What Are Nuclear Weapons For?

By Ivan Oelrich, 202-454-4682, ioelrich@fas.org

After a two decade gestation since the end of the Cold War, the question of what nuclear weapons are for is finally emerging into a wider public debate. If I took the title of this essay literally, this essay would be very brief indeed because I believe the proper role of nuclear weapons is extremely limited. The United States should declare the narrowest possible mission for nuclear weapons, take its weapons off alert, bring all forward-deployed nuclear weapons home, and reduce its total arsenal to a thousand or fewer unilaterally and then engage the other nuclear powers in further reductions, leading eventually to a world free of nuclear weapons.¹ Thus, much of what follows is a discussion of what nuclear weapons are not for.

The very question, “What are nuclear weapons for?” steers us toward the wrong answer. Nuclear weapons are immensely powerful and are particularly effective at blowing things up. So if we set off to find some military mission that nuclear weapons can fulfill, we will always find something. If we start, however, from the question of what are the nation’s and the world’s security challenges and how can those best be met, then following those lines of inquiry will rarely if ever lead to any plausible, much less optimal, solution that includes nuclear weapons.

Setting aside for a moment the substance of the discussion, the current debate’s nature and tone alone tell us a great deal about the talismanic power of nuclear weapons. Much of the discussion seems to treat nuclear weaponry as a force of nature, and the question is how we should cope with it. Nuclear weapons exist, true, and cannot be “uninvented” but we must remember that people made them and control them. We must not forget that how nuclear weapons might be used and for what purposes is always someone’s decision.

The Cold War still permeates thinking about nuclear weapons in two ways. First, directly: much of the vocabulary and logic of nuclear weapons that is used today was developed specifically to address the challenges of a nuclear Cold War. We must be careful not to apply the shorthand developed then to today’s radically transformed world. Second, our thinking is also affected by the physical legacy. Almost all nuclear weapons in existence today are left over from the Cold War. This legacy subtly, but powerfully, shifts the presumptions of the debate about nuclear weapons. Specifically, it shifts the burden of proof onto those who would shatter the status quo. Arguments that would be dismissed out of hand if used to justify building a nuclear arsenal up from zero are good enough to justify keeping the nuclear arsenal we have. A related logical sleight of hand is to make some argument in favor of nuclear weapons and allow it to imply that we should stick more or less with the status quo of thousands of weapons left over from the Cold War, even if the argument really justifies having only a handful of weapons.

Discussion of the purpose of nuclear weapons usually begins and ends with deterrence. Deterrence and nuclear weapons have become thoroughly entangled in our thinking and rhetoric. Much talk of deterrence is breathtakingly vague. “Deterrence” is rarely defined; how it is achieved is even more rarely discussed. That deterrence is something we need, and nuclear weapons automatically generate it, is typically simply assumed or asserted. Indeed, sometimes nuclear weapons just become a deterrent as when, for example, the nuclear weapons on missiles in submarines are called the “sea-based deterrent” or nuclear weapons in general are referred to as our “deterrent forces.”

Deterrence is, in theory, quite simple. You might be tempted to do something that I do not want you to do; so I must be able to plausibly threaten you with some punishment, to threaten to inflict some pain, such that your action will, on balance, not seem worthwhile.

When discussing the mechanisms of deterrence, advocates for a nuclear deterrent make several logical errors. The first is to carry over the zero-sum, game-theory thinking of the Cold War. If the goal of deterrence is to make your action seem not to be worthwhile, then the value of the action has to be taken into account. During the Cold War, two world systems, liberal capitalism and totalitarian communism, felt that they were locked in a struggle both for their own survival and for control of the future of the whole world. If the prize is the whole world, that is, everything, then I must threaten to inflict near infinite pain, total, nation crushing pain, to make seizing that prize seem like a bad deal. Moreover, in a truly global struggle, there is no out-of-bounds, which means outcomes are not measured in absolute terms but in relative terms. Indeed, during the Cold War, our war plans included not just destroying the Soviet Union but making certain that it could not recover faster than we could. This goal, in turn, means that damage to my opponent can seem like a positive to me. Thus, if we suffer ten million dead and the Soviets suffer one hundred million dead, we somehow come out, not ten million down, but ninety million ahead because we are ahead, not in any absolute sense, but compared to our global foe in a closed system. The zero-sum, no out-of-bounds nature of the nuclear stand-off allowed Cold War thinkers to abstract the competition from any outside context; the contest lent itself to analysis by game theory, computer simulations, and mathematical models of stability. Much of that thinking and vocabulary inappropriately carries forward today.

