations that legitimate their vision for what the discipline should look like. They then deploy them as a legitimating warrant – not only for their own approach, but as the proper foundation for IR as a whole. In doing so, they transform their vision into an imperial foundational project. But any such project begets others to engage in the same enterprise – deploying an alternative foundation in defense of their own vision. Unsurprisingly, the result is a debate that produces more heat than light.
In sum, while importing foundational arguments from the PoS may have been historically necessary to break the monopoly of a dominant (narrowly positivist) foundation, the way this was done (and continues to be done) remains problematic. Much as the foundational debate may appear to have been productive, it was parasitic on the substantive and political debates that led to the introduction of new – and putatively more secure – PoS positions. Whatever productive role the foundational debate may have had in the development of IR as a discipline derived from the behaviorists’ original sin of supporting their rise on foundational arguments imported from the PoS. And since the foundational debate in the PoS is itself stuck, with several logically incompatible positions vying for endorsement as the proper foundation for science, importing these positions into IR has ultimately contributed to some of the stickiest problems in the discipline – first and foremost, IR’s inability to resolve the science debate. The result is a fractured field, with little or no dialogue between different approaches to the same substantive questions.
The problem with foundations in IR
The foundational debate in IR hinders the development of the field in two ways – one visible and superficial, the other less obvious but more damaging.
At the surface level, the PoS debate in IR continues to propagate in the form of articles, books, and panels at professional meetings. IR scholars continue their quest for an unshakable philosophical foundation, capable of settling once and for all the most fundamental questions in the field – its scope, goals, criteria, standards, and methods. Invariably, each proliferating piece of foundational scholarship in IR endorses one of the many positions available in the PoS – each of which promises a different way to fix IR’s scientific insecurity. To quote the title of Ferguson and Mansbach’s most recent entry into the foundational fray, The Elusive Quest Continues (Ferguson and Mansbach, Reference Ferguson and Mansbach2003). Furthermore, the debate has taken an increasingly acrimonious tenor. In Wight’s words, ‘despite the appearance of philosophical sophistication, the field has moved from throwing philosophical hand grenades to a largely untargeted artillery barrage against an ill-defined series of enemies’ (Wight, Reference Wight2002: 33).
But the problems generated by the foundational quest in IR are deeper and more pervasive than those reflected in the ongoing debate. The divisions created by continuously importing PoS foundations cut across several aspects of IR – meta-theoretical and methodological.
The main meta-theoretical divide remains that between positivists and anti-positivists. By conceiving of their disagreements as cutting all the way down to the foundations of knowledge, each side makes it more difficult to engage in substantive dialogue with the other. Therefore, incompatible PoS commitments have reinforced divisions along paradigmatic lines. Realists, neoliberals, IR constructivists, post-modernists, feminists, and others, have often invoked arguments from the PoS in defense of their own approach to IR, tying acceptance of substantive theory to meta-theoretical agreement (Hollis, Reference Hollis1996; Wight, Reference Wight1996). This deepens the substantive distinctions between these paradigms to the philosophical level, making inter-paradigmatic dialogue harder. It also makes the much-touted question-driven approach to IR more difficult to implement.
At the methodological level, PoS divisions have entrenched the split between quantitative, qualitative, and interpretive methods. These distinctions hinge in large part on the question of whether the social sciences are fundamentally different from the natural sciences. Those who believe they are similar endorse methods originating in the natural sciences: testing hypotheses – quantitatively or qualitatively – against empirical observations in order to discover law-like regularities. Those who, on the contrary, believe the social world is fundamentally different from the natural world – more contingent, the product of human understanding and meaning-making – endorse interpretive methods in the study of IR.Footnote 13 There are, to paraphrase Hollis and Smith (Reference Hollis and Smith1990), always (at least) two stories being told – one ‘explaining’ international politics, the other interpreting or ‘understanding’ it – with the result that proponents of different camps increasingly see each other as inhabiting and studying fundamentally different worlds.
In short, the use and abuse of PoS arguments in IR has aggravated the problems it was supposed to solve and inhibited intra-disciplinary dialogue – leading to a discipline divided against itself. Compounding the problem, IR scholars often make non-trivial, a-historical, and decontextualized misreadings of nuanced philosophical arguments (Wight, Reference Wight2002: 26). In the PoS, foundational arguments are immersed in a rich context of reflection and argumentation. It is to this context that we now turn.
Philosophy and the foundations of science
The PoS fulfills two roles in relation to science. First, in its descriptive role, the PoS attempts to distill the essence of successful scientific practice by observing and analyzing what scientists do, laying out how science is conducted. Second, in its prescriptive role, it offers foundations for successful scientific practice, laying out how science should be conducted. It is this second role that foundational arguments invoke, granting to the PoS the status of a special discourse – a normative discourse capable of evaluating science and legislating its boundaries and practices.Footnote 14
Foundational issues are particularly important in the social sciences. Two reasons account for this. First, the realm of study in the social sciences is populated by reflective entities, so that questions of meaning and interpretation play an unavoidable role in social–scientific inquiry. Therefore, the social sciences cannot so easily avoid the metaphysical questions relegated to secondary status in the natural sciences (Schatzki, Reference Schatzki1988).Footnote 15 Second, the social sciences have not generated technological advancements as momentous and self-evident as those produced by the natural sciences. In the absence of a legitimating warrant provided by technological progress, the social sciences have turned to the PoS in search of scientific credentials. As a result, while the natural sciences largely ignore foundational questions, philosophical debates continue to arise in the social sciences – including IR.
