The debate surrounding U.S. nuclear policy focuses too narrowly on reducing the number of nuclear weapons in the American arsenal toward zero. More important is preventing the use of nuclear weapons in whatever numbers they exist. President Barack Obama should articulate a narrowed framework for the legitimate use of nuclear weapons that the United States believes would be defensible for others to follow as long as nuclear weapons remain.

A More Defensible Nuclear Doctrine

Threat assessment: The first use of nuclear weapons is unnecessary or irrelevant to defeat threats to the territory of the United States today. However, some U.S. allies face potential threats that they rely on the United States to deter, including via possible first use of nuclear weapons. The United States and other states tend to exaggerate the threats that justify their reliance on first-use nuclear deterrence, but all nuclear-armed states can do more to clarify that they will not seek or employ capabilities that could cause others legitimately to use nuclear weapons in self-defense. 

The proposed policy in a nutshell: The United States should declare that it possesses nuclear weapons only to respond to, and thereby deter or defeat, threats to its survival or that of its allies, particularly stemming from any use of nuclear weapons.

Differences from existing policy: This policy would raise the threshold of nuclear use to “threats to survival” instead of “extreme circumstances.” The first use of nuclear weapons would be allowed only in response to existential threats to the United States or its allies, eschewing attempts to conduct disarming first strikes against Russia or China. The policy would be more consistent with U.S. interests in strategic stability and more consonant with just war doctrine and international law. 

Objectives of Nuclear Policy

  • Contribute to overall deterrence of threats to the survival of the United States and its allies.
  • Minimize probability of any nuclear use and escalation.
  • Reduce incentives for other states to acquire or expand nuclear arsenals.
  • Enhance credibility of the deterrence policy by making it a model that the United States would recognize as morally and legally defensible if other nuclear-armed states copied it.

Guidelines for Implementation

  • If non-nuclear means fail, nuclear weapons could be used to defeat, through direct destruction of military forces and demonstration of escalatory risks, any existence-threatening incursion into the territory of the United States or that of an ally, and eliminate the adversary’s will to continue the war. 
  • Disarming first strikes against Russian or Chinese nuclear forces would be eschewed because they are not feasible, and if they were feasible, or were perceived to be feasible, they would drive Moscow and Beijing to seek countervailing capabilities and policies that would make the United States less secure. 
  • U.S. retaliation for an adversary’s nuclear first use would be directed to destroy the military and security apparatus and leadership of the attacking state. 

Concrete Steps Forward 

  • This policy could be conveyed in the Pentagon’s Quadrennial Posture Review, which is due in 2014.
  • The policy could be central to the debate on reducing the role of nuclear weapons at the 2015 Nuclear Non-Proliferation Treaty Review Conference. 

More Than Numbers: Preventing Use 

George Perkovich
Perkovich works primarily on nuclear strategy and nonproliferation issues; cyberconflict; and new approaches to international public-private management of strategic technologies.
More >

In Prague on April 5, 2009, President Obama declared that the United States is committed to seeking “the peace and security of a world without nuclear weapons.” All presidents since Truman had said more or less the same thing, except George W. Bush. Obama was perceived to really mean it.1  But he added a caveat that proponents and opponents of Obama and of disarmament tend to neglect2 —as long as nuclear weapons exist, “the United States will maintain a safe, secure and effective arsenal to deter any adversary.” 

Over the next twenty months, the Obama administration achieved a number of objectives in what became known as the Prague agenda.3  The administration released a new Nuclear Posture Review that acknowledged the decreasing contingencies in which the United States might need nuclear weapons to defend itself and its allies. Obama convened leaders of 46 other countries, including 38 heads of state, in a Washington summit to strengthen cooperation in preventing nuclear terrorism. U.S. leadership helped broker agreement by 191 states in May 2010 to reaffirm their commitment to the Nuclear Non-Proliferation Treaty (NPT). And in December 2010, the U.S. Senate ratified the New START Treaty, further verifiably reducing the number of strategic weapons to be deployed by the United States and Russia. 

Then progress stopped.

Four years after the Prague speech, Iran continues to defy the United Nations (UN) Security Council and the International Atomic Energy Agency by refusing to clarify its past nuclear activities and to take steps to increase international confidence that it will not build nuclear weapons. North Korea, the only state that has ever violated the Nuclear Non-Proliferation Treaty and acquired nuclear weapons, remains intransigent. China, expanding and modernizing its nuclear arsenal, competes with its neighbors with increasing intensity over uninhabited islands in the South and East China Seas. Factions in Japan and South Korea, mindful of North Korea and China, urge greater defense preparedness and privately caution the United States against further reductions in the role and numbers of its nuclear weapons. Pakistan is developing battlefield nuclear weapons. 

Russian military and political leaders exaggerate threats from the United States and the North Atlantic Treaty Organization (NATO) and trumpet new programs to muscle-up Russia’s nuclear arsenal. NATO’s easternmost members seek clearer commitments of American military power and resolve, urging retention of U.S.-NATO nuclear weapons in Europe. France remains ready to block any initiatives within NATO to reduce the role of nuclear weapons. 

The multilateral goals of the Prague agenda remain unfulfilled, too. There is little prospect that the Comprehensive Nuclear-Test-Ban Treaty will enter into force anytime soon. The United States, China, India, Pakistan, Israel, North Korea, Iran, and Egypt are among the states that have not ratified the treaty and are required to do so in order for it to enter into force. Negotiations of a treaty to end the production of fissile materials for military purposes still have not started. A much-vaunted conference on creating a zone free of weapons of mass destruction in the Middle East, which was to be held in 2012, was instead postponed indefinitely. 

In Washington, proponents of new nuclear weapons and opponents of arms control have effectively mobilized against the president’s Prague agenda. These antagonists either genuinely fear that Obama will put the U.S. arsenal on a slippery slope toward zero or tactically insinuate that this is the case in order to weaken the president. They prefer to fund and develop new nuclear warheads and delivery systems better suited for use in the post–Cold War world, which they view as more credible deterrents. 

Others are less hostile to Obama and arms control and do not advocate new nuclear-weapon capabilities but believe that the president must display greater commitment to modernizing the aging U.S. nuclear-weapons complex and arsenal. These voices of resistance mobilized quickly in February to challenge Obama’s new call for further reductions in the deployed U.S. and Russian nuclear arsenals. 

Unless all of these actors find a way to cooperate, nuclear arsenals will not be reduced globally and nonproliferation rules will not be strengthened or robustly enforced. The U.S. president alone cannot make this happen. Even if he overcame the internal conflicts within Washington, he would be unable to shift all the relevant calculations in Moscow, Beijing, Tehran, Pyongyang, Islamabad, Tel Aviv, Seoul, Tokyo, and other capitals. 

Much can still be done, though.  While the stalemate persists over whether and how to lower the number of nuclear weapons, the underlying goal of preventing these weapons from being detonated can be pursued in other ways. Nuclear-armed states cannot be forced to relinquish their weapons, but they can be deterred by the military, political, economic, and moral costs of being the first to use them. Military deterrence will operate as long as nuclear weapons and the knowledge to make them exist. What is needed now is the added deterrent power of international political, moral, and economic pressure on any actor that would break the established taboo against the first use of nuclear weapons.4  

A taboo is an especially powerful norm that is produced to protect individuals or societies from exceptionally dangerous behaviors. The nuclear taboo against first use has emerged over time through frightening episodes of near misses and through conscious actions of individual leaders and social movements.5  Strengthening the first-use taboo requires parallel focus on delegitimizing and reducing threats that can cause nuclear use to appear necessary and justifiable. All states have the right to self-defense, guided by the UN Charter, the laws of war, the precepts of just war that state “the anticipated collateral damage” of an action cannot far exceed the expected military advantage to be gained, and other civilized norms and laws. In practice, this means any society confronting aggression on a scale that could threaten its survival will do what it can to deter or defeat such a threat. If nuclear weapons were the only means sufficient to defeat such aggression, it would be reasonable to expect a state to use them, though such reprisal would not absolve the state from responsibility to minimize the effects, especially on noncombatants. 

Thus, the vital objective of preventing the first use of nuclear weapons requires concerted political and diplomatic exertions to induce states to clarify in word and deed that they will not threaten the survival of others, including Taiwan and Palestine. This is less quixotic than concentrating only on reducing the number of nuclear weapons to zero or, alternatively, believing that first-use nuclear deterrence will prevent the spread and use of these weapons forever. 

An effective nuclear policy for the United States would serve the following imperatives: 

  • Contribute to overall military deterrence of threats to the survival of the United States and its allies 
  • Minimize the probability that the United States and any other state will initiate use of nuclear weapons
  • Minimize the risks of escalation if first use occurs
  • Reduce incentives for other states to seek or expand nuclear arsenals 
  • Enhance international respect for the laws of war, just war, and international humanitarian law.6  

Clearly, there are tensions among these imperatives. No nuclear policy is free of ambiguities, contradictions, and uncertainties.7  In a world where no state enjoys a monopoly on nuclear weapons, initiating use of these weapons against a nuclear-armed adversary (or its allies) carries inherent risks of mutual devastation. And in a world with a moral-political taboo against using nuclear weapons to attack non-nuclear-weapon states, the consequences of such use would, over time, also be self-defeating. Yet, as long as these weapons exist, their possessors must devise policies that legitimately deter threats that could cause nuclear weapons to be used. Such policies should follow the principle the United States has recently suggested for the use of drones: “if we want others to adhere to high and rigorous standards for their use, then we must do so as well. We cannot expect of others what we will not do ourselves.”8

A policy that can meet these objectives better than alternatives should be based on an analysis of the threats that could legitimately necessitate the first use of nuclear weapons and recognition of the domestic dynamics that influence how threats and nuclear policies to meet them are defined and articulated. In the United States and elsewhere, the formulation of nuclear policy reflects domestic factors at the individual, bureaucratic, commercial, and partisan levels as much as it reflects informed assessments of how actual adversaries are likely to respond to a particular nuclear doctrine and force posture. 

Thus far, these dynamics have led to a policy that declares that the “United States “would only consider the use of nuclear weapons in extreme circumstances to defend the vital interests of the United States or its allies and partners . . . [and] will not use or threaten to use nuclear weapons against non-nuclear-weapon states that are party to the NPT and in compliance with their non-proliferation obligations.”9  The policy recommended here would declare instead that the United States possesses nuclear weapons only to respond to, and thereby deter or defeat, threats to its survival or that of its allies, particularly any use of nuclear weapons. The primary differences between the current policy and the proposed one concern the threshold of “extreme circumstances” as compared to “threats to survival” and the question of first use as compared to retaliatory use. Both of these differences have implications for compliance with the precepts of just war and international humanitarian law. 

While the recommended policy, unlike current policy, would eschew first strikes to disarm Russia or China (because these two states have or can be driven to acquire the capacity to engage the United States in a destabilizing nuclear competition), it would leave open the possibility of first use in response to a non-nuclear aggression that threatened the survival of the United States or its allies. This is the major difference between the recommended policy and that of no first use. It is an alternative to declaring that the sole purpose of U.S. nuclear forces is to deter nuclear use by others.

The timing of this intervention may seem peculiar. It was only three years ago that the Obama administration issued its Nuclear Posture Review. Bureaucratically, it would be very difficult to make major revisions so soon. However, this does not preclude further analysis and debate over how U.S. declaratory policy should evolve. The parties to the NPT will meet in 2015 to review progress on implementing the treaty and consider means for strengthening it. As always, the vast majority of parties will seek evidence of progress in nuclear disarmament. Such evidence will be scant if the United States and Russia do not overcome obstacles to further reductions of strategic and theater nuclear weapons; if China, Russia, India, and Pakistan continue to expand and modernize their nuclear arsenals; if the United States, China, India, Pakistan, Israel, Iran, and other states do not ratify the Comprehensive-Test-Ban Treaty; and if negotiations of a fissile-material-production-cutoff treaty remain blocked. 

Meanwhile, a significant number of “middle powers” are organizing to press for further reductions in the role of nuclear weapons in national security policies and for acknowledgment that humanitarian law applies to the potential use of nuclear weapons (including the Non-Proliferation and Disarmament Initiative led by Australia and Japan and a joint statement on the humanitarian dimension of nuclear disarmament signed by over 30 countries).10  If these initiatives are ignored or rebuffed, many important states and civil society leaders will conclude that the disarmament-nonproliferation bargain at the heart of the nonproliferation regime is defunct. Unfairly or not, these states and international civil society will focus more on the perceived shortcomings of the United States than of the other nuclear-armed states and Iran, especially if Washington does not demonstrate renewed vigor in shrinking the shadow of nuclear threats. 

One of the few ways that President Obama could restore confidence in U.S. intentions would be to update the declaration of the role of nuclear weapons in U.S. security policy, including in defense of its allies. In his searching Nobel Peace Prize speech, Obama recognized the occasional inescapability of war and the imperative of waging it justly. So, too, Obama now could examine how the ongoing existence of nuclear arsenals, even if temporary, can be reconciled with the moral-strategic imperative to prevent their use. The president could articulate a limited framework for the legitimate use of nuclear weapons that the United States believes would be defensible for others to follow as long as nuclear weapons remain. Such a nuclear policy could then be conveyed in the Pentagon’s Quadrennial Posture Review, which is due in 2014. Debate on the relative merits of the current U.S. policy and possible alternatives may encourage movement in this direction.

Threats That Require Nuclear First Use

Public discourse assumes that the United States wields nuclear weapons for deterrence. This is usually understood to mean that the United States will retaliate in kind if someone attacks it with nuclear weapons and that this reality should make others eschew aggression against America and its allies. However, this common understanding is somewhat mistaken. There can be strategic advantages to using nuclear weapons before the other side has done so. 

A full discussion along these lines requires a detailed understanding of what constitutes first use. The matter is not self-evident. For purposes of public debate, first use is when the United States would order the release of some fraction of its nuclear weapons toward targets before it detected that the adversary had ordered the release of its nuclear weapons to strike the United States, its armed forces, or its allies. 

The United States reserves the option of using nuclear weapons first in several scenarios. In response to massive non-nuclear aggression, the United States can use one or a few nuclear weapons against the adversary’s conventional forces to demonstrate resolve and seek a de-escalation of the war. Alternatively, Washington can launch a much larger and riskier preemptive first strike against the adversary’s nuclear forces and command and control centers. U.S. nuclear war plans contain multiple variations of first-use options with the purpose of putting the burden on adversaries not to undertake major aggression against the United States or its allies. Thus, U.S. nuclear action would be “retaliatory” in the sense that the adversary’s initial aggressive act would have started the escalatory process but not in the commonly assumed sense of responding to the adversary’s use of nuclear weapons. 

There are reasons for the persistence of the somewhat-misleading notion that U.S. policy is to use nuclear weapons only in retaliation for the adversary’s use of nuclear weapons. It would be problematic for America’s moral self-regard to trumpet that it plans to use nuclear weapons before the other guy. First use also highlights the horrifying implications of nuclear war: if the adversary has a survivable nuclear arsenal, the disarming first strike would almost guarantee that the United States or its allies will be hit by nuclear weapons in return. 

Much more limited first use, such as a “demonstration” detonation or limited attack on advancing conventional forces, could be attempted to compel a nuclear-armed aggressor to reverse course and terminate a conflict. But the adversary could choose to counter in an escalating cycle. In the case, a U.S. decision to use nuclear weapons first would begin a mutually suicidal nuclear chain reaction. The escalatory implications of first use are at least partly why no state has tried to use nuclear weapons against a state that could retaliate in kind.

Nevertheless, policymakers have preferred to leave adversaries to contemplate the possibility of first use while obscuring it from their own populations. For this reason, the U.S. Nuclear Posture Review does not explicitly use the terms “preemption” or “first use,” but both are “allowed” under its terms.

Notwithstanding the benefits of obfuscation, first use is clearly the central issue that needs to be understood and debated more than retaliation for another actor’s nuclear attack. Exploring the dilemmas arising from first use is not necessarily to advocate the wisdom and efficacy of declaring a no-first-use policy. Rather, the purpose of such exploration is to question when, where, and why it would be necessary and justified to be the first to use these weapons in a conflict and to determine the implications of so doing. 

In the 2010 Nuclear Posture Review, the United States declared that it “will not use or threaten to use nuclear weapons against non-nuclear weapon states that are party to the NPT and in compliance with their non-proliferation obligations.” Thus, the nuclear deterrence challenge for the United States relates to states that either possess nuclear weapons or are violating their nonproliferation obligations. 

Under this policy, the United States could use nuclear weapons first. Most likely, this would occur in a conventional war that was escalating in destructiveness to the point where U.S. leaders would calculate that the adversary saw no other option but to use nuclear weapons to forestall defeat. In this scenario, the United States should deter or preempt such a move either by detonating a limited number of weapons on attacking conventional forces or by carrying out a large, disarming strike on the adversary’s nuclear forces. 

