THE IDEAS DEBATE IN INTERNATIONAL AND EUROPEAN STUDIES: Towards a Cartography and Critical Assessment Andreas Gofas & Colin Hay 2008/11 Andreas Gofas Institut Barcelona d'Estudis Internacionals (IBEI) agofas@ibei.org Colin Hay Department of Politics University of Sheffield C.Hay@sheffield.ac.uk ISSN: 1886-2802 IBEI WORKING PAPERS 2008/11 The Ideas Debate in International and European Studies: Towards a Cartography and Critical Assessment © IBEI, de esta edición Edita: © Andreas Gofas y Colin Hay CIDOB edicions Elisabets, 12 08001 Barcelona Tel. 93 302 64 95 Fax. 93 302 21 18 E-mail: publicaciones@cidob.org URL: www.cidob.org Depósito legal: B-21.147-2006 ISSN:1886-2802 Imprime: Color Marfil, S.L. Barcelona, January 2008 THE IDEAS DEBATE IN INTERNATIONAL AND EUROPEAN STUDIES: TOWARDS A CARTOGRAPHY AND CRITICAL ASSESSMENT Andreas Gofas & Colin Hay Abstract: The appeal to ideas as causal variables and/or constitutive features of political processes increasingly characterises political analysis. Yet, perhaps because of the pace of this ideational intrusion, too often ideas have simply been grafted onto pre-existing explanatory theories at precisely the point at which they seem to get into difficulties, with little or no consideration either of the status of such ideational variables or of the character or consistency of the resulting theoretical hybrid. This is particularly problematic for ideas are far from innocent varitendency along with the limitations of the prevailing Humean conception of causality, and assoables - and can rarely, if ever, be incorporated seamlessly within existing explanatory and/or constitutive theories without ontological and epistemological consequence. We contend that this ciated epistemological polemic between causal and constitutive logics, continue to plague almost all of the literature that strives to accord an explanatory role to ideas. In trying to move Humean, conception of causation. If we reject this in favour of a more inclusive (and ontologically realist) understanding then it is perfectly possible to chart the causal significance of constitutive processes and reconstrue the explanatory role of ideas as causally constitutive. beyond the current vogue for epistemological polemic, we argue that the incommensurability thesis between causal and constitutive logics is only credible in the context of a narrow, Key words: Ideational Analysis, Causal and Constitutive Processes, Ontology, Epistemology, Rational Choice Theory, Constructivism, Morphogenesis, Uncertainty. -3- 2008/11 • IBEI Working Papers 1. Introduction The appeal to ideas as causal variables and/or constitutive features of political processes increasingly characterises international and European studies. Yet, perhaps because of the pace of this ideational intrusion, it is undoubtedly still the case, as it was in 2001, that “studies of ideational variables add up to less than the sum of their parts” (Berman, 2001: 231). Too often ideas have simply been grafted onto pre-existing explanatory theories at precisely the point at which they seem to get into difficulties, with little or no consideration either of the status of such ideational variables or of the character or consistency of the resulting theoretical hybrid. This is particularly problematic for, as we shall see, ideas are far from innocent variables – and can rarely, if ever, be incorporated seamlessly within existing explanatory and/or constitutive theories without ontological and epistemological consequence. In effect, we suggest, the burgeoning literature on the role of ideas has tended to lack solid, coherent and explicitly stated theoretical underpinnings. As a consequence, and despite the proliferation of ostensibly ‘ideational approaches’, there remains little clarity about just what sort of an approach an ideational approach is, about the range of ontological and epistemological positions with which such an approach might be compatible, and about what it would take to establish the kind of fully-fledged ideational research programme many seem to assume they have already developed. These questions and dilemmas provide the context and backdrop for the present contribution. Our aim in this paper is at least to begin to develop a cartography of existing ideational approaches and to submit these to critical assessment by teasing out their often implicit ontological and epistemological assumptions. To be clear from the outset, our aim is not to engage in ontological proselytising (on the dangers of which see Hay 2005). Though we defend against potential objections the critical realist perspective we ultimately prefer, we see this as but one of many (potential) ways to take forward ideational analysis in political science and international relations. In this respect, though arguing in general for a recognition of the significance and salience of ideational factors, we welcome theoretical plurality and diversity when it comes to dealing with such ideational factors. Yet plurality is one thing, eclecticism another. Too much ostensibly ideational scholarship in political science and international studies, we argue, is ontologically and/or epistemologically inconsistent – and it is so precisely because it does not reflect adequately on the consequences of ‘taking ideas seriously’. In order to differentiate between theoretical diversity (a good thing) and theoretical inconsistency/eclecticism (a bad thing), we need to be able to isolate, to spell our clearly and to assess the internal consistency of the ontological and epistemological assumptions underpinning various approaches to ideational analysis. That is the principal task of the present contribution. It entails opening up the ‘black box’ of constructivism – a term which has become something of a catch all concept spanning all ideational analysis and which, as Onuf rightly suggests, is “in danger of becoming all things to all scholars, finally suffering the fate of all fads” (2002: 130). It is in recognition of this potential danger and out of a desire to -4- A second ambition of this paper is to take us beyond the current vogue for epistemological polemic, typically pitting causal versus constitutive logics. This, we suggest, has tended to produce rather more heat than light and has tended also to fragment the discipline. Despite the undeniable importance of epistemological considerations and despite our own emphasis on the need for consistency in such matters, we suggest that the frequently claimed incommensurability of causal and constitutive logics is misplaced. More specifically, we argue, the incommensurability of causal and constitutive logics is only credible in the context of a narrow, Humean, conception of causation. If we reject this in favour of a more inclusive (and ontologically realist) understanding then it is perfectly possible to chart the causal significance of constitutive processes.1 The third purpose of this paper is to elaborate and defend such a position. Here we turn to the promise (still, to a significant extent, unrealised) of critical realism in the development of an approach to political analysis that is both sensitive to the significance of constitutive logics yet unapologetically explanatory in its concerns. In so doing we build from a critical engagement with the ‘morphogenetic approach’ of Margaret Archer. Given that critical realism and the morphogenetic approach have make sense of the diversity of approaches to ideational analysis and explanation that this paper proceeds by way of a stock-taking exercise of existing approaches. In so doing our aim is threefold. First, and as an immediate corollary of the above, we wish to challenge the overhasty and unfortunate tendency to lump together all existing ideational approaches as if they presented a single alternative to a equally singular (positivist and materialist) mainstream. This exercise in retrieval is not an easy one, as the distinctions do not involve stable and fixed theories that can be directly contrasted. In effect, it risks creating ideal-type categorisations that obscure, rather than illuminate, fruitful tendencies in research. Theories and theorists are never as neatly compartmentalised as cartographical narratives would have them to be. Moreover, the allocation of different research approaches to discrete intellectual boxes is an inevitably artificial process fraught with subjective bias (Jarvis, 1998: 107). As Jacobsen (2003: 39) ruefully notes, “the exercise can be a subtly negating one that simply fits an adversary’s argument inside one’s own framework in order to tell them what they meant to say if only they had sufficient rigour and wit to do so”. We do not pretend to overcome these problems; we simply wish to acknowledge from the outset the limitations inherent in such an exercise. But, such issues notwithstanding, their continued use throughout the social sciences surely bears testimony to the utility of schematic typologies when deployed as heuristic tools. It is the contention of this paper that some inevitable simplification is a price well worth paying for a sharper and more focused delineation of the ontological and epistemological issues involved in gauging the significance of ideational factors. As such, we make no apology for our inevitable failure to reflect the full subtlety and range of existing ideational scholarship; nor do we claim that our review is exhaustive. IBEI Working Papers • 2008/11 1. Realism here refers to the philosophical position (known also as critical or scientific realism) and is not to be conflated with realism in International Relations theory and political science. -5- had a relatively limited influence in the ‘ideas debate’, some might object to its inclusion in an exercise whose main purpose is cartographic. However, it is included here two straightforward reasons. First, the critical realist prioritisation of ontology reminds us that this is the level at which theories genuinely clash. Consequently, it may contribute to a redirection of the focus away from the current “epistemological paradigm wars that have riven the field in the past decade” (Wendt, 1998: 102). Second, we suggest, some of the core analytical concepts of critical realism (that of ‘emergence’ and the appeal to a non-Humean conception of causality, in particular) offer a particular purchase on ideational variables. After the preceding preliminaries, it is important to establish the organising principals which underpin our mapping of the diversity of existing ideational approaches. As noted elsewhere, the controversy over the role of ideas in political analysis “tends to resolve itself into the question of whether ideas should be accorded a causal role independent of material factors or not” (Hay 2002: 205). This controversy is actually double faceted – with distinct (if, of course, related) ontological and epistemological dimensions. 2008/11 • IBEI Working Papers The ontological controversy concerns the relationship between the material and the ideational. Positions adopted range from simple views, such as idealism and materialism, which assume a rigid ontological distinction between material and ideational factors and privilege one or other moment, to more dialectical approaches, like constructivism and critical realism, which posit the complex interdependence of the two. The simple view is predicated on an ontological or ontic dualism. This conceives of social reality in terms of a rigid division between the separate and opposing realms of the ideational and the material. The dialectical view, by contrast, posits either an ontic or an analytical duality. Whether this duality is ontic or analytical, the realms of the ideational and the material are seen to have no independent existence their existence is relational and dialectical. For those committed to an ontic duality, the ideational and the material are nonetheless real and distinct realms of the social. Yet for those committed to an analytical duality, the distinction between the material and the ideational is itself an abstraction. It is an analytical device rather than an appeal to innate constituent units of social reality. The epistemological controversy concerns the appropriateness of the appeal to ideas in causal and constitutive logics a mirror image of the explaining versus understanding controversy (see, for instance, Hollis and Smith 1990). The distinction is well captured by Suganami (2002: 32). While a causal logic describes a ‘mechanistic relation of production’, constitutive logics describe a ‘relation of logical or conceptual enablement’. In other words, within a causal logic ideas are treated as distinct ‘variables’ whose power can be established only by demonstrating a mechanistic and autonomous effect of ideational factors on specific (political) outcomes. Within a constitutive logic, by contrast, ideas provide the discursive conditions of possibility of a social or political event, behaviour or effect. A very similar rendition of the distinction between causal and constitutive logics leads Wendt (1998; 1999: 79-88) to argue that constitutive logics violate two central assumptions of causal explanation – namely, the independent existence of, and -6- This already serves to indicate why it is that Wendt sees the constitutive and the causal as independent and mutually exclusive. Consider the following example. We might be tempted to argue that the (supposedly) anarchic character of the international system causes the state to adopt an identity of enmity towards other states. However, since the structure of the international system cannot be separated, either ontologically or chronologically, from the identity of state, this cannot be conceived – in Humean terms – as a causal relationship. This is an important point. For it serves to indicate how the commensurability or incommensurability of causal and constitutive logics depends on one’s understanding of causation. This is a theme to which we return in greater detail presently, but it is important to introduce it here for one simple reason. In the following section, in which we categorise and review approaches to ideational analysis, we replicate the stark distinction drawn in the existing literature between constitutive and causal approaches to ideational analysis – considering each separately. Yet it is important to emphasise that, in so doing, we are by no means committing ourselves to such a dichotomisation of the causal and the constitutive. Indeed, towards the end of the paper, where we set out our own perspective, we reject precisely such a rigid distinction between the causal and the constitutive, arguing for the causal significance of constitutive logics. In the process we come to reject the Humean conception of causation which both dominates and, we suggest, continues to hold back ideational analysis in political science and international relations. the temporal asymmetry between, cause and effect. Yet this, we contend, relies on a narrow – if nonetheless highly conventional Humean understanding of causality. As Parsons puts it (2007: 105), within such a framework, “for A to be a cause of B, A must exist independently of B, A must occur before B in time, and all instances of A must be followed by the appearance of B”.2 IBEI Working Papers • 2008/11 2. The Role of Ideas in Contemporary Political Analysis In the following pages we discuss four currently influential perspectives on the relationship between the ideational and the material – rational choice theory, thick and think constructivism and critical realism. Their core ontological and epistemological assumptions and their approach ideas are outlined in the most general terms in table 1. 