This approach is irrelevant today because the threat I need to pose is tied to the prizes that you may try to seize, and today the fate of the world is not being contested; the prizes in play are much smaller. At first blush it seems reasonable that if the Russians are tempted to hurl a thousand nuclear bombs at us, we have to threaten to hurl eleven hundred back to deter that attack. That only works if destruction of Russia is our gain and outweighs equivalent pain on our part, which it does not. We have to ask why Russia is hurling missiles at us, what are the stakes in play, and what would make seizing those stakes seem unattractive. We can imagine that Russia and the United States could fall into a war. Just as hypothetical examples, Russia might make an incursion into one of the Baltic countries, which are NATO allies, because of mistreatment of the Russian minority, or into Kosovo in anticipation of NATO’s military enforcement of Kosovo’s independence from Serbia. But how many nuclear weapons have to go back and forth before the destruction makes the importance of the original issue pale in comparison? I cannot say exactly but the number may well be one. If the Russians in any case hurl a thousand missiles at us and we throw eleven hundred back, that is not really deterrence; that is nuclear war-fighting or revenge or something else, and we should not confuse ourselves by calling it deterrence. It is precisely because Russia today has the option of destroying the United States with thousands of nuclear weapons, and vice versa, even though no rational, sane situation would call for such an act, that we should find ways of dramatically reducing Russian and American arsenals.

Nuclear weapons are sometimes promoted as essential to deterrence because of their unique military capabilities. In a remarkable essay written after the end of the Cold War, Stephen Younger, the former associate director of Los Alamos National Laboratory, wrote that to effectively deter the Russians we have to be able to destroy anything and everything in Russia and only nuclear weapons can hope to do that.² If fact, in only one case does he suggest the possibility of negotiating limits with the Russians and that is to limit the possibility that they might build something that we could not destroy.

Drs. John Foster and Keith Payne, in their essay in this series [Physics & Society, October 2007], make essentially the same logical error, arguing that nuclear weapons may be needed for deterrence because some targets cannot be destroyed any other way. But this assumes that my enemy, not I, gets to decide how I inflict pain on him. Their assertion implies that I may be able to destroy North Korea’s army, navy, air force, its infrastructure, economy, transportation, food supply, indeed, its entire population, but if there is some tunnel somewhere that I cannot destroy, then deterrence might fail. This proposition is indefensible on several counts. If nothing else, it is utterly contrary to historical experience. No nation at war has ever had as a goal the utter destruction of every possible enemy target; no war has ever been won or lost on that basis. Moreover, if survival of some targets makes deterrence impossible then deterrence is impossible, first, because certain targets, such as deep tunnels, are immune to attack even by the most powerful nuclear weapons and, second, because we cannot destroy targets that we cannot find (or, perhaps, are not even aware oft). Remember that Saddam Hussein was at large for some time even though we had occupied his country and had troops on the ground. Saddam was finally captured by a soldier with a pistol, not destroyed by a nuclear weapon. With or without nuclear weapons, there will be targets that are immune to attack. Nuclear weapons cannot be essential for an essentially unobtainable goal.

Some argue that nuclear weapons have a special character that makes them the only instruments that can deter in some cases; again, Drs. Foster and Payne’s essay is a particularly clear example of this position. The special cachet of nuclear weapons may be completely illogical—after all, why should a potential enemy care how I destroy targets and inflict deterring pain? —but we are dealing with human beings so perhaps perceptions create their own reality and logic does not always apply. This is a proposition that I believe is impossible to unambiguously prove but, once accepted, equally impossible to clearly disprove. Yet, careful examination undermines the premise that nuclear weapons have some special role in deterrence.

One problem with any historical analysis of deterrence is that successes can be hard to see, but failures are painfully obvious. Every day that a war does not break out can be claimed as a deterrent success, but was war avoided because of the threat of nuclear retaliation, or of conventional retaliation, or because of domestic political considerations, or any of a thousand other reasons, or was war never really seriously taken under consideration, so never really deterred?

