Revised to reflect speech as given
Hobbes’s Dilemma and the Liberal Quest for World Order
Talk by Professor Robert O. Keohane (Princeton University; Shimer BA 1961) for Shimer College, at the Standard Club, Chicago, Illinois February 21, 2013.
Tonight I am going to talk about multilateral institutions, such as the World Trade Organization, the World Health Organization, the various institutions designed to cope with climate change, the European Union, or the United Nations. But I am not going to discuss specific institutions in detail. Like a good Shimerian, I’m going to start with the Great Books. They do not provide answers for the 21st century, but they ask some of the right questions, in very profound ways.
Many of you have read parts of Hobbes’s Leviathan in Soc 2. Hobbes is usually read in a domestic context, as arguing that anarchy is so horrible that obedience to authoritarian government is justified as long as that government protects its subjects’ lives. At the domestic level, what I call Hobbes’s Dilemma is the dilemma facing people who would prefer to be autonomous, but who face great uncertainty and a high probability of sudden death unless they submit to a sovereign whose actions they cannot control.
One side of the dilemma is that there is no security without government. The reason is that people are self-interested, they seek power and glory, and they are fearful of one another. Hobbes was certainly right about this: modern anthropology has demonstrated that stateless societies have much higher rates of death from warfare than do societies with states.
Rousseau disagreed with Hobbes’s solution to his dilemma and took a radical approach: the ideal community was very small and self-sufficient. But Rousseau’s approach is infeasible in a globalized society, although it might have worked for the Switzerland of the 18th and 19th centuries. Remember Adam Smith, since The Wealth of Nations is an underappreciated and usually misunderstood book. Smith, writing at about the same time as Rousseau but not responding directly to him, said that the division of labor is at the heart of prosperity and “the division of labor depends on the extent of the market.” In other words, Rousseau’s “solution” would condemn everyone in the world to poverty, even if it could succeed politically. Without both domestic and international exchange, there would be less division of labor and therefore less prosperity, worldwide.
The other side of the dilemma can be expressed as follows. Because people are self-interested and power-loving, unlimited power for the ruler implies a predatory, oppressive state. Hobbes was willing to accept this side of the dilemma, but most of us today are not.
Hobbes’s Dilemma is highly contemporary: it has faced people recently in Bosnia, Libya, and Mali. Should one submit to an authoritarian government that, by and large, offers protection to its citizens in return for submission, but is likely to exploit and oppress its citizens; or should one nevertheless seek autonomy even at great risk to oneself? Hobbes leaves little doubt of his choice: one must submit, to avoid “the war of all against all.” Otherwise, “the life of man is solitary, nasty, brutish, and short.”
To anyone who has been exposed to political liberalism, Hobbes’s solution to his dilemma also seems unsatisfactory. But liberals, unlike Rousseau, have a viable answer. In Soc 2 Shimer students also read Locke, Montesquieu, and the American founders, who argued that there is another option – neither submission to authoritarian rule or retreat into a largely mythical small democracy. The solution is constitutionalism and the division of powers, in which “power checks power.” In Federalist # 10, James Madison – the great American political theorist – argues that large republics with representative government can avoid Hobbes’s Dilemma:
“Extend the sphere, and you will make it less probable that a majority of the whole will have a common motive to invade the rights of other citizens; and if such a common motive exists, it will be more difficult for those who feel it to discover their own strength, and to act in unison with each other.”
Long experience, however, has shown that it is difficult to establish such a republic: the people must be willing to submit to constitutional restraints and they must not be divided into permanent factions on the basis, for instance, of religion. Instead, there must be pluralism, with different groups of people working together on different issues. And there needs to be a measure of trust on the part of each group that others will also abide by the constitutional restraints when they are in power. As Samuel P. Huntington showed 20 years ago, it is rare for new democracies to get this right the first time around; more frequently, there is at least one lapse into populism or authoritarianism before stable democracy emerges. Naively, the American public and American leaders often think otherwise – and are therefore surprised by reversions to authoritarianism – whether in some of our Asian allies in the 1960s and 1970s, Latin America especially in the 1970s, or the countries of the “Arab Spring” today.
This is the domestic story of how to respond to Hobbes’s Dilemma: the story of constitutional liberalism, its promise, and the difficulty of fulfilling that promise. But my focus tonight is on world politics.