2. The conventional view of Humean metaphysics (certainly within political science and international relations) is that his account of causation is nomothetic each effect must always have the same cause. Under this ‘successionist’ or ‘billiard-ball’ model, the demonstration of causal laws consists of the recording of empirical regularities. Yet recent philosophical scholarship has served to challenges such a reading. For a review of the literature pointing to a ‘New Hume’, see, inter alia, Read and Richman (2000). -7- 2008/11 • IBEI Working Papers Table 1: Varieties of ideational explanations Ontology (relationship between the ideational and the material) Epistemology (appeal to causal and/or constitutive logics) approach to ideas rational choice theory rigid ontological distinction (dualism) between material Humean - independence and temporal priority of and ideational factors; cause over effect prioritises material explanations material and ideational ontologically indistinguishable; materiality denied causal capacity accords an ontological equivalence to material and ideational factors; ontological dualism of material and ideational factors assigns an ontological equivalence to material and ideational factors; ontological (or analytical) duality of material and ideational factors ideas as auxiliary variables appealed to when and if materialist explanation falters thick (radical) constructivism rejects causal in favour of 'ideas all the constitutive logics way down' thin (Wendtian) constructivism rigidly demarcates causal and constitutive logics whilst appealing to both; retains a Humean conception of causality ideas as constitutive; retains a residual (causal) materialism and tends towards a structural idealism critical realism incorporates constitutive logics within an expanded and non-Humean conception of causation ideas as causal, constitutive and (potentially)3 causally constitutive 3. Rational Choice Theory: Ideas as Auxiliary Variables Whilst the current ‘ideational turn’ is invariably presented in terms of the challenge it poses (or is seen to pose) to a prevailing materialist orthodoxy, it is important to acknowledge that by far the most influential attempt to ‘take ideas seriously’ has come from within that orthodoxy. Indeed, however unlikely it might seem, it has come from within rational choice theory itself. Yet ideational factors are not easily incorporated within the parsimonious theoretical core of rational choice. Indeed, the recognition of the potential significance of ideas may serve in itself to undermine the very elegance and analytical simplicity that has long constituted rational choice’s principal appeal. If, for instance, to render material interests actionable, actors must interpret them, and if they may do so differently, then the context in which actors finds themselves is no longer a guide to their behaviour. Their behaviour is, in short, no longer a logical correlate of the materially-given rational course of action. Perhaps more damagingly still for rational choice’s pretensions to offer a predictive science of politics, actors’ conduct is now contingent upon their interpretation of both their interests and the context in which they find themselves. As this suggest, an ontological commitment to the significance of the ideational is not easily reconciled with an epistemological positivism. 3. As developed in the final section of this paper. -8- Understandably, perhaps, even those rational choice theorists seemingly most enthusiastic to embrace ideas have stopped some way short of tracing the full implications of so doing in this way. Consequently, as a number of commentators have noted, the attempt to accord a role for ideas in rational choice has tended to result in a loose and internally contradictory hybrid in which ideas are grafted on to existing rationalist accounts as auxiliary variable only if, as and where necessary (see also Blyth 1997; Hay 2007). IBEI Working Papers • 2008/11 As this in turn suggests, however much rational choice may have served to direct a welcome attention to ideas in contemporary political analysis, it hardly offers a natural home for the development of a distinctly ideational research programme. Indeed, that so much of the ostensible attempt to take ideas seriously has taken place within the confines of rational choice theory has contributed to a failure to reflect systematically on the ontological and epistemological implications of giving to ideas a causal (or, indeed, a constitutive) role. If we are, then, to develop an ideational research programme in political science and international relations, then, we must look beyond rational choice theory. In particular, we suggest, we must reject two core features of rational choice theory which prevent a greater recognition of complex interdependence of the ideational and the material. These are: (i) its ontological commitment to a rigid and dualistic distinction between material and ideational factors, as reflected in its irreducibly materialist conception of (self) interest; and (ii) the epistemological commitments which follow from its adoption of a Humean conception of causation in which the cause must always be antecedent to, and independent of, the effect. In this respect, a recognition of the causal significance of ideas is a very major, if largely unacknowledged, concession for rational choice theory. For it cuts to the heart of rational choice’s most cherished, basic and fundamental of assumptions – about the context-dependence of rationality, about the irreducibly material nature of context, and about the predictably rational character of all human conduct. Rational choice theory makes a virtue of parsimony and it achieves that parsimony through a series of simplifying assumptions (Hay 2004). Though rarely acknowledged explicitly, one of those assumptions is that ideas do not matter. Put slightly differently, rational choice holds that the distortion injected into its models by the assumption that ideas do not matter is a price worth paying for the analytical simplicity and parsimony such an assumption offers. However innocent it may seem, then, to concede a causal role for ideas even as auxiliary variables is, in a sense, to acknowledge the inherent limitations of the rational choice as a research programme. We examine each in turn before proceeding to a more detailed assessment and critique of rational choice theory’s engagement with the ideational. -9- 2008/11 • IBEI Working Papers 3.1 Rational choice theory’s commitment to ontological dualism Rational choice theory is committed to a rigid and dualistic distinction between the ideational and the material. This ontological dualism is evidenced in: (i) its disavowal of any role for ideas in the formation (and/or constitution) of an actor’s (material) interests; (ii) its commitment to a methodology of counterfactual argumentation; and (iii) its exclusive focus on the individual, as distinct from, collective and/or structural properties of ideas (see also Ruggie, 1998: 20). Goldstein and Keohane’s Ideas and Foreign Policy is widely, and rightly, regarded as the most systematic and thoughtful statement, from a rational choice perspective, of how ideas influence policy. Interestingly, it proceeds from a quite explicit challenge to rationalism itself. As Goldstein and Keohane write, “[a]lthough we concede that the rationalist approach is often a valuable starting point for analysis, we challenge its explanatory power by suggesting the existence of empirical anomalies that can be resolved only when ideas are taken into account” (1993: 6, emphasis added). Yet this challenge is strictly limited. For, as they go on to elaborate, their null hypothesis is “that variation in policy across countries, or over time, is entirely accounted for by changes in factors other than ideas” (6, original emphasis). In other words, though the failings of rationalism are clearly acknowledged, there is no attempt to address these at source. Materialist explanations are privileged and potential ideational variables are only considered at the point at which all (evident) materialist options have been exhausted. There is a clear hierarchy of preference, then, between non-ideational factors and variables (which are to be preferred) over ideational factors and variables (to be considered only as and when necessary). Yet it is not just the materialism of rational choice that Goldstein and Keohane reproduce. Despite their ostensible ‘challenge’ to the ‘explanatory power’ of rationalism, they remain unbowed in their (rationalist) commitment to interests being given exogenously of the actors themselves. Indeed, it is presumably precisely because they remain wedded to both a rump materialism and a rump rationalism that ideas are brought into the analysis only if, as and when a materialist explanation proves inadequate. As Mark Blyth has suggested, ideas are reduced to a deus ex machina conjured to ‘mop up’ residual variance or to deals with the discomforting deficiencies of an existing research programme (Blyth, 1997). Goldstein and Keohane, it seems, freely concede what they clearly regard to be the inherent limitations of rationalism – limitations they associate with the incapacity to accord a genuinely causal role to ideas. Yet, rather than revisiting and revising the core premises of rational choice theory such that they no longer preclude an appropriate role for ideas, they seem to prefer instead a strategy of patching up otherwise pristine rational choice models at the point at which they reveal themselves to be flawed empirically by the inclusion of variables (of an ideational kind) that would not normally be sanctioned. A rather arbitrary and ad hoc approach to political analysis is the result. Interestingly, in response to similar criticism, Keohane has protested that it is to exaggerate and caricature the position set out in the Goldstein-Keohane volume to see it as according only a residual role to ideas within an otherwise untainted -10- The rationalist commitment to a dualistic ontology of the material and the ideational is also evident in the logic of counterfactual argumentation that Goldstein and Keohane deploy. Thus, in her own work, Goldstein (1993) uses a deductive counterfactual method to explore the degree to which ideas, rather than material interests alone, have influenced US trade policies historically. She begins by deducing, from a consideration of structural imperatives, the policy preferences of actors following their exogenously given interests. These she then compares to exhibited policy outcomes, attributing any difference between the two to the degree to which ideas rather than interests were determinant. This is undeniably neat, but it is to commit the analyst to precisely the conception of interests as ‘ideas-free baselines’ that Keohane denies. Moreover, it is also to set up the problem as one of gauging the relative significance of ideational and material factors in determining specific outcomes – a problem, as we have seen, that Keohane dismisses as ‘misguided’. Finally, to deduce actors’ interest simply from the structural imperatives of the context in which they find themselves is also misleading. As Campbell suggests (2000), the crucial question is not what actors’ material interests or structural imperatives are, but how they come to perceive of their interests and imperatives in the way in which they do – and, it might be added, the extent to which such largely instrumental considerations come to inform their behaviour (it being far from clear that all motivations for action are narrowly instrumental). rationalist ontology. Interests, he states, are “certainly not idea-free baselines” (2000: 128). Moreover, as he continues, “the question – do ideas matter more fundamentally than material forces? – is … misguided … what is important is how ideas and material capabilities intermingle” (129). Taken on their own, we have considerable sympathy with both comments – though the former perhaps begs the question of what, precisely, interests are if they are not ‘idea-free baselines’? But the point is that it is very difficult to reconcile either remark with the theoretical statement to which they purportedly refer. Indeed, they would seem to be flatly contradicted by Goldstein and Keohane’s statement of the ‘null hypothesis’ that we have already considered. But this notwithstanding what Keohane’s comments do perhaps serve to indicate is a desire, certainly on his part, to accord a rather more significant explanatory role for ideas than is consistent with an unqualified commitment to rationalism. IBEI Working Papers • 2008/11 As this suggests, however impoverished and underdeveloped existing ideational analysis in political science and international relations may presently be (and this is a theme to which we return in later sections), the kind of methodological bracketing of material and ideational factors that Goldstein engages in here is certainly not a part of the answer. A third problem with rationalism’s commitment to a dualistic ontology of the material and the ideational, is what John Ruggie terms the “neo-utilitarian misspecification of … ideas” (Ruggie, 1998: 20). By this he refers to the characteristic focus (both ontologically and methodologically) on the individual – here, specifically, on the ideas (or cognitions) that actors hold in their heads rather than on the structural and/or collective properties of ideas. This ‘misspecification’ -11- of ideas follows directly from the methodological (and ontological) individualism to which rationalism subscribes. Indeed, in this respect, it is less a misspecification than it is a direct and logical correlate of a prior ontological commitment – if one is committed to a socially atomistic ontology (such as rationalist individualism) then ideas are (and can be) no more and no less than individual cognitions. But the point is that most rational choice theorists who write about ideas do so in such a way as to indicate that they regard ideas to have an irreducibly social (or intersubjective) dimension. Again, it seems, we run into the inherent incommensurability of a rationalist ontology on the one hand and an often intuitive and largely untheorised hunch as to the causal significance of ideas in processes of social and political causation on the other. If ideas are irredeemably social, then they are not easily assimilated within a rationalist individualist ontology; nor are they easily analysed by an approach with subscribes to a methodological individualism. 2008/11 • IBEI Working Papers Rational choice theorists, it seems, must at pain of self-contradiction omit from their analyses the social context of ideas and neglect what many of them may well regard as a fact namely that ideas exist as collective structures of meanings that are connected, but irreducible, to the actors who draw upon them. As Emmanuel Adler puts it (1997: 327), “intersubjective meanings are not simply the aggregation of the beliefs of individuals … Rather, they exist as collective knowledge … This knowledge persists beyond the lives of individual social actors, embedded in social practices as they are reproduced by interpreters who participate in their production and working”. Though such remarks are likely to prove unobjectionable to many of the rational choice theorists most prominent in the ideational turn, they are in fact a statement of a distinct ontological position. And, as we have sought to show, that ontological position is profoundly antithetical to rational choice theory. If we are right, then there is less space within rational choice theory than many recent scholars have assumed, and continue to assume, for ideas. Ultimately, something will have to give – for, we contend, to accord ideas an independent causal role is to violate the core ontological commitments of rational choice theory. If we are to develop an ideational research programme in political science and international relations, it cannot be constructed from within rational choice theory. 