The Logic of Anarchy: Neorealism to Structural Realism

Barry Buzan, Charles Jones and Richard Little

New York

Columbia University Press

1993

9. Continuity and Transformation in the International System

 

We shall now relate the previous discussion about the intimate relationship between agent and structure to the issue of continuity and transformation in the international system. As we observed in the opening chapter to this Section, the competing theoretical perspectives adopted by Wallerstein and McNeill throw very different light on the meaning of continuity and transformation provided by the Neorealists. The simplest framework is Waltz's, who argues that an international system comes into existence only when states start to coact. From his perspective, it is possible to envisage an international system steadily expanding as an increasing number of states come into contact with each other. But it is also possible on the basis of this perspective to conceive of independent international systems coexisting simultaneously. In either event, as far as the Neorealists are concerned, transformation can be said to have taken place only if the ordering principle of anarchy is replaced by hierarchy. It follows that if the Roman Empire is conceived to be an independent system that did not coact with other political entities, then in terms of the Neorealist framework, the emergence of the empire can be discussed in transformational terms.

This is indeed how Wallerstein describes the formation and the collapse of the Roman Empire; but clearly Waltz does not picture the Roman Empire in this way. As far as he is concerned, the Roman Empire was constantly interacting with other political entities. Waltz's position is partially endorsed by McNeill, who postulates that there were, from a very early stage, links between the political communities right across the Eurasian land mass. The Romans, for example, are seen to have coacted directly with the Parthians and the barbarian tribes on the periphery of the empire's borders and indirectly with the Chinese. But McNeill's framework is also very different from the Neorealist one because he conceives of the Eurasian landmass as being occupied by four competing cultural areas separated by the tribes of barbarians. Structural Realism aims to take advantage of the insights offered by McNeill while acknowledging the importance of parsimony in the Neorealist position.

 

Distinguishing System and Structure

Central to our theoretical framework is the conception of an international system. Waltz argues that a system consists of a set of interacting units and a system-wide component--its structure--that makes it possible to think of the system as a whole (1979:79). This definition is more elaborate than most; a system is usually defined simply as "a set of interacting parts" (Reynolds 1980:185). It could be argued that Waltz's notion of a structure does no more than spell out the criteria that identify the interacting parts as a set. But as we have seen, this is not the case. By looking in detail at Waltz's conception of structure we have shown that it is inextricably tied to the conception of agency. Rational agents, according to Waltz, are constrained by the structure of the system to act in a way that reproduces the structure of the system. His conception of structure, therefore, points in the direction of structuration theory rather than system theory. Systems theorists argue that a system can be identified by the fact that "change in any of the components, or in the interactions among them, produces changes throughout the system, or its breakdown" (Reynolds:1980:186). Such a formulation does not require the existence of rational agents and it can apply to natural as well as social systems.

It is important to recognize, however, that in the first instance our initial conception of system coincides with system theory rather than structuration theory. Consequently, in a primitive international system, it is not necessary for all the political units to be directly linked with or even aware of each other. So, to take an extreme example, a chain of states, where each state has contact with, and knowledge of, only its immediate neighbors, nevertheless constitutes an international system. The point about such a system is that goods and knowledge can potentially pass from one end of the chain to the other, bringing about change in each state in the process. Provided that all the information or goods do move along the chain in this way, then all the states are, albeit indirectly, linked and thus constitute an international system. Waltz, of course, would deny that such an arrangement constitutes an international system. From his perspective, the component states are not responding to or reproducing an international structure. They are simply interacting with their neighbors and Waltz is quite clear that interaction of this kind does not generate a systemic structure.

It is evident, for example, that a balance of power could not be developed in this system. Although there could be a process of internal balancing, it would take place on a sequential basis, as one neighbor emulated another along the chain. Even this possibility presupposes that the states in the chain are all identically structured and are equally receptive to new developments. If one state hiccupped and distorted the change carried out by its neighbor, then it would be the distortion that would be emulated further along the chain. As in the game of Chinese whispers, in such a system it is likely that any beneficial change introduced by a component state would be rapidly modified or distorted as it was passed along the chain. There could, moreover, be no effective provision for external balancing. It follows that if, in line with the analogy of Chinese whispers, the absolute power of states was steadily magnified as one neighbor endeavored to emulate another, the state at the end of the chain would be stronger than every other neighboring state and it could, slowly but surely, and with increasing ease, absorb each successive and weaker neighboring state.

