The normative problem of nuclear war in the thought of Kenneth Waltz and John Mearsheimer

Man, the State, and War

Kenneth Waltz founded Structural Realism in his 1959 book, Man, the State, and War, just, incidentally, as the nuclear revolution was taking shape in the Cold War world. Though he would not develop a complete structural model until his Theory of International Politics two decades later, the philosophical and methodological foundations of this new school of thought can be found in the first book. ⁴

Waltz was dissatisfied with the ‘classical realism’ (as it would become known later) of older Realists like Reinhold Niebuhr and Hans Morgenthau. In criticising their work, and that of other scholars in the Liberal and Marxist traditions, Waltz put forward two overarching arguments in Man, the State, and War that established a new way of thinking about international relations and led to the formation of structural realism.

First was his identification of anarchy as the ordering principle of international politics and as the root cause of the recurrence of war among major states throughout history. Earlier scholars had located the causes of war in human nature, or in belligerent national ideologies. Waltz effectively demonstrated why both of these explanations fail to account for the recurrence of war throughout time, and proposed that a better answer lay simply in the anarchical structure of the interstate system: the fact that there was no world government to prevent interstate violence. Anarchy, by itself, could explain the recurrence of war without relying upon reductionist appeals to human nature or warlike ideologies. ⁵ As long as the world remained anarchical, Waltz maintained, war would recur: this is why the last word in the title of the book was ‘war’.

Second was his call for an analytic, scientific study of international relations. A key component of this project was the rejection of normative scholarship, in favour of a rigorous social science which sought to explain and describe behaviour rather than seeking to apply it towards moral ends or policy prescription. Too many of his contemporaries, Waltz argued, undermined the coherence of their scholarship by calling for particular outcomes rather than simply seeking to explain them. Foreign policy, he continued, should not derive from normative objectives but rather ‘embodies a reasoned response to the world around us’. ⁶

Waltz’s core argumentation in Man, the State and War delivered a clear verdict on the question of nuclear war, a problem that had risen to the top of Cold War politics at the time of his book’s publication. As long as the world remained anarchical, there would sooner or later be a major war, and that now meant nuclear war. The structure of international politics made that inevitable, and individual leaders and citizenries would be powerless to prevent it. What is more, it was pointless and unscholarly to raise normative objections to this eventuality. Fears that the nuclear revolution had made war too dangerous to wage failed to heed the logic of the international and violated the demands of positivist scholarship. As Waltz wrote:

Each major advance in the technology of war has found its prophet ready to proclaim that war is no longer possible. Alfred Nobel and dynamite, for example, or Benjamin Franklin and the lighter-than-air balloon. There may well have been a prophet to proclaim the end of tribal warfare when the spear was invented and another to make a similar prediction when poison was first added to its tip. Unfortunately, these prophets have all been false. ⁷

In Theory of International Politics, written 20 years later and widely regarded as the foundational text of structural realism, Waltz transformed the structural logic of Man, the State and War into a rigorous social science theory. For our purposes, the basic claims in this book can be summarised quickly. Waltz assumes that states seek to survive, to continue to exist, in an environment in which no higher entity guarantees their security. He argues that international systems amidst this condition of anarchy ‘shape and shove’ the foreign policies of states, and especially the major powers, towards strategies of self-help, whereby they balance against potentially threatening adversaries by establishing alliances and/or building up their own military capabilities. Perhaps his most important claim was that bipolar systems, such as the Cold War, are more stable than multipolar systems, because the superpowers need not worry about shifting alliances and the possibility of defection. Thus they concentrate on ‘internal balancing’, on ensuring that their own military power is not falling behind their rival. This is an easier task than alliance management and helps to explain the stable Cold War order.

Much of the book is concerned with methodology, and this is more important for the argument here. Waltz maintains his position that the purpose of theory is, above all, explanation. He compares his theory to that of microeconomics, the analysis of the effect of market systems upon the survival strategies of firms. Economic theory, and as Waltz argues proper international relations theory, takes no normative position on what firms, or states, should or should not do: rather, the theoretical objective is to explain how the system pushes these units towards strategies of survival. In the anarchical world of international politics, this means balancing, the emergence of a balance of power, and so the ever-present possibility of conflict and war. ⁸

Waltz’s attack on reductionist approaches to the international illustrate his anti-normative approach aptly. He argues that any systems theory cannot derive from actions or intentions of the units, in his case states, and especially the great powers. The condition of anarchy explains the recurrence of balancing alliances and great power war over the millennia, and the job of systems theory is to explain why this is so and how it works – a task that requires the theorist to regard the units (states, and people) as both undifferentiated and doomed to eventual conflict. Waltz asks: ‘What do I mean by explain? I mean explain in these senses: to say why the range of expected outcomes falls within certain limits; to say why patterns of behaviour recur; to say why events repeat themselves, including events that none or few of the actors may like’. ⁹

