The nuclear age was inaugurated by the largest and most advanced industrial power in history, the United States. Only the greatest of great powers had the capacity to mobilize the vast and diverse resources necessary to acquire the first nuclear weapons. The U.S. then announced the advent of the new age with a big bang, over Hiroshima and again over Nagasaki.

The next power to acquire nuclear weapons was the only other superpower, the Soviet Union. Although never as super as the United States, the Soviets possessed the second-largest industrial capacity in the world, so it was altogether natural that the USSR took silver in the nuclear-weapons race.

Neither was it surprising that the next states to join the club were also major industrial powers, although not really super ones—Britain in 1952, France in 1960, and China in 1964. When India tested its “nuclear device” in 1974 (it did not then call it a nuclear weapon), even this merely developing economy did not seem to be very far below the previous capacity standard. Clearly, as the nuclear age was advancing in years, the required capacity for acquiring nuclear weapons was becoming smaller. Finally, in 1998, when Pakistan tested its first nuclear weapon, it demonstrated that nuclear weapons could be acquired by a state that was hardly an industrial power at all. The nuclear tests by North Korea in 2006 emphasized this new reality, and if and when Iran develops its own nuclear weapons, that will underscore the point.

In the course of the first six decades of the nuclear age, advances in technology and, more importantly, in the ability to obtain technologies already developed by more advanced nations, have steadily lowered the threshold for acquiring nuclear weapons. Technological advances have driven the necessary capacity downward, successively from superpowers, to major powers, to minor powers.


Now, in the seventh decade of the nuclear age, the great fear is that we will soon see—perhaps with another big bang—the next step in this drive downward. Nuclear weapons will be acquired by an organization that is no power or state at all, a subnational but transnational terrorist network, such as al-Qaeda, which has already said that it is intent upon using nuclear weapons against America.

Not being a state, a sub- and transnational network does not possess territory or population for which it would be responsible—assets that it seeks to preserve and protect, which would be the targets of retaliatory attacks by other states. Thus sub- and transnational networks cannot be the objects of classical, state-against-state deterrence, a principal foundation of the international order (such as it is) that we have been living in ever since the advent of the nuclear age.

Moreover, one can imagine that technological advances will eventually drive the capacity threshold even lower, from the sub- and transnational network or group to its logical endpoint: just one or two persons by themselves. The long-predicted age of the “super-empowered individual” would at last have arrived.

At present, however, we are only in the midst of the downward transition to the sub- and transnational network. That is still enough to get our attention. Unless deterrence can be re-invented to fit transnational Islamist terrorist networks, we are about to witness the end of an epoch—and the end of many Americans as well.

The reality of the threat from transnational Islamist terrorist networks was clearly revealed on Sept. 11, 2001. But the potential for that threat to also become a nuclear one has been ticking away ever since Pakistan acquired nuclear weapons in 1998. Not only was the Pakistani bomb the first “Islamic bomb” (if not quite an Islamist one), but Pakistan acquired it with the aid of an extensive transnational nuclear network orchestrated by its chief nuclear scientist, A.Q. Khan. It became obvious that Islamists in Pakistan might use a similar transnational network to pass on Pakistani nuclear weapons to other Islamists elsewhere. In any event, other Muslim countries, most obviously Iraq and Iran, seemed to be good candidates to acquire the next Islamic bomb and perhaps to pass it on.

Confronted with these ominous developments, this demonic dynamic that undermined and disoriented classical deterrence, the Bush administration chose to replace deterrence with preemption. This choice and the reasoning behind it were spelled out in September 2002 in the administration’s National Security Strategy of the United States. In principle, the preemption strategy appeared reasonable, but when applied to Iraq, it required making two dubious arguments—Saddam Hussein was acquiring nuclear weapons, and he was also supporting al-Qaeda—and then constructing a dubious connection between them. One might have thought that the debacle of applying preemption to Iraq would have discredited the doctrine, but the Bush administration is now seeking to do much the same with Iran.

