Abstract
Taking off from a recent article in this journal by Barry Buzan and Mathias Albert, I present a framework of structural differentiation as an alternative to IR’s standard tripartite conception of the elements of structure. Rather than employ ideal type models, though, as Buzan and Albert do, I present a more open-textured, multidimensional account of differentiation. I also emphasize the systemic nature of structural analysis. The elements of structure are interdependent parts of wholes (systems) — not the independent variables implied by standard formulations such as ‘the effects of anarchy’. A multidimensional systemic approach directs our attention to the diversity of and change in international systems and their structures, which mainstream structural IR typically ignores or obscures. I illustrate both the regularity of extensive structural change and the analytical utility of my differentiation framework with case studies of post-World War II international society and contemporary processes of globalization.
Keywords
Introduction
In a recent article in this journal, Barry Buzan and Mathias Albert argued that differentiation, a concept widely used in Sociology and Anthropology, has considerable untapped potential in IR. They also presented ‘an analytical framework of differentiation’ (2010: 315). I share their sense that differentiation has much to offer our discipline. I argue, though, for a different kind of framework.
I begin by introducing the concept of differentiation and showing its close connection to standard understandings of structure in IR. I then advance a multidimensional framework of structural differentiation as an alternative to the dominant tripartite (ordering principle, functional differentiation, and distribution of capabilities) conception of the elements of structure. I show that at least six dimensions are required to depict the structure of social and political systems: stratification (vertical differentiation), functional differentiation, segmentation (unit differentiation), polarity, geography and technology, and norms and institutions.
I further argue that mainstream IR misrepresents how the elements of structures function. Anarchy, which does almost all the work in standard structural explanations, typically is treated not as an interacting part of a system but as an independent variable with determinate effects; ‘the effects of anarchy’. So-called structural analysis in IR is in fact reductionist analysis that in principle cannot capture ‘system effects’. My framework, by contrast, provides an entry into the study of system dynamics.
A multidimensional systemic conception of structure shifts our focus from allegedly universal features of international systems to structural diversity and change — which, when we look for it, is substantively at least as striking and analytically at least as important as any generic similarities across international systems. The final two sections examine post-World War II international society and globalization to illustrate both the regularity of extensive structural change and the utility of differentiation approaches to structure.
Differentiation, structure, and IR
Social differentiation is the process by which ‘social groups become dissociated from one another, so that specific activities, roles, identities, and symbols become attached to them’ (Yoffee, 2005: 32). Differentiation establishes the ‘relations and mechanisms that constitute social categories and structure social boundaries’ (Juteau, 2003: 7). Through differentiation, ‘the main social functions or the major institutional spheres of society become disassociated from one another, attached to specialized collectivities and roles, and organized in relatively specific and autonomous symbolic and organizational frameworks’ (Eisenstadt, 1964: 376). 1
Understanding social structures as ‘multidimensional space[s] of different social positions’ (Blau, 1977: 4), differentiation creates and fills the positions that comprise structures. If structural theories provide a ‘purely positional picture of society’ 2 (Waltz, 1979: 80), then differentiation is what structural theories depict. Social structures are networks of differentiated social positions. Social structures exercise their constitutive and regulative effects, ‘shaping and shoving’ actors (Waltz, 1986: 343), through differentiated social positions.
The preceding paragraphs would seem obvious to most sociologists and anthropologists. Mainstream IR, however, views functional differentiation as the only form of structural differentiation and holds that ‘anarchy … precludes functional differentiation’ (Weber, 1990: 62). Differentiation thus would seem to have nothing to contribute to IR.
I agree with Buzan and Albert that, on the contrary, differentiation is central to sound structural theory. I focus, though, on another approach to differentiation.
Types and dimensions of differentiation
The Blackwell Dictionary of Sociology identifies two senses of social differentiation: ‘[1] the tendency of social systems to become increasingly complex as they develop, in particular through specialization … [2] the general social process of distinguishing among people according to the social statuses they occupy’ (Johnson, 2000: 88–89). These senses are associated with what I will call ‘type’ and ‘dimension’ approaches. (Waltz similarly considers both types of political structures [anarchic/international and hierarchic/national] and elements of structure [ordering principle, functional differentiation, and distribution of capabilities].)
Type approaches, on which Buzan and Albert focus, assume that ‘different types of social structure can be classified according to their dominant principle of differentiation’ (2010: 317). They also assume that there are few types; in Buzan and Albert’s case, only segmentary, stratificatory, and functional differentiation. ‘All other variants vary within these three principles’ (2010: 317).
Dimensional approaches, which I will emphasize, see societies as differentiated along multiple dimensions that often do not coalesce into a small number of types. Societies ‘have the annoying habit of possessing traits of [multiple] types’ (Yoffee, 1993: 64–65). We cannot assume that ‘the key question [is] which form is dominant in shaping the social structure as a whole’ (Buzan and Albert, 2010: 319). We must first determine whether there is a dominant form.
Even where there is, a single feature specified at a very high level of generality provides limited analytical leverage. Dimensions of differentiation permit more precise and powerful structural accounts. They also allow us to investigate, systematically, both ‘secondary’ dimensions and deviations from type. Dimensions thus are, at minimum, a necessary complement to types. Types might even be considered special cases in which a particular dimension happens to predominate. 3
Systems, structures, and differentiation
My goal is to present an empirically accurate and analytically fruitful account of the dimensions or elements (I use the terms interchangeably) of the structures of social and political systems. By ‘system’ — or, more precisely, ‘complex system’ — I mean a bounded set of elements characterized by internal relations that create a whole with properties that are not reducible to those of the parts and their interactions. By ‘structure’ I mean the arrangement or positioning of the parts of a system. These understandings, although not entirely uncontroversial, are fairly standard in IR and other social sciences. 4
Structural analysis, thus understood, addresses the positioning of actors in social and political systems; the properties and relations that make them parts of a system. 5 This is another way of talking about system differentiation. Differentiation defines and populates social positions; that is, arranges the parts of a social system. If ‘a structure is defined by the arrangement of its parts’ (Waltz, 1979: 80), differentiation defines structure. The principal dimensions of differentiation delineate a typology of elements of structure.
International structure: Variable not constant
In the standard Waltzian account, international systems are largely undifferentiated — and thus pretty much all the same. In international orders ‘the units of the system are not formally differentiated’ (Waltz, 1990b: 31; cf. Waltz, 1979: 59, 97). Functional differentiation, Waltz argues, is a consequence of ordering principle. (Hierarchic systems are functionally differentiated. Anarchic systems are not.) Different distributions of capabilities do produce secondary differences in international systems. Structural analysis in IR, however, is largely about ‘the effects of anarchy’ in ‘anarchic orders,’ ‘anarchic systems,’ or ‘anarchic structures.’
