What are philosophical foundations?

The disciplining gesture that the authors (and I) oppose is inextricable from a thorny philosophical problem. The only reason why anyone in IR now cares about the ‘scientific’ status of an empirical claim is that for the past several hundred years, European philosophers have held science to enjoy a privileged status when it comes to the production of knowledge. Scientific research was held to be a procedure that would allow investigators to clear away arbitrary assertions and replace them with better-grounded factual statements. This was part of the broader Enlightenment project of attempting to place everything – politics, morality, and knowledge itself – on a firmly reasonable basis, and to discard anything that could not be justified on that basis. Rene Descartes’ famous declaration of intent to ‘withhold my assent no less carefully from opinions that are not completely certain and indubitable than I would from those that are patently false’ (Descartes, Reference Descartes1993: 59) serves as a good expression of the methodological skepticism that both posed the problem of foundations in its most acute form and foreshadowed its solution. Scientific inquiry, proceeding from the principle of doubting everything, would eliminate false claims and replace them with true ones.

But the problem is that Descartes did not succeed in providing absolutely secure foundations that could stand up to critical examination; in fact, he needed to rely on an extremely convoluted argument about the existence of God to restore his confidence in his knowledge of an external world in the first place. This failure led to what Richard Bernstein refers to as ‘Cartesian anxiety’, the fear that without absolutely secure foundations knowledge will collapse into ‘intellectual and moral chaos’ (Reference Bernstein1992: 17). This fear is made particularly acute by the specific character of the challenge confronting the knowing Enlightenment subject to ensure that her or his knowledge accurately captures the way that things are – or, put slightly differently, to make sure that the knowledge produced by the mind corresponds to the nature of mind-independent reality. This is a dualist formulation of the problem of knowledge, relying on a presumption of a separation between the mind and the world for its acute philosophical urgency. The fear is that without firm foundations we will be entrapped in fanciful worlds of our own imagination or delusion – mind, completely divorced from the world. This fear engendered several centuries of speculation about whether sensory inputs (Hobbes), natural law (Locke), habit (Hume), a critique of pure reason (Kant), or spirit (Hegel) – to name only a few options – were sufficient to bridge the gap between the mind and the world, and place knowledge of the world back onto firm foundations.

Philosophical foundations for knowledge, then, are intimately connected with the specific problem of guaranteeing or warranting knowledge of a mind-independent, external world. Put another way, foundational claims for knowledge are designed to solve a problem of traditional epistemology, allowing us to trust our knowledge in the face of skeptical doubts. This is only an urgent problem if one is confronting Cartesian anxiety, and trying desperately to ground even the grounds of knowledge lest one slip away into a subjectivist abyss. The quest for philosophical foundations of knowledge is thus a specific cultural and intellectual project of the European Enlightenment, driven by a loss of confidence in the power and validity of established systems of social practice (MacIntyre, Reference MacIntyre1984: 37–39). It is only from such a perspective that foundational claims appear to be both urgently needed and difficult to provide.

Only one of Monteiro and Ruby’s three PoS positions easily qualifies as ‘foundational’ in this specific and restricted sense: Instrumentalism,Footnote 3 which appears in their account as a somewhat confused amalgam of empiricism, covering law explanations and predictions, an analyticist as-if attitude towards the question of the reality of theoretical entities and falsification. Despite the authors’ cautionary note that they are ‘glossing over myriad distinctions that exist and continue to emerge in the philosophical debate’ (p. 26), this is a very mixed bag of commitments; any position that can pass seamlessly from empiricist skepticism to Waltz’s more ideal-typical form of theorizing is clearly not placing much of a premium on logical or philosophical coherence. Regardless, Instrumentalism is clearly a foundational position, since it begins with a dualist formulation of the problem of knowledge and seeks to provide a warrant for our empirical knowledge of the external world in the procedures of providing operational definitions, looking for observable consequences of theoretical deductions, and regarding one’s theoretical concepts as merely a useful way of thinking about the world. If one accepts Instrumentalism, then knowledge can be founded on practices of good empirical research – and other kinds of claims about the world can be justifiably dismissed as part of the self-policing of the scientific endeavor.