Waltz’s theoretical position thus remains clear on the question of nuclear war – an outcome that very few actors would like. By attributing causation to structure rather than states and their leaders, Waltz is stating that even if every nuclear power adopted a policy of nuclear-war avoidance, and every citizen and leader were determined to prevent such a catastrophe, the system would overrule them, sooner or later. Anarchy and the perennial quest for national survival explains the recurrence of war, and this remains true in the nuclear age.

In this sense, structure – the ‘third image’ of international relations as he defined it in his first book, is amoral, uncharacterisable in normative terms. ¹⁰ Both the first and second images are inseparable from normative analysis, because these images attribute conflict and war to belligerence, whether it is found in the DNA of human beings or to warlike national ideologies. ¹¹ By attributing war to structure, Waltz removes the normative problem from his level of analysis, not only because it is beyond the control of human agency but also because, as a construct, structure can have no moral agency itself. It just obtains, like the weather: there is no one to blame if it rains. As long as the anarchical structure of international politics endures, Waltz states, ‘war occasionally occurs’. ¹²

However, Waltz makes two concessions in the latter part of Theory of International Politics. The first is his acknowledgement that the Cold War had been kept peaceful not only by the stability induced by bipolarity, but also by the unit-level factor of nuclear-war aversion. He insists that bipolarity is the more important of the two factors, but stresses that the ‘cataclysmic’ spectre of nuclear war bolsters the stable peace between the two Cold War superpowers. ¹³ The second is his striking last chapter, in which Waltz argues that the US and the USSR can take advantage of this durable stability to ‘manage’ international affairs: like oligopolies in other realms, they have the ability to address collective action problems by cooperating on matters of common global concern, including, Waltz stresses, the mutual interest in avoiding war. ¹⁴ He is careful to couch his discussion in this chapter in speculative and general terms, rather than calling overtly for global Cold War management. Nevertheless, he does admit that the unit-level factor of nuclear weaponry plays an important role in relations among the major powers, and makes a case for superpower cooperation one of whose benefits would be a markedly reduced chance of nuclear war.

Waltz could, and did, contend that these concessions did not undermine the larger logic of his theory. ¹⁵ Both points can be seen as additions to the house he built rather than the main edifice. But after the publication of Theory of International Politics, and particularly after the end of the Cold War, Waltz abandoned his rejection of the normative on the question of nuclear war.

More May be Better

In his many writings on nuclear politics since 1980 (he died in 2013), Waltz makes two arguments which move even further way from the structural fatalism of his first book. The first is a ‘deterrence absolutism’ based upon a common-sense perception of the catastrophic dangers of nuclear war. ¹⁶ For Waltz, it is simply obvious that states recognise that these dangers mean that it must be avoided: this is a fact of international life that inclines them to caution and automatically dissuades nuclear states from attacking one another.

In his prize-winning 1990 article ‘Nuclear Myths and Political Realities’, he resorts to this line of argument on several occasions. Why was there no nuclear war during the Cold War, and why will it be avoided in the future? It is because it is ‘obvious to everyone’, from great-power leaders to ‘the man on the street’, that nuclear war is a catastrophe that must be prevented. ¹⁷ During showdowns, Waltz imagines, leaders do not, as in previous times, calculate rationally about the merits of escalation or compromise: they ask themselves ‘one question: how do we get out of this mess?’ ‘Nobody but an idiot’, Waltz states in his debate with Scott Sagan, fails to see the sense of finding some way to stay out of a nuclear war. ¹⁸

In attributing nuclear war-avoidance to the common sense of unit-level actors such as states and people, Waltz makes two exceptionally important moves. Most notably, perhaps, this approach enables him to ignore, and even disdain, official policy-making. Yes, states develop strategies of nuclear victory and build arsenals accordingly, but this can be written off as ‘human error and folly’. ¹⁹ Leaders and governments may plan for nuclear war, but really matters is not what they do in preparing for it, but how they behave when a nuclear war actually becomes possible. For Waltz, the salient behaviour of states in the nuclear age comprises decisions about war or peace rather than formal policy. Waltz brushes aside objections about what states should do when facing defeat in a conventional war, or if they are attacked with a limited nuclear strike, because the iron logic of deterrence absolutism suggests that these scenarios are not going to eventuate in the first place. Nuclear states will not initiate major conventional wars against another for fear of nuclear escalation; even less likely are they to launch a limited nuclear strike. The nuclear history of the Cold War, and the post-Cold War era, reveals this truth. ²⁰

