Nuclear Weapons in a Changing Global Order

The Interplay Between the International System and the Global Nuclear Order

Back to table of contents
Steven E. Miller, Robert Legvold, and Lawrence Freedman
Meeting the Challenges of the New Nuclear Age

Lawrence Freedman

A political order is a top-down concept. It assumes that a political system can be designed in such a way so that conflicts are managed expediently and normally without violence. How it is structured may be the result of deliberate action by the major powers using treaties, conventions, and new institutions to encourage restraint and establish competent mechanisms for resolving disputes. Alternatively, there may be a favorable concatenation of circumstances that results in order. Either way the test lies in durability. An “order” shapes the choices of individual actors by means of informal or formal rules, to the point where over time patterns of behavior become embedded. Even as parties enter into a dispute, they will have expectations about how the dispute can be resolved.

Viewed from the bottom up rather than the top down, individual actors still have choices. If they decide in sufficient numbers that the system generates unfair outcomes or that disruptive behavior will yield more advantages, then the order will start to become vulnerable. At such moments, those who have the greatest stake in the system’s continuity must decide whether to enforce its rules. If that is their decision, then there can be no half measures. The stakes will be too high, and such determination can lead to major war. Alternatively, they may conclude that the tide of history is against them and not worth resisting. They may seek instead some sort of accommodation or compromise to facilitate a shift in the international hierarchy, as the United Kingdom did with the United States.

The yearning for a more orderly world becomes most intense during the course of long and painful wars. Such wars prompted deliberate action that established rules to prevent further carnage—for example, the 1648 Peace of Westphalia after the Thirty Years’ War, the 1815 Congress of Vienna after the Napoleonic Wars, the League of Nations after World War I, and the United Nations after World War II. An example of an effort to consolidate an existing order, taken without the prompt of a major war, was the package of measures associated with the East-West détente from 1969 to 1975. Here the aim was to take advantage of forms of political order that had emerged more as a result of a balance of power than by any deliberate design. The further shift in the balance of power that took place with the implosion of European communism from 1989 to 1991 created new opportunities to put in place a more satisfactory political order, based more on shared values than mutual fear.


The Cold War Order

The Cold War order is usually attributed to a convenient bipolarity and a shared fear of nuclear war. The two superpowers, the United States and the Soviet Union, each led an ideological bloc and both were armed with substantial nuclear arsenals. If the two alliances clashed, then mutual destruction was the likely result. Winston Churchill made this point in his last significant speech as UK prime minister when he observed how though a “sublime irony . . . safety will be the sturdy child of terror, and survival the twin brother of annihilation.”1 Thus a stable political order was the result of a stable nuclear order, which depended on mutual deterrence. During the years of the Cold War, a new line of strategic reasoning developed to explain how this depended on a shared vulnerability to nuclear devastation and it explored whether this might be upset by new technologies. A critical point came in the early 1960s when, after the shocks of the Berlin and Cuba crises and a cool appraisal of the consequences of vast numbers of offensive systems and the limited capabilities of defensive systems, the two superpowers began to accept that both sides were locked into a condition of mutual assured destruction (MAD). The search for a plausible first-strike strategy lingered on before it eventually subsided, but caution was now the norm in superpower relations.

To what extent did a political order depend on a nuclear order? At one level, there is no reason to doubt the reported views of political leaders on both sides of the Cold War divide— that they feared nuclear war sufficiently to moderate their geopolitical demands on each other and to exercise caution at moments of crisis. Part of this was a developing conviction on both sides that, whatever their clever strategists might tell them, there was no reliable route to a nuclear victory.2 Any victory would depend on a surprise attack that would either leave the enemy disarmed or disoriented but any initiation of nuclear war promised to be suicidal. The only circumstances in which one side might take a nuclear initiative would involve the intense emotions, panic, and uncertainty of another major war. Irrational decisions might be made in irrational situations—so best not to tempt fate by exploring how far it was possible to go in a large-scale war without triggering a nuclear conflagration. It became hard to see how and why a third world war might start.

It is possible the same effects could have been achieved with conventional weapons, as John Mueller has argued, because two earlier world wars had demonstrated that they were destructive enough and that victory was unlikely to be the result of anything other than a long and painful slog.3 This is a counterfactual assumption that cannot be proven one way or the other.

