Tuesday, September 11, 2001

Robert O. Keohane

The attacks on the United States on September 11, 2001, have incalculable consequences for domestic politics and world affairs. Reliable predictions about these consequences are impossible. However, it may be worthwhile, even at this early point, to reflect on what these acts of violence reveal about the adequacy of our theories of world politics. In what respects have our assumptions and our analytical models helped us to understand these events, and responses to them? And in what ways have we been misled by our theories?

In this short article, I will not attempt to be comprehensive. Instead, I will focus instead on specific issues on which my commentary may be of some value, without presuming that these are the most important issues to address. For instance, the attacks of September 11 reveal that all mainstream theories of world politics are relentlessly secular with respect to motivation. They ignore the impact of religion, despite the fact that world-shaking political movements have so often been fueled by religious fervor. None of them takes very seriously the human desire to dominate or to hate - both so strong in history and in classical realist thought. Most of them tend to assume that the world is run by those whom Joseph Schumpeter (1950 [1942]: 137) called "rational and unheroic" members of the bourgeoisie. After September 11 we need also to keep in mind another motivation: the belief, as expressed by Osama bin Laden, that terrorism against "infidels" will assure one "a supreme place in heaven."1 However, since I have few insights into religious motivations in world politics, I will leave this subject to whose who are more qualified to address it.

In the next section of this article I define the phrase, "the globalization of informal violence." In referring to a general category of action, I substitute this phrase for "terrorism," since the latter concept has such negative connotations that it is very difficult to define in an analytically neutral and consistent way that commands general acceptance.2 Even as the United Nations Security Council has passed resolutions against terrorism, it has been unable to define the term. Since everyone is against terrorism, the debate shifts to its definition, as each party seeks to define its enemy's acts, but not its own, as terrorist. Nevertheless, deliberately targeted surprise attacks on arbitrarily chosen civilians, designed to frighten other people, are clearly acts of terror. The attacks on the World Trade Center of September 11, 2001, were therefore terrorist acts and I refer to them as such.

This paper has three themes. First, the events of September 11 illustrate starkly how our assumptions about security are conceived in terms of increasingly obsolescent views of geographical space. Secondly, the globalization of informal violence can be analyzed by exploring patterns of asymmetrical interdependence and their implications for power. Thirdly, United States responses to the attacks tell us quite a bit about the role of multilateral institutions in contemporary world politics.

My argument is that our theories provide important components of an adequate post-September 11 conceptualization of world politics, but that we need to alter some of our assumptions in order to rearrange these components into a viable theoretical framework. Effective wielding of large-scale violence by non-state actors reflects new patterns of asymmetrical interdependence, and calls into question some of our assumptions about geographical space as a barrier. Responses to these actions reveal the significance of international institutions as well as the continuing central role of states. In thinking about these issues, students of world politics can be usefully reminded of Judith N. Shklar's concept of the "liberalism of fear," and her argument that the most basic function of a liberal state is to protect its citizens from the fear of cruelty.

1. The Globalization of Informal Violence and the Reconceptualization of Space

The various definitions of globalization in social science all converge on the notion that human activities across regions and continents are being increasingly linked together, as a result both of technological and social change (Held et al.: 15). Globalism as a state of affairs has been defined as "a state of the world involving networks of interdependence at multicontinental distances, linked through flows of capital and goods, information and ideas, people and force, as well as environmentally and biologically relevant substances" (Keohane and Nye 2001: 229).

When globalism is characterized as multidimensional, as in these definitions, the expansion of terrorism's global reach is an instance of globalization (Held et al. 1999: 80; Keohane and Nye 2001: 237). Often, globalism and globalization have been defined narrowly as economic integration on a global scale; but whatever appeal such a definition may have had, it has surely disappeared after September 11. To adopt it would be to imply that globalized informal violence, which takes advantage of modern technologies of communication, transportation, explosives, and potentially biology, somehow threatens to hinder or reduce the level of globalism. But like military technology between 1914 and 1945, globalized informal violence strengthens one dimension of globalism - the networks through which means of violence flow - while potentially weakening globalism along other dimensions, such as economic and social exchange. As in the past, not all aspects of globalization go together.

I define informal violence as violence by non-state actors, capitalizing on secrecy and surprise to inflict great harm with small material capabilities. Such violence is "informal" because it is not wielded by formal state institutions and it is typically not announced in advance, as in a declaration of war. Such violence becomes globalized when the networks of non-state actors operate on an intercontinental basis, so that acts of force in one society can be initiated and controlled from very distant points of the globe.