Scientific Realism presents a slightly more challenging case. As Monteiro and Ruby correctly point out (pp. 30, 31), Scientific Realism rests on a transcendental argument seeking to explain the evident empirical success of scientific endeavors as something other than a miracle. Scientific Realists thus respond to Cartesian anxiety by pointing to the many evidently successful efforts to generate valid knowledge of the world, a response that shifts the philosophical problem just slightly: instead of providing a warrant for knowledge claims, Scientific Realists advance an argument about the possibility of knowledge. But Monteiro and Ruby point out that accepting this formulation still requires something of a leap of faith in a knowable external reality (p. 34), and still engenders critiques of alternative positions for their alleged failures to acknowledge the reality of the referent of the theoretical terms with which they operate (p. 31). These moves only make sense within a dualist context (Jackson, Reference Jackson2008b: 138–143). As such, even though Scientific Realism is not quite the same kind of foundational claim as Instrumentalism, they are broadly similar enough to fit into the same category.

One of these things is not like the others

Not so the third PoS position, Social Constructivism.Footnote 4 A quick glance at the table of PoS positions that Montiero and Ruby draw up (p. 33) illustrates the point: while Instrumentalism and Scientific Realism both rely on observation and have a progressive view of science, Social Constructivism does not. In addition, looking along the ‘Notion of truth’ row, while the ‘correspondence to reality’ upheld by Scientific Realism is an elaboration of Instrumentalism’s ‘empirical adequacy’, Social Constructivism’s ‘accepted belief’ seems like a horse of a very different color. In fact, the Social Constructivist position seems wholly unconcerned with warranting or grounding knowledge in a philosophical way, and thus appears to do nothing to render knowledge justifiable in any ultimate sense.

Indeed, despite the authors’ assertion (in footnote 14) that Social Constructivism differs from the sociology of scientific knowledge in that the former provides normative prescriptions about how science should operate while the latter does not, it is difficult to see any such prescriptions in the work of anyone that they cite as taking the Social Constructivist position. It is notable that the best Social Constructivist prescription that the authors can come up with is a vague appeal to the importance of context (p. 30). It is also notable that the authors’ main PoS source for the articulation of a Social Constructivist position is an encyclopedia article written in 2002 on scientific realism by the scientific realist Richard Boyd. Thus, the authors’ claim that Social Constructivists are skeptical ‘about the human ability to know the world independently of the social context that gives it a particular and contingent meaning’ (p. 28) replicates the way that Scientific Realists generally understand the Social Constructivists’ claim about the conceptual relativity of knowledge. Reading this way, Social Constructivism would appear to be based on a consensus theory of truth,Footnote 5 and to lead to the implication that Social Constructivism is a form of radical subjectivism that grounds knowledge on unconstrained individual interests.

But this is definitely a slanted reading. Monteiro and Ruby claim that all Social Constructivists share an ‘unmasking impulse’, whereby ‘any positivist foundation’ is demonstrated to be ‘untenable’ (p. 29). But even if this were the case, it would not follow that such a demonstration constitutes a foundation for knowledge! Instead, it dissolves the philosophical problem of foundations, redirecting our attention to the somewhat different issue of how particular claims come to be regarded as true. An investigation into how any particular claim comes to have the status of ‘truth’ – an investigation that must be empirical rather than conceptual, since the question asked concerns the social status of a claim rather than its ultimate worthiness – has no bearing on whether that claim is properly grounded philosophically. As such, it is deeply problematic to equate the kind of philosophical foundations claimed by Instrumentalism and Scientific Realism with the institutions and social practices cited by Social Constructivists as part of their explanation for how claims come to be taken as true. If there are ‘foundations’ in Social Constructivism, they are operative foundations, parts of an empirical account rather than justifications for how valid knowledge can and should be produced.

The authors also fail to appreciate the fact that the Social Constructivist dissolution of the traditional Enlightenment problem of philosophical foundations places Social Constructivism outside of the entire debate about the status of ‘science’. It is not the case that Social Constructivists ‘deny the possibility of objectivity and therefore science’, (p. 17) unless one (wrongly) interprets Social Constructivism as a form of subjectivism – which is to say, unless one maintains that Social Constructivists begin with a dualist formulation of the problem of knowledge. The subjectivist interpretation of Social Constructivism presumes the traditional Cartesian problem of a mind separate from a mind-independent world, and poses the problem of knowledge as a problem of bridging the mind–world gap. If one starts with this Cartesian problem, Social Constructivism does look like an argument in favor of letting the mind-independent world go and staying on the ‘mind’ side of the mind–world gap.