The key property of nuclear weapons was that they left no room for doubt about their effects and so reduced any temptation to try a cunning strategy that might just catch a complacent enemy unawares. More important, however, than the question of whether the stability resulting from a nuclear balance might also have resulted from a conventional balance was the question of whether either might have worked without the existence of an alliance. The Atlantic Alliance was the result of a reflection on the twin failures of deterrence in 1914 and 1939 in the face of German aggression. The view was that the problem was not so much the absence of a super-destructive weapon and more the absence of the United States. This was the failing remedied by NATO. The value of a counter to Soviet power was confirmed by the Berlin airlift, leading to the formation of NATO in 1949. It took time before the Soviet Union could organize its satellites into something comparable—the Warsaw Pact was formed in 1955 in response to West Germany joining NATO.

A nuclear-armed superpower led both alliances, which divided Europe neatly down the middle. But this apparent symmetry was between two alliances that were fundamentally different in their political composition. The question of order was dependent on whether the two sides could accept a Westphalian Peace, based on respect for sovereignty and no interference in internal affairs, or what Moscow described as “peaceful coexistence” between two opposing systems. This was itself dependent upon the internal cohesion of the two ideological blocs.

Thus, the Cold War began with one form of domestic political order, based on liberal democracy, fearing another, based on tight control by vanguard communist parties. Liberal capitalism flourished, benefiting from the security and cohesion provided by an alliance, while state socialism progressively lost its legitimacy. The West agreed not to interfere directly in the affairs of the Soviet bloc as a precondition for détente. Nonetheless, the existence of a freer and more prosperous way of life, about which people in communist countries became increasingly aware, undermined the basis of Communist Party rule. The Cold War was sufficiently established so it could continue for some time, even into the twenty-first century, without major mishap, were it not for the fact that the domestic order of one side proved to be extraordinarily fragile.

The Warsaw Pact was a coercive alliance, in that most (if not all) of its members would not have chosen to be aligned so closely (including ideologically) to the Soviet Union. When individual states began to step out of line—Hungary in 1956, Czechoslovakia in 1968, or Poland in 1981—force was used to get them back. NATO was more voluntary, and so the benefits of alliance had to be regularly restated and demonstrated. Although in practice the Warsaw Pact was more brittle, it was always NATO that appeared to suffer most from internal rows. In addition, it was assumed for most of the Cold War that the advantages in conventional war lay on the Soviet side. This introduced one of the most challenging features of the relationship between nuclear order and political order—the impact of MAD on the credibility of an alliance.

The logic of the nuclear stalemate was to neutralize the effects of the arsenals. There was no premium in initiating a nuclear war. Each arsenal cancelled out the other. There was an obvious stabilizing benefit with self-neutralization so long as the only conflict in mind was one between the two superpowers. This reflected the view that the only possible rational reason to use nuclear weapons was in retaliation to a nuclear attack on the homeland.

Actually, looked at coolly as a rational calculation, there was no inherent logic to automatic retaliation once one’s own society had been lost. Nonetheless, the safest assumption was that the impulse would be irresistible.4 This was not an assumption that was going to be sensibly tested. But would that impulse extend to allies? Would treaty obligations, such as NATO’s Article 5 on collective defense, mean that any attack against a European state would be treated as if it had been an attack on the United States, even if it could not be resisted by conventional means?

Whatever the reassurances given in public, senior American policy-makers in private doubted that it would ever make sense to risk the destruction of the homeland by initiating nuclear war on behalf of allies. President Richard Nixon is reported to have said during an early White House discussion, with his typical elegance, “Nuclear umbrella in NATO is a lot of crap. Don’t have it.”5 According to the famous “Healey theorem” (named after a British defense secretary), it was far easier to deter Moscow than reassure the allies: “it takes only five percent credibility of American retaliation to deter the Russians but ninety-five percent credibility to reassure the Europeans.”6 The alliance held. Whatever their doubts, most alliance members concluded it was safer and certainly cheaper to rely on the U.S. nuclear guarantee than to seek out alternative security arrangements. The only exception was France, which developed its own nuclear arsenal and moved into a semi-detached relationship with NATO, largely because after the Suez crisis of 1956 France did not trust the United States to back its wider interests. Successive UK governments may have had similar thoughts, but care was always taken to present the national nuclear force as a supplement to that of the United States and as a contribution to NATO.