The implications of the globalization of formal violence were profound for traditional conceptions of foreign policy in an earlier generation, particularly in the United States, which had so long been insulated by distance from invasion and major direct attack. The great expositors of classical realist theories of foreign policy in the United States, such as Walter Lippmann, began with the premise that defense of the "continental homeland" is "a universally recognized vital interest." Before World War II, threats to the homeland could only stem from other states that secured territory contiguous to that of the United States or that controlled ocean approaches to it. Hence the Monroe Doctrine of 1823 was the cornerstone of American national security policy. As Lippmann recognized in 1943, changes in the technologies of formal violence meant that security policy needed to be more ambitious: the United States would have to maintain coalitions with other great powers that would "form a combination of indisputably preponderant power" (Lippmann 1943: 88, 101). Nevertheless, Lippmann was able to retain a key traditional concept: that of a geographically defined defensive perimeter, which can be thought of as a set of concentric circles. If the United States were to control not only its own area but the circle surrounding that area, comprising littoral regions of Europe and Asia, its homeland would be secure.

The American strategists of the 1950s - led by Bernard Brodie, Thomas Schelling, and Albert Wohlstetter - had to rethink the concept of a defensive perimeter, as intercontinental ballistic missiles reduced the significance of distance: that is, as formal violence became globalized. John Herz (1959: 107-108) argued that nuclear weapons forced students of international politics to rethink sovereignty, territoriality, and the protective function of the state:

"With the advent of the atomic weapon, whatever remained of the impermeability of states seems to have gone for good.... Mencius, in ancient China, when asked for guidance in matters of defense and foreign policy by the ruler of a small state, is said to have counseled: 'dig deeper your moats; build higher your walls; guard them along with your people.' This remained the classical posture up to our age, when a Western sage, Bertrand Russell, could still, even in the interwar period, define power as a force radiating from one center and diminishing with the distance from that center until it finds an equilibrium with that of similar geographically anchored units. Now that power can destroy power from center to center everything is different."

September 11 signifies that informal violence has become globalized, just as formal, state-controlled violence became globalized, for the superpowers, during the 1950s. The globalization of informal violence was not created by September 11. Indeed, earlier examples, extending back to piracy in the 17th century, can be easily found. But the significance of globalization - of violence as well as economically and socially - is not its absolute newness but its increasing magnitude as a result of sharp declines in the costs of global communications and transportation (Keohane and Nye 2001: 243-45).

Contemporary theorists of world politics face a challenge similar to that of this earlier generation: to understand the nature of world politics, and its connections to domestic politics, when what Herz called the "hard shell" of the state (Herz 1959: 22) has been shattered. Geographical space, which has been seen as a natural barrier and a locus for human barriers, now must be seen as a carrier as well.

The corollary to the barrier conception of geographical space was scorn for the geopolitical significance of weak countries, without nuclear weapons, far from the defensive perimeter, consigned to a "the obscurity that they justly deserve." One of the finest hours of American realism was its opposition to the war in Vietnam on just these grounds. Walter Lippmann and Hans J. Morgenthau opposed the war not for moralistic reasons, but because of Vietnam's unimportance to the national interests of the United States. For Lippmann, the key to a successful foreign policy is achieving a "balance, with a comfortable surplus of power in reserve, [between] the nation's commitment and the nation's power" (Lippmann 1943: 9). Going abroad in search of monsters to destroy upset that balance.

The globalization of informal violence, carried out by networks of non-state actors, defined by commitments rather than by territory, has profoundly changed these fundamental foreign policy assumptions.3 On traditional grounds of national interest, Afghanistan should be one of the least important places in the world for American foreign policy - and until the Soviet invasion of 1979, and again after the collapse of the Soviet Union in 1991 until September 11, the United States all but ignored it. Yet in October 2001 it became the theatre of war. Globalization means, among other things, that threats of violence to our homeland can occur from anywhere. The barrier conception of geographical space, already anachronistic with respect to thermonuclear war and called into question by earlier acts of globalized informal violence, was finally shown to be thoroughly obsolete on September 11.4

2. Interdependence and Power

Another way to express the argument made above is that networks of interdependence, involving transmission of informal violence, have now taken a genuinely global form. Using this language helps us to see the relevance for the globalization of informal violence of the literature on interdependence and power, which was originally developed to understand international political economy. In that literature, interdependence is conceptualized as mutual dependence, and power is conceptualized in terms of asymmetrical interdependence.5 This literature has also long been clear that "military power dominates economic power in the sense that economic means alone are likely to be ineffective against the serious use of military force" (Keohane and Nye 2001: 14).

