Today I want to introduce three ideas. First, our discipline has a mandate to help us human beings govern ourselves. Second, we should be focusing, more than we do now, on creating legitimate coercion. Third, we have neglected an important source of legitimate coercion: negotiation to agreement.
This analysis is both descriptive and aspirational. It is analogous to a lawyer’s brief before the Supreme Court (you, my colleagues in political science, are the court), pointing out a thread in its past decisions and arguing that this thread should serve as a guide to the future.
The Mandate
Having posed the question, “What is political science for?” I propose the answer: “To help us govern ourselves.”
It is true, of course, that political science is not always “for” anything. Political scientists often just pursue explanation and understanding without expecting the quest to serve any larger goal. It is marvelous to take satisfaction in solving—or just making progress on—a puzzle simply for the intellectual thrill of it—because, like Everest, “it is there.” I have done this, I love it, and I love seeing others do it. In the puzzles that we take on in political science, we have the added satisfaction of knowing that we are deepening human understanding—about the human condition, about the tradeoffs we face in political institutions, and about our past histories.
Yet if political science is “for” anything, I think it is, and should be, for helping us to govern ourselves. Political science is the only academic discipline specifically organized to study this question. Other disciplines—law, psychology, sociology, economics—make valuable contributions to answering it. But we, as a discipline, are organized around the issue of governing. Because we have consciously created social structures that let us think together, and because of our specialized toolkits, we—of all the people in the world—are the best organized to help do this. The world certainly needs that help.
Compared to our needs, we know very little about how to govern ourselves. We don’t know how to coerce ourselves into giving up what we need to give up in order to stop global warming. We don’t know how to stop nuclear proliferation. We don’t know how to transition from autocracy to democracy without descending into violence. Closer to home, we don’t know how to tax ourselves sufficiently to keep our infrastructure from crumbling or how to pay for the rising medical costs of an aging population. We don’t know how to produce laws in a polarized Congress or how to reduce that polarization. We don’t know how to keep ourselves from drifting into greater and greater inequality. At this moment of great need and relative ignorance, political science is the one academic discipline explicitly organized to study how we make our collective decisions on these matters, and how we can make them legitimately.
Creating Legitimate Coercion
Legitimate coercion is the fundamental problem of governance. How can large, highly interdependent structures produce sufficient legitimate coercion to solve their collective action problems? Such structures are evolving everywhere—even internationally—although in the case of nuclear proliferation and global warming we seem to be losing the race against time. On the domestic front, large-scale representative democracy has evolved only relatively recently, in the last 300 years, and we are continuing to experiment with the form. In the US, innovations in the primary system, redistricting, and Congressional rules have recently affected the way our democracy works. Elsewhere in the world nations and regions are experimenting institutionally with a realistic sense of urgency. Yet the democratic theory we have is not fully up to the job of providing normative guidance in the midst of these changes.
In its first incarnations, the democratic theory that accompanied representative democracy focused largely on preventing tyranny, not on legitimating the coercion needed to solve collective action problems. The social contract tradition arose in opposition to the rule of emperors and kings. Even Madison, Jefferson, and Hamilton, concerned though they were to establish a system of effective government, made the prevention of tyranny a central goal. What I have called the “resistance tradition,” rooted in contract theory and focused on combating too much power, has dominated much of democratic theory ever since. Footnote 1 Today, however, the challenge of creating legitimate coercion is at least as great as the challenge of resisting illegitimate coercion.
My central argument regarding coercion has three parts. First, in a world of increasing interdependence we face an almost infinite number of collective action problems that are created when something we want or need is a “free access” good. Second, solving most of those collective action problems requires coercion, which is the basic reason for government. Third, the best coercion is legitimate coercion. Important as legitimate coercion is, we do not know how to do it very well. The fundamental job of political science, I believe, is to help societies create and properly use legitimate coercion.
What do I mean by legitimate coercion? Coercion is relatively easy to define. By coercion I mean the threat of sanction (“Leave this room or I’ll shoot you”) or the use of force (carrying you out of the room kicking and screaming). Footnote 2 A sanction can be as small and informal as a dirty look or as large and formal as life imprisonment. Defining legitimacy is harder, and I will return to that.