3.2. The epistemological underpinnings of rational choice theory’s turn to ideas Given the anchoring of the rationalist ideational programme in a dualistic ontology which counterposes the ideational and the material, and given the directional dependence of ontology and epistemology, it should come as no surprise that there are also epistemological impediments to a rationalist turn to ideas. These are most evident in rationalism’s commitment to a Humean logic of causation. This stipulate not only a mechanistic notion of constant conjunctions according to which each cause must always produce a specific effect but also, and more significantly, that the cause must be independent of, and antecedent to, that effect. This, we -12- suggest is limiting – especially when it comes to a consideration of ideas. Indeed, rationalism’s epistemology is limiting in two respects: (i) as we show presently, its Humean understanding of causation is overly restrictive, effectively precluding a full recognition of the ways in which ideas in particular may prove causally effective; and (ii) its disavowal of the epistemological significance of constitutive logics in the narrow pursuit of simple logics of cause and effect leads it to overlook the constitutive power of ideas. IBEI Working Papers • 2008/11 Here Wendt is again interesting. Despite sharing with rationalism a Humean conception of causation, he argues for a recognition of the importance of both causal and constitutive logics – though he sees these as radically independent and incommensurate forms of reasoning (1998: 77-88). Moreover, he suggests that despite the differences between the causal and the constitutive and the tendency to associate the ideational with the latter, we should not exclude the possibility of ideas as causes (even in the narrowly Humean sense of the term he deploys). This is an important point with which we largely agree. But equally important, we contend, is that ideas are no only seen as causal in a narrowly Humean way. Indeed, this for us is the principal epistemological impediment within rationalism to the development of a more genuinely ideational perspective. The issues at stake here are important and take us to the heart of the epistemological fault-line that has tended to bifurcate the study of ideas in contemporary political science and international relations. And we will return to them when we discuss Wendt’s work. For now we restrict ourselves to a discussion of the limits of the Humean conception of causation as it is reflected in the rationalist turn to ideas, taking as a starting-point Goldstein and Keohane’s (1993) three-fold typology of the ways in which ideas can be invoked in causal logics. In the first of these, under conditions of uncertainty, ideas guide actors towards strategies for attaining their goals and realising their interests. Here ideas affect (or, at least have the capacity to affect) policy “by providing principled or causal road maps” (1993: 8). As Goldstein and Keohane themselves note, even the most ardent and purest of rational choice theorists tend to accept that individuals lack complete information. To the extent that this is the case (or is assumed to be the case), it is important to understand the mechanisms in and through which actors negotiate informational asymmetries. Such mechanisms are likely to be especially significant in periods of uncertainty, in which actors’ behaviour can only be informed by the anticipated consequences of potential courses of actions (ibid: 13). In such scenarios ‘causal ideas’ serve to resolve the incompleteness of information “by stipulating causal patterns” and by suggesting “strategies for the attainment of goals”. ‘Principled ideas’ do not resolve informational incompleteness. But, by providing “compelling ethical or moral motivations for action”, they allow actors “to behave decisively despite causal uncertainty” (1993: 16). The argument can be presented schematically as follows. a) Ideas as ‘Road Maps’ -13- 2008/11 • IBEI Working Papers Figure 1: Ideas as road maps guiding strategic action s uncertainty h ideas a p i n strategic action g policy outcomes Problems • Failure to specify the mechanism(s) of ideational (and hence strategic) selection - in effect, the social context of ideas is omitted; • The influence of ideas restricted to periods of uncertainty – a context-based ontology. This type of conception has proved immensely influential both within and, indeed, beyond rational choice theory. Yet it is far from unproblematic and far from unambiguous, as even Goldstein and Keohane at times seems to concede. Thus, as they point out, their approach offers no analysis of “which ideas are available”, where they come from and how they become influential and/or persuasive (ibid: 12, emphasis added). When it is considered that, even in their own terms, once a set of ideas proves influential and is selected the pathway it opens up closes off other possibilities, this is a major lacuna (see also Yee 1997: 1024). Yet this should come as no great surprise. For wedded, as they remain, to an ontological and methodological individualism, they can offer little if any analytical purchase on what might be termed the ‘social context of ideas’. As Yee notes, ideas are treated as essentially free-floating entities, the “properties that give causal and principled ideas their causal effects” simply being presupposed (ibid: 1024). Yet this is principally an error of omission rather than one of commission. Altogether more problematic is the reliance of this stylised model of the influence of ideas on assumptions about uncertainty. In effect, Goldstein and Keohane treat ideas as random products of uncertainty – wherever uncertainly persists, ideas (both causal and principled) will proliferate. As already noted this generates problems about where such ideas come from, how, in a context of uncertainty, they acquire authority and how they are ultimately selected. Yet there are bigger issues at stake here. Goldstein and Keohane, like many others before and since, vacillate uneasily between a conception of uncertainty as a universal and time-invariant human condition and as a periodic interruption to business as usual. They begin their reflections on ideas as road maps, for instance, by noting that virtually no-one accepts the (nonetheless still prevalent) assumption of complete information. This surely implies that uncertainty is rife. Yet they go on to associate conditions of uncertainty quite explicitly with exogenous shocks – such as depressions, crises, wars and the overthrow of a government (ibid: 17). If uncertainty is the exception rather than the norm, as this implies, and the degree of uncertainty defines the extent to which ideas matter, then a rump materialism may be defended and ideas may indeed be treated as residual or auxiliary variables, to be summoned only in disequilibrium scenarios. In effect, an ideationally-augmented rational choice theory can be reserved for the (presumably rare) special cases in which equilibrium assumptions are violated; a rational choice without ideas will suffice for all other scenarios. But what this surely necessitates is an account of the conditions of existence of certainty and uncertainty. Such an account is notable only by its absence; and in its absence we have a tautology – ideas are important where there is -14- uncertainty; and uncertainty is made visible to us by the failure of materialist accounts (and hence the need to import auxiliary ideational variables). IBEI Working Papers • 2008/11 Put slightly differently, if it is ideas that render interests actionable, then neither is ontologically primitive and there must be a general relationship between them that is constantly at work. If so, then bringing in ideas as a random product of uncertainty is suspect both methodologically and theoretically. It is, in effect, to commit a form of omitted variable bias. For it is to privilege arbitrarily ideas over interests when and if an interest-based explanation is insufficient. And, as this implies, it is also ontologically inconsistent (Gofas, 2001: 9; Hay, 2002: 214-5). Yet this is by no means the only problem with such a formulation. For, if complete information is nothing more than a convenient fiction to make the algebra simpler, as Goldstein and Keohane clearly imply, then ideas must ‘matter’ all the time. Yet if that is conceded – and, for rationalism, it is a very major concession then ideas must be causally prior to material interests. For, in the absence of complete information, actors can never know what their genuine material interests are; and must, as a consequence, always be reliant on a perception of their interests. If we are to understand and explain behaviour we do not need to know what an actors’ material interests are, merely what they are perceived to be. The inexorable link between context and conduct opened up by the assumption that behaviour is a logical correlate of contextualised rationality is shattered – and, arguably, with it rational choice’s pretension to offer a predictive science of politics. Acknowledging the ontological significance of ideas has major epistemological consequences. While limits of space do not permit a full treatment, it is important to note both the significant role assigned to uncertainty in circumscribing the degree to which ‘ideas matter’ in current debates, and the minimal attention the concept has thus far received (for a partial exception see Blyth 2002a: 31-3, 267-70 and, for a critique, Hay 2004). This neglect is unfortunate. Moreover, as we shall see presently, the concept of uncertainty is widely invoked even by trenchant critics of rational choice theory as a necessary precondition of ideational influence. Thus, Mark Blyth, a leading figure in the ideational turn and advocate of a constructivist institutionalism argues that “while ideas do not ‘matter’ all the time, in certain specific circumstances – particularly in moments of economic crisis, ideas, and the political control of those ideas matter most of all” (2002a: 16; emphasis added). This temporal unevenness, in which the ontological status of ideas (and hence the nature of the relationship between the ideational and the material) varies, we categorically reject (see also Gofas 2003). We see it as an unfortunate residue of the rationalist terms of reference in which much of the ideational turn to date has been couched. To sum up, for those wishing to convince a sceptical mainstream of the place of ideational analysis, it may well be convenient to appeal to a temporally inconsistent ontology that ascribes significance to ideas only under (periodic and infrequent) conditions of uncertainty. But such a move can only serve to reintroduce a dualistic ontology of the material and the ideational through the back door. It also serves (perhaps inadvertently) to commit those who deploy it to the assumption that, other than in a few exceptional moments, the complete -15- information assumption holds. Such a simplistic and convenient fiction may well be justified in rational choice theory in terms of the greater capacity for formal modelling that it offers. But no such utilitarian defence exists for those not wedded to the deductive inference from stylised assumptions. A more (ontologically) consistent and fruitful conceptualisation, we suggest, construes of uncertainty as both a universal human condition and as a discursive regime that imposes distinct dispositional logics upon actors. Understood in such terms, the appeal to uncertainty is as much as anything else a political and discursive device in the process of enlisting subjects to alternative visions of change. In this conception, the future is always uncertain because it is contingent upon social and political behaviour – behaviour irredeemably informed by ideas. This view we elaborate in the final section of this paper. A second, and in fact radically different, way of conceptualising the relationship between the material and the ideational appeals to ideas as ‘focal points’ in Goldstein and Keohane’s terms. In effect, ideas serve to ‘define cooperative solutions’ to problems by selecting between multiple Pareto-optimal potential outcomes. Here ideas are accorded a far more residual and second-order role and are clearly posterior to material interests. As Goldstein and Keohane put it, in such a causal sequence “ideas affect strategic interactions, helping or hindering joint efforts to attain more efficient outcomes – outcomes that are at least as good as the status quo for all participants. Here ideas contribute to outcomes in the absence of a unique equilibrium” (1993: 12, emphasis added). b) Ideas as Focal Points 2008/11 • IBEI Working Papers Figure 2: Ideas as focal points interests multiple (Pareto-optimal) equilibria Focal points ideas policy outcome Problems • The environment within which ideas operate is one of exogenously given interests. • No mechanism explains how ideas serve to select amongst potential Pareto-optimal equilibria - consequently, the model is indeterminate. The problems with this approach are principally two-fold. Again, it seems, ideas are invoked merely as an auxiliary variable conjured, rather arbitrarily, to resolve a given theoretical puzzle – here a problem of indeterminacy in a more conventional formal model. There is a certain unevenness about this. For once one acknowledges a role for ideas in selecting amongst multiple equilibria, might one not also acknowledge a role for ideas rather earlier in the causal sequence? Indeed, is it credible to retain a purely material conception of self-interest whilst assigning entirely to ideational processes the resolution of any indeterminacy in the resulting model that follows from such a materialist understanding of self-interest? Is it legitimate (and/or ontologically consistent), in other words, to circumscribe the role of ideas seemingly purely for the convenience of the model in this way? Moreover, as Blyth -16- Second, although the appeal to ideas is here intended to resolve the indeterminacy of a formal model, it is not at all clear that this is achieved. For, as Goldstein and Keohane themselves concede, “which idea is chosen as the focal point cannot be fully explained by rational choice theory” (1993: 19) – indeed, it is not at all clear that it can be partly explained by rational choice theory either. For, again, no mechanism is identified that might explain how and why ideas select amongst potential multiple cooperative equilibria (Bernstein, 2000: 481; Blyth, 2002a: 26; Yee, 1997: 1025-7). In the absence of such a mechanism, the outcome appears random. At best, then, all Goldstein and Keohane have done is to tell us what type of variable might explain the outcome; this is a long way short of providing an explanation for a specific outcome. We are forced to agree with Mark Blyth “the problem of multiple equilibria persists despite the invocation of ideas to solve it” (Blyth, 2002a: 26). The final way in which Goldstein and Keohane incorporate ideas within a causal sequence is through the mediating role of institutions (seen as instantiations and embodiments of ideas) on policy outcomes. As they put it, in this conception, “the impact of some set of ideas may be mediated by the operation of institutions in which the ideas are embedded” (1993: 20). Such a causal sequence is outlined schematically in figure 3. c) Ideas Encased Within Institutions notes, if one’s analytical starting point presupposes the fixity and materiality of interests, then it is in fact quite unnecessary to accord any role to ideas in the translation of multiple equilibria into a unique outcome. For, as he puts it, “more traditional instruments such as side payments can perform the same function more efficiently” (2002a: 26). IBEI Working Papers • 2008/11 Figure 3: Ideas encased within institutions Mediating 'filters' ideas institutions policy outcome Problems • Despite impressions to the contrary, ideas are reduced to an auxiliary variable whose content may be wholly determined by material factors. • Analysis is confined to identifying the institutional preconditions of ideational receptivity. Eventually, it would seem, we have a model that assigns causal priority to ideas – after all, they appear first in the causal sequence described above. Yet this is potentially misleading. For, as any closer analysis cannot fail to reveal, once again ideas are reduced to second-order and auxiliary variables. There are in fact two reasons for this: (i) the content of the ideas which appear in Figure 3 as causally prior is itself invariably seen to be determined by material factors (such as interests); and (ii) even in the absence of this, it is not at all clear whether it is the ideas instantiated in the relevant institutions or the institutions which instantiate the ideas that carry the causal significance. -17- Of course, though undoubtedly influential and expressive of something of a developing rational choice institutionalist orthodoxy, Goldstein and Keohane’s variant of institutionalism is not the only one on offer. Limits of space prevent a detailed consideration of a now breathtaking range of alternatives (for a review of which see Schmidt 2007). Suffice it to note for now that many of these are restricted in the role they assign to ideas by the notion that ideas must ‘fit’ with pre-existing institutional arrangements if they are to be selected. This, we suggest, makes it far easier to deal with path-dependent institutional change than it does path-shaping institutional change – which, as a consequence, tends to be exogenised. In its focus on ideas, then, much of the new institutionalism restricts itself to a consideration of the institutional preconditions of ideational receptivity. There is certainly much value in this existing institutional arrangements do exert path dependent effects and they do so, at least to some extent, by setting the ideational parameters of procedural and policy choice (Fligstein and Mara-Drita, 1996: 4-5). But they also have a key role to play in the process of institutional formation and in post-formative yet path-shaping institutional change. Indeed, it is in such moment that their transformative (as distinct from their reproductive) power is truly revealed (Blyth, 2002a: 20-23). This far too much institutionalist scholarship denies. This shows that Goldstein and Keohane assign no independent role to ideas in the process of institutional formation (in path-shaping institutional change). Yet their perspective might allow ideas a role in post-formative or path-dependent institutional change.4 Yet further reflection reveals that this, too, is denied. For, as they elaborate, it is the “institution that embodies the ideas”, rather than ideas themselves, that do the explaining (Goldstein and Keohane, 1993: 21, n. 34). Thus, as Yee has suggested (1996: 89), it is surely both “more accurate and less cumbersome to say that these institutions affect policies, rather than to argue that ideas affect policies by being encased in these institutions”. In effect, and despite superficial impressions to the contrary, this is not a form of ideational explanation at all. Thus, although Goldstein and Keohane (1993: 13) argue that it is “ideas embedded in institutions [that] specify policy in the absence of innovation”, they go on to explain that “this institutional path says nothing about why [or, indeed, which] ideas were originally adopted; in fact, ideas may become important solely because of the interests and power of their progenitor”. What is given with one hand, is taken away with the other. For if ideas are nothing more than an expression of the (given) interests of actors and if any clash of ideas is understood to be resolved simply in favour of the more (materially) powerful and resource-rich, then ideas are entirely empty variables with no independent causal significance. 