But the replacement of deterrence with preemption was premature. The full potential of deterrence had not been explored or exploited. Instead, it is possible to argue that classical deterrence—the deterrence of states—can still be used against one version of the Islamist terrorist threat: Shi’ite terrorism, which emanates from Iran. This is a simpler problem, and it largely can be addressed in the old-fashioned way.

The other version of the Islamist terrorist threat is Sunni terrorism, which emanates from transnational networks, particularly al-Qaeda and its affiliates. This is a complex problem, and to address it, we need to reinvent deterrence and redirect our approach downward from states to communities, be they ethnic groups, tribes, or clans.

In searching the Islamic world for places in which traditional deterrence can be effective against the threat of nuclear terrorism, the obvious place to look is strong states that can be held responsible for their own actions and for the actions of the people (including potential nuclear terrorists) who live within their territory. There are not many such strong states in the Islamic world. Rather, for reasons that seem to be intrinsic to Muslim societies, the normal pattern is authoritarian states that are brutal but also too weak to control all of their territory (e.g., Sudan, Yemen, Saudi Arabia, and Pakistan). Moreover, as the inevitable decay of any regime works its way, the weak state often becomes a failed one (e.g., Somalia, Pakistan as it seems to have now become). There is one very good example of the kind of state we are looking for, however: Iran.

When most Americans think of Iran, they think of a terrorist state, one that promotes transnational terrorist organizations such as Hezbollah and which, through its continuing nuclear enrichment program, is steadily laying the groundwork for obtaining nuclear weapons. Together, these two features will give Iran the capacity for nuclear terrorism.

Iran is certainly all of this, but it is also something more. Like the Soviet regime when it supported an international movement and like the Chinese Communist regime when it supported communist insurgencies in Southeast Asia, the Iranian Islamist regime wants to preserve its state, its territory, its resources, its people, and above all itself, while at the same time promoting revolution abroad. But the goal of preservation takes priority over the goal of promotion. Since the Iranian regime has a great deal to protect, it has a great many assets that are hostages to U.S. retaliation. Thus it is a good candidate for classical deterrence.

Iran is also the major Shi’ite state. Indeed, it is the only Shi’ite power. Its territory of 636,000 square miles and its population of 61 million are immensely larger than any other Shi’ite state. This means not only that the Iranian regime has significant territory and population that it wants to preserve, but it is also the only substantial and sustained supporter of Shi’ite transnational terrorist networks. That is why virtually all Shi’ite terrorists are concentrated within Hezbollah, which Iran supports (and can control). This means that by deterring Iran, it is also possible in effect to deter Hezbollah.

Now if Iran were the only potential source of nuclear terrorism, it would certainly present a grave deterrence problem. After all, deterring the Soviet Union posed a very serious problem indeed, and deterring the Chinese Communist regime still does. However, the problem of deterring Islamist nuclear terrorism can be thought about in the old-fashioned, classical way. We would not be wrestling with what appears to be a new and intractable puzzle.

It is not the established regime holding many valuable assets but the transnational network holding few or none that poses this problem. The danger comes not from the strong state that can control its territory and population but from the failed or failing one that cannot. Here we see a major and important contrast between Shi’ite and Sunni Islamist terrorist networks. Since the Shi’ite networks are in large measure controlled by Iran, we can hold Iran responsible for their actions. Sunni networks, in contrast, are in large measure controlled by no strong state. Rather, they are supported either by substate and transnational actors operating within weak states (like the Wahabist foundations, tribes, and clans within Saudi Arabia) or by official units of the weak states (like the Inter-Services Intelligence agency of Pakistan). If the United States tries to apply the methods of classical deterrence against the Saudi state or the Pakistan state, it is applying pressure at the wrong point; it is like pushing on a wet noodle. By pressing on the weak state, the United States is pressing only indirectly on the actual source of the nuclear terrorist threat, and the pressure is dissipated within the political and bureaucratic miasma that is a failing government.