Scale is vitally important to the question of first use. Against a nuclear-armed adversary that can predictably retaliate, the only threat against which first use would be credible is one that is so existentially great that the risks of not using U.S. nuclear weapons first are worse than the risk of the nuclear exchange that could follow, either to the United States or its allies. Using nuclear weapons to deter or defeat threats that are less than existential invites harm greater than that posed by the initial injury. 

Moreover, first use carries a much greater strategic, moral, and political burden than retaliatory or second use. The laws and norms of necessity, reprisal, proportionality, and discrimination limit the usability of nuclear forces because behaving justly is vital to sustaining the morale of the U.S. military and winning public support in the United States and among allied and other states. Ultimately, U.S. power depends on it. This in turn affects the credibility of the U.S. deterrent. 

Fortunately, it is very hard to find non-nuclear threats to the United States or its allies that are so grave that Washington would justifiably and credibly threaten to use its nuclear weapons first to defeat them. These few scenarios involve extended deterrence—that is, the commitment the United States makes to numerous allies to come to their defense if they are attacked. Indeed, there seems to be a near consensus among cognoscenti that the extended-deterrence challenge is the most credible one on which U.S. nuclear policy is now based. 


Russia remains a principal object of U.S. nuclear deterrence. It is difficult to find a knowledgeable, responsible person who thinks Russia will initiate large-scale conventional or nuclear aggression directly against the United States, though there are some exceptions.11  Nevertheless, the number and operational plans of U.S. and Russian forces continue to follow a Cold War paradigm, giving each president the option to preemptively attack the other side’s nuclear forces before they can be launched. Besides being anachronistic and futile in the U.S.-Russian political-strategic context, this paradigm also distorts perceptions of the nuclear forces and doctrines required to deal with states other than Russia.12  

Realistic American officials and experts focus more on the possibility that Russia could blackmail or otherwise aggress against new, relatively weak NATO members on its borders. Russia possesses conventional military forces theoretically sufficient to threaten the sovereign existence of the smallest neighboring NATO states if such an aggression could be completed before NATO mobilized its full capacity to defeat it. 

Some cite the 2008 conflict between Russia and Georgia as an example of this threat, perceiving the event as Russian aggression. On the night of August 7, Georgian forces entered South Ossetia, a small territory that was seeking independence from Georgia with Russian encouragement and where Russia and Georgia both had deployed peacekeeping forces after a conflict in the 1990s. Georgia said it was intervening to support Georgian villagers in the region. Russia disputed this claim, saying the Georgians were seeking to reconquer by force a land they had lost in 1992. The Georgian intervention caused casualties to Russian peacekeepers. Moscow immediately mobilized ground, naval, and air forces that expelled Georgian forces from South Ossetia and penetrated into undisputed Georgian territory. The European Union under French presidency negotiated a ceasefire on August 12. Eventually, Russian forces withdrew from undisputed Georgian territory but remained in South Ossetia and Abkhazia, which Moscow recognized as independent states once the conflict was over. 

The Georgian case raises questions about extended nuclear deterrence. Georgia was not a NATO member. (At the NATO summit in Bucharest four months before the conflict broke out, NATO members stopped short of offering either Georgia or Ukraine a Membership Action Plan but made an unclear pledge in their final declaration that “these countries will become members of NATO.”) If Georgia had been a member of NATO, Russia would have had to calculate more carefully whether and how to intervene militarily, in part because of the nuclear shadow. However, it is also true that if Georgia had been a member of NATO, Washington and other allies probably would have pressed it much more rigorously desist from provocative behavior. 

This raises an issue that is often neglected: extended deterrence is a two-way relationship. The provider of security guarantees—in this hypothetical case NATO, with the United States in the leading role—must demonstrate its capabilities and resolve to bear the risks of military action in order to deter the adversary and reassure the protected ally. But the ally should also reassure its protectors that it will take all steps necessary to contribute to collective defense and manage relations with potential adversaries in ways that minimize the potential for conflict. 

Moreover, investigations do not confirm that Russia fired the first shots in the 2008 conflict. The matter remains ambiguous. Uncertainty over who started the fighting surely would have complicated NATO’s and Washington’s willingness and justification to respond militarily, let alone with first use of nuclear weapons. This sort of ambiguity surrounding the causes and initial actions of conflict is quite likely to exist in potential scenarios in East Asia, Europe, and the Middle East where the United States extends its deterrent power.

The limited scale of Russian objectives and operations in the Georgia conflict also raises doubts about the relevance of nuclear use.13  Few reasonable people would have argued that this was an occasion to break the nuclear taboo. 

Resurgent Russian nationalism heightens concerns that Russia could undertake non-nuclear aggression against tiny NATO-member Estonia that could trigger Estonian and American calls to initiate nuclear use. Perhaps the most likely scenario would be a crisis that began with the perceived maltreatment of the Russian minority in Estonia. Mutual recriminations would ensue, with Russian demands that the offense be redressed. It is unlikely that Estonia, in such a scenario, would be contrite enough to appease Moscow. Russia might then mobilize its nearby conventional forces, seeking to intimidate leaders in Tallinn. NATO would in turn be prompted to stand up for Estonia, communicating the gravity of the situation to Moscow and mobilizing NATO military assets to demonstrate resolve. 

It is impossible to predict whether and how such a test of wills would be managed diplomatically, but it can be imagined that Russia would press the issue by moving forces into Estonia to occupy a small piece of its territory before NATO forces were mustered to the tiny country. Moscow would seek to compel Estonian authorities to make amends to the Russian population. Russia’s local military advantages could make such an incursion possible before NATO was sufficiently prepared politically or militarily to prevent it. Estonia and other NATO states might not know whether Russia’s military objectives were limited or instead posed a grave threat to Estonia’s survival as an independent state. If NATO’s conventional mobilization were insufficient or too late to stay Russia’s hand, Estonia could request that the United States or NATO issue a nuclear threat to motivate Russia to desist and withdraw. 

Presumably, though, NATO leaders would calculate that if the initial Russian aggression were limited and not aimed to destroy the sovereign Estonian state or its people, first use of nuclear weapons by NATO would increase the probability that Estonia would be destroyed in the ensuing conflict. In this case, nuclear first use would be a cure worse than the initial threat. The more prudent step if Russia had a conventional advantage would be for NATO to fire a nuclear “warning shot,” perhaps at a Russian naval target or other installation where casualties would be minimal. How Russia (and NATO’s 28 member societies) might react to that form of first use is anyone’s guess. 

Regardless of whether an allied nuclear threat would be forthcoming—and whether it would compel Russia to withdraw—this scenario is the sort that makes some NATO and U.S. officials and experts conclude that the first use of nuclear weapons should remain a viable option.14  However unlikely and problematic, it is one of the strongest that proponents of first use now proffer. 

Clearly there is a moral-strategic hazard in relying on the first use of nuclear weapons to deter (or defeat) threats of non-nuclear aggression against Estonia or other small NATO states (as well as countries in East Asia). For NATO, the hazard is that relying on the effectiveness of the nuclear first-use deterrent might tempt members to engage in risky behavior with the expectation that first use of nuclear weapons will bail them out from the consequences. Another hazard would be underinvestment in reserves of conventional military strength and preparedness. For Russia, the hazard is that leaders relying on nuclear weapons to compensate for overall conventional military disadvantages compared to NATO could precipitate conflicts that Russia would ultimately lose, either through conventional war or mutually devastating nuclear exchanges that would hardly be worth the putative gains sought by the initial coercion. It would also be hazardous for Russia not to clarify through rhetoric and diplomacy that whatever grievances it might have with NATO states, the Kremlin will not initiate military action against them. 

These moral-strategic hazards of relying too heavily on first-use nuclear deterrence resemble what befell overleveraged international financial institutions in 2008. Many investment institutions’ risk models were faulty and emboldened lenders to make bets that they lacked the cash reserves to cover. The risk assumptions in first-use deterrence models may be similarly flawed. The corrective, which would be analogous to increasing cash reserves in banks and exercising more rigorous controls on lending, is to take identifiable steps to bolster non-nuclear means of deterrence and alliance reassurance and to preserve stable relations between NATO states and Russia. 


U.S. attention is pivoting to international security challenges in Asia. Here, in the words of former director of national intelligence Admiral (retired) Dennis C. Blair, 

[T]he United States has the capacity to achieve its war aims in conflict with North Korea, China and Iran without the use of nuclear weapons. . . . The most likely circumstances of nuclear exchanges in these wars arise from American military superiority at the conventional level of war. With the United States on the way to victory, the governments of North Korea, China or Iran might threaten or actually use nuclear weapons to attempt to stop the war short of complete defeat.15 

In each of these possible cases of war, the United States presumably would be involved due to commitments to defend the security of allies and friends such as Taiwan, Japan, South Korea, the Gulf Cooperation Council states, Israel, and Turkey. 

Admiral Blair’s scenario posits that the conventionally inferior nuclear-armed adversary would initiate the use of nuclear weapons, against which the United States would retaliate. This does not mean that Washington would not use nuclear weapons first to preempt the adversary’s first use or would rule it out. But Blair does reflect the American military’s confident belief that it can fight and win without first use of nuclear weapons.

China is rising in importance as an object of extended nuclear deterrence for the United States. A relatively well-studied scenario in which both the U.S. and Chinese nuclear deterrents are relevant involves Taiwan. Since the late 1970s, Washington, Beijing, and Taipei have evolved a modus vivendi to prevent instigation of military conflict. The United States and Taiwan have an understanding that Taiwan will not do anything dramatic to change the status quo and precipitate a military conflict, for example, by declaring independence. In parallel, China will not initiate the use of force against Taiwan, again partly to avoid the risk of nuclear war with the United States. 

The operability of this understanding is neither inevitable nor necessarily permanent. If Taiwan were to break its pledge, Washington’s obligation to defend it is left deliberately ambiguous. Yet, if China were to launch massive military operations against Taiwan unprovoked, the United States would be obligated to help Taiwan respond, and this could lead to an escalatory process. Even here, however, it is far from clear that Washington would find it advisable to use nuclear weapons first, given the likelihood of Chinese retaliation.16  China has always insisted that it possesses nuclear weapons to deter others from making nuclear threats to coerce it and that it would not be the first to use nuclear weapons in a conflict. But if Beijing contravened that stance and decided to use nuclear weapons first, which is far from certain, Washington would feel relatively free to retaliate in ways that it judged most likely to terminate the war as favorably as possible and with the least damage to the United States and Taiwan.

In any case, the first-order objective regarding Taiwan is to motivate China and Taiwan not to take actions that could trigger a military escalation process. All parties recognize these imperatives. Nuclear weapons today are an important background condition of this motivation, but they are not the foreground priority. Nor is first use necessary for either side to maintain current deterrent effects.

Of course, pressure could mount to increase reliance on the potential first use of U.S. nuclear weapons if the combined conventional military power of the United States and its regional partners is allowed to become weaker relative to China’s, which is growing. However, this potential challenge is contingent on choices that the United States, its allies, and China will make in the future. They have options. The most desirable would be to combine constructive political vision, diplomacy, and careful defensive acquisitions to promote stability in Northeast Asia and resolve the security dilemmas that are growing more acute there. 

The Obama administration has recognized the necessity of preventing decisive conventional imbalances. If the conventional balance shifts significantly due to Chinese exertions and technological breakthroughs or due to U.S. funding and policy decisions, Taiwan reasonably could fear that the decades-old modus vivendi with China was threatened. In that case, Taipei would press Washington to reestablish a stout conventional balance. If that failed, Taiwan could be expected to become more accommodating to Beijing or, alternatively, to seek a nuclear deterrent of its own. 

The United States would then need to consider carefully whether such a Taiwanese move would be less desirable than alternatives. China would have to make a similar calculation. It is likely that Beijing would conclude that preventing Taiwan’s acquisition of nuclear weapons was imperative, which would mean China should take care to moderate its own acquisitions and deployments of conventional military capabilities directed against Taiwan and to affirm the wisdom of maintaining the current Chinese-Taiwanese-U.S. understandings and policies. Indeed, the admittedly provocative scenario sketched here underlines the importance of reinvigorating commitment to the existing tripartite modus vivendi.

Of growing salience today is the emerging problem of the disputed islands in the South and East China Seas. China claims a number of these islands and surrounding seabeds, which contain fossil fuel reserves and other resources. Some states in the Association of Southeast Asian Nations (ASEAN) and Japan have countervailing claims. The parties have not agreed on submitting these disputes to UN or other international arbitration. The United States generally has not taken a legal view as to whose claims in the various disputes are more or less valid, but Washington has said that its commitments to defend its allies apply to potential conflicts over these islands. 

In the midst of such ambiguity, is it conceivable that the United States would use nuclear weapons first in a campaign to prevent or remove Chinese forces that had occupied these islands? Would such use be necessary and proportional to the threat, given the risks of escalation? A senior U.S. military official recently told the Washington Post, “I don’t think that we’d allow the U.S. to get dragged into a conflict over fish or over a rock. . . . Having allies that we have defense treaties with, not allowing them to drag us into a situation over a rock dispute, is something that I think we’re pretty all well-aligned on.”17  

Far from weakening extended deterrence, this military official was injecting necessary realism to convey the two-way responsibilities that extended deterrence entails. He was protecting against the potential moral and strategic hazard of allies relying on a nuclear-armed protector undertaking first use to bail them out of crises that should be averted in the first place by both diplomacy and the deployment of more credible means of conventional defense. It is imprudent to think that the U.S. Congress and public would support initiation of nuclear war in a conflict triggered by a dispute between China and Japan over uninhabited or lightly populated islands that the vast majority of the population could not name or find on a map. The possibility that Washington would be put in a position of threatening first use of nuclear weapons in such a conflict would undermine the United States’ overall power and position in the world. 

Rather than apply nuclear deterrence to the problem of territorial disputes in the South and East China Seas, U.S. policy should concentrate on inducing allies to develop conventional capabilities. This would involve enhancing the policies, training, and other forms of cooperation useful to strengthen conventional deterrence while seeking diplomatic ways for China and its neighbors to lessen tensions and security dilemmas. It is a false form of reassurance to allow allies to think that the United States would or should initiate the use of nuclear weapons in a conflict escalating from a dispute over uninhabited islands whose sovereignty has not been vetted by relevant international bodies. 

North Korea, Pakistan, and Iran

Other nuclear-armed states pose challenges to the United States and its allies that may be defined as threats. North Korea and Pakistan come readily to mind, as does Iran if it were to acquire nuclear weapons. But analyzing the nature and scale of the challenges posed by these states further narrows the relevance and effectiveness of first-use deterrent postures and doctrines in countering them.

North Korea could attack South Korea or Japan with nuclear weapons or take operational steps to prepare to do so imminently. This could invite the use of American nuclear weapons in response, which is comparatively straightforward and therefore relatively unlikely. 

More complicated would be a potential massive conventional-artillery campaign by Pyongyang against South Korea. Theoretically, such an assault could be so existentially threatening to the Republic of Korea, given the vulnerability of Seoul because of its proximity to North Korean artillery, that the first use of nuclear forces in retaliation could be justified in imaginable circumstances, although the danger of radioactive contamination of South Korea would remain an inhibiting factor. 

Indeed, the possibility of nuclear use in retaliation for massive North Korean non-nuclear aggression adds to Pyongyang’s already-strong incentives not to initiate aggression on this scale. This scenario is also why the 2010 Nuclear Posture Review did not declare that the sole purpose of U.S. nuclear forces is to deter nuclear attacks. In preparing the review, the Obama administration conducted extensive dialogue with allies. South Korean officials fretted that if the United States declared the sole purpose of its nuclear weapons was to deter nuclear use by others, the South Korean population would feel that overall deterrence of North Korea was being weakened. “Seoul” trumped “sole” in the formulation of U.S. policy. 

However, North Korea’s actual capacity to conduct and sustain a devastating conventional military attack on Seoul in the face of U.S. and South Korean counterattacks is dubious, as Admiral Blair and others have noted. To see this point from another angle, if Pyongyang did not possess nuclear weapons, it is practically inconceivable that the United States would break the nuclear taboo and initiate use of nuclear weapons to defeat a purely conventional North Korean assault on Seoul.

In any case, the nuclear deterrents of the United States, China, or Russia will not stop Pyongyang from proliferating missile technology and perhaps even fissile material or the equipment to produce it; selling illegal drugs and counterfeit currency; conducting low-level military operations against South Korean ships or outposts; or horrifically mistreating millions of North Korean citizens. These behaviors are invidious and offensive, but they do not rise to the scale that makes first use of nuclear weapons credible to deter or defeat them.

