The Web Site for Critical Realism The WSCR Archive
[The following article, originally published by First Press: Writing in the Critical Social Sciences (2001), is archived here with the kind permission of the author, Colin Hay, who retains the copyright.]

What Place for Ideas in the Structure-Agency Debate? Globalisation as a 'Process Without a Subject'



Ideas are invariably the spectre at the feast in the debate on structure and agency. From sociology to international relations theory the tendency has been to keep separate the twin conceptual pairings structure/agency, material/ideational. Given the considerable influence of avowedly constructivist perspectives in international relations' structure-agency debate this is all the more remarkable. Yet the issues of structure and agency, the ideational and the material cannot be separated and should not be 'bracketed'. Though analytically separable, they are ontologically intertwined. The argument is illustrated with respect to globalisation a term and related literature whose silence on the questions of structure and agency and the relationship between the ideational and the material has resulted in clear tendencies for obfuscation. Globalisation tends to be depicted as a process without a subject. Restoring subjects to the process of globalisation and assessing the extent to which their behaviour is informed by constructions of globalisation is, I suggest, crucial to the broader task of demystifying globalisation and of challenging its perceived logic of no alternative.



"Given the supposedly abortive attempts at solving the structure-agency problem, one is tempted to conclude that sociologists are not smart enough to solve the problem or that the problem itself is spurious".

It is surely tempting to suggest that the extent to which the terms structure and agency are regarded as 'problematic' gives an indication of the extent to which a discipline is in crisis. On such a reckoning sociology and social theory are in a permanent state of crisis, political science periodically so. Moreover, international relations, by such a token, is in a condition of intense and rare, indeed, unprecedented crisis.

This judgement may seem be a little harsh, at least on sociology. For arguably that sociologists and social theorists have mired themselves in controversy over the relationship between the individual (or collectivity) and the social context is less an indication of the poverty of sociology than of the innate difficulty of its subject matter and its perhaps laudable desire to deal explicitly with issues the other social sciences prefer to leave unasked or, at best, implicit. This tendency to rely upon unexplicated and unacknowledged assumptions about structure and agency is especially characteristic of political science. Until recently it was equally typical of international relations. It evidenced (and in the case of political science continues to evidence) a certain ontological and epistemological complacency on the part of the discipline's principal protagonists. In international relations that confidence has long since evaporated. The discipline's core is up for grabs. In such a context it is certainly appropriate though no less unexpected for that that debate has turned for the first time in a long time to the 'big questions': those of structure and agency and of the material and the ideal. Alexander Wendt's identification of the need for a Social Theory of International Politics is here symptomatic of contemporary develoments. As a title it would have been unthinkable ten years ago.

In this paper my aim is to take the discussion of structure and agency in a slightly different direction, reminding us of the sometime spectre at the feast the realm of ideas. Whether in sociology, social and political theory, political science or, now, international relations, the tendency has been to keep separate the twin conceptual pairings structure/agency, material/ideational. Given the considerable influence of avowedly constructivist perspectives in international relations' structure-agency debate this is all the more remarkable. I suggest that the issues of structure and agency, the ideational and the material cannot be separated and should not be 'bracketed'. To deal adequately with the 'structure-agency problematique', as it has come to be known in international relations scholarship, is also to deal with the 'material/ideational problematique'. Though analytically separable, they are, I suggest, ontologically intertwined.

In pushing the debate in this direction, I seek to develop further a distinctive strategic-relational or critical realist approach. I do so, however, in such a way as to give rather greater sensitivity to the role of ideas in the unfolding and dynamic relationship between agents and the environment in which they find themselves. This perspective I illustrate with respect to globalisation a term and related literature whose silence on the question of structure, agency and, indeed, causation has resulted in clear tendencies for obfuscation. Restoring subjects to the process of globalisation a process without a subject is, I suggest, crucial to the broader task of demystifying globalisation and of challenging its perceived logic of no alternative.

The argument unfolds in three stages. In the first section, after a series of conceptual and theoretical preliminaries, I seek to establish the distinctiveness of the strategic-relational perspective developed here, situating it in terms of the broader discussion of structure and agency in contemporary international relations theory. Having done this, I seek to demonstrate how such a perspective might approach the question of globalisation, restoring conceptions of subjectivity to a process without a subject. I conclude by assessing the implications of this for contemporary international/global political economy, examining in particular the causal significance of the discursive construction of globalisation and the rhetorical appeal to globalisation as a logic of no alternative. I begin, however, with some general comments on the question of structure and agency within international relations scholarship.

What is not at stake in the agent-structure debate: Preliminary remarks on the 'structure-agency problematique'

At the outset it is imperative that we establish (with appropriate apologies to David Dessler) precisely what is at stake in the agent-structure debate and, more significantly, what is not. All too frequently sociologists, political scientists and international relations scholars present their chosen 'solution' to their particular rendition of the 'structure-agency problematique' as a universal panacea for all social science ailments, be they ontological, epistemological or methodological. To put it bluntly, whilst the agent-structure debate is important, it is not that important; moreover, it is imperative that we get its significance in some kind of perspective. This is the task I set myself in these necessarily brief preliminary remarks.

The unproblematic nature of the 'agency-structure problematique'

It is perhaps appropriate that we begin with Steve Fuller's superficially attractive and seemingly well-founded scepticism. Sociologists, he suggest, have spent two hundred years on the issue of structure and agency. Yet they have got no further than Marx's truism "men make history, but not in circumstances of their own choosing". This must surely lead us to question their intelligence either for failing to provide a solution to their question or for posing themselves such an imponderable in the first place. Moreover, such reflections would suggest that there is little to be gained by international relations theorists and political analysts more generally in following their sociological forebears into this cul-de-sac of obfuscation and meaningless abstraction.

However tempting it may be to concur and to terminate the discussion at this point, a couple of observations need first be noted. The initial problem here may well be the language in which Fuller's comments are couched a language familiar from much of the debate itself, particularly that which currently engulfs international relations. That language is one of problem and solution. Attempts at 'solving' the 'structure-agency problem' may well have proved abortive; yet this is hardly surprising to those for whom the issue of structure and agency is simply not a 'problem' in the first place.