The obvious solution for the United States is instead to press directly on the actual source of the nuclear terrorist threat. This means descending from the higher but superficial and artificial level of the state down to the lower but substantive and real level of the substate organization, ethnic community, or even local tribe or clan. In other words, just as technological advances have driven nuclear acquisition capacity downward, so too and in response, we will have to drive nuclear deterrence strategy downward.

Consider a vertical dimension of deterrence, ranging from the national state at the top down through the successive levels of the ethnic community and the tribe to the clan at the bottom. At the same time, imagine a horizontal dimension, spanning the wide array of entities that conflict with each other on any particular level: Iran versus Saudi Arabia; Shi’ite community versus Sunni community; Sunni tribe versus another Sunni tribe, and so on. It is the nature of societies composed of ethnic communities, tribes, or clans that on each of these levels there have been long standing conflicts.

Strategists are, of course, very familiar with state-versus-state conflicts. They are also well versed in the strategies of balance of power and divide and rule, by which one state or power preserves or advances its interests by playing different states against each other. What is less well known is that historically powers have also used the same strategies at the lower levels of ethnic community, tribe, and clan. This approach was fundamental to the ways that the British and the French once ran their empires, particularly in the Islamic world. And although it is now virtually forgotten, such strategies were also fundamental to the way that the young United States in the 19th century dealt with the many conflicting Indian tribes (sometimes called “nations”) on the Western frontier.

When we put the two dimensions together, we can see a sort of matrix in which we could, first, identify a particular entity that surrounds and supports either a terrorist group or a local node of a transnational terrorist network, and, second, locate the other entities that surround and conflict with the first entity. The most effective deterrence would be to zero in on the right point in the matrix and also on any adjacent points that can be mobilized against it.

Every Sunni terrorist network, no matter how transnational (like al-Qaeda and increasingly its franchises and protégés), is embedded within a community of people who sustain and support it. Islamist terrorists, just like Mao’s guerrillas, have to swim in a sea of people, and this is true of nuclear terrorists, too.

In the Islamic world, that community almost always has a self-conscious collective identity, with the members thinking of themselves first as a community and not as individuals. When Westerners gravitate toward groups, they define themselves within such weak or empty “communities” as the “urban community,” the “African-American community,” or the “gay community.”

But in the Islamic world, the collective community identity is often so strong as to be institutionalized by the pervasive practice of endogamous marriage. Indeed, in this kind of culture first cousins are often the preferred marriage partner.

This collective identity may exist only at the low level and narrow scope of the clan or the tribe, or it may be at the higher level and broader scope of the ethnic community. With respect to our problem of deterrence, the important level and scope are where the sanctuary and support of the Islamist terrorist network are found. It is here that we find the collective entity that should be held responsible for the actions of the terrorist network and should be the object of the full panoply of deterrence methods before an attack.

One of these methods, of course, is to directly threaten the entity with massive retaliation if a nuclear attack emanates from a terrorist network that it supports. Another and perhaps better method is more indirect: to threaten the entity by supporting and empowering its surrounding and conflicting entities. These other players usually have ample historical reasons for engaging in their own conflicts if empowered to do so.

The place where most of these problems come together in a big and bad way is the wild frontier region between Afghanistan and Pakistan. Neither of the two states can control its part of this region, thus the area makes a good base for the archetypal transnational Islamist terrorist network that is al-Qaeda.

The two parts of this frontier region are the southern and eastern provinces of Afghanistan and the neighboring Northwest Frontier Province and Federally Administered Tribal Areas of Pakistan. The inhabitants are overwhelmingly Pashtuns, and although these territories are formally divided between two (failing) states, they have historically formed one cultural region and, in the mind of the Pashtuns, one country—Pashtunistan.

When the British ruled India, they considered the Pashtuns (whom they called the Pathans) a notoriously unruly people—indeed, they called them “ungovernable.” And so the Pashtuns have remained, right down to the present day. We might now properly call them a rogue people.