Moreover, Waltz’s unit-level actors not only are decisive in preventing nuclear war; they also operate in the knowledge that they can pursue these decisions because ‘everyone knows’ that nuclear war is a catastrophe to be avoided. What Waltz means by ‘everyone’ is precisely that: all nuclear states, large or small, rich or poor, can be trusted to understand the vivid dangers of war, a fact that enables, or ought to enable, nuclear states to avoid intense security competition and arms racing. In his debate with Sagan, Waltz stresses this point. Any national leader, of a large democratic state or a small authoritarian one, will be able to see the folly of nuclear war and avoid getting into one. Nuclear wars threaten national existence, and rulers, no matter how radical, ‘want to have a country they can continue to rule’. ²¹

The second component of Waltz’s approach to nuclear war is his normative opposition to it. What this means is not simply that nuclear states will recognise that war is insane and avoid it, but that it is good that they do so and policies should be advocated that enhance deterrence and war-avoidance.

Of course, Waltz does not openly call for nuclear peace or go on at length about the evils of nuclear war, though he routinely refers to it as a ‘catastrophe’, ‘annihilation’ or a ‘global tragedy’. In a 1988 article, he states bluntly that ‘nuclear weapons make the implications of victory too horrible to contemplate’. ²² Indeed, a normative opposition to nuclear war, if one accepts (as Waltz does) the nuclear revolution position that any nuclear war is likely to lead to omnicide, might be consistent with structural realism’s emphasis on state survival. But in his case for the spread of nuclear weapons, Waltz develops an explicitly normative argument that goes beyond simple war avoidance. In his many published debates with Sagan, Waltz repeatedly states that the gradual acquisition of nuclear weapons by certain states would be a good thing because it would contribute to stability and keep the peace. This normative position is most apparent in his (in)famous last writing, a Foreign Affairs article entitled ‘Why Iran Should get the Bomb’. ²³ His case for this controversial view is wholly prescriptive: he argues that the United States should stop trying to prevent an Iranian bomb, because such acquisition would stabilise both Iran and the Middle East, thus reducing conflict and danger. Waltz attempted to downplay his normative agenda in early works: in his original 1981 paper ‘More May be Better’, he asked, awkwardly, what would the spread of nuclear weapons ‘do to the world?’ ²⁴ In his ongoing debates with Sagan, and finally with the Iran essay, Waltz abandoned any pretence of neutrality and argued explicitly in favour of policies which he believed would minimise the chances of nuclear war.

In a short 2004 essay, Waltz responded to his critics. By this time, criticisms of structural realism had become, perhaps, the most prevalent theoretical branch of argumentation in IR. A leading form of attack on structural realism, he noted, was its indifference to moral concerns. ‘Some have complained that normative considerations are omitted’, Waltz wrote. Should they be added? To ask the question is like asking whether we should add to a theory that explains gravity a warning that it is unwise to fall from high buildings. ²⁵ The irony, of course, is that this was precisely what he had been doing for the previous quarter-century.

The tragedy of great-power war

In The Tragedy of Great Power Politics, first published in 2001, John Mearsheimer sought to revive positivist structural realism for the post-Cold War world. The core argument of the book is straightforward: only a theory that assumes that the major states seek to maximise their power, not just preserve it relative to their rivals, is consistent with the logic of structural realism. Working from this fundamental assumption, Mearsheimer aims to establish a general theory of international politics which can explain great power behaviour across all time and space. He argues that the structural pressures upon states to increase their power at the expense of their rivals leads inexorably to a ‘harsh world of security competition and war’. ²⁶ Barring the establishment of an authoritative world government, the world is ‘condemned to perpetual Great Power competition’, and that ‘conflict and war are bound to continue as large and enduring features of world politics’. This is what he means by the word ‘tragedy’ in his title.

For Mearsheimer, the intellectual starting point for his theory is his argument that defensive realism, as advanced by Waltz in his book Theory of International Politics cannot explain the actual behaviour of the great powers. In a devastating article, published a few years after the first edition of Tragedy, Mearsheimer spells this reasoning out.