Now, let me begin to walk through my argument.
Free-Access Goods Create Collective Action Problems
First, as most people here know, collective action problems arise whenever a “free-access good” must be produced. By a “free-access” good I mean a good that, once produced, anyone can consume even if they did not contribute to producing it. If common defense against enemies is provided, anyone can benefit even if they did not contribute. So too with law and order, or a toll-free road. Once some group of people has provided such a good, anyone can use it. Everyone is therefore tempted to free-ride, using the good without helping to create or to maintain it by contributing effort or money.
I use the term “free-access good” instead of the economists’ terms “non-excludable good” and “public good” both because free access is easier to remember and for technical reasons. Footnote 3 All that matters for this analysis that when a good has the characteristic of being free access—that is, open to all potential users—it tends to be underproduced.
A collective action exercise. I would like to ask even those of you who are deeply familiar with the collective action problem to participate with me in a brief collective action exercise. Imagine that you have $100 in your pocket. Now please take out a real piece of paper and write on it either zero dollars or one hundred dollars, for the contribution you will make to the common pot. Just zero or 100, please, nothing in between, for simplicity. I will be a doubling machine. I will double everything in the pot, with no effort on your part, and return that amount to everyone equally. That doubled money is a free-access good. If you give me zero, you get to keep your $100 and also get an equal share of whatever others put in the pot. But if everyone keeps his or her $100, you will all lose the chance to double your money without effort. This is the “common pool” version of the collective action problem. The point is that no matter what the others do, if you do not contribute you will end up with more than if you do contribute. Now please write either zero or 100 on the paper, fold it over so that no one can see what you wrote, and pass it to the aisle and then to the front of the room. Thank you.
While the counters are tallying up the results, let me assume that 80 percent of you contributed, using imaginary dollars that cost you nothing. [In the event, 76 percent of the 472 attendees contributed to the common pool, a higher percentage than in most such situations, explained by self-selection.] The percentage you give will normally decrease rapidly when the money you are asked to give is real and as the amount you are asked to give grows larger. But the precise percentage is irrelevant to my point. We can learn two things from this exercise.
First, we learn that many of you contributed $100. Assuming that you were confident of your anonymity, your motives were some mixture of solidarity and duty, two separate motives on which polities can build. Footnote 4 You may also have looked around and quickly calculated that more than half of the people in the room would contribute, so that you would at least not lose your stake—and psychologists have provided ample evidence that human beings hate losses even more than we like gains. Footnote 5 That kind of trust in the others in your community is another significant base on which polities can build.
Second, we also learn that some of you, even though the dollars were imaginary, have contributed zero. Some of you may have gone along with the assumption that the dollars were real and not wanted to risk that stake. Others may have been making a political point that you should not be asked to give up your $100 for a collective cause. Whatever the reason, at least 20 percent of you probably did not contribute. If we repeated this exercise over time your refusal to contribute would begin to undermine the contributions of others. As Margaret Levi has pointed out in developing her concept of “contingent consent,” people who think that others are paying their taxes are themselves less likely to cheat. When they think that others are free-riding, they often begin to free-ride too. Footnote 6
We can conclude both that we ought not to undermine the solidarity and conscience that lead some to contribute and that we need to find some way of getting the non-contributors to contribute—unless they had a good reason for not contributing, in which case we want the contributors to hear, understand, and accept that reason. In a small setting, contributors can use some version of the informal coercion of a dirty look to punish the non-contributors. Sometimes even large groups can handle a collective action problem through a combination of critical mass, solidarity, conscience, social sanction, and reciprocity based on reputation, for example in the strategy that game theorists call “generous Tit for Tat.” Footnote 7 But in a large and relatively anonymous setting, where many people do not know the others’ reputations, we usually need more formal coercion to solve collective action problems. Government is the set of institutions we create to develop and administer that coercion. Footnote 8
Understanding the dynamics of the collective action problem is, in my view, a logical breakthrough comparable to understanding the implications of how supply and demand interact to determine price. But the collective action problem is a bit harder to understand than supply and demand. Until the last half of the twentieth century, no human being had access to this logic. Rousseau and Hume had developed partial formulations of one version of the problem, but only in the 1950s did scholars in the three separate fields of game theory, public finance, and fisheries management begin to work out its full logic. Footnote 9 The economist William Baumol was the first to stress that the problem must be solved with what Garrett Hardin later called “mutual coercion mutually arrived at.” Footnote 10 For the next several decades the implications of the various forms of the collective action problem swept through the social sciences. Footnote 11 Yet the full implications have still not been assimilated into general knowledge. Even today I doubt that more than a small percentage of political science students understands the logic of collective action, its dependence on a free-access good, and its implications for coercion. Nor do they understand its corollary, that solving collective action problems is the most significant reason for government.