Away from Europe, the Soviet Union also suffered its most serious alliance indiscipline with China. Although it had tolerated the early development of a Chinese nuclear capability, as a result of the 1958 Taiwan Strait Crisis the Soviet leadership concluded it did not wish to link its fate with a Chinese leadership that was careless about nuclear danger. The Chinese pressed ahead regardless with their own nuclear program while forcing a split in the Communist world that soon raised the possibility of a Sino-Soviet war. In this respect, the two blocs were quite different in their ability to manage proliferation. NATO held together despite the Gaullist challenge; the Communist bloc turned in on itself as a result of the Maoist challenge.

For the United States, the challenge of alliance management lay in the credible extension of deterrence. It was an issue not only with regard to NATO but also to the other American allies in the Asia-Pacific region—South Korea, Japan, Taiwan, and Australia. If they were not reassured then the most likely alternative was to build their own nuclear deterrents along the lines of France and China. It was partly to close off this possibility that the superpowers began to edge toward a nonproliferation treaty in the 1960s. This would oblige countries to accept publicly that they would not develop their own arsenals.

The success of the effort, in the form of the 1970 Nuclear Nonproliferation Treaty (NPT), provided the second pillar of the global nuclear order, according to William Walker.7 Limiting the number of nuclear players through the nonproliferation regime complemented the first pillar of mutual deterrence restricting the likelihood of war among the established nuclear players. Whereas a strong nonproliferation regime supported mutual deterrence by keeping strategic relationships relatively simple, it was not the case that mutual deterrence necessarily supported nonproliferation, precisely because it undermined the credibility of alliance guarantees. The key proliferation bargain was to eschew national arsenals in return for superpower security guarantees. Proliferation was more likely when that did not occur.

This provides a different twist to the standard understanding of the proliferation bargain as expressed in the NPT, which says that states should only be expected to give up their nuclear option if they still get the benefits of peaceful nuclear energy (which generally was not problematic), and as long as the nuclear weapons states agreed, under Article VI, to make their own best efforts to end the arms race and achieve general and complete disarmament (which has been extremely problematic). Article VI was designed to address the evident unfairness of an unequal treaty. This intent is stressed, not surprisingly, by proponents of disarmament. They warn that if arsenals are not cut back then proliferation will result.

Yet Article VI demands no more than best efforts and then sets a standard  that goes well beyond nuclear abolition to the effective elimination of major conventional forces as well (an idea that was still current in the 1960s). If, however, the United States really began to disarm substantially, its allies might begin to fret about whether their security guarantees were still in place. Of course, if their security problems had been dealt with along the way (which is what the grander schemes for disarmament often suppose) then this might not be a problem, taking us back to the question of whether it makes much sense to attempt to reduce armaments while underlying conflicts are still unresolved. For the relevant countries NATO’s Article 5 trumped the NPT’s Article VI.



  • 1Winston Churchill, “Never Despair,” speech in UK Parliament House of Commons, March 1, 1955, International Churchill Society, 1946-1963-elder-statesman/never-despair.
  • 2McGeorge Bundy, “To Cap the Volcano,” Foreign Affairs 48 (1) (October 1969): 1–20.
  • 3John Mueller, Retreat from Doomsday: The Obsolescence of Major War (New York: Basic Books, 1990).
  • 4Rose McDermott, Anthony C. Lopez, and Peter K. Hatemi, “‘Blunt Not the Heart, Enrage It’: The Psychology of Revenge and Deterrence,” Texas National Security Review 1 (1) (December 2017): 68–88.
  • 5“Minutes of National Security Council Meeting,” February 19, 1969, in Foreign Relations of the United States, 1969–1976, Vol. XXXIV, National Security Policy, 1969–1972, ed. M. Todd Bennett (Washington, D.C.: U.S. Government Printing Office, 2011).
  • 6Denis Healey, The Time of My Life (London: Michael Joseph, 1989), 243.
  • 7William Walker, “Nuclear Order and Disorder,” International Affairs 76 (1) (October 2000): 703–724.