In South Asia, Pakistan’s military and intelligence services have facilitated, or at least not acted decisively to prevent, low-intensity attacks on India by jihadi groups operating from Pakistani soil. Pakistani actors used these tactics from 1947 through the 2008 Mumbai attacks. More recent attacks by the Haqqani network and the Taliban on Indian and American interests in Afghanistan have also been attributed to Pakistani facilitation. Pakistani military leaders may be emboldened to believe that they can fight unconventionally against India and then rely on nuclear deterrence to keep India from retaliating conventionally.

While the United States has much at stake in the region, the U.S. nuclear deterrent cannot make Pakistan desist from using subconventional aggression or compel the Pakistani army and Inter-Services Intelligence to act with genuine determination to curtail the recruitment and operation of jihadi groups in Pakistan. Nor can the U.S. nuclear arsenal compel Islamabad to do all the things Washington wants it to do to secure Pakistan’s nuclear materials and weapons or stop building up its nuclear arsenal. 

The India-Pakistan competition is a different matter, but it is not relevant to the discussion of U.S. first use. India rightly insists that Pakistani authorities should act decisively to curtail the recruitment, training, and operations of actors in Pakistan that seek to commit aggression against India. India has struggled to develop conventional forces and doctrines that could retaliate against future terrorist attacks and thereby compel Pakistan to adopt a genuine and comprehensive antiterror policy. Pakistan argues—genuinely or not—that it does not control the anti-Indian militants and therefore that it would treat Indian military action against Pakistan as aggression and, if necessary, would initiate use of nuclear weapons to defeat Indian forces. This deterrence-compellence problem operates on a subconventional-conventional nuclear spectrum that is unprecedented and exceedingly difficult to manage. The problem is vital for peace and security in South Asia and for the prospect of eliminating all nuclear weapons but not for determining U.S. nuclear deterrence policy.

Iran poses challenges somewhat similar to Pakistan, though more manageable. Regardless of whether Iran’s leadership decides to acquire nuclear weapons, informed experts, including in Israel, do not believe Tehran would initiate use of nuclear weapons to achieve some offensive goal, including the destruction of Israel. Nor does Iran have conventional military capabilities sufficient to invade and occupy any of its neighbors.

Rather, the realistic concern is that a nuclear-armed Iran could become more emboldened to use organizations allied with it, such as Hezbollah, Hamas, or Islamic Jihad, to intensify violence against Israel or neighboring states. Gulf states fear that Iran would become more willing to push the Shia majority in Bahrain and minority in Saudi Arabia to subvert those two governments (though recent research challenges the assumption that the acquisition of nuclear weapons will make states more likely to undertake “blackmail” and that such nuclear blackmail would be effective).18  Were Iran and its followers to create acute security crises in neighboring states or Israel, the concern is that the United States would be deterred by even a tiny Iranian nuclear arsenal from projecting military power to help its favored governments defeat insurgencies.

It is difficult to see the relevance of the U.S. nuclear arsenal and first use in deterring Iran from pursuing even the most likely of these scenarios. The United States and many others, including some Iranians, want Iran to change its behavior—or, even better, to transition from its current system of government to a more genuinely democratic one. Yet, it is far from clear how U.S. nuclear weapons are necessary or useful to this end. In any case, for the foreseeable future the United States will retain a nuclear arsenal for other reasons, and this arsenal is more than sufficient to respond to Iran if necessary. 

For its part, Israel has distinct considerations in retaining and potentially using its nuclear arsenal. This arsenal is as sufficient as any in deterring an existential threat from Iran. Thus, while Israel’s nuclear policies are central to regional security and to the challenge of creating the conditions for eliminating all nuclear weapons, affinity for Israel does not determine the necessity or wisdom of the U.S. nuclear deterrent and first use regarding Iran. 

Hitler’s Ghost

Faced with the current paucity of realistic scenarios in which the United States (and the West more broadly) would find it necessary and justifiable to initiate the use of nuclear weapons, some analysts respond that the future is unpredictable and monstrous new threats could arise. As the late, great English defense official and nuclear strategist Sir Michael Quinlan put it, “Even if all nuclear weapons had been scrapped, there could never be total assurance that a new Hitler would go down to defeat without building some and using them.” 

The ghost of Hitler—or a supposed modern avatar such as Mahmoud Ahmadinejad—is also invoked by other nuclear-armed states that enjoy conventional military superiority over their competitors. Officials and pundits in the United States, France, NATO, and Israel often point to Hitler’s ghost as a reason to oppose nuclear disarmament. When the destructive potential of a country’s nuclear weapons looks disproportionate to the threats it realistically faces—and that country does not want to say that the sole purpose of its nuclear weapons is to deter the use of nuclear weapons by others, because this would mean that if others were to eliminate nuclear weapons, it should too—that country needs a monster over the horizon. Hitler is the specter that wins the debate. 

However, Hitler’s Germany was defeated (without nuclear war).19  And nuclear weapons were not necessary to defeat Japan in August 1945, whatever the merits of the justification of their use. The Soviet Union, too, collapsed on the contradictions of its own overmilitarization and totalitarian predations. It has become evident that conventional military balances and coalitions of nonaggressive states must be maintained to deter and defeat any actor that would seek to mobilize aggressive power on a Hitlerian scale. NATO remains and expands partly to deter any such future threat and would be expected to do so whether nuclear weapons were eliminated or not. Asia-Pacific states and India are tightening political-security bonds with the United States as China’s composite power grows.

None of the countries that the United States seeks to deter from committing aggression against its allies and itself represents the combination of economic and military power, territorial hunger, fanatical ideology, and political cohesion that Nazi Germany did. The scenarios of potential conflict emanate from the near borders of Russia, Iran and the Persian Gulf, the Chinese littoral, and the Korean Peninsula. There are good and legitimate reasons why the United States projects its military power to help defend its allies along these tense peripheries of states thousands of miles from the American homeland, but it also is natural that this projection unnerves Russia, China, Iran, and North Korea. Their governments do not share Washington’s trust in its own righteous defensive motives. Unlike Hitler’s Germany, these governments know they are relatively weaker than the combined power of the United States and its allies and friends. 

Washington may provide a valuable neighborhood-watch function in these places, but here the question of regime change becomes vital. If the United States is seen to be projecting its military power thousands of miles from home against the vital interests of one of these nuclear-armed governments, these states may fight back. If the ensuing fight occurs without the UN having authorized American military action and this action is seen as a campaign of regime change rather than a proportionate response to aggression, the justifiability of the United States then resorting to first use of nuclear weapons would be extremely problematic. This would be true even if U.S. leaders understood first use to be necessary to accomplish the regime-change mission. 

The recent salience and complications of U.S. policies toward regime change pose an underaddressed challenge to U.S. nuclear policy, especially first use. If efforts to militarily overthrow the government of another state (without UN authorization20) provoke defensive actions against U.S. forces or allies that could in turn create what Washington perceives as a necessity to initiate nuclear use, the United States could be held responsible for breaking the nuclear taboo, which would have lasting political-strategic ramifications. Moreover, threatening to initiate military action to destroy a regime undermines the central premise of deterrence: if a state desists from conducting aggression, it will be spared. Focusing on first use sharpens the issues in ways that have not been appreciated in debates on post–Cold War U.S. nuclear policy. 

Domestic Imperatives 

The conditions under which the United States would realistically and justifiably initiate the use of nuclear weapons have nearly vanished. This is good news. But it has not been celebrated in Moscow, Beijing, Pyongyang, or Tehran. Those capitals fear that the U.S. advantages in non-nuclear offensive and defensive military capabilities will weaken their capacity to deter the United States and will embolden Washington to coerce them or perhaps overthrow their regimes. Reducing the shadow of U.S. nuclear use may disinhibit the United States from being more assertive, in this view. Curiously, though, the relative advantage of the United States in an environment lacking threats that would necessitate nuclear first use has not been celebrated in Washington either. To understand why, it is necessary to look within. 

A relatively small, specialized community of experts and officials shapes U.S. nuclear policy. These individuals work, for example, in Department of Defense and National Nuclear Security Administration offices responsible for nuclear doctrine, forces, and infrastructure; in nuclear-weapons laboratories and production facilities and their corporate contractors; in consulting firms and governmental and private think tanks; and on congressional committees overseeing the armed services and nuclear establishment. These specialists gather annually at major conferences in Omaha and Washington organized, respectively, by Strategic Command and the Los Alamos, Lawrence-Livermore, and Sandia National Laboratories. 

Members of this community often portray threats and advocate particular deterrence policies and forces to meet them in ways that reflect political-economic interests at the individual, bureaucratic, and party levels.21  Such interests do not mean that protagonists in policy debates are intellectually dishonest or ill intentioned. Far from it. The issues are complicated. Reasonable people can and do disagree on analysis and policy prescriptions. The distribution of perspectives and inclinations in this community resembles the shape of a wave rather than a bell curve, rising and leaning toward the “right,” but it is not a wall. 

Still, incentives often favor erring on the side of overstating threats and the consequent arsenal and policies “needed” to deter or defeat them. These domestic considerations usually are better informed and more potent than the participants’ understandings of how foreign adversaries will respond to the policies and postures that result from the Washington competition.22  Many participants in these debates know surprisingly little about the history, psychology, and political dynamics of China, North Korea, or Iran. However, they are acutely aware of how they as individuals will be treated in their professional and political communities if they favor positions that display less robustness than their competitors. Their stances are not primarily informed by calculations of what will deter Russia, China, North Korea, or Iran. Rather, the central motivator is a sensitivity to what will deter their domestic rivals from attacking them as too weak to hold office or the drive to defeat such attacks if they cannot be deterred. 

Robert Jervis, who has written seminal studies on perception and deterrence and the challenges of intelligence analysis, recently reflected on what he has learned in his lifetime of study: 

Internally generated impulses can . . . drown out a concern for what others are saying and doing. . . . The desire of democratic leaders to gain and retain power can guide foreign policy; concentrated and well-organized interests can trump or constitute the national interest; struggle and compromises within the bureaucracy can shape the information and options displayed to leaders. . . . The external world is glimpsed only dimly and in distorted form, and states may be reacting more to themselves than to others. Although deterrence theory and the security dilemma interpret arms competition differently, they both see states reacting to what others are building and doing. . . . But this may be more of a rationalization—sometimes without leaders being aware that this is the case—and the driving forces may be lodged within the state’s own political economy.23 

Jervis concludes, “It is surely a truism that the US government spends at least as much time negotiating with itself as it does negotiating with other countries—and the domestic struggles seem even more bitter.”

Rituals of display are part of this process, as anyone who has spent much time in nuclear policy discussions and debates in Washington is aware, if only subconsciously. Legendary elders like Herman Kahn proudly reveled in thinking the unthinkable about nuclear war and became iconic role models for some. The aura of brilliant bravado of the early nuclear strategists was captured well by Fred Kaplan in The Wizards of Armageddon. The most avid champions of the U.S. nuclear enterprise still display a peculiar courage in urging resolve to unleash nuclear war as the best way to deter adversaries from making it necessary to do so. 

Of course, deviations occur, too, especially in a community where intelligence and creativity are important. Occasionally, leaders of the tribe have a change of heart or mind and question the nuclear orthodoxy they once displayed. Usually such changes are expressed after leaders retire from official service. This prompts criticism that they are turncoats or have “gone soft.” However, such derision obscures the more telling implication of such changes, which is that the pressures to conform to orthodoxy stifle expression of alternate views when one is operating within the system. 

There are personal risks in breaking ranks. In 1996, Lee Butler, shortly after retiring as the four-star air force general at the head of U.S. Strategic Command, famously called for phasing out the U.S. nuclear arsenal. He was immediately vilified within the defense establishment. In 2012, James Cartwright, the vice chairman of the Joint Chiefs of Staff and a former commander of Strategic Command, co-authored a report recommending that the United States reduce its nuclear arsenal to 900 total weapons (including all categories) over the next ten years. His name now evokes sneers when mentioned in nuclear establishment circles, and his views on nuclear policy were invoked relentlessly in the campaign to block Senator Chuck Hagel’s nomination to the post of secretary of defense because Hagel was associated with him. (This is not to say there was no basis for challenging Hagel’s nomination, only that the nuclear policy issues were not seriously argued.)

Less illustrious participants in the nuclear policy arena know that the surest way to weaken their standings and reputations is to question whether threats are exaggerated, numbers of weapons are unnecessarily high, or new warhead capabilities are needed. Moderate Republicans fear that if they do not display great vigilance against threats and eagerness for more capabilities, they will be ostracized by more bellicose members of their party. In deliberations on ratifying the Comprehensive-Test-Ban Treaty in 1999, for example, Senator Richard Lugar was informed by Senator Jon Kyl that Kyl had secretly secured agreement from 34 Republican senators to oppose the treaty and that Lugar would risk his standing in the party if he did not go along. Similarly, liberal Democrats fear more hawkish Democrats plus all but a few Republicans. Individuals who aspire to executive branch positions requiring Senate confirmation learn to keep their record clear of statements and actions that could be portrayed as too “soft” to be entrusted with safeguarding national defense. 

The debate over Iran illustrates these dynamics. A recent expert speaker at a large conference on deterrence organized by the U.S. Strategic Command insisted that U.S. reluctance to enforce “redlines” in Iran’s nuclear activities has encouraged Iranian leaders to resist actions that would build confidence that Iran will not acquire nuclear weapons: 

The Iranians look back and see that Reagan pulled our Marines out of Beirut and that the United States left Vietnam without a victory, that we didn’t act very decisively in the straits in 1987 and 1988 when we ran into Iranian mines and had skirmishes at sea with them. They have crossed some red lines in supplying the insurgents in Iraq and sending in some of their soft forces into that area and also they seem to have crossed the red lines in the pursuit of their nuclear program. . . . Your deterrence credibility may be weak if you haven’t acted very decisively in the past when they crossed red lines.24 

Such inward-looking arguments do little to further sound U.S. strategy. Iranians who have dealt with Supreme Leader Ayatollah Ali Khamenei say that he recites a long litany of U.S. failures to live up to earlier proffered agreements on cooperation. All Iranians vividly recall the U.S. role in the 1953 overthrow of Mohammad Mossadegh’s government and Washington’s support for Mohammad Reza Shah Pahlavi’s repressive rule. The United States, in Khamenei’s view, failed to live up to President George H. W. Bush’s promise to reset relations when Iran helped gain the release of hostages from Lebanon. During the Iran-Iraq war of the 1980s, the United States sided with Saddam Hussein even as he used missiles to attack Iran with chemical weapons. Washington refused to apologize in 1988 when the USS Vincennes shot down an Iranian commercial airliner, killing 290 civilians. After Iran quietly cooperated with the United States in December 2001 at the beginning of the war in Afghanistan, Khamenei was shocked to hear President George W. Bush include Iran in the “Axis of Evil” in his January 2002 State of the Union address. 

Thus, the process of defining and articulating threats tends to be driven more by internal considerations in Washington than by an interest in understanding and addressing the perceptions and interests of those who are to be deterred.

The recent NATO military campaign in Libya that resulted in the killing of Muammar Qaddafi reinforced previous lessons and led Iranian hardliners to conclude that this is what happens when one gives up a nuclear-weapons program. This view, of course, does not recognize the reality that Libya had no significant nuclear capability that would have altered NATO’s decisionmaking and that Qaddafi himself triggered the intervention by violating all precepts of civilized governance. Nevertheless, Iranians are not alone in thinking that countries that acquire nuclear weapons (such as North Korea and Pakistan) do not become victims of U.S.-led regime change, while countries that do not have these weapons do become targets.

Khamenei, according to former spokesman of the Iranian negotiating team Seyed Hossein Mousavian, frequently chided his diplomats: “you are fools if you think the Americans will live up to anything they promise. They want to overthrow the Islamic Republic and will not stop until they achieve this. If we make a concession on the nuclear issue, they will simply demand more in other areas, and it will never end until they eliminate us.”25  Khamenei knows the destructive power of the U.S. military. He saw how quickly it defeated Saddam Hussein. The more difficult problem for the West is not to display the many ways it could destroy Iran but rather to convince Iran’s exceptionally isolated leader that the United States will respect the sovereignty of the Islamic Republic, which constitutionally vests enormous power in the supreme leader.

More recently, for much of the spring and summer of 2012, the U.S. presidential campaign featured debates between Republicans and Democrats and between Obama and Israeli Prime Minister Benjamin Netanyahu over redlines and specifically over what level of Iranian nuclear capability would prompt the United States or Israel to attack Iran. These debates almost never addressed whether and how publicly declaring a particular redline would affect Iran’s strategy and behavior. Rather, the willingness to proclaim redlines, and where one drew them, were expressions of the relative strength, courage, manliness, resolve, and therefore leadership suitability of Barack Obama, Mitt Romney, Bibi Netanyahu, and Ehud Barak. The political gamesmanship involved here became almost laughably obvious when Governor Romney shifted his redlines forward and backward as he transitioned from the Republican convention through the debates with President Obama. 