Nonetheless, collective action problems have become much more central to human life in the last hundred years. As we increase in number, free-access goods that were earlier supplied by nature (clean air, clean and sufficient water, fish in the sea) require more and more human action to maintain or produce them. As human beings also produce more complex goods and develop more refined demands (like blueberries in the winter), we become more and more interdependent. And as we become more interdependent, we require more free-access goods, such as contract enforcement and certain forms of reliable knowledge. To get these free-access goods, we need more legitimate coercion. Footnote 12
Coercion Helps Solve Collective Action Problems
Take the looming catastrophe in climate change. Many people still do not see a reduction in global warming as a free-access good that we will have to coerce ourselves to produce. At the other end of a long spectrum, take the trivial case of blueberries in the winter. To get them on our tables, state coercion helps at every stage. In January, more than half of US blueberries come from Chile. Footnote 13 Even before they are planted, the Chilean Agricultural Ministry gives farmers information about the crop and the Chilean Plant and Animal Health Policy, which is one of the strictest in the world, helps keep dangerous organisms out of the agricultural system. Getting the berries to the table also takes a large number of free-access goods, each provided through some sort of state coercion, including the coercion needed to collect taxes. Highways, ports and airports, emissions and pollution regulations, safe seas, law and order, property rights protected by the courts—the list of free-access goods that we need to get those berries, accurately labeled and safe to eat, to our tables, runs on and on. The market depends on the coercion that makes those free access goods possible.
To keep this whole complex of free-access goods going smoothly, the necessary coercion must be relatively well designed, in recognition of its costs. For example, extrinsic motivation such as coercion tends to drive out intrinsic motivation. Footnote 14 Fining people if they pick up their children late from daycare increases rather than decreases the number of late pickups. Footnote 15 Paying Swiss communities to accept nuclear waste makes them less likely to accept it than asking them to do so for the nation’s good. Footnote 16 So, on the implementation side, we often need more persuasion and less reliance on coercive power. As Elinor Ostrom and others have shown, even necessary coercion should be minimal, graduated, appropriately designed for the specific situation, and both formulated and applied by those who will have to live under it. Footnote 17 Yet Ostrom never confronted—I would say she avoided—the problem of legitimacy.
The Best Coercion Is Legitimate Coercion
Many studies have shown that people are more likely to obey a law they consider legitimate. Footnote 18 The more legitimate they think the coercion is, the less often sanctions need to be applied. Thus the best coercion is legitimate coercion. Less legitimate coercion throws sand in the cogs, the system begins to grind more slowly and less well, and the product becomes more expensive—sometimes too expensive to compete.
As I said earlier, legitimacy is harder to define than coercion. The term “legitimacy” refers to two different things. “Empirical” legitimacy arises when a group of people believes that something is legitimate. You can measure its existence by asking people questions about their beliefs. “Normative” legitimacy exists when what we believe can be justified with good reasons that withstand collective scrutiny. Empirical legitimacy does all the work in backing the coercion that solves collective action problems. In a good world, however, our belief in the rightness of a given system would also be backed by the actual rightness of that system. Normative legitimacy would underpin empirical legitimacy.
The centrality of legitimacy to the entire enterprise is, I believe, why political science (unlike the other social sciences) has always included inquiries that are explicitly normative, addressing issues related to the ideals that make coercion legitimate. It is the job of democratic theory, among other things, to explore democratic ideals, ferret out their implications, reveal their contradictions, and either underscore or challenge the conclusion that they deserve our allegiance.