In fact, though, anarchy and polarity do not even come close to depicting the arrangement of the parts of international systems. Waltz has confused the definition or delimitation of international systems with their structure. Even if all international societies are in some particular way(s) structurally different from all national societies, 6 these shared ‘defining’ features simply do not (alone) define the structure (differentiation; arrangement of the parts) of international societies.
International social and political structures are, at minimum, differentiated along, and vary across, the dimensions listed in Table 1. Before showing this, though, I want to consider three objections to my project.
Dimensions of international structural differentiation
Anticipating three objections
Many will question my framework’s (lack of) parsimony. An analytical framework of differentiation, however, requires as many dimensions as necessary to depict the important positions in close to the full range of systems to be analyzed. Substantive theories may fruitfully narrow their attention to a few dimensions. One might even say that explanatory theories should err on the side of parsimony, if err they must. Typologies or analytical frameworks, however, should err on the side of inclusiveness.
In any case, though, my framework is not too inclusive. I will show that each identified element is needed to depict accurately the fundamental positions and their relationships in social and political systems. 7 I will also show that the apparent parsimony of the mainstream approach is mostly misleading appearance.
My concern with accurate depiction might seem to ignore the fundamental distinction between description and explanation. The positivist conception of explanation as deductive prediction irrespective of descriptive accuracy, however, is misapplied to a framework of differentiation. In stating what differentiates social positions — what the elements of social and political structures are — a considerable degree of descriptive accuracy is essential. And when the arrangement of the parts is doing the explaining, ‘as if’ explanation is no explanation at all. 8
My attention to Waltz might also be challenged — as misguided, excessive, even obsessive. But I criticize not structural realism, which indeed may be a long-since played-out topic, but the Waltzian tripartite conception of structure, which continues to dominate mainstream IR. (Ask a PhD student in IR the elements of structure and the almost unthinking answer typically is ‘ordering principle, functional differentiation, and distribution of capabilities.’)
Most realists, both structural and neo-classical, accept Waltz’s conception of structure. Neo-liberals define structure in Waltzian terms (e.g. Keohane, 1989: 8; Keohane and Nye, 1989: 260) and study the ways in which institutions, understood as non-structural, mitigate the (structural) effects of anarchy. Actor-oriented rationalists, when they do not abstract from structure or consign it to a largely undifferentiated environment, typically refer to it in either the Waltzian or the ordinary language sense of the term.
Even constructivists have advanced no widely endorsed alternative conception of the elements of structure. Identity and norms, on which much constructivist work focuses, are often presented as ‘structural.’ That term, however, is usually employed in a loose, ordinary language sense. The focus is usually on establishing that they are structural (without situating them within a broader framework of elements or types) or on their explanatory power (which is rarely considered in the context of interacting elements of a system’s structure).
Consider, for example, Wendt’s emblematic account of cultures (not structures) of anarchy, which are defined largely by identity and associated ‘role structures’ (1999: ch. 6). Although Wendt admits that ‘in what follows … the structure of the international system is [treated as] its “culture” even though in reality social structure is more than that’ (1999: 249), he gives no real attention to what that ‘more’ is and how culture relates to it. Although I agree with Wendt that ‘social regularities are determined primarily by shared ideas that enable us to predict each other’s behavior’ (1999: 251), my principal attention is instead on the fact that those shared ideas are not reducible to culture or role structure and that there is a material dimension to structure as well — which points to the need for a comprehensive multidimensional account of structure.
My goal, to repeat, is (neither more nor less than) to present, and illustrate the utility of, a differentiation account of the elements of social and political structures as an alternative to mainstream IR’s tripartite conception.
Dimensions of differentiation
This section introduces the six dimensions of social and political structures identified in Table 1. I show how each positions actors in social and political systems. I also show that even the notoriously parsimonious Waltz in practice employs all of them, providing powerful unexpected support for my framework.
I divide these elements into ‘form’ and ‘content.’ 9 Social and political structures cannot be understood independent of the substance of the positions actors occupy. For example, a formal account of stratification, placing actors higher, lower, or equal to one another, is incomplete without a substantive account of the basis for that stratification and the rights and obligations associated with particular super- and subordinate positions.
For reasons of space, this section too often reads like a list, with only the briefest, allusive illustrations. Two sections below, however, the framework is explicitly deployed to analyze structural change in post-World War II international society.
Stratification (vertical differentiation)
Waltz labels his first element of structure ‘ordering principle’ and identifies two types: hierarchy and anarchy, in which units do and do not stand in formal relations of superordination and subordination. This is much better described as stratification (or vertical differentiation), avoiding any possible implication that order can be reduced to stratification.
Stratification in states systems
Pace Waltz, stratification is a central feature of international systems. Even if ‘each state, like every other state, is a sovereign political entity’ (Waltz, 1979: 96), sovereign statehood is a particular status, with rights, powers, and obligations denied to other actors:
States set the scene in which they, along with nonstate actors, stage their dramas or carry on their humdrum affairs.… [States] set the terms of the intercourse, whether by passively permitting informal rules to develop or by actively intervening to change rules that no longer suit them. When the crunch comes, states remake the rules by which other actors operate. (Waltz, 1979: 94)
In other words, states are differentiated from and superordinate to non-state actors. States are legally entitled to command non-state actors, which are required to obey. Georg Schwarzenberger thus calls states ‘the aristocracy of international society’ (1951: 102).
States are also stratified among themselves. ‘Inequality … is inherent in the state system; it cannot be removed. At the pinnacle of power, no more than small numbers of states have ever coexisted as approximate equals’ (Waltz, 1979: 132). ‘Great powers … dominate and shape international politics’ (Mearsheimer, 2001: 17). They occupy ‘extraordinary positions in the system’ (Waltz, 1979: 199). Schwarzenberger thus identifies, among the aristocracy of states, an ‘international oligarchy’ of great powers, whose members ‘have arrogated to themselves … supreme control over affairs within international society’ (1951: 113).
‘International politics can be studied in terms of the logic of small-number systems’ (Waltz, 1979: 131) only where international societies are thus doubly stratified (i.e. where states are vertically differentiated from non-state actors and great powers from lesser powers). That such stratification usually is taken for granted should not lead us to denigrate its significance — let alone deny its existence. Quite the contrary, precisely because it is assumed to be fundamental, our account of structure should highlight rather than obscure it.