The central claim of this article is that the quest to ground IR inquiry on philosophically secure foundations is not likely to succeed. Perhaps ironically, the strongest evidence supporting our claim comes from the PoS as a field. In fact, the PoS is itself divided on questions of foundations and the epistemological and ontological commitments presented to secure inquiry as ‘scientific’. But, in the PoS unlike in IR, these divisions are substantive and do not prevent dialogue among philosophers.
What are philosophical foundations?
Philosophical foundations are ontological and epistemological starting points for inquiry, that is, for producing and justifying knowledge about the world (Honderitch, Reference Honderitch1995b). Foundational commitments define the scope of scientific knowledge.
Epistemology refers to the study of ‘the nature of knowledge, its possibility, scope, and general basis’ (Honderitch, Reference Honderitch1995a). Epistemologies are most usefully conceived of as theories of knowledge, and epistemological questions are questions about how and to what extent we can know the world. In their epistemological role, foundations traditionally establish the conditions under which knowledge is possible, that is, whether humans experience the world directly (as it is) or whether (and how) experience is mediated, giving us only indirect access to the object of study.
Ontology explores ‘the nature of existence and the categorical structure of reality’ (Honderitch, Reference Honderitch1995c). Ontologies are most usefully conceived of as theories about what the world is made of in a literal sense. Ontological debates in the social sciences revolve around the questions of what the social world is made of (e.g., individuals, actions, social forces, institutions, etc.), and whether it is fundamentally different from the natural world. In their ontological role, therefore, foundations define the kinds of things that exist in the world, which become the basic building blocks on which theories rely to explain phenomena.
What makes a set of epistemological and ontological commitments foundational is that there are no further commitments to which one might turn to justify knowledge. Foundational arguments posit that ‘knowledge of the world rests on a foundation of indubitable beliefs from which further propositions can be inferred to produce a superstructure of known truths’ (Honderitch, Reference Honderitch1995b). According to Hollis, ‘[f]or knowledge to be possible, there must be some propositions which can be known without proof or evidence’ (1994: 62). For example, there is no logical proof capable of justifying induction that does not itself draw on an inductive inference, that is, induction seems to grant us reliable knowledge of the world, and therefore, by inductive inference, should be (some argue) the foundation for science (Okasha, Reference Okasha2002: 29–33). There is also no proof to anchor the rules of logic essential to either inductive or deductive inference. Thus, induction and logic, commitments upon which the edifice of scientific knowledge is constructed, are beyond proof (Hollis, Reference Hollis1994: 62). In short, ‘foundations’ are themselves necessarily without foundations. They are by definition constituted by a priori knowledge and, therefore, their truth-status is unascertainable: they cannot be proven true or false.
And herein lies the crux of the PoS problem in IR. All foundational positions in the PoS are sets of epistemological and ontological commitments that are presented as necessary conditions for knowledge and yet are beyond proof. Epistemological and ontological positions, portrayed as bedrock foundations about the structure of the world and our knowledge of it, are beyond refutation. Consequently, one foundational argument cannot uproot another. It follows that the divisions resulting from the foundational quest – divisions rooted in commitments to incompatible PoS positions – are not likely to be resolved at the foundational level.
The three main foundational positions
The primary issue distinguishing the three main foundational positions in the PoS is how each of them conceives the relationship between epistemology and ontology. Put differently, having a strong view on the relationship between the two already goes a long way to determine one’s foundational position. Despite claims to the contrary by some IR scholars, the relationship between epistemology and ontology is itself essentially contested. Indeed, misrepresenting the philosophical debate on this topic as having reached any consensus is the source of many of the foundational problems plaguing IR.Footnote 16
In what follows, we survey the philosophical lay of the land. Philosophical arguments on the foundations of science have increasingly gravitated around three general positions: Instrumentalism, Social Constructivism, and Scientific Realism. These are also the three PoS positions most often endorsed in IR.Footnote 17 For space considerations, we treat each of these positions as ideal types, glossing over myriad distinctions that exist and continue to emerge in the philosophical debate. Although our simplifications are often ruthless, we attempt to distil the essence of each of the three positions, portraying them in ways recognizable to their respective proponents, even if only broadly so.Footnote 18
Instrumentalism prioritizes epistemology over ontology. Specifically, Instrumentalism is grounded on an empiricist epistemology according to which reliable knowledge comes only from what can be observed via the senses. Since the existence of that which cannot be observed cannot be verified, truth-claims about unobservables are beyond the scope of scientific knowledge. Observable phenomena are the only objects that science can reliably know to exist, and theories are judged in relation to how well their predictions accord to the observable world.