When we turn from domestic to world politics, the first horn of Hobbes’s Dilemma remains the same: there is no security without government. International relations is prone to conflict and war as a result of human self-interest, pervasive uncertainty, and the resultant desire for “power after power,” generating destruction and death. The “war of all against all” is tempered by the protection offered by the sovereign state, but war is always a possibility since there is nothing to prevent it. And even though everyone knows that war is costly, it can still occur because of lack of full information about others’ intentions and behavior, or because of the inherent difficulties of making credible commitments in world politics.
The International Dimension: the Liberalism of Progress?
The situation gets worse when we take into account authoritarian rule: rule by what Keynes called “madmen in authority.” These madmen may lead political movements that seek regional or even world domination, or groups motivated by extremist and exclusivist religions. Or they may be isolated leaders of countries with nuclear weapons, or charismatic but highly aggressive and destructive leaders such as Adolph Hitler Think about Bolshevik Russia, al Qaeda, Iran and North Korea, or Hitler’s Germany. The sources of threat are diverse but genuinely frightening. War is an ever-present possibility, only mitigated by the protection that what has been called the “hard shell” of the nation-state. Hobbes took comfort from this hard shell; but unlike in Hobbes’s time, weapons can now penetrate well inside nation-states.
This discussion brings us to liberals such as John Locke, who saw constitutionalism as helping not only to constrain would-be autocrats but also ordinary people who might, instead of staying on the right path, fall into ditches or get caught in hedgerows. Such constraining institutions would not be necessary if, as Rousseau thought, human beings were fundamentally benign creatures. We would only have to liberate our true selves. I have provided some reasons why I do not share such a secular faith.
The Liberalism of Fear
The milder but still unattractive version of this existential nightmare is a world dominated by defensive autocrats seeking to hold onto power despite frequent popular unrest. Such protective sovereignty involves a number of authoritarian states, defensive in orientation but armed to the teeth against potential aggressors. Think of the Soviet Union and China in the late 1960s – they fought a border skirmish that could be considered a war. This “solution” also fails to provide decent conditions of life. War remains an ever-present possibility, either from miscalculation or because some of these defensive authoritarian states turn aggressive. And even if there is peace among such autocracies, the resulting barriers to exchanges of goods or people block the realization of prosperity. Smith’s argument that prosperity requires a division of labor, which requires a large market, makes this option look pretty bleak itself.
But at the global level the second horn of Hobbes’s Dilemma vanishes. It cannot be predatory authoritarian global rule because there has never been a world state and there is no prospect of one. In other words, on the classic, so-called “Realist” account there is no dilemma, simply an existential situation, what some international relations theorists have referred to as “anarchy.” So we shift from a concern with Hobbes’s Dilemma to what I will call Hobbes’ Existentialism.
One possible way to avoid Hobbes’s Existentialism at the international level is to deny his view of human nature and to put one’s faith in what could be called the “liberalism of progress.” In this view, human nature either is benign or at least subject to radical improvement. Much American liberalism has taken this view. If everyone only sought “life, liberty, and the pursuit of happiness,” and did not define those goals in opposition to the goals of others, the liberalism of progress would be valid. The next step internationally would be to argue that democracy is necessarily peaceful – and some Americans, led by Woodrow Wilson, who sought “a world safe for democracy,” have taken that line. Inconveniently for this thesis, however, the United States is one of the most war-prone states of the last century. Between 1913 and 2013 the United States has been involved in two world wars and two other major wars – in Korea and Vietnam – that cost hundreds of thousands, probably millions, of lives. It has invaded Iraq twice and several countries in the Caribbean. The United States also engaged in military intervention in the Soviet Union in 1919, was heavily involved in the Chinese civil war, and is still involved militarily in Afghanistan. This list does not include covert interventions involving the use of force in Chile, Iran, Cambodia, Laos, Nicaragua, and a variety of other countries.
I am not arguing that all of these interventions were wrong – far from it. I believe that some were justified and others –including the war in Vietnam and the 2003 invasion of Iraq – were not. My point here is simply that the unqualified view that democracies are necessarily peaceful is mistaken.