If threats necessitating nuclear responses are exaggerated and blustery responses to them are generally counterproductive, why do otherwise intelligent people repeat this behavior? One answer, again, is that participants in the policy debates are more preoccupied with defending themselves and defeating their domestic competitors than they are with figuring out what’s really going on in Russia, China, Iran, North Korea, and other countries and how to influence it. Fear and well-armed wariness trump dispassionate analysis of the perceptions and interests of opponents.

These patterns are reinforced in struggles within Congress to decide whether and how to fund nuclear-weapons programs. To justify spending scores of billions of dollars to modernize the nuclear-weapons complex and develop new bombers, submarines, and ballistic missiles when resources are scarce, it is necessary to portray pressing and grave threats along with the corresponding need for new or additional capabilities to meet them. Proponents of more robust nuclear capabilities often argue they are necessary to hedge against a future threat environment that could be more menacing, mindful of the long lead times necessary to generate new nuclear capabilities.26  All of this heightens the projection of nuclear threats and the display of American resolve and capabilities to defeat them.

Russian, Pakistani, and North Korean policymakers do the same thing, of course. They often do not appear to assess or care how their expressions of nuclear resoluteness actually affect the adversaries they seek to deter. (The United Kingdom, France, China, India, and Israel have been more circumspect.) President Vladimir Putin and the Russian military leadership appear not to consider how their bellicose declarations and weapons-development programs affect the United States and NATO. Otherwise, they would notice that NATO became less interested in withdrawing U.S. nuclear weapons from Europe from 2009 to 2012 as Russian strategic bombers sortied close enough to Norwegian air space to prompt interceptors to scramble, and Russian military leaders warned of nuclear strikes against missile defense sites in Eastern Europe if NATO went ahead with planned deployments. 

Moderate voices in Moscow’s nuclear policy circles report how vulnerable they feel in questioning the bellicosity that has been commonplace in the past few years. Given the one-party domination of the Russian state, leaders like President Putin need not fear that rivals in Moscow could harm them for moderating their chest-thumping rhetoric and displays of nuclear power. Rather, the motivation of these powerful figures seems to be convincing Russian society that its state is still mighty, still a world power, still capable of inspiring fear abroad, still capable of standing up to the West. Thus, Russian television in October 2012 broadcast pictures from what was reported as the largest nuclear missile command exercise in history, conducted “under Putin’s personal control,” according to Putin’s spokesperson. 

The internal dynamics described here are natural. However, human beings also have the capacity to recognize if and when their displays of strength are unnecessary or counterproductive and to adopt different behavior. Some states that possess nuclear weapons do not engage in the sort of displays that Americans and Russians (and Pakistanis and North Koreans) do. Israel does not even admit to possessing nuclear weapons, let alone boast and display this capability. Yet, no one questions whether Israel’s deterrent is as effective as those of others are. Indian leaders, for the most part, have been similarly reticent. Chinese leaders have for decades been content to retain a much smaller nuclear arsenal than the United States and Russia and to insist with some evidence that they would not use nuclear weapons first in a conflict. They have been comparatively reticent and modest in flexing their nuclear muscles, although this may be changing. Within weeks of becoming general secretary of the Communist Party of China, Xi Jinping promoted the general in charge of the army’s Second Artillery Corps, which operates the nation’s land-based ballistic and cruise missile forces. Days later, Xi met with the corps’ top leadership, heralding their force as “a strategic pillar of our great power status.”27  

Rather than searching creatively for threats that might somehow justify the first use of nuclear weapons, American policymakers and analysts would better serve the nation’s and the world’s interests by identifying ways to accomplish defense objectives without having to resort to first use. The same holds for nuclear officials and experts in Russia, France, and Pakistan. Indeed, emphasizing a retaliation-only role for nuclear weapons could increase public willingness to fund the infrastructure necessary to maintain an effective, albeit smaller, arsenal. President Obama demonstrated this political logic in the process of ratifying the New START Treaty in 2010. To win Republican support, Obama delivered Democratic support for increasing the budget to refurbish the nuclear-weapons complex, which the Bush administration and Congress had neglected. By more clearly manifesting determination to avoid the use of nuclear weapons, American officials would build confidence that the infrastructure necessary to maintain these weapons is an essential last resort. 

A Revised Declaratory Policy 

Where does all this lead? The conditions under which the United States would conceivably need to initiate use of nuclear weapons are blessedly few and improbable. The evolving laws and norms of armed conflict make first use of nuclear weapons increasingly difficult to justify. Washington should take these developments into account in defining and conducting its nuclear policy in ways that would be acceptable for others to follow. 

All states have the right to self-defense. In exercising this right, states have recognized the interest in creating and upholding norms and rules to limit the destructiveness of warfare. The laws of armed conflict, humanitarian law, and the doctrine of just war have evolved along with technology to create today’s environment, in which the United States takes care to limit even the use of discriminating drone weapons. The principles of necessity, discrimination, and proportionality are central to this framework. 

At the same time, any society confronting massive aggression that could destroy much of its population and the social-physical infrastructure on which its security and way of life depend will do whatever it can to deter or defeat such aggression. No state can be expected to deny itself the right and means to defend its people against such threats.28  Any peacetime declaration to the contrary would not reasonably be expected to obtain in the midst of an actual threat to the survival of a state. 

This probable reality was recognized reluctantly by the International Court of Justice in its 1996 advisory opinion on the “Legality of the Threat or Use of Nuclear Weapons.” In a nonbinding judgment, the majority concluded that “the threat or use of nuclear weapons would generally be contrary to the rules of international law applicable in armed conflict, and in particular the principles and rules of humanitarian law.” However, the court could not “conclude definitively whether the threat or use of nuclear weapons would be lawful or unlawful in an extreme circumstance of self-defence, in which the very survival of a State would be at stake” (emphasis added).29  Nuclear-weapon states almost certainly would not bind themselves to follow this or any other international body’s judgment that barred the potential use of nuclear weapons. But rule of law is a central pillar, if not the central pillar, of modern civilization and of American and other states’ political philosophies. As power is distributed more broadly in the international political order of the twenty-first century, the United States will have a keen interest in the predictability that comes from widespread adherence to rule- or law-based “regimes” to manage international affairs. The value in relating U.S. nuclear policy to international law is, among other reasons, to hold other nuclear-armed states to similar standards.

The most realistic politically and legally defensible formulation that the United States (and any other state) could declare to guide its potential use of nuclear weapons is that it possesses nuclear weapons only to respond to, and thereby deter or defeat, threats to its survival or that of its allies, particularly any use of nuclear weapons.

The formulation recommended here would be more restrictive than the declaratory policy of the 2010 Nuclear Posture Review but not as restrictive as the leading “sole purpose” alternative. The review states: 

  • The United States will continue to strengthen conventional capabilities and reduce the role of nuclear weapons in deterring non-nuclear attacks, with the objective of making deterrence of nuclear attack on the United States or our allies and partners the sole purpose of U.S. nuclear weapons.
  • The United States would only consider the use of nuclear weapons in extreme circumstances to defend the vital interests of the United States or its allies and partners.
  • The United States will not use or threaten to use nuclear weapons against non-nuclear weapons states that are party to the NPT and in compliance with their nuclear non-proliferation obligations.

There are two major differences between the Nuclear Posture Review and the declaratory policy recommended here. First, the existing policy places the nuclear threshold at “extreme circumstances,” whereas the alternative favored here pegs nuclear use to “threats to the survival of the United States or its allies.” Second, the policy recommended here posits use of U.S. nuclear weapons only in response to existential threats, whereas the Nuclear Posture Review preserves preemptive first-use options against adversaries’ nuclear forces or to defeat extreme non-nuclear threats. Though it does not explicitly mention preemptive first use, perhaps to avoid moral or political debate on the issue, preemption remains central. The policy recommended here, by contrast, would eschew preemptive, disarming nuclear first strikes against Russian and Chinese nuclear forces. 

“Extreme Circumstances” vs. “Threats to Survival”

Whether “extreme circumstances” or “existential threats” are the threshold, both formulas are imprecise. Ambiguity in such matters is unavoidable and prudent. The key issues when it comes to strategy are the degree to which a particular formulation leans toward or away from first use, the implications for deterrence stability, and the consequences of war if deterrence fails. 

What is an existential threat? The United States lost 1 percent of its population in the Revolutionary War. Few people looking back on that liberating conflict would say that if the United States had possessed nuclear weapons, it should have used them to defeat British forces or to attack London, even if the British could not have retaliated with nuclear weapons of their own. Yet, 1 percent of the American population in 1776 amounts to three million people today, as a share of total population. A president facing an imminent threat to the lives of three million citizens today could be forgiven for contemplating the first use of nuclear weapons in order to prevent such a loss, even though the United States would survive the death of 1 percent of its population. 

Against a non-nuclear-armed foe, the temptation would be greater. Yet, if one assumes that the stated option to use nuclear weapons against states that are not complying with their NPT obligations would only be exercised if the United States judged that such states had acquired nuclear (or, in one reading, biological) weapons, it is safe to say that Washington rules out nuclear use against non-nuclear-armed states. And initiating nuclear use against a nuclear-armed adversary with the capacity to retaliate, even in response to a truly mass-casualty non-nuclear aggression, would be to invite even greater destruction. Fortunately, there is no non-nuclear military threat to the United States today that could cause a loss of life equivalent to the experience of the Revolutionary War. 

Other historical episodes reinforce the subjectivity of defining an existential threat. In World War II, U.S. incendiary bombing campaigns destroyed 41 percent of the six most important Japanese industrial cities and 48 percent of the urban areas of an additional 57 cities before atomic bombs destroyed Hiroshima and Nagasaki.30  Yet Japan survived. Still, if the United States or any other nuclear-armed state suffered the level of destruction that Japan did, the pressure to resort to nuclear weapons to make the adversary stop would be enormous. Nor is it clear that the world would condemn a victim of such a massive assault for using the minimal number of nuclear weapons required to end such destruction if no other means were available. (The chief prosecutor of the Nuremburg War Crimes Tribunal [check], Telford Taylor, reflected, “the Nuremberg and Tokyo judgments are silent on the subject of aerial bombardment. . . . Since both sides [in World War II] had played the terrible game of urban destruction—the Allies far more successfully—there was no basis for criminal charges against Germans or Japanese, and in fact no such charges were brought.” A Japanese tribunal found the United States guilty of war crimes for the Hiroshima and Nagasaki attacks but lacked the power to make this judgment effective.31

Governments might make other claims of existential threat that rightly can be rejected. Kings, autocrats, theocrats, and dictators may believe the saying commonly attributed to Louis XIV—“L’État, c’est moi” (I am the state). They could claim that even relatively small-scale threats (in terms of physical destructiveness) against them and their ruling positions are threats to the survival of their states. However, in the twenty-first century, justifying the first use of nuclear weapons on this basis would be deemed a criminal act by the civilized world. The rule of autocratic leaders who equate themselves with their states is typically challenged by many of their own citizens. Such leaders generally retain power by violating international standards of human rights. This does not necessarily justify or recommend military intervention by others to unseat such regimes, but it does negate claims to the legitimate first use of nuclear weapons by governments to save their own autocratic positions. Thus, the United States could declare that it only possessed nuclear weapons to retaliate against threats to its survival or that of its allies. This would raise the threshold of social destructiveness so high as to avoid creating a model that unrepresentative governments could legitimately invoke to justify first use of nuclear weapons to save their regimes in a military conflict of otherwise limited scale. 

Leaders of democracies also have occasionally described threats as “existential” in ways that would be unacceptable as justification for the first use of nuclear weapons. Winston Churchill said that the survival of the United Kingdom depended on retaining imperial control of India. He clearly believed it, but he was wrong. Had the United Kingdom acquired nuclear weapons before 1947, it surely would have been a crime against humanity to use them to stop the Indian independence movement.

The benchmark of threats to the survival of a state, by contrast, connotes a massive scale of physical destruction of the population that the state is legally bound to defend. The scale implied is quantitatively and qualitatively greater than what may be suggested by “extreme circumstances”—a term that is also difficult to define a priori but the vagueness of which appears to offer more leeway for leaders to decide subjectively than does “existential threat.” 

Positing, as the United States does, that potential scenarios of nuclear use are limited to conflicts involving other nuclear-armed states, the question becomes whether it makes sense to take an action¬—starting nuclear war—that would threaten one’s own existence if a threat were extreme but not existential. Threatening to trigger nuclear war to deter less-than-existential harm may be tempting because it could work as a bluff, but actually carrying out the threat of first use would invite a graver risk. Indeed, there’s little reason to think that threatening first use makes deterrence any more effective than the more credible threat of retaliatory use does.32  Nuclear deterrence has not spared the United States, Russia, France, the United Kingdom, India, Pakistan, and Israel from numerous wars where the stakes were not great enough to make nuclear use credible. Yet, any state contemplating major aggression against a nuclear-armed state would have to assume that there is a very good chance of being hit with nuclear weapons in response. 

This insight is reflected in the declaratory nuclear policy of Russia. Russia’s 2010 Military Doctrine states that “the Russian Federation reserves the right to utilize nuclear weapons in response to the utilization of nuclear and other types of weapons of mass destruction against it and (or) its allies, and also in the event of aggression against the Russian Federation involving the use of conventional weapons when the very existence of the state is under threat.” In World War II, the Soviet Union lost 26.6 million citizens, constituting 13.5 percent of the 1939 population, but the country survived.33  Yet the scale and horror of that destruction make it almost inevitable that a Russian leadership now faced with a full-scale invasion would strongly consider the use of nuclear weapons. 

Russia’s current leaders sometimes issue blustery threats that imply a willingness to use nuclear weapons in scenarios in which Russia’s existence is not remotely threatened. But the Kremlin’s historic behavior demonstrates awareness that beneath the muscle-flexing bravado lies an understanding that nuclear war would be suicidal and therefore only credible against aggression of an existential scale, as Russian doctrine says. Similarly, Russia allows for nuclear use in response to chemical and biological attack. But it is nearly impossible to imagine that NATO, the United States, or China would use such weapons against Russia or that Russia would risk the strategic and political consequences of breaking the nuclear taboo in response to a chemical or biological attack that much weaker adversaries could someday theoretically mount. By emphasizing that the only potentially legitimate use of nuclear weapons would be in retaliation for threats to national survival, the United States would be staking a position that Russia could share. And if stray Russian officials were to rattle nuclear sabers in contravention of this policy, other states and international society would have a solid basis for stigmatizing them. 

Both China and India have no-first-use policies for weapons of mass destruction. Adversaries may question how faithfully China, India, or any other state would adhere to their declaratory policies in an actual conflict. But both China and India today possess relatively small nuclear arsenals deployed in ways more suited for retaliatory second-strike use than for first use. Both have been notably restrained in issuing nuclear threats, which affirms the sense that their threshold for use is at the level of existential threats of the kind that only other states’ nuclear weapons pose to them. Beijing and New Delhi should welcome a move in this direction by Washington.

Former French president Nicolas Sarkozy declared that his country’s “nuclear deterrence protects us from any aggression against our vital interests emanating from a state—wherever it may come from and whatever form it may take. Our vital interests, of course, include the elements that constitute our identity and our existence as a nation-sate, as well as the free exercise of our sovereignty.”34  With this policy, France does not limit its potential first use of nuclear weapons to action against other nuclear-armed states or non-nuclear threats of an existential scale. This is problematic on strategic, moral, and international political grounds. 

However, discussions with French strategists indicate greater circumspection. “Existence” does seem to be the underlying benchmark for determining if a threat would warrant nuclear use. This can be detected in their rebuttals to arguments that nuclear weapons are immoral. “No,” they say, “nuclear weapons are not a moral problem. Rather, they are a solution to the moral problem posed by the type of aggression that Hitler projected against France and the world.” Leaving aside that France survived Hitler and could not expect to fare as well in a nuclear war, the central point is to deter threats of the scale of World Wars I and II. However, failure to make this clear, and instead declaring the much lower threshold of threats from “any aggression against our vital interests,” invites the impression that France could rely on first use of nuclear weapons to deter or defeat threats that do not rise to the proportion warranting it.

Pakistan’s nuclear policy falls most short of the “existence” threshold. Pakistani military leaders say their state would use nuclear weapons if Indian conventional forces crossed the international border and overran Pakistani forces. The pending development and deployment of battlefield nuclear weapons and short-range missiles confirms an apparent willingness to climb the escalation ladder to nuclear war to deter or defeat threats that in and of themselves could be disproportionately small. 

