THREE APPROACHES TO SYSTEMS WITH PREEMINENT POWERS

Mainstream approaches to systems with single, preeminent powers fall into one of two frameworks—although scholars often draw upon both in their substantive analysis (Ikenberry 2002, 8–10). The first, unipolarity, assumes that states are unitary actors operating within an anarchical environment. Some see unipolar systems as inherently unstable: they argue that the concentration of power in the hands of a single political community eventually triggers counterbalancing behavior (e.g., Waltz 2000). Others believe that unipolar systems are stable: because no state or coalition of states will be able to challenge the preeminent power, few are likely to even attempt to engage in balancing behavior (e.g., Wohlforth 1999).

The second, hegemonic-order theory, covers a broader array of approaches, including hegemonic-stability theory, power-transition theory, and constitutional-order theory (a distinctive subtype of hegemonic order). The fate of hegemonic systems depends largely on the interaction of three factors: the ability of the hegemon to sustain its economic, military, and technological leadership; the degree to which potential challengers perceive themselves as benefiting from the existing hegemonic order; and the propensity for hegemonic overextension. Hegemonic-order theorists tend to treat states as unitary actors, although their vision of hegemonic systems involves elements of both anarchy and hierarchy (e.g., Gilpin 1981, 26–27, Ikenberry 2001).

Given the intensity of recent debates about “American Empire,” one might expect that “empire” would constitute a third framework for analyzing systems with preeminent powers. Since the publication of Michael Doyle's (1986) seminal Empires, however, mainstream international-relations analysts displayed minimal interest in developing a spare set of propositions concerning the dynamics of empires. Only a few scholars tackle aspects of imperial dynamics in the context of international-relations theory, and they seldom focus on the broader question of how those dynamics compare to unipolar and hegemonic orders (e.g., Cooley 2005; Spruyt 2005).

One reason for this lacuna stems from the continued prevalence of the states-under-anarchy framework in international-relations theory. We analyze unipolarity and hegemony as structural arrangements, but we often treat “empire” either as a question of the domestic organization of units operating under anarchy or as a particular outcome of state expansion. Predominant accounts of empires in mainstream international-relations theory, for example, often assess why states opt for imperial expansion (e.g., Lake 1996; Snyder 1991) or analyze the broader conditions that facilitate the rise of empires (e.g., Gilpin 1981; Hui 2005).

Many popular assessments of the question of American empire reflect this focus. The Bush administration, as we have seen, claims that the United States is not an empire because it eschews territorial conquest. Kagan (2002, 16) similarly argues that “American policy makers have a fundamentally different goal from Roman emperors” in that they seek peace and prosperity among sovereign states rather than the annexation and direct domination of independent polities. But the basic questions raised by the American empire debate concern the degree to which relations between the United States and other polities have imperial characteristics. In other words, it calls attention to imperial orders as multiply realizable structural arrangements. Imperial orders may be unplanned or planned; they may emerge through conquest or voluntary amalgamation. They may, in turn, involve a wide range of ideologies and norms (Lake 1996, 8; Motyl 1999, 133–36). Although the motives of political elites and the causes of imperial expansion surely matter a great deal for the rise and decline of imperial orders, they do not provide direct evidence for, on the one hand, the existence or nonexistence of imperial relations or, on the other, what kinds of dynamics follow from the existence of imperial orders (Barnett and Duvall 2005, 63).

Our general neglect of the structural dimension of empires accounts for the fact that much of what passes for analysis of imperial dynamics begins with propositions derived from hegemonic-order theory. Both Gilpin (1981) and Kennedy (1987) stress fundamental continuities between empires and great powers in their accounts of hegemonic cycles. Although Gilpin sees contemporary hegemony as a crucial break from the older pattern of imperial formation and decline, his decision to place both empires and hegemonies within a similar theory of international-political change continues to shape many contemporary treatments of empire. Even some of the more probing assessments of the “lessons of empire” for American foreign policy tend to elide the distinction between imperial and hegemonic systems (Cohen 2004, 56).

Some analysts, in fact, define empires as merely extremely powerful states or preeminent powers (e.g., Lefever 1999). Others treat “empire” as little more than a synonym for unilateralism, preventive war, the promotion of market economies, and high levels of military spending (e.g., Chace 2002; Kagan 1998). Their arguments usually focus on whether the United States is at risk of hegemonic over- or underexpansion, of provoking countervailing coalitions, and of alienating allies (e.g., Snyder 2003). The more general concern—about whether the United States should conserve the existing international order or reconstruct it in light of new challenges—is indistinguishable from classic questions in hegemonic-order theory (Motyl 2006).