It can, of course, be argued that this chain model is unhelpful because social systems always do possess a structure. But such an assessment can be challenged on the grounds that units do interact in the absence of a systemic structure. The chain model, as a consequence, does not have to be treated as counterfactual. If the chain is turned into a circle, for example, it has more than a passing resemblance to the Kula Ring discussed in chapter 8. It can also be suggested that missing from the initial discussion was the recognition that the Kula Ring represents an anarchic system which lacks an anarchic structure. In contrast to the Neorealist model of an international system, therefore, it can be suggested that the tribal agents were not rationally endeavoring to ensure the survival of their political units; their rational interest was in trade. The survival of the anarchic tribal units can thus be seen as an unintended consequence of the complex transnational trading links associated with the kula. The resulting international anarchy, therefore, is a second-order and contingent consequence of the kula. As a result, it cannot be seen as a structure, unintended or otherwise, constraining the participants in the kula. On the other hand, if the units became centralized and generated a security dilemma with associated balance of power behavior, then it would be appropriate to suggest that the anarchy had developed structural properties which would necessarily constrain the behavior of the tribal units. (See Section III for a discussion of necessary and contingent anarchy). It would also be appropriate to suggest that the system has undergone a transformation.

Very far from being counterfactual, it could be argued that the chain model outlined above sheds considerable light on the early evolution of the international system, when levels of interaction capacity were low. It suggests that, in the first instance, it is appropriate to think of the early international system as one without a political structure. To help in the analysis of such a system, it is useful to draw upon the notion of sectoral analysis. The system can be seen to lack any political or strategic dimension. But it does contain a societal and economic dimension because it is the circulation of knowledge and goods that makes it appropriate to identify the political entities as forming an international system. But it can now be argued that such a system is necessarily subsystem dominant, as indicated in Section I, because the system itself lacks any overarching structure.

It is an anarchic system in which the component units interact in the absence of an anarchic structure. Waltz is unable to incorporate this idea in his theory because he assumes in the first place that the state is inevitably a repository of power. As the Kula and Carthaginian examples suggest, however, this is not necessarily the case. Waltz also fails to distinguish between different types of coaction. By focusing exclusively on coaction of a strategic and political nature, he inevitably postulates a structured anarchy.

Of course, states can develop as repositories of power, and when they do so, they will start to coact on a political and strategic basis. Under these circumstances, Neorealist theory can be seen to come into play. But in terms of the evolution of the international system, it must be recognized that coaction of this kind emerged initially on an isolated or local basis. In other words, within the unstructured international system there emerged regional or local subsystems with an identifiable structure. This structure constrained the behavior of the rational agents and the structure of the system was thereby reproduced. Neorealist theory applies to these subsystems whenever this structural condition can be seen to be in operation. Because these structured subsystems were part of the unstructured international system, however, unit behavior was necessarily affected by environmental interaction with the international system as well as by the structure of the subsystem. Inevitably, there would be occasions when structural constraints conflicted with environmental interaction. Under these circumstances, rational agents could endeavor either to minimize the effect of environmental interaction and maintain the structure of the subsystem, or respond to the environment and accept that the subsystem would lose its identity. The expansion of the Roman Empire at the expense of the Carthaginian and Hellenistic subsystems helps to throw light on these possible responses.

 

The Reproduction and Decline of the Roman Empire

It has already been suggested that it is unhelpful to think of the rise of Rome taking place in what Waltz identifies as a closed and structured international system. Roman agents were aware of a wider system, but their actions were systematically constrained only by the structure of the local subsystem within which they operated. As the interaction capacity in the Western Mediterranean subsystem rose, however, coaction with the neighboring Hellenistic subsystem increased. This eventually had the effect of dissolving the boundary that separated the two subsystems. As discussed earlier, the Macedonian Empire, conscious of the growth in power of Rome, allied with Carthage, while the Greek city states, disturbed by the intrusive power of Macedon, were more than ready to take up the offer of an alliance with Rome. With the development of these balance of power maneuvers, the two subsystems merged under a common anarchic structure.