September 11 revealed how much the United States could be hurt by informal violence, to an extent that had been anticipated by some government reports but that had not been incorporated into the plans of the government.6 The long-term vulnerability of the United States is not entirely clear, but the availability of means of mass destruction, the extent of hatred for the United States, and the ease of entering the United States from almost anywhere in the world, all suggest that vulnerability may be quite high.

If the United States were facing a territorial state with conventional objectives, this vulnerability might not be a source of worry. After all, the United States has long been much more vulnerable, in technological terms, to a nuclear attack from Russia. But the United States was not asymmetrically vulnerable. On the contrary, the United States either had superior nuclear capability or "mutual assured destruction" (MAD) kept vulnerability more or less symmetrical. Russia has controlled great force, but has not acquired power over the United States from its arsenal.

With respect to terrorism, however, two asymmetries, which do not normally characterize relationships between states, favored wielders of informal violence in September 2001. First, there was an asymmetry of information. It seems paradoxical that an "information society" such as that of the contemporary United States would be at an informational disadvantage with respect to networks of individuals whose communications seem to occur largely through hand-written messages and face-to-face contacts. But an information society is also an open society. Potential terrorists had good information about their targets, while before September 11 the United States had poor information about the identity and location of terrorist networks within the United States and other western societies. Perhaps equally important, the United States was unable coherently to process the information that its various agencies had gathered. Secondly, there is an asymmetry in beliefs. Some of Osama bin Laden's followers apparently believed that they would be rewarded in the afterlife for committing suicidal attacks on civilians. Others were duped into participating in the attacks without being told of their suicidal purpose. Clearly, the suicidal nature of the attacks made them more difficult to prevent and magnified their potential destructive power. Neither volunteering for suicide missions nor deliberately targeting civilians is consistent with secular beliefs widely shared in the societies attacked by al-Qaeda.

The United States and its allies have enormous advantages in resources, including military power, economic resources, political influence, and technological capabilities. Furthermore, communications media, largely based in the West, give greater weight to the voices of people in the wealthy democracies than to those of the dispossessed in developing countries. Hence the asymmetries in information and beliefs that I have mentioned are, in a sense, exceptional. They do not confer a permanent advantage on the wielders of informal violence. Yet they were sufficient to give the terrorists at least a short-term advantage, and they make terrorism a long-term threat.

Our failure to anticipate the impact of terrorist attacks does not derive from a fundamental conceptual failure in thinking about power. On the contrary, the power of terrorists, like that of states, derives from asymmetrical patterns of interdependence. Our fault has rather been our failure to understand that the most powerful state ever to exist on this planet could be vulnerable to small bands of terrorists due to patterns of asymmetrical interdependence. We have overemphasized states and we have over-aggregated power.

Power comes not simply out of the barrel of a gun, but from asymmetries in vulnerability interdependence - some of which, it turns, out, favor certain non-state actors more than most observers anticipated. The networks of interdependence along which power can travel are multiple, and they do not cancel one another out. Even a state that is overwhelmingly powerful on many dimensions can be highly vulnerable on others. We learned this lesson in the 1970s with respect to oil power; we are re-learning it now with respect to terrorism.

3. Institutions and Legitimacy

Institutionalist theory implies that multilateral institutions should play significant roles wherever interstate cooperation is extensive in world politics Yet a reader of the American press immediately after the September 11, 2001, attack on the World Trade Center and the Pentagon, might well have thought this claim weirdly divorced from reality. Immediate reactions centered on domestic security, military responses, and the creation of a broad international coalition against terrorism. Although the United Nations Security Council did act on September 12, passing resolution 1368, its response attracted relatively little attention. Indeed, President Bush's speech to Congress of September 20 did not mention the United Nations, although the President did praise NATO and made a generic reference to international organizations. And coverage of the United Nations was virtually nonexistent in the New York Times.

