Spurred by the progress of weapons programs in North Korea and Iran, nuclear proliferation is once again at the top of the U.S. national security agenda. Practically all of the discussion about the issue has centered on how to prevent proliferation. Hawks have pushed for regime change or military strikes, whereas doves have favored arms control and negotiation. Even though none of these measures is likely to solve the problem, few observers have spent much time considering what a postproliferation world would look like.

Those who have done so can be divided into pessimists and optimists. The pessimists assume that the dangers of a nuclear confrontation will increase exponentially as the number of nuclear powers grows and that a future catastrophe is all but certain. Since little can be done to avert such a terrible outcome or mitigate its consequences, the argument goes, efforts to stop proliferation in the first place must be redoubled. The optimists, by contrast, assume that the stability that nuclear weapons seem to have brought to the superpowers' Cold War confrontation will be replicated. Far from being a sure disaster, they argue, the spread of nuclear weapons could be a relatively cheap and easy (albeit nerve-racking) solution to the age-old problem of war.

Actually, however, a postproliferation future is likely to be far more complex than either the pessimists or the optimists believe. In a multipolar nuclear world, international politics will continue but in an environment dominated by fear and uncertainty, with new dangers and new possibilities for miscommunication adding to and complicating familiar ones. As a result, many of the military plans, defense policies, and national security doctrines that officials in the United States and other countries now take for granted are likely to become obsolete and will need to be revised significantly.


Assume, for the sake of argument, that within the next decade Iran manages to acquire a few crude nuclear weapons and that these can be delivered by ballistic missiles within the Middle East and by clandestine means to the United States and Europe. Assume also that Saudi Arabia and Turkey, out of fear or competitive emulation, also develop their own nuclear arsenals. How would strategic interactions in this new world play out?

During the Cold War, the small number of nuclear states meant that the identity of any nuclear attacker would be obvious. Preparations could thus be made for retaliation, and this helped deter first strikes. In a multipolar nuclear Middle East, however, such logic might not hold. For deterrence to work in such an environment, there would have to be detection systems that could unambiguously determine whether a nuclear-armed ballistic missile was launched from, say, Iran, Turkey, or Saudi Arabia. In earlier decades, the United States spent an enormous amount of resources on over-the-horizon radars and satellites that could detect the origin of missile launches in the Soviet Union. But those systems were optimized to monitor the Soviet Union and may not be as effective at identifying launches conducted from other countries. It may be technically simple for the United States (or Israel or Saudi Arabia) to deploy such systems, but until they exist and their effectiveness is demonstrated, deterrence might well be weak; it would be difficult to retaliate against a bomb that has no clear return address.

It gets worse. During the Cold War, most analysts considered it unlikely that nuclear weapons would be used during peacetime; they worried more about the possibility of a nuclear conflict somehow emerging out of a conventional war. That scenario would still be the most likely in a postproliferation future as well, but the frequency of conventional wars in the Middle East would make it a less comforting prospect. If a nuclear-armed ballistic missile were launched while conventional fighting involving non-nuclear-armed ballistic missiles was going on in the region, how confident would any government be that it could identify the party responsible? The difficulty would be greater still if an airplane or a cruise missile were used to deliver the nuclear weapon.

One of the greatest fears about Iran's possible acquisition of nuclear weapons, moreover, is that Tehran might give them to a terrorist group, which would dramatically increase the likelihood of their being used. Some argue that the Iranian government would never condone such a transfer; others that it would. There is no way of knowing for sure. What can be said, however, is that the likelihood of a clandestine transfer to radical Islamist terrorists will increase if the number of Islamic nuclear powers grows, if only because it would get more difficult to identify the state responsible for the transfer so as to punish it.

If an Islamist terrorist group acquired fissile material or a nuclear bomb today, it would be hard to determine with certainty which country had provided it. Attention would focus on Pakistan, the only Islamic state currently in possession of nuclear weapons. But uncertainty would grow if more Islamic states went nuclear, and retaliation would become all but impossible unless one were willing to strike back indiscriminately at all suspect states.


During the Cold War, the United States and the Soviet Union engaged in an intense arms race and built up vast nuclear arsenals. Other binary nuclear competitions, however, such as that between India and Pakistan, have been free of such behavior. Those states' arsenals have remained fairly small and relatively unsophisticated.