Pakistani military authorities and strategists frequently invoke NATO’s Cold War nuclear doctrines and forces as a legitimating model for their approach. But in doing so, they neglect many important realities. India’s conventional forces and (ill-defined) doctrine do not pose a viable threat of the scale and quality that the Warsaw Pact posed. India’s restraint after two major terrorist attacks attributed to Pakistan in 2001 and 2008 reflects not only nuclear deterrence but also India’s deep interest in concentrating on internal development and avoiding war. Democratic India evinces no ideological or material interest in incorporating Pakistani territory or “converting” the Pakistani polity, as the Soviet Union once did in Europe. Moreover, NATO eventually recognized the folly of battlefield nuclear weapons, and the United States and the Soviet Union undertook multiple forms of nuclear arms control and removed intermediate-range ballistic missiles. 

Nevertheless, Pakistan will resist adoption of the declaratory policy recommended in this essay for the United States and others. If international advocacy of such a policy discomfits Pakistani leaders, this will create opportunities for dialogue to identify what measures could be taken in South Asia to enable Pakistan to move closer to an international norm.

First Use vs. Retaliatory Use 

The second major difference between current U.S. policy and the existential threshold recommended here is the question of first use. To understand the difference, some distinctions need to be made. A policy of using nuclear weapons only to retaliate for threats to the survival of the United States or its allies would not necessitate preemptive first strikes to destroy an adversary’s nuclear forces. Indeed, this policy should eschew such strikes. The existential threshold would leave open the possibility of using nuclear weapons first in retaliation for a non-nuclear aggression that threatened the survival of the United States or its allies. This is where the recommended policy differs from no first use or a policy to use nuclear weapons solely for the purpose of deterring nuclear use by others. 

The fundamental objective of this policy would be to further delegitimize and deter threats to the survival of states. First use of nuclear weapons is the clearest and most present threat to the existence of the United States and its allies, as well as to countries now in tense relations with nuclear-armed states. This is due to the inherent destruction that can be caused by first use of nuclear weapons against targets in populated areas and to the escalatory process that would likely ensue after any such use. In declaring a policy of using nuclear weapons only to deter (by retaliation) threats to survival, Washington would highlight a fundamental imperative of the nuclear age: not to act in ways that threaten the survival of any state so as to avoid the potential catastrophe of nuclear war. 

The United States and the broader international community have an interest in clarifying the liabilities of first use, even while not renouncing it in all circumstances. U.S. nuclear policy in preceding decades has not done this sufficiently. Instead, with varying degrees of explicitness or implicitness, the United States has relied more heavily than is necessary or wise on preemptive first use of nuclear weapons. 

In a potential war with an adversary that possesses the capability to retaliate for a nuclear first strike, there are trade-offs in weighing whether to attempt to disarm the opponent or to limit possible first use to attacks on the military and infrastructure assets involved in the aggression that threatens the survival of the United States or its allies. In weighing these trade-offs, the baseline questions are: First, are the anticipated military threats so great that preemptive nuclear strikes against the adversary’s deployed nuclear forces seem necessary to defeat and therefore deter the potential aggression because conventional defenses (plus missile defenses) are unlikely to do the job? Second, if conventional forces alone cannot defeat the enemy aggression, will the first use of a very limited number of nuclear weapons against the aggressing forces be sufficient to cause the opponent to seek to terminate the conflict before it escalates to nuclear exchanges of the scale that disarming a counterforce first strike would be expected to prompt?

If nuclear first use, particularly of the disarming counterforce variety, does not appear necessary or credible against realistic threats, the next question is whether the perceived residual benefits of positing such use outweigh the negative effects of doing so. The main benefit of first-use threats is the omnibus deterrent effect they might have on the minds of potential adversaries (even if nuclear first use would not be necessary to defeat that adversary or would not be credible for other reasons). But there is practically no threat to the United States or its allies that would necessitate first use or against which first use would not invite risks greater than the initial threat. (Biological and cyberweapons today do not change this calculus, though more debate should be expected on this from nuclear-weapons establishments.) This opens the way to consider the potential risks of maintaining a preemptive first-use policy for other reasons, even if it is not necessary to do so. 

The first risk of preemption is exacerbating instability in crises. As Michael S. Gerson, writing from the Center for Naval Analysis, noted, “Given U.S. quantitative and qualitative advantages in nuclear forces, and given that current and potential nuclear-armed adversaries are likely to have nuclear arsenals with varying degrees of size and survivability, in a future crisis an adversary may fear that the United States could attempt a disarming first strike.”35  In a crisis, an adversary fearing a U.S. first strike could take steps to increase the survivability of its nuclear forces to preserve retaliatory options. Such steps could include dispersing forces, raising alert levels, mating warheads to delivery systems, or predelegating authority to use nuclear weapons to field commanders.36  

The United States could detect such moves and interpret them as preparations by the adversary to “go first,” increasing the incentive for Washington to preemptively pull the nuclear trigger even more quickly. If the United States did so, the adversary would perceive an existential threat, prompting it to release whatever nuclear forces it hopes would survive a U.S. first strike. Clearly, within this dynamic there are many risks of unintended or mistaken nuclear use and escalation. Again, for the United States, the question is whether the nature of the threats it and its allies face today—and the combined relative advantages it and its allies have in conventional capabilities—make it prudent to run such risks of nuclear instability.37 

Next is the risk of escalation. Once nuclear weapons are unleashed, a cascade of confusion, emotions of fear and revenge, uncertain command and control, and pressures to “use them or lose them” rushes forward. The chaos raises grave doubts that calibrated intrawar escalation could be maintained. Hundreds of studies and debates have been devoted to the challenge of limiting nuclear war. None allow more than hope that it could be done. This (often vague) hope is sustained by the possibility that the doubtfulness of limiting nuclear war will motivate actors not to initiate armed conflicts that could lead down this horrifying path. 

While this possibility invigorates faith in nuclear deterrence of general conflict, it fails to provide counsel for circumstances when deterrence fails and armed conflict begins. Then, the questions are whether to initiate use of nuclear weapons with little valid reason to think the ensuing nuclear war would be limited and whether certain forms of first use would be more or less likely to cause escalation. The United States would be more confident in managing the risks of escalation against states with much smaller and less resilient nuclear arsenals, insofar as Washington could destroy much more of the adversary’s nuclear capability than the adversary could destroy on the U.S. side. Yet, beyond Russia and China, the two most relevant potential candidates for conflict with the United States are North Korea and Iran, and Washington could defeat both without resort to nuclear first use. (Of course, while the United States and its allies would be confident of defeating either country without resorting to nuclear first use, situations theoretically could arise in which first use would be advocated to hasten the outcome and minimize the damage that North Korean or Iranian forces could do to U.S. allies in the meantime.)

But the consequences of even a few of the adversary’s nuclear weapons surviving a U.S. first strike and detonating on U.S. or allied cities should be sobering. Moreover, retaining or enhancing U.S. preemptive first-use options may drive adversaries such as Russia and China to augment their own capabilities to withstand and counter U.S. options. This in turn could undermine stability and increase the destructiveness of nuclear war should it occur. 

Whereas instability and escalation are well-studied challenges, much less attention has been paid to the legality and morality of first use. Ethical-legal considerations are sometimes derided in defense establishments as “soft.” However, they have important implications for U.S. power and leadership (and that of other states). The laws of armed conflict, just war, and humanitarianism can be stretched to encompass the use of nuclear weapons only for retaliatory purposes and against military targets away from civilian population centers. The U.S. military claims to subscribe to these laws and norms. The United States and all other parties to the NPT affirmed at the 2010 Review Conference “the need for all states at all times to comply with applicable international law, including international humanitarian law.” 

Whether and how nuclear weapons realistically could be targeted in compliance with these laws and norms remains highly problematic, especially with high-yield weapons that are central to Cold War–legacy arsenals. It is tempting for political leaders and military commanders to soothe themselves and their publics by expressing their “intention” only to destroy legitimate military targets when in fact their targeting plans, if carried out, would exceed the bounds of law, ethics, and morality by using weapons whose destructiveness exceeds that which is necessary to destroy given targets located near large civilian populations. U.S. nuclear-war planners in the Cold War and today emphasize that they do not target civilians. Yet scenarios for conducting large strikes against Russian and Chinese nuclear forces and command and control facilities entail detonating perhaps scores of weapons in major cities.38  Performing such preemptive first strikes would be extremely difficult to justify in terms of just war and humanitarian law, as would be less extensive first use of nuclear weapons in response to anything short of existential threats. Minimizing this hazard is one reason why use of nuclear weapons should be limited to only non-preemptive retaliation for existential threats. 

The benefits of confining the use of nuclear weapons to retaliation against forces committing aggression against the United States or its allies outweigh the risks. The main counterargument is that removing threats of first use—either preemptive or in response to less-than-existential threats—may embolden potential adversaries to become more aggressive and perhaps more ready to use their own nuclear weapons first, believing that they could gain a decisive advantage that would realistically mitigate the counter-risk of U.S. nuclear retaliation. 

However, this assertion overlooks the reality that retaliatory use is a stronger and more credible deterrent than first use. Indeed, China demonstrates this. Experts debate whether China would use nuclear weapons first, regardless of its pledges not to, but no one doubts that China would use its comparatively small survivable nuclear arsenal it if it were attacked by nuclear weapons first. 

U.S. administrations of both parties and various ideological inclinations have acknowledged a condition of mutual deterrence with China. This has instilled noticeable caution in the way Washington manages relations with Beijing (and with Taiwan). There is no reason to think that China or other potential adversaries would eschew caution in dealing with the United States and its allies if Washington had the doctrine recommended here. In fact, the U.S.-Chinese strategic relationship would be more stable in a crisis if China were confident that the United States did not have the capability or intent to conduct disarming first strikes against China’s retaliatory nuclear forces and command and control centers.

Nonetheless, some U.S. allies could become unsettled by a move to a retaliatory doctrine and posture. Some officials and experts—by no means all—in Japan, South Korea, and among the Eastern European members of NATO fear that an American no-first-use or sole-purpose policy would weaken deterrence against North Korea, China, or Russia. The Obama administration consulted with these allies extensively in developing the 2010 Nuclear Posture Review. These consultations and parallel unofficial interactions among American experts and counterparts from these countries need to continue. There is still an underappreciation of the risks and costs of first use, especially in its preemptive form. It is easy in the abstract to bemoan the potential loss of what seems to be an extremely robust deterrent threat. However, the downsides of first use are rarely injected into the discussion, in part because the United States currently maintains a first-use policy and would seem to undermine it if it explicitly reminded allies of the risks. 

But intellectual and political-strategic honesty requires less prejudiced debate. Opponents of declaring a no-first-use policy assume that enemies do believe the United States would use nuclear weapons first—that this is a credible, effective threat. Yet, there is some evidence that potential adversaries are dubious. Chinese General Xiong Guangkai famously is believed to have said in the midst of the 1995–1996 Taiwan Strait crisis between the United States and China that the United States would not risk Los Angeles for Taipei. Charles DeGaulle felt similarly that France needed its own nuclear arsenal because the United States would not trade New York or Washington for the defense of Paris. 

Moreover, there is an underaddressed risk of reliance on first use, especially preemptive first strikes. Russia or China might be tempted to conduct probing conventional attacks on U.S. allies to create crises in which some American or allied official would raise the prospect of nuclear first use. The prober (Russia or China) would expect American or allied publics to express grave fears of nuclear war, which in turn could weaken collective resolve. (Russian or Chinese leaders would be less constrained by public opinion and could calmly proclaim that they had no intention of provoking war and that the United States was aggressively overreacting.) A more restrained nuclear doctrine would reduce adversaries’ incentives to conduct such provocations meant to stir nuclear crises and self-deterrence of the United States and its allies. 

The Survival Threshold and First Use 

In any case, the “survival threshold” recommended here does not limit the potential use of U.S. nuclear weapons only to retaliation against another state’s nuclear attack. Rather, it allows nuclear retaliation for a non-nuclear aggression that threatens the survival of the United States or its allies. Those allies that favor a lower threshold for nuclear use—for example, first use against any form of aggression against them—need to be reminded that if the adversary is also nuclear armed, it would most likely respond to U.S. nuclear first use with retaliatory nuclear strikes on the allied state. Running this risk would be imprudent for stakes lower than survival. 

In other words, in frank and full debate, a policy limiting the use of nuclear weapons to retaliation for threats to the survival of allies should be more persuasive than either the current U.S. policy or the sole-purpose alternative. At the very least, this web of issues should be disentangled through sustained official and unofficial dialogue and debate with Russians, Chinese, Iranians, Koreans, and U.S. allies.

The domestic dynamics of the U.S. nuclear establishment pose another major obstacle to the proposed policy. Adopting this policy is not a security risk but a career one for a president and others who want to rise in the national security establishment. Proponents of potential new nuclear-weapon systems that putatively would be ideal for preemptive first strikes will mobilize resistance, including from contractors and laboratories that generally fear any move to reduce the role of nuclear weapons. Proposing a policy change that would be portrayed as weakening the power and resolve projected by the United States invites ostracism. The problems of instability, escalation, legality, and morality that would derive from initiating nuclear use are not as important as the perceived resoluteness of threatening first use. However, again, the policy recommended here is not no-first-use or sole purpose. It is consonant with what the American public assumes U.S. policy is: to retaliate as a last resort against an aggressor that threatens the survival of the United States or its allies. 

There is another potential benefit of the proposed policy that could attenuate the political costs of pursuing it. Some champions of ballistic missile defenses, echoing Ronald Reagan, see them as a means to transition from a deterrence paradigm of mutually assured destruction to one in which offensive (first-use) nuclear doctrines could be superseded by defensive postures. This is not feasible when competitors possess large numbers of nuclear weapons and preemptive first-use doctrines and postures, as the United States and Russia do. In this circumstance, defenses are seen to augment the temptation to conduct nuclear first strikes and then rely on defenses to blunt the effects of whatever retaliatory strikes the adversary may be able to muster (after suffering the first strike). However, if competing states shifted their doctrines and force postures from first strike to retaliation only, it would become more plausible for missile defenses to be accepted as stabilizing. This could be attractive to champions of ballistic missile defenses. 

Sole Purpose

The most powerful resistance to declaring that the United States possesses and potentially would use nuclear weapons only to respond to, and thereby deter or defeat, threats to its survival or that of its allies will come from people who think this formulation is too constraining. Yet, some will argue that it is not constraining enough. They would rather declare that the sole purpose of U.S. nuclear forces is to deter nuclear attack on the United States or its allies and partners. 

This sole-purpose formulation excludes using nuclear weapons to retaliate against non-nuclear threats, apparently with no exceptions. Thus, if a non-nuclear threat did emerge to the existence of the United States or its allies, and conventional military action was insufficient to deter or defeat it, Washington should nonetheless eschew using nuclear weapons to try to end such a war. 

The merit of sole purpose hinges on whether the United States and its allies do not face existential military threats aside from nuclear weapons. No such non-nuclear threats to the U.S. homeland exist today, but U.S. allies bordering Russia and China (and perhaps someday a nuclear-armed Iran) believe that these more powerful neighbors could possibly threaten their existence by non-nuclear means. 

The governments and populations of Estonia, Lithuania, Japan, South Korea, and Taiwan may not be correct. Their fears may exceed the reality of the non-nuclear harm that Russia and China are willing and able to inflict on them. Indeed, there is very little high-quality analysis and debate on these questions. As in many other states, discussions of security threats tend to project internal dynamics more than rigorous assessments of potential adversaries’ capabilities and interests. Still, U.S. policymakers are responsible for assuring the safety of these allies. Failure to do so would expose U.S. officials to political charges of weak leadership. It is strategically reasonable and politically expedient to heed the perceptions of allies in formulating U.S. nuclear policy. This is why the 2010 Nuclear Posture Review stated that conditions were not yet ripe for declaring that “the ‘sole purpose’ of U.S. nuclear weapons is to deter nuclear attack on the United States and our allies and partners.” 

The declaratory policy and underlying principles recommended here are in the U.S. interest. Some of the other nuclear-armed states may not see them in their interests—Russia, Pakistan, and France come most readily to mind. These and perhaps other states may perceive efforts to stigmatize first use as a plot by conventionally stronger competitors. Yet, like the United States, these discomfited states will still retain nuclear arsenals. It is difficult to see why the rest of the world should not make it harder for them to justify their attachment to potential first use in situations other than when national survival is at stake. 

If the principles discussed here are closer in accord with justice and international law, they will be more politically compelling to the neighbors of these nuclear-armed states and to the broader international community. This in turn can create strategic pressures to move toward adopting these principles and add urgency to diplomatic resolution of underlying conflicts, thereby reducing the overall role and threat of the first use of nuclear weapons. Moreover, the credibility of a deterrent based on retaliation and the discriminatory application of force is greater than that which comes from positing preemptive first use with its attendant implications for crisis instability and escalation. 

Targeting and the Arsenal

It is the job of military specialists to turn declaratory policy into targeting requirements, force posture requirements, and operational plans, following guidance from the president. This work is done in strict secrecy. An outsider can only suggest general parameters that should follow if the president were to revise U.S. declaratory policy in the way recommended here. 