More plausibly, it has been argued that if democracy were universal, peace would follow, since democracies do not fight each other. This generalization has some merit, although one has to define “democracy” in a special and somewhat peculiar way to make it work. Germany had a working parliament and fair elections in 1914, along with a free press; Pakistan and India have fought three times, which is hard to attribute simply to Pakistan being under a military government, since they have come close to war when Pakistan was democratic; Finland, a democracy, was allied with Germany in World War II; and Serbia had a democratic form of government, as did Bosnia, when they fought in the 1990s. Most US interventions have indeed been against authoritarian regimes or movements, although the regimes that it attacked in Chile and Nicaragua were arguably more democratic than their opponents.
Furthermore, democracy can turn bad, leading to coups and support for authoritarian government. Think of the French Revolution, leading to Napoleon; the Weimar Republic leading to the election of Hitler in a free election in 1933; or the widespread middle class support for military coups in Brazil, Chile, and Argentina in the 1960s and 1970s. I think that there has been progress in human history, but not because human nature is benign or improving. We can hope for the best but we need to take specific institutional action to make our hopes realistic.
In other words, we need to temper the Liberalism of Progress with what the political theorist, Judith Shklar, called “The Liberalism of Fear.” Shklar was a refugee from both the Nazis and the Soviets, so she knew what fear means. And as she said, “no liberal ever forgets that governments are coercive.” The Liberalism of Fear is horrified by the atrocities of Rwanda, Bosnia, Libya, and Syria but unlike the Liberalism of Progress its view of human nature is not challenged by these terrible events. Nor do they shake its liberalism – its belief in limitations on the powerful – which was forged in the searing recognition that humans can act in terrible ways. So let me discuss the Liberalism of Fear.
Belief in the possibility of progress and retention of active fear of autocracy and misgovernment are two sides of the same liberal coin. From both perspectives, we need to understand how otherwise unattractive human passions can nevertheless promote the general good. In Smith’s worldview, progress is generated by a desire for comfort, even luxury, and by greed. In Locke’s view, fear can be reduced by constitutional constraints. So the combination of a belief in the possibility of progress and continual fear of both autocracy and war provides the basis for a realistic but forward-looking liberalism. James Madison is the American father of such a realistic liberalism, but it has deep roots both in English utilitarianism, going back to Hobbes, and in French thought.
Neither Madison nor Smith indulged in the more utopian dreams of the believers in progress. Even though potential gains from trade, combined with advancing technology, make it possible for all economies to prosper simultaneously, the Hobbesian desire for “power after power" gets in the way. So does greed. People often seek to gain distributional advantages not by being more productive but by gaining control of public policies in order to capture rents. Nevertheless, economic protectionism has been proved to be bankrupt, and the institutions of liberal democracy have limited, although they have not eliminated, the success of rent-seeking. Smith and Madison would not be fully satisfied, but they would be gratified by the partially successful institutionalization of their ideas.
Progress, it seems to me, is due principally to the development of institutions, including the state and multilateral organizations. Indeed, the liberalism of progress and the liberalism of fear both emphasize the need for institutions. Smith's liberalism calls for institutions to promote exchange; Shklar's for institutions to control human vices and those individuals among us whose vices are most dangerous to others. For these institutions to be morally acceptable, they must rest both on humane beliefs and substantial mutual trust. The Mafia is not better than anarchy; the people who live under either find themselves impaled on one horn or the other of Hobbes’s Dilemma. And those people condemned to an existential reality of war and destruction are no better-off.
In domestic society, “the people,” defined in some way, come together to create the “benign constraints” that limit both themselves and their rulers. The Constitutional Convention of 1787 is a classic example, also studied in Soc 2. There needs to be what Benedict Anderson called an “imagined community” or what Karl Deutsch referred to as “we-feeling” among the people who come together. When such beliefs prevail in a geographical area, and the people have the capacity – for instance, through successful revolution – to create a constitutional state, they can enforce their rules on themselves. At the domestic level, this can solve Hobbes’ Dilemma. But as I have emphasized, the international situation is different, since there are no realistic prospects of a universal state, democratic or not.
So we need to ask: How can institutions help avert the existential nightmare of a global system characterized by violence and conflict.?
We have seen that institutions can provide an escape from Hobbes’ Dilemma, at the domestic level. Do they also help us escape from the existential reality of war, and threatened war, at the international level? If there are no powerful aggressive states bent on war and conquest, my answer is yes. But the wary and self-intersted states that I have discussed have to do so differently – not through a social contract among members of a coherent public, but through the development of institutions as a result of repeated interaction among separate social groups, typically organized as states.
