Democracy and the Political Unconscious

By now it should be clear that my ideas of deliberative democracy depart considerably from those of Habermas and company. What I am sketching is more compatible with poststructuralist notions of subjectivity and discourse. The thread that runs from that kind of continental philosophy to this kind of democratic theory is the practice we call talk, which political leaders and traumatized body politics shudder at engaging in. How are we to talk with the other who seems so unreasonable? How are we to talk when we still cannot fathom what has befallen us? In this chapter I aim to sharpen the difference between what I am calling an integrative theory of deliberative democracy and other versions. I want to show that this ideal is something toward which those hoping for a better democracy might aim.

I draw on my experience working at the intersection of three models of deliberative democracy: the preference-based model held by many deliberative theorists in the social sciences; the rational proceduralist model suggested by John Rawls’s political philosophy and Jürgen Habermas’s discourse ethics; and what I have been calling an integrative model, which has been overlooked in the literature but can be seen at work in most actual deliberative forums composed of members of a polity deliberating on that polity’s direction. The latter includes the National Issues Forums (NIF), described earlier, a network of civic organizations that run deliberative forums consonant with a quasi-Deweyan approach to public deliberation. My aim in this chapter is to see the extent to which any or all of these models can be mapped onto actual deliberative forums, including deliberative polls, the method developed by James Fishkin (1991 and 1995). The three models are not mutually exclusive. A deliberator might see herself engaged in more than one at a time (perhaps testing out, as in the second model, whether a justification for a policy is acceptable to all, while hoping to find some integration even where participants cannot reach accord, as in the third model). Any combination could work, in practice, even though some of the methods may, in practice, work at cross-purposes. For example, focusing largely on the normative aims of the second model might minimize the empirical facts of people’s actual, strategic aims, of which the third model is highly aware. I want to draw out the theoretical differences among these approaches and show how the differences matter in practice.

My own “intersection” among the three approaches is rather makeshift: I happened to begin working with Professor Fishkin on a deliberative poll we called the National Issues Convention (NIC) while I was a graduate student at the University of Texas (writing a dissertation, in part, on Jürgen Habermas). I was never Fishkin’s student; our collaboration began because of my association with the Kettering Foundation, which is a major force behind NIF.¹ Fishkin became allied with NIF and the Kettering Foundation because of their shared interests in deliberation and because Kettering offered support in finding trained moderators and putting together issue briefing books. Most of the deliberative theory swirling in the air drew on the resurgence of political philosophy brought on by both Rawls’s work and Habermas’s notion of reasoning, that in moral, ethical, and political discourse participants should try to offer justifications for their favored policies that would be agreeable to all others affected by them.

The intersection of deliberative thought in deliberative polling, normative political theory, and actual deliberative practice is a more general phenomenon. All draw on the key term “deliberation,” and observers expect a commonality because of this. But this intersection is not altogether seamless. Many who take part in deliberative experiments have rather different ideas of what “deliberation” means, but the term often gets used as if everyone agrees. The differences are not merely semantic; they are rooted in very different conceptions of politics. Because it operates at the intersection of these differences, deliberative polling, specifically the two National Issues Conventions held in the United States, offers a useful case study of how these approaches converge and diverge. In this chapter I describe the three models I see at work and offer some preliminary ideas of how they enter into deliberative polling. My goal is not to offer an encyclopedic account, but enough details to flesh out the key differences in their orientations and goals and to show the virtues of an integrative model of deliberation.

The first model I consider comes out of the social sciences, primarily via political scientists’ adoption of the language and theoretical structures of economics.² From the point of view of classical economics, a human is homo economicus, a being who sees the social world as a market in which he or she tries to maximize his or her own preferences. Political science takes this notion and makes it democratic by saying that a democracy would be rule by the people in a way that helps them maximize individuals’ preferences as much as possible. But given that one person’s aims will no doubt conflict with another’s, democracy calls for some way to compromise or to aggregate preferences while treating every individual as an equal, respecting the preferences of all. Though aggregating, e.g., voting, seems to be a very democratic decision procedure, it has its problems, especially when individuals’ rankings of options are somehow incoherent (for example, ranking a conservative option first, a liberal one second, and a moderate one third) or when a group of individuals’ rankings show no clear winner (for example, when one person prefers A to B, another B to C, and a third C to A).