This important, if potentially rather cryptic comment, requires some unpacking. To appeal to the issue of structure and agency as a 'problem' with a potential 'solution' such that one could speak, in Fuller's terms, of progress towards a solution over time is to conflate the epistemological and the ontological. Claims as to the relative significance of structural and agential factors are founded on ontological assumptions as to the nature of a social and political reality. As such they simply cannot be adjudicated empirically. It is, presumably, this which frustrates Fuller's desire for a solution after two hundred years of reflection and debate. Consequently, structure-agency it is not so much a problem as a language by which ontological differences between contending narratives might be registered.

The closest we are likely to get, then, to 'solutions' to the agency-structure problematique are evolutionary cycles of fashion for particular ontological positions, the appeal of which may well reside in their ability to present themselves as 'solutions'. In short, there are as many different solutions to the structure-agency problem(atique) as there are authors holding different ontological assumptions. The language of structure and agency provides a convenient means of registering such ontological differences in a systematic and coherent manner. It should not be taken to imply an empirical schema for adjudicating contending ontological claims.

The epistemological fallacy: seeking empirical solutions to ontological questions

This immediately raises a closely related and no less significant point. If the issue of structure and agency is an ontological one, then, as noted above, it cannot be resolved by the weight of evidence. For, precisely the same catalogue of events can be accounted for differently within different ontological schemas. What in one account is attributed to the burden of structural or contextual factors which the immediate actors bear might, in another, be seen as the product of free choice on the part of an unconstrained, autonomous and reflexive actor. Consequently, it is imperative that we avoid claims, such as frequently appear within the existing literature, that either: (i) the author's (general) theory of the structure-agency relationship is a simple reflection of an external empirical reality (the empirical induction of ontological assumptions); or (ii) that the relationship between structure and agent varies between cases and is, as such a context-dependent empirical question. Although both statements commit an epistemological fallacy (namely, the attempt to derive an ontology from an epistemology) and although they are mutually incompatible (for if, as in the second formulation, the relationship between structure and agency varies from situation to situation there can be no general theory of the relationship between the two), both types of statement frequently occur in the same piece. Thus, despite Wendt's avowed (philosophical) realism, which would surely lead him to place ontological concerns prior to epistemological ones, he suggests (in an otherwise admirable piece co-written with Ian Shapiro),

"The differences among 'realist' models of agency and structure and among them and their individualist and holist rivals are differences about where the important causal mechanisms lie in social life. As such, we can settle them only by wrestling with the empirical merits of their claims about human agency and social structure These are in substantial part empirical questions".

Wendt and Shapiro are surely right to note that ontological differences such as those between, say, intentionalist and structuralist accounts, tend to resolve themselves into differences about where to look for and, indeed, what counts as, important causal mechanisms in the first place (i.e.: ontology proceeds epistemology). Yet they then go on to commit a cardinal conflation of epistemology and ontology, by suggesting that we can effectively make a choice between contending ontologies on the basis of what we observe empirically (i.e.: epistemology proceeds ontology). This is to have one's cake and eat it. Either we make prior ontological choices (with epistemological implications) and seek to operationalise these in concrete theoretically-informed inquiry (a rationale consistent with a realist approach) or we follow an empiricist rationale, inductively inferring our ontological preferences (if, as and when required) from a surface description of the empirical domain.

Wendt and Shapiro's conflation of ontology and epistemology is further compounded in the passage which immediately follows.

"The advocates of individualism, structuralism and structuration theory have all done a poor job of specifying the conditions under which their claims about the relationship of agency and social structure would be falsified".

Here we see appeal made to the possibility (and desirability) of an epistemological refutation of ontological propositions. A similar conflation underpins Wendt's recent prescriptive suggestion that "ontology talk is necessary, but we should also be looking for ways to translate it into propositions that might be adjudicated empirically". If only this were possible. When, as Wendt himself notes, ontological sensitivities inform what is 'seen' in the first place and, for realists, provide the key to peering through the mists of the ephemeral and the superficial to the structured reality beneath, the idea that ontological claims as to what exists can be adjudicated empirically it rendered deeply suspect. Quite simply, structuration theory, as an ontology, cannot be falsified for it makes no necessary empirical claim. It is for precisely this reason that the logical positivists reject as meaningless ontological claims such as those upon which realism and structuration theory are premised. Moreover, it is presumably for this reason that Wendt is keen to emphasise that "constructivism is not a theory of international politics".

The ontological fallacy: seeking ontological solutions to empirical questions

If we must, then, be careful not to imply (the potential for) an empirical resolution to ontological dilemmas, it is equally important that we resist the temptation to assume that our social ontologies provide ready-made solutions to empirical questions. Structuration theory cannot tell us whether the course of human history would have been different had Adolf Hitler never been born, just as a social ontology of predestination cannot tell me the final form this article will take. It is to Wendt's credit that he is rather more attuned to the dangers of this ontological fallacy than he is to the epistemological fallacy detailed above. Thus, in noting the perils of philosophical (for which read ontological) solutions to epistemological questions, he argues (again, with Ian Shapiro),

".. many proponents of neo-Marxism and structuration theory ... suggest that realism either entails their social theories or predisposes us to accept the truth of them. The notion that there is such a thing as 'realist social theory' reflects a more general point about the nature of philosophy that we wish to reject: that its task is to generate knowledge about the world independent of the practices of science".

As they go on to suggest, "realism is not a social theory [far less a theory of international politics] and as such does not entail the truth of any particular hypothesis about the causal structure of society". This is a perceptive point; quite how it might be reconciled with Wendt's prescription that we seek to translate ontological claims (such as those which underpin realism) into substantive claims which might be tested empirically is another matter. As I have elsewhere remarked, albeit in a somewhat different context, "we should be extremely wary of any suggestion that essentially interpretative disputes can simply be adjudicated by importing abstract considerations of structure and agency".

