Randall Schweller

Contrary to what Schweller suggests, my book is not simply an attack on realism, nor is it an attempt to replace realism with some other doctrine. If it is an attack on anything, it is an attack on something that might be called tribalism among writers about international politics, that is, the never-ending unproductive conflict among the supporters of competing ‘paradigms’. The thesis of the first chapter of my book is that this conflict is sustained by a common willingness to accept invalid arguments. The fact that a senior scholar in the field such as Schweller can refer to a careful reconstruction and evaluation of some of the core arguments that have been advanced by supporters of all these schools of thought as a ‘diatribe’ against realism that is ‘relentlessly pedantic and sometimes strident’ illustrates the problem, which is one of the main themes of the book.Footnote 2

Realism plays a central role in the book because the theses advanced by writers who identify themselves as realists have been the focus of conflicts among these warring tribes, and liberalism and constructivism are largely defined by their opposition to realism. However, I argue that members of those camps have misidentified the problems with realism, and opposed it with equally invalid arguments. None of these doctrines could have attracted as many supporters as each of them has done if it were entirely wrong, and therefore it is not surprising that a careful evaluation of them would actually provide better reasons for believing one or another of their theses than the members of these warring tribes have done. Schweller chides me for just restating familiar realist ideas. Other readers have said I am actually a liberal, or a constructivist. I am in fact none of the above, nor am I interested in forming yet another tribe. My goal is just to identify the right answers to the questions that the members of these tribes have been debating so unproductively. Some parties in these disputes have claimed that these are empirical questions, and the right answer can be found only by confronting the competing claims with the data. But that is not possible if no empirical predictions can actually be derived from these competing schools of thought. Thus, although not a substitute for empirical research, valid reasoning is a necessary precondition for it.

One of the core claims advanced by people commonly identified as realists is that anarchy, or the absence of a world government, is the most fundamental explanation for the occurrence of war, and therefore any attempt to eliminate war by means short of the creation of a global state is utopian. The connection between anarchy and war is said to be something called the security dilemma. As Schweller says, ‘The security dilemma is a structural theory that purports to explain why wars can happen even among states that seek nothing more than their own security’. This thesis was most explicitly articulated by John Herz, but it is based on ideas found in the writings of Thomas Hobbes and Jean-Jacques Rousseau.

I show that no valid argument has ever been given showing that premises that describe the properties of anarchy and the security dilemma imply anything about the occurrence of war. James Fearon and Robert Powell have shown, of course, that what Fearon labeled as the commitment problem can explain ‘why wars can happen even among states that seek nothing more than their own security’. But the commitment problem is not what Herz defined as the security dilemma. Moreover, it is neither a necessary consequence of anarchy, nor restricted to relations among independent states. In fact, some realists have recently unwittingly discovered that anarchy can be a solution to the commitment problem, by noticing that the partition of warring groups into independent states might make possible self-enforcing agreements between or among them, while acceptance of a common government would not. This means that the entire basis for the structural realist distinction between anarchy and what Waltz misleadingly called ‘hierarchy’ is an illusion founded on an invalid argument, and realists do not actually have an argument to show that the factors emphasized by liberals and constructivists could not lead to a peaceful global order consisting of independent states. But neither have liberals and constructivists have offered valid arguments for the contrary view. My claim is that a necessary condition for progress toward answering this question is that we abandon Waltz’s question-begging distinction between anarchy and hierarchy, and study the problems of domestic and international order together.

An issue that Schweller belabors at some length is whether the Prisoner’s Dilemma game could provide the missing link between anarchy and the security dilemma, on the one hand, and war on the other. The book explains why it cannot, and also why neither offensive advantages nor incentives to attack first necessarily imply the existence of a commitment problem.Footnote 3

Michael Williams

I am grateful for Williams’ (Reference Williams2010) support for much of what I wrote, as well as his sympathetic criticisms. However, I do not think that he has made a compelling case for his criticisms.

The issue has to do with the relation between the ideas developed by European writers in the reason of state tradition and modern ideas about reason and rationality (which is the issue that has been raised by writers about international politics who call themselves constructivists). Contemporary constructivists charge both realism and liberalism with being ‘utilitarian’ or ‘rationalist’. As many of the ideas of realists and liberals can be traced back to Thomas Hobbes and the writers he influenced, this charge exemplifies the treatment of Hobbes by both Meinecke and Talcott Parsons, who regarded Hobbes as a utilitarian and argued that utilitarianism could never explain how societies were held together (Wagner, Reference Wagner2007: 70–71).