2008/11 • IBEI Working Papers 4. On the distinction between path-dependent and path-shaping logics and see Hay, 2006b. -18- IBEI Working Papers • 2008/11 4. From Dualism to Duality: Constructivisms, Thick and Thin Having discussed in some detail the problems with rationalist-inspired readings of ideas, we turn now to those ‘dialectical’ conceptions of the materialideational relationship which posit some form of (ontological or analytical) duality. In so doing, we again strive to resist the tendency to treat all such formulations as if they constituted a single alternative to the mainstream. The move from a position of ontological dualism to one of ontological duality is invariably presented, by those who wish to take it, as the move from a simplifying distortion of ‘reality’ to one of subtlety and sophistication. This kind of ontological proselytising we resist. It is nonetheless the case that many of those ostensibly committed to a dualistic and materialist treatment of the material-ideational relationship do feel the need to concede a role for ideas as causal variables (auxiliary or otherwise). In so doing, as we have shown, they invariably accord an ontological status to ideas that is incompatible with their commitment to materialism (rump or otherwise). Where this is indeed the case, we suggest, it is reasonable to suppose some disparity between ontological assumptions chosen (quite legitimately) for practical or analytical reasons and those truly reflective of the analyst’s own worldview. The latter are, we contend, far more likely to acknowledge a causal and/or constitutive role for ideas than the former. If this is correct, then however dominant it may appear, materialism is in fact far less prevalent in mainstream political science and international relations than is often assumed. Moreover, its seeming prevalence owes far more to the parsimonious analytical modelling that it facilitates than it does to its innate ontological credibility. By the same token, however, there is an analytical price to be paid for adopting a more credible and complex set of ontological assumptions. Either we choose our ontological assumptions in such a way as to close off the need to deal analytically with the complex consequences of ‘taking ideas seriously’ or we face such analytical difficulties head on. There are advantages to each strategy – though, as we have seen, much of the existing attempt to take ideas seriously seems to vacillate uneasily between the two. In what follows we consider the work of those who have tended to refuse to choose their ontological assumptions for analytical convenience. The point is that the move to a dialectical understanding of the materialideational relationship almost inevitably entails a loss of parsimony and a seeping away of often cherished former epistemological certainties. As John Ruggie puts it, “the strength of neo-utilitarianism lies in its axiomatic structure, which permits a degree of analytical rigour, and in neoliberal institutionalism’s case also of theoretical specification, that other approaches cannot match” (1998: 36). But the strength of that agenda is simultaneously its weakness – namely, its commitment to an ‘axiomatic structure’ and an largely asocial ontology that seemingly lacks credibility even amongst those ostensibly most wedded to it (ibid.: 36). -19- Most such positions are, at least within political science and international relations, now labelled constructivist. This, in itself, is confusing – since many of these positions bear little or no relationship to the long-established tradition of constructivism in the philosophy of science (Berger and Luckmann 1966; Hacking 1999; Kukla 2000; and for a review, Hay 2002: 197-204). Indeed, the range of positions labelled constructivist in contemporary international relations theory alone is far greater and far more internally diverse than the equivalent range in the philosophy of science literature. What such positions nonetheless tend to share is a rejection of the stark differentiation between ideas-based and interest-based approaches. For most constructivists, ideas provide the framework through which interests are defined. As Finnemore puts it, “structures of shared knowledge and intersubjective understandings may shape and motivate actors. Socially constructed rules, principles, norms of behaviour and shared beliefs may provide states, individuals and other actors with understandings of what is important or valuable and what are effective and/or legitimate means of obtaining those valued goods” (1996: 15). Thus, constructivists have long argued that we should eschew thinking in terms of an oppositional dualism between ideas and interest-based approaches; the question is how the two are related, rather than which of the two is the more important. 2008/11 • IBEI Working Papers Yet if this might be seen to indicate a common core to constructivism, it cannot and should not serve to hide the vast chasm that has opened up between what are now invariably referred to as ‘thick’ and ‘thin’ variants of constructivism. Yet whilst is it relatively easy to differentiate between these twin constructivisms, in part precisely because the distinction has served to provide labels of self-identification, there is a danger that all internal divisions within a very broad church are presented in dualistic terms. In particular, we suggest, there is a danger that the constructivism of critical realism is overlooked. It is certainly unconventional to refer to critical realism as a form of constructivism and it is certainly true that, as a philosophy of (social) science, it long predates the ideational turn in political science and international relations. Yet, in the relatively loose and inclusive sense of the term used in international relations theory, critical realism is undoubtedly constructivist in its ontological commitments. Moreover, it is not just that critical realism can be labelled constructivist. It is, we suggest, useful to consider it as a variant of constructivism in reviewing the range of approaches to ideas in political analysis. In the proceeding discussion we follow this conventional distinction, differentiating between thin and thick constructivisms. A second distinguishing feature of constructivism, despite its inner diversity, is its focus on the social and structural rather than the individual properties of ideas; that is, its focus on intersubjectivity rather than subjectivity. As Adler explains, “intersubjective meanings are not simply the aggregation of the beliefs of individuals … Rather, they exist as collective knowledge … This knowledge persists beyond the lives of individual social actors, embedded in social routines and practices as they are reproduced by interpreters who participate in their production and working” (1997: 327). -20- IBEI Working Papers • 2008/11 4.1. ‘Thick’ constructivism: the challenge of postmodernism Once upon a time, a valiant fellow had the idea that men were drowned in water only because they were possessed with the idea of gravity. If they were to knock this position out of their heads, say by stating it to be a superstition, a religious concept, they would be sublimely proof against any danger from water. His whole life long he fought against the illusion of gravity, of whose harmful results all statistics brought him new and manifold evidence. This honest fellow was the type of the new revolutionary philosophers in Germany. (Marx and Engels, from the ‘Preface’ to The German Ideology) The above passage, from The German Ideology, has become a key reference in ontological and epistemological disputes invariably summoned to indicate the author’s contempt for, and exasperation with, the self-evident idiocy of an alternative ontological and/or epistemological position. Yet we cite it here for rather different reasons. For we eschew the kind of ontological and epistemological evangelism with which this passage is now almost inevitably associated and into which debate about postmodernism and thick constructivism seems invariably destined to descend. Indeed, our aim is precisely the converse – namely, to show that passages such as these, which increasingly litter the critical reception of thick constructivism, are little more than exercises in ontological evangelism. Yet our disavowal of ontological proselytising does not make us enthusiasts nor, indeed, ‘apologists’ for thick constructivism either. Our concern is not with whether the thick constructivists, like their young Hegelians predecessors, choose, or refuse, to acknowledge natural forces like gravity or, indeed, as Wendt would have it, the existence of cats and dogs (Wendt 1999: 49). Nor is our concern with postmodernists that they “have spread the disease of unwarranted doubt that has decayed the foundations for a community of scholars … destroyed the bridges between alternate paths to knowledge, eroded the bases for interdisciplinary compromises and respect for the pursuit of understanding, and undermined the confidence in the value of research programs” (Kegley, 2002: 66). Our problem is rather different and it has little or nothing to do with the ontological status of cats and dogs, nor the damage that a contagious epistemological scepticism might do to cherished claims about progress and the scientific status of political analysis. Nor, finally, does it entail caricaturing postmodernism as a form of naïve idealism that denies all materiality. Yet, for us, postmodernism remains a form of idealism nonetheless. By denying causal efficacy to material factors, it tends towards a simple view of the relationship between the material and the ideational. And this, in turn, leads it to jettison the baby with the bath water ultimately disavowing all appeals to causal logics (whether ideational or material). It is this that concerns us – not because it is innately wrong when judged against some universal ontological and epistemological standard, but because it is decidedly and unnecessarily limiting and because it imposes, analytically, a far more virulent form of idealism than is in fact entailed by its ontological assumptions. -21- To see why it is useful to begin by considering what is, and what is not, at stake in the existing debate on the ontological status of cats and dogs (see Wendt, 1999: 49; Doty, 2000: 138; Kratochwil, 2000: 95). The debate proceeds by way of an engagement with Wendtian constructivism (a position we consider in more detail in the following section); and it is important, as a consequence, to understand where Wendt is coming from in such exchanges. As we shall see presently his over-riding aim is to build bridges between a prevailing positivist orthodoxy which, partly as a consequence, he sees as defining the legitimate terrain of scientific enterprise and a decidedly thin constructivism which he grounds (or claims to ground) in a variant of critical or scientific realism. This bridge-building exercise brings him into inevitable conflict with postmodernists – for his intention, as much as anything else, is to show that to embrace (thin) constructivism is not to take the first step down the slippery slope to postmodernism. This he does by seeking to demonstrate the natural affinities, as he sees them, between positivism and constructivism and by seeking to show the natural incommensurability of both with the nihilistic epistemological relativism he ascribes to postmodernism. It is this playing to the galleries, we suggest. that leads him to the rather overstated assertion that, for postmodernists, “we cannot even know if seemingly observable entities, like cats and dogs, exist out there in the world” (1999: 49). 2008/11 • IBEI Working Papers Doty’s response to his accusation is more careful and temperate. “Though I know of no postmodernist who has actually written about the existence or nonexistence of cats”, she confesses, “I think it is safe to suggest that for a postmodernist it is not the physical existence of furry little animals with whiskers that would be in question, but what these furry little creatures are, e.g. sacred and revered deities as in ancient Egypt or over-breeding pests that befoul the streets and alleys and should be euthanized” (2000: 138). In similar vein, Kratochwil reminds us that, “these descriptions are not neutral and somehow objective but embrace all types of social practices and interests that then make the things into what they are called or referred to. Thus, while ‘dog’ might be the name for an animal, it is not a description which is appropriate under all circumstances. Part of what defines a ‘dog’ is not only its zoological characteristics, but a socially significant property such as ‘tameness’, which brings a dog under the description of ‘pet’” (2000: 95). First, Wendt is clearly expressing and, indeed, expressing clearly the concerns of the mainstream. From such a perspective the problem with postmodernism is, as Osterud puts it, its “retreat from [the] basic norms of science and professional scholarship”. As he continues, “if these norms are disregarded, professional scholarship and research are out. If the text and the discourse is all, then the Earth is flat if you say so” (1996: 389). So, what are we to make of this and other exchanges like it? But, as Doty’s and Kratochwil’s responses both indicate, the postmodernism to which they subscribe is not the naïve idealism Wendt assumes it to be. But it is a form of idealism nonetheless. Thus, as suggested elsewhere, whilst “it would be a travesty to label … [the postmodernist position] simple, [its] stance on the relationship between the material and the ideational is far from complex, even if [its] proponents’ motivations for holding it are not” (Hay, 2002: 205). Or, put slightly differently, by -22- restricting its analysis to a description and elucidation of the internal workings of discursive formations and practices, postmodernism reduces, perhaps inadvertently, a two-way relationship between the material and the ideational to a one-way street (see also Joseph, 2007: 353). IBEI Working Papers • 2008/11 Thus, Wendt is surely right to suggest that in its almost exclusive emphasis on social meaning, postmodernism tends to fail to consider (even where it is prepared to acknowledge) “the resistance of the world to certain representations” (Wendt, 1999: 58). Yet, by the same token, the repertoire and range of representations on which we draw in narrating our experiences is also limited by pre-existing discursive formations. And if Doty and Kratochvil have little to say about the former, Wendt is just as silent on the latter. Both processes of mediation and conditioning might be seen as forms of discursive selectivity (Hay 2002: 212). In the former, material factors select for, but by no means entirely determine, the meanings we assign to our experiences. In the latter, pre-existing discursive formations select for, but again by no means determine, the ways in which we attribute meaning to social events, processes and behaviours. Thus, as Joseph suggests, “not just anything can be articulated as a nuclear weapon; it has to have certain material properties….