The reification of the state

If it is important to establish what is, and what is not, at stake in the agent-structure debate at the outset, it is also crucial that we consider the specificity of the debate within international relations scholarship. This is above all crucial if we are to make the link to international political economy and globalisation in particular. Rather more so than in cognate disciplines and sub-disciplines (such as sociology and political science), the structure-agency controversy in international relations tends to be couched in terms of individualism vs. structuralism or holism. The irony is, of course, that the notion of the individual deployed here is little more than a synonym for the nation-state, structure similarly little more than the system of states. This raises the thorny 'levels of analysis' problem (if problem it really is). Frequently, as Wendt notes, the levels of analysis 'problem' is confused with the question of structure and agency. If the issue is set up as one in which there are essentially three levels that of the individual, that of the state and that of the system of states and, in addition, the state is reified as a unified and rational subject qua individual, this is by no means surprising. A number of points might nonetheless here be noted.

First, given the long-time ascendancy of realist and neorealist perspectives within international relations scholarship it is not at all surprising that meta-theoretical discussions should have tended to replicate realism's ontological statism and associated reification of the state. It is important to note, however, that the methodological primacy accorded to the state and the conception of the state as a unified, rational and utility-maximising subject are in no sense entailed by considerations of structure and agency. Indeed, as I will go on to argue, the question of structure and agency may be used to provide a useful corrective to tendencies to project in an unproblematic fashion human qualities onto complex collective actors like the state.

Similarly, there is nothing inevitable about the (similarly realist) tendency to attribute a narrow means-end rationality to actors (or states as actors), nor the insistence upon an ontological and methodological individualism by which explanations are only deemed adequate if they are reduced to the level of singular actors (be they individuals or states). For, whilst it is important to ask how seemingly collective actors (such as socio-economic classes) acquire the capacity to act collectively (a question no less significant for the state itself), it would be both arbitrary and unnecessarily restrictive to assume that only individuals (and institutions like the state conceptualised in such terms) can exhibit agency.

If we are to be able to fashion any understanding of complex developmental trajectories, tendencies and counter-tendencies such as those associated with globalisation, it is imperative that we resist the realist temptation to prejudge the issue of causal significance by restricting our focus only to state or state-like actors.

Finally, if we are simply to describe, let alone explain, the frequently contested contours of complex institutional and ideational change, it is imperative that we resist any temptation to restrict our focus to one spatial scale or level of analysis. For, as we shall see, it is in the dynamic relationship between levels of analysis and spatial scales that processes such as globalisation must be sought. As elsewhere noted, if we are adequately to conceptualise globalisation "political scientists and those ostensibly concerned with domestic political and economic processes should acknowledge the international conditions of domestic political and economic dynamics". Similarly, "international relations scholars and those ostensibly concerned with international/global processes should acknowledge the domestic conditions of international/global political and economic dynamics".

The reification of theoretical labels: beware of gurus, beware of jargon

A final brief observation is perhaps also warranted at this stage. As Martin Hollis and Steve Smith note, we should be extremely wary of gurus in the structure-agency debate. More than most controversies, the existing literature is characterised by a proliferation of bon mots and bon noms. Consequently, positions are often evaluated and adjudicated not on the basis of what the author says, but what has already been attributed (fairly or otherwise) to those favourably cited. This is especially true where reference is made to the seminal contributions of Giddens, Bhaskar and, more recently, Archer, and to the terms structuration theory, scientific/critical realism and the morphogenetic approach. Frequently, what is intended by such terms and appeal to such authors is different from what is read into them. Given the indulgently inpenetrable prose of Bhaskar and the sheer complexity of the issues dealt with in rather more accessible terms by both Giddens and Archer, this should come as no great surprise.

It is for precisely this reason that, in the sections which follow, I avoid seeking justification for the position developed by reference to the usual gurus. This should not be read as a judgement of their contributions. As it happens, I have considerable sympathy for the work of all three, despite more or less serious reservations in each case. Like other positions in the debate, what follows should be assessed on its own terms, not in terms of its affinities to positions which are hardly unambiguous and which have been interpreted in a great variety of often mutually incompatible ways.

'Bringing ideas back in': Towards a strategic, relational and ideational approach

If we are to demonstrate the centrality of the realm of ideas and ideational processes to the relationship of structure and agency, then it is crucial that we begin with the simplest statement of the ontological assumptions which underpin the strategic-relational approach.

Structure and agency: a purely analytical distinction

The first of these, which places the strategic-relational approach (at least as developed here) in opposition to much of the existing literature is that the distinction between structure and agency is a purely analytical one. This assumption (for it can only ever be an assumption) renders meaningless Margaret Archer's insistence, for instance, that structure and agency reside in different temporal domains, such that the pre-existence of structure is a condition of individual action. This may seem like a purely semantic point, but it has important theoretical implications as we shall see. Interestingly, however, as Anthony King has recently noted, Archer's own position on this question seems to have shifted over time. For, in her first book, Culture and Agency, she refers to this temporal divide as purely 'analytical', whereas in Realist Social Theory it acquires the characteristics of a profound ontological dualism. Archer's view, then, is that structures pre-exist agents (or subjects); the view developed here, by contrast, is that structures can only be said to exist by virtue of their mediation of human conduct structures constitute both the medium and condition of human agency.

In this sense, then, neither agents nor structures are real, since neither has an existence in isolation from the other their existence is relational and genuinely dialectical. Whilst it may be useful analytically to differentiate between structural and agential factors, then, it is important that this analytical distinction is not reified and hardened into a rigid ontological dualism. As I have argued elsewhere, structure and agency are best seen, not so much ( la Giddens) as flip-sides of the same coin, as metals in the alloy from which the coin is forged. From our vantage-point they do not exist as themselves but through their relational interaction. Structure and agency, though analytically separable, "are in practice completely interwoven (we cannot see either metal in the alloy only the product of their fusion)".

How then is it that, from the vantage-point of a particular actor, the world does indeed appear to be pre-structured, such that structure and agent inhabit different temporal domains? The problem here, as Anthony King perceptively notes, is a perspectival one. From the vantage-point of a singular actor, social structures do indeed appear external and temporally independent. Yet, a subtle change in vantage-point alters this. As King explains,

"the key error which Archer makes in her derivation of social structure is to draw the sociological conclusion of the existence of a social structure from the perspective of a single individual if she had de-centred her perspective to see that the constraint which I face is other individuals and no less serious for that just as I form some of the social conditions which mutually constrain these others, she would not have fallen into ontological dualism".