Denying stratification in depicting the structure of international systems is not a defensible analytical simplification. It is both wrong and wrong-headed. As Waltz notes, ‘international politics is mostly about inequalities’ (1979: 94). Super- and subordination, both formal and informal, are central to the structure of most international societies — including modern international society.
Over the past decade, contemporary IR has increasingly come to recognize ‘hierarchy’ as a regular feature of anarchic orders (e.g. Donnelly, 2006; Hobson and Sharman, 2005; Lake, 2009). Hierarchy, however, is too blunt an analytical tool. And with anarchy no longer the opposite of hierarchy — that is, once we recall that absence of government entails no particular form of stratification — we have no way to describe a system without super- and subordination.
In place of an anarchy–hierarchy dichotomy (or continuum), I have elsewhere (Donnelly, 2009: 55–71) identified (and illustrated) three principal forms of stratification:
Societies may be unranked (‘egalitarian’), either because no one has any rank/status (as in states of nature) or because all have the same rank (as in systems of sovereign equality).
‘Singly-ranked’ (‘hierarchical’) societies have one axis of super- and subordination. Waltz’s hierarchic/national orders and the international system of the Roman Empire are examples.
Societies may also be multiply ranked (‘heterarchic’), with different hierarchies operating in different domains (see Donnelly, 2009: 63–70). Medieval Europe provides a particularly complex example.
Force, authority, and stratification
The Waltzian conception of structure is also seriously flawed in (quite ‘unrealistically’) restricting ‘ordering principle’ to formal authority (1979: 81, 88) and separating it from ‘distribution of capabilities.’ Stratification properly links ‘force’ (coercive material capabilities) and authority as complementary sources and dimensions of vertical differentiation.
The relative importance, coherence, and fungibility of force and authority vary across systems. Realists model power as largely a matter of force; ‘authority quickly reduces to a particular expression of capability’ (Waltz, 1979: 88). Liberals stress institutional authority. Constructivists often emphasize normative authority. Whether any of these accounts applies to any particular international system is an empirical question. We need a conception of the elements of international structures that treats it as such.
Functional differentiation
Societies differ in how they define and allocate social activities. Waltz’s claim that ‘the units of an anarchic system are functionally undifferentiated’ (1979: 97) is rebutted by his own analysis.
An adequate international theory, Waltz insists, must ‘show how international politics can be conceived of as a domain distinct from the economic, social, and other international domains’ (1979: 79). International relations, in other words, is functionally differentiated, into political and other subsystems (cf. Buzan and Albert, 2010: 316, 318, 326).
Waltz also recognizes functional differentiation among states. Chapter 9 of Theory of International Politics addresses ‘The Management of International Affairs,’ which is neither left undone nor performed by undifferentiated units. ‘Some of the parties emerge as specialists in managing system-wide affairs’ (1979: 197). Great powers ‘take on special responsibilities’ (1979: 198). ‘Managerial tasks are performed’ in anarchic no less than in hierarchical orders ‘but in markedly different ways’ (1979: 204).
Barry Buzan and Richard Little (2000: 77) conceptualize international functional differentiation in terms of military, political, economic, societal, and environmental sectors. Another (more complex) approach would be to ask, as an empirical question, how particular societies draw boundaries between functions and what they place within each. For example, market and non-market ‘economies’ are very different things. ‘National security’ and ‘human security’ define ‘the security function’ in different ways.
How functions are allocated to actors is also structurally important. For example, classical liberalism, embedded liberalism, and neo-liberalism differently allocate economic activities among states and markets. Feudal lords leading vassals and retainers, mercenaries employed by dynastic kings, and citizen armies defending territorial states produce rather different systems of security politics.
What Waltz says of domestic polities is equally true of international societies: ‘Political actors are formally differentiated according to the degrees of their authority, and their distinct functions are specified’ (1979: 81). Structural analysis in IR must ask, as an empirical question, who performs which functions under what authority — rather than pretend that stratification and functional differentiation do not exist.
Segmentation (unit differentiation)
The Waltzian language of ‘units’ seeks to strip reference to the character of the actors from the account of structure. Waltz claims that ‘the logic of anarchy obtains whether the system is composed of tribes, nations, oligopolistic firms or street gangs’ (1990b: 37; cf. Waltz, 2000: 10). In fact, however, different types of units act and interact in systematically different ways, making segmentation or unit differentiation — the definition of the identities and boundaries of groups — structurally central.
Waltz acknowledges this. ‘Systems populated by units of different sorts in some ways perform differently, even though they share the same organizing principle’ (Waltz, 1990b: 37). Of special importance in international systems is the definition of dominant unit type(s). ‘International political structures are defined in terms of the primary political units of an era, be they city states, empires, or nations’ (Waltz, 1979: 91, emphasis added; cf. Waltz, 1979: 94). For example, modern international relations is typically associated with the modern state. Robert Gilpin (1981: 40–42) even defines a change of system (in contrast to changes within a system) as a change in dominant unit type. (Standard anthropological and archeological typologies also define societal type by units [e.g. band, tribal, chiefdom, and state societies].)
‘State’ in the broad sense of polity (‘unit’) is a variable rather than a constant. In the narrow sense of a particular type of political entity, separate from both society and the ruler, most historical international systems have not been states systems. And even among states systems, the type of state matters. For example, 17th-century dynastic states, 19th-century national states, and 20th-century territorial states — not to mention the ‘warring states’ of ancient China or the city-states of Classical Greece — behaved in fundamentally different ways.
International societies have also been segmented by civilization (civilized versus barbarian was a central distinction in imperial China, Classical Greece, and modern Europe alike) and by religion (e.g. Christendom and the heathen world or the Dar al-Islam and the Dar al-Harb). Furthermore, segmentation need not follow a single principle. Medieval Europe is particularly interesting for its multiple cross-cutting segmentations, based on birth (noble versus common), function (the three orders of knights, priests, and workers), lineage (dynasty), feudal allegiance, type of polity (e.g. kingdoms, principalities, duchies, communes, free cities, bishoprics, electorates) and the biblical distinction between the things of God and the things of Cesar. Globalization similarly involves the multiplication and diversification of sub-national, transnational, and international actors.
We can also note that segmentation provides an alternative to anarchy (ordering principle) for defining international systems. International societies are societies whose members are societies (cf. Rosenberg, 2006: 308).
Polarity
I have already suggested that ‘distribution of capabilities,’ Waltz’s third element of structure, is logically a dimension of stratification. Furthermore, treating ‘distribution of capabilities’ as the number of great powers in a system ignores most of the distribution — contributing to mainstream IR’s misspecification of stratification and functional differentiation.
System polarity, however, is an important dimension of differentiation. For example, polarity distinguishes oligopolistic from free markets. Or consider the differences between two-party and multiparty democracies.