Uncertainty, Repetition and Reputation
In game theory, there is a famous game called “Prisoners’ Dilemma.” Two alleged burglars are caught and put in separate rooms. Unlike the situation in American law, they are not informed of their right to be silent. Each is told that if he confesses, and testifies against the other one, who refuses to confess, he will get a light sentence; but if does not confess but is convicted through the testimony of his partner, the judge will “throw the book at him.” Both prisoners know that the police do not have sufficient evidence to convict them; but they also cannot communicate.
The solution of this game is that both prisoners confess and they both receive substantial jail terms. That is, they do not cooperate with each other since they fear the worst outcome if they attempt to do this unilaterally. It is a paradox: they get a worse outcome than they could have received if both had refused to talk – that is, if they had cooperated. But they are driven to behave in this way by the combination of self-interest and inability to make credible promises to each other. Both prisoners are better off confessing, in a single-play situation, no matter what their partner does. Once they figure this out – and their interrogators are bound to inform them – they will confess. That is, they will fail to cooperate with each other.
You can think of this situation as analogous to two mutually suspicious but not fundamentally aggressive states facing one another when each side sees large advantages to striking first and uncertainty about the other side’s intentions is high. Both states wish to avoid war, but the worst outcome for each of them is to be attacked by the other. Both have incentives to attack.
What changes this suboptimal outcome?
First, if the two parties can credibly communicate their benign intentions, they can achieve a better result. If the two prisoners can credibly promise not to “rat” on the other one, they will be more likely to be released or acquitted. It is hard, however, to ensure credibility when the temptation to defect is so great. Indeed, one of the functions of the Mafia or similar organizations for criminals is that it makes such promises credible – the “stool-pigeon” does not survive long. At the international level, guarantees by great powers or formal treaty pledges to international organizations composed of great powers can perform the same function: it now becomes irrational to renege.
The second way to change the outcome is to have repeated play over a long period of time. Suppose we play Prisoners’ Dilemma three times a day for the rest of our lives. In each play of the game, someone can win $10 by reneging on a promise – so his partner gets nothing – instead of cooperating and winning $6. But if the other player plays “tit for tat,” he will renege on the next move and will keep doing so as long as the first player keeps reneging. $10 plus zero indefinitely is a lot less than $6 for a very long series of plays. This is to say, in repeated games, as in world politics, reputation matters. Indeed, one thing that multilateral institutions do is to ensure repeated play. An obvious example is the World Trade Organization. But look also at the Security Council. Issues involving Korea, Suez, the Congo, Cuba, Iraq, Libya, Iran and Syria have arisen. On each issue some great power wants action; others may be reluctant. Cooperation is promoted by the knowledge on the part of reluctant participants that they may want help from the Security Council on some unknown future issue – and by their desire to avoid a reputation for obstructing peace.
Now we can see why institutions are so important: they can both increase the credibility of commitments and make reputation matter more. More generally, they reduce the sort of uncertainty that tempts states to engage in unnecessary conflict. Think about arms control treaties – reducing fear of a first strike by an enemy – or trade organizations, reducing fear of sudden protectionist measures.
Institutions also reduce uncertainty, and increase credibility, by lowering the costs of negotiating agreements. Consider the Great Depression of the 1930s and the Great Recession of 2008-9 – which was much milder and shorter. Much was similar about the period before the crash: bubbles in stocks and real estate, over-optimism by banks that became over-committed; a lack of effective regulation by governments. But the period after the crash – of 1929 and 2008 – looked very different. One difference is that policy-makers had better ideas about how to react and made fewer errors in more recent crises. But an equally important difference was that institutions for cooperation were almost absent in 1929 but well-developed in 2008. The countries in the G-8 and G-20 met quickly after the Lehman Brothers crash and acted together to expand their money supplies and enlarge their fiscal deficits, pumping money into their banks and into their economies in general. The chain-reaction collapses that made the recession of 1929-30 into the Great Depression were avoided in significant part because institutions were available to facilitate cooperation.
At the global as well as national level, we need institutions to make mutually beneficial cooperation possible – to prevent both war and depression. Such cooperation does not assume benign human nature and it does not involve altruism. It arises from discord, not from harmony: that is, it arises from conflicts of interests. But if institutions structure incentives properly, they provide incentives for independent, self-interested states to cooperate with one another and a setting that provides relevant information, enhances credibility, and that, through repeated play and reputation, gives them the ability to do so.