My only difference with King's view as expressed here is with his rather condescending suggestion that Archer has made a simple conceptual 'error' in viewing the structure-agency relationship from this particular vantage-point. Whilst it strikes me as more useful to regard structures an agents as abstractions from a reality in which they are complexly and necessarily interlinked, this is, in the end, an ontological choice not a matter of empirical record.

The material and the ideational: a purely analytical distinction

If it is controversial to insist that the distinction between structure and agency is an analytical rather than a real one, then at least equally controversial is the second ontological premise of the strategic-relational approach (as developed here). That is that the distinction between the material and the ideational is also purely analytical. Just as structures and agents do not exist in isolation, so too the material and the ideational are complexly interwoven and mutually interdependent. What is likely to make this ontological insistence yet more controversial is the popular idea that the material itself circumscribes the realm of the real. Whilst this may well be a convenient assumption to make for those keen to retain their access to a positivist methodology, it is somewhat difficult to reconcile with our own experiences and is, above all, limiting. For, whether we come to reject them at some subsequent date, the ideas we hold about our environment (about what is feasible, possible and desirable, for instance) have substantive effects. Moreover, they do not do so independently of that environment itself both the effects themselves and the ideas we fashion in the first place are mediated by the context in which we find ourselves. Consequently, as with the question of structure and agency, whilst it may be useful to distinguish analytically between the material and the ideational, it is important that an analytical strategy does not crystallise into an ontological dualism.

Structure, strategy and agency

Having established some fairly fundamental ontological premises we can move to build up a picture of the relationship between structure and agency. Actors are conceptualised as conscious, reflexive and strategic. They are, broadly, intentional in the sense that they may act purposively in the attempt to realise their intentions and preferences. However, they may also act intuitively and/or out of habit. Nonetheless, even when acting routinely they are assumed to be able to render explicit their intentions and their motivations. Actors are assumed to monitor the immediate consequences of their actions, whether intuitively or more deliberately, and to be capable of monitoring the longer-term consequences of their actions. Though actors are conceptualised as intentional and strategic, their preferences are not assumed to be fixed, nor to be determined by the material circumstances in which they find themselves. Different actors in similar material circumstances (exposed, perhaps to different ideational influences) will construct their interests and preferences differently, just as the same actors will review, revise and reform their perceived interests and preferences over time (as material circumstances and ideational influences change). Accordingly, in monitoring the consequences (both intended and unintended) of their actions, actors may come to modify, revise or reject their chosen means to realise their intentions as, indeed, they may also come to modify, revise or reject their original intentions and the conception of interest upon which they were predicated.

No methodological or ontological individualism is assumed. Consequently, actors are neither presumed to be individuals nor (as in much realist and neorealist scholarship) to behave 'as if' individuals. Nonetheless, where collective actors are identified or where agency is attributed to complex and institutionally-differentiated bodies (such as states or trans-national corporations), it is important that a series of subsidiary questions can be answered. These concern the mechanisms by which, say, a collective actor comes to acquire and, over time, to reproduce and potentially to transform its collective identity, and the processes by which (and the conditions under which), say, states come to acquire the characteristics of agency.

If this may seem to present an unduly agency-centred perspective, then it is important that any such impression is quickly tempered by a consideration of the context in which agents find themselves a context about which they are assumed to know much. The key to the relationship between structure and agency within the strategic-relational approach, is, unremarkably, the concept of strategy. Actors, as noted above, are presumed to be strategic to be capable of devising and revising means to realise their intentions. This immediately implies a relationship, and a dynamic relationship at that, between the actor (individual or collective) and the context in which she finds herself. For, to act strategically is to project the likely consequences of different courses of action and, in turn, to judge the contours of the terrain. It is, in short, to orient potential courses of action to perceptions of the relevant strategic context and to use such an exercise as a means to select the particular course of action to be pursued. On such an understanding, the ability to formulate strategy (whether explicitly recognised as such or not) is the very condition of action.

>At this point it is important to deal with a potential objection. For, it might be suggested, there is a certain danger here of so closely eliding strategy and agency as to imply that all action is the product of overt and explicit strategic calculation (just as rational choice theorists attribute an instrumental utility-maximising means-end rationality to all actors). The argument being made here is, in fact, somewhat different. What I am suggesting is that all action contains at least a residual strategic moment though this need not be rendered conscious. This makes it important to differentiate clearly between intuitively and explicitly strategic action.

  1. Intuitive, routine or habitual strategies and practices are based upon perceptions (accurate or otherwise) of the strategic context and the likely consequences of specific actions. As such they can be regarded as strategic insofar as such practices are oriented towards the context in which they occur. However intuitive, the act of crossing the road so as to avoid oncoming cars and pedestrians contains an inherently strategic moment. Although such an understanding and lay knowledge can be rendered explicit, invariably it remains unarticulated and unchallenged. Note, however, how effectively a close shave on a zebra crossing brings to the surface previously unquestioned strategic calculations. Insofar as the assumptions which implicitly inform such routines, habits, rituals and other forms of unreflexive action can be rendered explicit, these practices contain a significant strategic component. Such strategy is manifest in 'practical consciousness'.

  2. Explicitly strategic action also relies upon perceptions of the strategic context and the configuration of constraints and opportunities that it provides. Yet here such calculations and attempts to map the contours of the context are rendered explicit and are subjected to interrogation and contestation (particularly in the formulation and reformulation of collective strategies) in an overt and conscious attempt to identify options most likely to realise intentions and objectives (whether individual or collective).

These are, of course, ideal types. Any specific action is likely to combine both intuitive and explicit strategic aspects, though to differing degrees. Even the most explicit strategic calculation is likely to be infused with intuitive assumptions at the level of 'practical consciousness'. In addition to an emphasis upon the explicitly identified strategic calculations of reflexive subjects, then, a strategic-relational methodology must also interrogate the frequently unarticulated and/or uncontested assumptions which define the parameters of the strategic imaginary.