Post–Cold War Disorder

The Cold War concluded with a victory for the West as a demonstration of the strength of liberal capitalism and alliance cohesion. Unlike the end of hot wars, in which the victors were careful to avoid triumphalism. Russia was now encouraged to serve as an active partner in the creation of what President George H. W. Bush called a “new world order.”8 There was a period in the early 1990s when would-be architects of this new world order designed structures that would meet all the criteria of peace, justice, and durability. The designs required that some old institutions be discarded while others found new roles. Additional institutions with their own rules might need to be created so that all conflicts could be resolved and rivalries calmed.

Ideally this new world order could manage without a new nuclear order. Instead of relying on the MAD of the Cold War, which had resulted in at best a prudent peace, the aim was now to have a more positive, liberal peace based on partnerships, economic interdependence (globalization), and the rule of law, in which nuclear weapons had no role at all. This view was reflected in schemes for nuclear abolition, which were predicated on achieving a new world order, since ways had to be found to encourage nuclear powers to disarm completely without any loss of security. As these schemes developed it became evident they depended on hitherto intractable issues—such as the eccentricity of North Korea or the tension between India and Pakistan—being sorted out on the way to a nuclear-free world. In practice, most of these schemes ended up describing stages in this process, building upon existing treaties and pushing for the reduction of arsenals and less dangerous operational procedures, which soon appeared to be challenging enough by themselves, falling well short of complete abolition.9 

One reason for the disappointing lack of progress on an even more modest arms control agenda (after an initial burst of disarmament in the 1990s) is that a new world order did not develop as hoped. The term reflected a misapprehension about the series of events that began with disaffection in Eastern Europe, marked by the breach in the Berlin Wall of November 1989 and the breakup of the Soviet Union at the end of 1991. This period was, and still is, celebrated as the “end of the Cold War.” This moment is now fixed in our mental chronologies along with the end of the World Wars as one of the great punctuation points of history. Yet the Cold War had been through many stages before this point, and its end had also been celebrated, prematurely, in the early 1970s. The relationship between the two superpowers was quite different in 1989 from what it had been in 1947, 1956, 1961, 1972, or 1983. What had really come to an end was European communism—which was finished as a political force—and a Russian empire—which had been built up over centuries and extended after 1945.

Within a short time frame, Russia lost its governing ideology, economic bearings, other former Soviet Socialist Republics, and the satellite states of the Warsaw Pact. All the new states to its west began to gravitate toward NATO and the European Union (EU), and most of them ended up in both organizations. As the new leaders in Moscow became alarmed at the assertiveness of the West and tried to push back they still often had to give way, for example over Kosovo. It was only as the West allowed itself to get overextended, in Afghanistan and Iraq, and as the Russian economy grew with higher oil and gas prices, that Moscow was able to recover some lost ground. In addition, it had clung to its position on the UN Security Council and to its nuclear arsenal, both of which were seen to be increasingly valuable assets (which is one reason for the growing Russian disinterest in plans for nuclear abolition). Russia was left bereft of its allies. Its recent international behavior can be explained in part by its anxiety about the loss of one client, Ukraine, and the potential loss of another, Syria.

Thus what took place in the 1990s was a massive shift in international power. It was not the end of Russia, but its power was greatly diminished, soon leaving it feeling insecure. It was a fresh start for the United States since it could spend less time on some long-standing concerns, but its global position was still founded on its existing network of alliances. The main change was that the United States took on more allies as a result of the surge of new applicants from the Soviet bloc. The United States also enjoyed a degree of ideological hegemony and a greater freedom of political maneuver (although both turned out to be far less than what was assumed in the 1990s). It was not so much that the United States and its allies did not accept the value of a rules-based international order, especially when it came to the economic domain, but that they felt they could rewrite the rules, for example with regard to humanitarian intervention.

One reason for this outcome may be found in the continuing repercussions of the other great feature of the post-1945 world: namely, decolonization. While the old European overseas empires were all essentially dismantled by 1975, the collapse of the Soviet Union and Yugoslavia could be seen as a continuation of the same process, resulting in far more independent states. From the fifty-one states that signed the UN Charter in 1945, membership in the United Nations grew to 193. Many new states turned out to be economically weak with immature political institutions and deep social cleavages, generating a slew of new security challenges.