Unobservables, however, still matter for scientific explanation. Instrumentalists recognize that scientific theories typically rely on assumptions and causal mechanisms that cannot be observed. If reliance on unobservable (theoretical) entities improves scientific theories, these entities are treated ‘as if’ they were true. Instrumentalists are thus agnostic on whether theoretical entities are actually real. In one of the most radical formulations of ‘as if’ theorizing, Milton Friedman (Reference Friedman1953: 40–41) argued that there need be no relationship between reality and theoretical assumptions – the only thing that matters is that the theories based on such assumptions lead to valid predictions. Theories are useful instruments in improving our understanding of the observable world – thus the name ‘Instrumentalism’. If science is better able to generate explanations and predictions by thinking of assumptions and other theoretical entities ‘as if’ they were true, then these entities have been proven useful, which is all that can be asked of them.
Instrumentalists therefore endorse a deflated view of truth defined in terms of the empirical reliability of theories. Hypotheses deduced from theories can be tested on their ability to explain or predict observable phenomena. Science is thus a privileged form of inquiry, with scientific progress understood to mean an increase in the scope of useful and reliable theories about the world. But the theories themselves are neither true nor false, merely better or worse at capturing ‘objective’ empirical outcomes. It is this combination of empiricist skepticism and a sanguine attitude vis-à-vis the possibility of scientific progress that marks Instrumentalism as a form of positivism.
The quintessential example of Instrumentalism in IR is found in the opening chapter of Kenneth Waltz’s Theory of International Politics (1979). In what remains one of the most nuanced treatments of meta-theory – and, specifically, philosophical foundations – in IR literature, Waltz reflects instrumentalist skepticism regarding our ability to know whether the theoretical entities postulated by a theory (i.e., its assumptions and causal mechanisms) mirror reality: ‘The question to ask of an assumption, as ever, is not whether it is true but whether it is the most sensible and useful one that can be made’ (Waltz, Reference Waltz1979: 91). This leads him to endorse treating the unobservable constructs of a theory ‘as if’ they exist, that is, as instruments for explaining and predicting observable phenomena. Waltz advocates instrumentalist views on evaluating theories based on how well hypotheses derived from them capture observable phenomena. In his own words: ‘[Theoretical] assumptions are not factual. One therefore cannot ask if they are true, but only if they are useful’ (Waltz, Reference Waltz1979: 8–10, Waltz’s emphasis). For Waltz, then, a theory is merely an instrument for the apprehension of reality (1979: 124). If IR theories yield better explanations of international relations, whether they are actually true or not is beside the point.Footnote 19
A second central position in contemporary PoS is Social Constructivism. For social constructivists, knowledge is socially constructed – constituted by discourse and social practices all the way down.
Like instrumentalists, social constructivists start from an a priori epistemological commitment – in this case to a social epistemology. According to this view, language and social context mediate all human experience. Language gives meaning to the world, not vice versa. The world is therefore principally a social one in which the ‘objects’ of scientific inquiry are not independent of the social context and discourse that defines them. This means that ontology (what there is) is enmeshed with epistemology (how we know it) to a degree that makes the distinction meaningless. It also leads to skepticism about the human ability to know the world independently of the social context that gives it a particular and contingent meaning.Footnote 20
Social-constructivist skepticism is based on the history of science and scientific progress explored by sociologists and historians of science. Observing that long periods of scientific consensus are punctuated by moments of radical change in which taken-for-granted scientific truths are overturned, social constructivists conclude that it is social acceptance, not the world per se, that establishes ‘scientific truths’ (Kuhn, Reference Kuhn1996).
This emphasis on language, meaning, and social context leads Social Constructivism to see ‘scientific truth’ as a function of what is commonly accepted by scientists to be true.Footnote 21 Crucially, the role played by such an acceptance view of truth in Social Constructivism is radically different from that played by the other two positions’ favored theories of truth. While instrumentalists try to achieve empirical reliability (and, as we will see, scientific realists look for correspondence with reality), social constructivists try to reveal the processes through which scientific consensuses are produced. They are not themselves in search of truth. Rather, their purpose is to unmask how claims that other positions find ‘true’ – and thus take for granted – are in fact the results of a socially produced consensus.
Acceptance, and therefore also truth, is a function of the social context (or, to follow Kuhn, ‘dominant paradigm’) in which scientists define their research goals, methods, and standards of evidence. Social Constructivism in effect inverts the traditional causal logic of belief, with the social fact of belief causing the content of the believed statement to be true, not the other way around. In short, for Social Constructivism, science invents truths; it does not discover them.
Denying the possibility of objective knowledge of the world, Social Constructivism challenges science’s privileged status and takes an anti-positivist stance. Science is a social enterprise that is not driven by the search for truth; rather it is driven by the particular goals of different scientific communities, laden with the interests of the social actors participating in it, and the power relations between them. It follows that science cannot be progressive in an objective sense of producing increasingly more accurate or reliable representations of the world.
In IR, scholars who share Social Constructivism’s unmasking impulse come from a variety of intellectual traditions, among them critical theory, feminism, post-modernism, post-structuralism, and Pragmatism.Footnote 22 Though there are important differences between these approaches (Brown, Reference Brown1994), they all share the view that any positivist foundation is untenable. This, in turn, has several implications for how substantive IR questions can and should be approached. First, the emphasis on the discursive construction of reality lends philosophical support to using interpretive approaches to the study of international relations. Scholars such as Ashley and Campbell believe that explaining the behavior of states and statesmen requires understanding the contingent social and political conditions constituting the field of possibilities for actors and their behaviors.Footnote 23 Second, Social Constructivism’s emphasis on the socially constructed nature of science legitimates a critical discourse on the practices of IR itself. In effect, it legitimates approaches critical not only of the way international relations is practiced, but also of how it is studied and represented by IR scholars. In this way, the discipline’s attachment to particular logics of inquiry and methodology itself becomes of substantive importance because the practices of IR scholars help shape the world they claim to be only observing.