Unlike these other writers, however, Williams claims to have found ideas in Hobbes’s writings that are neither utilitarian nor rationalist, which Williams believes Hobbes thought could help one understand how human beings could become organized in the state of nature. Williams claims that Hobbes thought that ‘it was belief and not reason that founded societies, and that provided the basis for collective coercion which then became a part of the equation’. Thus, we need ‘not only a theory of coercive logics, but also one of symbolic logics. The challenge is that the latter may not … conform to the same logic as the one based on material fear and coercion …’.

Thus, Williams seems to believe that I think political order is based on rational coercion, and his criticism is that this view cannot be supported. There are two issues here: one is the role of coercion as the basis for political order and the other is the role of rationality in constructing a theory of it.

One of the main themes of my book is that Hobbes’s image of the sovereign and his sword as the foundation of political order is a false clue in thinking about the distinction between politics within states and politics among them. Governments do coerce people, of course, but the coercive power of governments is the result of how people are organized (or not), and must therefore be supported by a complex set of equilibrium expectations. Thus, political order is based not on fear, but on what everyone expects everyone else to do (or not do). When these expectations shift, then the governments that once were powerful can ‘wither away’ (to coin a phrase), as happened spectacularly to the Soviet Union.

Thus, the fact that a peaceful equilibrium within states is based partly on expectations that support a stable government does not imply that there could not be a set of expectations about how independent governments will deal with each other that would form the basis of a peaceful equilibrium at the global level. This seems to me to be a good way of understanding Kant’s critique of Rousseau’s pessimistic view of international politics.

However, equilibrium expectations are beliefs, and thus it is not clear why there should be a conflict between belief and reason as the basis of social order. Moreover, it is clear that human beings are willing in some circumstances to sacrifice themselves for others, and that the discovery by leaders of violent organizations of how to take advantage of that willingness is an important part of the explanation of how large-scale violence becomes possible. This is not inconsistent with what I wrote, and its importance is in fact mentioned in the book. Unfortunately, discussions of such issues are hampered by the fact that the words ‘reason’, ‘rational’, and ‘rationality’ are so ambiguous as to be practically meaningless. A failure to recognize this fact is the basis for unproductive debates among students of international politics about something called ‘rationalism’. I regret not saying more about this latest ‘ism’ in my book. For an attempt to make up for this omission, see Wagner (Reference Wagner2010).Footnote 4

Tarak Barkawi

Like Williams, Barkawi (2010) claims that I have left something important out of my analysis, but his list is both longer and less well defined. He says that my analysis is Eurocentric, and stuck in the eighteenth century. But the reason for my emphasis on eighteenth century ideas is that the debate between realism and liberalism has not advanced very far beyond the ideas originally stated by Hobbes, Rousseau, and Kant, nor has constructivism had much to say about them. My goal was to go back to the origin of those ideas in order to advance beyond them. He claims that I understate the role of European imperialism in the formation of the modern state system. If so, this is a consequence of both the state of the existing literature, and my own unfortunately limited knowledge of it. However, the aim of my book was not, as Barkawi suggests, to offer ‘a theory of international politics’, but to define a subject that is broader than the ones defined by any of the existing political science subfields, and invite students of international politics to investigate it. It seems clear to me that imperialisms of all sorts must be an integral part of that subject, and I do not see how anything I wrote could be interpreted otherwise. Finally, he says, ‘in turning war into “violent bargaining”’, I reduce ‘war to a species of peacetime politics’, and ignore both the uncertainty it entails, and its effects on the development of the organizations that participate in it. I do not see how anyone who has actually read the entirety of my book could make such a statement.

James Fearon

I was relieved that Fearon (Reference Fearon2010) did not find any major mistakes in the book, though he doubts whether some of the things I said in it could withstand careful criticism. But his biggest criticism is that ‘while the argument in the first chapter of the book is clear and important …, it is not so clear what the core question is in the subsequent chapters, or what Wagner’s answer is’. Fearon’s interpretation of the core question is ‘Does anarchy, in the sense of no world government, imply that interstate relations must be…plagued by interstate war?’ He interprets my answer to be, ‘probably not’, but says that it is not clear why I think that. He identifies two lines of argument that he thinks might bear on the question. One is my suggestion that ‘the absence of third-party enforcement at the level of states might make war less rather than more likely…’. The other is my observation that ‘within states there is no higher power that can enforce agreements between the government and its citizens…’, so ‘some of the causes of interstate war…may also be operative within states’. But it is not clear, he says, what these lines of argument imply about the answer to the core question, and he complains that the book’s concluding chapter ‘seems to trail off with the observation that the problems…are complicated and that the valid arguments are hard to construct…’.