It is important to note that the merger of subsystems can bring about structural transformation. The Romans quickly discovered the desirability of taking advantage of the existing international society in the Hellenistic subsystem. In the West, where the behavior of the barbarian tribes was unconstrained by any established diplomatic or political structures, the Romans had no alternative but to use brute force to establish control. In the East, by contrast, the Romans quickly saw the benefits to be derived from the existing diplomatic rules that helped to sustain the patron-client relationships which prevailed in the area. In the Hellenistic subsystem it was possible for the Romans to use power rather than force to establish a hierarchical structure in the subsystem. As Luttwak notes, "The rulers of Eastern client states and their subjects did not have to see Roman legions marching towards their cities in order to respond to Rome's commands, for they could imagine what the consequences would be" (1976:32).

Although the system of client states in Asia may have come into existence for essentially pragmatic reasons, Roman agents, realizing the effectiveness of the system, came to use it as the basis of their strategy for organizing and defending the empire as a whole. As a consequence, there was no demarcated and defended frontier at this time. Instead, defense against external enemies was left to client states and beyond them to client tribes on the very periphery of the empire. Roman troops were retained within a core region where the possibility of internal threats to the empire still persisted. But these troops also served as a vehicle of power and force. They could be sent to the periphery of the empire to bring new territory within its ambit. They could also be used to give assistance to the troops of client states unable to cope with internal or external threats.

The concentration of troops at the center of the empire provided a source of power. Fear of Rome's troops served to keep client states and tribes compliant. The clients recognized that Rome could not only provide assistance if they were in danger of being overwhelmed but also annex the territory if a client stepped out of line. During the era of expansion, therefore, the Roman troops operated as mobile striking forces and were concentrated along the major routes that led outward to the unconquered territory and inward to the areas of potential unrest. The routes outward were referred to as limes and they ran perpendicular to the secured territory. Underlying this strategy there was the assumption that the empire would continue to expand until it embraced the world. Mann asserts that Tacitus believed that any Roman leader who did not actively promote the expansion of the empire "was guilty of criminal irresponsibility" (1986:177). The Romans, therefore, very effectively managed to incorporate units from the unstructured anarchy as well as units in anarchically structured subsystems into their hierarchically structured subsystem. The empire provides a clear example of subsystemic transformation taking place in the context of the unstructured anarchic international system.

While the Roman Empire was expanding there was a constant interaction between the empire and political communities on its periphery. Civil conflict and determined external opposition, however, were to pose major threats to the hierarchical structure of the system. These two threats played a major role in bringing about a shift in the strategy designed to ensure the survival of the empire. As the opportunities for expansion dwindled, a strategy capable of defending a delimited empire evolved. Symptomatic of the change was the redefinition of established terminology. Instead of denoting a route to an as yet unconquered land, the term "limes" now came to identify the defense barrier enclosing the established territory of the empire. Client states were absorbed and the empire came to be demarcated by defended borders, by deserts and rivers and seas.

Luttwak insists, however, that it is a mistake to see the Romans suffering at this time from some kind of irrational Maginot mentality. He argues that the defense tactics adopted by the Romans at that time can be understood in terms of a rational response by agents confronted by a new security situation which generated new security goals. In the earlier phase, it was intended that the Romans could, ultimately, maintain the security of the Empire, although the unguarded borders meant that in the frontier zones insecurity was endemic. Areas could readily be attacked before assistance could be given. In the second phase, however, there was a demand for what Luttwak calls "continuous security" for life and property (1976:78). The aim now was not to rely on non-Romanized clients. It was intended instead that there should be a sharp divide between those who lived within the empire and barbarians who lived outside. Within the boundaries of the empire there was economic development, urbanization, and political integration. These activities could not go on in the absence of guaranteed security.