Thus, like the advocates of Instrumentalism in IR, social constructivists are drawing on the PoS in an attempt to make IR better. But, contrary to instrumentalists, social constructivists do not argue for better methods or stricter standards of objectivity. In their view, this would be a futile exercise. Instead, they recognize the role social context plays in the substantive domain of international politics as well as in the practices of scholars. From this perspective, interpretation and critique is the best we can do.
The third major position in the PoS debate is Scientific Realism. While Scientific Realism comes in many guises, all share one basic intuition – that science is capable of getting at the way the world works. In a nutshell, scientific knowledge mirrors reality (Rorty, Reference Rorty1979; Boyd, Reference Boyd2002).
In contrast to both Instrumentalism and Social Constructivism, Scientific Realism prioritizes ontology over epistemology. Scientific realists start with a foundational commitment to the world as knowable in principle, and it is this commitment that defines the scope of legitimate scientific inquiry. The argument is that, since science has achieved progress, a mind-independent world accessible to scientists must exist – and thus epistemological optimism is warranted.
The logic of the scientific-realist challenge to any sort of epistemological skepticism is best captured by the so-called ‘no-miracles argument’, also known as the ‘ultimate argument for realism’ (Musgrave, Reference Musgrave1988; Wendt, Reference Wendt1999: 64–67; Okasha, Reference Okasha2002: 62–66). According to this argument, the tremendous technological success of modern science can only be explained without reference to miracles if scientific theories (including their unobservable elements) correspond to actual entities and mechanisms operating in the world. In essence, Scientific Realism claims to be the only ‘no miracles’ explanation for the technological success of modern science. To the degree that unobservables are part of a theory that constitutes the ‘best explanation’ for the observable evidence, we are warranted in believing these unobservables to be true, rather than thinking about them in merely ‘as if’ terms. Indeed, they must be true; otherwise their repeated explanatory success would be miraculous. Scientific realists thus challenge the epistemic skepticism of both instrumentalists and social constructivists.
Correspondence to reality – both in its observable and unobservable elements – is the scientific-realist criterion for theory-choice. The goal of scientific theories is not only empirical reliability, but truth about (correspondence with) all aspects of the world.Footnote 24
Like Instrumentalism – and in contrast to Social Constructivism – Scientific Realism is committed to (indeed founded upon) the possibility of scientific progress. Science progresses as theories increasingly capture both the observable and hidden aspects of how the world works. Science can and does occasionally get it wrong, but in the long run technological progress justifies accepting scientific theories as true. The goal of science is to represent reality and, with time and testing, science is increasingly successful in achieving this aim.
In recent years, and due mostly to Wendt’s work, Scientific Realism has gained significant traction within IR as the best foundation for a truly question-driven social science. By privileging ontology and rejecting any a priori limitations on what we can know about the world, Scientific Realism is presented as capable of legitimating all IR approaches. Such apparent neutrality has contributed to Scientific Realism being couched as a (post-positivist) solution to the seemingly irreconcilable foundational split in IR between pro-science (positivist) and anti-science (anti-positivist) factions – carrying with it the promise of theoretical and methodological pluralism. Specifically, Scientific Realism promises to bridge methodological divides, placing quantitative, qualitative, and interpretive methods on equal scientific footing. Moreover, Wendt argues that Scientific Realism captures what IR scholars of all stripes already do, allowing him to argue that ‘non-[scientific] realists are usually “tacit realists” in their own scientific practice, and … this only makes sense if realism is true’ (1999: 67, Wendt’s emphasis). According to this line of reasoning, IR scholars have nothing to lose and much to gain by openly endorsing Scientific Realism (Wendt, Reference Wendt1987: 351). Ironically, it is this purportedly non-prescriptive character of Scientific Realism that bolsters the ranks of IR scholars who prescribe it.
Scientific Realism has thus entered the science debate in IR promising to be a via media capable of providing the definitive foundation for the scientific study of international relations. But its promise of epistemic neutrality (i.e., that all approaches to studying IR are equally scientific) hinges on accepting a highly contested claim: that science grants us access to a mind-independent world. Scientific Realism, therefore, is not a via media. If anything, it is a via maxima, hotly disputed in both IR and the PoS.Footnote 25
Table 1 summarizes the stance each of the three main positions takes on the core issues at stake.
Table 1 Main foundational positions in the philosophy of science
The problem with foundations in the PoS
Although Instrumentalism, Social Constructivism, and Scientific Realism are internally coherent, one central problem makes them inadequate as foundations for science: each requires at least one leap of faith, precisely what foundations were meant to avoid in the first place. Because there is good reason to believe that the philosophical problems plaguing the three positions will never be overcome, these leaps of faith will never become unnecessary. If this is so, the quest for a single secure foundation for IR is quixotic. We discuss each position’s core leap of faith below.