But theory is not tested by the immediate reactions of policymakers, much less by those of the press. Social science theory purports to elucidate underlying structures of social reality, which generate incentives for action. Kenneth Waltz rightly looks for confirmation of his theory of the balance of power "through observation of difficult cases." The theory is confirmed, he claims, where states ally with each other, "in accordance with the expectations the theory gives rise to, even though they have strong reasons not to cooperate with one another" (Waltz 1979: 125). Realists rightly argue that if leaders seem to be compelled toward actions that theory suggests - as, for instance, Winston Churchill was when Britain allied with the Soviet Union in 1941 and American leaders when they built NATO after World War II - this counts for their theory. Indeed, the most demanding test of theory comes when policymakers are initially unreceptive to the arguments on which the theory is based. If they nevertheless turn to the policy measures that the theory anticipates, it gains support.

The terrorist attacks of September 11 therefore pose a fruitful test for institutionalist theory. Before September 11, the Bush Administration had been pursuing a notably unilateralist policy with respect to several issues, including global warming, trade in small arms, money laundering, and tax evasion. Its leading policymakers all had realist proclivities: they emphasized the decisive use of force and had not been public supporters of international institutions. Their initial inclinations, if their public statements and those of the President are any guide, did not lead them to emphasize the role of the United Nations.

Nevertheless, the United States returned to the Security Council. On September 28, 2001, the Security Council unanimously adopted Resolution 1373, on the motion of the United States. This resolution used the mandatory provisions of Chapter VII of the United Nations Charter to require all states to "deny safe haven" both to terrorists and to those who "provide safe haven" to terrorists. Resolution 1373 also demanded that states prevent potential terrorists from using their territories, and "prevent and suppress the financing of terrorist acts." It did not, as noted above, define terrorism. Furthermore, the United States continued to engage the United Nations, indeed delegating to it the task of bringing Afghan factions together in Germany in a meeting that culminated in an agreement in December 2001.

Why should the United States have relied so extensively on the United Nations? The UN, in Stalin's famous phrase, has no divisions. The United States, not the UN, carried out the significant military actions. Transnational banks, central banks, and states in their capacities as bank regulators, froze funds allegedly belonging to terrorists. Even before the September 28 Security Council resolution, allies of the United States had already invoked Article 5 of NATO's Charter.

Inis L. Claude proposed one answer almost 35 years ago (Claude 1967). States seek "collective legitimation" for their policies in the United Nations. Only the UN can provide the breadth of support for an action that can elevate it from the policy of one country or a limited set of countries, to a policy endorsed on a global basis. In contemporary jargon, the "transaction costs" of seeking support from over 150 countries around the world are higher than those of going to the Security Council, ready to meet at a moment's notice. But more important than these costs is the fact that the institution of the United Nations can confer a certain degree of legitimacy on a policy favored by the United States.

What does legitimacy mean in this context? Legally, decisions of the United Nations Security Council on issues of involving the use of violence are legitimate since members of the United Nations, through the Charter, have authorized such decisions. In a broader popular and normative sense, decisions are legitimate for a given public insofar as members of that public believe that they should be obeyed. As Weber pointed out, the sources of such legitimacy may include tradition, charisma or rational-legal authority (Weber 1978: 954); they may also include appeal to widely accepted norms. People in various parts of the world may believe that their governments should obey decisions of the Security Council because they were made through a process that is normatively as well as legally acceptable. Or they may regard its decisions as legitimate insofar as they are justified on the basis of principles - such as collective opposition to aggression - that they regard as valid.

Why is legitimacy important? In part, because people will voluntarily support a legitimacy policy, without requiring material inducements.7 But it would be naive to believe that leaders of most countries will be persuaded, by Security Council action, of the wisdom or righteousness of the policy and will therefore support it for normative reasons. To explain the impact of Security Council resolutions, we need also to look for self-interested benefits for leaders.

Even if the leaders are entirely cynical, the adoption of a legitimate UN resolution will change their calculations. If they lead democratic societies in which publics accept the legitimacy of UN action, they will benefit more politically from supporting policies endorsed by the United Nations, than from supporting policies not so endorsed. If they exercise rule over people who are unsympathetic to the policies and who do not accept them merely due to UN endorsement, the legal status of Security Council resolutions may change their calculations. Chapter VII decisions are mandatory, which means that states defying the Security Council run the risk of facing sanctions themselves, as in the Gulf War. Leaders of countries with unsympathetic populations can point out that, however distasteful it may be to take action against Osama bin Laden and his network, it could be more costly to be cut off from essential supplies and markets, to suffer disruption of transportation and banking services, or even to become a target of military action.