This has made it even harder to configure a new world order. Interstate relations have never been more varied and complex. International institutions still work reasonably well away from the security sphere, for example in managing trade, and coped better than might have been expected with the 2008 financial crisis. But it was always the security sphere that was supposed to provide the test of a new world order, and the results continue to disappoint. This does not mean the system is doomed to violence and disarray. Most states, most of the time, work out their relationships with each other. The U.S. alliance network has a profound influence on many of the potential arenas of conflict. What we do not have is a new, top-down international order designed for contemporary purposes.



  • 8George H. W. Bush, “Address Before a Joint Session of the Congress on the Persian Gulf Crisis and the Federal Budget Deficit,” September 11, 1990, American Presidency Project, http://
  • 9See, for example, Gareth Evans and Yoriko Kawaguchi, Eliminating Nuclear Threats: A Practical Agenda for Global Policymakers (Canberra, Australia: International Commission on Nuclear Non-proliferation and Disarmament, 2009). Gareth Evans (Australia) and Yoriko Kawaguchi (Japan) were the cochairs.

The Nuclear Order and the Political Order

It is remarkable that more than a quarter century after the end of the Soviet system the network of alliances forged for Cold War reasons is still in place. I have argued that the key proliferation bargain was to eschew national arsenals in return for superpower security guarantees. Proliferation was more likely when that did not occur. Does the persistence of alliance networks therefore help explain why nuclear proliferation has been less than feared?

Certainly if stopping proliferation had depended on progress toward disarmament as envisaged in Article VI of the NPT, then there should already be many more nuclear powers. Even without much happening on this particular Article, the NPT provides a normative framework that might influence some potential proliferators. In addition, the entry costs into the nuclear business remain high, requiring advanced technological and industrial capabilities. These capabilities are hard to construct covertly. When the effort has been made it tends to get revealed before the projects are complete. It is particularly difficult now if a state has signed on to International Atomic Energy Agency (IAEA) safeguards. The agency is a sophisticated organization that has learned from past mistakes. If the IAEA is prevented from doing its work, or there has been a formal withdrawal from the NPT, that provides a damning clue about intentions. Since it takes time to produce fissionable material and to develop and then realize weapons designs, alert neighbors may start to become anxious years before the capability is in place, creating a risk of preemption. Whatever the limits of the Iran nuclear deal, the fact that a country with radical regional aspirations accepted it in the end is a testament of sorts to the pressure would-be proliferators face.10 At the same time, one of the reasons why President Trump abandoned the deal in 2018 was evidence that despite the deal, which was still working as intended, it had done nothing to impede Iranian missile development and regional militancy.

Because of the costs and risks of proliferation the benefits need to be serious. There were always three sorts of benefits. The first was to break open the international hierarchy. One objection to the NPT was that it confirmed a hierarchy in which the five designated nuclear powers were also the permanent members of the UN Security Council. Many would argue that there is a case for reform of the Security Council to include the major nonnuclear powers in order to break that link. But since reform has proven to be so elusive, the permanent membership has stayed the same. In practice, however, economic strength is usually far more relevant to most international groups.

A second benefit was prestige. To be able to master the technology demonstrated that a country was on the first rank in its science and advanced industrial capabilities. This argument is still current. It appears to have been important in the Iran case and was part of the rationale for countries such as Brazil, which put considerable effort into their nuclear programs. But nuclear energy is no longer seen to be vital to wider industrial success.

Whatever the prestige or civil energy arguments for developing an advanced capability short of an actual weapon, the strategic value lies in hedging since the breakout time is reduced if a weapon is desired. This was a major issue in discussions of the quality of the Iran deal.

Enhancing security must be the main potential benefit of a nuclear capability. Because of the costs and risks of a nuclear program, the security gain must be high. It is evident that with many security challenges, especially if they involve terrorism or insurgencies, nuclear weapons are irrelevant, and even with interstate warfare they are very much in the last resort category. Those countries that have opted for their own nuclear capability do have serious strategic rationales. Even then they still seek to soften their dependence on nuclear deterrence by building up their conventional capabilities, or looking for ways to ease tensions with their neighbors, or strengthening their ties with friendly great powers. One reason Israel refuses to brag about its nuclear options is that it does not want to jeopardize the far more important alliance with the United States, and its nuclear capability is less relevant to most of the security challenges Israel currently faces.