Such analytical slippage between “hegemony” and “empire” raises particular problems because the debate concerns, by and large, the possible existence of an informal imperial order. The United States' putative empire contains juridically sovereign states. The relevant question, therefore, is the degree to which specific sectors of international politics reflect not principles of anarchy, but of the hierarchical domestic organization of empires (Lal 2004, 78; Rosen 2003; Wendt and Friedheim 1995). Informal empires, as forms of political structure, occupy a conceptual space somewhere between ideal-typical empires and hegemonic orders (e.g., Hurrell 2005; Lake 1996). In principle, they combine elements of both systems. In the absence of a well-defined “empire framework,” then, it should not be surprising that concepts and dynamics drawn from the analysis of hegemonic systems dominate substantive analysis of the possible implications of American Empire.

COMPARING UNIPOLARITY, HEGEMONY, AND EMPIRE: A RELATIONAL PERSPECTIVE

We can tackle these problems by treating unipolar anarchies, hegemonic orders, and empires as ideal-typical forms of political organization. Researchers construct ideal types to create an idealization of a phenomenon's characteristics which can then be compared against other, related ideal typifications. A particular ideal type will never accurately or exhaustively describe the concrete manifestations of a specific phenomenon (Ringer 1997, 5–6, 17; 110–16). But ideal-typical analysis, by itself, only represents a first step toward developing an “empire framework” that can be compared and contrasted with hegemony and unipolarity.

International-relations scholars have already developed well-known ideal-typifications of unipolarity and hegemony. Some even suggest ideal-typical definition of empire. These ideal-typifications usually catalog the categorical attributes of each type: some, for example, define “unipolarity” as the existence of a single great power in the absence of a common authority and “hegemony” as a condition in which a single great power establishes “rules of the game” for economic and political cooperation (e.g., Mastanduno 2005, 179). A focus on the categorical attributes of ideal types makes it difficult to assess the degree to which specific, often local, patterns of interaction match or combine those associated with unipolarity, hegemony, or empire. If we combine ideal-typical methodology with relational accounts of structure, however, we can recode unipolarity, hegemony, and empire as structural arrangements that vary in terms of their network properties. Doing so allows us to link specific variations in the organizational structures of ideal-typical unipolar, hegemonic, and imperial orders to distinctive dynamics.

Unipolar Orders

Scholars interested in the dynamics of unipolar systems derive most of their assumptions from structural-realist theory: states are unitary, thinly rational actors operating in an anarchical environment (Waltz 1979, 104–105). In ideal-typical unipolar orders routine interstate ties are extremely weak and sparse. They contain no significant vectors of authority (1979, 88). As Keohane and Nye (1989, 24–25) argue, such an ideal type presumes that multiple channels fail to “connect societies,” whether in the form of “interstate,” “transgovernmental,” or “transnational” ties. In other words, the relations among states is one of “billiard balls colliding” against one another (see Figure 1a).

(a) Unipolar Anarchy, (b) Hegemonic Order, (c) Constitutional Order, and (d) Imperial Order

Ties within states are another matter entirely. According to Waltz (1979, 105), “the division of labor across nations … is slight in comparison with the highly articulated division of labor within them. Integration draws the parts of a nation closer together.” In ideal-typical unipolar orders, furthermore, strong collective identification within states contrasts with weak (or nonexistent) collective identification between units. State identities, at least implicitly, are heterogeneous with respect to one another: each state claims to represent a different national or ethnic group (e.g., Lapid and Kratochwil 1996).

The central dynamics of ideal-typical unipolar orders need little elaboration: the key issues concern whether power asymmetries trigger counterbalancing by second-tier political communities, or whether power is so unbalanced that no such counterbalancing is possible (e.g., Layne 1997). Note that this is perfectly consistent with a social-network account of ideal-typical unipolarity. Collective mobilization among actors will be subject to strong Prisoner's Dilemma logics in the absence of dense ties and cross-unit collective identification. Heterogeneous identities work against trust, whereas sparse networks prevent effective monitoring of cooperation and defection. Given that collective action is (comparatively) easy within units, we should see a strong bias toward within-unit provision of collective goods, such as security. The ubiquity of standard collective-action problems among political communities, as many realists argue, is one of the key factors that determine whether unipolar systems will generate counterbalancing (e.g., Wohlforth 1999).