The general point is one that has often been made by institutional theory: international institutions work largely by altering the costs of state strategies. Of course, there is no guarantee that institutions will be sufficiently important to ensure that strategies change: they are only one element in a mixture of calculations. Yet as the use of the United Nations by the United States indicates, they are an element that should not be overlooked.

The general issue of whether the United States should secure multilateral endorsement of its policies takes us back to issues of interdependence and power. The focused terrorist attacks on the United States have made the United States more dependent on other states for assistance in its struggle. Reciprocity requires that in return for such assistance, the United States provide benefits to its partners. Some of these benefits are bilateral, but others are more efficiently provided through multilateral institutions. Using these institutions, the United States can send stronger signals and make more credible commitments. Hence terrorism directed against the United States is likely to make the United States more responsive to its partners - both partners in crucial regions, such as Pakistan, and the most important entities on a global basis, such as China, Russia and the European Union. Terrorism may also make the United States more receptive to the use of multilateral institutions even when they limit American freedom of action.

In the short run, multilateralism may be a beneficiary of the globalization of informal violence. A major reason for a government to commit itself to a multilateral policy is that it can thereby induce other governments to make valuable reciprocal commitments. The United States has greater need for commitments from other states now than it had before September 11, and is therefore more inclined to make reciprocal commitments of its own. Greater multilateralism with respect to security issues is the direct result. Indirectly, due to issue-linkage, more American concessions on other issues can also be expected. These concessions will surely entail, as an indirect result, more multilateralism on these issues, as a way of improving relations with states whose help the United States needs to fight terrorism, and the support of whose people the United States seeks.

Yet at this point a note of caution is necessary. The United States has been notably reluctant to permit the United Nations, or its own allies, to restrict its military freedom of action in Afghanistan. In fact, requests by Great Britain to send in troops to protect relief operations were rebuffed by the United States on the advice of its military commanders. Another interpretation of the delegation of inter-factional negotiations to the United Nations is that the United States seeks to be able to leave Afghanistan promptly after the defeat of the Taliban and the capture or killing of Bin Laden and al-Qaeda fighters. Even more cynically, it can be feared that American policymakers wish to be able to lay the blame for inevitable political difficulties as the door of the UN.

One can easily imagine a more pessimistic scenario for the next few years. The United States government could decide that its security required radical measures that would not be supported even by many of its NATO allies, such as an attack on Iraq without strong evidence of Iraqi complicity in prior attacks on the United States. In such an eventuality, American actions would not be legitimated either by the United Nations or by NATO. Having acted unilaterally, the United States would not be moved to rely more heavily on international institutions, and multilateralism could suffer a serious blow.

Even if the multilateral path is chosen, it is hardly likely to be sufficient. It is unlikely that multilateral organizations will be the key operating agencies in dealing with the globalization of informal violence: they are too cumbersome for that. The state, with its capacity for decisive, forceful action and the loyalty it commands from citizens, will remain a necessary part of the solution to threats of informal violence. Jejeune declarations of the "death of the state" are surely among the casualties of the terrorist offensive But multilateral organizations will be an essential part of the process of legitimizing action by states.

It should be evident that these arguments about multilateral institutions and networks are not "anti-realist." On the contrary, they rest on an appreciation of the role of power, and of state action, in world politics; on an understanding that new threats create new alliances; and on a belief that structures matter. Analysts who are sensitive to the role of multilateral institutions need not regard them as operating independently from states, nor should they see such institutions as a panacea for our new ills. But sensitivity to the role of multilateral institutions helps us see how these institutions can play a role: not only by reducing transaction costs but also by generating opportunities for signalling commitments and providing collective legitimacy for effective action.

4. The "Liberalism of Fear"

Judith N. Shklar's "liberalism of fear" envisages liberal democracy as "more a recipe for survival than a project for the perfectibility of mankind." It seeks to avoid the worst outcomes, and therefore declares that "the first right is to be protected against the fear of cruelty" (Shklar 1984: 4; 237). The liberalism of fear certainly speaks to our condition today, as it did to that of victims of the Nazis such as Judith Shklar. It raises both an analytical and a normative issue. Analytically, it leads us to ask about the protective role of the state, facing the globalization of informal violence. Normatively, it should make us think about our own role as students of world politics.