Democracy is ideal based. In the advanced industrial democracies of the early twenty-first century, democratic systems are legitimated not by one ideal, but by a constellation of ideals. These include republican liberty (by which I mean autonomy or self-rule), “liberal” liberty (by which I mean the ideal of non-interference), a form of community grounded in equal respect, and various forms of equality based, among other things, on human dignity and formal justice. Footnote 19 These ideals derive from human experience, have evolved over time, and resonate among the marginalized as well as the powerful. They anchor the legitimacy of democratic systems. They also conflict with one another. Indeed, the ideals on which democracy is based are all what Kant called “regulative ideals”—ideals toward which we should strive, all other things equal, but which we can never expect to reach fully. Footnote 20 They are impossible to reach in all their fullness partly because of the nature of reality and partly because they do often conflict with one another. This means that no democracy is ever fully legitimate. Most people can tolerate this failure because they are practical. They don’t expect things to be perfect. The feature that further legitimates democracy, when this is in fact the case, is that the institutions are designed to be likely to move, when possible, closer to these animating ideals.
In 1989, Robert Dahl listed and described many of the ideals on which democracy is based, along with some of the practices to which they have given rise. Footnote 21 His analysis has served as the touchstone for much subsequent empirical work. But our understanding of these ideals is still evolving, and so are the practices anchored by these ideals. It is to those practices that I turn now.
Negotiating to Agreement
We have learned a great deal in the last fifty years about the legitimacy-inducing power and shortcomings of democratic mechanisms such as unanimity and majority rule, deliberation, and many forms of electoral representation. Recently we have even begun to understand better the legitimacy that can be based on representation by lot. Footnote 22
One legitimating mechanism, however, has been surprisingly neglected both empirically and normatively, namely negotiation to agreement. Negotiation to agreement is an important source of legitimacy in a world that greatly needs more legitimate coercion to solve its growing list of collective action problems.
Negotiation to agreement is possible only when issues are “tractable.” That means there must be some area in which two (or more) negotiating groups can get better outcomes together than they could working separately. In politics, negotiation to agreement is most important in non-Westminster systems of government. It is a large part of politics in systems with multi-party proportional representation, where parties must constantly negotiate to form and maintain coalitions. It is also a large part of politics in systems with a strong separation of powers, where parties must negotiate around many constitutional veto points. In the United States, as Thomas Edsall said recently, “politics is negotiation.” Footnote 23
A more inductive approach to politics would immediately reveal that negotiation plays a major role in domestic political decisions. But we have little empirical work in this field. The best work mostly comes from the rational choice tradition, which models what the incentives provided by specific institutions would induce rational but self-interested political actors to do. Footnote 24
For the other aspects of negotiation in domestic politics, political science must build on the last half-century of scholarship in labor studies, in business, law, and policy schools, and in our own subfield of international conflict resolution. One big lesson that these fields teach—and that we should take seriously in designing institutions—is the importance of what negotiation theorists call “expanding the pie,” which primarily means bringing in new issues on which the different parties have different priorities, allowing each side to trade its low priority items for higher priority items that matter less to the other side. However, the textbooks and courses in business, law, and policy schools aim primarily at teaching individuals how to be good negotiators, not at identifying institutions that facilitate successful negotiation. That is a task for political science.
Because negotiation is such a large part of politics, we need to identify institutions that help negotiators bring in new issues and make good trades. And because one function of institutions is to help correct individual mistakes, we need to figure out which institutions help participants combat the cognitive and emotional barriers to successful negotiation. Self-serving bias, for example, means that we tend to select our facts and even our conceptions of justice and the common good without noticing that we are selecting facts and concepts that benefit us more than others. Footnote 25 Well-designed institutions help participants agree on facts and see that their conceptions of justice are not shared by everyone else.