Social choice theorists have tried to solve such problems, attempting to discover how social or public policies that respect and preserve the preference rankings of the individuals within a polity can be devised. There are two sides of social choice theory: individuals ranking their preferences between two or more policy options and social planners devising ways to meld these numerous, individual rankings into one rank ordering of options. Yet social choice theorists have yet to find a nonproblematic way to do this (Elster and Hylland 1986:2). Most agree that people’s individual, given preferences should be aggregated in some way. But how? What kind of voting system would ensure that “the will of the people” really does emerge, especially when there is no clear first choice? For example, what happens when the option that got the second highest number of votes is nearly everyone’s last choice? Our winner-takes-all system leads to all kinds of counterintuitive inconsistencies and difficulties, and social choice theorists have taken it on themselves to try to solve these problems, often by mustering intricate formalisms and tackling logical minefields. Decades of failure have led to the view that there is no “will of the people” that can be objectively put forward. Any aggregation scheme introduces its own shape to what this will seems to be. Moreover, no scheme seems to do a good job of illuminating social preference without being vulnerable to individual voters manipulating the system to get their favorite candidate chosen. Perhaps the whole enterprise of trying to develop a public policy that is consistent with individual preferences is doomed, along with democracy in general.³

Certainly by the 1970s, the science of politics had led to the view that democracy is a vain hope, inconsistent and absurd. This was an odd place for a discipline to land, especially one that began in part as an attempt to understand the mostly American democratic project (Smith 1997). Perhaps in response, a more optimistic area of study has emerged in political science departments since about the mid-1980s: deliberative democratic theory.

Social scientists who have taken the deliberative turn reject the following views: that individual preferences are fixed prepolitically; that they are primarily self-regarding; that individuals are rational to be ignorant—meaning that they’d just as soon not bother to inform themselves since they think it wouldn’t make any difference—and hence their preferences are ill informed; and that each individual set of preferences will likely remain incoherent. Jon Elster has argued that deliberation is a means for transforming individual preferences (Elster 1998:1). Fishkin and his colleagues argue that deliberation can help people develop opinions that are more informed, reflective, and considered (Luskin et al. 2002). Because they retain the social science focus on individual opinions and preferences, I call this model the preference-based view. Still, there are key differences between deliberative theories of preference and the old classical economists’ notion. The latter holds that preferences are given in advance of the political process and that each individual’s preferences are primarily self-regarding—that individuals tend to put their own desires before others. Hence politics is an arena for getting what one wanted before entering into that arena. From a deliberative standpoint, preferences are not fixed in advance; they can be informed with balanced briefing materials and expert knowledge and transformed through deliberations with others, making them other-regarding, not just self-regarding. In short, these deliberative theorists think that people can transform their preferences for the better during deliberative, informative discussions with others in the community, making them more reflective, informed, and cognizant of others’ concerns.⁴ Such preferences would not be so difficult to aggregate rationally and democratically. Hence democracy becomes a possibility, as a kind of governance in which preferences transformed through deliberation are the basis for public policy.