One thing that can be proven is that nuclear weapons are not sufficient for deterrence. Since the United States has had nuclear weapons, it has experienced major deterrence failures in China, Korea, Vietnam, Iraq, and numerous lesser cases. What also seems inescapable is that every time there is an aggression that is not met with nuclear weapons, the credibility of nuclear weapons as a deterrent for that type of event is further reduced. It has been over six decades since the United States has used nuclear weapons. Is their use still plausible in response to another event like the Iraqi invasion of Kuwait? Will it be after a hundred years of non-use? The Department of Energy, in justifying the need for a so-called Reliable Replacement Warhead, claims that U.S. warheads, now 98-99% reliable, need to be more reliable—as though the difference between 99, 95, or 90% could make any conceivable difference in any potential enemy’s deterrent calculation—when any technical difference is completely swamped by the implausibility of use created by decades of non-use.

This does not mean that nuclear weapons have no deterrent value. As the physicists among the readers know, we do not measure time directly, we count off some event that we assume is regular, whether it is the rising of the sun, the swing of a pendulum, or the oscillation of the magnetic moment of a cesium atom. If we measure “deterrence time,” not in the passage of years, but in the passage of events, then much time has passed in terms of Koreas, Vietnams, and Iraqs. Even more time has passed in terms of Haitis, Panamas, Rawandas, and Dafurs. And as time passes without nuclear use, the plausibility of nuclear use continues to decline. Thus, it is inescapable: The only way to make the use of nuclear weapons more plausible in these types of cases is to occasionally use them in these types of cases, and I know of no one advocating this. But no time at all has passed in terms of nuclear attack on the United States or its allies. Thus, a nuclear response to nuclear use is as fresh and intensely plausible today as it would have been in 1945. And this is the only justifiable use for nuclear weapons, the use for which a few should be reserved, as a response to nuclear use by others.

The nuclear “posture” we have today, the combination of weapons, their number and characteristics, that we keep them on hair-trigger alert, constantly deployed, many on submarines forward deployed off the coasts of Russia and China just minutes from their targets, demonstrates that the United States maintains nuclear war fighting options including disarming first strikes. Reserving nuclear weapons solely for the mission of responding to nuclear attack, thereby deterring such an attach in the first place, implies a decisive no-first-use posture, weapons off alert, perhaps even stored separately from their delivery systems. And since the pain that must be inflicted today should be proportionate to the stakes in play, not a potential enemy’s arsenal, the number of weapons needed is almost certainly only in the double digits.

Nuclear weapons once dominated security thinking but, as instruments of national power, their time has come and gone. The United States and the Soviet Union once had nuclear-armed surface-to-air missiles and air-to-air rockets, nuclear depth charges and torpedoes, nuclear land mines and demolition charges, and nuclear-armed rockets that could be launched from the back of a jeep. All of these missions have fallen away, not because of arms control agreements or political pressure but because nuclear weapons have been displaced in each case by technologically and militarily superior solutions made available by advances in miniaturized sensors and computers. Nuclear advocates are forced to ever more contrived and convoluted missions to justify nuclear weapons, for example, the nuclear bunker buster, which required very cooperative enemies who buried vital targets just out of reach of conventional attack but not so deep that they were out of reach of even nuclear weapons. Nuclear weapons have simply become almost entirely obsolete.

Moreover, nuclear weapons are no longer morally acceptable except to deter nuclear use. The United Stats used fleets of B-17s and B-29s to carpet bomb German and Japanese cities in World War II in part because that was the greatest degree of targeting discrimination that the technology of the day allowed. Using B-17s against Baghdad in the same way in the recent war would have been universally denounced as a war crime because today militarily effective alternatives of greater discrimination exist; similarly, using nuclear weapons when any other alternative is available, now that alternatives are available, is immoral.

Nuclear weapons loom so large in the national security calculus today primarily because of inertia, because of the legacy of the Cold War. The question in the title of this series, “What Are Nuclear Weapons For?” is the wrong question. Any analysis that involves nuclear weapons will, of course, find missions for them. But any analysis that starts with security challenges facing the world and rationally examines alternatives will rarely lead to nuclear weapons as the optimal solution. There will be non-nuclear alternatives that are better, whether measured by military, technical, cost, moral, or political criteria. With repeated iterations of the process of elimination, we are finally left with virtually no missions for nuclear weapons at all. The United States should lead the world toward their elimination.

Ivan Oelrich is the Vice President for Strategic Security Programs at the Federation of American Scientists. He received his BS from the University of Chicago and a PhD from Princeton University, both in chemistry. He has held research positions at the Technical University of Munich in Germany, the Institute for Defense Analyses, the Center for Science and International Affairs at Harvard University, the Congressional Office of Technology Assessment, and the Advanced Systems and Concepts Office of the Defense Threat Reduction Agency in the US Department of Defense. He is an Adjunct Professor in the Security Studies Program at Georgetown University.