[because] the physical properties of something, far from being meaningless outside articulation, are the very things that make social construction possible” (2007: 353, emphasis added). Yet similarly, whether a nuclear arsenal is constructed as a means of security or as a potential terrorist threat will depend on (amongst other things) the constraints that prevailing national and international discourses impose. -23- Thus most postmodernists would surely concede that, by virtue of its (for want of a better term) species being, we are unlikely to assign the status ‘pet’ to an untamed lion. But this is more of a concession than it might seem. For if one rejects a purely idealist ontology, then prior to the question of discursive articulation there is the question of “what is it about the physical or material properties of something that allows it to lend itself to a particular form of social construction?” (Joseph, 2007: 353). The propensity of humans to drown in large volumes of water most certainly lends credibility to the ‘construction’ of gravity as, indeed, it might countless other constructions. And, of course, it matters whether we attribute drowning to gravity, bad luck, the incapacity of humans to breathe underwater, a vengeful God, or some combination of these constructions. But so, too, does the drowning such constructions narrate. It is not difficult to see how. Thus, if we return to their exchange with Wendt, Doty and Kratochwil’s terms of engagement are set in such a way that they focus exclusively on the social properties assigned to ‘cats’ and ‘dogs’ whilst giving absolutely no consideration whatsoever to the extent to which the latter’s materiality might set limits on the social properties assigned to them. To be clear, it is not that postmodernists are incapable of acknowledging that such limits might exist (this is where Wendt gets it wrong), but that their analytical commitments seem to accord them no significance and largely prevent them from becoming the subject of analysis. The result is an immaterial form of analysis – and analytical idealism, if you like but one that need not (and, as in the case of both Doty and Kratochvil, does not) rest on a purely idealist ontology. 2008/11 • IBEI Working Papers As this suggests, if we are to respect analytically, let alone operationalise methodologically, an ontology that neither privileges the material nor the ideational, we cannot but reject the analytical idealism of postmodernism and its disavowal of causal logics. But if this is so, might a thinner form of constructivism provide that still elusive via media? 4.2. ‘Thin’ constructivism: all things to all people? Alexander Wendt’s Social Theory of International Politics (1999) has, rightly or wrongly, become the central reference point for all debate in international relations theory, at least, on the merits and demerits of constructivism. And whatever else one might say about it, it would be churlish not to recognise that Wendt’s intervention is a major one (indeed, if there is any kind of a problem here it is not with Wendt himself, but with the industry his work has spawned). What is perhaps most significant of all about Wendt’s intervention, is that it has served to (re)-position social theory as an indispensable element of international relations theorising and, in so doing, to break a long-standing tradition of disciplinary insularity that has tended to insulate the field from wider developments in the social sciences. But it is with his defence of a constructivism of a particular kind that we are here principally concerned. Yet his ability to deliver a genuinely dialectical approach to the relationship between the ideational and material is, we suggest, ultimately compromised by the two central features of the ‘via media’ that he offers: namely, (i) his proposed reconciliation and synthesis of scientific realism and constructivism; and (ii) the rigid and dualistic treatment of the distinction between the causal and constitutive effects of ideas. Ontologically, we argue, he provides a failed marriage of inconsistent Wendt’s ‘rump materialism’ thus aims to strike a balance (or, as he prefers, to provide “a philosophically principled middle way” [2]) between a vulgar materialism, according to which ideas are epiphenomenal and have no impact on politics, and a radical constructivism that denies any causal efficacy to materiality. As he puts it, “material forces are not constituted solely by social meanings, and social meanings are not immune to material effects” (111-2). As aphorisms go, it is difficult to imagine a clearer statement of a dialectical understanding of the relationship between the material and the ideational. Stripped to its ontological core, Wendt’s position is and is explicitly presented as a form of ‘rump materialism’, in which the distribution of capabilities (principally, geographical and resource constraints) and the composition of capabilities (which he sees as largely technologically determined) are understood as material phenomena (1999: 109-113). This he does to remind us, as he puts it, that “constructivism should not proceed as if nature did not matter” (111). Thus, although ideas and culture give impetus and direction to the on-going human effort to transcend material constraints, “whether we like it or not, the distribution and composition of material capabilities at any given moment help define the possibilities of our action” (113). -24- premises rather than a “philosophically principled middle way”. Moreover, despite the appeal of his attempt to develop an ‘epistemological Westphalia’, he introduces an artificial and unnecessarily polarising dichotomy between the causal and constitutive effects of ideas which, instead of transcending current disciplinary divisions between causal and non-causal forms of theorising, actually serves to reproduce them. We consider each in turn. As already noted, if we are fully to understand the at times somewhat quirky character of Wendt’s hybrid constructivism, it is important to acknowledge from the start the task he sets himself. As much as anything else that task is not to upset mainstream international relations theorists by convincing them, as best he can, that they have little or nothing to fear from embracing the kind of thin constructivism he offers. Thus, from the outset, he declares his commitment to a science of politics and international relations, emphasising in terms comfortingly familiar to his mainstream audience, the scientific realist character of his constructivism.5 And judged simply in terms of the welcome afforded him in mainstream IR, his strategy has undoubtedly succeeded. Stephen Krasner, for instance, suggests that Wendt’s “discussion of scientific realism ought to be required reading for any student of international relations, or political science for that matter” (2000: 131). Similarly, and perhaps more significantly still, Robert Keohane declares that Wendt “has shown convincingly that one does not have to swallow the contaminated epistemological water of postmodernism in order to enjoy the heady ontological wine of constructivism” (2000: 129). But is this right? Can ideas be incorporated so seamlessly, and without any epistemological and/or ontological fall-out, within an otherwise untainted international relations scientific orthodoxy? This we contend. To see why it is necessary to explore Wendt’s argument in a little more detail. a) The Marriage of Critical Realism and Constructivism IBEI Working Papers • 2008/11 The first problem we detect in Wendt’s hybrid constructivism is, quite simply, the incommensurability of the ontological and epistemological commitments from which it ostensibly builds. To claim to reconcile critical (or scientific) realism and positivism might seem like a tall order; but to claim to reconcile positivism, postpositivism and critical realism simultaneously stretches credulity. But Wendt could not be clearer on the subject. Thus, in summarising his position he states, “epistemologically I have sided with positivists … On ontology –which is to my mind the more important issue I will side …with post-positivists”, before going on to suggest that “scientific realism plays an essential role in finding [the] … via media between positivist epistemology and post-positivist ontology” (Wendt, 1999: 90, 91). The problems with such a formulation are principally three-fold. First, Wendt here seems to imply that ontological choices are epistemologically neutral and vice In such a context, one can only surmise that the term ‘scientific realism’ to which Wendt refers is likely to give off more of the right kind of signals than the more familiar ‘critical realism’. 5. -25- versa. This is patently not the case as, indeed, scientific realists like Bhaskar (1989: 67 ff.) have spelled out very clearly (uncharacteristically clearly, as it happens). If ontology concerns the nature of social and political reality itself, then clearly it sets limits on what one can hope to know about that reality. Thus, as elsewhere argued, “if we are happy to conceive of ourselves as disinterested and dispassionate observers of an external (political) reality existing independently of our conceptions of it [an ontological position] then we are likely to be rather more confident epistemologically that if we are prepared to concede that: (i) we are, at best, partisan participant observers; (ii) that there is no neutral vantage-point from which the political can be viewed objectively; and (iii) the ideas we fashion of the political context we inhabit influence our behaviour and hence the unfolding dynamic of that political context” (Hay 2002: 63-4). Ontology, as this suggests, exerts a ‘direction dependence’ on epistemology (see also Hay 2007). Thus, to adopt a positivist epistemology is to commit oneself to a view of social and political reality (a social ontology) that is consistent with, at minimum, the law of induction and an unqualified naturalism (such as might permit a unity of method and epistemological standards between the social and natural sciences). No post-positivist ontology of which we are aware is consistent with either. Indeed, Bhaskar himself demonstrates at some length how a critical (or scientific) realist reflection on the ‘limits of naturalism’ must logically entail a rejection of positivism (see, for instance, 1979; 1989, 1994; see also Collier 1994: 237 ff.; Hay 2002: 85-6). A logical corollary of this is Colin Wight’s suggestion that “one cannot be both a scientific realist and a positivist” (2002: 36; see also Smith 2000: 153). But if that is the case then it is difficult to see how the former might provide any via media between positivist epistemology and postpositivist ontology. Second, as already hinted, it is somewhat confusing to refer to positivism and post-positivism as ontological commitments – such that one might refer to one’s ontology as positivist or post-positivist. Positivism and post-positivism are epistemologies compatible, like other epistemologies, with a range (albeit a limited range) of social ontologies. And although it is fair to suggest that positivism perhaps imposes greater limits on one’s ontological latitude than any other epistemological stance, this does not make it an ontological stance in itself. Indeed, the ontology with which it is invariably associated is naturalism. By the same token, post-positivism is consistent with a potentially breathtaking range of ontological stances – from Bhaskar’s qualified naturalism to a radical perspectivism. It makes little or no sense, then, to speak of a post-positivist ontology. 2008/11 • IBEI Working Papers Third, and relatedly, one does not need a via media, in Wendt’s terms, to connect up ones ontological and epistemological premises – for they do not come (or, at least, they should not come) separately packaged and in need of bolting together. Yes, a given ontology is likely to be consistent with a range of epistemological commitments the relationship is one of directional dependence, but that dependence is by no means fully determinant. But the point is that one’s ontological and epistemological choices are interdependent. Thus, when it comes to a consideration of the role of ideas in social and political dynamics, the analysts faces a series of ontological choices about the extent to which, and the ways in which, ideas matter. These, in turn, have epistemological implications which one cannot simply choose to -26- ignore in preference, say, for a continued commitment to a positivist epistemology. No via media, scientifically realist or not, is capable of delivering simultaneously the epistemological self-confidence of positivism (which is sustained, as much as by anything else, by a disavowal of any causal role for ideas) and the (ontological) view that “social life is ‘ideas all the way down’ (or almost, anyway)” (90). This is to insist on both having one’s cake and eating it. IBEI Working Papers • 2008/11 In each of these respects Wendt’s hybrid constructivism is internally inconsistent. Yet it is important to be clear about the precise implications of the preceding discussion. Though we fail to see how a constructivist ontology that accords an independent role to ideas in the constitution of social and political reality can be reconciled with naturalism (and hence with a positivist epistemology), we see no such logical incommensurability between constructivism and critical realism. Indeed, and as already suggested, we see critical realism as a potentially potent vehicle for further exploring the constituted nature of social and political reality and the causally constitutive role of ideas in political processes. Finally, it is important to see how this hybrid of contradictory premises leads Wendt’s constructivism into a starkly dualistic (indeed, unapologetically Cartesian) distinction between the ideational and the material. Here it is necessary to consider a second reason for Wendt’s turn to scientific realism – namely, his need to justify philosophically the study of unobservables, such as the state, the state system and, of course, ideas. In anticipation of the familiar empiricist objection, he invokes the critical realist claim that we can establish the reality of unobservables by means of their observable effects. This, we suggest, is a valid and helpful move, but the way in which Wendt develops it draws him into an unnecessarily dualistic distinction between the material and the ideational realms. According to Wendt, “if we are to be scientific realists about social life” we have to recognise that material forces and ideas are “constituted as different kinds of independently existing stuff”. And, as he goes on to acknowledge, “this formulation of the materialism-idealism problem is ultimately Cartesian insofar as it separates the world into two kinds of phenomena – in effect, mind and body” (1999: 112). Yet this, we contend, is based on a further misrepresentation of critical realism. For even if ideas (or culture) are accorded an ‘objectified’ status, such that might be see to possess, in realist terms, structural emergent properties, this need not entail a rigid ontological dualism. Wendt, it seems, assumes that to accord ideas and materiality sui generis properties it is to adopt a rigid Cartesian dualism. Yet, once we takes account of the realist notions of emergence and conditioning, this need not be so – the dualism is dissolved. Thus, according to the principle of emergence, structures (whether material or ideational), although irreducible to agency, are