In this view, though, public policy is not formed in the deliberations. Given that deliberators will rarely unanimously agree on what policy is best, a deliberative polity still needs some kind of external decision-making procedure (Miller 1992). This might be a direct vote or a matter of transmitting up the political ladder the new, improved set of individual preferences. Unlike conventional democratic politics, where policy makers make policy on the basis of unreflective preferences captured in standard public opinion polls, this model offers policy makers a snapshot of what a deliberative public thinks. That is how John Dryzek characterizes deliberative polling:

From the point of view of deliberative democracy, ordinary opinion polls are pointless because they register only unreflective preferences. The idea of a deliberative poll is to assemble a random sample of members of the public, have them deliberate about the key issues of the election, poll them on their positions on the issue, and publicize the results. The intent here is to model the distribution of opinions that the general public would hold if they were able to engage in genuine deliberation, a far cry indeed from the unreflective preferences which ordinary opinion polls register. (Dryzek 2000:55)

Without diminishing the importance and usefulness of deliberative polling, I want to highlight one of its self-imposed limitations (which others might take to be a benefit): it truncates the political task of trying to turn individual views into public judgments. The end product of a deliberative poll is not any kind of collective judgment about the nature of the problem and what ought to be done. Moderators are specifically instructed to avoid questions that might, to some, seem at all coercive, such as, “What do you think we [meaning the political community] should do?” or “Is there any common ground for action on this problem?” The result is not a public expression about what might be the best course of action, only an indication of the distribution of individual opinions. However considered these are, they do not equal an integrated policy. Even aggregating the results does not lead to a coherent, democratic policy or even to the will of the people, as social choice theorists well know. A legislature might take on the political task of trying to integrate the various needs, aims, and constraints into something like a coherent public policy. If it does so on the basis of individuals’ considered judgments, so much the better. But we should be keenly aware that the political work occurs at this higher level, not at the level where deliberators work on transforming their own preferences. The preference-based view shows how individual opinions are transformed into superior opinions, but not into public policy.

Why do deliberative polls shun any deliberation aimed at developing a public voice on an issue? Like other deliberative theorists, preference-based adherents are committed to democracy. Their commitment is shaped by the view that democracy calls for respecting individual preferences and that anything that exerts any “untoward” (e.g., coercive) force on individual preferences is undemocratic. Such forces include factions, the “tyranny of the majority,” social pressure, and the like. Hence, there is some tension inherent in an individualist, preference-based model of deliberation, for the more people deliberate in public with others, the more likely they are to be moved by these others in their midst. Therefore, preference-based deliberative theorists try to guard against public pressure on individual deliberations, a real problem in the setting of public deliberation. Their goal is for participants to use deliberative settings to transform their preferences without being unduly swayed by others. In their view, deliberations should be geared toward giving participants full information and a clearer picture of how each option on the table would or would not satisfy each participant’s preferences;⁵ they should supply expertise and an appreciation of others’ concerns, not social suasion. These theorists tend to worry that deliberations might lead participants to conform to others’ expectations rather than to refine their own preferences (Elster 1998:15; Sunstein 2003), which they think would result in a kind of “false consciousness” where participants are not fully aware of their own, true self-interests and opinions. For them, democracy means rule by fully informed individuals. Autonomy equals not being unduly influenced by anyone else.

The social scientists’ approach to deliberative theory goes a long way toward turning homo economicus into an other-regarding democratic citizen. But at the end of their deliberations, there is no discernible public transformation. Each individual might transform his or her opinions, in light of more information and exposure to others’ perspectives, but there’s no expectation that the result might be public views; they will just be more rational and considered views. At the end of the deliberations, individuals’ views still have to be transformed into some kind of social ordering. If everyone were to agree on the nature of a problem and what policy best addresses it, there would be no difficulty. But most anyone steeped in the facts of public life, Elster included, thinks this is unlikely (only the normatively oriented, to whom I turn in the next section, think this is a possibility). Given the fact of disagreement, some way needs to be found to make collective decisions. If the way is through voting, then the preference-based theorists have come full circle to the problem of articulating “the will of the people” out of a set of individual preferences. Bound to individualism, preference-based thinkers still face the challenge of social ordering.