Why do political scientists know so little about negotiation empirically in our domestic democratic institutions? Most negotiation takes place behind closed doors and leaves no easily-quantifiable record. The light from the lamppost does not reach there. We could nevertheless triangulate from retrospective interviews, as journalists have done. Footnote 26 Or we could analyze the letters that participants assumed were private but that were later made public, as Daniel Naurin has done in the EU. Footnote 27 This wide-open field poses many unanswered empirical questions. Footnote 28
Normative theorists have also neglected negotiation. We have tended to see deliberation to agreement (based on common interests) and majority rule (based on conflicting interests) as the only two democratic mechanisms for generating legitimate coercion. Negotiation to agreement is another such mechanism, with a unique normative configuration. It captures the legitimating force of agreement—like the consensus that sometimes ends deliberation. But like majority rule, negotiation to agreement also allows the parties to recognize their sometimes on-going conflicting interests. Footnote 29
The fact that negotiation is often about conflicting and untransformed material interests may even have helped stymie significant normative inquiry on the subject. A long tradition, stretching back at least to Aristotle, identifies regimes as perverse or tyrannical if they are not aimed at the common good. As a result, negotiations over conflicting material interests are tainted. In 1962, for example, Jürgen Habermas roundly condemned “compromise [that] literally had to be haggled out, produced temporarily through pressure and counterpressure and supported only through the unstable equilibrium of power constellations between state apparatus and interest groups.” Footnote 30 Habermas later changed his mind on compromises and bargaining, but his early focus on the common good and excoriation of both self-interest and material interests has also characterized parts of the work of many other contemporary political theorists, including many whom I admire. Footnote 31
Once we abandon this perspective and accept negotiating to agreement as an important way of legitimating state coercion, many questions arise. How do we legitimate the negotiation? Negotiation involves both power and persuasion (“bargaining and arguing” some call it). Footnote 32 It is far more open than majority rule to domination by the powerful. This feature of representation raises major normative issues. First, who is doing the negotiating? The kinds of representatives we want when the representatives are acting behind closed doors are different from the kinds we may want when we can monitor every important move. Footnote 33
Second, how do we make the relations among the negotiators “democratic”? Like majority rule, democratic negotiation must presumably be legitimated by its approximation to the ideal of equal power, or at least the proportional representation of interests. No theorist disagrees with the general thrust of these criteria. But it is not clear if “equal power” implies equal individual power or the equal power of competing groups, let alone proportionality for the affected interests. Nor is it clear which threats of sanction are allowable—perhaps only those that offset existing power disparities in order to create a more equal negotiation. Much at the very base of what we might mean by a democratic negotiation still needs to be explored.
In addition, negotiation poses a host of subsidiary normative questions. The privacy that allows negotiators to speak freely is antithetical to the widely accepted democratic norm of publicity. The ongoing relationships that produce sufficient mutual trust among legislators to generate creativity in negotiation often require long incumbencies, which are antithetical to the widespread democratic norm of frequent closely contested elections. The tradeoffs on unrelated issues that make package deals work are antithetical to the classic focus on the common good. All of these issues need careful normative attention before we can understand how best to use negotiation to generate democratically legitimate coercion. Footnote 34
Conclusion
In a Venn diagram, the two concepts “collective action” and “need for formal government” have only a partial overlap. In the first place, as your recent voluntary donations to the common pool illustrate, we can largely solve some collective action problems without the formal coercion of government. In the second place, we need government for issues other than collective action problems. Government can act as a coordinator without coercion, as a model for right behavior, and as a voice for collective ideals. Government institutions can serve as deliberative arenas and focal points for collective struggle over who we are. We may also need the formal coercion of government to accomplish other goals, such as justice, that are arguably not collective action problems. Footnote 35
Yet, while I am not contending that the only function of government is coercion, I believe that coercion is government’s most important and most controversial function. Most of that coercion is required to solve collective action problems. Surprisingly few people today understand the basic point that whenever you want to produce a free-access good—and they are all around us, everywhere, not least in our changing climate—you will often need some coercion to help make it happen. For large scale problems, you will very often need state coercion.
We do not like the word or, quite reasonably, the fact of “coercion.” The Nobel Prize Committee, in praising Elinor Ostrom’s pathbreaking work, implied that she was showing how to do without the state. Footnote 36 In fact, however, Ostrom always made the “polycentric” point that for informal or local self-governing institutions to work well, they usually had to be “nested” within larger state entities that could, among other things, police agreements. It has not been widely noted, but in her most famous case, the groundwater agreement in southern California, local and private negotiations were successful only because the state Supreme Court had threatened to impose its own plan if the parties could not agree. Footnote 37
It would be astonishing if the use of coercion did not, as a general rule, favor the more powerful. It would be astonishing if even the most well-meaning bureaucracies did not want to advance their missions with ever more personnel and ever more access to coercive power. So coercion, whether state or private, will tend to be unequally applied and greater than necessary. Resistance to state coercion is therefore always an important part of governing ourselves. Even in relatively uncorrupt and just governments the threat of resistance plays an important role in keeping them relatively uncorrupt and just. Resistance has an especially important role when governments are corrupt, self-serving, unjust, or deeply misguided.