If the declared policy were to deter, by retaliation, threats to the survival of the United States or its allies, U.S. conventional, nuclear, or other military forces and those of its allies should be able to accomplish the following objectives: (1) defeat, through direct destruction of military forces and demonstration of escalatory risks, any existence-threatening incursion into the territory of the United States or that of an ally, and (2) eliminate the adversary’s capacity and will to continue the war.

The force used to accomplish these two objectives should:

  1. Minimize risks of the use of nuclear weapons against the United States and its allies
  2. Spare noncombatants in the aggressing state to the fullest extent possible
  3. Establish a precedent that the United States would find most acceptable for other states to follow in similar circumstances.

U.S. policymakers and armed forces naturally seek to accomplish these objectives without using nuclear weapons. Indeed, under current policy, nuclear weapons would be considered for first use only against Russia, China, and North Korea (with Pakistan and Iran on the equivalent of a “watch list” for the future). This stems from the Nuclear Posture Review’s declaration that the United States today possesses nuclear weapons only to deter or defeat nuclear-armed states or those not complying with nonproliferation obligations. The United Kingdom, France, India, and Israel fit the nuclear criterion for being subjected to U.S. nuclear threats but are not envisioned to be potential military adversaries. 

Russia, China, and North Korea lack the capability to invade or occupy the United States. Therefore, U.S. forces seek to deter them from aggression against U.S. allies, possible incursion into territory claimed by allies, and related attacks on U.S. armed forces as they are deployed to defend U.S. allies or friends. Less likely would be Russian, Chinese, or North Korean strikes on the U.S. homeland, unrelated to a regional war and with the purpose of coercing the United States to some end. (This posits that North Korea is deterrable, based on evidence that its leaders possess nuclear weapons to deter threats to their regime’s survival and also to extort payoffs that help preserve their regime.) 

Accordingly, the primary aim of U.S. military action would be to reverse attacks on and incursion into allied territory and to end hostilities in ways that would leave the aggressor significantly worse off than it was before the war. This would accomplish the immediate military objective and also augment deterrence of future aggression. First use of nuclear weapons by the United States should not be necessary to accomplish these aims. All other means should be used before resorting to nuclear first use. 

But, if non-nuclear capabilities proved insufficient to end an aggression that threatened the survival of the U.S. ally (or allies), leaders presumably would consider first use, albeit in a limited manner that minimized the probabilities of nuclear retaliation against the U.S. homeland. And if the adversary initiated nuclear use against a U.S. ally or U.S. forces defending an ally, then the United States would be free to respond in kind. In both potential scenarios, Washington should be guided in using nuclear weapons by the subsidiary objectives of minimizing risks of nuclear escalation, sparing noncombatants, and establishing acceptable precedents. Targeting would concentrate on armies, naval forces and ports, air forces, internal security forces, and leaders responsible for authorizing the initial aggression.

In a war grave enough to threaten the survival of a U.S. ally—or of the United States itself—Americans would want to remove the aggressing government and replace it with a more cooperative one. The impulse for justice and revenge would be natural and strong. If such an aggressor also initiated use of nuclear weapons, the demand for regime change would be even greater. Nevertheless, practical realities would have to be taken into account in guiding U.S. war aims. It should be remembered that massive bombing campaigns were patently insufficient to cause the aggressors of World War II to surrender. Territorial invasion by ground forces was absolutely necessary in Germany; in Japan, the threat of imminent Soviet invasion was vital. Similarly, boots on the ground were necessary to overthrow the regimes of Saddam Hussein and, more recently, Qaddafi, both of whom were orders of magnitude more vulnerable than the leaders of Russia or China. 

Unlike Japan in World War II, Russia and China possess significant nuclear capabilities to deter military campaigns and occupations aiming to remove their governments. It is difficult to imagine U.S. and allied forces invading and occupying these countries on the scale necessary to forcibly topple their regimes. Russia and China are enormous, distant countries. The size and organizational capabilities of the Chinese population and armed forces make a land war with it inconceivable. The prospect of removing the Russian government by force would also be overwhelming. 

The objective of coercing regime change in these two gigantic states would be complicated still further if the United States had used large numbers of nuclear weapons to eliminate their capacities to resist a regime-changing occupation. The physical environment would be extremely hazardous and inhibiting, as would be the attitude of the devastated population toward American forces. (North Korea presents a different scenario. Its scale and relatively weak capacity to withstand an American or allied invasion to remove its government would make regime change a likely objective if Pyongyang should precipitate a major war.)

Nevertheless, for purposes of deterrence, the United States and its allies should retain plans and capabilities to escalate attacks to weaken the military, security services, and leadership of aggressing governments to such a degree that decisionmakers would choose not to risk aggression in the first place. The vulnerability of a regime to its own citizens and outside powers at the end of an escalatory process would be intended to make the risks of war-precipitating aggression greater than the potential benefits. But threatening to punish potential aggressors in this manner should not be confused with the more ambitious objectives of actual U.S. intervention to replace an odious regime with one more to its liking. 

These considerations—simplified and abbreviated here and inviting fuller debate—suggest that the realistic objective of U.S. operational nuclear planning would be constricted to causing an adversary to reverse its aggression and end hostilities on terms that leave Washington confident that aggression will not be repeated. Against adversaries as large and capable as Russia and China, little more could be expected than denying local victory on the territory of a U.S. ally and deterring escalation by threatening the state’s security apparatus and leadership. 

Guidelines for Implementation

To minimize escalation and the risks of nuclear weapons being 52 on U.S. and allied territory, it is advisable to move away from preemptively targeting Russia’s and China’s nuclear forces. The U.S. nuclear counterforce arsenal and related first-use doctrine are driven by scenarios with Russia largely left over from the Cold War. Each targets the other’s nuclear forces, meaning that as long as one has 1,000 deployed weapons, say, the other must have at least as many. To have the possibility of destroying the other side’s nuclear weapons, each must be prepared to launch its weapons first and deliver them to their targets before they can “escape.” If either country perceives the other to be planning first strikes, both must increase their preparedness to beat their adversary to the punch. This anachronistic, self-perpetuating interaction between the two countries’ nuclear forces and first-use doctrines is destabilizing in crises. It is exacerbated by Russia’s exaggerated perceptions of threats from U.S.-NATO conventional forces and ballistic missile defenses, which motivate Russian leaders to threaten to build new missiles and delivery platforms suited for first-use counterforce plans.39  These Russian tendencies are reinforced by the military-industrial complex’s bureaucratic interests. The United States cannot indefinitely ignore these Russian developments, reinforcing the circularity of counterforce arsenals and doctrines. 

China is an additional driver insofar as its relatively small and “relaxed” nuclear posture makes it tempting to contemplate counterforce attacks with conventional or nuclear weapons, backed someday by effective ballistic missile defenses. Unlike Russia, China is still far from having a theoretical capability to preemptively destroy significant elements of the U.S. nuclear arsenal. Nevertheless, history suggests that China will not allow the United States to gain an effective capacity to negate its nuclear forces. China has the funds and the know-how to prevent this through measures such as building a larger nuclear force and enhancing its survivability and perhaps by developing asymmetric antisatellite and cyber capabilities. Thus, the United States should pursue by example and diplomatic exertion a gradual shift in nuclear forces and targeting doctrine to induce parallel Russian behavior and to encourage China to maintain its self-restraint in building up nuclear forces and retaining a no-first-use policy. 

Rather than focusing on disarming first strikes targeting Russia’s and China’s nuclear forces and command and control centers, the deterrent recommended here would focus on non-nuclear military and infrastructure capabilities that the adversary would rely upon to conduct, sustain, and possibly escalate its aggression against a U.S. ally and U.S. theater forces. It would specifically focus on those capabilities that could not be destroyed in a timely manner by non-nuclear means. Military capabilities to be targeted would be defined broadly enough to accomplish the objective but narrowly enough in early stages to minimize nuclear escalation, civilian casualties, and damage to the United States’ standing as a morally responsible great power. Indeed, U.S. moral-political standing and therefore power would be served by demonstrating restraint throughout any conflict, including after the nuclear threshold were crossed.

As Elbridge Colby, a Republican strategist and adherent of just war doctrine, has suggested, such “limited nuclear strikes could . . . focus on targets linked directly to the nature of the aggression while minimizing collateral damage—for instance striking at bases clearly and directly associated with the opponent’s initial attack on one’s vital interests. Other criteria could include: selecting targets away from national and strategic command and control and warning facilities, population centers, and strategic force bases or supporting facilities; selecting targets within a clearly defined theater of conflict or some other recognizably limited physical space; using lower-yield weapons.”40  Because the imagined military operations would be on or near the territory of U.S. allies, the imperative of minimizing damage would be strong. 

Some will be tempted to assert mistakenly that the shift recommended here from (moral) nuclear counterforce targeting must mean a move to (immoral) “countervalue” targeting of large civilian populations. This dichotomy is false in two ways. First, counterforce targeting with more than 1,000 nuclear weapons would cause millions of civilian casualties even if this is not the specific intention. Second, eschewing disarming first strikes against Russian or Chinese nuclear forces and concentrating instead on targeting conventional military forces and other assets related to the military aggression that triggered a U.S. nuclear response need not involve more nuclear weapons being detonated in populated areas than would be the case under current counterforce plans. In any case, rather than invoking the dubious outdated dichotomy between “counterforce” and “countervalue” targeting, detailed comparative analysis should be conducted on the effects of disarming first-strike plans and those that would have different objectives. 

Of course, the United States already prioritizes finding non-nuclear ways to destroy or disable targets, all the more so when targets are located in population centers. To suppress air defenses and destroy urban military targets, highly precise cruise missiles with conventional weapons, aircraft-delivered non-nuclear bombs, and electronic warfare presumably would be used. Nonexpert discussions of nuclear policy in Washington often overlook how post–Cold War advances in technology, particularly accuracy of delivery systems and effectiveness of intelligence gathering to pinpoint targets, enable non-nuclear weapons to destroy targets that before would have been assigned to nuclear weapons. 

But employing more discriminating technology can be more difficult than it seems. Much of the current U.S. nuclear arsenal (and the Russian and probably Chinese arsenals) consists of weapons whose yields are higher than necessary or morally or legally defensible. As the accuracy of delivery systems improves, the destructive yield of warheads that would be necessary to eliminate targets has been or could be dramatically reduced. But were the United States, or any other state, to propose building new, lower-yield weapons, international outcry would be severe and could undermine nonproliferation policies. Critics would argue that nuclear-weapon states were restarting arms races (in violation of the NPT) and that the reduction of yields would increase the usability of these weapons, thereby increasing the risk of nuclear war. Advocates of lower-yield weapons would respond that increased credibility of use would enhance deterrence and thereby reduce risk of use. 

Here the distinction between first use and retaliatory use would be more important than is commonly appreciated today. If lower-yield weapons were embedded in an overall policy of retaliatory use only, and the force were downsized, postured and operated in ways consistent with such a declaration, the deployment of lower-yield weapons would not necessarily be perceived to lower the threshold for their use, especially if the reduction in yield were achieved by modifying existing weapons rather than building new ones with additional military capabilities. However, this web of political, strategic, and technical issues would be extraordinarily difficult to disentangle in the real world.

These complications notwithstanding, even in retaliation for nuclear attacks, the United States would have a wider range of options to defeat the adversary than Cold War–era notions of massive retaliation suggest. Moreover, if the United States were retaliating to the adversary’s aggression and found that some targets could only be destroyed by nuclear weapons, the imperative would be to use weapons of the lowest yield necessary to destroy the target. And, since the number of such targets would be much lower than under current conditions because the arsenal would not be sized and aimed to preemptively destroy Russia’s similarly oversized arsenal, the overall threat to civilian life and the environment would be significantly reduced.

Elbridge Colby, again, reflects the logic being argued here: 

Limiting the purposes for and conditions under which nuclear weapons may be used . . . could contribute to a morally tolerable form of nuclear deterrence. In terms of ius in bello, certain constraints on the use of nuclear weapons might also be imposed. Warnings, for instance, could be strongly encouraged to enable the civil population to avoid harm. Tolerance might be given in extreme circumstances to limited, essentially demonstrative, employment of nuclear weapons against targets isolated from substantial noncombatant populations. If such efforts are unavailing, targeting that focuses on those in power (and so responsible for the extreme measures being taken) and what they value, or on essential military facilities, as opposed to the general population, might also be tolerable in some particularly grave situations. In all cases the purpose of the strikes would be to prevent some grave evil, to deter further aggression or escalation, and to bring the war to a tolerable conclusion as rapidly as possible.41 

To be sure, as Colby recognizes, “even if such restraints were imposed, nuclear deterrence would, in all likelihood, still rely on the threat of cataclysmic destruction. Indeed, in all honesty, the possibility that conflict might escalate to such a level lies near the root of the effectiveness of nuclear deterrence.”

The imperative to raise the threshold for use of nuclear weapons by the United States, and to narrow the destructiveness of potential use, reflects and reinforces the U.S. interest in setting a precedent for nuclear use that would be relatively tolerable if other states were to act similarly. U.S. leaders would be conducting defensive war with the aim of minimizing damage to the United States, its allies, and even the society of the adversary so that international life would resume as propitiously as possible, although inherent risks of escalation cannot be precluded. Washington and its allies would want to revive and probably enhance the civilizing effects of international law, the laws of war, and humanitarian law. Presumably, the United States would also seek to preserve the postwar legitimacy of its possession of nuclear deterrent forces and those of, at least, the United Kingdom, France, Israel, and perhaps India. These vital postwar objectives all would be served by conducting with unimpeachable restraint the imagined war in which the United States had been compelled to use nuclear weapons. Working back from the imagined postwar environment to develop the U.S. deterrent and war-fighting guidelines today is a clear national interest. 

Numbers of Weapons

The size of nuclear arsenals is not as important as how they are used. Logically, the size of the U.S. arsenal should be determined by targeting requirements and need not be equal to that of Russia or any other adversary if preemptive targeting of adversary nuclear forces is eschewed. However, political realities dictate that the size of the operationally deployed U.S. strategic arsenal will remain on par or larger than any competitor’s. “Second to none” in overall effectiveness is a political imperative. Thus, there is little point in suggesting a particular number of weapons that would seem from a generalist’s perspective sufficient to deter Russia or any lesser nuclear-armed state from committing aggression that would threaten the survival of the United States or its allies. The number of weapons Washington retains will be determined by the arms control process with Russia, which in turn is affected by the psychology and technology of ballistic missile defenses, the implications of new conventional-strike capabilities, Russia’s relationships with the NATO states on its borders, and so on. 

However, the declaratory policy and subsidiary targeting concept and implementation guidelines proposed here would require significantly fewer nuclear weapons than Washington will deploy when the New START Treaty is implemented fully. In principle—and recognizing that in practice the United States would only reduce in parallel with Russia—a force closer in numbers and doctrinal governance to that of China today should be sufficient. This is not to minimize the real challenges that would have to be met to actually move to such a force. China would have to eschew major expansion of its nuclear arsenal and introduction of new forces that would indicate an operational departure from its long-standing declared policy of no first use. South Korea, Japan, and other U.S. friends and allies in Asia would have to be persuaded that the postulated changes in U.S. policy and forces would not increase the probabilities of Chinese (and North Korean) aggression against them. Meeting these conditions would be difficult and take considerable time and deft leadership at the highest levels of the U.S. government as well as a greater degree of bipartisan cooperation in Washington than can be imagined today. 

Numerous issues and interests now impede efforts to reduce nuclear arsenals further, to negotiate an end to further production of fissile materials for nuclear weapons, and to bring the Comprehensive-Nuclear-Test-Ban Treaty into force. In this situation, greater effort should be made to raise the doctrinal, political, and moral barriers to the first use of nuclear weapons, whatever numbers of these weapons exist. 

Resisting the Temptation to Abandon Principles

Among the many reasons why the nuclear policy recommended here would augment U.S. and international security is the interest Americans share with their allies and even the populations of adversary states to minimize the harm of warfare, especially to innocent civilians, and to promote justice and international law. Nuclear policies that do the most possible to achieve these ends will also be more credible, thereby strengthening deterrence. 

Matching military practice with moral-political values and principles is challenging. It is all too easy for governments to traduce ethical barriers to destructive escalation. World War II is instructive in this regard, when governments lapsed in the application of principles they professed to hold essential to their own identity and purpose. This historical record underlines why limitations should be sought on the use of weapons of mass destruction and how much conscious effort is necessary to maintain such limitations at moments when they would be most tested. 