activitydependent. Consequently, they are not analogous to biological systems, possessing no self-reproducing powers of their own. Similarly, according to the principle of conditioning, such power as they might be seen to possess is always mediated in and through human agency. As this suggests, critical realism, contra Wendt, maintains a degree of independence among the realms of the material and ideational without introducing a rigid distinction between the two. -27- 2008/11 • IBEI Working Papers Wendt’s constructivism, as we have seen, entails an ultimately dualistic ontological distinction between the material and the ideational. Yet this is by no means the only dualism on which it is built. Indeed, certainly no less significant is the epistemological dualism he introduces between logics of causation and logics of constitution. This distinction is central to his theoretical enterprise and the analytical utility he sees it as providing is offered as an index of the value-added of thin constructivism. Causal theories, he suggests, ask the ‘why?’ questions, while constitutive theories pose the ‘how?’ questions. According to such a logic, to say that ideas constitute social situations by providing meaning to material forces and factors is not to make a causal claim (Wendt, 1999: 78). b) On Causal and Constitutive Logics This is undoubtedly neat. Moreover, in serving effectively to bracket off – as non-causal – much (if perhaps not all) of the role he assigns to ideas, it is decidedly non-threatening to a mainstream thus far reluctant to embrace enthusiastically constitutive logics. And, if for no other reason than this, is has proved extremely influential. Yet the underlying rationale, we contend, is misleading. A number of points might here usefully be made. First, and at the risk of stating the obvious, most causal theories in the natural sciences when engaging in what they regard to be explanatory exercises invariably provide answers to ‘how’ rather than to ‘why’ questions. In fact, that is not quite strictly true. For what they in fact do is to provide ‘how’ answers to ‘why’ questions. Thus, when asked to explain why a particle dropped from a tower falls to the ground, most scientists (and, indeed, nonscientists) will undoubtedly invoke the law of gravity. But in so doing they offer, as an explanation, a re-description of the exhibited behaviour of the particle in the context of a more general theory. In short, they answer not a ‘why’ question but the question ‘how can we best describe the behaviour of the particle?’ or ‘what is the appropriate theory to describe the exhibited behaviour – under which governing law does it fall?’ At best this merely defers the ‘why’ question, though, unless one has small children, it will usually suffice. As this suggests, much (if perhaps not all) explanation in the natural sciences involves the redescription of an observation or occurrence in terms of a more general governing law; as such, most ostensibly explanatory natural science is non-causal. This, of course, does not invalidate Wendt’s distinction between the causal and the constitutive in the social sciences, but it does serve to contextualise it rather differently. More significant is that Wendt’s rigid distinction rests, and is indeed logically entailed by, a prior (if undeclared) commitment to a narrow, Humean conception of causality. If this is a conscious choice on his part (and given the powerful hold that such an understanding of causality now has in international relations theory, it may well not be), it is puzzling move for someone so exercised by the philosophy of science. For it sits awkwardly, at best, with his declared commitment to scientific realism. Thus, whether he chooses to ignore it or not, it is difficult to imagine that Wendt is unaware that critical realism adopts a much more open view of causation. As Milja Kurki’s excellent and exhaustive survey of the topic -28- As this example suggests, causal and constitutive logics are only at loggerheads as long as causality is understood in Humean terms of constant conjunctions. Indeed, within a consistently critically realist view of the social world in which we cannot presuppose the existence of constant conjunctions, Kurki is surely right to argue that “we would be better placed to deal with the social world … if we saw constitutive theorising as an indispensable part of causal theorising” (2006: 212; see also Hay 2004; Hay and Rosamond 2002). shows, for critical realists causation is understood “rather loosely as all those things that bring about, produce, direct or contribute to states of affairs or changes in the world” (Kurki, 2006: 202). Such a conception is, in fact, far closer to the lay understanding of the term and is entirely consistent with the causal significance of constitutive logics. Thus, for instance, the narration of a set of (perceived) policy failures as constitutive of a crisis may serve to shape decisively the pace, and indeed the content, of subsequent policy change (Blyth 2002; Hay 1996). In this respect, the constitution of crisis is of considerable causal significance. IBEI Working Papers • 2008/11 4.3. Critical realism and cultural conditioning: the morphogenetic approach The attempt to construct a critical realist via media capable of negotiating a path between positivist epistemological and post-positivist ontological commitments was almost bound to fail simply by virtue of the impossibility of the task. And the above analysis confirms, in our view, that failure. Yet this does not, of course, exhaust the potential of critical realism to develop a genuinely dialectical understanding of the material-ideational relationship and to illuminate the way in which ideas may be causally constitutive of social processes. But, to assess that potential, we must turn to authors, like Margaret Archer, who have built their approach to such matters from ontological premises that are both more consistent, and more consistently critical realist.6 Although Archer’s approach to the question of structure and agency has become increasingly influential in international relations and political science, her approach to culture and agency has yet to receive sustained reflection in the ideas debate (for a partial exception see McAnulla 2002). This is unfortunate. For although her approach is not without its problems, it offers perhaps the most theoretically sophisticated and ontologically consistent attempt to date to develop an approach to the ideational-material relationship which privileges neither moment. The analytical starting point for the morphogenetic approach is the attempt to overcome what Archer terms the ‘fallacy of conflation’. By this she refers to any conceptualisation of the relationship between the material (‘structure’) and the 6. Though Archer’s approach is undoubtedly critical realist in its inspiration and undoubtedly consistent in its critical realism, it is by no means the only consistently critical realist perspective – nor, as will become clear presently, is it how we would choose to develop a critical realist perspective on the materialideational relationship. We focus on it here since it provides the most systematic and clearly stated critical realist treatment of ideas and culture. -29- ideational (‘culture’) which, by blurring the distinction between them, precludes an examination of their interplay. 7 Conflation, she suggests, is problematic since it either treats “structure and culture as partners in a zero-sum relationship” or sees their partnership “as one of such tight mutual constitution that neither has any degree of autonomy from the other” (Archer 1988/99: 274). The former reminds us of the ideas- versus material interests dichotomy imposed by the rationalist reading of ideas, while the latter is close to some versions of postmodernism that deny any degree of autonomy to material factors. Thus, for Archer, the first challenge in conceptualising culture, structure and agency is to overcome the conflation of structure and culture and that of culture and agency so as to provide an approach that is capable of linking structure and culture to agency without sinking their differences (xiv). Any failure to do so can only result in an elision of “the material and ideational aspects of social life” such that they become “clamped together in a conceptual vice” (xi). 2008/11 • IBEI Working Papers But, as this suggests, and as is so often the case in such matters, the proof of this particular pudding lies in the eating. Thus, as she continues, “it is one thing to accept the advantages of drawing such distinctions in principle”; it is quite another to operationalise the distinction in such a way as genuinely to enhance our analytical purchase on the ideational-material relationship (xviii, emphasis added). Her focus throughout is on the additional purchase as practical social scientists that we gain from such abstract analytical considerations. If it is to offer genuine valueadded, she suggests, her proposed analytical dualism has to be followed through to the identification of the different and distinct properties and dynamics of the terms comprising the dualism. Moreover, these properties must be ascribed in such Archer’s solution is to propose that we conceptualise the relationship between the material and the ideational in terms of an analytical dualism. This emphasis on the need for a clear analytical separation of the material and ideational is distinctive to her approach. It provides, she suggests, the methodological springboard to an examination of the interplay, rather than the interpenetration, of the material and the ideational over time. Yet this analytical dualism, as Archer is keen to emphasise, does not entail the adoption of a Cartesian (ontological) dualism, as Wendt suggests. Thus, contra authors like Wendt, Archer insists that proposing an analytical dualism is “emphatically … not the same thing as” proposing a philosophical (i.e.: ontological) dualism, “for there is no suggestion that we are dealing with separate entities, only analytically separable ones and ones which it is theoretically useful to treat separately” (xvi, emphasis added). Archer clearly sees this as a very important clarification, for it is repeated at several key points in her account. Analytical dualism, she emphasises, is a heuristic devise and “an artifice of convenience” (143). 7. Rather unfortunately, to our way of thinking, Archer insists on referring to the relationship between the ideational and the material as that between culture and structure. This formulation we resist in our own work, since we prefer to see social structure as comprising both material and ideational elements. Consequently, wherever possible in what follows (i.e.: except where making reference to direct citations), we will substitute the ‘ideationalmaterial’ pairing for Archer’s ‘culture-structure’ pairing. -30- a way that the analytical dualism is respected and retained – the (potential) additional analytical purchase offered by an analytical dualism is lost if, in operationalising the distinction, dualism gives way to duality (105). Two principles guide Archer in this task. The first is that she must guard against what she terms the ‘the myth of cultural integration’. This, she suggests, “perpetuates an image of culture as a coherent pattern, a uniform ethos or a symbolically consistent universe” (xvii). The danger of this is that it forecloses the possibility of contradictions within the ideational/cultural realm, with the result that we may tend to emphasise path dependent processes of ideational/cultural reproduction at the expense of path-shaping processes of ideational/cultural transformation. For Archer the cultural/ideational system acquires an existence of its own; it is real and objective. Its components (theories, beliefs and so on) have emergent relational properties that condition (in a causally constitutive way) specific courses of action. Moreover, each such system is characterised by an internal logical consistency which gives a unity to the parts out of which it is constituted. IBEI Working Papers • 2008/11 Second, and in line with the principle of analytical dualism, Archer seeks to keep analytically separate the cultural and the agential. This she does by ensuring that the nature of the relations between the components of a culture are distinguished clearly from those pertaining to the actors. The former relations are captured by the notion of causal consensus, defined as “the degree of social uniformity produced by the imposition of culture … by one set of people on another” (4). By contrast, the relations between groups and individuals are, for Archer, ultimately “rooted in ... material interests” (xxii). We have now established the core principles, both ontological and analytical, from which Archer develops her distinctive approach to the interaction between culture and agency. This is outlined schematically in Figure 5 and takes the form of the ‘morphogentic cycle’, as Archer rather grandly puts it. This cycle is, in fact, comprised of three overlapping but temporally distinct processes. These Archer terms cultural conditioning, socio-cultural interaction and cultural elaboration; they mirror equivalent processes in Archer’s treatment of the structure-agency relationship (1995: 193, 309; for a critique of which see Hay 2002: 122-6). We consider each in turn. Figure 5: The morphogenesis of culture (Archer, 1995: 193) Cultural Conditioning T1 Socio-Cultural Interaction T2 T3 Cultural Elaboration T4 1. Cultural Conditioning. Given that, for Archer, the cultural system predates the socio-cultural action which ultimately reproduces or transforms it, the -31- 2. Socio-cultural integration. Were the analysis to stop at this point, a form of cultural determinism would be implied. This Archer eschews by incorporating the role of the reflexive actor in a second phase of the cycle. Situational logics, though conditioning are never fully determinant of a reflexive actor’s conduct. Recognition of this serves, then, “to bring the people back in, not merely as static upholders of this or that idea, but as active makers and re-makers of their culture” (184). But recall here that Archer conceives of the relationship between actors as of a qualitatively different character to that between elements of the prevailing cultural system. Indeed, as we have seen, for Archer the orderly or conflictual nature of the relations between actors at this level reflects “structural antagonisms based on material interests” (187). As this suggests, social conflict may be rife even in situations characterised by high levels of cultural consistency. Similarly, cultural contradictions may arise in situations characterised by high levels of social order and the absence of social conflict. This directs Archer to the importance of the “discrepancies between the degree of integration at the two levels” (186) and to the affect of such discrepancies on the third phase in the cycle, the process of cultural elaboration. 3. Cultural elaboration. It is the outcome of the process of cultural elaboration that determines whether the overall cycle is ‘morphogenetic’ (i.e.: transformatory) or ‘morphostatic’ (i.e.: reproductive) of the prevailing cultural system. As this makes clear, there is no in-build prejudice towards change in the model. Indeed, whether the outcome of the cycle will be cultural transformation or cultural reproduction is determined, for Archer, by the distribution of contradictions and complementaries between the two systems. In general, and at the risk of some simplification, if the degree of integration at both levels is high then nothing points towards change; similarly, if the degree of integration at both levels is low then the converse is likely to apply. In either case, the outcome of the process of cultural elaboration completes the morphogenetic/static cycle and marks the beginning of a new cycle. As the above paragraphs hopefully serve to indicate, Archer pursues her project with an unusually high degree of conceptual rigour and theoretical consistency. Yet, above all, her approach is distinguished by a refreshing desire to -32- morphogenetic cycle inevitably takes off from cultural conditioning. This refers to the process by which the cultural context (itself an outcome of contextualised agency in previous cycles) serves to impose situational logics upon behaviour. For Archer such cultural effects are largely determined by the degree of logical consistency exhibited by the culture in question. In this formulation, culture is perceived as a dispositional register that conditions agents by providing them with specific situational logics. If the degree of cultural logical consistency is low then a condition of constraining contradictions emerges. The appropriate situational logic in such a context is one which demands that actors attempt to resolve or repair these contradictions. Yet, if the degree of internal consistency is high then a condition of concomitant complementarities emerges. Here the appropriate situational logic is one in which actors are mainly concerned to ensure the continuation of a condition of cultural consistency. 