Abstract
While generally accepted as an interpretive theory, Bull’s emblematic text demonstrates strong structural characteristics. Subsequent attributions move between the interpretive or ‘reflexive’ and the institutional and structural. Recently, however, the idea has come forward that English School theory is, and maybe have been from the beginning, a form of structuration theory, a theory in which structures are not quite the hard determinants generally understood in structural theories, and interpreting agents are not quite so free to interpret structures in any tradition that seems appropriate to a matter at hand.
While it is quite possible to read English School (ES) theory as an interpretive exercise, it is not the only reading available, particularly if we consider The Anarchical Society (TAS) as its emblematic text. When TAS was ‘received’ in America, it was received as a structural, not as an interpretive argument (see Morse’s, 1977, review in Foreign Affairs). Bull’s taxonomy of international system, international society and world society, with their ‘systemic effects’, is generally presented as the heart of the ES theory, and it is generally classed as system-level in the European theoretical glosses (see, for example, Lechner, 2017). Barry Buzan in his 1993 article formally blessed the structural approach, as opposed to Suganami’s case for Verstehen (see below), by dividing the ES theoretic corpus into Gemeinschaft and Gesellschaft traditions, deriving a gemeinschaft interpretivism from the approach of Martin Wight and a gesellschaft structuralism from the legacy of Hedley Bull. His purpose was to ‘improve’ American structural realism via an injection of a European sociological understanding of structures, and American regime theory via an understanding of the interpretivist nature of regimes, but the effect was to ensconce two traditions of ES theorising, allowing those working within the ES tradition to variously identify.
Opinion by the main ES theorists aside, there are other reasons for doubting that one can simply choose interpretivism not least the central proposition of the ES theory – that anarchy is the defining characteristic of the international order, and moreover, that it has consequences and sets limits to aspirations. In other words, anarchy as understood by Bull, Manning, the British Committee (BC) and its heirs is a structural concept. One can of course choose to understand anarchy in an interpretive fashion, as a social construct (e.g. Donnelly, 2015), but that does not in the least challenge a central insight of the ES, that anarchy causes things to be a certain way and not another, not to mention the ‘social fact’ that state sovereignty is an embedded social institution with real-world consequences. To ignore or deny this simply eviscerates the ES theory and turns it into a variant of Mark Bevir’s radical anti-structuralism.
If one cannot simply choose interpretivism, does that mean that one is stuck with the irresolvable tensions of a theory with different and contradictory historical roots? There has been a strong temptation to do just that – to welcome the tensions in the theory as somehow analytically creative (Buzan, 2014: 81; Zhang, 2016: 104). Such an inclination has, however, left ES theory conceptually un-rooted. The other way is to interrogate more closely what the ‘classical school’ of ES theorists might have meant by ‘structures’ (a word they little used), what they meant by ‘institutions’ and how they understood agency. Recently, the idea has come forward that ES theory is, and maybe have been from the beginning, a form of structuration theory, a theory in which structures are not quite the hard determinants generally understood in structural theories, and interpreting agents are not quite so free to interpret structures in any tradition that seems appropriate to a matter at hand.
This interrogation will proceed in three parts. First, I will begin with interpretivism and examine how various ES theorists have actually understood interpretivism. Different ES theorists have had different ideas of what constituted interpretation and some have used the term or terms associated with an interpretivist approach to signal nothing more than an anti-positivist agenda. Second, I will take on the more strenuous task of understanding what the classical theorists might have meant by structures or as they termed them ‘institutions’. The classical texts are tantalisingly illusive concerning this central concept; for decades, it remained largely unexamined, and the question of the ontology of ‘institutions’ almost deliberately ignored. The third part will look more closely at the general understanding of institutions as represented in the empirical literature, to see if it provides a clearer guide. A final section will explain co-constitution and why the ES theory might be best understood as a theory in which agents and structures co-constitute.
The interpretivist side
The interpretivist possibilities have been explored by Hall and Bevir above. In the Introduction, they call attention to the first clarion call of British international relations (IR) against American ‘scientism’. This was Bull’s (1966) ‘International Theory: The Case for a Classical Approach’, written after a semester in America absorbing the new ‘scientific’ literature on IR. In it, he castigated science in IR as remote from its subject, as unable to deal with central questions of political order and as unlikely to realise its foresworn goals, and he defended ‘the approach to theorising that derives from philosophy, history, and law’. ‘Philosophy, history and law’, taken together, point to a reflectivist, interpretivist approach, particularly when contrasted with ‘science’. Bull appeared to argue that interpretivism was a British, and perhaps even a European preference, as against an American positivist agenda.