In the second phase, therefore, Roman strategy embraced two potentially contradictory requirements. On the one hand, as in the earlier period, there was a need for large striking forces that could deal with major attacks on the empire. It was still essential that the Romans could ensure the ultimate survival of the empire against any form of external attack. On the other hand, it was now necessary to reinforce this requirement with a capacity to prevent isolated raids across the frontiers, disrupting day-to-day security of the population living in the frontier zones. To deal with this kind of threat it was necessary to disperse the troops around the frontier zones. Luttwak argues that the Romans developed a frontier strategy designed to overcome potential conflict between these two very different requirements (1975:75). In the first place he infers from the record that the population living within the empire was fully integrated and that there was no attempt to extend the frontier if in doing so it would embrace a population that would be difficult to Romanize (1976:96).

The Romans then developed a multifaceted approach to defending the frontier that allowed them to deal with low level threats in an economical fashion: in the areas where there were natural barriers, for example, the emphasis was on surveillance, by means of watch towers and signal stations; in desert areas, where the population was concentrated into oases, defense was concentrated instead on the areas of habitation; in more populated regions the Romans relied on the fortified frontier. The fortification did not simply consist of a physical barrier. It also embraced the means for effective surveillance and good communication routes on either side of the barrier to permit rapid movement and concentration of forces. The essential aim of this strategy was to curb environmental interaction. It was being dictated not by the structure of the international system, because there was no structure, but by the desire to eliminate contact with the system. The agents of the state devised a strategy, in other words, designed to reproduce a hierarchical structure in the context of an unstructured anarchic system.

The ultimate collapse of the empire in the West and its survival in the East reveals very clearly the dangers of restricting the analysis of how the state reproduces itself to the idea of a grand strategy. When attention is focused on the division between East and West, it becomes apparent that there were major differences in the way that the two sectors were reproduced in the international system. A comparison can help to explain why the Western sector collapsed while the Eastern sector survived for another thousand years.

Once the comparison is made, it becomes apparent that the familiar explanation, which links the collapse of the empire to the waves of barbarians who swept into the empire, is inadequate. The barbarians had always existed on the other side of the frontier; they were not a new phenomenon. Moreover, the Eastern Empire was also confronted by enemies, among whom were not only the barbarians but also the Persian Empire--a much greater military threat. Indeed, Goffart has gone so far as to question the whole idea of a barbarian invasion. He argues that "Hardly anything has done more to obscure the barbarian question than the talk and images of wandering peoples tirelessly battering down the Roman frontiers and flooding into the empire" (1989:10). He denies, moreover, that the idea of the barbarian invasions represent "a direct reflection of the experience and thinking of those immediately involved in the events" (1989:113). Our image of the decline of the Western Empire is seen to have been largely dictated by Medieval historians. By contrasting the fragmentation of the Western Empire with the persistence of the East, it becomes possible to give a clearer picture of what happened in both sectors.

From the onset of the Empire, the East was more urbanized than the West, with its wealth coming from trade and industry. The agricultural base was also different in the two sectors. In the East, land was owned by independent peasants; in the West it was controlled by the wealthy aristocracy who relied primarily on slave labor. Once the empire stopped expanding, however, slaves became an increasingly expensive commodity. The population decline, which had begun in the third century, also had a bigger impact on the less populated West. Wealth, derived from agriculture, decreased during this period; at the same time, the taxes to pay for an enlarged army increased. By 350 A.D., the land tax had increased three times within the memories of the oldest taxpayers. While the peasant farmers in the East were able to raise enough money from the sale of corn to pay these taxes, the peasant farmers in the West were unable to do so (Brown 1971:36). Anderson notes how the free tenants in the West "fell under the 'patronage' of great agrarian magnates in their search for protection against fiscal exactions and conscription by the State, and came to occupy economic positions very similar to those of ex-slaves" (Anderson 1974:94). The effect of taxation in the West, therefore, was to concentrate wealth even further into the hands of the rich. In the East, by contrast, wealth was more widely distributed and the peasants were able to preserve their independence.

The attitude to the empire in the East and West also increasingly diverged. In the East, where there had always been a tendency to deify the rulers, the provincials became committed "Romans" in late Antiquity. Brown argues that the provincials felt this loyalty "not through the brittle protocol of senatorial or civic institutions, but directly--by falling on their knees before statues and icons of the emperor himself" (1971:42). Christianity reinforced this strong sense of identity in the East where Brown argues there could be observed "violent waves of xenophobia and religious intolerance." (1971:111) In the West, on the other hand, Rome was still a semi-pagan city in the year 400 and the aristocracy subscribed to pagan rather than Christian beliefs.