Despite the logic that might warn against relying on the United States to provide deterrence, given the risks to which that would expose the United States, alliance continues to trump proliferation. France has come back into the NATO fold. In the 1974 Ottawa Declaration, it allowed the force de frappe to be presented, along with the UK’s Polaris fleet, as contributions to NATO’s deterrent. No other U.S. ally has built a nuclear weapon. There have been regular reports that America’s Asian allies, and in particular Japan and South Korea, have kept the matter under review, but as long as the United States appears ready to honor its obligations in the face of Chinese and North Korean threats, the Asian allies choose to stay with the status quo. The uncertainties resulting from the policies of President Trump, including on notionally unrelated areas such as trade, have led America’s allies to wonder about the solidity of alliance ties. Allies are clinging to the fact that they are supported by official policies even though this is rarely reflected in the President’s statements. If he wins a second term, it is hard to see how the current alliances can survive in their current form, although Trump is not wholly consistent and future crises may change the context.

What then of the former Soviet allies? We have already noted that China broke with the Soviet Union completely over the nuclear issue. North Korea also decided to go it alone in the early 1960s when Moscow refused to help it develop its own arsenal. The most intriguing case was the unplanned nuclear status of Ukraine, Belarus, and Kazakhstan in 1991 when the Soviet Union came apart. Whether or not at the time the nuclear-tipped missiles that were left in bases on their territory were close to being operational, it is reasonable to assume that this could have been achieved at some point. In order to get these newly independent countries to join the NPT and relinquish any claims to nuclear status, security assurances were offered.

The 1994 Budapest Memorandum, a document that has now become a byword for ineffectuality, offered Ukraine in return for joining the NPT a reaffirmation from the treaty’s depository powers, inter alia, to respect its independence, sovereignty, and existing borders, and to refrain from the threat or use of force. None of this did Ukraine much good twenty years later when Crimea was annexed by the Russian Federation. Indeed the fact that Russia had felt able to invade non-NATO Georgia (in 2008) as well as non-NATO Ukraine, while it had held back from acting against such NATO members as Estonia and Latvia, with whom it had comparable arguments, can be taken as a demonstration of the importance of alliances as a source of security. At the same time, Russia has drawn regular attention to its nuclear strength when warning outsiders not to intervene actively on behalf of Ukraine. “Russia’s partners,” Putin was reported saying in August 2014, “should understand it’s best not to mess with us,” with a reminder that “Russia is one of the leading nuclear powers.”11

We can develop this theme further by looking at the other proliferators: India, Pakistan, and Israel. These countries do not have formal alliance arrangements. Pakistan and the United States once had an alliance arrangement, and Israel is all but an ally with the United States. Israel was able to use the threat of a nuclear program to persuade the Kennedy and Johnson administrations in the 1960s to reverse past policies of transferring conventional arms. When it became apparent that Israel had a covert nuclear program, the Nixon administration agreed not to make a fuss so long as the program stayed covert. As Or Rabinowitz has shown, this approach extended to a degree to Pakistan and to South Africa (before South Africa decided as it entered its political transition to abandon its weapons program).12

In addition to looking after its own nuclear status, Israel pursued a very active regional anti-proliferation policy. It destroyed an Iraqi nuclear reactor in 1981 and a Syrian one in 2007, and threatened to strike against Iran. After the 1991 Gulf War, the United Nations terminated Iraq’s revived nuclear program. Iran joined Libya, whose ambitions were exposed with the A. Q. Khan network, in agreeing to abandon, or at least defer, its nuclear ambitions in return for relief from sanctions. Neither the Libyan nor Iranian experiences are straightforward. The program Libya abandoned in 2003 was still in its early stages. It is unlikely it would have been completed by 2011, but if it had been then it might have led to doubts in NATO about whether it was so wise to support the rebellion against the Libyan leader, Colonel Gaddafi, that led to him being killed. In the case of Iran, critics argued from the start that easing economic sanctions meant that in return for apparently temporary restrictions on its nuclear program it would be able to push ahead with its missile program and pursue aggressive regional policies, including in Syria. While supporters pointed out both the durability of the restrictions and Iran’s compliance, the critics persuaded President Trump to abandon the deal in May 2018. At the time of writing, Iran’s intentions with regard to its nuclear program are unclear, but the episode demonstrates the difficulty of separating non-proliferation measures from broader judgments on foreign policy. If Iran does revive its program, it is also not clear whether this will lead to American military action and whether those Sunni Arab states opposed to Iran will begin to hedge with their own nuclear programs.