The second model I want to lay out here comes from a different direction, not the supposedly empirical and normatively agnostic orientation of the social scientist but the normatively steeped orientation of the philosopher.⁶ This model sets a very high bar for what kinds of reasons deliberators should offer and accept: participants should deliberate on the basis of reasons that are rational and acceptable to all. This view specifies what can count as a good reason and what kind of procedures should be in place to ensure a good outcome. Accordingly, I call it the rational proceduralist model of deliberation.

In this second model, citizens are guided by a will to come up with universalizable norms—or at least norms that are acceptable to all affected by any given policy. This view can be traced back to both the social contract tradition of consent theory and the Kantian normative claim that we, as rational individuals, can act morally by only acting on the basis of maxims (or policies) that can be rationally universalized, i.e., applied to all equally and without contradiction. Here we have an explicitly philosophical conception of autonomy: it means acting and choosing on the basis of universalizable norms; true self-rule is to live by rules that hold for all, not just for oneself. Rational deliberators offer arguments concerning justice and the public good. In this view, motivations of self-interest stand in the way of developing legitimate public policies, which should be good for all. This theory encourages deliberators to adopt an impartial, objective point of view and to offer reasons (not rhetoric) that all others would find compelling. Otherwise, a policy would not attract general consent—consent of all those rational agents affected by it. Irrational agents need not be heeded; in fact, according to Gutmann and Thompson (1996), they should not even be in the room.⁷ Though this view sees deliberators as always already in community, it does tend to think of them as capable of imagining themselves stripped of affective and communal associations, roles, and conventions so that they might be able to deliberate objectively and impartially. They need to use their reason so as to imagine how a policy would affect anyone else. They are rational agents.

In this model, deliberation is a way individuals collectively decide whether a policy is legitimate. A policy or law is just only if all those affected by it have an opportunity to consider, collectively, whether or not it is just and if all those affected assent to it. Deliberation, then, is the process through which people decide whether a proposal is normatively or ethically right. It involves the back and forth of argumentation, with everyone having an equal opportunity to put forward his or her own case. Ultimately, the “unforced force of the better argument” will prevail; that is, all the participants should ultimately agree on which proposal is most rational and right. The rational procedural model considers deliberative democracy as a way to create legitimate public policy, that is, policies that all citizens would, under ideal conditions, have authored themselves.

Both Habermas and the political philosopher John Rawls contribute to this approach. As Elster writes, “the arguments advanced by Habermas and Rawls do seem to have a common core: political choice, to be legitimate, must be the outcome of deliberation about ends among free, equal, and rational agents” (Elster 1998:5). It is no wonder then that Habermasians like Seyla Benhabib and Rawlsians like Joshua Cohen arrive at roughly the same philosophical position on deliberation. Whether one adopts the Habermasian regulative ideal of the ideal speech situation or the Rawlsian regulative ideal of the original position, in democratic deliberations all those affected should recognize the outcome as in keeping with what they would have chosen had they had an opportunity to participate.

For Habermas, the operative form of reason in this democratic setting is communicative, internal to discursive practices. Though Habermas’s theory of communicative action draws heavily on American pragmatist conceptions of self, truth, and action, ultimately it looks very much like Enlightenment rationality with its claims to universality, impartiality, and, in Benhabib’s hands, reversibility of perspectives (see McAfee 2000:ch. 1). Habermas’s theory and allied theories of public reason-giving are put to use in deliberative theory, laying out the limits of the kind of talking (that is, reasoning out loud) that ought to occur in a deliberative forum.

In this model, partiality is a serious fault. One should be able to see the whole picture, not just one’s own arena. It seems that one need not actually consult others to find out what the world looks like from their perspective; each sovereign citizen should have a mental map of the whole. Accordingly, the ideal deliberator reasons publicly with others not to get more information about how policies would affect others, but to get their consent. This model holds out hope that decisions can be reached by consensus. If all agree on what policy is best, then there will be no need for social-choice type aggregation, bargaining, or voting.