Viscerally and experientially I identify with resistance. Women, for example, have won most of our gains in the last two centuries by resisting the domination of men. My generation grew up with resistance and our political theory was largely about resistance. Nevertheless, I think the western democratic tradition, anchored in resistance to kings, has focused too much on the possibilities and actualities of tyranny and domination and not enough on the equally important problem of how to create legitimate coercion for collective action. In this talk I have focused on legitimate coercion because of its centrality to what political science should be about. I do not mean to exclude protest and resistance from the concept of governing ourselves; I just want to strike a better balance.
More broadly, on the mission of political science, I contend that political science has always organized itself at least in part to help human beings govern themselves. It is in our DNA. Aristotle, lecturing on politics, wanted his listeners not only to hear, but also, after hearing, to govern themselves well. So too Hobbes, Locke, Montesquieu, and Madison wrote to help their readers do better in this business of governing.
When political science first became an academic discipline, Frank J. Goodnow, the first president of the APSA, recommended in his 1904 presidential address that the association bring together scholars and practitioners so that the scholarship would not “conduce to the adoption of impracticable and unworkable methods.” He gave a learned lecture on administrative law and concluded with recommendations aimed at his “hope to secure the highest public welfare in the industrially and socially complex age in which we are living.” Footnote 38
In England in 1926, when Harold Laski was appointed to the first chair of political science in that country, he said in his inaugural lecture that in political science we have to “define aims . . . [and] discover both the institutions through which those aims are likely to be realized, and the methods by which they are to work.” His first question was, “Do we need a parliamentary system?” Footnote 39
Democracy, as the pragmatists point out, lends itself particularly well to experimentation. Footnote 40 But experiments often need people to evaluate, sift, and pass on to others the results of the experiments. Political science performs this task.
All of the subfields in political science are involved in the process of trying to improve the processes by which we govern ourselves. We need to explore the ideals we have—or think we have—about how we should govern ourselves. We need to explore the polity we know most intimately, whether it be the United States of America or another polity, to understand it in greater depth. We need to compare existing governments to one another, to ferret out their greatest strengths and weaknesses. We need to understand better how states and other entities relate to one another and how they can do so more productively.
We also need to use all of the methodologies we have developed—from the close reading of thinkers whose subtleties have often been forgotten or never understood as the intellectual wave of the moment swept past them, to the stringent analysis through formal models of the possible interactions of individuals motivated only by self-interest, to experiments in labs and in the field, to interpretive immersion in the processes of governing and being governed, to the quantitative analysis of large data sets.
What we have in common, I believe, is that we can all try to help the human race figure out how to institute sufficient legitimate mutual coercion to stave off the catastrophes toward which we are heading—as well as to move toward goals that would be good, by and large, for us all. We are the only discipline organized to address these questions. If political scientists do not try to do this, who will?
Jane Mansbridge is Charles F. Adams Professor of Political Leadership and Democratic Values at the John F. Kennedy School of Government at Harvard University (jane_mansbridge@harvard.edu). She is the immediate past president of the American Political Science Association (2012–2013) and the author of Beyond Adversary Democracy, an empirical and normative study of face-to-face democracy, Why We Lost the ERA, a study of anti-deliberative dynamics in social movements based on organizing for an Equal Rights Amendment to the U.S. Constitution, and other works on feminism, oppositional consciousness, self-interest, representation, democratic deliberation, everyday activism, and the public understanding of collective action problems. She thanks André Blais, Martha Field, Chase Foster, Inge Hoffmann, Morton Horowitz, Jeffrey C. Isaac, Hélène Landemore, Cathie Jo Martin, Nancy Rosenblum, Duncan Snidal, Kathleen Thelan, Melissa Williams, and Robert Young for criticism, suggestions, and support at various stages in the evolution of this address.