The erosion of the concept of a protected homeland within a defensive perimeter, discussed above, makes the "liberalism of fear" more relevant to Americans than it has been in almost two centuries. Suddenly, the task of protecting citizens from the fear of cruelty has become a demanding project for the state, not one that a superpower can take for granted.

Judith Shklar looked to the state as the chief threat. "No liberal," she declared, "ever forgets that governments are coercive" (Shklar 1984: 244). In this respect, the "liberalism of fear" shares a blind spot with the most popular theories of world politics, including realism, institutionalism and some forms of constructivism . All of these views share a common fault: they do not sufficiently take account of how globalization facilitates the agency of non-state entities and networks. After September 11 no liberal should be able to forget that non-state actors, operating within the borders of liberal states, can be as coercive and fear-inducing as states.

Recognition of the dangers of informal violence may lead the United States toward a broader vision of its global interests. As we have seen, classical realist thinking drew a bright line between geographical areas important to the national interest and those parts of the world that were insignificant from the standpoint of interests. Now that attacks against the United States can be planned and fostered within countries formerly viewed as insignificant, this bright line has been blurred.

One of the implications of this blurring of lines is that the distinction between self-defense and humanitarian intervention may become less clear. Future military actions in failed states, or attempts to bolster states that are in danger of failing, may be more likely to be described both as self-defense and as humanitarian or public-spirited. When the only arguments for such policies were essentially altruistic ones, they commanded little support, so the human and material price that American leaders were willing to pay to attain them was low. Now, however, such policies can be framed in terms of American self-interest, properly understood. Sound arguments from self-interest are more persuasive than arguments from responsibility or altruism.

More generally, recognition of the dangers of informal violence will force a redefinition of American national interests, which could take different forms. Such a redefinition could lead Americans to support measures to reduce poverty, inequality and injustice in poor countries. The Marshall Plan is a useful if imperfect analogy. In 1947 the United States redefined its self-interest, taking responsibility for helping to build a democratic and capitalist Europe, open to other capitalist democracies. The United States invested very large resources in this project, with great success. The task now in the less developed countries is much more daunting, both in sheer magnitude and since the political systems of most of these countries are weaker than those of European countries in 1947.8 But the resources available to the United States and other democratic countries are also much greater than they were in 1947.

Any widely appealing vision of American interests will need to be based on core values that can be generalized. Individual freedom, economic opportunity, and representative democracy constitute such values. The ability to derive gas-guzzling sports utility vehicles (SUVs) does not. In the end, "soft power" (Nye 1990) depends not merely on the desire of people in one country to imitate the institutions and practices prevailing in another, but also their ability to do so. Exhibiting a glamorous lifestyle that others have no possibility of attaining is more likely to generate hostility and a feeling of "sour grapes" than support. To relate successfully to people in poor countries during the 21st century, Americans will have to distinguish between their values and their privileges.

The attachment of Americans to a privileged lifestyle raises the prospect of a defensive and reactionary broadening of American national interests. Recall that a virtue of classical realism was to link commitments to a relatively limited set of interests, defined partly by geography. Ideology and a self-serving attempt to preserve privileges could define a different set of interests. Opponents - not merely those who have attacked the United States - would be demonized. Deals would continue to be cut with corrupt and repressive regimes to keep cheap oil flowing to the United States. The United States would rely exclusively on military power and bilateral deals rather than also on economic assistance, trade benefits, and efforts at cultural understanding. The costs would include estrangement from our democratic allies and hatred of the United States in much of the world. Ultimately, such a vision of national interest is a recipe for isolation and continual conflict - an environment in which liberal democracy could be threatened by the emergence of a garrison state at home.

Normatively, thinking about the "liberalism of fear" reminds our generation that in a globalized world, we cannot take liberal societies for granted. People such as Judith Shklar, who experienced Nazism, understood the fear of cruelty in their bones. Those of us who grew up in the United States during the Cold War experienced such fear only in our imaginations, although nuclear threats and wars such as those in Korea and Vietnam gave our imaginations plenty to work with. The generations that have come of age in the United States since the mid-to-late 1980s - essentially, those people under 35 - have been able to take the basics of liberalism for granted, as if the United States were insulated from the despair of much of the world's population. The globalization of informal violence means that we are not so insulated. We are linked with hateful killers by real physical connections, not merely those of cyberspace. Neither isolationism nor unilateralism is a viable option.