Instrumentalism demands a leap of faith in that, while espousing the belief that we can only know that which is observable, it requires certainty about the notion of ‘observability’. Yet, the property of observability is itself not observable, and therefore not scientifically knowable in the context of a philosophical position (Instrumentalism) that limits knowledge to that which can be observed. In other words, there is (using instrumentalist logic) no scientific basis for judging something as observable. ‘Observability’ must be defined a priori, in a pre-scientific way (Fine, Reference Fine1984). A leap of faith is thus required when, as instrumentalists do, the observable/unobservable distinction is used to anchor both a theory of truth (empirical adequacy) and the goal of science (utility and reliability). In effect, the skeptical empiricism that instrumentalists endorse based on the observable/unobservable distinction makes the boundaries of knowledge arbitrary.Footnote 26
Social Constructivism also entails its share of faith. Social constructivists claim that there is nothing more to a truth-claim than the fact that it is accepted as such. But what warrant do we have for believing this? What basis do social constructivists – who emphasize the context-dependence of scientific knowledge – have for making trans-historical claims about the context-dependence of scientific knowledge? Social Constructivism presents itself as an Archimedian point while denying the possibility of just such a position (Okasha, Reference Okasha2002: 89–90). Equating truth with acceptance thus cannot be justified on logical grounds (Fine, Reference Fine2007: 5). In sum, the claim that knowledge is socially constructed requires faith in that, despite the social nature of knowledge, claims about the social construction of knowledge are themselves not socially constructed.
Finally, Scientific Realism is perhaps the foundational position requiring the deepest faith. As we have seen, Scientific Realism has at its core a correspondence view of truth. But this view, despite being intuitively appealing, is plagued by a problem of logical circularity. Recall that Scientific Realism justifies the correspondence between unobservables and reality based on the observation that science is successful. But what explains the success of science? Their answer is that unobservable entities referenced in successful scientific theories must really exist in the world and operate in the ways stipulated by the theory; otherwise, the success of science would be miraculous. In other words, Scientific Realism infers that the best explanation for the success of science is that inference to the best explanation works. But this is using inference to the best explanation to prove the validity of inference to the best explanation – an instance of circular reasoning.Footnote 27 The correspondence view of truth that anchors scientific realism as a distinct PoS is itself without a secure logical foundation. In short, correspondence must be accepted on a leap of faith – faith that the success of science is a function of correspondence and that this, in turn, justifies a correspondence view of truth.Footnote 28
To summarize, each position in the PoS requires its adherents to accept at least one core proposition on a basis of faith. None can claim to offer foundations invulnerable to challenge. It follows that none can fulfill the promise of philosophical foundations. Following the philosopher of science Arthur Fine, we argue that scientific knowledge is unavoidably built on ‘shaky’ ground (1986).
Foundational prudence
The meta-theoretical fragmentation of IR is the result of an open debate in the PoS between three incompatible foundational positions and the deployment of these foundations in IR as necessary for the discipline. Our goal has been to demonstrate that IR cannot be grounded on a single PoS foundation as if philosophers have figured out what science is and we should all take them at their word. As long as the PoS remains a contested discipline, we should not be surprised that the quest for foundations in IR will perpetuate rather than reconcile divisions. Moreover, the cost entailed by the quest for foundations is high. As it stands, the discipline is marked by acrimonious exchanges, an absence of productive dialogue, and the foreclosure of potentially fruitful avenues of research.
There is, however, an alternative – one that dissolves rather than solves the foundational debate. Instead of endorsing a ‘position in’ the foundational debate, IR scholars should adopt a prudent ‘attitude towards’ foundations as a whole. An ‘attitude’ can be distinguished from a ‘position’ in that, while it acknowledges the role of PoS arguments in illuminating scientific problems and challenges, it does not force IR scholars to choose one foundation over another.Footnote 29 Thus, we argue for an attitude of foundational prudence in IR.
Our argument for foundational prudence is built on two empirical planks. First, we have shown that each of the three major philosophies of science – Instrumentalism, Social Constructivism, and Scientific Realism – requires at least one leap of faith to justify its claim to explain science and prescribe its goals and limits. This suggests that each of the available PoS positions can be challenged on philosophical grounds, and therefore cannot serve as a secure foundation for IR. Attempting to find a legitimating warrant for IR as a ‘science’ by recourse to the PoS will fail, for no such single legitimating warrant exists. Furthermore, if the PoS debate were ever to be settled – an unlikely prospect to be sure – IR scholars having hung their theoretical jacket on the wrong philosophical hook would lose their foundation.Footnote 30 Second, the very fact of the ‘science’ debate suggests that the plurality of approaches is not captured by any of the three foundational positions. Substantive contributions in the field are made by scholars coming from a wide range of meta-theoretical perspectives that cannot simply be subsumed under a single philosophical warrant. Limiting the discipline to one of these perspectives would stifle our potential for understanding international relations. In a nutshell, given the divided state of the foundational debate in the PoS, adopting a prudent attitude towards philosophical foundations is the only way to go.
Broadly speaking, foundational prudence requires that we be agnostic about the foundation of IR. It does not, however, commit the discipline to the impossibility of foundations. It is post-foundational, not anti-foundational. As such, foundational prudence recognizes the value of philosophically informed diversity while at the same time abandoning attempts to quash it at the foundational level.