Without addressing IR’s substantial literatures on polarity, here I will merely note that the vital issue of whether the system is highly polarized — that is, whether it has great powers — is rarely addressed. For example, the Hobbesian state of nature is explicitly a system without great powers (all the actors are materially and morally equal) (Leviathan, ch. 13 par. 1–2, ch. 14 par. 4) Hobbes, (1986 [1651]) and yet widely assumed to be a model of wide applicability. Conversely, it is often assumed, as illustrated by quotes above from Waltz and Mearsheimer, that international systems are great power systems — a clearly inappropriate assumption for a general framework of the elements of structure.
Material content: Geography and technology
Waltz is often depicted as presenting a material conception of structure. In fact, though, neither ordering principle nor functional differentiation is centrally material. And in most structural accounts, anarchy — which, it bears repeating, is not a material feature 10 — does most of the analytical work.
Having considered polarity separately and understanding stratification as material no less than normative — the distribution of capabilities and the distribution of material resources often differ significantly — I suggest thinking of material content primarily in terms of geography and technology (cf. Deudney, 2007).
Physical location — spatial differentiation — often is an important feature of social positions. ‘Geopolitics’ is a classic IR example. Nineteenth-century European empires were structured around a differentiation between metropolis and periphery that had a central spatial dimension. Much talk about globalization references spatial reconfigurations.
The structural impact of technology is evident, for example, in the offense–defense balance. 11 Furthermore, geography and technology often operate interactively. For example, the stopping power of water depends on the form of transport to be stopped. The technological capability to deliver nuclear weapons by plane and missile eliminated the traditional geographical protection of the United States and circumvented the Soviet Union’s hard-won land buffer — with decisive structural effects.
Normative content: Rules, roles, and practices
Finally, we turn to norms and institutions, the normative complement to geography and technology. Waltz, despite excluding norms and institutions from his tripartite conception, argues that ‘structures may also be changed by imposing requirements where previously people had to decide for themselves’ (1979: 108, emphasis added). ‘Within an international order, risks may be avoided or lessened by moving from a situation of coordinate action to one of super- and subordination, that is, by erecting agencies with effective authority and extending a system of rules’ (1979: 111).
Norms and institutions are major components of social positions in domestic societies. Therefore, they must be part of any adequate analytical framework — even if they happen to ‘drop out’ in many, most, or even almost all international systems. And where they do drop out, that is an important feature of that society’s structure. This is how Waltz, properly, treats functional differentiation. (He insists on its categorical significance, denying only its empirical existence in anarchic orders.)
Even if norms and institutions are in most international systems ‘limited in extent and modified from what they might otherwise be’ (Waltz, 1986: 336), this is a contingent empirical fact, not a necessary consequence of anarchy. Furthermore, ‘limited’ need not mean insignificant — weaker does not mean powerless; less does not mean none — and ‘modified’ suggests that we should explore their character rather than ignore their presence.
Glenn Snyder similarly labels norms and institutions ‘structural modifiers’ because they are ‘structural in their inherent nature but not potent enough internationally to warrant that designation’ (1996: 169). Potency, however, is an empirical issue. Because norms and institutions arrange the parts of social and political systems, they belong in any framework for structural analysis. They are, as Snyder says, inherently structural.
‘Norms and institutions’ range over a wide field. I suggest a two-dimensional categorization: they are made up of rules, roles, and practices that in international systems principally concern international legitimacy, internal legitimacy, cultural values, and foundational regulative practices.
Rules, roles, and practices
‘Rules’, ‘roles’, and ‘practices’ are very different types of norms and institutions. Rules (and institutions understood as persistent sets of formal and informal rules) have been extensively examined in contemporary IR — although rarely as elements of structure. Roles and practices, although less familiar, are no less important.
Roles — behavioral norms and expectations associated with social positions — are so vital to arranging and positioning actors that Peter Blau defines a social structure as ‘the differentiated social positions or statuses in a collectivity and the role relations of people as influenced by their statuses’ (1975: 131). Roles, however, have received little attention in contemporary IR other than in fairly peripheral areas of foreign policy analysis and in Wendt’s cultures of anarchy (1999: ch. 6).
Note, though, that great power is a role — although rarely explicitly considered as such. (Great powers, as China is learning, are expected to behave in particular ways.) And roles, when we look for them, prove to be regular features of international societies. For example, Britain’s role as a balancer gave the classical European balance of power a distinctive character. A hegemon differs from a great power by role not power. Friend and enemy are roles, the substance of which changes with time and place.
Roles, beyond their intrinsic significance, are of special interest as points where structure and agency meet. They also capture some of the structural dimensions of ‘identity,’ providing a bridge to bodies of work often not thought of as structural.
The language of ‘social practices’ highlights features such as inarticulate dispositions, habits, customs, ontological and ethical assumptions, and standard operating procedures, 12 the significance of which IR is only beginning to appreciate (see Adler and Pouliot, 2011; Pouliot, 2010). Social ‘institutions’ are not just a matter of — often not even primarily about — ‘rules,’ in the sense of (formal and informal) directives that are followed more or less intentionally. They draw on, instantiate, and help to (re)produce complex ways of seeing, doing, and being. Pierre Bourdieu (1990 [1980]: 103–104) talks of ‘practical sense, which does not burden itself with rules or principles (except in cases of misfiring or failure), still less with calculations or deductions,’ resulting in a ‘kind of acquired mastery, functioning with the automatic reliability of an instinct.’ For example, the practice of marriage goes well beyond rules about who may marry, how, and with what consequences. The practice of diplomacy is only hinted at in its rules.
The constitutional structure of international societies
Rules, roles, and practices, which both constitute actors and regulate their actions, can be seen as operating principally in four domains that together make up what I will call the ‘constitutional structure’ of international societies:
Principles and practices of international legitimacy. Membership in international society, and its associated rights and duties, is obviously structurally central (and helps to provide content to segmentation, stratification, and functional differentiation). For our purposes here, Daniel Philpott’s ‘three faces of sovereignty’ pose the essential questions. ‘Who are the legitimate polities? 13 What are the rules for becoming one of these polities? And, what are the basic prerogatives of these polities?’ (2001: 12).
Principles and practices of internal legitimacy. How actors, and especially the dominant actors, understand and justify their rule — what Christian Reus-Smit calls ‘the moral purpose of the state’ (1999) 14 — also typically has important structural effects. For example, dynastic states fight over matters of kinship and inheritance, as in the Wars of the Spanish and Austrian Successions, whereas national states fight over the fate of co-nationals and creating or maintaining nation-states. Ideas of equal human rights and democratic self-rule (or at least self-determination) helped to undermine first European overseas empires and then the Soviet Union and its informal empire. Welfare states approach international economic relations in a distinctive fashion.