The Middle East is thus an area of high conflict in which alliance relationships are tentative, and the country that is closest to a superpower is still hedging with a covert nuclear arsenal. One cannot demonstrate that the Middle East would be more orderly if the key antagonists were nuclear powers. Many of the conflicts are the result of fractures within states rather than between them. This region would offer a striking test of the thesis that nuclear weapons can bring stability. At the moment it is certainly not at all stable. One can argue that the Israeli nuclear arsenal, though never officially declared, has weighed on the strategic calculations of Arab countries, and that more nuclear arsenals might encourage more caution. But this is a region where fraught situations develop that could overcome the natural caution of governments.

In Asia it is arguable that a nuclear order has contributed to a political order, in that a degree of caution has been injected into the India-Pakistan relationship, although there is an important proviso about the consequences of internal turmoil should weapons and facilities fall into the hands of extremist groups. The most curious case is that of North Korea, a country that does not feel itself bound by any rules or norms and has set up a separate existence for itself detached from the rest of the international community. During the course of 2017, the first year of Donald Trump’s administration, Pyongyang accelerated its nuclear testing and, most alarming for the United States, its intercontinental ballistic missile (ICBM) program. This put the continental United States directly at risk, which could be seen as a challenge to the U.S. nuclear guarantee to South Korea. Yet it was not as if the North Koreans had been without retaliatory threats up to this point, including a formidable conventional deterrent in the mass of artillery that had Seoul in range. It was therefore not clear what the United States could do to prevent North Korea from deploying ICBMs without taking enormous risks. This became a question of whether the United States was prepared to risk North Korean retaliation against South Korea (and Japan) in order to remove a threat that undermined its ability to deter attacks on South Korea.

Not surprisingly, the South Koreans engaged in direct diplomacy of their own with the North in order to calm tensions.13 Despite the bellicose rhetoric (“fire and fury”) of President Trump and of the North’s leader, Kim Jong-Un, this led to a surprise move for a summit between the two, which took place in Singapore a month after Trump had abandoned the Iran deal. His stance this time was quite different. Trump built up Kim as a worthy partner in the pursuit of peace, and accepted vague promises of “complete denuclearization.” The provisions were less firm than past agreements that Pyongyang had violated. The new situation was preferable to an arms race and heightened tensions, and meant that North Korea was unlikely to develop a direct threat to the continental United States, but was also likely to remain a nuclear power for some time. This created its own risks if the Americans came to feel that they had been cheated once again. Moreover, it demonstrated what a small nuclear arsenal might do for an otherwise insubstantial country in raising its international standing

Compared with past projections of how the nuclear age might develop, one could argue the situation is better than expected. There have been no nuclear weapons used in anger since 1945, credible ways to win a nuclear war against another nuclear power have yet to be developed, and relations among the main nuclear states have therefore been stable, while the nonproliferation regime has been surprisingly successful. Yet the underlying issues of extended deterrence remain in play. Russia has shrunk as a power but it is seen still as a threat by neighbors, including a number that were part of the same state not long ago, while China has grown as a power, and the threat it poses to its neighbors is assumed to have intensified. In these respects, therefore, the nuclear order has not changed so much, except that there are now additional issues revolving around India and Pakistan, North Korea, and Israel. A number of countries expected to loom larger in these calculations, notably Iran, having officially left the nuclear business for the moment, might now return.

The foundation for containing nuclear proliferation therefore remains the American network of alliances. There are two potential and related risks in this situation. The first is that the United States no longer wishes to play a role it has now accepted for almost seventy years and requires its allies to make other arrangements. The second is that the allies believe they must make new arrangements anyway, because they no longer trust American security guarantees, even if Washington tries to reassure them that nothing has changed. One of these developments could trigger the other, especially if it was the result of a major crisis.