But Social Constructivists do not start with the Cartesian problem and its dualist formulation of the problem of knowledge, a fact that Monteiro and Ruby acknowledge in their discussion of how Social Constructivists find the ontology–epistemology distinction meaningless (p. 28), but then forget in treating Social Constructivism as an effort to answer the same question as Instrumentalism and Scientific Realism. The supposed Social Constructivist leap of faith about ‘the context-dependence of scientific knowledge’ (p. 34) is cashed out as a denial of the possibility of successfully crossing a mind–world gap that Social Constructivists do not, in fact, posit in the first place. Instead of beginning with dualist presuppositions, Social Constructivists operate with the monist commitments: knowledge is always and already implicated in the world, and as such cannot be sensibly thought of as referring to some kind of mind-independent reality (Jackson, Reference Jackson2008b: 133, 134, 146–149). Social Constructivists’ accounts of how scientific knowledge (or, more generally, a claim taken to be ‘true’) is practically produced should not be read as critiques of science or as efforts to place science on firmer foundations (pace p. 30), but as consequences of the fact that philosophical validity cannot be taken to be either necessary or sufficient to explain a claim’s social status. Instead, the Social Constructivist wager is that knowledge claims have operative foundations – processes and practices that give them particular social statuses – but this says nothing about any philosophical foundations that those claims might (or might not) have.

Towards a genuinely post-foundational IR

With these distinctions in mind, consider Monteiro and Ruby’s proposal for an IR that can ‘stand on its own, without recourse to philosophical foundations’ (p. 37). They suggest that the failure of any of the existing PoS positions to provide unquestionably secure foundations is a sufficient warrant for what amounts to a mutual demilitarization pact. Partisans of particular PoS positions would need to accept the legitimacy of other positions, and agree to give up the exclusive claim to the label ‘science’ – because, after all, their own claims to truly be a science are no more secure than the claims of partisans of other positions. Monteiro and Ruby thus call for an IR freed from the demarcation problem – the problem of distinguishing science from non-science (Lakatos, Reference Lakatos2000: 22–24), and thus, in principle, free of the disciplinary politics and practical boundary work (Taylor, Reference Taylor1996: 88, 89) surrounding the designation of any particular piece of work as ‘scientific’. Instead, they argue that IR scholars should be discussing ‘how an argument is internally consistent and externally valid, that is, supported by empirical evidence’ (p. 37), and hence whether the argument enhances our understanding of world politics. The phrase ‘externally valid’ here is striking, since it seems to point in the direction of the traditional Cartesian problem of knowledge and its dualistic concern with relating knowledge (mind) to reality (world). This in turn seems to tilt their putatively pluralistic solution away from any non-dualist conception of knowledge, including Social Constructivism.

Further evidence for this bias can be glimpsed in the authors’ call for specific adjustments by partisans of all three PoS positions (pp. 41–42). To contribute to an IR based on ‘foundational prudence’, Instrumentalists would have to admit the possibility of real-but-unobservable entities, Scientific Realists would have to back off from their ‘epistemic optimism’ and stop claiming that ‘all IR scholars are at least tacit realists’, and Social Constructivists would have to accept ‘the possibility (though not the obligation) of taking scientific knowledge as objective’. The adjustments asked of Instrumentalists and Scientific Realists are not philosophically onerous, especially given the protorealism of some of the seminal logical positivist (Hempel, Reference Hempel1965) and post-positivist (Popper, Reference Popper1979)Footnote 6 philosophers when it comes to unobservable entities, and given that Scientific Realists are in effect being asked to tone down their public rhetoric about their superior foundational solution (note that they are not asked to give up that solution, just to stop imposing it on everyone else). But the Social Constructivists are asked to abandon their entire approach to knowledge-production, since they are required to accept a dualistic formulation of the problem of knowledge and countenance the possibility that their empirical claims are unwarranted speculations – unlike the ‘objective’ scientific knowledge they are required to make room for. Monteiro and Ruby’s understatement – ‘Social Constructivists would perhaps be those who would have the most difficult time accepting our argument’ – does little to soften the blow. The purportedly level-playing field of foundational prudence seems slanted in a decidedly Scientific Realist direction.