Hegemonic cultural values. International societies often are embedded in broad foundational normative complexes. Ideas of the Middle Kingdom and the mandate of heaven helped to structure Chinese international society across the rise and fall of several dynasties. Medieval Europe was centrally shaped by principles of religion and hierarchy. The ‘Stanford School’ of ‘world culture theory’ argues that scientific rationality has fundamentally (re)shaped understandings of legitimate actors and action in modern international society (Meyer and Jepperson, 2000: esp. 106–108; Meyer et al., 1997: esp. 148–151; Thomas, 2010). Such values may be more or less hegemonic, penetrating more or less deeply in different places or parts of society. Counter-hegemonic values may also exist. Whatever form they take, though, they help to shape social positions and relationships.
Foundational regulative practices. How differentiated functions are performed is also structurally significant. Barry Buzan (2004: ch. 6) and Kalevi Holsti (2004) have advanced interesting accounts of primary or fundamental international institutions, focusing on the practices of modern international society. We might also try to capture this structural dimension by identifying common functional domains. For example, to Hedley Bull’s (1977) ‘primary values’ of life, truth, and property (which regulate the use of force, making agreements, and the possession of things), we might add practices that regulate conflicts, facilitate communication, and allow collective action.
The multidimensionality of structure
Mainstream IR not only has an inadequate conception of the substantive elements of social and political structures, but it misrepresents their operation.
Systemic and analytic (reductionist) analysis
Differentiation and structure are features of systems, which, as Waltz notes, are ‘both complex and organized’ (1979: 12). Systems produce their effects through the often-nonlinear interaction of interrelated parts and relations. Novel phenomena emerge that could not be predicted from the separate parts. Large changes in some variables have little or no effect while seemingly small changes may tip the system into a fundamentally different configuration or lead to its breakdown. Examining the parts individually — employing analytic (reductionist) methods — thus usually yields inadequate, and often misguided, explanations.
In analytic (reductionist) explanations, ‘the whole is understood by studying its elements in their relative simplicity,’ and one examines ‘the attributes and interactions of two variables while others are kept constant’ (Waltz, 1979: 39, 12). This is how mainstream IR treats anarchy: the absence of hierarchy explains certain properties and behaviors of the units, conceptualized in the simplest possible terms as undifferentiated polities, with everything else held constant. As the resulting anarchic international orders are neither complex nor organized, there is no need for, or actual recourse to, a system or its structure. ‘Structural’ international theory is largely a linguistic pretense.
Systems are composed of interdependent variables. Mainstream IR, however, treats anarchy as an independent variable. Leaving out — abstracting from — some of several independent variables may prove a fruitful simplification. Such ‘simplifications,’ however, when dealing with multiple interdependent variables are likely to leave us unable to determine the source or character of system effects.
Mainstream ‘structural’ explanations in IR are cast at the level of ‘the international system.’ They conceptualize that ‘system,’ however, in non-systemic (reductionist, analytic) terms. Systemic analysis is distinguished by what does the explaining — complex organized systems — not level of analysis. It simply is not true that ‘theories that conceive of causes operating at the international level … are systemic’ (Waltz, 1979: 18). Analytic approaches may operate at all levels of analysis. Mainstream ‘structural’ IR in practice conceptualizes anarchy and polarity analytically, as independent variables with (Humean) causal effects (that happen to operate at the level of ‘the international system’).
Much can be learned from independent-variable analyses of ‘system-level’ influences on state behavior. They should not, however, be confused with systemic/structural analyses. Explanation in terms of wholes (systems and their structures) and of separate parts are radically different types of explanation — as we can see clearly in the case of polarity.
Polarity
IR typically assumes the effects of anarchy, holds everything else constant, and then attempts to trace the implications of a particular polarity. But if multiple structural elements interact complexly, polarity (alone) is likely to explain nothing of interest. And a particular polarity is likely to be associated with different behavioral patterns.
For example, Waltz argues that ‘bipolarity has been proof against war between great powers’ (1979: 182). The Peloponnesian War, however, took place in a bipolar system. The Punic Wars involved the two great powers of the 3rd century
Even the absence of war between the Cold War superpowers cannot plausibly be attributed to bipolarity (alone). Technology was also essential. ‘Nuclear weapons dissuade states from going to war more surely than conventional weapons do’ (Waltz, 1990a: 743). ‘States armed with nuclear weapons may have stronger incentives to avoid wars than states armed conventionally’ (Waltz, 1979: 174–175). The mutually reinforcing interaction of bipolarity and nuclear weapons, it seems to me, explains more than the sum of polarity and technology considered separately. (And adequately accounting for the difference between the bipolar worlds of Thucydides and Waltz requires recourse to other dimensions of structure as well, such as cultural attitudes toward hegemony.)
Waltz also argues both that ‘balancing is differently done in multi- and bipolar systems’ (1979: 163) — in bipolar systems only internal balancing is important — and that nuclear weapons make external balancing unimportant, because ‘nuclear forces do not add up’ (1979: 181). Had the Cold War world been non-nuclear, superpower alliances in Europe probably would not have been inconsequential. The Berlin Crisis probably would have unfolded in a different way. And there would have been no Cuban Missile Crisis. Again, the interactive effects of technology and polarity seem greater than the sum of the separate parts.
Standard discussions of the sustainability of unipolarity are similarly misspecified. Although anti-hegemonic balancing operated relatively effectively in the modern European international system, elsewhere and at other times unipolarity has proved robust, even normal (Kaufman et al., 2007; Watson, 1992). For example, a cultural predisposition to hierarchy facilitated hegemonic and imperial projects in what archeologists call the Chinese interaction sphere. The collapse of the Zhou dynasty in the 8th century
Polarity actually explains behavior only in conjunction with other structural elements. And the same is true of other dimensions as well.
Multidimensionality and ‘ordering principles’
I have in effect argued that mainstream IR in practice treats ‘anarchy’ not as one element of structure, but as a summary statement of a particular type of complex international order; an ‘ordering principle’ in the ordinary language sense of that term. Beneath the simple surface of Waltz’s tripartite conception we can see outlines of a complex, although implicit and radically underdeveloped, multidimensional conception of structure. Waltzian anarchic international orders, when we look carefully at the details, rest on and arise from the interaction of particular forms of stratification (the double stratification of states and great powers), functional differentiation (a separate international political realm managed by great powers), segmentation (a states system), polarity (two or a few — not one or many/no — great powers), technology (e.g. in defensive realism, that offense does not have a huge advantage), and rules, roles, and practices (e.g. sovereign statehood, no systemic restrictions on the aggressive use of force, and a hegemonic diplomatic culture centered on the autonomy and formal equality of the great powers).