Hence, the liberalism of fear means that we who study international interdependence and multilateral institutions will need to redouble our efforts. We should pay less attention to differentiating our views from those of other schools of international relations; more to both synthesis and disaggregation. We need to synthesize insights from classical realism, institutionalism, and constructivism, but we also need to take alternative worldviews - including religious worldviews - more seriously. We need to examine how purposes are shaped by ideas and how calculations of power interact with institutions, to produce outcomes in world politics. We need, at the same time, to disaggregate strands of asymmetrical interdependence, with their different implications for power; and to differentiate international institutions and networks from one another, in their effects and their potential for good or ill.

Conclusion

The terrorist attacks on New York and Washington force us to rethink our theories of world politics. Globalism should not be equated with economic integration. The agents of globalization are not simply the high-tech creators of the internet, or multinational corporations, but also small bands of fanatics, travelling on jet aircraft and inspired by fundamentalist religion. The globalization of informal violence has rendered problematic our conventional assumptions about security threats. It should also lead us to question the classical realist distinction between important parts of the world, in which great powers have interests, and insignificant places, which were thought to present no security threats although they may raise moral dilemmas. Indeed, we need to reconceptualize the significance for homeland security of geographical space, which can be as much a carrier of malign informal violence as a barrier to it.

Most problematic are the assumptions in international relations theory about the roles played by states. There has been too much "international relations," and too little "world politics," not only in work on security but also in much work on international institutions. States no longer have a monopoly on the means of mass destruction: more people died in the attacks on the World Trade Center and the Pentagon than in the Japanese attack on Pearl Harbor in 1941. Indeed, it would be salutary for us to change the name of our field, from "international relations" to "world politics."9 The language of "international" relations enables us to slip back into state-centric assumptions too easily. Asymmetrical interdependence is not merely an interstate phenomenon.

Yet as the state loses its monopoly on means of mass destruction, the response to terrorism is strengthening the powers of states, and the reliance of people on government. Even as states acquire more authority, they are likely to cooperate more extensively with each other on security issues, using international institutions to do so. Ironically, as states acquire more authority, they will be forced to learn better how to relate to networks - both hostile networks and networks that they may use instrumentally - and to rely more heavily on multilateral institutions. These institutions, in turn, will have to define their tasks in ways that emphasize their advantages - in conferring collective legitimacy on actions - while minimizing the impact of their liabilities, as cumbersome organizations without unity of command.

One result of these apparently paradoxical changes is closer linkages between traditional security issues and other issues. The artificial but convenient separation of the field into security and political economy may be one of the casualties of the struggle against terrorism. Areas formerly seen as "non-security areas," such as air transport, transnational finance, and migration, have become more important to security, and more tightly subject therefore to state regulation.

Finally, the globalization of informal violence indicates how parochial have been some of the disputes among various schools of international relations theory. Analysis of the ramifications of the attacks on the United States must come to grips not only with structures of power, but also with changing subjective ideas and their impact on strategies. It must be concerned with international institutions, and with non-state actors and networks -- elements of world politics emphasized by different schools of thought. And it must probe the connections between domestic politics and world politics. We do not face a choice between these perspectives, but rather the task of synthesizing them into a comprehensive, yet coherent, view.

Our understanding of world politics has often advanced under the pressure of events, such as those of World War II, the Nuclear Revolution, and the growth of economic interdependence over the last fifty years. Perhaps the globalization of informal violence will refocus our attention for a new period of intellectual creativity, as sober thinking about global governance and classic political realism converge on problems identified so well by the "liberalism of fear."

Acknowledgements

I am grateful for comments on earlier versions of this paper to Carol Atkinson, Hein Goemans, Peter Gourevitch, Nannerl O. Keohane, Lisa L. Martin, Joseph S. Nye, John Gerard Ruggie, and Anne-Marie Slaughter, as well as to participants at seminars at the University of Pennsylvania, October 18, 2001; at the University of Amsterdam, November 2, 2001; at Duke University, November 16, 2001; and at the University of Tokyo, December 10, 2001. At the Amsterdam colloquium I benefited particularly from the comments of Gerd Junne and at the Tokyo colloquium from the comments of Yasuaki Osuma.