Imperial Orders

Ideal-typifying unipolar and hegemonic orders in terms of their social-network structures was a relatively straightforward task. We merely recoded existing accounts. Not only do we lack, by and large, such accounts of imperial orders in international-relations theory, but also empires are, even in categorical terms, notoriously difficult to define; those political communities we commonly refer to as “empires” differ a great deal from one another (Howe 2002, 9–29). Despite a great deal of ambiguity surrounding the concept, however, scholars and commentators agree on many characteristics of empires. Most argue that empires are multiethnic polities, that they involve the dominance of one polity over other political communities, and that core–periphery relations are an important component of empire (e.g., Spruyt 2005, 3–4; Watson 1992, 16). By themselves, however, these characteristics fail to distinguish “empires” from “hegemonies.” They also do not, in of themselves, suggest a spare set of structural dynamics.

An ideal-typical approach to empires resolves some of these problems. From the perspective of ideal-typical analysis, we need not be overly troubled by the fact that some political communities we think of as empires will not fit our conceptualization, and that some we think of as nonimperial will. What matters is that we construct an inductively plausible ideal-typical account of empires that is also distinctive from ideal-typical hegemonic and unipolar orders. In fact, existing comparative-historical scholarship already supplies the building blocks for such an ideal type. A good starting place is Galtung's (1971, 89) insight that empires involve a distinctive core–periphery structure from those found in other institutional forms. Imperial cores are not merely differentiated from peripheries, but peripheries are differentiated—or segmented—from one another. As Motyl (2001, 16–20) argues, the “most striking aspect of such a structure is not the hub and spoke, but the absence of a rim… of political and economic relations between and among the peripheral units or between and among them and non-imperial polities…. [emphasis added].”

The rimless-hub-and-spoke structure of ideal-typical empires derives, in part, from specific aspects of imperial rule. As Tilly (1997, 3, 3) notes, imperial cores rule their peripheries through the “retention or establishment of particular, distinct compacts” and exercise power “through intermediaries who enjoy considerable autonomy within their own domains” in exchange for “compliance, tribute, and military collaboration with the center.” Empires combine heterogeneous contracting—the basic institutional bargain between the core and each periphery involves distinctive attributes—with some degree of indirect rule. The features, in combination, distinguish imperial structures from those associated with unipolar anarchies and hegemonic systems.

Indirect Rule

A great deal of imperial historiography defines systems of “indirect rule” as those in which empires recruit intermediaries from peripheral populations, such as local chieftains or princes. “Direct rule,” in this terminology, means administration by officials recruited from the imperial core (Mamdani 1996).¹

In a very few cases, the British genuinely utilized direct rule; the core determined policies and few, if any, imperial agents operated on the ground. See. Lange 2003, 393.

In our usage, both systems are variations in indirect rule, i.e., rule through autonomous and quasi-autonomous intermediaries (Lieven 2000, 28; Tilly 1997, 3).

The scope of indirect rule varies across empires. In some, local intermediaries are granted very little discretion over a wide range of policy formulation and implementation. In others, local intermediaries merely have to guarantee a regular flow of resources, access to markets, or minimal compliance with general imperial rules (Newbury 2003). During much of the eighteenth century, for example, the North American colonies of the British empire were essentially self-governing (Lenman 2001). Under the Golden Horde (a typical tribute-taking empire), the princes of the former Rus' could do pretty much what they pleased as long as they supplied regular tribute (however burdensome) to their imperial rulers, did not unduly interfere with the trade routes that provided a crucial source of revenue to the horde, and so forth (Halperin 1983, 242, 250). During the northeast expansion of Muscovy, central authorities would often issue orders of astounding specificity concerning the movements of populations, livestock, and resources. Muscovite agents often replaced local elites (Keenan 1986, 131). In Korea, Japanese “governors-genera…functioned as imperial pro-consuls, rather than as mere agents of civil government.” In Taiwan, they were “semiautonomous and highly authoritarian.” Elsewhere, they had far more limited autonomy (Peattie 1984a, 25–26).

Informal empires generally involve very high levels of intermediary autonomy. Just as they often are in many tribute-taking empires, intermediaries in informal empires are local elites who have their own independent power-base among members of the local population: Elected officials or other individuals who rule in their own right occupy the position of “local intermediary.” The degree of intermediary autonomy in informal empires, as in formal empires, may vary not only across cases but also among peripheries within a single empire.