While it is thus unsurprising to find that the Western Empire was much less united than the East, it should be noted that the West had always been much less stable. The difficulties can be traced back to the Republican period when the Western aristocracy had been unwilling to pay the legionnaires adequate compensation for the victories they won for Rome. As a consequence, the soldiers gave greater loyalty to their generals than to the state. As Anderson notes, "soldiers looked to their generals for economic rehabilitation, and the generals used their soldiers for political advancement" (1974:68). Although the reforms of Diocletian and Constantine improved the situation, the West, "with its strong armies guarding Britain and the Rhine frontier, remained a point of instability, a springboard for generals with the ambition to make a grab for power" (Goffart 1989:19).

Goffart goes on to argue that a very different attitude to the question of security began to develop in the two halves of the empire. He suggests that the dangers posed by the barbarians had to be measured against the instability threatened by a strong army in the West. What began to happen was a conscious decision to transfer military control to Gothic, Frankish, and Vandal chieftains in preference to strengthening the army in the West. In earlier centuries, the Romans had always endeavored to drive the barbarians beyond the frontiers of the empire. But beginning in 382, when the Romans first negotiated a settlement of this kind, the main instrument for ending barbarian aggression was to "grant the offending tribe an area of settlement within imperial territory" (Goffart 1989:14). By 450 a set of autonomous Gothic, Vandal, and Burgundian districts had been established in the Western Empire.

Like Goffart, Brown insists that the barbarians did not engage in perpetual or destructive "organized campaigns of conquest." He views them instead as "gold rush" immigrants from underdeveloped regions who were attracted by the wealth of the empire. But Brown goes on to argue that in the East there was a greater sense of the empire's vulnerability. The "East Romans came to learn that their empire was one state among many, in a world that had to be scanned anxiously and manipulated by adroit diplomacy" (1971:139). When confronted by the Visigoths the East responded with a "combination of force, adaptability and hard cash" (Brown 1971:124). By contrast, the aristocrats in the West preferred to work in cooperation with the barbarians. As Brown notes, "The idea of a united western empire was increasingly ignored by men who genuinely loved the smaller world of their province" (1971:129). The notion that the Western Empire was engulfed by a barbarian invasion is, from this perspective, a historical invention of a later age.

Having explored the rise and fall of the Roman Empire, it is now possible to conclude this Section by spelling out the relationship which exists in Structural Realism between transformation, continuity and deep structure in the international system.

 

Transformation, Continuity, and Deep Structure

The aim of this Section has been to demonstrate that our definition of deep structure as incorporating both the organizational principle of anarchy and the differentiation of units is compatible with the theory of structuration that has been advanced as a way of resolving the problem of structure and agency. But it has also become apparent that the consequences of introducing structuration theory into the study of international relations are both complex and profound.

Three main points about deep structure can be drawn out from our discussion. The first relates to the importance of recognizing that there is an indissoluble link between actions that reproduce the deep structure of the state and those that reproduce the deep structure of the international system. Our emphasis on structure and agency forces this conjunction to the surface. Human agents mediate between the state and the international system. The structures defining the state and the system are constituted or reproduced by the practices established and implemented by these agents. For Neorealists, the link between state and system is indissoluble because the same set of practices are involved. The central task of any political leader is to ensure that the state will survive against internal and external threats. But the leader's actions, argues Waltz, unintentionally reproduce the deep structure of the international system. This line of argument can be consistently maintained because Neorealists identify an invariant relationship between the structure of the state and that of the international system. In reproducing hierarchy within the state, its agents will always be simultaneously and unintentionally reproducing anarchy within the international system.

Structural Realism breaks the extraordinary coherence of this position by introducing the idea of differentiation into the deep structure of the international system. The effect is to open up the idea of divergent political units within it. As Ruggie notes, in the Medieval period the major units in the international system "were known as civitates, principes, regni, gentes, and republicae, the common element among them, the idea of statehood, not yet having taken hold. To these must be added cities, associations of trades, commercial leagues, and even universities, not to mention the papacy and the empire--all of which, for some purposes, were considered to be legitimate political actors" (1986:155). The important consequence which follows the decision to locate both the principle of anarchy and functional differentiation in the deep structure is that the actions reproducing different kinds of political actor must, according to Neorealist logic, necessarily affect the nature of the international arena which is simultaneously generated by these actions. It follows that if the actions required to reproduce the United States in the contemporary international system involve the penetration of other political units, as Strange suggests, then this must affect the way the international system is being reproduced.