Lessons From World War II

Before World War II erupted with Germany’s invasion of Poland in September 1939, Italy had airdropped mustard gas and other munitions on civilian targets in Ethiopia in 1936. Japan had bombed civilians in China throughout its 1937 invasion. A photograph of a Chinese infant sitting alone screaming amid bombing rubble around a railway line became a memorable feature of newsreels seen by 136 million people around the world, according to Life magazine.42  

Referring to such “ruthless bombing from the air of unfortified centers of population,” U.S. President Franklin D. Roosevelt on September 1, 1939, issued an appeal immediately at the onset of the new war urging all antagonists to commit to “under no circumstances undertake the bombardment from the air of civilian populations or unfortified cities, upon the understanding that these same rules of warfare will be scrupulously observed by all their opponents.”43 Initially, Britain, France, and Germany responded by vowing to bomb only military targets and eschew targeting civilians and cultural property. 

It is difficult to judge precisely who first broke this vow and began the bombing of civilian populations. Germany in early 1940 bombed legitimate military and military infrastructure targets in English cities. The Luftwaffe conducted these strikes in daylight with evident effort to concentrate damage on legitimate targets. Still, civilians were killed and injured.

The United Kingdom was inhibited initially in its efforts to bomb targets in German towns. In a June 1940 directive, the chief of Bomber Command was ordered to concentrate attacks on German oil resources, with the injunction, “In no circumstances should night bombing be allowed to degenerate into mere indiscriminate action, which is contrary to the policy of His Majesty’s Government.”44 

But countervailing desires for revenge and for visiting destruction on the German population were naturally expressed. Winston Churchill, taking the prime ministership on May 10, 1940, wanted to hit the German heartland. On the second night of Churchill’s premiership, “We began to bomb objectives on the German mainland before the Germans began to bomb objectives on the British mainland,” a British air-power strategist recalled.45  

Days later, Churchill worried that bombing German cities would alienate Roosevelt and the American people, whose entry into the war England desperately sought. Churchill asked his minister of information to “arrange that discreet reference should be made in the press to the killing of civilians in France and the Low Countries, in the course of German air attacks.” The prime minister specifically asked that the word “retaliation” not be used to describe the British bombing of German cities then under way because it would not withstand scrutiny.46 

Tactical considerations also compelled moves to bomb German towns and cities. Bombers found it exceedingly difficult to hit specific targets with any accuracy, especially at night. But only at night could British bombers proceed to target areas with manageable rates of attrition from German fighters and air defenses. During the first year of bombing southwest Germany, for example, 50 percent of the bombs landed in open country. Twenty percent hit residential areas, and only 5.2 percent landed on military objectives.47  The general inaccuracy of night bombing, paired with the location of military targets in sparsely populated areas, meant that most bombs failed to accomplish objectives.48

Inaccuracy also provided an opportunity. If crews aimed at military targets in cities, bombs that predictably missed their preferred targets would likely hit populations. This, it was believed, would serve the strategic objective of weakening enemy morale. The benefit of turning misses into hits when attacking cities was noted in British military debates over whether to concentrate bombing on German shipbuilding and aircraft-manufacturing facilities, oil production and storage installations, railways, or “morale,” which was the euphemism for populated areas. Rather than targeting military facilities and related infrastructure, in the words of a 1940 order to the deputy chief of the Air Staff, “It is desired that regular concentrated attacks should be made on objectives in large towns and centres of industry, with the primary aim of causing very heavy material destruction which will demonstrate to the enemy the power and severity of air bombardment and the hardship and dislocation that will result from it.”49  

In late August 1940, Germany dropped its first bombs on London. The next night, August 25, British bombers attacked Berlin. Churchill “suggested that Bomber Command should henceforth spread its bombs as widely as possible over the cities of Germany.” According to the official history, “This idea did not yet, however, appeal to the Air Staff.”

As towns and cities became a clearer focus of British bombing in late 1940, official rhetoric still conveyed the principle of targeting military objectives and not populations per se. This was useful for several reasons. It maintained moral propriety for the nation as a whole and for the armed forces. And it exerted pressure on Germany to be restrained in its bombing practices, as was seen in a speech Hitler gave in Munich on November 8. Hitler declared he had allowed German bombing attacks only during the day, to avoid indiscriminate damage. “Then it suddenly occurred to Mr. Churchill to attack the German population at night. . . . I watched for eight days. They dropped bombs on the people of the Rhine. They dropped bombs on the people of Westphalia. I watched for another fourteen days. I thought that the man was crazy. He was waging a war that could only destroy England. I waited over three months, but then I gave the order. I will take up the battle.”50 

Churchill’s claims of propriety and discrimination were a fiction (as were Hitler’s more broadly, of course). But this did not stop the prime minister from repeating them. On October 17, 1940, after Britain had begun intentionally bombing population centers, Churchill was drinking port in the House of Commons smoking room when a conservative member of Parliament approached him, as Harold Nicolson recounted in his diary, and urged the prime minister to heed public demand for unrestricted bombing of Germany. Churchill replied, according to Nicolson, “My dear sir, this is a military and not a civilian war. You and others may desire to kill women and children. We desire (and have succeeded in our desire) to destroy German military objectives. I quite appreciate your point. But my motto is ‘Business before Pleasure.’”51  In fact, by that time, according to the official British history of the strategic air offensive against Germany, “the fiction that the bombers were attacking ‘military objectives’ in the towns was officially abandoned” in private, if not publicly.52  But Churchill still repeated the “business before pleasure” quip in March 1941 when asked at a private lunch why Britain was not making German citizens suffer from urban bombing as the British were. Churchill’s wife, Clementine, laughed and said to her husband, “You are bloodthirsty.”53  

As the war unfolded through 1943 to 1944, England and the United States made firebombing of densely populated targets a central element of their strategies and practices. The policymakers who authorized these attacks—and the military personnel who conducted them and the press that reported on them—still took pains to say they were militarily necessary, but private records revealed the truth. In a telling summation of the collective obfuscation involved in waging mass destruction, John Dower relates that Winston Churchill in late 1945, weeks before Germany’s capitulation, wrote to his military staff: “It seems to me that the moment has come when the question of bombing German cities simply for the sake of increasing the terror, though under other pretexts, should be reviewed. . . . I feel the need for more precise concentration upon military objectives, such as oil and communications behind the immediate battlezone, rather than on mere acts of terror and wanton destruction, however impressive.”54  Revealingly, Churchill withdrew this memorandum four days later at the request of the Air Staff. A new, sanitized memorandum was substituted that eliminated references to wanton terror.55 

The lineage between firebombings and the use of nuclear weapons is well-known, as is the role of U.S. General Curtis LeMay in both. In World War II, LeMay directed the XX Bomber Command in China and then the XXI Bomber Command in the Pacific, from which he led the initiation of massive incendiary bombing campaigns against 64 Japanese cities. Official documents and historical narratives demonstrate that LeMay was uninhibited by humanitarian restraint in the conduct of his job. Indeed, as Robert McNamara, who had been an aide to LeMay in World War II, recounted in the Academy Award–winning documentary Fog of War, “LeMay said, ‘If we’d lost the war, we’d all have been prosecuted as war criminals.’” 

However, as Dower again documents, it would be wrong for anyone to dismiss LeMay as a bloodthirsty aberration. On May 30, 1945, the New York Times’s front-page headline read: “Marines Crash Into Shuri, Win All North Part of Naha; Tokyo Erased, Says LeMay.” Two articles ran below the headline, and the one about Tokyo being erased ran second. Its own headline read: “51 Square Miles Burned Out in Six B-29 Attacks on Tokyo—LeMay Backs Figures With Photos of Havoc—1,000,000 Japanese Are Believed to Have Perished in Fires.” This was not a typographical error. The Times was reporting in a third-tier subhead that U.S. attacks had killed one million Japanese in a single day and night. The paper did not report that anyone expressed misgivings about it. The report was erroneous (the attack killed approximately 100,000 civilians), but this does not explain the nonchalance of its delivery and public reception.

The Taboo Matters—Strengthen It

The world has come a long way since 1945. Indeed, the distance can be marked partly by the abhorrence readers of the New York Times would feel today at news of airstrikes killing one million people. However, it is easy to slip and slide away from principles and best intentions. England and the United States were the “good guys” in World War II, the forces of civilization. The short time between Roosevelt’s 1939 proclamation and the terror bombings of German and Japanese cities is one reminder of how easy it is to fall, as is the pretense of leaders that only military objectives were being targeted. 

The adjustability of principles may be a relief to those who fear that self-imposed constraints on the use of nuclear weapons will make us weak and vulnerable and will embolden unscrupulous enemies. But this is literally a demoralizing attitude. It undermines the identity on which civilization and the rule of law broadly depend. The United States promotes civilized norms around the world. Indeed, the taboo on the first use of nuclear weapons is the most portentous of these norms, for it serves to inhibit the act that would likely escalate to the greatest burst of self-destructiveness in human history. 

The nuclear taboo is reified in the widespread view that a nuclear weapon is not just another weapon and is not merely defined by its physical destructiveness. As former secretary of state and chairman of the Joint Chiefs of Staff Colin Powell said in the aftermath of the 2002 Indian-Pakistani military standoff, which many feared would lead to nuclear war, the nuclear bomb “is not just another weapon in a toolbox of weapons. It crosses a line that the world does not want to see crossed in 2002. And the condemnation that would go against whichever country did it would be worldwide and it would be immediate.”56  

Norms may seem soft, but they matter. The Nobel-laureate strategist Thomas Schelling has said the evolution of the nuclear taboo is “as important as the development of nuclear arsenals.”57  The United States, Russia, China, the United Kingdom, France, India, Pakistan, and Israel could use nuclear weapons to destroy non-nuclear adversaries with relative ease and without fear of retaliation, but they have not. A strong political-moral injunction—a taboo—creates the widely shared perception that it would be beyond the pale to initiate use of nuclear weapons against a state that poses no similar-scale threat, even if one could do so without risk of being destroyed in return.

This does not mean that officials, strategists, and citizens of nuclear-armed states refrain from thinking and talking about using nuclear weapons. Taboos exist precisely because people are tempted to do the taboo thing, and sometimes they talk about it. If the temptation were not great, there would not be a need for inhibition as strong as a taboo. The tension between temptation and inhibition can be seen in the ways that American officials obfuscate the question of first use in U.S. nuclear policy. 

Of course, the nuclear taboo is backed by unsurpassed hard power. The nuclear taboo is against first use. It is a moral-political deterrent backed by a physical threat. The consequence of breaking it most likely would be nuclear retaliation. It would be naive to gainsay the degree to which prudential, hard-power calculations have contributed to the non-use of nuclear weapons since 1945. But rather than disprove the reality of a taboo against first use, acknowledging the inhibiting effects of the fear of retaliation reinforces the reasons why the taboo emerged in the first place. Particular acts are taboo precisely because they are so dangerous to the species. 

The taboo against nuclear first use has at least one major implication that is inadequately recognized and discussed: actors have a corollary obligation not to incite someone to break the taboo. If the nuclear-first-use taboo has been constructed and strengthened over the past sixty-seven years, its continuation will be augmented by conscious efforts to inculcate the underlying imperative for states (and nonstate actors) to forbear from threatening the survival of other nations. Such threats are inherent in the use of nuclear weapons and also in the sorts of massive aggression that could trigger legitimate nuclear use. 

Limiting the scale and destructiveness of warfare is a long project of civilization, as Steven Pinker has recorded in his massive study The Better Angels of Our Nature: Why Violence Has Declined.58  Efforts to prevent the use of nuclear weapons would be strengthened by explicitly promoting the underlying intolerability of conducting aggression on a scale that would threaten the survival of any nation. 

Aggression in any form is illegitimate, of course, as recognized by the UN Charter. But aggression often is not easy to define. Lower-scale forms of it unfortunately are not rare. There is strategic, political, and moral meaning in distinguishing among scales and types of aggression. This is what the laws of just war, humanitarian law, the genocide convention, and treaties, such as those banning chemical and biological weapons, do. The singular potential destructiveness of nuclear weapons invites efforts to specifically anathematize violence of a scale that threatens another state’s survival because the dangers of nuclear war and massive aggression are causally almost inseparable. 

Focusing international attention on the taboo against the first use of nuclear weapons and actions that could cause the taboo to be broken would be more than a rhetorical initiative. There are observable, important actions and policies that states can take to demonstrate their commitment not to threaten the survival of others, even while retaining capabilities and policies for lower-level warfare. 

Policies declaring nuclear use only in response to existential threats would provide a benchmark for other states and international experts to evaluate whether the overall military forces, doctrines, and behaviors of states are consistent with declaratory policy. 

States could clarify that they have no demand or policy to annihilate any other people or, more pertinently, to initiate conflict to forcibly remove a nation or state from disputed territory. (This is the basic logic on which the United States, China, and Taiwan have based their security relations since 1979. The injunction against using force to “resolve” territorial disputes is now salient in East and Southeast Asia.) 

Correspondingly, states could acquire and deploy conventional forces in quantity and quality that convey defensive intentions and are unsuitable for invading, occupying, and incorporating the sovereign territory controlled by others. States could conduct military exercises that reflect their intentions not to operate military forces scaled and directed to threaten the survival of their neighbors or others. 

Such policy declarations and actions would still leave open possibilities of military initiative, including a sort that some would call aggression. The point here is that in working from the objective of reinforcing the taboo against nuclear first use and other threats to survival, states could design a realistic web of policies and actions to pursue this objective. Indeed, in some ways, this is what NATO and Moscow did in the period between 1987 and the 1999 Kosovo intervention, when Russian politics turned noncooperative in response to the perceived overassertiveness of NATO. One of the missed opportunities of this period, and one of the reasons why the cooperation did not last, was that neither side renounced its preemptive first-use doctrine and force postures and encouraged the other to follow suit. The United States and Russia reduced and limited conventional and nuclear forces, increased transparency, managed the peaceful realignment of Russia’s relations with its former satellite states, and generally changed history in a previously unimaginably positive way. But they did not get out from under the threat of nuclear first use—preemptive or otherwise. The fact that the two countries’ nuclear relationship appears tragically anachronistic to many in both societies suggests that it still could be changed, perhaps with more encouragement from others. 

Debates over nuclear proliferation, deterrence, and disarmament, including in the context of NPT Review Conferences, often focus narrowly on nuclear weapons and not enough on the threats to security that animate the acquisition and potential use of these weapons. President Obama, as the single most influential figure in global nuclear affairs, could advance the purpose of his 2009 Prague speech by inviting explicit international debate on the role of nuclear weapons and the shared responsibility of all states not to cause these weapons to be used. Obama’s Nobel Prize acceptance speech, which challenged and perhaps surprised its audience by reminding them that war is sometimes necessary and can be just, is a good benchmark. He could do this in relation to whatever nuclear policy he chooses to declare, but the recommended formulation would invite international attention to the threshold-crossing effects of threatening the survival of another nation and the global interest in perpetuating the taboo against the first use of nuclear weapons. 

Reinforcing the nuclear taboo is all the more important if the further proliferation of nuclear weapons remains a threat. In this case, it is prudent to raise the moral-political costs and inhibitions of initiating use of nuclear weapons. All who seek power recognize the need to convince followers that their actions are just and moral. If the taboo against the first use of nuclear weapons becomes stronger, people contemplating actions that could cause first use would risk losing the mantle of justice and morality on which their power would at least in part depend. 

It is difficult to see how national and international security would be weakened if state leaders and international civil society focused more on stigmatizing threats to the survival of states and on reinforcing the taboo against the first use of nuclear weapons. Russian leaders may privately protest that reinforcing the taboo against nuclear first use is intended to disadvantage Russia and privilege the United States and NATO, which enjoy non-nuclear military superiority. Pakistani military leaders will think similarly regarding India. French strategists bridle at any efforts to diminish the perceived value of nuclear weapons. These concerns need to be managed in ways that do not exacerbate security dilemmas. 

No one can take away the nuclear weapons of Russia, Pakistan, France, or any other state. The nuclear taboo cannot physically prevent these states from using their nuclear arsenals to defend themselves. But strengthening this taboo can and should drive all states to eschew actions that will put them on a course toward experiencing the costs and consequences of breaking it. If major powers of the twenty-first century are to avoid the destructiveness of the twentieth century, leaders will have to concentrate actively and assiduously on removing the temptation to initiate use of nuclear weapons.


1 Ronald Reagan meant it, too, but many American and international observers did not perceive this.

2 Some proponents of nuclear disarmament don’t like the provisional, but clearly long-term, endorsement of nuclear deterrence and the corollary upkeep of the nuclear-weapon complex. Some opponents neglect the Prague caveat because it contradicts their narrative that Obama is a dangerous unilateral disarmer.

3 The Prague agenda entails reducing the role of nuclear weapons in national security strategies; reducing nuclear arsenals; ratifying and entering into force of the Comprehensive Nuclear-Test-Ban Treaty; completing a treaty to end production of fissile materials for military purposes; strengthening the Nuclear Non-Proliferation Treaty (NPT) by enhancing the authority and resources behind international inspections (the International Atomic Energy Agency) and establishing consequences for countries caught breaking the rules or seeking to leave the NPT without cause; building a new framework for civil nuclear cooperation, including an international fuel bank; and engaging Iran in diplomacy to comply with its obligations as a non-nuclear-weapon state.