Heterogeneous Contracting.

Empires, like all political systems, are based on bargains that specify rights and obligations. For instance, imperial bargains may involve an exchange of basing rights in the periphery for access to markets in the core, or they may specify the adoption of specific forms of government or legal orders in a periphery. The benefits derived from peripheries may be diffuse, or they may be oriented toward a specific elite (colonial or indigenous), ethnic group, or subregion. Such relations are always asymmetric, backed by the threat of imperial sanction and negotiated based on the superior resources of imperial cores, but they are contracts nonetheless. To reduce empires to coercive arrangements is to miss the degree to which the benefits of these contracts, and their ultimate legitimacy, underpins imperial orders (e.g., Daniels and Kennedy 2002; Lendon 1997).

Relational-contracting theories of international hierarchy provide the most developed account of the relationship between interstate bargains and forms of hierarchical control (e.g., Lake 2003; Weber 1997). As Lake notes, “all relationships, whether entered into voluntarily or as a result of coercion, can be considered as based upon some ‘contract’ between the two parties” that specifies “the residual rights of control retained by each.” He identifies a continuum of security contracts in which empires “lie at the hierarchic end of the continuum” and occur “when one partner cedes substantial rights of residual control directly to the other” (Lake 1996, 7–8). This conceptualization, however, inadequately distinguishes empires from some other hierarchical relationships. Federations and nation-states, for example, also involve the transfer of substantial residual control from one party to another—including, with a few notable exceptions, control over security (e.g., Spruyt 1994).

Relational-contracting theorists, in fact, tend to collapse indirect and direct rule into the question of residual rights. Empires, as we argued earlier, may involve the transfer of substantial residual rights of control from a periphery to the core but, in practce, delegate decision-making to local intermediaries. Such theories, at least as presently constituted, also neglect a crucial dimension of hierarchical contracting: the degree to which contracts vary between rulers and different subordinate polities (Galtung 1971, 89–90). As Tilly notes, what differentiates ideal-typical federations from ideal-typical empires is that the former involve relatively homogenous contracting. All peripheries agree to, or have imposed on them, a generalized set of contractual rights and obligations (e.g., the United States Constitution). Empires, in contrast, involve heterogeneous contracting: cores develop distinctive bargains with each periphery under their control (Tilly 1997, 3). The contractual relationships between Britain and, for example, its various American colonies, the British East India Company (as “rulers” of Britain's developing Indian empire), and Ireland differed significantly from one another (e.g., Lenman 2001).

The Network Structure of Imperial Orders.

The combination of indirect rule and heterogeneous contracting generates a particular kind of ideal-typical structure (see Figure 1d): one in which (1) central authorities are connected by ties of authority to local intermediaries who, in turn, exercise authority over a variety of other local actors and (2) routine political ties among peripheries are sparse and weak. In terms of social-network theory ideal-typical empires have a very familiar “star-shaped” or “spoke” network structure (Montgomery 2005, 169–70; Padgett and Ansell 1993, 1278) combined with specific patterns of indirect ties that position local intermediaries as brokers between central authorities and local populations (Gould and Fernandez 1989).

Figure 1 illustrates the differences and similarities between empires and the three ideal-typical structures discussed above. It should be rather evident that the network structure of ideal-typical empires is quite distinctive from unipolar and hegemonic orders. Peripheries are tied to the core but segmented from one another; although the preeminent power occupies a brokerage position (see next) in ideal-typical hegemonic orders, its relative centrality is far less than that of central authorities in imperial systems. In ideal-typical accounts of hegemonic orders, the brokerage position occupied by “local intermediaries” drops out entirely. With such differences in mind, we now turn to the distinctive dynamics that follow from imperial orders.

Our more detailed representation of empires in Figure 2 involves three types of actors: core imperial authorities (IA), local intermediaries (L1–L4), and local actors (A–H). To simplify matters, we have included only two local actors per periphery. One should also assume a greater number of peripheries than the four represented here. This network represents only routine authority relations. Imperial authorities may, for example, bypass local intermediaries and negotiate directly with local actors, but doing so involves a significant break from the normal processes of imperial rule. Other kinds of ties may also exist in the network (e.g., economic, social, and kinship), but we will exclude consideration of such noninstitutional ties for the present.