It is important to note, however, that Bull was not representing his own views in the famous article (on more below), or even the activities of the British IR community at large, but what he took to be the main thrust of discussions as they were proceeding in the BC on the Theory of International Relations. In 1966, when the pivotal Diplomatic Investigations was published, the BC was already in the seventh year of discussions, under the guidance of the historian Herbert Butterfield, in which discussions Martin Wight was given extraordinary prominence and rendered considerable deference, not least because Wight had the only claim to being a theorist in the group and the only one who had given sustained thought to the question of an IR theory (in a series of lectures, then in its 10th year, titled the ‘Theory of International Relations’). In the context of the Committee, he was moving from ‘traditions’, that is, groups of thinkers, to focus on ‘ideas’; that is, categories of thought or ‘unit ideas’ as Hall (2006: 150–155) prefers, and he was close to developing the idea that ideational structures formed the basis of international orders. Wight was also the most determined enemy of ‘science’ in the group. Bull was representing not merely the general tenor of the Committee, but the particular influence in it of Martin Wight and the direction in analysis Wight was taking.
Behind Wight lay the formidable figure of Charles Manning who ruled the Department of International Relations at the London School of Economics (LSE) for 37 years. Manning arguably invented the concept of international society as a set of social practices 1 and he certainly gave it its ideational colouring. Aalberts (2010) argues, correctly in my view, that he viewed international society through the lenses of linguistic philosophy; that is, as a set of linguistic constructs or ‘games’. Manning considered the proper disciplinary foundations of the subject to be ‘sociology’, by which he would appear to have meant society as a set of social constructions in the contemporary sense. (It is worth noting, however, that Dunne, 1995, disagrees; see below). New members of the department were expected to follow his lead, and he preferred them not to have been trained up in diplomatic history or any cognate subject such as politics. He hired Wight, a medievalist, to teach a course on International Institutions and Bull, a philosopher and a logical positivist, to follow in his stead. 2 Wight’s and eventually Bull’s use of the term ‘international society’ came directly from Manning.
It was Wight, pace Manning, who first formalised the idea of the ‘institutions of international society’ – what would come to be understood as the heart of the ES theory. Hall (2006) has pointed out that Wight had worked with Margery Perham on colonial constitutions, and was exposed to British imperial thinking as it was developing within the imperial reform movement, both of which directed attention to institution-building. Soon after Bull arrived at the LSE, Wight gave a departmental seminar signing off on this first set of lectures, on International Institutions, whose preparation must have been a fairly sorry task given the early days of the phenomenon (Wight was reduced to beginning with the fifteenth-century Conciliar Movement). In the signing-off, he used the term ‘pseudo-institutions’ (in other contexts ‘epiphenomenal’; Hall, 2006: 123) to refer to international organisations. As opposed to organisations like the United Nations (UN), he named the true institutions of the international order as alliances, diplomacy and war (Bull, 1977a: 6). In his essay on ‘Western Values’ they became the diplomatic system, the balance of power, international law and ‘social and technical interdependence and the functional international organizations established latterly to regulate it’ (Wight, 1966b: 96–97). Bull’s famous ‘five’ were not in fact Wight’s creation; they had all appeared in the German historian A.M.L. Heeren’s account of the European system as it had met and contained the French Revolution, 3 an account of which the Committee made much (Vigezzi, 2005: 5–6). However, Bull generally referred his own usage, and his incorporation of them into his account of order in IR, to Wight. (Bull had received no training in IR much less diplomatic studies before he arrived at the LSE, and he adopted many of Wight’s categories to correct the deficiency.)
On the ontology of such institutions, there was near-silence. Wight had referred to them as ‘doctrines’’ and ‘patterns of thought’ in the essay on ‘Western Values’, and to their origins in ‘habitual intercourse’. ‘[M]anifest in the diplomatic system, in the conscious maintenance of the balance of power . . .; in the regular operations of international law’ (Wight, 1966b: 96; my italics). Accordingly, they could be understood as constructions in the constructivist sense. But on whether they were rules, rational precepts or slow-growth social customs, and on whether diverse practices could in fact be rationalised into regularised ‘institutions’, there was no word. As for Hedley Bull’s contribution to their elucidation, the silence is complete. The institutions of international society were introduced into TAS by the stroke of a pen, not by any consideration of their nature. In TAS, they are presented as some kind of self-evident ‘social fact’, a lapse noted by Holsti (2009) who has drawn attention to the lacunae in Bull’s theory in rather emphatic terms (see also James, 1993). Their elucidation as ‘interpretations’ did not come from within the BC, much less from Bull; it came from outside, and it came slowly and piecemeal.
The first assignment of an interpretive identity to the institutions of international society was in Suganami’s (1983) article, which referred to ‘British mainstream international relations’ as sociological and institutional in orientation – institutions that included guiding precepts and practices. (In a private communication of 12 May 2018, Suganami attributed his neglect of TAS’s structural aspects to the fact that, at the time, it seemed a formalised version of Wight.) Suganami’s understanding seems to have been pretty thoroughly interpretive (he reported his reliance on the emerging distinction between ‘understanding’ and ‘explaining’, and since concern with explanation did not seem to be high on Bull’s agenda, assigned Bull’s theory to the ‘understanding’ school). Alongside his characterisation of the approach as Verstehen, his reference to cultural frameworks and ideal-type analysis point clearly to an interpretive approach. But 20 years later, he admitted that ‘there was not clear awareness’ at the time of a particular identity, much less a social constructivist or interpretivist identity (Suganami, 2003). The clearer identification of the ES with interpretivism came not from any ES theorist but from the liberal institutionalist Robert Keohane in his presidential address to the International Studies Association in 1988.