Waltz assumes that states simply seek survival. ²⁷ They just wish to be left alone, but in an anarchical environment they must concern themselves with the capabilities of other leading states, in order to maintain a balance of power. War may occur if the balance of power unravels and states overreact to the threat posed by their rivals, or if reckless leaders, driven by domestic pathologies, seek conquest. For Waltz, therefore, great-power wars are always caused by foolish decisions and overreactions made by reckless leaders. Because states seek only survival, international politics ought to remain stable and peaceful. Wars are a result of dysfunctional behaviour, and those who start them are inevitably punished by ‘the system’ – that is, the balancing alliance that arises to stop them. ²⁸

Because international politics until 1945 (a point to which we will return shortly) was replete with the initiation of major wars, from Napoleon to Bismarck to Hitler, Waltz is forced to write them off simply as consequences of reckless behaviour. These acts are unstrategic, peculiar to the actors at the time. They defy the logic of the system, which rewards defensiveness and produces stability. His theory, which operates entirely at the structural level and does not account for the foreign policies of particular states, cannot and does not seek to explain these particular reckless acts at odds with nations’ security goals. ²⁹

The problem becomes apparent. If the top story of international politics through the Second World War is that of states launching wars that make no sense, then Waltz is acknowledging that his theory is incapable of explaining manifestly important great-power behaviour. Waltz effectively creates in this way of thinking an unfalsifiable hypothesis: if states act defensively and peace is maintained, it reveals that they are responding to the structural environment as he would predict; if they launch wars, it is because of domestic pathologies that are outside of his theory’s purview. Defensive realism cannot be proven wrong.

In the most profound insight of an article that contains many, Mearsheimer concludes that Waltz’s defensive realism is at heart a normative theory: it paints a picture of how great powers ought to act in a condition of anarchy, not how they do act. Rational states and their leaders, pursuing only survival, should avoid initiating a major war that will only lead to their undoing. If they followed the logic of defensive realism, the world could see perpetual peace. Wars occur because states do not do that. ³⁰

By characterising Waltz as a normative theorist, Mearsheimer is clearly saying that, in contrast, he is not. Defensive realism is normative, offensive realism, scientific. It is telling, however, that Mearsheimer does not explore why Waltz had adopted a normative theory. Given Waltz’s copious writings on the nuclear question since 1980, and the fact that normative opposition to major war since Waltz began writing is inescapably opposition to nuclear war, Mearsheimer could not have been in any doubt about the meaning of Waltz’s normative move. Yet he does not even raise this point. As we shall see, there is an explanation for this omission.

In Tragedy, Mearsheimer develops a theory of offensive realism that avoids the trap Waltz finds himself in. Great powers, he argues, often act aggressively not because of domestic belligerence or irrationality, but because it is a rational strategy to acquire as much power as possible lest one’s rival does the same. Thus major powers seek hegemony, the domination of global spheres of influence, not out of folly or imperial lust but because it is the surest way of guaranteeing their survival. States must have the power to deter reckless revisionist states at all levels of hostility, even if by doing so they may appear reckless themselves. The optimal strategy therefore is to bid for global supremacy and full spectrum military dominance, but this has been impossible because of the difficulty of projecting hegemonic power across the oceans, what Mearsheimer calls the ‘stopping power of water’. ³¹

With this move, Mearsheimer is able to revive structural realism. He can explain the long history of great power aggression and war by characterising it as a rational response to structural pressures rather than the result of unit-level foolishness, which is precisely what makes it a structural theory. At the same time, he abjures the implicit normativism of Waltz and other defensive realists, stating clearly that great-power competition and war cannot be wished away.

The demarcations between offensive and defensive realism are thus clear. Defensive realists are optimistic about international politics and believe that peace is basically assured as long as states act as the logic of the international system dictates. ³² As Waltz argued, this condition is only intensified by the spectre of nuclear war, which presents great powers with a clear picture of the apocalyptic costs they will incur should they act foolishly.

On the other hand, the offensive realist Mearsheimer is pessimistic about international politics. He sees the world as rife with security competition and war, and rejects the implicitly normative reasoning of Waltz in favour of an explanatory theory that predicts belligerent great power behaviour for good systemic reasons. Therefore, Mearsheimer cannot agree with Waltz’s positions on deterrence absolutism and the iron logic of MAD, because if great powers are rewarded for aggressive behaviour, they can and will not accept a condition in which such behaviour will lead to their total destruction, and so will seek new strategies that will allow them to overcome it. ³³ To put it another way, if Mearsheimer allows that MAD obtains in contemporary international politics, then we will never be able to confirm whether great powers act as his theory predicts. Either they remain forever defensive and avoid war; or they do not and the world is destroyed.

To put it bluntly, if contemporary great power politics is rife with conflict and war, and these powers are wired to adopt aggressive rather than defensive strategies, then the event which would best confirm offensive realism would be the deliberate (rather than inadvertent) waging of nuclear war by an aggressive great power. Does Mearsheimer accept this? Nowhere in his writing does he confront this question directly.