2008/11 • IBEI Working Papers ensure that philosophical commitments and analytical strategies are followed through to deliver practical social scientific insights. As she emphasises repeatedly, ontological exertions, important as they are, cannot suffice. “The practical analyst of society needs to know not only what social reality is, but also how to begin to explain it” (Archer, 1995: 5, original emphasis). There is much in this from which contemporary political science and international relations can learn. IBEI Working Papers • 2008/11 This brings us to a third insight of Archer’s account, namely its emphasis on the emergent properties of ideas. Archer argues that the myth of cultural integration is problematic because it forecloses the possibility of contradictions between ideas as components of a culture. Indeed, as she maintains, ideas develop relational emergent properties of logical consistency or inconsistency which, in turn, condition the cultural context within which action takes place. In the process she comes to ascribe to ideas causal powers. Such insights are undoubtedly important, but they shouldn’t lead us to overlook the problems with Archer’s perspective. For, ultimately, we contend, it fails to deliver the full potential of the critical realism to which it is clearly wedded to advance a truly dialectical perspective on the ideational-material relationship. A second, and no less significant, insight is the need to guard against the myth of cultural integration the tendency to conceive of culture as a seamlessly integrated and hermetic inter-subjective consensus. Such a tendency, we would again contend, is no less prevalent in international relations and political science than it is in sociology – and it is certainly no less problematic. Indeed, there is a second dimension to this. This is overlooked by Archer, but it is perfectly consistent with her approach. This is the degree of correspondence between ideas at different levels of generality within the same culture or discourse. As Craig Parsons puts it, albeit in a slightly different context, “actors who share one idea as the ‘master frame’ for their action do not necessarily agree on more detailed actions therein” (Parsons 2003: 8). Yet it is with the specific contribution that the morphogenetic approach makes, or might make, to ideational analysis that we are here principally concerned. Again, there is much to commend Archer’s perspective. First, she is undoubtedly right to point to the tendency just as prevalent in international relations as in sociology for the question of structure and agency to overshadow the relationship between the material and the ideational. She is also surely right to see the two issues as intimately interlinked (see also Hay 1999; 2002). This we see as the first key insight to be gained from Archer’s analysis. Archer sets out to examine the interplay between culture and agency but also between culture and structure. Indeed, she notes that the key issue animating her approach “is how to disentangle the interplay of interests, power and ideas” (1988/96: 188). Yet this remains, in our view, at best an unrealised aspiration, as her theoretical efforts tend to privilege the relationship between culture and agency at the expense of that between culture and structure. -33- So what is so problematic about this? By restricting the interplay of culture and structure to something that occurs only in the second ‘socio-cultural interaction’ phase of the morphogenetic cycle, Archer downgrades its significance. Moreover, and more problematically still, since she conceives of actors in the process of socio-cultural integration as animated ultimately by their material interests, she leaves entirely unchallenged an ultimately rationalist conception of the human subject. Thus, despite all her efforts to develop a dynamic and avowedly dialectical understanding of the material-ideational relationship, she falls back on precisely the same rationalist foundations with which we began our consideration of the ideational turn. Thus, first and last, the picture she paints is of rational actors pursuing their materially-given self-interests. And, as already argued, to conceive of interests as materially determined in this way is in direct violation of the commitment to a dialectical understanding of the relationship between the material and the ideational. It is also to turn an analytical dualism into precisely the ontological dualism Archer sought to overcome. Yet the problem here is not just Archer’s perhaps surprising commitment to a purely materialist conception of interests. Equally problematic, from a perspective seeking to explore the relationship between the ideational and the material without privileging either moment, is that actors are seen to be empowered by their position in the social hierarchy – where the latter is, again, understood in purely material terms. Thus, despite setting out to explore the ‘interplay of interests, power and ideas’, both interests and power are understood in purely material terms – there being no recognition, for instance, that actors may be -34- McAnulla’s view is that the morphogenetic approach is ambiguous about the relationship between the cultural and the structural over time and that, “although Archer has offered some useful, if brief, reflection on the nature of this relationship, it is not something made clear” (2002: 290). This is overly generous. Indeed, the only element of McAnulla’s assessment with which we agree is that Archer’s treatment of the relationship between culture and structure is brief. For it is, we suggest, unambiguous – and unhelpful. McAnulla can find no developed account of the relationship between culture and structure over time because, quite simply, there is none. And this is so not because Archer ’s work is theoretically underdeveloped on this point nor, indeed, because of some ambiguity in the text, but because Archer explicitly denies such a diachronic interplay between the two. In fact, she is crystal clear on the subject: “When both structure and culture are conceptualised from the morphogenetic perspective … the two intersect [only] in the middle element of the basic cycle” (Archer, 1988/96: 282). Moreover, in contrast to more sympathetic critics, we detect a more fundamental problem here. For we do not see this, like McAnulla (2002), as an underdeveloped aspect of her work. Rather we see it as evidence of a residual philosophical dualism between the material (structure) and the ideational (culture) that violates her ostensible commitment to analytical dualism and that is in marked contrast to her more effective treatment of the relationship between culture and agency. This, as we shall see, can ultimately be traced to the unfortunate conflation of the material and the structure. 2008/11 • IBEI Working Papers Archer’s restrictive conception of reflexivity is further revealed in her account of the way in which specific situational logics are imparted to actors according to the degree of logical integrity and coherence of a given cultural system. This is unduly mechanistic and, again, belies any notion of actors’ reflexivity. If contradictions and complementarities are to motivate behaviour they have to be perceived as such. It is what actors – not analysts make of them that matters here. Whether accurate or otherwise, it is the contradictions we perceive that motivate us to seek an alternative. By the same token, the absence of logical contradictions does not preclude the possibility that we will taken action against the contradictions we perceive to exist. -35- A second – and closely related problem with the morphogenetic perspective is its restrictive treatment of actors’ reflexivity. Archer insists that every form of determinism which might be seen to be implied by her conception of cultural conditioning should be avoided since it denies the “quintessential reflexive ability of human beings to fight back against their conditioning” (Archer, 1988/96: xxvi). Few contemporary theorists would disagree with such a statement. Yet, as operationalised by Archer, agents are able to ‘fight back against’ their cultural conditioning only in so far as it clashes with their material interests and only in so far as their material position gives them the power resources to do so. There is nothing very reflexive about this, certainly by any conventional understanding of the term. Reflexivity certainly doesn’t extend to any capacity of the actor to interpret or to re-interpret his or her material interests. In this respect Archer’s position is eerily reminiscent of Goldstein and Keohane’s argument that ideas “become important solely because of the interests and power of their progenitor” (1993: 13). The source of all of these problems is common the separation of culture and structure in the morphogenetic approach. For it is this that prevents them from intersecting at any point other that in the second phase of the morphogenetic cycle – and arguably then only in a rather circumscribed way. It is precisely for this reason that, as noted earlier, we prefer to talk of a social structure with interrelated, if nonetheless analytically separable, ideational and material elements. Ultimately, then, it is Archer’s conflation of the material and the structural that is the source of her failure to respect the principles of analytical dualism and yet ontological duality she espouses. Indeed, one might go further here. The problem, for us, with Archer’s approach is, in the end, her commitment to analytical dualism. This, we suggest, is ultimately unnecessary. For, whilst it is indeed important to avoid the conflation of the ideational and the material, the agential and the structural, and thus to retain the analytical separability of the terms comprising such pairings, we do not accept that this can only be achieved by imposing an artificial analytical dualism. Moreover, there are associated dangers with such a strategy – namely, that by assuming that all emergent properties of social and cultural systems are internal to them, we fail adequately to consider the interplay between them. In this way an analytical dualism all too frequently gives way to an ontological dualism. powerful not (purely) by virtue of their material capabilities but also by virtue of a widespread system of meanings that accords them status, legitimacy and/or authority. IBEI Working Papers • 2008/11 2008/11 • IBEI Working Papers 5. Conclusions: Towards a Critical Realist Constructivism We have travelled a considerable distance in the preceding pages and yet we have still come up short – in the sense that a consistent approach to the relationship between the ideational and the material that fails to privilege one or other moment and is capable of recognising the causally constitutive role of ideas remains frustratingly elusive. Yet we are rather better placed now to take stock of the failings of existing approaches and perhaps to go one better than we were at the outset. Those are the twin goals we set ourselves in this final concluding section. If we consider first the limitations of existing perspectives, a number of common themes emerge from our analysis. Four in particular stand out. 1. A failure to acknowledge the ontological and epistemological implications of ‘taking ideas seriously’. Mainstream political science and international relations theory is undoubtedly more enthusiastic today than it has been for several decades to accord a genuinely explanatory (and hence causal) role to ideas and ideational variables. Indeed, since the mid 1990s a veritable academic industry has developed as analysts have increasingly turned (or, perhaps more accurately, returned) to ideas. Yet, to date at any rate, this literature has been decidedly less enthusiastic to explore fully or adequately the ontological and epistemological implications of according to ideas an explanatory role. Far too frequently, ideas have simply been grafted on to preexisting theories whose ontological and epistemological premises accord no role to ideas. They have, as a consequence, been accorded a rather limited status as auxiliary, residual or supplementary variables – incorporated only at precisely the point at which pre-existing theories run in to explanatory difficulties. Though convenient, this is deeply problematic. For perhaps more so than any other category of variable, to accord ideas an explanatory or, indeed, a constitutive role (where previously such a role was denied to them), is to revise fundamentally ones ontological commitments with major epistemological consequences. In particular, as we have been at pains to suggest throughout, to posit ideas as causes or to see ideas as (partly) constitutive of the social is to reject naturalism and with it the positivist epistemological self-confidence it has sustained. As this suggests, ideas cannot be incorporated within existing mainstream perspectives without ontological and epistemological residue. 2. The problem of invoking uncertainty as a necessary condition of a causal role for ideas. It is perhaps not entirely fair to suggest that all mainstream attempts to accord to ideas a greater role fail to acknowledge the epistemological implications, in particular, of so doing. But those that do have invariably sought to limit such epistemological implications by specifying the boundary conditions of a causal role for ideas. In almost all such cases ideas are held to become causally significant only under conditions of uncertainty. This is undoubtedly neat. Moreover, it would certainly appear to allow us to accord to ideas a causal role whilst, at least for the most part, retaining our former materialist ontological commitments and the epistemological self-confidence they in turn facilitate. Yet this, we contend, is illusory – or, at least, it -36- 3. The limits to the Humean conception of causality. Time and again our analysis has revealed the limitations of the Humean conception of causality which, we contend, continues to plague almost all of the literature that strives to accord an explanatory and/or constitutive role to ideas. The Humean notion of causality relies on the principle of constant conjunction – such that, if A causes B, then “all instances of A must be followed by the appearance of B” (Parsons 2003: 105). This, we have argued, is fundamentally incompatible with the non-naturalist ontology that results from according to ideas (any) explanatory significance. For, quite simply, if actors’ behaviour is shaped to any extent at all by the ideas they hold, and such ideas are not simply reducible to the context in which they arise, then there are no constant conjunctions to which a Humean conception of causality might appeal. As this suggests, to accord an explanatory or constitutive role to ideas is to disavow naturalism; and to disavow naturalism renders a Humean conception of causation redundant. Yet tempting at this point though it might be to follow many postmodernists in rejecting causal in favour of purely constitutive logics, this too we reject. For there is another option here. That is to reject instead the narrowly Humean conception of causation that continues to dominate the existing literature and to prefer, instead, a broader critical realist conception of causality (see, for instance, Collier 1994: 122-30). This drops the constant conjunction requirement, replacing it with a conception of ‘generative mechanisms’. The key point in the present context is that this allows a recognition of the causal significance of constitutive processes. 