Footnotes

1 Statement by Osama bin Laden, New York Times, October 8, 2001, p. B7.

2 The best definitional discussion of terrorism that I know of us by Alex Schmid, who defines it as "an anxiety-inspiring method of repeated violent action, employed by (semi)clandestine individual, group or state actors, for idiosyncratic, criminal or political reasons, whereby - in contrast to assassination - the direct targets of violence are not the main targets" (Schmid 1993: 8, 12).

3 A few pessimistic and prescient observers understood that terrorism could pose a threat to the United States homeland despite our dominance in military power. See Carter and Perry 1999, and the Hart-Rudman Report, Phase I, September 15, 1999, Conclusion 1.

4 Another implication of this change is that the bright line between humanitarian intervention - to save others from human rights abuses - and self-defense - to protect ourselves - has become blurred.

5 In 1977 Keohane and Nye distinguished between two types of dependence, which they labelled (following the contemporary literature on economic interdependence) sensitivity and vulnerability dependence. Sensitivity dependence refers to "liability to costly effects imposed from outside before policies are altered to try to change the situation." Vulnerability dependence, in contrast, refers to "an actor's liability to suffer costs imposed by external events even after policies have been altered. This language seems inappropriate in the contemporary situation, since in ordinary language, the attacks on an unprepared United States on September 11 demonstrated how vulnerable the country was. But the distinction between levels of dependence before and after policy change remains important. See Keohane and Nye 2001: 11; the text is unchanged from the 1st edition, 1977.

6 My colleague Ole Holsti has pointed out to me that in surveys conducted by the Chicago Council on Foreign Relations in 1994 and 1998, the public more often regarded international terrorism as a "critical" foreign policy issue than did leaders. Indeed, 69% and 84%, respectively, of the public regarded terrorism as a critical issue in those years, compared to 33% and 61% of the elites. See Holsti 2000: 21.

7 Douglas North links legitimacy to the costs of enforcing rules. "The costs of maintenance of an existing order are inversely related to the perceived legitimacy of the existing system. To the extent that the participants believe the system fair, the costs of enforcing the rules and property rights are enormously reduced." North 1981: 53.

8 It is tempting in hindsight to forget that the political systems of European countries were not terribly strong in 1947. Germany was still under occupation, Italy had recently been Fascist, and France and Italy had very large, pro-Soviet communist parties. Nevertheless, these countries had relatively highly-educated populations, they had some history of democratic or at least liberal politics, and their administrative bureaucracies were quite effective.

9 This is a point that the late Susan Strange repeatedly emphasized.

Bibliography

Carter, Ashton B. and William Perry. 1999. Preventive Defense: A New Security Strategy for America. Washington, DC: Brookings.

Claude, Inis L. 1967. The Changing United Nations. New York: Random House.

Hart, Gary, Warren Rudman, et al. 1999. Phase I Report on the Emerging Global Security Environment for the First Quarter of the 21st Century, United States Commission on National Security/21sst Century (Washington: September 15). "Hart-Rudman Report."

Held, David and Anthony McGrew, David Goldblatt, and Jonathan Perraton. 1999. Global Transformations. Stanford: Stanford University Press.

Holsti, Ole. 2000. "Public Opinion and Foreign Policy." In Robert Lieber, ed., Eagle Rules? (New York: Longman): 16-46.

Keohane, Robert O. and Joseph S. Nye. 2001. Power and Interdependence. 3rd edition. New York: Addison Wesley Longman.

Lippmann, Walter. 1943. U.S. Foreign Policy: Shield of the Republic. Boston: Little, Brown.

North, Douglas C. 1981. Structure and Change in Economic History. New York: W. W. Norton.

Nye, Joseph S. 1990. Bound to Lead: The Changing Nature of American Power. New York: Basic Books.

Schmid, Alex P. 1993. "The Response Problem as a Definition Problem," in Schmid and Crelinsten, 1993: 7-13.

Schmid, Alex P. and Ronald D. Crelinsten,. 1993. Western Responses to Terrorism. London: Frank Cass.

Schumpeter, Joseph. 1950 [1942]. Capitalism, Socialism and Democracy. New York: Harper and Row.

Shklar, Judith N. Ordinary Vices. 1984. Cambridge: the Belknap Press of Harvard University Press.

Waltz, Kenneth N. 1979. Theory of International Politics. Reading, MA: Addison-Wesley.

Weber, Max. 1978. Economy and Society, edited by Guenther Roth and Claus Wittich. Berkeley: University of California Press.