Keohane was responding not to any ES insights directly, but to the early stirrings of American constructivism. Titled ‘International Institutions: Two Approaches’, his address contrasted ‘the assumption of substantive rationality’ that had hitherto dominated the study of international organisation with a ‘challenging sociological and reflective approach’ that ‘stresses the impact of human subjectivity’. Referencing the German Kratochwil and the American Ruggie as writers that ‘emphasise that . . . even states develop within the context of more encompassing institutions’, he distinguished between institutions in the sense of a ‘general pattern or categorization of activity’ (my italics) and institutions in the sense of ‘particular human-constructed arrangements’. In the former, he included Bull’s institutions of international society, citing sovereignty, diplomacy and statehood, which he characterised as ‘persistent and connected sets of rules (formal or informal) that prescribe behavioural roles, constrain activity, and shape expectations’. Examples of the second class; that is, ‘particular institutions’, were the international nuclear non-proliferation regime and the General Agreement on Trade and Tariffs, which were ‘specific institutions with life-histories . . . which depend on the decisions of particular individuals’ (Keohane, 1988: 383–385).
Keohane drew on an over-arching ‘practice of sovereignty’ to suggest a meta-theoretical realm of social construction, noting, almost in passing, that ‘the specific institutions of world politics, with their challengeable rules [are, ed. CBN] embedded within more fundamental practices’. He slightly elaborated this theoretical ‘stub’ by reference to the philosopher John Rawls’s distinction between rules as summations of previous rules (i.e. rules simple and changeable) and ‘the practice conception of rules’. He related the ‘practice conception of rules’ to games of understanding in the sociological sense, which require conformity to rules to achieve the objective of the ‘game’ (e.g. those aspiring after sovereignty must cast their efforts in terms of the fundamental rules defining sovereignty if they hope to gain sovereignty) (Keohane, 1988: 384–385). Such an ordering suggested that the ‘more encompassing institutions’ may be prior to organisations, as sorts of foundational understandings, without which agents could not start to build organisations, rendering institutions in Bull’s sense the foundations for organisational developments. But Keohane did not go so far and apparently did not want to go so far. The essay ended with a defence of the rationalistic approach and with a challenge to ‘reflectivists’ to get their research programme in order; that is, with two sets of institutions as parts of two different research programmes and in an indeterminate relation to one another.
The notion of a relationship between practices, customs and organisational rules and deeper organising social principles appeared a year later, again out of the constructivist school. In 1989, Alexander Wendt and Raymond Duvall, collaborating to clarify some initial constructivist ideas that might be applied to IR, seized upon Keohane’s distinctions (and the ES notion of sovereignty, diplomacy and war as institutions) to illustrate a proto IR theory of social construction. Using the term ‘fundamental institutions’, they identified some international institutions as fundamental in that they ‘represent the shared intersubjective understandings about the . . . precondition for meaningful state action’ (Wendt and Duvall, 1989: 53). These were ‘constitutive’ in that they, among other things, defined and empowered the agents of international society. They distinguished these from the ‘actual practices that are institutionally organised and selected’. Institutionally organised practices, they suggested, ‘constitute the medium through which the social or constitutive side of international institutions is reproduced and/or transformed’, and they added, suggestively, ‘thereby completing the circle of structuration’ (Wendt and Duvall, 1989: 62).
However, this sociological and practice conception of fundamental institutions (and their instantiation in organisation), did not immediately gain much purchase with the British scholars associated with the ES. 4 Following Keohane, Hurrell (1993) authored an ES contribution on international law for the Rittberger and Mayer (1993) volume on regimes that sourced law in a common culture and common interests, but with only a brief reference to practice. He distinguished between ‘the procedural rules of state behaviour and the structural principles which define the character of the system and the identity of the players’ (Hurrell, 1993: 59). But he did not elaborate on the distinction, nor did he explain the difference between a ‘structural principle’ as understood in constructivism and a ‘structure’ as understood in a true structural theory. Evans and Wilson (1992) referred to the ‘interpretive elements’ in ES thought but presented no theory of interpretation and seemed to imply that law was the basis of international order, but again with no theory of law.
The ES theorist who argued ES theory as a form of social construction was Dunne (1995) in the European Journal of International Relations. Drawing on the early Wendt, he recast the by-then familiar distinction between understanding and explanation into ‘subjectivism’ and ‘objectivism’ and related it to the levels of analysis choice, whether to start holistically from the system to it parts or individually from agents to the system. This allowed him to distinguish between different inclinations in subjectivism/interpretivism – one that sees social structures as the result of collective social action – that is, social construction, and another quite different form of subjectivism that sees it in the interaction of individual impulses, which is phenomenology (Dunne, 1995: 372). These are in fact two of the main methods in interpretive analysis (along with ethnomethodology and labelling), and Dunne correctly identified the ES, if it were a form of interpretivism, to align with social construction in method and assumptions. He proceeded to underscore the idea, intimated by Wight and ambiguously endorsed by Bull, that international society emerged out of processes of ‘conscious deliberation’ and put forward the formal argument that ‘[international] society is what states make of it’, dismissing Manning as an important forbearer in the process (‘Manning does not adequately address the key question of the degree to which the actions of the players are determined by the rules and conventions of the game’). He not only distinguished but also put on the same explanatory plane, intersubjectivity and structures, underscoring the argument that ‘intersubjective games can be just as powerful as system structures’ (Dunne, 1995: 378).