Monteiro and Ruby’s post-foundational IR, then, seems more like (non)foundational IR; foundational in as much as it has a tacit bias toward Scientific Realist foundations, and non-foundational in that it would prevent anyone from raising philosophical questions about foundations in the future. So, what would a genuinely post-foundational IR that contained room for both dualist and monist positions on knowledge production look like? One clue might be found in Vincent Pouliot’s presentation of a post foundationalFootnote 7 IR constructivism centered on social facts – IR ‘constructivists have no need to be foundationalist because social agents already are’, Pouliot argues, and because ‘agents continually essentialize reality’, it is possible to observe their practices without generating philosophically foundational accounts of the outcomes (Pouliot, Reference Pouliot2004: 330; see also Brubaker, Reference Brubaker2006). It follows that analysts can side-step the issue of philosophical foundations by bracketing the validity or invalidity of foundational commitments in favor of an empirical analysis of what the commitments do in practice – what pragmatists would call the ‘cash value’ of those commitments (James, Reference James1978: 430; see also Hellmann, Reference Hellmann2009).

Although Pouliot advocates this bracketing largely when dealing with the social agents under study in our empirical work, it is perhaps even more important that we bracket when dealing with other IR scholars. Such an abandonment of the futile quest for a universally compelling set of philosophical foundations of knowledge allows us to recognize even more clearly that foundational commitments have a profound effect on how particular scholars carry out their empirical work. Scientific Realists do not engage in research in the same way as Instrumentalists or Social Constructivists do, and remaining strictly and pragmatically non-committal about all of these claims may allow us to appreciate these differences more clearly – and to critique any instance of empirical social–scientific work when it seeks to overrun its proper boundaries and tries to imagine itself a purely atheoretical account of reality (Jackson, Reference Jackson2008a). In this way, both dualism and monism can be treated as working assumptions or wagers (Jackson and Nexon, Reference Jackson and Nexon2009), and evaluated for their analytical productivity rather than in terms of their ultimate philosophical validity.

Whether the result of such a bracketing of philosophical validity leads to a single ‘set of foundational assumptions that can consistently on its own terms give rise to the pluralism that characterizes international relations’ (Bially Mattern, Reference Bially Mattern2008: 696) or to a typology of philosophical and social theoretical foundations that gives IR scholars commonplaces about which to disagree (Jackson, Reference Jackson2009) is, to my mind, an open question. But this kind of a post-foundational IR would, I think, better fulfill Monteiro and Ruby’s vision of an IR in which scholars would stop deploying ‘foundational arguments to show how their scholarship is “scientifically” superior to that of others’ (p. 36), because it would provide a built-in check against any such philosophical hubris. To dispel the disciplining gesture based on strategic misreadings of the philosophy of science, we need sufficient critical consciousness of different PoS positions along with their practical strengths and limitations. This is unlikely to be achieved by ruling out philosophical argumentation in IR, but it might be achieved by redirecting that argumentation in a more pragmatic direction.Footnote 8

As it stands, however, the place that Monteiro and Ruby leave us is somewhat depressingly familiar. By calling for empirical work that does not explicitly consider its own philosophical foundations, they imply that the debunking of poor arguments ‘on theoretical, methodological, or empirical bases’ (p. 42) can proceed in the absence of any explicit consideration of philosophical issues – as if IR scholars would naturally and easily come to consensus about good and bad arguments. Theory and methodology without an explicit consideration of the philosophical issues involved in knowledge production become nothing but ways of organizing empirical data, and thus evaluating any argument becomes purely an empirical question. Such an empirical evaluation requires a strict separation of theories, methods, concepts, methodology, etc. from empirics; gone is the widely shared PoS notion that perception and observation are themselves theory-laden activities. But this return to a kind of flat-footed empiricism justifies not ‘a plural, inclusivist IR’ (p. 44), but a neo-behaviorist IR completely devoid of anything but endless correlations of causal factors in a probably futile (Alker, Reference Alker1966) effort to cumulatively generate knowledge. Such behaviorism would be foundational, not post-foundational, as it would reject out of hand any argument that did not conform to its very narrow understanding of proof and disproof. I doubt very much that Monteiro and Ruby would like to inhabit that kind of an IR field; but, in banishing discussions of philosophical foundations, they leave us no real alternative. IR does not need the repeated dismissal of arguments for supposedly not being properly ‘scientific’, but it most certainly does need discussions of the foundations of knowledge.