Some argue that President Obama’s failure to manage the Syrian crisis effectively and his willingness to cut a nuclear deal with Iran has had a deleterious effect on relations with allies in the Middle East. I contend that doubts about the U.S. nuclear umbrella guarantee have been around since the 1950s. So long as the Americans professed to take it seriously, the allies professed to be reassured. Is this a spell that might one day be broken? As argued, the policies of President Trump represent a challenge to past assumptions about the conduct of American foreign policy and its solidarity with traditional allies. For the United States to sustain its network of alliances does not intrinsically depend on a readiness to offer a nuclear guarantee, although in practice one comes with the other. The United States—with its allies—could deal with most likely conventional threats without having to escalate. But the thought that this might lead an enemy to escalate will always be there.

Alliance works by reminding strong regional powers that they risk a fight with the United States if they start to pick on any of its allies. One argument against a system so dependent on the United States is not only the weight this places on the quality of American leadership, but the implication that any local conflict could get a whole lot worse, because it might turn what would otherwise be nasty but contained into a nuclear confrontation. Putin is not the only one who argues that American interference can turn what might otherwise be peaceable parts of the world into places of strife and discord. I have concentrated on one objection to this view, namely, that a world without an alliance system would be more, not less, conflictual, and more prone to nuclear proliferation. The other difficulty is that while it might be preferable to have a world system based on rules and strong institutions, somehow rules would have to be enforced. It is hard to see why the United States would play that role if it had abandoned its alliance network, which provided it with considerable influence over the course of events.

During the Cold War, especially in its early stages, political order was a possible casualty of nuclear disorder; now it is more likely that the nuclear order can be put at risk as a result of political disorder. Over time the deterrent effect of nuclear weapons became less dependent on the credibility of conditional threats and more on the possibility, however tiny, that they would be used in a major war. Some states depend on that possibility more than others and therefore will be inclined to stress that potential if they are edging toward a substantial conflict. There are risks when the United States has the least interest—because of its conventional strength—in the possibility of escalation, but the need exists for its nuclear threat to engage with the greatest range of contingencies because of its alliance commitments.

For now nuclear weapons continue to underpin one aspect of the current international system, the alliance network of the United States, without getting drawn regularly into everyday discourse on international affairs. It is an issue lurking in the background and only rarely pushed to the fore. When it is pushed to the fore—for example by North Korea when it wishes to draw attention to itself, or Russia when it wants the West not to intervene too strongly on behalf of Ukraine—anxieties are soon raised. It is difficult to see the dynamics that would lead any of the current conflicts to escalate into nuclear war, especially when the major powers have avoided skirmishes for some time, yet the anxieties that surface whenever this seems at all possible confirm how much the dreaded fear of a total war still weighs heavily on the international consciousness.

Recent wars have been less about great confrontations between states and more about civil wars, or ragged conflicts with many elements fighting within and over the same territorial space, often with links to criminality as well as terrorism. My earlier discussion on the Cold War stressed that the major transformation of the post-1945 order was the result of not so much change in interstate relations as in the “domestic order” of individual states. The progressive loss of legitimacy in Communist states was part of this. So too was the anti-colonial movement. Since 1991 most conflict has occurred within states rather than between them. These conflicts have on occasion led to considerable concern about the security of nuclear assets. President Obama made this a general theme of his administration, looking at a variety of ways in which terrorists and other non-state groups might get access to nuclear materials or weapons. The most dangerous cases would be those in which a state is breaking up—hence the concerns as the former Soviet Union was in disarray in the early 1990s about “loose nukes”; the fears about who is looking after Pakistani nuclear systems and how, as that country tries to cope with Islamist insurgents; or the uncertainty about the role played by nuclear weapons in North Korea’s internal politics.

The great shifts in international politics start in individual states. As I noted in the opening of this essay, a political order is a top-down concept, yet political change tends to start from the bottom up, with a radical movement or populist politician or economic discontent or social tensions that prove hard to contain. The history of nuclear weapons encourages us to think of them from the top of the international hierarchy, as the defining feature of the superpowers and as the great deterrent against a third world war, but we also need to keep an eye on them from the bottom as well. However stable any order—including the nuclear order—may look from the top, its greatest vulnerabilities may well be found at the bottom.

Author’s Note

I am grateful to Francesca Giovannini, Robert Legvold, Jeff Michaels, Matthew Harries, Heather Williams, and Daniel Salisbury for their advice on an earlier draft of this essay.