4 Paul Bracken, in his much-noted new book, The Second Nuclear Age (New York: Times Books, 2012), emphatically urges the president of the United States to “solemnly declare, ‘The United States will not be the first to use nuclear weapons under any circumstances (262).’” Bracken, generally regarded as a conservative strategist, argues “‘No first use’ means just that. It says the bomb is a weapon that should never be used. Yet it does not take the utopian leap, to getting rid of it, that is, to disarmament (263).” “No country benefits more from a tradition of non-use than the United States. Given its tremendous investment in conventional forces, and given the unique history of having used the bomb twice, the United States isn’t going to go nuclear on account of a tactical expediency. I cannot imagine any North Korean or Iranian bunker that is so important that the United States would break a seven-decade taboo against nuclear use (264).” Bracken’s no-first-use position goes further than the policy recommended in this essay, as becomes clear below, but his reasoning buttresses the general thrust of arguments made here. One of the challenges that makes a strict no-first-use position problematic is the lack of confidence that major global powers, including the permanent members of the UN Security Council, would coalesce quickly and decisively enough to mobilize international pressure and military strength to deter or defeat aggression of a scale that could prompt the first use of nuclear weapons. 

5 Nina Tannenwald, The Nuclear Taboo (New York: Cambridge University Press, 2007), 10.

6 The action plan agreed upon by states at the 2010 NPT Review Conference registers “the need for all states at all times to comply with applicable international law, including international humanitarian law.”

7 For outstanding treatments of contradictions and uncertainties, see Robert Jervis, The Illogic of American Nuclear Strategy (Ithaca, N.Y.: Cornell University Press, 1984); Robert Jervis, The Meaning of the Nuclear Revolution (Ithaca, N.Y.: Cornell University Press, 1989); Michael Quinlan, Thinking About Nuclear Weapons (Oxford: Oxford University Press, 2009). 

8 John O. Brennan, “The Ethics and Efficacy of the President’s Counterterrorism Strategy,” speech at the Woodrow Wilson International Center for Scholars in Washington, D.C., April 30, 2012,

9 United States Nuclear Posture Review, April 2010,

10 The Non-Proliferation and Disarmament Initiative also includes Canada, Chile, Germany, Mexico, the Netherlands, Poland, Turkey, and the United Arab Emirates.

11 See, for example, Peter Pry, “Dark Days for America’s Strategic Deterrent,” Journal of International Security Affairs, no. 19 (Fall/Winter 2010): 57–68.

12 In the words of Alexei Arbatov and General (retired) Vladimir Dvorkin, “the strategic relationship between Russia and the United States will undergo essentially no significant change for the foreseeable future, leaving neither side with the capability to conduct a disarming first strike under any scenario of nuclear conflict.  According to models advanced by independent experts, under any conditions of a counterforce attack the defending side would still have several hundred nuclear warheads with which to retaliate, which would be sufficient to inflict unacceptable levels of damage on the hypothetical aggressor.” Alexei Arbatov and Vladimir Dvorkin, “The Great Strategic Triangle,” Carnegie Paper (Washington, D.C.: Carnegie Endowment for International Peace, April 2013).

13 To be sure, the scale of the stakes involved in Georgia could have grown. If the United States or NATO had intervened to stop the Russian intervention—or, more likely, intervened afterward to try to push Russian forces back—Russia would have faced an imperative to fight back or risk humiliation. In the process, Moscow could have worried that NATO—or, more likely, the United States—would issue nuclear threats to compel Russia to desist or (less realistically) withdraw. Weaknesses in Moscow’s intelligence and warning capabilities could have gravely lessened Russian leaders’ confidence that they would be able to detect NATO or U.S. preparations to launch nuclear weapons. These Russian leaders then could have felt it necessary to threaten their use of nuclear weapons to counter this compellence. If NATO were involved at this point, discord would have emerged within the alliance. Some members would have rejected any moves that would risk nuclear war over a situation this murky and relatively minor compared to the risk of nuclear war. But in the fog of such a crisis, would all sides, especially Russia, have seen each other’s capabilities and intentions clearly enough to manage it prudently?

14 In the early 1980s, in the midst of the debate over the morality of nuclear deterrence that was sparked by a pastoral letter of the U.S. Catholic bishops, the strategist Albert Wohlstetter wrote: “We should not and do not rely on the threat of losing control to deter either nuclear or conventional attack. But MAD [mutually assured destruction] and the fictions of uncontrollability it has propagated encourage us to rely on the threat of losing control as a substitute for dealing with the dangers of conventional conflicts. In short, they have led us to be less serious about conventional war as well.” Albert Wohlstetter, “Bishops, Statesmen, and Other Strategists on the Bombing of Innocents,” Commentary, vol. 75, no. 6 (1983): 15–35, reprinted in Nuclear Heuristics: Selected Writings of Albert and Roberta Wohlstetter, edited by Robert Zarate and Henry Sokolski (Carlisle, Pa.: U.S. Army War College, Strategic Studies Institute, 2009), 583. Wohlstetter was trying to reassert the feasibility of limited nuclear war in Europe and NATO first use. In doing so, he highlighted the related necessity to enhance conventional military defenses to mitigate self-inhibiting fears of nuclear escalation. Whereas in the Cold War, the vital importance of defending Germany and Western Europe against massive Soviet conventional forces made nuclear use relatively credible, the Estonia scenario adduced here indicates that the wisdom and credibility of relying too much on nuclear deterrence and first use has become much more problematic. Security and reassurance require greater attention to strengthening conventional military capabilities and plans or diplomacy to mitigate threats in the first place.

15 Dennis C. Blair, “Nuclear Deterrence and War Plans,” in In the Eyes of the Experts: Selected Contributions by Experts of the Congressional Commission on the Strategic Posture of the United States, edited by Taylor A. Bolz (Washington, D.C.: United States Institute of Peace, 2009), 51–52.

16 However, China’s belief that the United States would not initiate nuclear use in such a conflict is attenuated by fear that Washington could use powerful non-nuclear weapons to conduct a disarming first strike on Chinese nuclear forces and command and control centers and then rely on regional and national missile defenses to blunt China’s ragged retaliation. The United States today does not have such a preemptive conventional capacity, and it is very possible that effective ballistic missile defenses are not technologically feasible or could be readily countered by adaptations of the retaliatory Chinese arsenal. However, the United States’ refusal to consider policies to limit the potential scale and capabilities of conventional strategic weapons and missile defenses intensifies China’s concerns about strategic stability.

17 “Panetta to Urge China and Japan to Tone Down Dispute Over Islands,” Washington Post, September 17, 2012,

18 For an outstanding treatment of this issue, see Todd S. Sechser and Matthew Fuhrmann, “Crisis Bargaining and Nuclear Blackmail,” International Organization 67, no 1 (2013): 173–95,

19 If the West alone had possessed nuclear weapons, it is reasonable to think Hitler would not have attacked it. This is an argument for nuclear nonproliferation. If both Hitler and the West (and the Soviet Union) had possessed nuclear arsenals, Hitler probably would not have attacked either. This is an argument for nuclear deterrence. But, if Hitler nonetheless had attacked, the casualties of the ensuing nuclear war could have been at least as horrible as those of World War II. This is an argument for using the lessons of the nuclear age to create the conditions necessary to remove and preclude the reemergence of forces and policies that would threaten the survival of states.

20 In American politics it is common to dismiss the importance of UN authorization, but all recent administrations led by both parties have taken pains to seek such authorization, for example, regarding Iraq twice and Libya. From another direction, advocates of the Responsibility to Protect could bridle at the notion that avoiding conflicts that could escalate to nuclear use should inhibit efforts to remove exceptionally abusive governments. A tension arises here between the Responsibility to Protect and the taboo against first use of nuclear weapons. Given the probable scale of destruction in nuclear war and its implications for international security, precedence should be given to reinforcing the nuclear taboo in instances where intervening to protect civilian populations could result in the use of nuclear weapons. 

21 Disclosure: judging by the company the author often keeps with current and former officials in the U.S. nuclear establishment, the author self-identifies as part of this community. He supposes that many members of this community, especially those who work in the nuclear-weapons laboratories, the Defense Department, the National Nuclear Security Administration, and their contractors and congressional supporters, consider the author to be on the misguided-but-well-intentioned pro-disarmament margin of the community. Still, members of the “nuclear disarmament establishment” inside and outside the United States may view the author as too tolerant of states’ attachments to nuclear deterrence. All of this affirms the larger point: when you are in a community like this, you know and viscerally feel where you stand and how others regard you. You tacitly know where the lines of acceptance and expulsion are, how your words could and would be used against you, depending on the audience. We also get paid for our work. Most members of the nuclear policy establishment live on the budgets allocated for the overall nuclear weapons enterprise—directly or indirectly. A smaller number, including the author, are funded through donations from individuals, foundations, and governments that define their missions in terms of promoting international peace, preventing proliferation and nuclear war, and so on. We all think we are objective, of course, but the best we can do is to subject our analyses and perceptions to full and open debate.

22 Proponents of no-less-than-the-status-quo nuclear forces and policies would say conversely that advocates of more restrained postures and policies underestimate the nefariousness of adversaries and the necessity of robust deterrence and defense. They (“deterrent hawks” for shorthand) assume that adversaries will behave “better” the more formidable and forward U.S. forces and doctrine are. But they rarely offer detailed evidence to back up their assumption that leaders of opposing states are affected meaningfully by knowledge of the number and types of nuclear weapons the U.S. deploys and the doctrine guiding their potential use. Nor do they produce evidence that U.S. deployment of larger numbers and a first-use posture causes these leaders to be more restrained (deterred) in pursuing aggression against the United States or its allies than they would be if Washington deployed fewer nuclear weapons and a doctrine that foreswore disarming counterforce first strikes. Deterrence hawks generally do not know enough about adversaries to empirically justify these assumptions, but they do have empirical evidence (their own experience) that their position makes them more secure from domestic attack. 

23 Robert Jervis, “Psychology and Security: Enduring Questions, Different Answers,” Yale Journal of International Affairs 7, no. 2 (Summer 2012): 13.

24 Barry Schneider, remarks at STRATCOM conference, August 8–9, 2012. 

25 Conversation with former Iranian nuclear negotiator Seyed Hossein Mousavian. For a similar view, see President Mahmoud Ahmadinejad’s interview with David Ignatius, “Iran’s Mahmoud Ahmadinejad on Israeli Threats, Nuclear Program and Syria,” Washington Post, September 23, 2012,

26 In emphasizing the need for increased nuclear capabilities amid uncertainty, nuclear champions tend to underappreciate the value of pursuing arms control and other diplomatic means to ameliorate security dilemmas and often adopt rhetoric that impedes such efforts.

27 Jane Perlez, “New Chinese Leader Meets Nuclear Military Officers,” New York Times, December 5, 2012.

28 UN Charter, Article 51, “Nothing in the present Charter shall impair the inherent right of individual or collective self-defence if an armed attack occurs against a Member of the United Nations, until the Security Council has taken measures necessary to maintain international peace and security.”

29 For an outstanding analysis of the International Court of Justice’s advisory opinion, see Burns H. Weston, “Nuclear Weapons and the World Court: Ambiguity’s Consensus,” Transnational Law & Contemporary Problems, 7, no. 2 (Fall 1997): 371–400.

30 H. P. Willmott, The Great Crusade: A New Complete History of the Second World War (Dulles, Va.: Potomac Books, 2008); Neil Sheehan, A Fiery Peace in a Cold War (New York: Random House, 2009), 134.

31 Telford Taylor, Nuremberg and Vietnam: An American Tragedy (Chicago: Quadrangle Books, 1970), 140–41.

32 The main counterargument that the United States and (less so) Russia could make, thanks to their enormous nuclear arsenals and targeting capabilities, is that threatening preemptive first use to limit the potential retaliatory damage the adversary could inflict would deter that adversary more effectively than retaliation. However, as James Acton has persuasively argued in Deterrence During Disarmament (London: International Institute for Strategic Studies, 2011), it is technically implausible that the United States could preemptively destroy enough of Russia’s or China’s capacity to retaliate so that such damage-limiting preemption would be strategically reasonable. Moreover, forces and doctrines meant to conduct damage-limiting first strikes exacerbate crisis instability and drive arms racing. Acton, Deterrence During Disarmament, 44–52.

33 Michael Ellman and S. Marksudov, “Soviet Deaths in the Great Patriotic War,” Europe-Asia Studies 46 no. 4 (1994): 671. Stalin, prior to the war, presided over policies that directly or indirectly killed millions of Soviet citizens, as did Mao in China. Among other things, such murderous policies indicate the elasticity of leaders’ assessments of what levels of death constitute threats to national survival.

34 Nicolas Sarkozy, speech at presentation of “Le Terrible” submarine, Cherbourg, March 21, 2008,

35 Michael S. Gerson, “No First Use: The Next Step for U.S. Nuclear Policy,” International Security, 35, no. 2 (Fall 2010): 37.

36 Ibid.

37 The destabilizing effects of preemptive first strikes against Russian and Chinese nuclear forces would obtain if the United States developed and deployed conventional prompt strike capabilities that could perform the same mission. U.S. strategists, military leaders, and congressional overseers are aware of this, which is one reason—in addition to technological feasibility and costs—why they are approaching such conventional capabilities cautiously. 

38 American nuclear strategists, politicians, and pundits sometimes argue that reductions to “low” numbers of nuclear weapons would require a shift from targeting adversary military forces to targeting “nursery schools” and therefore would be less moral than current U.S. policy. However, proponents of this position ignore the scale of civilian deaths that would attend the execution of “counterforce” strikes with an arsenal of, say, 1,000 strategic weapons. Nor do they defend on moral grounds Russia’s effort to restore a large counterforce arsenal as preferable, say, to a much smaller Russian force. China currently deploys several scores of nuclear weapons that could hit the continental United States, an insufficient number to target U.S. nuclear forces. This probably means the Chinese forces are targeted at major American cities and ports. Yet few would argue that it would be morally better, and desirable from the perspective of saving civilian lives, for China to expand its arsenal dramatically so that it could be targeted at U.S. nuclear forces and command and control facilities. 

39 Such tendencies to exaggerate potential threats are not limited to Russia, as discussed in “Domestic Imperatives.”

40 Elbridge A. Colby, “Defining Strategic Stability: Reconciling Stability and Deterrence,” in Strategic Stability: Contending Interpretations, edited by Elbridge A. Colby and Michael S. Gerson (Carlisle, Pa.: U.S. Army War College: Strategic Studies Institute, 2013), 59. 

41 Elbridge A. Colby, “Keeping the Peace,” First Things (January 2011): 6.

42 John Dower, Cultures of War (New York: W. W. Norton, 2010), 159.

43 Ibid., 160.

44 Directive, June 4, 1940, Air Vice-Marshal W. S. Douglas (Deputy Chief of the Air Staff) to Air Marshal C. F. A. Portal, in Sir Charles Webster and Noble Frankland, The Strategic Air Offensive Against Germany 1939–1945, vol. IV (London: Her Majesty’s Stationary Office, 1961), 113.

45 James Spaight, quoted in Nicholson Baker, Human Smoke: The Beginnings of World War II, the End of Civilization (New York: Simon and Schuster, 2008), 178.

46 Baker, Human Smoke, 182.

47 Webster and Frankland, The Strategic Air Offensive Against Germany 1939–1945, vol. 1, 302.

48 Ibid., 150.

49 Ibid., 129; Directive from Air Vice-Marshal W. S. Douglas (Deputy Chief of the Air Staff) to Air Marshal Sir Richard Peirse, October 30, 1940.

50 Baker, Human Smoke, 247.

51 Harold Nicolson: The War Years 1939–1945 – Volume II of Diaries and Letters, edited by Nigel Nicolson (New York: Atheneum, 1967), 121–22.

52 Webster and Frankland, The Strategic Air Offensive Against Germany, vol. 1, 157.

53 Baker, Human Smoke, 292.

54 Dower, Cultures of War, 174, citing Charles Webster and Noble Frankland, The Strategic Air Offensive Against Germany, 1939–1945 (London: Her Majesty’s Stationery Office, 1961), vol. 3: 112.

55 Dower, Cultures of War, 175.

56 Cited in Tannenwald, The Nuclear Taboo, 361.

57 Thomas Schelling, “The Legacy of Hiroshima: A Half-Century Without Nuclear War,” Institute for Philosophy and Public Policy, report 20 no. 2/3 (Summer 2000),

58 Steven Pinker, The Better Angels of Our Nature: Why Violence Has Declined (New York: Viking, 2011).