How States Think: The Rationality of Foreign Policy, John Mearsheimer and Sebastian Rosato, Yale University Press, £25

Following Vladimir Putin’s failed attempt to seize Kyiv in 2022, liberals seemed to enjoy a period of vindication. Western leaders had been warning Russia in the run up to its invasion that such a move would be self-defeating, it would lead to quagmire, it would distract from needed economic measures and make both Russians and the wider world poorer – and all of these things seemed to be proved true. Meanwhile, the realists, who had argued for years that western support for Ukraine would provoke a war, and that in such a war Ukraine stood little chance, stood chastened and condemned before polite opinion, contemplating a scenario in which right seemed to make might.

That was then. As the summer of 2023 draws to a close and the Ukrainian counteroffensive grinds on with very modest territorial gains, the pattern of facts seems more favourable to the realists. To take a prominent example, witness John Mearsheimer’s long essay, published on his Substack, detailing at great length why the counteroffensive was ‘bound to lose’, and explaining, essentially, why his predictions and counsel had been right all along. Realists of Mearsheimer’s sort believe that they are the professional skunks at the garden party, responsible for bearing the unwelcome news about how the world really works.

Such an approach is taken up with gusto by Mearsheimer and his co-author Sebastian Rosato at the start of their new book, How States Think, which also came out last month. Citing a collection of Western think tankers, journalists, and politicians all making versions of the point that Putin’s invasion proves that he is in some way an irrational actor, Mearsheimer and Rosato propose that, on the contrary, Putin is very much a rational actor, a fact to which his liberal critics are blind because they do not understand what rationality means. Their short book proposes to correct this and similar misunderstandings by means of a self-described ‘radical’ project.

The project is radical because the challenge goes beyond limited claims regarding Putin or any other individual troublemaker. Mearsheimer and Rosato detect an increasingly strong acceptance among both academics and laymen that ‘states’ often behave nonrationally. In their view, the growing acceptance of this position is a serious problem both for the modern field of international relations theory, which has generally been premised on the notion that states and their leaders are rational actors, and also for political leaders, who faced with a world of madmen would be unable to make good decisions. As neither the study nor practice of international politics ‘can be coherent in a world where nonrationality prevails’, the authors intend to demonstrate that ‘states are rational most of the time’. They will do so by offering ‘a meaningful definition of rationality in international politics where none existed’, then comparing it to a series of cases where nonrational state behaviour has been claimed by others. The historical record will then show that most of these cases have been examples of rational behaviour after all.

What is this new, meaningful definition of state rationality? At the individual level, ‘Rationality is all about making sense of the world for the purpose of navigating it in the pursuit of desired goals.’ That is to say, rational people form theories of the world and make decisions in light of those theories. In the words of the authors, man is ‘homo theoreticus’. But not every theory is rational – it must be a credible theory. One gets no points for acting in accordance with the view that the moon is made of cheese. Furthermore, a ‘state is rational if the views of its key decision makers are aggregated through a deliberative process and the final policy is based on a credible theory’.

Clearly, a lot hangs on the word ‘credible’ here. The authors provide us with a three-part test of credibility: the theory must be based on realistic assumptions, must involve logical causation, and must be verified by empirical support in the historical record. Even more helpfully, they then list the theories that are credible in the context of international politics, and condemn the ones that are not.

Which theories make the cut? As it happens, they are the academic theories promoted by the major modern schools of international relations. Defensive, offensive, and hegemonic realism and their major subvariants are on the approved list, as are (ecumenically, it must be said, given the authors’ realism) democratic peace theory, economic interdependence theory, liberal institutionalism, and finally a set of theories falling within social constructivism.

How can a state hope to navigate rationally in the face of such generous pluralism? Well, as ‘no credible theory applies to all problems’, it’s anticipated that policymakers will mix and match among credible academic theories, which ‘find their way into the minds of aspiring decision makers before those individuals begin to make policy’. Moreover, the authors also helpfully provide a list of noncredible theories, ranging from the universally condemned (racial theories prevalent in scholarship a century ago) to upstart academic theories today that, in the authors’ view, fail some element of their three-part test (for example: neoclassical realism, audience costs theory, nuclear coercion theory, and several more).