But Dunne’s subsequent analyses did not themselves display any remarkably constructivist features. The major constructivist analyses of human rights in the Dunne and Wheeler edited (1999) volume on human rights was put forward by Ken Booth; Dunne’s own contribution to the volume was as part of a joint conclusion with Booth on the potential role of universities in reconstructing a salient human rights programme, and it displayed a rather anti-constructivist voluntarism, in the notion that universities had a role to play in bringing internationally protected human rights to the fore if they would but choose to. He has been the major defender of constructivist IR in the theoretical literature, but his empirical work is as much institutional in orientation as constructivist.
The two thoroughly interpretive approaches to have come out of the ES are Hidemi Suganami on The Causes of War and Robert Jackson’s The Global Covenant. Suganami argued that the effective causes of war are rooted conventionally, in the way the origins of wars are narrated. Comparing various narrative accounts of their origins, the book shows that enquiry into the causes of war is inseparable from the question of responsibility and who or what is responsible for war. Jackson explicitly defends an interpretivist method and identifies 10 procedural norms that lay down the ways and means of conducting inter-national relations (and that provide reference points and justifications for the activities of states and other international actors). 5
It was not until 2004, however, that an empirical understanding of the nature of ES institutions as ‘constructions’ in the interpretivist sense was established. The first move towards that end was in Holsti’s Taming the Sovereigns, where he established the critical distinction between ‘constitutive’ and ‘procedural’ institutions (but where the underlying theory appeared only in two footnotes; Holsti, 2004: 21–22). Buzan amplified Holsti’s insight in his 2004 From International to Global Society and gave added empirical thrust to it, arguing that ‘secondary institutions’, that is, international organisations and regimes, were preceded by ‘primary institutions’, that is, norms and organisational precepts, referring the insight to John Searle. But neither expanded on their ontology – Buzan indeed declared at the European International Studies Association (EISA) in Warsaw in 2013 that he had never worked it out. (Navari (2011) suggested their ontology in ‘The Concept of Practice in the English School’, where she related fundamental institutions to ‘practices’, casting both as teleological constructs and referred the latter to Wittgenstein and to systems of intelligibility. But she did not relate the concept of a practice to interpretivism as a methodology.)
In the interim, Wendt had published his Social Theory of International Politics, and the constructivist interpretation of the ES became, by association, widespread, the main question being how the ES differed from constructivism-simple. Suganami (2002), acute as usual, was one of the few who noticed Wendt’s desire to stay ‘within American scientific orthodoxy’ and who pointed out that Wendt’s theory was in fact a form of structural theory.
The structural side
The structural side appeared early, reflected in what would become the BC’s persistent use of the term state-system to describe and to characterise the object of their inquiries. Initially intended to identify and characterise different modes of interstate relationships – Wight (1977), who wrote the book, considered state systems as simply ‘a group of states having relations more or less permanent with one another’ (1977: 22) – they were initially ‘systemic’ only in the sense that they had permanent institutions – resident ambassadors, conferences and so on. But they became associated almost at once with defining and informing properties. War would be a persistent feature of all of them (‘the causes of war, like the need for diplomacy . . . will remain so long as a multiplicity of governments are not reduced to one government’; Butterfield, 1966: 138). Balancing power seemed to be a recurrent practice (notably within the Western state system) and, though demonstrated by Butterfield (1996: 144) to be a delightfully variable precept, periodically took on the features of a constraining institution: ‘it [the balance of power] operated to preserve freedom rather than peace’. The Western state system also ‘came nearer than any other’ to making a ‘structure of government’, that is, democracy, ‘a principle of international legitimacy’ (Wight, 1977: 41).
These early representations are probably best understood as ‘soft structures’ in the sociological/anthropological sense – collective judgements embedded in institutional practices with consequent effects; but in TAS, they took on a harder edge. Institutions became ‘a set of practices and habits . . . [that] give substance and permanence to their [states] collaboration . . . and moderate their tendency to lose sight of common interests’. A chapter is devoted to each of the five ordering institutions at the end of which the main functions that each institution performs are summarised. Bull goes out of his way to deny that he has a structural functional notion of order in mind or that ‘the primacy of the whole over its part [accounts] for what occurs within it’ (Bull, 1977b; 75). But there is no doubt that the ordering elements order; they set parameters, they limit action, and they entail obstacles that have to be overcome. Thus, the balance of power has ‘served to preserve the system from . . . domination by a . . . predominant power’ and ‘provide[s] the conditions . . . upon which international order depends’ (107). International law embeds sovereignty and bounds the state ‘by common rules’ (40). Among Great Powers, there may be ‘preponderance’, ‘primacy’ or ‘hegemony’, but each entails patterned relationships and constitutes an arrangement with effects. For example, hegemony ‘produces a kind of order . . . [t]he lesser states in each area cannot resort to force against each other, nor can their governments be overthrown, except by leave of the hegemonial power’ (215–219). Above all, there is the structural constraint of order over justice. Donelan (1978) noted Bull’s positivist leanings in his response to Bull’s review of The Reason of States (a work edited by Donelan on the part of the International Political Theory group at the LSE whose activities paralleled those of the BC).