Accountability and Democratic Governance
Democratic theory is even more demanding than constitutionalism. From a democratic standpoint it is not enough to have non-oppressive institutions that enforce rules. Accountability, participation, and persuasion are also essential. International institutions will probably never meet the standards of electoral accountability and participation that we expect of domestic democracies (Dahl 1999), so at best they will be low on a democratic scale. It is unfair to demand too much of them. But in the liberal democratic tradition that I embrace, voluntary cooperation based on institutions of accountability provides the strongest guarantee of a legitimate process. In this final part of my talk, I will focus briefly on accountability. The key question is: how can we go beyond a merely protective set of institutions that reduce the likelihood of oppression and war, to a set of institutions that fulfil wider human potential? So we now move briefly from an analytical perspective to a more normative and speculative one.Accountability is not necessarily electoral, so it is essential to explore other forms of it if we are to increase accountability in global governance. The world that I imagine would not be governed by a representative electoral democracy. States will remain important; and one state/one vote is not a democratic principle. National identities are unlikely to dissolve into the sense of a larger community that is necessary for democracy to thrive.
Accountability, however, can be indirectly linked to elections without a global representative democracy by strengthening mechanisms of domestic accountability of governments to their publics. Non-electoral dimensions of accountability also exist. The World Trade Organization holds governments accountable for their trade-related measures through a well-structured legal system. More ambitiously, global governance, combined with modern communications technology, can begin to generate a public space in which some people communicate with one another about public policy without regard to distance. Criticism, heard and responded to in a public space, can help generate accountability. Professional standards and markets constitute other forms of non-electoral accountability.
These mechanisms of accountability exist, in fragmented ways, at the global level, but they are disarticulated. They do not come together in a clear pathway by which laws are enacted and implemented. Chains of delegation are long, and some of their links are hidden behind a veil of secrecy. Incentives for politicians to hold leaders of other governments accountable are lacking. Public, professional groups, and advocacy networks can only punish leaders inconsistently. Governments, nongovernmental organizations, and firms that do not rely on brand names may be immune from market-based sanctions. In devising acceptable institutions for global governance, accountability should be built into the mechanisms of rule making and rule implementation.
I have been speaking tonight with three perspectives in mind: from political philosophy, political science, and policy analysis. Shimer is strongest on political philosophy, but political philosophy by itself is very abstract and not empirically focused. I believe in Shimer because I think that an education in the classics – emphasizing philosophy, political and otherwise – can provide a basis for further work in science and policy analysis. But such further work is equally important. We start with the Great Books but we cannot stop with them. And the greatest social scientists all lived at least in part in the 20th century. Science involves devising hypotheses and testing them rigorously: this can be done in social science as well as in the natural sciences, although in a different way. Policy analysis involves applying the lessons from science, in the light of the ethical insights of political philosophy and by exercising good judgment, to contemporary problems.
As students of political philosophy, a major objective of students of world politics should be to understand how to overcome the existential nightmare at the global level suggested by Hobbes’s Dilemma. When is governance necessary and what principles would make it legitimate? As positive political scientists, we need to continue to analyze the conditions under which different forms and levels of governance are feasible. As practitioners of a policy science, we need to offer advice about how institutions for global governance should be constituted. This advice must be realistic, not romantic. We must begin with real people, not some mythological beings of higher moral capability. But we need also to recognize, and seek to expand, the scope for reflection and the normative principles that reflective individuals may espouse. We should seek to design accountable institutions that reduce uncertainty and create incentives for cooperation.
The sakes in the mission I propose are high for the world. If global institutions are designed well, they will promote human welfare. But if we bungle the job, the results could be disastrous. Either oppression or ineptitude would likely lead to conflict and a renewed fragmentation of global politics. Effective and humane global governance arrangements are not inevitable. They will depend on human effort and on deep thinking about politics.
This deep thinking begins in the Academy. At Shimer it begins, as my talk did, with the classics – the Great Books. But it moves to the great works of contemporary times as well. I mentioned Hobbes, Rousseau, Locke, Madison and Smith in my talk. But there were sentences that relied on major 20th Century thinkers. The Great Books are the beginning of wisdom – not the end of it.