To be sure, he asserts that war remains a permanent feature of great-power competition: even a war between nuclear powers, he writes in Tragedy, is ‘still thinkable’. ³⁴ In a recent response to a book review roundtable, he sides with authors like Keir Lieber and Daryl Press who argue that the nuclear revolution is a ‘myth’ – that it does not explain actual behaviour, and that great powers are rational to pursue nuclear superiority. ³⁵

Yet in other places he seems to reject this conclusion. During the early post-Cold War period, he argued that nuclear proliferation to Ukraine and Germany would stabilise European international politics, agreeing with Waltz that nuclear arsenals provide states with a powerful means of security. ³⁶ This argument, as Waltz makes clear, depends upon the assumption that states will never try to conquer others in possession of a secure nuclear arsenal, which is the basis of MAD.

Mearsheimer also highlights the unique dangers of nuclear war on occasion. In Tragedy he argues that the ‘presence of nuclear weapons makes states more cautious about using military force of any kind against each other’, which clearly suggests that states are afraid to wage any war lest it escalate to the nuclear level. ³⁷ In a later book, The Great Delusion, he states that the consequences of nuclear war would be ‘horrendous’ and that nuclear weapons serve as the ‘ultimate deterrent’. ³⁸ In a recent interview, conducted only several months after he endorsed the Lieber/Press thesis, he goes considerably further: the US and the USSR steered clear of conflict in Europe, he states, because any war there was likely to become nuclear quickly. That never happened, he continues, ‘because nobody in his or her right mind, would start a war given the possibility of nuclear Armageddon’. ³⁹

Needless to say, his occasional endorsements of the logic of MAD pose a categorical problem for Mearsheimer’s theory. Is major war in the nuclear age an inexorable, if tragic, outcome of a great power order that incentivises aggression and expansion; or is it an ‘Armageddon’ that nobody in his or her right mind would ever start? Does his theory apply to great-power conflict in the nuclear age, or does it not? ⁴⁰

Mearsheimer’s logical inconsistency with respect to MAD can be illustrated more specifically by examining his analysis, or lack thereof, of two basic questions about the nuclear age. First, why is it that, after two great-power wars in the space of 30 years during the first part of the 20th century, there has been no such war during the past seventy-five? Second, why did the Soviet Union peacefully surrender, handing the West total Cold War victory, in late 1991? Both of these questions, concerned with war and great-power conflict in the most fundamental sense, should obviously be central to Mearsheimer’s inquiry. Yet he does not confront them.

During 45 years of Cold War, and for 30 years now afterwards, we have seen no great-power war. Indeed, since the Cuban Missile Crisis of 1962, more than 60 years ago, the great powers have assiduously avoided direct confrontation. This fact poses an evident problem to a theory which stipulates that great-power politics is rife with competition and war. Of course, many scholars have sought to answer this question. Notably, liberal theorists such as John Mueller and Stephen Pinker argue that peace among the major powers since 1945 can be explained by moral improvement in human and state behaviour; other liberals such as John Ikenberry and Bear Braumoeller point to the effectiveness of international institutions and trade. ⁴¹

MAD, of course, provides Realists (and others) with another, simple answer to this question. The great powers have avoided war since 1945 because in the nuclear age they have seen it as a threat to their survival – and possibly to that of humankind generally. Moreover, the acquisition of secure second-strike arsenals by both superpowers during the Cold War, and by the great powers of the post-Cold War era, establishes a condition of general deterrence and provides them with reliable security, a fact that, as any Realist would agree, further reduces the likelihood of war. But Mearsheimer does not explain the long peace by arguing that it was due to MAD, and indeed refrains from attempting to explain it at all. ⁴²

A similar interrogation of Mearsheimer’s evasive approach can be made in examining his treatment of the end of the Cold War. The peaceful surrender of the USSR, and so the termination of the bipolar order without the systemic war that Realists had long argued always characterises transitions from one great-power system to the next, has widely been seen as a devastating, and perhaps fatal, blow to structural realism. Great powers like the Soviet Union, still in possession of massive military capabilities when it raised the white flag, are not supposed to simply hand victory to their rivals without a fight. ⁴³

As with the broader question of the long peace, one would expect Mearsheimer to contend directly with a puzzle that manifestly appears to threaten his theory, and certainly to contend with IR scholars who see Soviet surrender as vindication of their own theories and a fatal blow to structural realism generally. Instead, in only 2 paragraphs of a 400-page book, Mearsheimer argues that realism can explain the end of the Cold War because Soviet leaders, witnessing their nation’s economic decline, chose Cold War defeat because they recognised that they could no longer compete with the West. ⁴⁴