Bull’s (1977b) account also set forth a set of alternative ‘structures’ that anticipated the by-now familiar triad of international system, international society and world society (1977b: 239–250). These were essentially imagined political orders, designed as thought experiments, to test the durability of the state system as constructed and to locate the sources of potential change and transformation within it, but they drew on contemporary and historical example to demonstrate that they were plausible alternatives. Thus, the ‘new medievalism’ drew on the contemporary dispersal of power to regional organisations, while the ‘World Political System’ drew on transnational cause groups and their activism within the UN system (Bull, 1977b: 261–263, 270).
Bull presented these imagined orders, somewhat confusingly, as of similar ontological status to his account of the existing international society, producing the triad of international system, international society and world society. He actually named them the Hobbesian, Grotian and Kantian orders, adapting Wight’s (1977) schema of international political thought, and he identified them with different structural aspects (1977: 24). He characterised the Hobbesian order as a condition of interstate conflict unmediated by rules and as a ‘purely distributive or zero-sum game’. Behaviour in such a system would incline to that as described by Hobbes. He suggested that the Kantian order was one of a necessary ideological conflict, which ‘enjoin[s] not coexistence . . . but rather the overthrow of the society of states and its replacement by a cosmopolitan society’, equally a structural concept. The Grotian imperative was to ‘enjoin . . . acceptance of the requirements of co-existence and co-operation’ (1977: 26–27).
But international system and world society, along with the new medievalism are not structures at all. They are and were intended by Bull as ideal-typifications, to be used for analytical purposes. The one that Bull considered to be the most realistic alternative (‘on the cards’ in the mid-1970s – Bull, 1977b: 258) was the decline of international society and its reversal to a system. This is a situation of patterned relationships whose pattern emerged not from conscious design but from fortuitous interaction: As he termed it, ‘causal connections of certain sets of variables to one another’ (Bull, 1977b: 13). It is clear that he understood the bi-polar nuclear balance to be more the result of an unintended stand-off than the creation of a Soviet–American sense of common interests (though the latter he reckoned played some role; Bull, 1977b: 259). He cites the continuing ideological differences between communist and non-communist states and the growing tensions between rich industrial states and poor agricultural states as well as the consequences of a past, and fears for a future war as conditions curtailing and limiting cooperative endeavours. In an international system, states lose a sense of common interests but are forced by their particular political or geographical placement to interact, whether they wish to or not; hence the designation as a ‘field of action’.
The idea of a field of action, and its association with the international system, was a borrowing from Morton Kaplan’s System and Process, a work that exercised a continuing fascination for Bull. He defended Kaplan’s concept in the first meeting of the BC on state systems in January 1965, having just returned from America, and referred to it repeatedly in his 1967 ‘Notes on the Modern International System’ (Vigezzi, 2005: 205). More significantly, he issued an apology for any previous denigration in his 1972 piece for the Australian Outlook (Bull, 1972). After Bull had been 7 years in Australia, returning via India, and remote from the European centre, he was more inclined to consider the system of states more in terms of an international system than an international society.
The dubious status of ‘international system’ as a structural concept was picked up by Alan James. In his 1993 ‘System or Society?’ James’ (1993) argument was that any set of relationships between any set of social bodies required mutually recognised rules, however minimal, and that ‘his [Bull’s] distinction between an international system and an international society is without substance’ (1993: 276). James did not argue that an international system could have no causal properties – that is, that anarchy between sovereign entities did not entail certain inevitable consequences. His argument was that anarchy between sovereign entities was a rule-bound condition, and hence not ‘systemic’ in any sense understood by the notion of a system (he famously used the example of the digestive system), and accordingly not structural in the sense intended by a true structural theory. Buzan (2004) would eventually join James in abandoning the concept of an international system, on the grounds that no historically existing systems had been completely asocial 2004: 98–101).
The confusion between structural entailments and ideal-typifications of international order reached new heights in Wendt’s (1999) Social Theory of International Politics, who postulated three ‘cultures of anarchy’ in a remarkable case of parallelism with Bull’s schema, characterising them as Hobbesian, Lockean and Kantian, each of which is a particular social structure made up of material factors, interests and ideas. His Kantian order differs from that originally proposed by Bull in that it has at its heart a pluralistic security community in which states precisely do not contest ideologically. But the Hobbesian and Lockean orders followed Bull’s originals. The subject’s position at the core of the Hobbesian culture of anarchy is ‘enemy’, while at the core of the Lockean order are ‘rivals’ and in the Kantian order ‘friends’. Wendt’s are all ‘structures’ in the formal sense that the logic of identity and in consequence conduct (and ‘interests’) are determined largely as a result of the nature of the structure, but they are equally ideal-typifications, which are meant to be suggestive of the analytical positioning the analyst should adopt in confronting international reality. None refers to an empirical reality.