For individual scholars, personal commitments to foundational arguments are fine inasmuch as one finds that a basis of beliefs about the world and how to generate knowledge serves the purposes of one’s research. In fact, one of our purposes is to raise awareness about the plurality of foundational PoS positions, their strengths, and weaknesses. Only by acquiring substantive knowledge of the PoS can IR scholars more coherently and consciously choose their individual foundational commitments, thus avoiding inconsistencies or even contradictions between, on the one hand, such commitments and, on the other, their theoretical and methodological choices. This is a crucial element of our post-foundationalism, distinct both from pre-foundational ignorance and from dogmatic anti-foundationalism.
Problems arise only when IR scholars deploy foundational arguments to show how their scholarship is ‘scientifically’ superior to that of others. We want the discipline to recognize that genuine and productive pluralism demands acknowledging that scholars are confronted with a genuine choice of foundational commitments and that there is no a priori basis on which to recommend one over the other. The only plausible criterion is that of reflecting upon how such commitments contribute to our understanding of international relations.
What, then, does foundational prudence mean for IR as a whole? First and foremost, the post-foundationalist character of our argument entails the rejection of any a priori universal or essential criteria for theory-choice. This emphatically does not mean that anything goes. Rather, it means that broad scientific traditions cannot be chosen or discarded according to previously existing algorithms. (More detailed arguments can, but that is typically done by resorting to theoretical, methodological, and empirical – rather than philosophical – arguments.)Footnote 31 The same applies to method choice. Foundational prudence denies the existence of a necessary connection between particular questions and specific methods. It opens the door to methodological pluralism – encompassing quantitative, qualitative, and interpretive traditions – while admitting that the scope of acceptable methods can be limited by substantive – but not philosophical – argumentation. Like theories, methods should not be chosen based on a priori foundational arguments about their purported ‘scientific’ character. Rather, they should be chosen for their contribution to our substantive understanding of international politics.
Foundational prudence accepts that science is performed in the vast and fertile ground between any reductionist view of truth and subjective matters of taste. Abandoning the former does not necessarily consign us to the latter. In other words, acknowledging the false promise of philosophical foundations does not condemn us to relativism. It is on the ground between these two poles that knowledge claims are – and should be – rationally debated and adjudicated. Foundational prudence thus encourages us to trust IR’s ability to stand on its own, without recourse to philosophical foundations.
Any approach to ‘doing IR’ must prove its own legitimacy by demonstrating its ability to further our understanding of international relations, not by asserting its superiority according to some foundational argument. Standards of scholarship should be defined within the IR community, based on how a particular argument relates to the general topics deemed relevant to the discipline and, given its relevance, how an argument is internally consistent and externally valid, that is, supported by empirical evidence. Philosophical foundations add nothing to the checks, balances, and safeguards the discipline has in place to ensure that the knowledge it produces is scientific. Consequently, scientific standards should not come from outside IR itself.
Endorsing a prudent attitude towards foundational arguments thus means recognizing that scientific standards do not naturally flow from the PoS. IR has its own internal standards and, as many before have pointed out, the individual scholar’s ability to function in the discipline does not seem to rely on having a proper understanding of, or a commitment to, particular positions in the PoS.Footnote 32 The procedures followed by the community of IR scholars to adjudicate between theories, validate observations, etc., should enable us to trust the results of that process, even if we cannot guarantee that, at the end of the day, the product of science is itself objective.Footnote 33
Our argument also bears on the debate about what are the criteria for progress in the field and whether particular research programs are, in light of such criteria, progressive (Elman and Elman, Reference Elman and Elman2003). Foundational prudence is agnostic on scientific progress. The impulse to assess progress in the social sciences stems in part from a desire to establish parity vis-à-vis the natural sciences, where technological development grants (at least the appearance of) both shared goals and their progressive achievement. We find this comparison with the natural sciences problematic. It emphasizes the purportedly unitary character of science, obscuring the diversity of goals, practices, standards, etc., that characterizes scientific activity. It also highlights the inferiority complex that has long plagued the social sciences, which have historically felt the need to emulate their allegedly more successful natural cousins. Progress in IR (as in any other science, social or not) should be measured by standards internal to the discipline, taking into consideration the particularities of the objects of study. Whether progress in IR is possible remains a question open to debate – and foundational arguments cannot settle it.
By making room for a diversity of purposes – to be determined and debated within the field by its practitioners – foundational prudence is also agnostic on the question of policy relevance. Policy relevance is an odd requirement for IR scholarship, but a perfectly fine goal among others. The key criterion is disciplinary relevance.
But perhaps the greatest impact of adopting a prudent attitude towards philosophical foundations would be on disciplinary politics.Footnote 34 PoS arguments have been among the most effective in dividing the field on what is considered legitimate IR scholarship. They have also been key in establishing and sustaining the dominant (positivist) approach to IR. The result is foundational clashes that do little to advance our knowledge of international relations.
It might be argued that foundational debates are useful precisely because they expose issues central to disciplinary politics. But this is fighting by proxy. Political issues – such as the interests motivating different types of IR research, their role in determining funding and appointments, and their impact on international relations – are best addressed head-on, not camouflaged as foundational arguments. Indeed, by fighting politics under the cover of foundations, IR scholars have deepened political divisions in the discipline to the PoS level, making them all but intractable. This is the result of attempting to settle political questions by invoking purportedly unshakable foundational arguments. Doing politics by discussing foundations is unnecessary, dangerous, and overly disciplinary. Instead of bringing politics out, it smuggles politics in.