This argument does not follow from the fundamental assumptions presented in Tragedy. For Mearsheimer, economic capabilities matter only insofar as they contribute to military power, ⁴⁵ and the Soviet Union possessed in 1991 the world’s largest army and a massive nuclear arsenal. It had been locked into a bipolar struggle with the United States, following its colossal victory over Nazi Germany, for almost 50 years, spending billions of rubles and engaging in security competition with the United States across the globe. If Mearsheimer is saying that a great power like this is free to surrender when its leaders perceive that their economy is faltering, to abandon its control over Eastern Europe and watch it join NATO, cease contending with the United States everywhere, open up its borders to the capitalist imperialists, and, over the next several years, cut its military spending by as much as 80%, then it becomes difficult to see what action a great power could take that could not be fit into his theory. ⁴⁶ Mearsheimer could do away with all of these problems by using MAD to explain great-power peace since 1945 and the peaceful disappearance of a superpower. I contend that he does not use this explanation because he understands that relying on MAD to account for these world-historical events endangers the salience of his theory.

Mearsheimer and the post-Cold War era

The end of the Cold War indisputably poses another grave challenge to structural realism. And as with the question of the long peace, Mearsheimer has access to a realist answer to the question: MAD. The Soviet Union chose to surrender because its other alternatives were to fall further behind the West economically, or to wage war, as declining great powers have done in the past, but which would this time culminate in nuclear annihilation. Moreover, because the Russian successor state would still possess a secure arsenal, Moscow would not have to worry about being conquered by the victors, as has also happened to collapsing great powers throughout history. Yet Mearsheimer declines to make this argument, relying instead on an interpretation of Soviet surrender that does not follow from his theoretical assumptions.

By making war suicidal and security easy, nuclear weapons and the condition of MAD they create provide realists with a plain means of explaining eight decades of great-power peace and the shocking superpower suicide of the USSR. It speaks equally to the post-Cold War environment. As is well known, Mearsheimer foresees an inevitable clash between the United States and China, and the return of great-power conflict after a unipolar interregnum. ⁴⁷ This new Cold War, Mearsheimer predicts, is certain to lead to intense security competition between the two states, and possibly war. But if MAD is as decisive as Mearsheimer sometimes suggests, would not the United States (and China) be deterred just as certainly as the two superpowers were during the Cold War? China currently possesses a secure-second strike arsenal, and that is all that is needed for MAD to apply. ⁴⁸

Mearsheimer’s recent argument that NATO expansion, the Bucharest declaration that Ukraine and Georgia would someday join NATO, and general Western hubris triggered justifiable security concerns in Moscow, leading to the present war, has made him one of the world’s most famous intellectuals. But what explains why he has taken such a controversial position? As his University of Chicago colleague Paul Poast points out, it clearly does not follow from his structural theory, in that Mearsheimer’s depiction of Russian policy is in accord with classic defensive realism; an offensive realist, Poast shows, would see Moscow’s aggression as a resumption of its Cold War attempt to dominate its region, not a reaction to Western policy as such. ⁴⁹ Structure, not fears of NATO or liberal infiltration, should, according to offensive realism, explain the war in Ukraine.

Mearsheimer might reply that his attacks on Western policy derive from his concern that the US should be focussing on China rather than Russia, an argument that is consistent with offensive realism, or from his ideological opposition to American liberal hegemony. But it is striking that, in recent work, he does not stress either of these first two points. Instead, he emphasises the danger of nuclear war. Like his predecessor Kenneth Waltz, his normative concern with this problem is becoming more transparent in the latter part of his career.

In an August 2022 essay in Foreign Affairs, Mearsheimer argues that the war in Ukraine is likely to last indefinitely, as both sides have put themselves in a position where backing down is politically difficult or impossible. In this piece, however, he does not lament this situation by insisting that the West is fighting the wrong enemy, as offensive realism might prescribe. Instead, the entire thrust of this piece, titled ‘Playing with Fire in Ukraine’, is that the problem with the West’s expansionism in Eastern Europe is that it has substantially increased the likelihood of a direct conflict between Russia and the US, and so the ‘nuclear annihilation’ that could follow. Even if this is avoided, Mearsheimer continues, Russia could use tactical nuclear weapons in Ukraine. In either event the results would be ‘catastrophic’. ⁵⁰