The most determined of the ES structuralists is Barry Buzan, strongly influenced by Waltz and the theory of structural realism: that a system had properties that would determine, or at least strongly influence, the behaviour of the agents has been a driving concern of his work. This has led him, however, into a concern with differentiation rather than structuration. Having abandoned the system–society distinction, his concern became to differentiate between international society and world society, which revolved around the analysis of the relevant units (Buzan, 2004). International society, according to him, was characterised by an interstate domain (defined by states), whereas world society was the home of an inter-human domain (defined by individual humans) and a transnational domain (defined by transnational actors). In essence, Wendt’s (and Bull’s) schema of three ideal-typical international societies was adapted, and transformed, by Buzan (2004) into interstate, inter-human and transnational societies. Each is not a single social order but rather an ideal-typical spectrum in which the relevant agents (states, humans and transnational groups) can go from enmity to relative amity. Buzan picked up from Wendt the idea that social structures can create both identities and interests, and this is reflected in his three types.
The one true account of a structure, according to the meaning of ‘structure’, to have followed TAS is Buzan’s (and Lawson’s) account of the structure of globalisation. Historically based on three dynamics (industrialization, rational statehood and ‘ideologies of progress’), these combined to generate a new basis for how power was constituted, organised and expressed. This new ‘mode of power’ dynamically favoured the West and pulled all previously existing ‘sub-systems’ into a global system of economic exchange (Buzan and Lawson, 2013). This political/economic structure was accompanied by a ‘social structure’ (which Buzan had previously identified) characterised in terms of interaction capacity – the level of transportation, communication and organisation capability in the system (Buzan et al., 1993, ch. 4). Interaction capacity had become ‘complete’ by the late-twentieth century, producing ‘globalisation’ as a political, economic and social structure. He completed the structure by arguing a ‘significant logic of cultural syncretism, which means that there is less of a contradiction between parochial and cosmopolitan identities than might at first glance seem to be the case’, allowing him to assert that a ‘kind of world society is possible’ (Buzan, 2010).
Buzan’s structural interpretation crystalized the latent tension in the theory. Dunne (2005) seized on it to claim it was a standard social science understanding of theory and as yet another misguided attempt to apply positivist methods ‘to comprehend a post-positivist ontology’ (2005: 162).
Squaring the circle: The new institutionalism?
If we set aside the meta-theoretical confusion and consider the empirical examples scattered through the basic texts, more pertinently, if we consider the secondary literature such as Gerrit Gong’s Standard of Civilization, a different and steadier picture emerges. It is generally one of the actors responding to the situations in which they find themselves, and those situations described in terms of major institutional arrangements within which they are, more or less, compelled to act. In Gong, a civilizational concept has invaded the corpus of international law ordering state relations between ‘advanced’ and ‘backward’ states, in the course of which different categories and rules come to be applied to each. He argues, convincingly, that these came to serve as the standard grid for incorporating foreign lands into the widening European ‘system’. This approach points to historical or sociological institutionalism; ‘Institutionalism’ uses institutions as subjects of study in order to find, measure and trace patterns and sequences of social, political and even economic behaviour. Historical institutionalism focuses on change across time; sociological institutionalism on synchronic developments. In IR, regime theory is a case in point: regimes have been defined by Krasner (1983) as a set of explicit or implicit ‘principles, norms, rules, and decision-making procedures around which actors’ expectations converge in a given issue-area’. The convergence of expectations explains the subsequent outcome in action, including new institutional formations, which produces a new set of conditions around which a new convergence of expectations occurs. Decision-making and new or reformed structures are heavily influenced, if not determined, by the shape and character of previous institutional arrangements.
A newer institutional theory moderates that picture slightly, in that it highlights cultural influences on decisions and organisational processes. Organisations, and the individuals who populate them, are suspended in a web of values, norms, rules, beliefs and taken-for-granted assumptions. In the new institutional literature, the web is more diffuse, more ideational and ‘at least partially of their own [individuals’] making’ (Barley and Tolbert, 1997: 93). The actors are active in creating influential cultural products such as norms and rules as well as being influenced by them. But equally, the cultural elements define the way the world is and should be. In the concept of the Stanford School (a paradigmatic institutional approach), ‘They provide blueprints for organizing by specifying the forms and procedures an organization of a particular type should adopt if it is to be seen as a member-in-good-standing of its class’ (Meyer and Rowan, 1977: 346). In the new institutional literature, values, norms and rules are the focus rather more than the formal institutions. But they perform the same functions as the formal institutions, in constraining or guiding the agents.
Bull and Wight resemble the ‘new institutionalists’ in their views of the structure of international society. Their structures are as much ideational as organisational or material and they are the results of collective action. There is a pervasive notion of constraints, particularly given the resilience of the notion of sovereignty, amounting even to some claims of little change – Wight detected remarkably little change in the fundamental institutions of international society bestowed by ‘Western values’ over 300 years. War remains a determinant of the system, and self-help makes war an ever-present possibility. A wide variety of historical influences are at play in the accounts, but the ones that matter are the ones that inform the secondary institutions. These can change, such as new treaty arrangements like the World Trade Organization (WTO) with its formalisation of the most-favoured nation rules (see, for example, Holsti, 2004: 225–226), the assignment of specific powers to the Security Council and the development of Responsibility to Protect. The agents can modify the arrangements, as for example, formalising great power responsibility within the Security Council, but within limits.