An Empire with Four Peripheries

One of the advantages of a relational approach to structure is that it calls attention to the ways political orders produce different, but nested, structural arrangements at higher and lower levels (e.g., our discussion of anarchy as a nested catnet structure). In this relatively spare account of ideal-typical empires, we confront two important “levels” at which we can view the network structure of empires: the aggregate structure of the entire network, and the local structure of core-periphery relations (e.g., IA, L1, A, and B).

Aggregate Structure.

At the aggregate level, IA is connected to L1, L2, L3, and L4, and indirectly connected to A-I. In contrast, no actors within a periphery are directly connected to any actors in a different periphery. For example, L1 is not connected to L2, L3, and L4; L2, C, and D are connected to one another, but are not connected to L2–L4, or to A, B, E, F… H. Thus, ties between peripheries are comparatively weak and sparse, ties within peripheries are comparatively strong and dense. This aggregate structure gives rise to a number of significant, though interdependent, network characteristics.

Centrality and Brokerage. Both in terms of the hierarchy of authority within empires and in terms of patterns of ties, core imperial authorities occupy a central position with respect to the overall network. In some respects this is just another way of saying that empires have a core–periphery structure, but it has important implications.

In network analysis, centrality is often a proxy for power and influence. The underlying mechanism involves informational asymmetries: actors with more connections have more information about the preferences and orientations of others than those with fewer connections (Freeman 1977, 35–36; Lange 2003, 375). In imperial structures, core authorities occupy a brokerage position between local intermediaries and aggregate peripheries: they negotiate relations among different peripheries, giving them a substantial advantage in terms of power and influence vis-à-vis actors in the rest of the network (Galtung 1971, 90). The fact that these networks involve authoritative relations of super- and subordination only reinforces those other structural sources of asymmetric power.

Peripheral Segmentation and Collective Mobilization. Figure 1 does not adequately capture one important consequence of heterogeneous contracting: the ties that run from IA through L1–L4 to local actors in each periphery all represent a different combination of rights, rules, and obligations. In other words, the categorical identity inhering in the relationship between IA and each periphery is different. This tends to prevent a concordance of interests among peripheries.

Moreover, the organization of empires involves structural holes among peripheries, that is, the comparatively sparse and weak interperiphery ties discussed earlier. Peripheries are disconnected or segmented from one another, at least with respect to authoritative or institutional ties. The Mongols, for example, divided their empire into khanates, and each khanate contained polities (such as Armenia) that were administered as distinctive entities (Allsen 1987). These structural holes work against cross-periphery coordination and collective mobilization (Lange 2003, 374). Because actors in each periphery are isolated from one another, peripheries will (all things being equal) tend toward greater institutional, attitudinal, and normative differentiation over time.²

Peripheries may themselves be segmented and categorically heterogeneous; they may even contain a variety of peripheries themselves. See, for example, Daniels and Kennedy 2002.

Local Structure

The local structure of imperial relations presents a different picture. In the relationship between core imperial authorities and any given periphery, local intermediaries display a higher degree of network centrality. L4, for example, is directly connected to IA, G, and H. G is directly connected to H and L4, H to G and L4, and IA only to L4. Local intermediaries occupy the key brokerage position in the relationship between imperial cores and their peripheries. In terms of local relationships, they are more powerful than central authorities (Stevenson and Greenberg 2000, 656), although that power will, in some respects, be undermined by their lack of authority over imperial central authorities.

THE CONSEQUENCES OF IMPERIAL STRUCTURES

The structure of transactions and the distribution of identities in empires help account for the four major dynamics discussed in the introduction: the workings of across-segment divide-and-rule, the dangers posed by cross-pressures and how they shape legitimating strategies, the tradeoffs of intermediary autonomy, the use of within-segment divide-and-rule to prevent the formation of local coalitions against imperial control, and how these factors may concatenate into imperial overextension. In this section, we elaborate on these dynamics and illustrate our claims—in appropriate ideal-typical fashion—with reference to a number of different historical cases.³

IMPLICATIONS FOR THE AMERICAN EMPIRE DEBATE

Real international systems involve different configurations of anarchy, hierarchy, interdependence, and other structural relations. Imperial structures, for their part, permeate many of the canonical “anarchical” cases of realist theory. The Habsburg bid for European hegemony, the nineteenth-century European balance-of-power system, and the bipolar Cold War, for example, took place in ages of formal and informal empire. Relational accounts of international politics highlight that these were not simply instances of empires relating to one another in an anarchical environment. If we add vectors of imperial control to ideal-typical anarchy, we alter the dynamics of the “international system” (e.g., Barkawi and Laffey 1999, Galtung 1971, 104–109).