4. The limits of a material conception of interests. A further and yet seemingly constant stumbling block to the attempt to accord to ideas a more significant explanatory role is the notion of interests as materially-given. Yet it is important to be clear here. For, in the end, whether interests are seen to be materially given or, as we prefer, as constructions which orient an actor to his or her context, is a largely semantic choice – though one loaded with ontological significance. As such it cannot easily be judged right or wrong; and it is not our argument that such a choice is, in itself, misguided. But and this is our point to conceive of interests as materially-given is profoundly anathema to the attempt to accord to ideas a causal significance. As we have sought to explain, the appeal of the material conception of interest is that, in combination with assumptions about rationality and complete information, it serves to render actors’ behaviour predictable given the context in which they find themselves. If actors are rational, have a clear and materially-given self-interest and -37- is in the absence of a theory specifying the conditions under which uncertainty arises and the conditions under which it is resolved. For, in the absence of such a theory, whether we are prepared to accord to ideas an explanatory role or not is essentially arbitrary – and to invoke uncertainty as the proximate justification for so doing is tautological. In effect, whenever we find that materialist explanations will not suffice we invoke uncertainty and seek an ideational solution; whenever they will suffice we assume the absence of uncertainty and refuse to concede the causal significance of ideas. This is clearly inadequate. Moreover, rather than seeing uncertainty as the exception to the rule, is it not more credible to see it as a universal and timeless human condition? Yet if we do so, then we should expect ideas to matter all the time – an ontological commitment with precisely the epistemological implications that the uncertainty clause was designed to protect us from! IBEI Working Papers • 2008/11 can be assumed to be accurate in their assessment of both the context in which they find themselves and that interest, then they will behave predictably – their behaviour is, in effect, determined by their surroundings (see also Hay 2004, 2008). Such a conception is conceptually neat, theoretically elegant and analytically parsimonious. More importantly still it holds open the prospect of a predictive science of politics – and it also enables the elucidation of causes in the Humean sense of the term. Yet, as the turn to ideas amongst rational choice theorists itself serves to indicate, its ontological premises are increasingly seen to lack credibility, chosen as they are largely for analytical convenience. More credible, it seems, to many (if not perhaps all) rational choice theorists today is that behaviour is not predictable given context, because of the intervening or mediating role of ideas. Yet rather than see interests themselves as irredeemably ideational, most analysts to date (rationalists or otherwise) still cling to a material conception of interests. In so doing they argue that actors do indeed have materially given interests but that it is their perceived interests rather than their true interests that ultimately motivates their behaviour. This, we suggest, is an unnecessary and unfortunate formulation which betrays its rationalist legacy and which can and should now be dispensed with. For it implies that, given complete information, identically located actors would reach identical assessments of their material interests – their perceived and true interests would be one and the same. This is yet another convenient fiction. Moreover, if one accepts that it is actors’ perceived interests that motivate their behaviour, it is a convenient fiction which no longer does any analytical work – for actors’ true interests are largely irrelevant to their behaviour. What this suggests to us is the value of a radical reassessment of the concept of self-interest itself. It is such a reassessment that we would place at the heart of a critical realist constructivism. Within such a formulation, interests do not exist, but constructions of interest do. Such constructions are, in turn, predicated on irreducibly normative conceptions of self good – of what it would advantage the individual to do or to have done either on her behalf or inadvertently by others. They are idealised extrapolations of subjective (or, if shared, inter-subjective) preferences and, as such, are different from immediate and/or particular desires. We may, after all, desire that which we acknowledge not to be in our perception of our own best interest. Conceptions of self-interest provide a cognitive filter through which the actor orients herself towards her environment, providing one (of several) means by which an actor evaluates the relative merits of contending potential courses of action. But such conceptions, though they arise out of an ongoing interaction with that context, are neither given or determined by it, nor given or determined by the actor’s knowledge of it. They have an autonomy from it. They reflect an actor’s subjective preferences as to the things she value and the relative values she assigns to the desires she can imagine. Thus, however conventional the actor’s conception of her own self-interest may be, it is hers alone. It is, crucially, about what she value and the relative value, in the abstract, that she assigns to the different conditions and possibilities that she can envisage at any given point in time. How might such a conception of interests inform differently constructivist analysis in political science and international relations? Consider the rise of -38- 2008/11 • IBEI Working Papers monetarism in the advanced liberal democracies out of the ‘crises’ of the late 1970s. This topic has, perhaps unremarkably, attracted considerable attention from constructivist institutionalists. For it sees a profound and rapid change in the ideas informing economic policy in these democracies, which more conventional approaches have generally failed to explain – save other than by appeal to exogenous shocks (which, themselves, remain unexplained). The key question in this literature is why business in particular came to back monetarism, acting both as an ideational entrepreneur and as a financial backer for it. And the standard answer, in so far as one can be found, is that it did so because monetarism served the interests of business. Yet this proposition is itself never really defended, it presumably being assumed that it is largely self-evident. For what it is worth, we think it is to concede far too much to monetarism to assume that it has led, for instance, to higher rates of profit than would otherwise have been the case. Yet the point is that we simply do not need to know whether monetarism was an expression of the ‘real’ interests of capital or not, in order to explain its origins and ascendancy. A number of points might here be made. First, the notion that monetarism was promoted by business because it served the interests of business to do so is simply wrong. It is wrong not because monetarism was not in the interests of business (that is almost impossible to adjudicate and, as we hope to show, an irrelevance here). It is wrong because monetarism was promoted by business not because it did serve business interests to do so, but because it was perceived to serve business interests to do so. That may seem like a very trivial and largely semantic distinction, but it is in fact crucial. And it is crucial because it changes the nature of the question we must now answer to explain the rise of monetarism. IBEI Working Papers • 2008/11 For a genuine constructivist, to explain monetarism is not to explain how and why monetarism was a reflection of the material interest of business (and/or its other backers), but to explain how and why business leaders (and others) came to perceive of their interests in such terms. That is, in fact, not a very difficult question to answer. But it is a question that largely remains unanswered within the existing literature. That business leaders backed monetarism, believing it to be in their interests to do so was, we contend, in large part because they had for a long time been accustomed to conceiving of their interests in a manner that gave a high value to the things monetarism prioritised (such as price stability) and a low value to those previous constraints on policy it tended to discount (consensus, employment, universal welfare free at the point of access, and so forth). Monetarism, then, didn’t resonate directly with a given set of extant material interests, but with a particular conception of such interests – a conception which, to be sure, it partly came to reconfigure. This, of course, is a far from complete or adequate answer to the question of how and why business leaders came to conceive of monetarism as a set of economic policies that it would benefit them from promoting. But it does hint at the way in which a more consistently constructivist approach might provide answers to some of the thorniest puzzles in political analysis – notably concerning the determinants of rapid institutional and ideational change which existing approaches tend to exogenise. It suggests, too, that such answers do not lie in a materialist conception of -39- self-interest and that the advantage of a consistently constructivist approach is that it does not need to deal with the question of adjudicating the real interests of the actors whose behaviour it studies. Finally, it is also suggestive of a research agenda for the further development of such a constructivist institutionalism. Such a research agenda would focus far more explicitly than it has to date on the social and political processes through which interests are identified, constructed and rendered, in Mark Blyth’s terms, ‘actionable’. As this would suggest, the future of constructivist institutionalism surely lies in detailed ethnographic research which maps and charts the development and redevelopment of interest perceptions rather than the abstract and deductive derivation from stylised assumptions of the ‘real’ interests of institutionally-embedded actors. 2008/11 • IBEI Working Papers References Adler, E. 1997. Seizing the Middle Ground: Constructivism in World Politics. European Journal of International Relations, 3 (3): 319-363. Archer, M. S. 1988/99. Culture and Agency: The Place of Culture in Social Theory. Cambridge: Cambridge University Press. Archer, M. S. 1995. Realist Social Theory: The Morphogenetic Approach. Cambridge: Cambridge University Press. Berger, P. L. and T. Luckmann. 1966. The Social Construction of Reality. London: Allen Lane The Penguin Press. Berman, S. 2001. Review Article: Ideas, Norms, and Culture in Political Analysis. Comparative Politics, 33 (2): 231-250. Bernstein, S. 2000. Ideas, Social Structure and the Compromise of Liberal Environmentalism. European Journal of International Relations, 6 (4): 464-512. Bhaskar, R. 1979. The Possibility of Naturalism: A Philosophical Critique of the Cotemporary Human Sciences, 3rd ed. London: Routledge. Bhaskar, R. 1989. Reclaiming Reality. London: Verso. Blyth, M. 2002. Great Transformations: the Rise and Decline of Embedded Liberalism. Cambridge: Cambridge University Press. Blyth, M. 1997. Any More Bright Ideas? The Ideational Turn of Comparative Political Economy. Comparative Politics, 29 (2): 229-250. Campbell, J. L. 2000. Convincing the Sceptics: Six Questions for Ideational Analysis. Paper presented at the conference on “Ideational Institutionalism: Perspectives on European Politics”, University of Birmingham. Collier, A. 1994. Critical Realism: An Introduction to Roy Bhaskar’s Philosophy. London: Verso. Doty, R. L. 2000. Desire All the Way Down. Review of International Studies, 26 (1): 137-139. Finnemore, M. 1996. National Interests in International Society. Ithaca: Cornell University Press. Fligstein, N. and I. Mara-Drita. 1996. How to Make a Market: Reflections on the Attempt to Create a Single Market in the European Union. American Journal of Sociology, 102 (1): 1-33. Gofas, A. 2003. Ideational Analysis and the Discursive Construction of Uncertainty: Some Conceptual Considerations. Paper presented at the BISA 28th Annual Conference, University of Birmingham, UK. Goldstein, J. 1993. Ideas, Interests and American Trade Policy. Ithaca: Cornell University Press. Goldstein, J. and R. O. Keohane. 1993. Ideas and Foreign Policy: An Analytical Framework. in J. -40- IBEI Working Papers • 2008/11 Goldstein and R. O. Keohane (eds). Ideas and Foreign Policy: Beliefs, Institutions and Political Change. Ithaca: Cornell University Press. Hacking 1999. The Social Construction of What? Cambridge, MA: Harvard University Press. Hay, C. 2007. Constructivist Institutionalism. In R. A. W. Rhodes, S. Binder, and B. Rockman (eds.). The Oxford Handbook of Political Institutions. Oxford: Oxford University Press. Hay, C. 2005. Making Hay…or Clutching at Ontological Straws? Notes on Realism, ‘As-If-Realism’ and Actualism. Politics, 25 (1): 39-45. Hay, C. 2002. Political Analysis: A Critical Introduction. Basingstoke: Palgrave. Hay, C. 1996. Narrating Crisis: The Discursive Construction of the ‘Winter of Discontent’. Sociology, 30 (2): 253-277. Hay, C. and Rosamond, B. 2002. Globalisation, European Integration and the Discursive Construction of Economic Imperatives. Journal of European Public Policy, 9 (2): 147-167. Hollis, M. and Smith, S. 1990. Explaining and Understanding International Relations. Oxford: Clarendon Press. Jacobsen, J. K. 2003. Duelling Constructivisms: A Post-Mortem on the Ideas Debate in Mainstream IR/IPE. Review of International Studies, 29 (1): 39-60. Jarvis, D. S. L. 1998. Postmodernism: A Critical Typology. Politics & Society, 26 (1): 95-142. Joseph, J. 2007. Philosophy in International Relations: A Scientific Realist Approach. Millennium: Journal of International Studies, 35 (2): 345-359. Kegley, C. W. Jr. 2002. Bridge Building in the Study of International Relations: How ‘Kuhn’ We Do Better?. In D. J. Puchala (ed.). Visions of International Relations: Assessing an Academic Field. Columbia: University of South Carolina Press. Keohane, R. 2000. Ideas Part-Way Down. Review of International Studies, 26 (1): 125-130. Krasner, S. D. 2000. Wars, Hotel Fires, and Plane Crashes. Review of International Studies, 26 (1): 131136. Kratochwil, F. 2000. Constructing a New Orthodoxy? Wendt’s ‘Social Theory of International Politics’ and the Constructivist Challenge. Millennium: Journal of International Studies, 29 (1): 73-101. Kukla, A. 2000. Social Constructivism and the Philosophy of Science. London: Routledge. Kurki, M. 2006. Causes of A Divided Discipline: Rethinking the Concept of Cause in IR. Review of International Studies, 32 (2): 189-216. McAnulla, S. 2002. Structure and Agency. In D. Marsh and G. Stoker (eds). Theory and Methods in Political Science, 2nd ed. Basingstoke: Palgrave. Onuf, N. 2002. Worlds of Our Making: The Strange Career of Constructivism in International Relations. In D. Puchala (ed.). Visions of International Relations: Assessing an Academic Field. Columbia: University of South Carolina Press. Osterud, O. 1996. Antinomies of Postmodernism in International Studies. Journal of Peace Research, 33 (4): 385-390. Parsons, C. 2003. A Certain Idea of Europe. Ithaca: Cornell University Press. Parsons, C. 2007. How to Map Arguments in Political Science. Oxford: Oxford University Press. Read, R. and Richman, K. (eds). 2000. The New Hume Debate. London: Routledge. Ruggie, J. G. 1998. Constructing the World Polity: Essays in International Institutionalisation. London: Routledge. Schimdt, V. 2007. Discursive Institutionalism: The Explanatory Power of Ideas and Discourse. Annual Review of Political Science, forthcoming. Smith, S. 2000. Wendt’s world. Review of International Studies, 26 (1): 151-163. Suganami, H. 2002. On Wendt’s philosophy: a critique. Review of International Studies, 28 (1): 23-37. Wendt, A. 1998. On Constitution and Causation in International Relations. Review of International Studies, 24 (5): 101-117. -41- 2008/11 • IBEI Working Papers Wendt, A. 1999. Social Theory of International Politics. Cambridge: Cambridge University Press. Yee, A. S. 1996. The causal effects of ideas on policies. International Organisation, 50 (1): 69-108. Yee, A. S. 1997. Thick Rationality and the Missing ‘Brute Fact’: The Limits of Rationalist Incorporations of Norms and Ideas. The Journal of Politics, 59 (4): 1001-1039. -42-