Equally, however, ES structures are not deterministic, and they are remarkably open to change even as parameters. Butterfield’s historic diplomacy morphs seamlessly into the ‘new diplomacy’. Wight’s ‘International legitimacy’ is presented as the subject of a sea-change, from the dynastic principle to the popular principle. Mayall’s (1990) nationalism is proposed by a few Central European expositors at the beginning of the nineteenth century, and within decades transforms the international system from a dynastic civilizational basis to a populist cultural basis. Also, the forces of change are ideational as much as interest-driven, material or politicist.
More to the point, such changes are understood in terms of complex interactions between the identities that institutions bestow and the agent’s self-identifications. Thus, Hudsen’s ‘genuine alliances’ are those with ‘precisely defined’ commitments that allow for ‘obliging allies’. In this framing, ‘obliging allies’ are defined by a certain nature of commitment and they become obliging allies through the creation of that kind of commitment (Hudsen, 1966: 178). So, Hudsen (1966) tells us that Britain in concluding the NATO, ‘as never before . . . has yielded her cherished freedom of action’ (1966: 179). Mayall’s nationalists become ‘state nationalists’ through their reforming of economic relations through state action, changing the state as well as their identities in the process.
The relation is presented in a manner such that agents, identities and institutions are often difficult to disentangle in any simple causal way. Thus, Wight’s (1966a) ‘so long as there is an absence of international government’, so long must powers be preoccupied with their own survival and so long will they seek to maintain ‘some kind of balance’ (1966a: 174). It is this melding of institution and agent that pushes the theory more in the direction of structuration.
Structures or structuration?
Structuration theory is that of Anthony Giddens. It has certain resemblances to the new institutional theory, in a correspondence between Giddens’ (1984: 2, 377) notion of ‘structure’ and the definition of institutions in the new institutional literature. Structuration also, like the new institutionalism, implies that ideas and norms work their way into organisations, which are powerful socialisers for those who inhabit them, informing identities. The difference is the emphasis on agency as much as on structures. Structuration theory is not only about how agents receive identities but also about how they renegotiate them and even transform the structures that they inhabit. The emphasis is less on constraints than on identities, and on transformational routes between old identities and institutions and new ones.
In method, structuration analyses both structure and agents without giving primacy to either; in other words, neither micro- nor macro-focused analysis alone is considered sufficient. At the macro-level, and at the heart of a social system, are shared rules and typifications that identify categories of social actors and their appropriate activities or relationships (Giddens, 1976: 29). With regard to IR, such typifications would include state, head of state, diplomat, secretary-general and so on. At the micro-level are actors who are knowledgeable (‘mutual awareness of . . . a common enterprise’; Giddens, 1984: 4, 375). At the micro-level the various processes at play focus on and highlight moments when identities are being adjusted and institutions reformed or even demolished. Analysis focuses on the strengths and weaknesses of institutions as they are experienced by the agents and how the agents perceive their identities within the institutions. In Giddens, this led to the view that agents and institutions were ‘two sides of the same coin’.
Most theories have intimations of both structure and agency, but in the ES corpus, these appear in a particular way. In fact, the corpus is largely agentic in orientation; that is, the dynamic elements in ES accounts of international life are either collective or individual agents who hold institutional positions. They are states, state leaders, nongovernmental organisations (NGOs), development officers, agency bureaucrats and so on. Moreover, the agents are, themselves, constantly involved in creating, elaborating, reaffirming or altering the institutional environments of their action. They are searching out institutional possibilities that will aid their specific objectives, in the process reaffirming them, while also creating or altering the structures within which they act. This points to the significance of structuring processes – successive actions that create stabilising effects and whose effects have consequences for what follows. Agents interacting with institutions place the ES in a family relationship with the ‘new institutional theory’, to be sure, but establish it more firmly as a type of structuration theory.
The idea that ES theory is a form of structuration theory is not new. Clark’s (2007) work on world society informing and impacting on international society, Spandler’s (2015) work on constitutionalisation and institutionalisation, Parrat’s (2014) work on institutions and actors modelling them, Holsti’s (2004) and Buzan’s (2004) works again on institutions implicitly or explicitly rely on structuration theory. But the recent study of the relations of the ES’s fundamental institutions and international organisations, carried out by Knudsen and Navari (2019) (the joint editors) with a team of young scholars, published as International Organization in the Anarchical Society develops the theory to its furthest point so far and models it.