Consider our historical illustrations from sixteenth- and seventeenth-century Europe. Balance-of-power dynamics shaped its dynastic rivalries, but so did the fact that dynasts controlled composite political communities with varying degrees of imperial organization. Spanish power depended on processes of resource extraction and imperial bargaining in its European and American empire. Dynasts supported resistance in one another's peripheries and transnational religious movements undermined the ability of various dynasts to divide-and-rule their peripheries (Nexon 2006). In the eighteenth century, as in the twentieth, attempts by great powers to fragment one another's empires reflected the mixed logics of international-relations (Hager and Lake, 2000). Past imperial orders, such as the Roman and ancient Chinese, display equally complicated patterns of anarchical, hegemonic, and imperial relations (see, e.g., Barfield 1989; Luttwak 1976). In all of these cases, not only by variations in anarchy and hierarchy but also by other factors—such as shifts in the nature of economic relations—impacted how various logics of political order translated into historically specific outcomes.

It follows that questions such as “is America an empire?” or “is the international system an imperial one?” obscure more than they reveal. The problem of contemporary empire (American or otherwise) hinges, rather, on the degree to which specific relations—among polities, between polities and nonstate actors, and in specific policy arenas—look more or less like those associated with ideal-typical empires (Cooley 2005). Whenever relations between or among two or more political organizations involve indirect rule and heterogeneous contracting, their interactions will develop at least some of the dynamics associated with imperial control. This is the case regardless of whether we are looking at “neo-trusteeship” overseen by international institutions (Fearon and Laitin 2004), the structure of political control in the sovereign-territorial states such as the Russian Federation (King 2003), or situations in which one state wields such influence over the security-specific policy arenas or functions of another that the latter is, in that respect, a “nominally sovereign but functionally dependent and therefore controllable agent” (Lake 1996, 9).

Thus, we take issue with those who argue that American foreign relations lack imperial qualities because “it would be news to Japan, South Korea, Germany, Kyrgyzstan, and others that Washington actually runs their domestic and foreign policies” (Motyl 2006, 244). Donnelly (2006, 164), for example, argues that not all influence translate into rule; he therefore claims that even contemporary Iraq should be categorized not as a member of an American informal empire but as a “semi-sovereign protectorate of the United States.” But influence, exercised routinely and consistently, becomes indistinguishable from indirect rule. In informal empires the lines between influence and rule necessarily blur. When actors believe that certain options are “off the table” because of an asymmetric (if tacit) contract, or consistently comply with the wishes of another because they recognize steep costs from noncompliance, then the relationship between the two becomes effectively one between ruler and ruled (Barnett and Duvall 2005, 63). Recall that one of the fundamental processes of imperial rule involves the ongoing negotiation of contractual bargains between a variety of actors. Intermediaries and local actors may, in theory, opt to reject or renegotiate any aspect of the imperial bargain. They may decide not to because they accept the legitimacy of the bargain, out of habit, or because they fear imperial sanction. The fact that such sanctions may involve the loss of crucial military, economic, or political support rather than the use of force does not render the relationship nonimperial (Barkawi and Laffey 1999).

These considerations shed important light on the salience of imperial structures and dynamics in American foreign relations. The American-led invasion of Iraq, for instance, currently positions the United States in an imperial relationship with that country. The United States negotiates and renegotiates asymmetric contracts with other states—such as its bargains with Pakistan concerning counterterrorism policy—that place foreign leaders in the structural location of local intermediaries between U.S. demands and their own domestic constituencies (e.g., Lieven 2002). Its basing agreements incorporate many of the hallmarks of imperial bargains (Johnson 2000). But “American empire” is not a phenomenon restricted to the post-Cold War or post-9/11 world. Most of the architecture of contemporary imperial relations in American foreign policy developed during the Cold War (e.g., Bacevich 2002). Decades-long geopolitical developments have, in fact, tended to render American relations less, rather than more, imperial in character.