Nuclear-armed countries in the Middle East would be unlikely to display such restraint. Iran and Iraq would be much too suspicious of each other, as would Saudi Arabia and Iran, Turkey and Iraq, and so forth. And then there is Israel. Wariness would create the classic conditions for a multipolar arms race, with Israel arming against all possible enemies and the Islamic states arming against Israel and one another.

Historical evidence suggests that arms races sometimes precipitate wars because governments come to see conflict as preferable to financial exhaustion or believe they can gain a temporary military advantage through war. Arguably, a nuclear war would be so destructive that its prospect might well dissuade states from escalating conflicts. But energetic arms races would still produce larger arsenals, making it harder to prevent the accidental or unauthorized use of nuclear weapons.

Nuclear arms races might emerge in regions other than the Middle East as well. Asia features many countries with major territorial or political disputes, including five with nuclear weapons (China, India, North Korea, Pakistan, and Russia). Japan and Taiwan could join the list. Most of these countries would have the resources to increase the size and quality of their nuclear arsenals indefinitely if they so chose. They also seem to be nationalist in a way that western European countries no longer are: they are particularly mindful of their sovereignty, relatively uninterested in international organizations, sensitive to slights, and wary about changes in the regional balance of military power. Were the United States to stop serving as guarantor of the current order, Asia might well be, in the words of the Princeton political science professor Aaron Friedberg, "ripe for rivalry" -- including nuclear rivalry. In that case, the region would raise problems similar to those that would be posed by a nuclear Middle East.

The United States has not been strategically affected by the peacetime arms races of other countries since the global competition for naval power and the European bomber contests of the 1920s and 1930s. Were such rivalries to emerge now, it is unclear how Washington would, or should, respond. During the Cold War, U.S. and Soviet strategists worried not only about how to protect their own countries from nuclear attack but also about how to protect their allies. Questions about the credibility of such "extended deterrence" were never fully resolved, but their urgency was lessened, in the United States at least, by Washington's decision to bind itself tightly to its NATO partners (going so far as to station U.S. nuclear missiles in West Germany and Turkey). Similar questions will inevitably return if proliferation continues. In a future confrontation between Iran and Kuwait, for example, a nuclear-armed Tehran might well try to coerce its opponent while treating Washington's protests and threats as a bluff. Would heading off such challenges require the formation of a new set of tight alliances, explicit security guarantees, and integrated defense structures?

Another Cold War concept, known as the stability-instability paradox, posits that actors take advantage of the very fear of nuclear war to pursue lesser sorts of conflict with impunity. This, too, might play out in the future. A nuclear Iran, for example, might support increased terrorism against U.S. forces in the region on the theory that Washington would be reluctant to escalate the conflict.


Nuclear weapons have not been used since 1945, and any further use would come as a profound shock. Yet some future nuclear actors might think that resorting to these weapons would serve their interests. It is not inconceivable, for example, that some state or group might want to show the rest of the world that it is willing and able to violate the most hallowed norms of the international system. Nazi Germany deliberately targeted civilian refugees in Poland in 1939 and in the Netherlands and France in 1940 as part of its strategy of Schrecklichkeit (instilling terror). A quasi-terrorist use of nuclear weapons, not by a small group but by a state that wanted to be recognized as fearsome, would be a contemporary analog.

What kind of state might attempt such a thing? If history is any guide, a state that openly rejects the existing international order, considers its opponents to be less than fully human, and seeks to intimidate others. Alternatively, internal conflicts could create hatreds so powerful that actors might resort to using nuclear weapons; consider, for example, how Moscow might respond if another Chechen attack killed hundreds of Russian children. Some states might also be tempted to use nuclear weapons in other ways. For example, before it started to abandon its nuclear weapons program, South Africa had planned to use its bombs if it was ever approaching military defeat, as a last-ditch effort to draw the superpowers into the conflict. If it were to cross the nuclear threshold, Taiwan might embrace a similar strategy.

By far, however, the most plausible use of nuclear weapons would involve a nuclear power that found itself on the losing side of a nonnuclear war. Such a state would be faced with a choice not between maintaining peace and initiating nuclear war but between accepting its impending defeat and gambling that escalation might suddenly end the fighting without defeat. Conflicts between Iran and Iraq or China and Taiwan are plausible candidates for such a nightmare.