The Pashtuns’ roguery has come at great cost to their neighboring ethnic communities: the Tajiks, the Uzbeks, the Hazaras, and the Punjabis. They are now also a rogue people to the rest of the world. They are virtually the only ethnic community in Afghanistan that supports the Taliban. Indeed, almost everyone in the Taliban is a Pashtun. It was, of course, the Taliban regime, and therefore the Pashtun community, that hosted and protected al-Qaeda before the American invasion of Afghanistan in 2001. And it is the Pashtun community in the Northwest Frontier Province and Federally Administered Tribal Areas of Pakistan that gives al-Qaeda a haven there today.

Like many close-knit ethnic or tribal communities, the Pashtuns have an intense sense of communal identity and almost no sense of individuality. They also naturally have an intense awareness of the communal identity, even the collective guilt, of their enemies. It is impossible to deal with the Pashtuns as if they were individuals, responding to calculations of individual benefits and costs. This is why, after more than six years, no one has stepped forward to turn in Osama bin Laden or Mullah Muhammed Omar, the leader of the Taliban, even though the United States has offered a $25-million reward for each. The Pashtun ethical code is “Pashtunwali,” their unique Pashtun way.

The only way to deal with the Pashtuns is the way they deal with themselves and everyone else, as a community, one that is capable of both collective honor and guilt. Perhaps the best way for Americans to think about the Pashtun tribes on the Northwest Frontier would be the way late-19th-century Americans thought about the Apache and Comanche tribes on their own Southwest frontier at that time.

The Bush administration committed a grave error in December 2001 and January 2002 when it did not pursue Osama bin Laden and Mullah Mohammed Omar until they were caught or killed or turned over by their Pashtun protectors to American officials. The administration correctly punished the Taliban for supporting and harboring bin Laden and al-Qaeda, at the level of the Afghan state. But it foolishly did not drive that punishment home, so the sanctuary and support merely migrated to the level of the Pashtun community and local Taliban tribes.

Had the Bush administration carried its retaliation to an effective and conclusive end, it would have achieved many good results. Among them would have been to lay a firm foundation upon which credible deterrence against Islamist terrorists could have been constructed for the future. For that grave act of omission, we have already paid a great price, and we are likely to pay a fortune in the future.

Because the Pashtuns at this new and lower level of community and tribes continue to shelter al-Qaeda, they would be a fitting target for retaliation if any new—or nuclear—attack occurs. This means that they would be a suitable object of deterrence policy today, before an attack.

Since the Pashtuns have been an ungovernable people, perhaps they should be assigned a territory where they can govern themselves and only themselves and no one else and where they can do so in their own Pashtunwali way. If that way includes learning how to construct a state that can be held responsible for the actions of itself and its people, that would be a good thing, and we can support it. If that way instead includes providing sanctuary and support for Islamist terrorists who are bent on acquiring and using nuclear weapons, this will be a bad thing, and we can destroy it.

Of course, it may now be impossible for Americans—with their ideals of individualism, liberalism, and democracy at the very core of their identity—to deal directly with the Pashtuns in such a communal and collective-guilt way. There are, however, other ethnic communities in Afghanistan—the Tajiks, the Uzbeks, and the Hazaras—and even in Pakistan, who have long been dominated or abused by the Pashtuns and who would be willing to do so, if this were allowed by the United States and the other NATO countries now operating in Afghanistan. Of course, to allow the local and historical adversaries of the Pashtuns to deal with them in the local and historical way—and the way of the Pashtuns themselves—would be repugnant to conventional standards of human rights and universal justice. However, sometimes local but generally held conceptions of justice are more fitting to local realities than universal, general ones.

The Pakistani state has always been artificial and brittle. It was created, as both West Pakistan and East Pakistan, in one bloody partition in 1947, and was recreated in reduced form, as West Pakistan alone, in a second partition in 1973. The United States has long tried to make Pakistan into a strong state, and the Bush administration is still trying. The administration’s efforts, however, are obviously failing, and the Pakistani state is failing, too. Even at its strongest, the Pakistani state has never been able to effectively govern the Federally Administered Tribal Areas, which everyone, including successive Pakistani governments, has officially recognized as “autonomous.” That is to say, the local, largely Pashtun tribes have pretty much been able to do as they please—and these tribes have been pleased to provide support for al-Qaeda and the Taliban, so much so that they have now spread out and are conducting terrorist attacks in the rest of Pakistan.