Wendt’s cultures of anarchy also present a simple, one-dimensional surface — identity as an independent variable in rationalist, ceteris paribus models — above a more complex, and somewhat less obscured, multidimensional account. For example, the ‘role structures’ of enmity, rivalry, and friendship define membership criteria and constitute states with particular kinds of self-interest (Wendt, 1999: 291–295). But identity/culture (alone) does not have the extensive and determinate foreign policy and macro consequences Wendt attributes to it (1999: 262, 265–266, 282–285). As with Waltz’s ‘effects of anarchy,’ Wendt’s ‘effects of culture’ in fact arise from a much broader and more complex set of interacting structural forces. Amity, rivalry, and enmity — much like ‘anarchy’ — are best seen as ‘ordering principles’ that summarize a particular configuration of multiple structural elements.
Similarly, Buzan and Albert’s types involve multiple dimensions of structure being shaped by a dominant ‘ordering principle’ of differentiation. For example, segmentary societies have minimal stratification and limited functional differentiation. In fact, their segmentary and stratificatory societies correspond to Waltz’s anarchic and hierarchic orders, to which they add functional differentiation as a third type.
However we formulate it, though, international societies and their structures are not simple things made up of a few independent variables, but complex systems in which multiple interrelated elements hang together with a certain unity or coherence — and that coherence, and its effects, are the principal subject of structural analysis.
Diversity and change in international systems
This understanding requires abandoning the Waltzian pretense that almost all international societies share the same structure. To draw a biological analogy, ‘international society’ is not a species but a kingdom. Type approaches to differentiation specify (at least some of) the phyla of international societies. Dimensional approaches provide the criteria by which subordinate taxa are differentiated.
Structures typically persist over extended periods. Nonetheless, structural change, in any plausible sense of that term, usually is at least as striking as structural continuity — when we look for it. Type and dimension approaches to differentiation provide tools to identify structural differences, compare international societies, and study structural change.
The transformation of post-war international society
We now turn to two empirical illustrations that seek to demonstrate both the regularity of extensive international structural change and the analytical power of differentiation approaches. This section uses my six-element framework to examine the post-World War II era. The next section considers globalization.
Because only polarity varies in the Waltzian conception of international structure, IR’s standard structural story of post-war international relations boils down to bipolarity, with ideological rivalry treated as a non-structural adjunct. This reading, I will argue, obscures the profound and extensive structural transformation of post-war international society.
Decolonization, non-aggression, and territorial integrity
Striking changes occurred in all four dimensions of constitutional structure identified above. Principles and practices of international legitimacy were fundamentally altered. Although states remained the principal members of international society, membership rules were radically transformed (cf. Philpott, 2001: Pt 3).
Modern international relations, right up to World War II, was dominated by imperial states formed primarily through dynastic ties and military contests. Lesser powers retained their independence, and varying degrees of autonomy, largely due to the limited military reach of great powers or as an unintended consequence of their power balancing. The number of Europe’s independent polities was steadily reduced, from several hundred in the 16th century to a couple of dozen in 1900, while most of the rest of the globe was formally incorporated into, or semi-formally subordinated to, European empire-states. (The principal exception was the Americas in the late 18th and early 19th centuries, although genocidal state expansion and the destruction of indigenous polities remained the 19th-century norm throughout the region.)
Beginning with India (1947) and Indonesia (1949), however, decolonization swept across Africa, Asia, and the Pacific (and the still-colonized parts of the Americas). The 1960 Declaration on the Granting of Independence to Colonial Countries and Peoples (General Assembly Resolution 1514) codified the norm of ‘self-determination’ — Western colonial empires were to be dismantled and new states created following the principle of uti possidetis — which became central to international legitimacy. Processes of largely peaceful, norm-governed political devolution — with major violent exceptions in Indonesia, Vietnam, Algeria, and the Portuguese empire — increased the membership of the United Nations from 51 in 1945 to 76 in 1955, 117 in 1965, and 144 in 1975; that is, by half in each of the first two post-war decades and by a quarter again in the following decade.
Principles and practices of internal political legitimacy underwent a parallel transformation. ‘Democratic self-rule’ and ‘welfare and development’ became hegemonic, replacing principles such as monarchy, a history of imperial association, and laissez faire. The details and implications of these values certainly were contested. Liberal democracies competed with peoples’ democracies. Three worlds of development were defined by competing visions of welfare and how it was to be realized. The mix between internal democracy and freedom from external rule varied considerably from region to region and country to country. Nonetheless, the legitimacy of post-San Francisco states was quite different from that of early-modern dynastic states, the monarchies of the first half of the 19th century, and the empires of the early 20th century. Perhaps most dramatically, citizens replaced subjects, with profound consequences for the nature of the national interest.
Equality and reason replaced hierarchy, religion, and tradition as hegemonic cultural values — not just in a privileged white ‘Western’ core, but across the globe. Racial, ethnic, and gender equality came increasingly to the fore in all regions. In international relations, centuries of formal sovereign inequalities were largely swept away and sovereign equality ‘attained an almost ontological status in the structure of the international legal system’ (Kingsbury, 1998: 600).
Perhaps the most profound changes concerned the legitimate use of force. Classical international law recognized a ‘right of war’; the right of each sovereign to decide when its vital national interests were threatened and what measures to take to protect them. Furthermore, coerced territorial adjustments were not merely legitimate but the norm. Article 2, Section 4 of the United Nations Charter, by contrast, states that ‘All Members shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state.’ And these idealistic words, penned in a moment of post-war optimism, rapidly became firmly established practice with respect to territorial integrity.
At most, one state (Tibet) has been eliminated by force since the conclusion of World War II (cf. Fazal, 2007) despite the creation of dozens of weak states that in earlier eras would have been ripe for the picking. Even coerced territorial changes have been largely abolished. Before 1945, about 80 percent of wars led to territorial changes. The percentage dropped to 30 in the half-century after 1945 — and to zero after 1976 (Zacher, 2001). Territorial war, a characteristic practice of most historical international systems, was largely eliminated — with profound human and political consequences.
The realist maxim that ‘survival dominates other motives’ (Mearsheimer, 2001: 31) became largely irrelevant to most states, whose survival and territorial integrity were effectively guaranteed by international legal norms and supporting political practices. Most of the world moved, in a period of about three decades, from ‘Hobbesian’ to ‘Lockean’ anarchy (Wendt, 1999: ch. 6).