Two closely related features of this strategic imaginary are the spatial and temporal horizons of action. A strategy, say, for securing economic growth in the context of a globalising international economic environment over an extended period of time will clearly differ markedly from one designed to produce short-term economic benefit in line with electoral cycles. Indeed, short-termism may well compromise the basis upon which sustained economic growth might be achieved. Thus the time horizons to which strategies are oriented will itself have important strategic implications.

If strategies are necessarily temporal and temporalising, they are also spatial and spatialising. The question at which spatial scale economic development strategies should best be pursued in the context of a global political economic environment is inherently strategic. Moreover, its answer is contingent upon the various political, economic, and social specificities of a particular context at a particular moment in time. This latter comment reveals that the temporal and the spatial are not separable. In the present example, the strategic choice of time horizon will itself in part dictate the appropriate spatial scale at which economic development initiatives should be sought. Conversely, the choice of spatial scale will in part determine the time horizons within which economic growth can be anticipated.

Strategic selectivity

This brings us to what is perhaps the definitive concept of the strategic-relational approach strategic selectivity. As the above discussion serves to indicate, structures are selective of strategy in the sense that, given a specific context, only certain courses of strategic action are likely to see actors realise their intentions. Social, political and economic contexts are densely structured and highly contoured. As such they present an unevenly distributed configuration of opportunity and constraint to actors. They are, in short, strategically-selective, for whilst they may well facilitate the ability of resource- and knowledge-rich actors to further their strategic interests, they are equally likely to present significant obstacles to the realisation of the strategic intentions of those not similarly endowed.

As the above discussion indicates, patterns of strategic selectivity and hence the complex configuration of constraints and opportunities that a context presents to a strategic actor are temporally and spatially specific. The strategic selectivity imposed by the financial markets looks rather different for an investor seeking an immediate return on her investment that it does for one projecting a similar return over a rather longer period of time. Similarly, the conditions of economic success for a small locally-based firm in a declining national economy are likely to prove fundamentally different from those of a trans-national corporation more free to relocate its productive capacity in line with changing labour costs and market share.

Discursive selectivity: the place for ideas

Thus far we have tended to assume that strategic actors have a fairly direct and unmediated access to the contours of the terrain they inhabit, such that they can effectively 'read off' the likely consequences of their action from their knowledge of the context in which they find themselves. This, however, is a most dubious premise, akin to the perfect information assumption much beloved of neoclassical economists and many rational choice theorists. Though convenient and parsimonious, it is unrealistic.

Actors are reflexive and strategic and they orient themselves and their strategies towards the environment within which their strategic intentions must be realised. Yet they are by no means blessed with perfect information of that context. At best their knowledge of the terrain and its strategic selectivity is partial; at worst it is demonstrably false.

Given, however, that actors are reflexive, routinely monitoring the consequences of their action, one might expect their perceptions of the context to evolve over time if not, perhaps, to a situation approximating complete information, then at least to one of relatively reliable reconnaissance. Yet a moment's reflection reveals that this too is unlikely. For whilst actors might well acquire cumulative knowledge over time in an environment that is essentially unchanging, this is rarely, if ever, the case in situations characterised by a density of existing institutions and practices and a proliferation of strategic actors. Invariably it is such contexts we are interested in analysing. Moreover, even were we to assume complete information of a current context (based, presumably, on extensive reconnaissance of prior strategic interventions), this would be insufficient to predict the likely consequences (even over the short-term) of a particular course of strategic action. For the effects of a specific and given intervention are not merely determined by the strategic selectivity of the context at the moment at which the action occurs. A range of additional and from the vantage-point of the actor about to make an intervention contingent and unpredictable factors are also relevant. These include strategic responses made to the intervention itself as well as the quite independent actions of others. In principle, this gives social and political interaction an inherently indeterminant, unpredictable and contingent quality. When the incomplete information of any given actor is also considered, it is hardly surprising that strategic action almost always includes unintended consequences.

Nonetheless, whilst all contexts exhibit this complex, contingent and unpredictable quality, some are clearly more contingent than others. Interestingly, arguments pointing to the globalisation of social, political and economic relations often identify the growing interconnectedness between once separate contexts. Insofar as such a claim is warranted, this suggests a tendency for ever escalating complexity, contingency and unpredictability.

In a world which exhibits such qualities, it should come as no surprise that actors routinely rely upon cognitive short-cuts in the form of more or less conventional mappings of the terrain in which they find themselves. Thus, for instance, policy makers typically conceptualise the policy-making environment through the lens of a particular policy paradigm such as Keynesian or monetarist economics. Once again, access to the context itself is discursively mediated. How actors behave the strategies they consider in the first place, the strategies they deploy in the final instance and the policies they formulate reflect their understanding of the context in which they find themselves. Moreover, that understanding may eliminate a whole range of realistic alternatives and may, in fact, prove over time to be a systematic misrepresentation of the context in question. Nonetheless, for particular ideas, narratives and paradigms to continue to provide cognitive templates through which actors interpret the world, they must retain a certain resonance with those actors' direct and mediated experiences. In this sense the discursive or ideational is only ever relatively autonomous of the material.

What the above discussion hopefully demonstrates, is the centrality of ideas to an understanding of the relationship between agent and structure, conduct and context. It also suggests the power of those able to provide the cognitive filters, such as policy paradigms, through which actors interpret the strategic environment. In sum, in the same way that a given context is strategically-selective selecting for, but never determining, certain strategies over others it is also discursively-selective selecting for, but never determining, the discourses through which it might be appropriated.

Power and Levels of Structuration

Thus far we have dealt with the complex interplay of structure and agency as if all aspects of context were potentially amenable to transformation by all actors. This is to exclude from the equation the crucial concept of power, understood here as the ability to shape the contexts within which others formulate strategy. We must then differentiate between what might be termed levels of structuration with higher levels of structuration relating to structures over which given strategic actors (over a particular time horizon) can be said to have minimal impact. Such structural constraints (and the opportunities they imply) are shaped and reshaped by the actions of the (more) powerful (whether intentionally or unwittingly), setting the (external) context for the (relatively) powerless. From the vantage point of such actors, these might be considered non-accessible levels of structure/structuration. They condition the possible range of strategies and actions within a specified social and political context but are not immediately accessible to transformation by the agents that they embed within such a context.