There is a reason behind the weakness, a method to the madness, of the Pakistani state with respect to the Pashtun Tribal Areas. Although almost no one in the U.S. foreign-policy establishment talks about it, the cause can be clearly seen with a look at the map. The strategic center of the rather large country that is Pakistan consists of a rather small area in the northern Punjab and is composed of the major cities of Islamabad (the capital), Rawalpindi, and Lahore. Lahore is only about 20 miles from the frontier with India, and the Islamabad and Rawalpindi border on the Northwest Frontier Province and are only about 50 miles from the border of the Tribal Areas [CHECK]. Moreover, this entire strategic core is only about 150 miles in width.

This means that there has long been an obvious strategy for the Indian military to pursue, if it should ever want to destroy Pakistan, and that is a massive military thrust across the strategic core. An obvious response for the Pakistan military is to seek strategic depth to its rear—and to the west—and that means in the Tribal Areas of Pakistan and even, since it could become essential in the event of an Indian invasion, in the Pashtun area of Afghanistan. Such a strategic rear requires a friendly local host population—the Pashtuns. This is why the Pakistani military, and its powerful Inter-Services Intelligence agency, has long insisted upon close and co-operative relations with the Pashtuns, be they in Pakistan or Afghanistan and no matter how dangerous they might be to everyone else. This strategic link between the Pakistani military and the Pashtun community is the fatal flaw, the toxic dump, of Pakistan as it relates to the rest of the world. Very likely, it could continue to emanate toxins—and terrorism—until Pakistan as a state is broken and dissolved.

Moreover, with a strong Islamist presence in the country and even in the military, Pakistan could one day become an Islamist state, one already possessing nuclear weapons. An Islamist Pakistan, with al-Qaeda operating on its territory, would probably be the most dangerous state in the world, a rogue state in the fullest sense of the term. If the United States should ever determine that this state had to be put to an end, India would obviously be the best one to do it, to “crack the Paks” and bring about a third partition of Pakistan.

In the ruins of this artificial country would be four or five separate ethnic states or provinces—most likely, the Punjab, Sindh, Baluchistan, and, if the Pashtuns should prove themselves capable of rising to the level of a state, Pashtunistan. Each could be reconstructed and ordered by the Indian Raj, with a mixture of direct and indirect rule not unlike the British Raj, which once ruled these very same provinces. And at long last, this artificial state with its wild regions and fatal flaws would be replaced with a strong state that could be held responsible for the Islamists under its rule and that would have every incentive to exercise this responsibility.

The essential U.S. strategic objective in the Islamic world should be the establishment of strong, responsible states that we can hold responsible for their own actions and for the actions of the Islamists who live within them.

The logic of this analysis leads to what many people—at least neoconservatives—will consider a perverse conclusion. We consider it to be more a discerning, if paradoxical, one.

A good example of such a state is Iran. When the United States is dealing with this troublesome country, the worst thing it could do would be to destroy the Iranian state totally so that Hezbollah and other Shi’ite terrorist networks would have no state to control them. They would become unguided missiles or loose cannons, careening around the Middle East and even the globe.

In regard to transnational Sunni terrorist networks, however, we do not now have any obvious candidates for strong, responsible states that can control them. Unless or until these are established, deterrence will have to point in a different direction.

For deterrence to survive in the new nuclear age—the age defined by Sunni Islamist terrorism—it must become focused upon ethnic communities. American strategists will have to learn about the features of specific communities and even tribes, just as they learned about specific states and nations in the old nuclear age. And if deterrence does not survive in this new age, neither will we.

James Kurth is the Claude Smith Professor of Political Science at Swarthmore College, where he teaches American foreign policy, defense policy, and international politics.