I have several responses to these comments. As I said in the preface to the book, my primary aim was, like Kenneth Waltz’s in his book, Theory of International Politics, to define a subject, and not to provide a definitive answer to all the questions one might ask about it. As Fearon points out, Waltz’s goal was to ‘stake out the autonomy of the field of International Relations’. I think this goal was misguided. In trying to accomplish that goal, Waltz introduced a distinction between anarchy and hierarchy, but never defined what he meant by those terms. They are not self-defining. A hierarch is a priest, and a hierarchy is a system of priestly rule. It can also be any ranked ordering of things. The organization of a business firm might be called a hierarchy. David Lake (Reference Lake2009) has defined it to mean an authority relationship, and has recently claimed to find hierarchies among independent states . ‘Anarchy’ can refer to the absence of a ruler, or chaos and disorder. It is often interpreted to mean just the absence of a world government. But for most of their time on earth human beings lived without governments of any sort, so if life without a government is ‘solitary, poor, nasty, brutish, and short’, as Hobbes put it, it is hard to see how humans survived.

One thing that should be clear from my book is that the difference between conflicts within states and conflicts among them cannot be the availability of a third party to enforce agreements. Governments do not make peace by enforcing agreements among rival mafias, thereby enabling them to limit their conflicts. Rather a government is a mafia without rivals, or, as Weber put it, it has a monopoly of the legitimate (or organized) use of force. Hobbes’s state of nature, by contrast, is most plausibly interpreted as a situation in which competing mafias (or protection organizations) vie for control over the same population. We cannot determine whether relations among independent governments could be reliably peaceful if we confuse relations among governments with the state of nature interpreted in this way.

The key to establishing reliable peace is to find a way by which people who might fight can commit themselves not to. One way to do that is for warring parties to disarm and submit themselves to an organization with a monopoly of the organized use of force. Because of the collective action problem they would face in reversing that decision, that is at least a weak commitment (which is why the contestants in civil wars are often reluctant to make it). However, the collective action problem can be overcome, and therefore a government may confront a lot of organized violence if it cannot commit itself not to use its monopoly in such a manner that some potential group would have an incentive to resist. In the right circumstances, liberal democracy has proved to be a way of solving this double commitment problem.

Kant seems to have thought that the same institutional arrangements can enable independent governments to commit themselves not to fight each other, but no one has offered a non-question-begging argument to show why that might be true. The answer I give at the beginning of chapter 6 is that it could be true if the liberal democratic contract was an ‘all for one and one for all’ commitment by a population to mobilize all available resources to defend everyone within a well-defined territory (sometimes called civic nationalism), but not to provide resources for territorial expansion. Then if such territorial commitments were consistent, democracy could allow the populations of democratic states to prevent their own governments from violating territorial agreements, and deter others from doing so. Such consistent territorial commitments could be supported by agreements facilitating interstate commerce, which provides a means of sharing the benefits of natural resources without paying the costs of directly controlling them in the face of local resistance.

When territorial commitments are inconsistent, however, as they are in China and Taiwan, in India and Pakistan, and in what used to be Palestine, then they can make compromise difficult and lead to intractable conflicts. Thus, the combination of democracy and ethnic nationalism may make a stable peace impossible.

Although some contemporary states might be defined by such liberal democratic domestic contracts, most are not, and much of the globe is not governed by effective states at all. To ignore this fact would be to engage in what E.H. Carr called utopian thinking. The rest of chapter 6, therefore, discusses the problems that arise when territorially satisfied democratic states confront territories that are not states, or states that are not defined by liberal democratic domestic contracts characterized by civic nationalism.

Thus, I say more about what Fearon regards as the core question posed by the book than he seems to have noticed. Nonetheless, I am under no illusion that I have done more than sketch out what might be the basis for a productive investigation of this question, and thus the final chapter just contains a restatement of what I think the subject of study should be and how it should be conducted.