He makes similar points in an interview with Freddie Sayers on the website UnHerd, also ominously titled ‘We’re Playing Russian Roulette’. Though Mearsheimer, as we have seen, has endorsed the recent argument that there has been no nuclear revolution, and that states engage in security competition and wage war just as before, he takes a different view here: If Russia were to use nuclear weapons in Ukraine, Mearsheimer states, the US ‘is not going to retaliate with nuclear weapons against Russia, because that would lead to a general thermonuclear war’. He repeats his endorsement of pure deterrence in a vivid passage:

I also think there’s a non-trivial chance that this will lead to nuclear war. And when you marry the consequences of nuclear war with the possibility, in my mind, that means you should be remarkably cautious. Let me illustrate this by this analogy. If I have a gun, and the barrel has 100 chambers, and I put five bullets in that barrel. And I say to you, Freddie, I’m gonna pull the trigger and put the gun up to your head. But don’t worry, there’s only a 5% chance that I will kill you . . . The question you have to ask yourself is, are you going to be nervous? Are you going to be scared stiff? . . . The consequences here involve nuclear war. So there only has to be a small probability that John is right. ⁵¹

Mearsheimer is most explicitly normative on the problem of nuclear war in an interview on the PBS News Hour in May 2022. His criticism of Biden administration policy on the war in Ukraine is entirely based upon his fears that this policy raises the possibility of such a war:

We know that the one circumstance in which a great power is likely to use nuclear weapons is when its survival is threatened, when it thinks a decisive defeat is being inflicted on it. And what the Biden administration is bent on doing is inflicting a decisive defeat on Russia. We are threatening its survival. We are presenting the Russians with an existential threat. And this, again, is the one circumstance where they might use nuclear weapons. And I think we should be going to enormous lengths to make sure that we don’t put them into a position where they even countenance using nuclear weapons, much less use them. ⁵²

It is difficult to overstate how plainly this argument departs from offensive realism. Attempting to inflict ‘a decisive defeat on Russia’ is, according to offensive realism, something the United States might well be expected to do. A policy of going to ‘enormous lengths’ to avoid antagonising Russia for the purposes of nuclear war-avoidance, on the other hand, is textbook defensive realism, if not something more dovish entirely. Mearsheimer might respond that he is opposed to Western policy on Ukraine for a number of reasons, and that the danger of nuclear war is just one of them. This may well be the case, but his recent highlighting of the nuclear danger strengthens the claim here that, like Waltz, Mearsheimer has gravitated towards a more explicitly normative position on this danger.

Mearsheimer recognises, as does everyone else who studies international politics, that the consequences of nuclear war are unprecedentedly destructive. Scholars debate whether some nuclear wars are worse than others, whether a nuclear war can be won, and whether states accept – or have ever accepted – MAD. On this question, as we have shown, Mearsheimer occupies a kind of middle ground, between ‘deterrence absolutists’ like Waltz and nuclear strategists, such as Lieber and Press, who believe nuclear wars can be fought and won. His theory of offensive realism, however, is inconsistent with this middle ground approach. Great powers as he depicts them will not accept the stalemate that MAD imposes; if they do, offensive realism will be unable to explain a fundamental aspect of great-power behaviour. Mearsheimer would therefore have much to gain theoretically by allying himself more firmly with the strategists. I have shown that he has been unwilling to do this, and suggested why this may be so.

The dilemma of structural realism

Waltz and Mearsheimer both theorise that structure, rather than unit-level behaviour, determines the outcomes of international politics. The most important of these outcomes, they both argue, is great-power war. As long as the anarchical system persists, such a war will eventually occur. Both scholars, however, have defied this theoretical position by advancing normative arguments designed to reduce the chances of a nuclear war occurring. Indeed, what they seem to be doing is to theorise that war will eventually occur, but in the meantime they are going to do their best to prevent it from happening – and, one must assume, that they think their efforts could have an effect. ⁵³ Presumably, they believe that others in the future should also do their best to prevent it from happening, and that their efforts also could have an effect. Their position on the problem of nuclear war, then, would seem to be hoping for its indefinite prevention, which can happen, if one takes them at their word, as a result of the normative actions of individual people.

Yet one must ask, in conclusion, if their larger conceptions of international politics, and in particular US foreign policy, are consistent with this normative position. In one sense, Waltz’s position is fairly consistent, because he lays out a clear case for avoiding nuclear war, even if he sometimes tries to conceal this. For Waltz, deterrence works perfectly, and so he would prefer that all nuclear powers accept MAD and avoid confrontation, as Robert Jervis argued in the 1980s. ⁵⁴ As we have seen, moreover, he believes that the spread of nuclear weapons would reduce global conflict and war – the logical conclusion of which would be a world in which every state had a nuclear arsenal and the possibility of war more or less disappears. The implications for US foreign policy are thus clear: America should forego any attempt to develop a war-winning nuclear strategy, scale back its arsenal to one of basic deterrence, and stop supporting the nonproliferation regime. Waltz would therefore have been a thoroughgoing critic of America’s actual foreign policy today – as his last piece on the Iranian bomb foretold.