Foundational prudence de-escalates disciplinary politics by preventing a whole category of powerful weapons – foundational PoS arguments – from being used in the fight. It de-claws foundational disputes, ruling out attempts to debunk alternative IR approaches solely on the basis of their meta-theoretical underpinnings. As long as IR scholars make clear their theoretical, substantive, and methodological choices, while simultaneously recognizing and acknowledging their limitations, the field will function unimpeded by the absence of secure foundations.
IR without foundations
Our argument will no doubt raise many questions among participants in IR’s foundational debate. First, foundational prudence might be seen as anti-philosophical. After all, without definitively refuting any of the foundational PoS positions, we argue for moving beyond foundations and for a cessation of foundational hostilities in IR. Furthermore, the distinction between science and philosophy is admittedly fuzzy. So how can we try to rule out philosophical argumentation in IR?Footnote 35
This criticism, though intuitively appealing, proves unwarranted. We put forth foundational prudence as the proper attitude for IR scholars, not philosophers. Nothing in our argument precludes the latter from continuing to devote themselves to finding a proper foundation for science. But for as long as the philosophical jury is out on what constitutes a proper foundation for science, foundational prudence enjoins scientists to face the shaky nature of scientific knowledge and drop the quest for foundational certainty. Positions contested in the PoS should not be deployed as definitive solutions to the ‘science’ debate in IR. In other words, they should not be turned into imperial foundational projects.
Furthermore, we are not arguing against philosophical argumentation per se in IR. Indeed, we have no problem with scholars making use of PoS arguments as part of the toolkit with which they try to understand international relations. Foundational prudence makes room for the use of the PoS in IR in non-imperial ways. Fleshing out one’s foundational commitments may facilitate a better understanding of the assumptions, goals, and implications of an entire research program or a particular piece of scholarship. It can also highlight inconsistencies (MacDonald, Reference MacDonald2003). (Whether it does is, in each case, an empirical question, which cannot be answered a priori.) Indeed, we believe one of the benefits of foundational prudence is the opening of more constructive and creative ways of using PoS arguments, potentially leading to new avenues of IR research.
A prudent attitude towards philosophical foundations is well served by philosophical awareness, which can only be engendered through knowledge of the PoS literature. What we oppose is the attempt to apply one’s foundational commitments as a basis for determining the scope of scientific IR. In this sense, our point is to push responsibility for determining the boundaries of science back on to the scientists themselves, and to urge them to do so without resorting to philosophical foundational arguments as legislative tools. We want ‘to let the chips fall where they may’ and see how we get going (Fine, Reference Fine1996: 13).
Second, critics might portray foundational prudence as anti-scientific. Isn’t it the case that, by undermining any candidate to serve as a proper foundational for science, we are undermining the privileged status of science itself? By depriving IR of its unshakable foundations, how can we account for its scientific status?Footnote 36 Once again, this criticism proves wrongheaded. It assumes that the current foundations of IR are unshakable, and that we would therefore be on shakier ground by abandoning them. But, as we have demonstrated, IR’s foundations are shaky. They add nothing to the substantive, methodological, and empirical justifications we may have for its privileged role as a science. As Wilfrid Sellars (Reference Sellars1963: 167) wrote, ‘science is rational not because it has a foundation, but because it is a self-correcting enterprise which can put any claim in jeopardy, though not all at once’.
So, what would adopting foundational prudence mean for IR scholars with a stake in a particular PoS foundation?
Scholars committed to Instrumentalism would have to refrain from dismissing arguments that use unobservable phenomena – in many respects the explanatory hallmark of social science – solely on a PoS basis. Instead, they would have to challenge such arguments on theoretical, methodological, or empirical grounds. If no such objections can be found, even committed instrumentalists should take seriously explanations that include unobservable entities or reveal the social basis of empirical reliability. Such an adjustment may seem relatively minor; but, if taken seriously, the instrumentalist mainstream of American IR scholarship would no longer be able to ignore the plurality of scholars that pursue questions, endorse theoretical perspectives, and make methodological choices incompatible with instrumentalist views.
Social constructivists would perhaps be those who would have the most difficult time accepting our argument. This is because Social Constructivism – which is predicated on the view that all knowledge is a function of social processes – has as its central purpose to unmask the social and political forces that constitute discourses as reasonable, true, and dominant. From such a perspective, our move, which upholds the possibility (though not the obligation) of taking scientific knowledge as objective, will inevitably evoke a debunking response on the grounds that we are ourselves naïvely fashioning a position with important political implications for the discipline. Social constructivists are likely to see in foundational prudence a weapon the positivist mainstream will be able to deploy against their critical approach. In fairness to Social Constructivism, we recognize that this is inevitably the case with any argument. In fairness to us, we ask social constructivists to recognize that, although politics is inextricably part of any discipline, this does not take away from the force of our claim about the shakiness of all PoS foundations – including Social Constructivism. We have no doubt that our argument can be scrutinized through the lens of a meta-discourse on foundations, and that such a meta-discourse can itself be analyzed through a meta-discourse about meta-discourses, and so on and so forth. But we disagree that such an escalation is productive with respect to our argument’s stated purpose – a truly pluralistic IR.