The heyday of American formal imperial control extended from its period of westward expansion through the aftermath of the Spanish-American war (Motyl 2006, 137). The apogee of American informal imperial relations spanned from the post-World War II occupations of significant portions of Western Europe and East Asia through the early decades of the Cold War during which time the United States restructured the domestic and foreign-policy orientations of large portions of Europe and East Asia (James 2006; Louis 1978). In Western Europe, the United States opted for a multilateral alliance that, over time, took on more of the characteristics of a constitutional order than an imperial one because it encouraged collaboration between segments through institutions such as the North American Treaty Organization (NATO) and the European Economic Community. In East Asia, in contrast, it chose a bilateral hub-and-spoke system (Hemmer and Katzenstein 2002). It did so in order to facilitate divide-and-rule strategies, that is, to avoid collusion between “local intermediaries” such as Chiang Kai-shek of Formosa and Syngman Rhee of South Korea (Cha 2004). American involvement in Central and Latin America also had a strong imperial cast. Since 1945, the United States has consistently aimed for, when it could, hegemonic or informal rule over subordinate entities rather than formal empire. This may reflect a cultural antipathy to formal empire, but it also may stem from an adaptive response to contemporary realities. The diffusion of nationalist ideology makes it very difficult to legitimate—and therefore maintain—formal empire in the current international system (Philpott 2001; Rosen 2003).

Yet the salience of even informal imperial relations in American foreign policy, as we noted earlier, may be in decline. Despite enjoying tremendous military advantage over its nearest peer-competitors, the United States currently exercises less control over the security policy of its Western European and East Asian allies than it did during the early Cold War. Its use of indirect rule in the economic sphere has also declined. The United States' disproportionate influence in the Bretton Woods system, particularly with respect to the special role of the dollar, once led other states—such as France and Britain—to complain bitterly even as they generally acquiesced to American demands. But the United States' relative economic position no longer commands such control. Most foreign currencies are no longer fixed against the dollar, the European Union is a coherent unit in trade and, in a more limited way, in monetary policy, and foreign central banks, particularly in Asia, have accumulated U.S. debt leading to an “economic balance of terror.” Current trends, moreover, point toward a diminishment in the ability of the United States to use its economic position to control the behavior of other polities. As other states, most notably China, gain greater economic leverage, they will be able to provide alternative sources of economic, military, and political support to actual or potential American peripheries (James 2006). Such concerns, familiar in discussions of American hegemony, also suggest an erosion of American informal empire.

Further research should more precisely establish the degree to which, and in what transactional spheres, the United States exercises indirect rule over other political communities through heterogeneous bargains. But we believe that a relational approach to imperial structures still shines a great deal of light on contemporary world politics. The most obvious examples include Iraq, which remains, at least for now, a more or less “controllable agent” whose regime depends on American support. Although the debate over how much autonomy the heads of the Coalition Provisional Authority actually enjoyed continues, the United States indisputably relied on an additional substratum of intermediaries who operated with often little oversight: private contractors. This led to a great deal of corruption, collusion, and inefficiency in the process of reconstruction. These problems ultimately harmed the legitimacy of the American “imperial bargain” with Iraq—and fueled the insurgency—because they undermined the ability of the United States to provide public goods. The costs of their agents' opportunistic exploitation outweighed the measure of plausible deniability that U.S. officials gained from the arrangement (Cooley 2005, 144–56).

American attempts to pursue “unite and rule” policies in Iraq have, so far, met with limited success. Some argue, in fact, that the United States should have focused on binding local potentates and other strategies of within-segment divide and rule (Hulsman and Debat 2006). In any event, the United States faces classic problems of peripheral management: how to navigate the tradeoffs created by intermediary autonomy while preventing the formation of hostile coalitions capable of undermining its objectives and the problems created by transnational identities that interpenetrate a periphery (cf. Diamond 2005). In Afghanistan as well, the United States and NATO currently struggle to legitimate a measure of indirect rule in the context of pivoting and binding strategies, as well as ongoing indigenous and even externally sponsored resistance (e.g., McFaul 2002). The risks of American political, military, and economic overextension follow, in such cases, less from traditional counterbalancing by independent polities than, at least in part, some combination of failures of imperial management and the exploitation of these failures by strategic rivals.

Attention to the dynamics of empires, in fact, helps clarify the dangers faced by the United States of strategic overexpansion. International-relations scholars tend to treat imperial overextension as a consequence of balance-of-power dynamics or of a variety of “myths” generated by core domestic political calculations, either of which leave cores without the fiscal and military resources to secure their commitments (Kupchan 1996; Snyder 1991). This focus mirrors much of the literature on the risks of American overextension (Bellah 2002; Snyder 2003). But, as Paul MacDonald (2004, 9) argues, imperial expansion is often “more the result of pressures in the periphery that lead to unintended, unanticipated political developments that generate reactions that pull great powers more deeply into the politics of other polities.” In the eighteenth century, for example, the interests of their colonists pulled the British into strategically peripheral struggles on the North American continent (Lenman 2001).