In other words, Mearsheimer and Rosato’s book is an attempt by pre-eminent international relations scholars to seize the mantle of who gets to define rationality in the context of international politics from scholars dealing with rational choice theory and from the political psychologists who had been enjoying that privilege. In the view of the authors, both of these schools are failing to live up to their responsibilities, in the first case by failing to describe a mental process in which rational decisions actually happen, and in the latter case by simply surrendering to the argument that irrational choices are prevalent – that people rely on analogies or heuristics rather than full-blown theories. The historical record simply shows, in the authors’ view, that this is not the case – state level decision-making seems to attract a clearer-headed sort of person who avoids these kinds of errors – and even, for the most part, avoids being overcome by his or her passions, as ‘instances of emotions driving the train [in international politics] are rare’.

So much for the liberal pundits, the economists, and the psychologists. If the approach described here seems solipsistic, or perhaps a bit circular – the approach of international relations theorists, which depends upon rational decision making by policymakers, is vindicated because in fact policymakers make rational decisions using international relations theories that are definitionally rational, say international relations theorists – well, the historical record will have to be the judge.

And it is to the historical record that our authors turn, in brisk essayistic fashion, taking on twenty-one cases of major state decision making that others have claimed are nonrational, and (with a few exceptions) begging to differ. It is here that the book’s issues come into starker relief – though in fairness, given Mearsheimer’s (and to an extent Rosato’s) prominence in the field, we should take the methods used here as a guide to what the field itself considers an acceptable, even exemplary standard of serious work – and so perhaps it is the field’s issues that are made clear.

Here’s an example. In the course of discussing crisis decision-making, specifically as it pertains to the conduct of war, Mearsheimer and Rosato take up the case of the decision of the United Nations—in effect, of the Americans—to cross the 38^th Parallel in the fall of 1950 in Korea. They select a target who has made the case that this decision was nonrational, in this case the scholar Irving Janis, who argued that this development was an illustration of groupthink and thus (in our authors’ terms) ‘nondeliberative,’ or failing an important part of the test of state rationality. Our authors then assert the contrary: ‘The American decision to move into North Korea was the product of a deliberative process, not the result of groupthink.’ They then take a paragraph to develop this point with a series of assertions – there were frequent meetings of the relevant leaders, there was no evidence of in group pressures, the participants were generally in agreement about something wrong, of course (Chinese intent and capabilities), but that’s not evidence of the lack of a deliberative process. The name ‘Douglas MacArthur’ is not mentioned. There is an end note to the paragraph which, beyond Janis, cites three books: David Halberstam’s The Coldest Winter (we are directed to part seven), Max Hastings’ The Korean War (Chapter six), and a chapter of a monograph on the Chinese decision to enter the war published sixty-three years ago. The authors then move on to the question of escalation in Vietnam.

It seems worth questioning whether the rhetorical method employed here is intended to persuade anyone who is not already quite sympathetic to the case being made. I should also add that my example, though crisp, is a little unfair, as it’s one of the shorter case studies in the book. Some cover several pages. For example, the authors’ treatment of Germany’s decision to launch Operation Barbarossa, commonly thought of as a paradigmatic example of Hitler’s chaotic decision making and, to the extent motivated by a theory, driven by racist and genocidal visions of German dominion over Slavs and the elimination of the Jews, includes quotes from many of the key German players – including Walther von Brauchitsch, Franz Halder, Alfred Jodl, and Hitler himself – that makes them sound very much like adherents of ‘straightforward realist theory’, which our authors assert was the dominant Nazi thinking driving the invasion. These quotes, with one exception, are pulled from secondary literature. There are no illustrations provided of any of the players thinking in racist or ideological terms, though our authors grant that such motivations did exist – but they assert such issues were ‘secondary factors’.

Strictly speaking, then, what we have is not really the judgement of the historical record. Our authors have taken their hypothesis and mounted a blitzkrieg through the secondary literature, searching for examples of IR theory at work in the wild. They are reporting back that they have found them, and yet it is impossible to escape the sense that the world that they are describing is uncannily like, but not actually, ours. It has been cleared of nuance and complexity, and we are essentially to take the authors’ word that their case is proved. This gets hard sometimes. Hitler’s decision to declare war on the United States after Pearl Harbor was ‘the product of a credible theory for defeating Germany’s great power rivals’? The authors do concede that they have discovered at least four cases of strategic irrationality, including the Bush administration’s decision to invade Iraq in 2003. Setting aside the merits, or lack there of, an analysis which rates Hitler’s invasion of the Soviet Union or declaration of war on the United States as more rational than the invasion of Iraq makes one inclined to tap the authors’ instrument panel to see if a needle is stuck.