What, if anything, can be done to prepare intelligently for such contingencies? Pessimists say preparation is pointless, and optimists say it is unnecessary. But there are several steps prudent officials can and should take now.

To bolster the efficacy of deterrence in a world of small, closely located nuclear powers, it would be necessary to deploy surveillance systems that could identify and warn against aircraft movement and missile launches. These systems might be operated on a national or a multilateral basis; in fact, a number of states in exposed regions could contribute to collective efforts to detect airborne threats.

The construction of such a regional surveillance system, moreover, would put in place much of the infrastructure needed to support another useful tool: some form of missile defense. Skeptics of missile defense have often ridiculed, with some reason, the notion that such systems can be effective against nuclear weapons or large numbers of missiles. What they overlook, however, is that even leaky or somewhat ineffective defenses can play a constructive role in deterring an attack from a nuclear power with a small arsenal or lowering the odds that a full-scale nuclear conflict will erupt from a single use (of whatever origin).

Other kinds of defense could also help lower the odds of an attack or mitigate its terrible consequences. U.S. officials should develop the capacity to evacuate those cities at risk of a direct attack or of being in the path of nuclear fallout, as well as stockpile radiation meters, build fallout shelters, and implement other measures first devised in the 1950s. Civil defense came to be seen as a grotesque joke when the Soviet Union acquired tens of thousands of nuclear weapons. But, like missile defense, it could play an important role in a world of smaller nuclear powers.

Should a nuclear bomb get through nevertheless, it would be critical for the government of the targeted state to respond with policies other than doing nothing or ordering indiscriminate retaliation. One option would be to launch a massive nonnuclear military campaign against the responsible party to make sure that such an attack was never repeated. But even with all the will and money in the world, such a response simply could not be summoned up out of the blue; it would require careful planning and preparation.

The United States was able to launch an extraordinary mobilization effort in 1942 not simply because the attack on Pearl Harbor had galvanized the American public, but also because people such as Vannevar Bush (as head of the National Defense Research Committee) and Robert Lovett (as an aide to Secretary of War Henry Stimson) had spent years laying the groundwork for organizing civilian scientists and the automobile industry for military purposes. Unfortunately, no one has thought about planning for such mobilization in the United States in 50 years. And yet in the aftermath of a nuclear attack on the United States, the American people would demand that the government try to transform the world so that such an attack would never occur again. Starting to think through and plan such an effort now would be cheap and doable, and could dramatically expand the U.S. government's options later.

Meanwhile, controlling arms races in the Middle East and Asia would be difficult. Traditional arms control, which requires that parties allow mutual inventories of their weapons and submit to verification procedures, seems unlikely to be of much use, both because of deep-rooted mutual suspicions in these regions and because small nuclear powers have an incentive to hide their weapons and exaggerate the size and the effectiveness of their arsenals in order to discourage opponents from attacking them. Agreement on serious verification measures would thus be difficult to reach, and that in turn would undercut the prospects for effective agreements more generally.

Still, a peculiar form of nuclear transparency might be possible and helpful. But it would require that all states realize that nuclear war, even with limited arsenals, would result in their own deaths. How can the point be driven home? There is historical evidence that seeing nuclear weapons tested had a powerful effect on many people in the United States and the Soviet Union. Thus, paradoxically, nuclear proliferation might have to continue a while longer before it can be halted or slowed down: were nuclear tests to be conducted in full view again, the current generation of policymakers and their constituents might realize that the use of even a small number of nuclear weapons would lead to intolerable destruction.

In short, if nuclear proliferation continues, the world is likely to move into a new era with challenges significantly different from those of the Cold War. Several more small nuclear powers will likely coexist along with the large ones, and the problems of nuclear deterrence, arms races, offensive and defensive weaponry, and appropriate retaliation will need to be worked out again in new and more complex conditions. From the 1950s through the 1980s, nuclear strategy was a lively field of inquiry of great practical importance. It nearly died out with the end of the Cold War, as the prospects of a superpower nuclear confrontation receded. Unfortunately, it is now due for a revival.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • Stephen Peter Rosen is Beton Michael Kaneb Professor of National Security and Military Affairs and Director of the John M. Olin Institute for Strategic Studies at Harvard University.
  • More By Stephen Peter Rosen