Overextension frequently results from local failures of imperial management rather than simply “foreign policy” dynamics. The Spanish Habsburgs' conflicts with England—which scholars often cite as a key factor in Spanish overextension—were, in part, a byproduct of a peripheral uprising in the Netherlands. Both Philip II and Philip III hoped that, by either conquering England or forcing it to capitulate to Spanish hegemonic control, they could cut off England's strategic support for the Dutch (e.g., Allen 2000). Sustained rebellions represent, in fact, only an extreme case of these dynamics. As resistance to imperial bargains grows, empires will find it more difficult to garner and direct resources—manpower, money, trade, and so forth—from and toward peripheries. As their political capacity to manage peripheries diminishes they will, in turn, be more likely to suffer from overextension. Those who currently advocate American—or American-backed Israeli—military action against Syria and Iran embrace very similar reasoning to that of the Spanish: they argue that American problems in Iraq, and in the entire region, might be resolved if only the United States could neutralize those regimes that sponsor resistance to its objectives (e.g., Kristol 2006).

Conceptualizing the international system as involving various configurations of structural logics may elucidate other important concerns in American foreign policy. Scholars operating through the lens of unipolarity and hegemony have tremendous difficulties assessing the implications of, for example, the rise of transnational movements opposed to American interests, changes in contractual relationships between the United States and other polities, and rising anti-Americanism throughout the globe. Some introduce notions of “soft balancing,” whereas others struggle to make sense of potential hegemonic overextension in the absence of counterbalancing (e.g., Paul 2005). Yet many of these developments represent familiar logics in imperial orders. Indeed, some of the key anxieties of contemporary American foreign policy may reflect the intersection of various processes identified by international-relations scholars—including balance-of-power and hegemonic-stability theorists—with the changing character, scope, and degree of informal American imperial relations.

Consider processes of globalization—such as greater economic and communications interdependence—that increase the strength and density of ties among many political communities. In doing so, they enhance the benefits peripheral actors may realize from indirect rule, but they also create conditions of possibility for more effective resistance to informal imperial control by eroding firewalls against transnational and transregional mobilization. The fact that many now view transnational religious and populist movements— violent and otherwise—as realpolitik power-political concerns suggests a conjunction of well-rehearsed imperial dynamics (Barkawi 2004).

The United States, in general, exercises less and less control of the flow of information—and the interpretation of that information—about its bargains and activities around the world. As James (2006, 113–14) notes, empires often deploy extraordinary levels of violence… usually remembered as a series of spectacular brutalizations: the Athenian destruction of Melos, the Roman eradication of Carthage, Oliver Cromwell at Drogheda on September 11, 1649, the German genocide of the Hereros in Southwest Africa, or the Armritsar massacre of 1919.” Frequently, these acts served to frighten locals into submission. But, when “the images of violence are distributed very widely, as they are in a modern imperial age, they undermine rather than strengthen the imperial power.” Both unintentional and intentional acts of violence and abuses of human rights, such as those associated with the Abu Ghraib scandal, generate similar effects: information spreads quickly across multiple channels and erodes the legitimacy of American policies.

Some of the problems in American–Islamic relations may be tied to cross-pressures: consider the difficulties created by close American ties to Israel for its relations with Islamic publics in Egypt, Pakistan, and elsewhere, and the difficulty the United States has had in legitimating its authority vis-à-vis these different audiences (Katzenstein and Keohane 2006, 22). Because the stability of empires depends, in no small measure, on the continuing legitimacy of imperial bargains, growing hostility toward imperial cores comprises a demonstrable threat to imperial order. What some scholars understand as “soft balancing,” therefore, may reflect informal-imperial dissolution that, in the future, could result in structural conditions more favorable to the workings of the balance-of-power mechanism.

Such developments, however, might also create incentives for the United States to expand the breadth and depth of informal empire. As new challengers to the United States offer alternative bargains to weak states, as increasing connectivity undermines its ability to control local intermediaries with a “light footprint,” and as continued disorder in peripheral regions threatens its strategic interests, American policy makers may find themselves returning—whether self-consciously or not—to more robust forms of imperial control (MacDonald 2004, 44–46; Rosen 2003). This was, in many respects, the pattern of the Cold War; it has also been the pattern of many systems in the past, in which empire, anarchy, and other structural conditions combined to produce textures of international politics not comprehensible through any one ideal-typical account of international politics.