It should immediately be emphasised, however, that this attribution of power to particular actors or organisations is conditional upon time horizon. Structures which might appear non-accessible to particular actors and organisations over a particular time horizon may well become subject to strategic transformation over a longer period. For, by identifying a collective interest, actors may overcome their powerlessness by pooling their resources and thereby constituting themselves as strategic actors at higher levels of structuration. Consequently, the attribution of power is dependent upon the context being interrogated, the form of that interrogation, the vantage point from which the context is viewed, and the time frame considered.

Over a given time-frame, the actions of the most powerful (A1) take place within a strategically-selective context which is 'always already there' (S1) and which favours certain strategies over others a context which is itself the product of previous displays of agency. This imposes strategic constraints upon even the most powerful of actors (collective or individual). Nonetheless, the strategic conduct of such actors has the effect of transforming (however partially) the contours of that strategic context. This action setting (S1) is thus dynamic and constantly evolving. Its strategic selectivity is not, however, purely the product of the effects (intended and unintended) of the actions of the powerful. It is also shaped by the actions of the (relatively) powerless and the contingent effects of their articulation. Thus, crucial aspects of the strategic selectivity of the terrain inhabited by the powerful include the likely reactions of the powerless to particular strategies, and their ability to mobilise strategic resources to empower themselves.

Nonetheless, from the vantage-point of the less powerful (A2, A3), the structures, say, of the global political economy appear as external constraints over which they can exercise minimal, if any, strategic or intentional influence. The powerless do not (by virtue of their powerlessness) exist as strategic actors able to make a decisive intervention on this terrain (at least within the given time horizon). The strategic context inhabited by less powerful strategic actors (like A3) thus comprises accessible levels of structure amenable to strategic action (S3), and non-accessible levels of structure beyond their immediate strategic reach (S2, S1). Power resides in the capacity (whether intentionally exercised or not) to transform aspects of the context in which other less powerful groups and individuals are constrained to formulate their strategies.

Globalisation: A process without a subject?

This brings us eventually to the question of globalisation. In the previous section I sought to develop a conceptual schema capable of interrogating the question of globalisation through a strategic-relational approach to structure and agency. It is perhaps important to emphasise before proceeding that what follows is an attempt to trace through the implications of a strategic-relational ontology for an analysis of globalisation. It is not an attempt to argue for the objective superiority of the resulting account nor for the ontology upon which it is predicated. As I have been at pains to demonstrate, different ontological assumptions entail different standards of explanatory adequacy and the choice between ontological assumptions is not one which is easily subjected to empirical scrutiny. Consequently, the case for an account of globalisation which seeks to place at centre stage the dynamic relationship between political actors and the contexts in which they find themselves must, in the end, be a normative one. It can be relatively simply stated. An emphasis upon the negotiated and, at least in part, discursively constituted nature of political and economic constraints places greater emphasis upon the capacity of political actors to shape the context in which they find themselves. In so doing, it offers the prospect of holding such actors accountable for their conduct in a way which is simply not the case for more structuralist accounts which present globalisation as an inexorable and non-negotiable external economic constraint.

Before proceeding to this attempt to restore notions of political accountability to the process(es) of globalisation, it is important also to note the problem of ontological consistency. In the context of discussions of globalisation this is, perhaps, particularly significant. There is nothing especially original about an account of globalisation linked explicitly to an ontology which emphasises the dynamic relationship between structure and agency. Yet, all too frequently the substantive claims made about globalisation in the name of such ontologies are profoundly structuralist and hence considerably at odds with their claimed ontological premises. Cerny's suggestive account of the development of the 'competition state' is a particularly prominent case in point. For whilst Cerny claims that his work (including that on globalisation) is informed by a 'structurationist' ontology, his analysis of the one can only presume inexorable and necessary transition from the 'welfare state' of the postwar period to the 'competition state' of today is deeply deterministic. It turns a well-observed empirical generality into a structuralist logic of economic determinism. In so doing, it increasingly seems to belie any notion of the negotiated nature of economic constraint. Similar observations might be made of Jessop's somewhat more complex, sophisticated and carefully qualified rendition of the transition from the Keynesian welfare state to the Schumpeterian workfare state. Yet, qualifications accepting, how is the non-negotiable and deterministic (even functionalist) logic of economically-driven political development to be squared with either author's attempts to develop a dynamic understanding of the relationship between context and conduct? In fact, it may not be impossible to reconcile what is, in the end, an empirical claim that globalisation entails a very significant loss of political autonomy at the national level with a strategic-relational or structurationist ontology. Nonetheless, the point is that neither author has made any attempt so to do. Consequently, the accounts they have offered seem to mutate from relational ontological assumptions to deterministic, even functionalist, substantive narratives.

If we are to develop a strategic-relational account of globalisation, what is first required is a theoretical apparatus capable of linking the relationships between agents and their structured contexts at a variety of spatial scales. This is developed schematically in Figure 3, in which different spatial scales (chosen on the basis of their heuristic utility) are treated as levels of structuration in the broader process of globalisation/counter-globalisation. The dynamics of the global political economy are captured in the complex and unpredictable interaction between these various levels of structuration.

Beginning at the level of sub-national processes and economic dynamics we can identify a range of strategic actors (for instance, businesses, governmental and extra-governmental actors) formulating a variety of strategies (be they specific production-distribution regimes or economic growth strategies). If such strategies are to prove successful in their own terms, they must be oriented not only to the configuration of opportunity and constraint provided by the immediate sub-national economic and political environment, but also to the broader national, regional and, indeed, global context. These higher levels of structuration are non-accessible to direct intervention by the sub-national actors themselves but nonetheless have a crucial bearing upon the strategic selectivity of the context in which they seek to realise their strategic intentions. Such actors can be assumed to have no privileged knowledge of the contours of the terrain they inhabit and, on the basis of partial information, may well come to misperceive the complex configuration of opportunity and constraint they face.