The second point directly follows from the first. It relates to the idea that cooperation and rules between states intentionally reproduce the deep structure of the international system. It can be suggested that an international system in which the deep structure is reproduced intentionally will take a very different form from the Neorealist one which is reproduced unintentionally. The difference corresponds to the distinction drawn by Bull and others between an international system and an international society. It is unlikely that there has ever been a structured international system which has operated at either extreme but it can be asserted that the modern international system is increasingly being reproduced on an intentional basis. Our sectoral analysis has an important bearing on this issue. The potential for rule formation extends across all the sectors. It may, in fact, be possible to discuss the reproduction of an international system by focusing on the political and strategic sectors. But to understand the reproduction of an international society, it is necessary to embrace every sector. Rules have been established between states to regulate activity in all the sectors and they cover everything from coffee prices to narcotics trafficking. In the process of performing specific functions in all these sectors, however, these rules also simultaneously help to reproduce the state and the international system/society.

Attention has been drawn to game theory to illustrate this point because of the suggestion that a common logic explains both conflict and cooperation. It is apparent that this argument does not neatly square with the Neorealist claim that the logic of anarchy drives states to pursue competitive policies. A simple matrix (figure 9.1), representing a two state system, reveals the inadequacies of the Neorealist line of argument.

The Neorealists assume the dominance of a strategy where interdependent actors seek to avoid their worst option: the hegemonic outcome where a subordinate state adopts a cooperative strategy, while the dominant state pursues a competitive one. If such a situation persists, the logic of anarchy would give way to that of hierarchy. To avoid this possibility, Neorealists insist that states have no alternative but to pursue mutually competitive strategies, and thereby generate a competitive anarchic system. But the matrix makes clear that there is an alternative outcome, whereby states cooperate and generate a cooperative anarchic society. It is now being recognized that the structure of an anarchic society has a very different form and logic to that of an anarchic system. In an anarchic system competition drives states to reproduce their autonomy, and in the process, reproduce the structure of the anarchic system. But in an anarchic society, the states are sovereign units and they are reproduced by the process of mutual recognition and common practice. The practices associated with sovereignty simultaneously and intentionally reproduce the anarchic society. Although it is often assumed that anarchy is incompatible with cooperation and interdependence (Milner 1991), it is now beginning to be acknowledged that the logic associated with competitive, autonomous states operating in an anarchic system needs to be distinguished from the logic associated with cooperative sovereign states operating in an anarchic society (Wendt, forthcoming).

The third and final point concerns the relationship between continuity and transformation in the international system. For Neorealists the question scarcely arises. As far as they are concerned, the international system has persisted throughout world history. It follows that their theory does not need to make provision for the emergence or the fragmentation of a world empire. A transformation of this kind can be depicted within the Neorealist framework in terms of a gestalt switch between hierarchy and anarchy, but such a switch is not considered to have taken place and is, as a consequence, left unexamined. Structural Realism introduces the idea of international subsystems that exist within the context of the international system and we have attempted to work out the implications of such a distinction. We can draw on the idea of structure to distinguish between system and subsystem dominance.

In cases of subsystem dominance, the international system lacks a political structure and thus cannot influence the behavior of the units in the system. The subsystems, by contrast, are defined by a political structure that can take the form of either a hierarchy or an anarchy. Transformation can occur within the subsystems but not in the international system itself. Transformation, therefore, can be seen to occur within the context of continuity. This situation changes only when a globally structured international system emerges. Identifying exactly when a fully global international system did come into existence is an important but as yet unfulfilled task. Establishing the answer would require both theoretical work (on the criteria distinguishing subsystem from system dominance), and empirical investigation (into levels and types of interaction, and responses to them). It could be suggested that such a system did not emerge until as recently as the end of the nineteenth century.