Yet Waltz’s vision of a world of nuclear ‘porcupines’ is, in the end, difficult to reconcile with his normative position. ⁵⁵ As Sagan stresses in his long-running debate with Waltz, a world of dozens or hundreds of nuclear states might well reduce the chances of deliberate interstate war, but it would certainly increase the likelihood of an accidental or inadvertent one. One need not adopt the paternalistic, or even racist, view that certain peoples are less capable of avoiding nuclear use than others ⁵⁶ to recognise that the sheer increase in numbers multiplies the chances of something going wrong. A regional nuclear war initiated by accident would lead to unprecedented disaster and potential environmental catastrophe, and it could easily escalate to a general war, particularly if the belligerent states are allied to larger powers.

What is more, if Mearsheimer is correct that Waltz attributes great-power war to recklessness and over-reaction, then his vision of a world composed of many nuclear powers requires him to insist that such foolish behaviour will never again occur. Waltz argues, correctly in my view, that the prospect of nuclear apocalypse makes states far more cautious and risk-averse than before, which means that war among the major powers is much less likely than it was in the pre-nuclear period. But the omnicidal consequences of general nuclear war mean that the stakes of Waltz being wrong, just once, become existential.

Waltz made his normative preferences clear and coherent on nuclear war-avoidance and tended to be less outspoken about US foreign policy. On the other hand, Mearsheimer, as we have seen, has had much less to say systematically about the nuclear problem, but has made his normative views about American policy very clear: the United States ought to regard China as an inevitable great-power rival and act to contain it immediately, because offensive realism predicts that China is destined to seek hegemonic domination over Asia.

How does this position align with Mearsheimer’s recent commentary on nuclear-war avoidance? As Jonathan Kirshner argues in a recent book, not well at all. ⁵⁷ This is because one of the surest ways to increase the chances of a deliberate general nuclear war – as opposed to a regional, or accidental one – would be for the United States to aggressively confront China in the Asian theatre. If the US chooses to start a second Cold War with China, as Mearsheimer both advocates and argues is inevitable, then logically it will find itself in a security competition similar to its conflict with the USSR during the Cold War. Everything in Mearsheimer’s reading of international politics would suggest that this security competition would lead, on occasion, to direct showdowns over key stakes, just as it did over Berlin or Cuba several decades ago.

Must the United States and China wage another Cold War? As Kirshner shows, the case for this on basic national security grounds for either state is not self-evident:

Is China’s “survival” really in jeopardy if it does not aggressively bid to dominate all of Asia? Will the United States not “survive” if it fails to reach across the Pacific Ocean to crush a rising China before it is too late? . . . What exactly threatens the survival of these great powers? Given their military establishments, their nuclear deterrents, their economic might, their continental size, and their vast populations, is their survival really imperilled if they do not act as offensive realists? ⁵⁸

It is difficult to see how Mearsheimer could deny that the course of action he demands the US undertake increases the likelihood of nuclear war, especially when contrasted with alternative approaches – most evidently, Waltz’s defensive realism. He would certainly respond that defensive realism does not accurately portray how the world actually works, and that a second Cold War between the US and China is unavoidable. This is certainly a logical position to take, but it cannot be reconciled with his normative views on nuclear war, which suggest both that war-avoidance must be prioritised, and that unit-level behaviour can have an effect on international politics. I have no doubt that Mearsheimer is aware of this problem. In the last sentence of the second edition of Tragedy, he writes: ‘Given the grim picture that I paint, let us hope that if China becomes especially powerful, the actual results of that development will contradict my theory and prove my predictions wrong’. ⁵⁹

Structural realists have long purported to adopt a scientific, morally neutral approach to international politics. They point out, following Thomas Hobbes, that ethical judgements are untenable with respect to an anarchical environment, where there is no universal arbiter and justice is defined to the liking of the powerful. This important insight, however, reaches its limit in the nuclear age. One might be able to remain neutral about genocides, invasions, imperial predations, or even the conquest and disappearance of entire nations (at least if they are not one’s own), but it is impossible, or at least absurd, to take no normative position on the possibility of human extinction that nuclear war raises. ⁶⁰ I have tried to show how both Waltz and Mearsheimer have taken such a position without reckoning with the damage this move does to the logical integrity of their theorising. In future, realists would do better by acknowledging the inescapably normative problem of nuclear war, and modifying their theories accordingly.