The study posits a reticular relationship between ‘fundamental institutions’; that is, embedded norms, values and practices, and international organisations. In the relationship, fundamental institutions not only inform but also are actually reposed within international organisations, which reflect them. So, for example, the sovereignty principle has been embedded in the UN Charter as Article 2.4 and informs the provisions of the Charter throughout. The agents of international society (states, heads of states, security agencies, NGOs, etc.) are constituted in significant ways by both the fundamental institutions and the ‘secondary’ organisations, and they are, importantly, empowered by them. But at the same time, the agents are constantly involved in institutionalising processes, which constitute new agents and elaborate fundamental institutions, as well as organisational processes. The model presented by the authors traces the relevant relationships and routes of influence (Figure 1).
The Institutional Structure of World Order.
The model is heavily institutional, and arguably structural, at two levels, both at a conceptual level of agreed norms embodied in practices and at an evidential organisational level. But the dynamic elements in the theory are the practitioners, the ‘interested agents in interaction’. It is their iterations that stabilise the institutional order and that indeed ‘institutionalize’ it. They institutionalize at both levels, creating new norms and principles at the fundamental or primary level of institutions and new rules and procedures at the organisational level. In other words, the agents create the structures within which they act.
At the same time, it is not the sort of agentic theory which presents the agents as determining their choices on the basis of an unlocated ‘freedom’. The agents are situated in the sense Bevir suggests in his 2017 article ‘Situated Agency’ – they have reflexivity and creativity while always enmeshed in specific historical contexts. But their situatedness is also related to and informed by their institutional identities in non-fortuitous and effectual ways.
More to the point, the structures in the model do not constitute the sort of structure that is demanded by a true structural theory. To be ‘structural’ in the social science sense requires a high degree of determinism – the structures must be doing most of the work in producing outcomes. This means that they must be uni-directional, and functionally purposive, the archetype being Marx’s class system, which in its evolution through capitalism produces the kind of working class that is capable of revolution. The fundamental institutions of international society – sovereignty, international law and diplomacy are not like that at all. They are protean composites, flexible and constantly being adjusted. They present different action routes and different possibilities of interpretation. Not everything is possible, but it is not possible to know before the fact what is impossible. What the structures in structuration theory do is to alert the agents to obstacles that have to be overcome and to involve them in heavy debates about the possibilities of overcoming them.
Aligning the ES with structuration inevitably burdens it with the limitations generally ascribed to Giddens’ theory; that is, under-determination, tendency to historicism and simplicity about the nature of ‘rules’ (Archer, 1982; Thompson, 1984). But then, these are the charges with which the ES has traditionally been taxed. An observation worth making is that the authors of the text within which the above model appears have improved on Giddens, as well as on previous ES accounts, in distinguishing between values-simple, values accepted as the norm, procedural norms and rules, and the rules instructing agents. The model also improves on Giddens and previous ES accounts by making it possible to distinguish between the reproduction of social structure in the general sense and the reproduction of institutions. As for under-determination, this is the logical entailment in moving from determinants to parameters in the framing of social action and should be understood as such by all theorists who move away from structural determinism.
Conclusion
Finnemore (2001), responding to Buzan’s (2001) call for more attentiveness to the ES, complains about a ‘lack of clarity about both [its] method and theoretical claims’ and that ‘simply figuring out what its methods are is a challenge’. One can sympathise. The Expansion of International Society, arguably the core text of the first generation of ES theorisers and a joint enterprise directed largely by Bull, is an example. It is an historical text, engaged in describing a critical historical process. This process is seemingly driven by a bewildering variety of propellants: ideas, institutions and agents are all at play, and ideas and actors jostle against one another for prominence. Institutions are presented at the same time as constraints but also as intense subjects of contestation, as well as objects of reform. It also occasionally descends from historical narrative to contemporary analysis, appearing to leave the historical story behind.
But there is order in the tale. First, note the emphasis on institutions. Expansion codified the idea that international society is constituted by institutions and that entities joined international society by conforming to its rules, practices and practical understandings, not least the rule that required state formation in order to be a member. This means that rules and institutions at the phenomenal level are the analytical referents, the focus and point of the exercise – in methods language, the explanandum. Second, in parallel with this focus, and keeping a steady eye on it, we may observe practitioners who are both creative and reactive, and who have a variety of intentions. In Expansion, these appear to be overwhelmingly politicist but with political motivations that are heavily informed by prevailing ideas of value – what Weber referred to as ‘ideal interests’. It is this aspect that opens the door to interpretivism, but an interpretivism that has to be understood in terms of the particular situatedness of the actor. The actors’ intentions are the explanator, but only in the context of the actor’s institutional identities and within a political game that has rules. Finally, the action stops occasionally for an analysis of the nature of the game, since this can also change. (In Expansion, those bits were written by Bull.)
Understanding ES theory in terms of co-constitution, in which situated agents constantly react against, support or renovate the institutions within which they act, clarifies much of both the ES theory and the thrust of its empirical studies. It also makes sense of the interpretive elements: the agents in ES theory are constantly interpreting their situations in the sense of ‘understanding’ – they are not the passive receivers of structural determinants. On the structural side, understanding the structures in terms of structuration makes clear what kind of ‘structures’ they are: on one hand, ‘institutions’ in the sense of conceptual and ideational deposits; on the other hand, rules. It also draws attention to the central dynamic that the ES theory proposes as the heart of international political processes: situated agents in organised settings pushing against rules.