Other dimensions of change
If this does not count as fundamental structural change (as it does not in mainstream IR’s standard account), then we need a different conception of structure (which I have tried to provide). And these substantive normative changes were paralleled by and interacted with changes in other forms of differentiation.
Decolonization brought two important changes in unit differentiation (segmentation): the demise of formal empire, a venerable type of great power unit, and the dramatic proliferation of what Robert Jackson (1990) has called quasi-states (whose only substantial power resource is international recognition). Dozens of new weak states altered the distribution of material capabilities. Patterns of stratification were reshaped by the near complete elimination of formal sovereign inequalities.
International functional differentiation expanded dramatically. The Waltzian logic of unit sameness is driven by a Hobbesian fear of violent death. With that fear largely eliminated for most states, substantial space opened for the pursuit of absolute gains (cf. Waltz, 1979: 71; Wendt, 1999: 282). This facilitated new forms of cooperation and interdependence — functional differentiation — evidenced in the rise of international economic and security regimes and the emergence of ‘new issues’ such as the environment and human rights.
International economic relations became increasingly differentiated from international security relations — and of steadily growing importance, eroding the old distinction between ‘high politics’ and ‘low politics.’ In the West, ‘embedded liberal’ welfare states drove creation of a new international economic architecture (designed in significant measure to protect employment and social benefits). In the Second and Third Worlds, the focus on welfare and development took different forms, but had comparable impact on international relations.
These changes also interacted with — both facilitated and were facilitated by — the rise of non-state actors. A world marked by the declining utility of force and the rise of international functional differentiation had much more space for transnational and supranational actors. And their actions reinforced and helped to deepen functional differentiation and peaceful conflict resolution.
Material inequalities certainly remained central. Exploiting these inequalities, however, was differently enabled and constrained, as talk of ‘neo-colonialism’ and the Soviet ‘informal empire’ indicate. These structural changes were also related to the material interests of powerful states. For example, decolonization was facilitated by the fact that the United States and the Soviet Union did not possess overseas colonial empires. (The areas into which they had expanded were geographically contiguous and had become generally recognized as integral political subunits.) There was a technological dimension as well. For example, nuclear weapons facilitated a status quo orientation toward territorial boundaries and improvements in communications and transportation made indirect economic and political influence a viable alternative to direct rule. Practices like territorial integrity and self-determination, however, were in no way reducible to material and technological forces. And the fact that ‘underlying’ material forces expressed themselves in these particular forms was crucial to their impact.
To say that post-World War II international relations remained anarchic but became bipolar is neither untrue nor without insight. It does, however, miss — even obscure — a wide range of structural changes that fundamentally altered the post-war international order. 15 Although ‘sovereign states’ remained privileged actors, these were not the same sovereign states of earlier eras. And the ‘anarchic order’ they inhabited was profoundly transformed.
Globalization and differentiation
We turn now to ‘globalization,’ understood as the decoupling of social relations from sovereign states and their reorganization on increasingly deterritorialized bases. For mainstream structural theory, globalization involves no structural change: the international order remains anarchic and globalization has no special connection with polarity. This analysis, to me at least, vividly illustrates the need for a more intelligent conception of structure — which differentiation provides.
The preceding sections emphasized the multiplicity of elements of structure and their interactions. Here I focus more on differentiation’s contribution to comprehending change. I also use globalization to compare type and dimension approaches.
Globalization as functional differentiation
Let us stipulate a ‘late-modern’ baseline for comparison; say 1815 or 1848 to somewhere between the early 1970s and late 1980s. In type terms, late-modern international society was segmentary (Buzan and Albert, 2010: 319, 321, 332; cf. Waltz, 1979: 95 n.*). If globalization merely further complicates that segmentary order, then a types approach provides little insight (because it largely abstracts from change within types). If globalization is leading toward a functionally differentiated international society, though, Buzan and Albert’s framework offers a distinctive, and quite interesting, analysis.
Late-modern states were omnifunctional entities that aspired to perform all important functions for themselves (and their subjects or citizens). They increasingly asserted a monopoly not only on the legitimate use of force, but also on law, finance, social services, and even identity, striving to organize and control an ever-growing proportion of social and political life.
In a transition-of-types scenario, various non-state actors provide goods, services, opportunities, and protections that states previously provided (or that previously had not been provided). This initially occurs for varied, largely unconnected, reasons. As such changes become interrelated and reinforcing, though, the re-parceling of functions and authority expands and accelerates. A tipping point is reached, after which state-centric segmentary international society gives way relatively quickly to functionally differentiated global or world society.
This analysis is directional: functional differentiation generates cascading functional differentiation, which reshapes segmentation, stratification, and rules, roles, and practices. 16 This gives the account considerable bite. Why, though, should we believe globalization to be directional — especially in the absence of any account of driving mechanisms? 17 It therefore seems to me at the very least prudent to explore non-directional scenarios — which are best approached dimensionally.
A multidimensional approach to globalization
Dimensional approaches see change as normal, varied, and open-ended (rather than rare, radical, and directional). This suggests challenging the idea of a typologically coherent and relatively constant ‘late-modern’ baseline — and thus a very different perspective on globalization.
Change and variety in late-modern international society
Great overseas empires were built — and then torn apart. Large land empires were broken into smaller and more unitary nation-states. And states were transformed from composite polities with limited capabilities into unified national or territorial states increasingly entwined in the lives of their citizens. In addition, commercial and non-commercial transnational actors acquired new prominence in the later 20th century, as did regional and global intergovernmental organizations and international regimes of various sorts.
Truly national economies were created and consolidated in Europe in the first half of this period, driven by and driving the political consolidation of territorial nation-states. They also increasingly penetrated, and gradually integrated, the rest of the globe. But by the end of the late-modern period a new international division of labor pointed to a global, rather than international, economy.
The concert system instituted after Vienna reverted, within a couple decades, to a more familiar multipolar great power order. After World War II, however, functional and regional regimes increasingly overlaid heterarchic multiple-rankings on the classic two-dimensional great power stratification.
The preceding section discussed the far-reaching normative transformations of the post-war period. Nearly comparable transformations took place in the mid-19th century under the pressures of nationalism and industrialization. Geotechnical transformation was almost as striking. A European international society, whose members acted in only limited ways outside the continent’s boundaries, became a single, increasingly integrated international system that spanned the globe.
For certain analytical purposes, one might prefer models of ‘anarchic international orders’ or ‘segmentary international societies.’ The more fine-grained picture presented by a multidimensional approach to international systems and their structures, however, deserves at least equal attention.