At higher levels of structuration (say, the national level) we can identify a range of potentially more influential actors, the effects of whose strategies may significantly alter the context in which, say, sub-national actors operate. They too, however, must adapt their strategies to the environment in which they find themselves. Once again, that environment contains significant external constraints passed down to them from higher levels of structuration (say, those operating at regional and/or global spatial scales). These aspects of context are, consequently, not amenable to direct intervention by the actors themselves, becoming effectively external constraints. As at lower spatial scales, these actors have no privileged access to the contours of the strategic terrain they inhabit and will tend to adapt existing understandings of the operation of the system to their own situation and experience as a guide to formulating strategy in a changing environment.

At the very highest spatial scales (for instance, the global) something very interesting and significant occurs we effectively run out of strategic actors capable of making decisive interventions at the level of the system itself. This raises a seeming paradox. For if no strategic global actors can (as yet) be identified, how it is that we can speak meaningfully of a process of (global) structuration at all, let alone restore actors to the process of globalisation? The answer is to be found in unintended consequences. For whilst there is a deficit of actors capable of purposively refashioning the global political economy itself, there is a whole host of actors capable of contributing unintentionally to a series of global processes, tendencies and counter-tendencies. Perhaps the classic example here is global environmental degradation. A complex array of sub-national, national, multi-national and trans-national actors contribute, through the more or less unintended consequences of their actions, to a genuinely global structuration process one of environmental degradation. Clearly the potentially global significance of individual acts of ecological despoliation hardly empowers the actors involved, but it does nonetheless contribute to a global process. Moreover, in the absence of genuinely global governance mechanisms or alternative strategies for ensuring concerted global solutions to such global 'bads', little is likely to be done to counter such a tendency. This is perhaps the first sense in which we can speak of globalisation as a process without a subject.

This observation has important implications for the attempt to restore active (and hence potentially accountable) subjects to a process of globalisation widely appealed to as a process without a subject. It suggests, in particular, the need to differentiate between: (i) genuine processes of globalisation in which, in the absence of genuinely global actors, developmental tendencies and counter-tendencies are generated and sustained through the unintended consequences of actions pursued consciously at lower spatial scales; and (ii) processes pursued consciously by strategic actors at the national or regional level which are (falsely) appealed to as processes of globalisation. In this second case, restoring subjects to the process of globalisation may entail challenging the appeal made by politicians and commentators alike to globalisation as the proximate cause of political decisions pursued for strategic advantage at the national or regional level.

It is interesting at this point to note, however, that although globalisation is frequently and, as we shall see, problematically referred to as a process without a subject, it is rarely in the sense identified above. In the final and concluding section of this paper we examine the popular invocation of globalisation in such terms and turn to the broader question of restoring active subjects to the process of globalisation.

Bringing the subject back in: Structure, agency and globalisation

Insofar as international relations theorists tend to speak of a structure-agency 'problem', it is because they think that they have a 'solution'. That solution invariably involves some claim to have identified a middle way (a Giddensian third way perhaps?) between the twin extremes of structuralism and intentionalism or holism and individualism. Such middle ways have the obvious appeal of acknowledging the dynamic interaction of structure and agency, of context and conduct. Consequently, they tend to place the emphasis not upon the explication of deterministic structural logics or the identification of hegemonic intentional actors but upon the elucidation of processes, in which structure and agent and intimately interwoven, over time. Given that globalisation is a process term, one might be forgiven for thinking that the analysis of globalisation naturally lends itself to a subtle and complex rendition of the structure-agency relationship to elucidate the causal mechanisms involved. Sadly, nothing could be further from the truth.

This is because, whilst globalisation may masquerade as a process term in both the popular and academic vernacular, it is a 'process without a subject'. It is, in short, a process to which no actors are linked, a process which rapidly becomes a deterministic logic of structural inevitability the closer one looks. Accordingly, the term globalisation as used in most popular and academic debate is an obfuscation, and a potentially dangerous one at that. For it tends to conjure a sense of inexorability, inevitability and immutability, mapping a path to an end-state (a condition of pure globalisation) never fully realised yet always in the process of being realised. This represents a dangerous conflation of process and teleology which can only serve to hide the complex causal processes which generate the evidence frequently cited in support of the globalisation thesis.

To point to the dangers of appeals to globalisation as a causal process is not to insist that globalisation is a figment of the imagination. It is merely to suggest the need for considerable caution in the use of the term if we are not to further mystify phenomena which might genuinely be regarded as evidence of globalisation. The challenge posed by this article (not, it should be noted, an entirely novel one) is to build upon the foundations laid by the critique of earlier, more cavalier, appeals to the notion of globalisation, to further unpack and demystify this 'process without a subject'. This in turn suggests that we should view globalisation not so much as a process or end-state, but, at best, as a tendency to which counter-tendencies may be mobilised. Once viewed as merely tendential, the challenge is to reveal the dynamic and contingent articulation of processes in certain spatial contexts at certain moments in time to yield effects which might be understood as evidence of globalisation. Such scrutiny invariably reveals the causal significance of processes operating at spatial scales below the global and for which strategic political actors might be held accountable. Our aim, then, must be two-fold: (i) to explain the complex causation (often by appeal to logics of unintended consequence) of those genuine if rare processes of globalisation which escape the intentions of the actors involved; and (ii) to demonstrate (where this is not the case) that key actors retain considerable control over what they choose to appeal to as a globalisation process which they claim to be powerless against.

If we are to do this, it is imperative that we reverse the conventional direction of causality appealed to in the academic literature as, indeed, in the popular discourse of globalisation.

We must ask not what globalisation might explain, but how we might account for the phenomena widely identified as evidence of globalisation. If we are to resist and reject the deterministic appeal to a process without a subject we must excise all reference to globalisation as an explanatory (or independent) variable. Within such a schema, the term globalisation becomes little more than a convenient short-hand for a confluence of processes which might together be seen as constitutive of any observed (and, presumably, contingent) globalisation tendency. The existing literature has, to date, given far too limited attention to such genuine (causal) processes, like financial liberalisation, to which actors might be linked directly. Financial liberalisation is, perhaps, a good example. For many accounts of globalisation's seeming 'logic of no alternative' or of neoliberal convergence, in the end, rely on claims not of globalisation per se but of financial liberalisation and consequent heightened capital mobility. In such cases, appeal to the term globalisation is quite simply an obfuscation if the causal agent is the (quite conscious political) decision to engage in a process of financial liberalisation, why not call it as it is?