Finally, scientific realists, especially those who uphold this position as a result of Wendt’s influential advocacy for it, will be inclined to believe Scientific Realism already accomplishes what we claim is the payoff of our ‘post-foundational’ attitude. Why not adopt Scientific Realism, they could argue, since it entails epistemic relativism, making space for a question-driven IR? To take foundational prudence seriously, scientific realists would have to acknowledge two points. First, that the epistemic optimism that characterizes Scientific Realism – the notion that we get things right, and that, when we do, we get at the way the world works – faces unresolved philosophical challenges that will likely prove intractable. Second, that, pace Wendt, not all IR scholars are scientific realists – not even tacitly. Many are self-professed, or implicit, instrumentalists and social constructivists, upholding foundational beliefs that are fundamentally incompatible with the core tenets of Scientific Realism. Asserting that all IR scholars are at least tacit realists is to assume the problem away. It amounts to claiming that were we all to agree (on Scientific Realism as the proper foundation for IR), we would all agree – a tautological truth, and one that fails to reflect the divisions among IR scholars that are the root cause of the foundational debate itself. Foundational prudence captures the normative intent motivating advocates for Scientific Realism without Scientific Realism’s meta-theoretical baggage.
In sum, a post-foundational IR rules out IR cookbooks. The core message of this article is that there are no universally accepted PoS recipes for how international relations should be studied. Philosophical arguments cannot and should not be used to legislate which research questions – or ways of answering them – are legitimate in IR. ‘Bad science’ can be debunked on theoretical, methodological, or empirical bases. Attacking scholarship on the basis of its failure to conform to a particular foundation adds nothing to the arsenal of tools IR scholars have at their disposal to weed out deficient arguments. Worse, it creates the temptation to dismiss on foundational grounds scholarship that does make a meaningful contribution. The mere fact of there being (at least) three reasonable yet incompatible foundational arguments suggests to us that a pluralism based on a single one-size-fits-all foundation is destined to fail – a claim supported by the history of the PoS debate in IR.
Conclusion
PoS arguments continue to shape IR by (i) granting the illusion of foundational security to the dominant modes of scholarship and, thus, (ii) requiring any strands of scholarship that aim at challenging the mainstream to advance an alternative foundation. In this sense, foundational arguments are what Joseph Rouse (Reference Rouse2002: 63) calls ‘vampires’: ‘philosophical undead that still haunt our concepts and interpretations of nature, culture, and science’. Foundations are arguments that have never been proven or disproved (and are indeed undecidable), but which emerge from time to time to haunt new, live ideas – in this case, ideas about international relations. If foundations are vampires, the purpose of this article is apotropaic.
The Sisyphean effort to place IR on unshakable PoS foundations is at least in part an attempt to boil down the complex relation between IR and international relations to the supposedly more straightforward relation between social science and philosophy. In other words, it is an attempt to turn a political problem into a philosophical/foundational one.Footnote 37 For proponents of this move – and they come from all foundational positions – knowledge produced by a science of IR that stands on unquestionable foundations can claim the status of a special discourse about international relations, one that cannot be challenged.
This reduction, however, is an over-simplification. There is no way to strip the relation between IR and international relations from its political dimensions. There is no cookie-cutter way to deal conclusively with this complex issue in an apolitical way – a questionable goal in itself. Foundational prudence recognizes the intrinsically political character of the relation between IR and international relations, and does not try to claim a privileged status for the scientific discourse produced by IR scholars.
This may seem bad news for advocates of ‘scientific’ IR, who demand a set of rigorous standards by which work in the field can be objectively evaluated, and theories, methods, etc., endorsed or discarded. But such concerns miss the point. Despite much anxiety the prospect of an IR without foundations may cause, foundations have not provided any firm bases for standards, have not settled any methodological debates, have not conclusively decided any empirical question, and have not managed to guarantee the policy relevance of our scholarship. What they have successfully accomplished is the unproductive division of the field along artificial fault-lines. By working as philosophical underlaborers, we hope to have cleared the foundational brush in order to make room for a more integrated discipline of IR and prevent another ‘great’ foundational debate. Great debates should be about substantive questions, not about inevitably shaky meta-theoretical positions. To put our argument in the simplest terms, we learn how best to do IR from doing IR, not philosophy (Schmidt, Reference Schmidt1997: 111).
Many have called for a question- or problem-driven IR. If this is indeed our goal – and we believe it should be – it is a mistake to think that the relationship between questions, approaches, methods, theories, etc., can somehow be specified a priori in order to secure the status of what we produce as scientific knowledge. The goal of a plural, inclusivist IR cannot be unshakably grounded. That is the false promise of philosophical foundations.
Acknowledgements
For helpful comments and suggestions on previous drafts, we would like to thank Daniel Abebe, Arthur Fine, Todd Hall, Anne Holthoefer, Rosemary Kelanic, Matthew Kocher, Adria Lawrence, Charles Lipson, Kevin Narizny, Jacob Schiff, Duncan Snidal, Matthias Staisch, Ian Storey, Lisa Wedeen, Alexander Wendt, the discussants and audiences at the University of Chicago PIPES workshop, the ISA and APSA 2007 conferences, and three anonymous reviewers.