A structural framework should not prejudge questions of continuity and change or consistency and diversity. The structural coherence of stable systems may lie not in approximating a typological ideal, but in the mutual adjustment of multiple elements driven by no particular teleology or overarching ordering principle. And even where a dominant dimension of differentiation does organize a society, a framework of the elements of structure is essential to drawing all but the broadest and most general substantive structural observations.
Thinking through globalization dimensionally
This view of late-modern international society suggests seeing globalization as a complex combination of the persistence, intensification, and transformation of existing patterns and trends. Rather than the unfolding of a typological ideal (such as ‘functional differentiation’ or ‘the logic of capitalism’), globalization appears as a contingent, path-dependent process arising from the interaction of multiple structural (and non-structural) forces. Dimensional approaches to differentiation provide a systematic framework for analyzing such cases by exploring the possible consequences of changes in individual elements and their interactions across the network of social positions.
For example, the continued growth of a truly global economy is likely to reduce the capabilities of many states to provide social welfare benefits to their citizens. But will guaranteed welfare provision decline, be redefined, or become shared with other actors? Which other actors? In what mix? How will this affect the privileged position of states in international society? What impact will it have on patterns of stratification? Treating these as open, empirical questions seems to me a fruitful analytical strategy.
Or imagine that neo-liberalism fails to establish its ideological hegemony; that the passing of ‘the Washington consensus’ was not just a temporary setback. One sort of globalization is likely if an economic cold war develops. Globalization will be very different, though, if, for example, a global embedded liberalism asserts itself — or if cosmopolitan welfarism begins to gain serious traction.
There are no answers in this kind of account. I am deploying an analytical framework, not a substantive theory. There is no specified direction to or driving force behind change. The framework does, however, provide a disciplined strategy for depicting dimensions and pathways of change, and possible ways in which that change will be shaped by the network of social positions. It helps us to ‘describe and understand the pressures states [and other international actors] are subject to’ (Waltz, 1979: 71).
As citing Waltz suggests, this is not very different from the mainstream approach. One cannot read off substantive theory from a typology of elements of structure or an analytical framework of differentiation. From IR’s tripartite conception of structure nothing more of substance follows than from my multidimensional framework — pace Waltz’s suggestion that it entails structural realism. Explanation or prediction is possible only when particular elements are combined in a particular fashion. A typology of elements of structure may delimit a range of structural possibilities. It does not imply any particular substantive theory — unless, perhaps, all systems have the same structure (which they do not).
A dimensional framework can also be used to add discipline to historical comparisons, which too often are impressionistic or focused on features taken out of context. For example, a neo-medieval (e.g. Bull, 1977: 254–256, 264–276) account of globalization singles out the absence of a dominant unit type and the presence of extensive functional differentiation and heterarchic stratification. But what happens when Christendom and hierarchy are replaced by an egalitarian global culture and universal human rights as an internal standard of political legitimacy? Or, how do the technologies of our era influence the ways functions are divided, how group actors are formed and aggregated, or the relations between individuals and authorities of various sorts?
Daniel Nexon suggests comparing globalization with early-modern Europe, especially because of the prevalence of composite states and the importance of transnational relations (2009: 19, 299–300). This was also a period when space and time were transformed and new forms of religious and cultural conflict emerged. I find particularly intriguing the idea of using a framework of differentiation to ‘unwind’ ‘Westphalia’ in order to think through possible ‘post-Westphalian’ futures.
Historical comparisons are, to be sure, ‘only’ analogies. But we should appreciate, and put to good use, the understanding and insight they provide — especially given the lack of a serious ‘stronger’ alternative. The mainstream Waltzian approach, for all its social-scientific pretense, amounts to analogical reasoning — employing a rather simplistic, ahistorical analogy at that. Rather than seek to depict accurately the actual structure of real international systems, ‘structural analysis’ has become a dubious exercise in determining how closely a system approximates a highly idealized one-dimensional type. A dimensional approach directs us back to dealing with the real complexities and varieties of international structures.
Conclusion
Both Waltz and Buzan and Albert assume that most historical international systems closely approximate a few ideal types (at most). I have argued that this assumption is without sound theoretical or empirical basis. Individual analysts or schools of thought may reasonably make such ‘bets.’ IR as a discipline, however, requires a multidimensional framework of differentiation if it wants to address the actual structures of (certainly many and probably most) real international systems.
Mainstream structural theory today is, at best, in the doldrums. After some early successes, structural realism is largely played out and of limited utility; even self-proclaimed structural realists offer what are in fact largely non-structural theories. 18 Among both realists and non-realists, structural analysis is used, if at all, primarily as a starting point; a device to pose problems that are explored through other lenses and with different analytical tools.
Buzan and Albert’s addition of functionally differentiated international societies certainly is an advance on the view that all international societies are anarchic/segmentary. I suspect, though, that ‘the logic of functional differentiation’ will prove to have the same severe limits as ‘the logic of anarchy.’ The extraordinarily high level of abstraction, combined with the reduction of system structure to a single dimension, renders such models of significant analytical value only for the broadest and most general ‘big’ historical comparisons. Such comparisons, although of great interest and value, cannot sustain structural theory in general, or even a particular progressive research program. For that we need to get down to multiple dimensions of differentiation.
Some might suggest that Waltz and Buzan and Albert illustrate structural theory’s inherent limits — which are much more constricting than its proponents realize. I have suggested instead that the problem is an inappropriately truncated conception of structure (whether in the form of a restricted list of elements or reliance on ideal types defined by a single dimension of differentiation). An accurate and insightful multidimensional depiction of the basic organization of a society and the fundamental relations among its principal parts promises not merely valuable but vital knowledge that is available from neither mainstream structural theory nor Buzan and Albert’s types.
Understanding structure as a multidimensional network of differentiated social positions opens up a new kind of mid-range structural international theory. It requires us to compare and contrast international societies across a fairly wide yet limited range of dimensions. Doing so will tell us not simply how closely a particular international society approximates some ideal type, but how that society is in fact structured — and how that structure matters for international relations within that system.
Footnotes
Acknowledgements
For comments on earlier drafts, I thank Mariano Bertucci, Barry Buzan, Claudia Fuentes, Chiung-Chu Huang, Raslan Ibrahim, Arturo Lopez-Levy, Jonathan Moyer, Hidemi Suganami, Matthew Weinert, and the editors and reviewers of this journal. I also received many useful comments at a workshop at the Wissenschaftzentrum Berlin für Socialforschung in December 2009. And I owe a special debt to Barry Buzan and Matthias Albert, who kindly shared multiple drafts of the paper that became the article from which mine takes off.