Recast in this way, there is no need to make essentialising and reifying assumptions about the effects, consequences, or even the very existence, of globalisation. For, in so far as globalisation can be identified, it is understood as a tendency the contingent outcome of a confluence of specific processes that are themselves likely to be limited in space and time. Globalised outcomes and effects might then be the product of very different, indeed entirely independent, mechanisms and processes of causation (financial liberalisation, European integration and policy transfer to identify merely three) that can only be obscured by appeal to a generic (and causal) logic of globalisation. Whilst problematising and interrogating the processes which underpin globalising tendencies, then, it is important to resist the temptation to appeal to globalisation itself as a causal factor or process working, apparently independently of the actions, intentions and motivations of real subjects. It is precisely this appeal to causal processes without subjects that summons the logic of necessity and inevitability so often associated with the notion of globalisation.

If we are then to demystify globalisation, we must ensure that in making what we think are causal arguments, we can identify the actors involved, thus giving due attention to the 'structuration' of globalising tendencies whilst rejecting structuralist or functionalist 'logics' operating over the heads or independently of social subjects.

It is only by paying careful attention to the problem of structure and agency in this way, dismissing accounts which privilege either structure or (far less frequently in analyses of globalisation) agency in the determination of outcomes, that the notion of globalisation might be used to open up and not merely to obfuscate the analysis of social, political and economic change.

It is one thing to restore subjects to the process of globalisation within academic debate. Yet, the use of the term globalisation is not merely confined to the common room. It has become a key referent of contemporary political discourse and, increasingly, a lens through which policy-makers view the context in which they find themselves. If, as I have argued consistently in this article, strategic actors have no privileged vantage point from which to view the strategic selectivity of their environment and, as the vast majority of commentators would surely concede, one of the principal discourses through which that environment comes to be understood is that of globalisation, then it is imperative that we consider the causal significance of ideas about globalisation for contemporary political and economic dynamics. As noted elsewhere, "in a context in which diagnoses of globalisation and the various imperatives it is held to summon proliferate, it is remarkable ... that so little attention has been paid to the independent causal and constitutive role of ideas about globalisation in the generation of the effects attributed to globalisation itself".

For, despite a growing literature which seeks to demystify the often exaggerated and distorted claims made by the business school 'hyperglobalisation' thesis, it continues to prove highly influential in elite policy circles. Consequently, as even the most cursory analysis of contemporary European popular and political discourse on the subject can hardly fail to attest, the ideas about globalisation which animate public policy making are frequently based on at best a casual and highly selective appeal to the empirical evidence. This has important implications when it is considered that governments acting in accordance with the hyperglobalisation thesis may well serve to summon precisely the consequences the thesis would predict. Thus, for instance, the hyperglobalisation thesis may have it that, in a (globalised) context characterised by the heightened mobility of capital, states simply cannot afford but to reduce the level of corporate taxation. Any failure to do so, it is argued, will result in a punitive depreciation in net revenue as capital utilises its new found exit option. The irony of such a thesis is that if governments believe it to be true, they will act in a manner consistent with its predictions, thereby contributing to an aggregate depreciation in corporate taxation whether it is true or not. This is but one example. What it, and others like it, suggest is that the discourse of globalisation may play a crucial independent role in the generation of the effects invariably attributed to globalisation and invariably held to indicate its logic of inevitability. This in turn suggests that not only must we give rather greater attention to the discursive mediation of political and economic change, but that we must retain the capacity to expose to critical public scrutiny the dominant ideas which pass in the name of globalisation. It is above all crucial that we differentiate clearly between the effects of globalisation itself and the effects of having internalised popular constructions of globalisation. All too frequently the latter is mistaken for the former.

One brief example will perhaps suffice. Consider the following statement, familiar from both the academic literature and the pronouncements of politicians on the subject: "globalisation places pressures on western states to roll back their welfare provision". Statements such as this imply a loosely-articulated explanation for welfare retrenchment along the lines, "globalisation causes (or, at least, necessitates) welfare retrenchment". Here, as is so often the case, globalisation is invoked as a process without a subject; no agent is identified. This, it need hardly be noted, is highly convenient for politicians wishing to legitimate otherwise unpalatable social and economic reforms. Yet if we seek to restore active subjects to this hypothesised process, its logic of inevitability and indeterminacy is rapidly tempered. Better, then, is the following: "the ability of foreign investors to move capital and assets rapidly from one national context to another undermines the state's capacity to raise revenue to fund the welfare state through corporate taxation".

Such a statement has the clear benefit of identifying a series of agents with the capacity to act. Yet there is still no direct attribution of causal agency to identifiable subjects. Moving further to restore actors to this process without a subject, we might suggest a second modification: "the perception on the part of many western governments that investors are mobile and will exit high taxation environments has driven a process of corporate tax cutting, thereby undermining the revenue basis of the welfare state". This is, once again, an improvement. We have identified a rather different set of potential actors rather closer to decisions relating to welfare expenditure and we have introduced their perceptions into the equation. Yet we have still not directly attributed welfare reform to identifiable subjects in a genuinely causal explanation. One final step fully restores agency to the (now considerably weakened) relationship between globalisation and welfare retrenchment: "government X, acting on its belief that (mobile) investors will leave high taxation environments for low taxation environments, has reduced the rate of corporate tax, with consequent effects for the revenue basis of the welfare state". This is what is meant by 'bringing the subject back in' to the logics and illogics of globalisation.

Again, this is but one example. Yet, as an example it suggests once again the centrality of ideas to the relationship between structure and agency and, more specifically, to the complex causation of the processes invariably attributed to globalisation.

Finally, and in a similar vein, it is important that we acknowledge the strategic use made of the rhetoric of globalisation. For, as a process without a subject, seeming to operate above the heads of elected officials it provides, or is capable of providing, a most convenient scapegoat for the imposition of unpopular and unpalatable measures. By restoring active and strategic subjects to the process of globalisation we can not only contribute to the demystification of this process without a subject, we can also contribute to the repoliticisation of political and economic debate.

Copyright © 2001 Colin Hay

The Web Site for Critical Realism The WSCR Archive