The unpersuasive quality of the argumentation is unfortunate, because at least one negative contention of Mearsheimer and Rosato here is very important: liberals do in fact have a tendency to understand the decisions of authoritarians and totalitarians as, essentially, irrational errors. Realism, which in its modern form exists always and everywhere as a corrective to liberalism, traditionally plays an important role in pushing back on that serious and systemic misunderstanding.

But the book’s rhetorical style is not its biggest problem. In addition to their negative case that claims of irrational behavior in international politics are overly subscribed, their positive case that states mostly behave rationally is meant to apply not only to means or ‘strategic’ rationality, but also to ends, to ‘goal’ rationality. Here, they argue, the record is clear: ‘Throughout history, states have almost invariably exhibited goal rationality.’ What this means is that they logically rank their priorities, and what that means is that they ‘almost invariably’ place survival first. ‘It is a matter of incontrovertible logic and evidence that survival is a prerequisite for pursuing any other goals a state might have. Other goals can be ranked in whatever order a state chooses since credible theories can be constructed to justify any ranking.’

One hates to quibble with both ‘incontrovertible logic’ and ‘evidence’ at the same time, but it seems necessary to point out that there is an essential tension here between ‘the state’ and groups within it, or between a state and a regime, and that the tensions between the goals and survival imperatives of the two may somewhat complicate the picture being drawn here. Indeed, the absence of domestic politics from the book’s account – symptomatic of the authors’ whole academic project – is a major contributor to the eerily sterile picture of history presented throughout.

Indeed, the one case in which they are prepared to concede state ‘goal nonrationality’ – German behaviour at the end of the Second World War – goes to the heart of the problem. Our authors claim to be perplexed by the fact that ‘Faced with the prospect of certain defeat, the Third Reich continued to fight rather than surrender, ensuring its own destruction.’ Setting aside for a moment the fact that such a counterexample exists seems like a rather large problem for the broader case (if after every ninety-nine apples, Newton saw the hundredth fly up into space, he might have had to rethink gravity) the reason Hitler behaved this way is not in fact so mysterious. His Reich was committed to a revolutionary vision, which was also a hegemonic vision –hegemony pursued *not* only for security but for deeper and darker reasons. Once it was unachievable the Reich couldn’t last anyway – a nonhegemonic Nazi Reich, answering indefinitely to the world’s inferior races was, ipso facto, no longer the same regime – so the Reich’s suicide and its murder of the German state are actually rational, or at least understandable, though not in terms Mearsheimer and Rosato are prepared to accept.

The authors can’t accept those terms because, as their use of survival as the be-and-end-all of human behaviour shows, they are simply liberals themselves, just of an older, more foundational variety. Thomas Hobbes, whom they cite in their prioritisation of survival among state goals, is sometimes thought of as a founder of modern authoritarianism, given his uncompromising vision that a powerful sovereign must exist to alleviate the violence and misery of the state of nature. But he is also a proto-liberal, a kind of early founder of liberalism, as the man who lowered the sights of politics from our highest aspirations – about which we men disagreed and took to violence over – to our lowest common denominator, namely, the fear of violent death. The varieties of liberalism and progressivism that succeeded him are all rooted in or responding to that refounding of politics, and so modern IR Realism, which looks to Hobbes as an inspiration as it seeks to correct the idealism and self-satisfaction of modern liberals, is in fact correcting a tradition from within that tradition. Because modern realism accepts modern liberalism’s deepest premises, it cannot be a true friend of liberalism. Both visions make the same errors. Both struggle to account for the role played by regimes in international politics, and both struggle to conceptualise ideologies not driven at core by fear of violent death. And so, while modern realism’s corrections can be helpful at times, they are only helpful to the extent that its insight about fear is true – which it is much of the time. But it is useless in the face of true revolutionaries, those who will to power not because they are afraid, but because they have a vision for which they are prepared to die.

Given that liberalism stands astride the world order, it seems that managing and suppressing the possibility of revolution should be of the utmost importance to its leaders and its friendly critics – and in this life-or-death matter, as Mearsheimer and Rosato inadvertently show, realism can be of little help. It is just as confused. What perhaps would be more helpful is scholarship of international politics that rejects liberalism’s self-satisfaction and academic realism’s reductiveness, while retaining what is useful of both, and supplying what both miss. A richer anthropology, involving some sense of how regimes matter, and how they form human intent, is indispensable. For such a project, literature, history (military, diplomatic, financial), and political philosophy are better starting points than international relations theory.