In its search for unanimity, deliberation becomes a contest, a battle of arguments in which the best argument wins. To be a contender, a policy needs to get universal consent. This kind of deliberation does not try to piece together second-best alternatives into something with which most everyone could live. It looks for policies that are simply the best. As a result, it is possible that the participants in such deliberative ventures are more interested in winning a contest than in solving problems. This model seems to lose sight of the reasons people enter into public deliberations—to work through traumas or other difficulties that have jammed public politics or because their communities are wracked by problems that politicians seem unable to solve. In actual community deliberations, participants are not looking for which claim is normatively right, but for which picture of the problem is most telling and which courses of action have promise.

Yet even though universality may be the guiding ideal, most deliberative proceduralists realize that consensus is rarely reached and some kind of vote will be needed. As a result, this view runs into the very same problems that social choice theory does: finding the best way, short of consensus, to articulate the will of the people (see Dryzek 2000:38–41).

Anyone familiar with deliberative theory probably recognizes the above two models readily, especially with the little bit of detail that I have provided. The third model has not been discussed in the literature, or certainly not to any great extent. It is the model that I came to know firsthand through observing the deliberative forums that are part of the National Issues Forums network. As mentioned earlier, some of the intellectual founders of this approach include Kettering’s president, David Mathews; the survey researcher Daniel Yankelovich; and the political theorists Benjamin Barber and Harry Boyte. Also instrumental have been the works of Hannah Arendt and John Dewey.

While NIF now has many sister organizations around the world, I trace the intellectual roots of this model to an American pragmatist tradition that is concerned more with “what works” than “what is true.” The model has some roots in the civic republican tradition as well, though its normative conceptions are not as strong as those criticized in Habermas (Habermas 1998:244–249). It sees the public as heterogeneous, not the collective actor of a Rousseauean model. This model is empirically observable, but participants have normative concerns: when people deliberate together about public matters they develop an interest in the public welfare (in solving public problems) that may override their particular preferences.

Even with all these theoretical resources, the National Issues Forums are primarily driven by the way people actually deliberate and what their aims and concerns seem to be when they sit down together and try to solve problems that resist solution. The deliberators are motivated by the need to find a way forward on problems that affect them and their communities. Here the reader might recall E. J. Dionne’s book, aptly titled Why Americans Hate Politics. Its main point is that Americans do not care about ideology, which seems to be the currency of conventional politics; rather, they want solutions to problems. This is an insight that has long steered NIF. David Mathews and I spent a few years writing guides for deliberation to be used by NIF convenors and moderators. We oriented the texts according to a framework we called “choice work.” The aim was for citizens to consider an array of policy options and, on each one, to spell out the costs and consequences as well as the trade-offs that would have to be made if the approach were adopted. Only by “working through” the various choices, grappling with what must be abandoned in order to proceed in a particular direction, do deliberators begin to develop a public judgment about what policy might be best. Though it often evokes the language of “finding common ground for action,” this approach does not aim for happy consensus. Rather, choice work engages deliberators in the pragmatic task of delineating what courses of action might work given polity members’ many aims and constraints.

According to the integrative model, deliberation is a process through which people grapple with the consequences of various public problems and proposals. Participants focus on solving public problems in ways that are consistent with their publicly formed understandings and ends. Instead of narrowly focusing on autonomy, this model sees democratic choice and action as practices that involve people considering how various options would affect their communities. The public dimension of deliberation is indispensable to the task of fathoming problems and forming a public that can respond. Instead of seeing politics as bargaining about preferences, people see it as a difficult matter of deciding what kinds of communities they are making for themselves. Instead of merely preferring, deliberators choose.

At the outset, I called this model “quasi-Deweyan.” By this I do not mean that Dewey spawned the integrative model but rather that his observations mesh with it uncannily well. I noted in the previous chapter that Dewey’s work in The Public and Its Problems intersects with the understanding of deliberation that I am laying out here. Both Dewey and NIF hold that public communication can be a way citizens simultaneously develop an understanding of public problems and of themselves as a public that can and should create sound and effective public policies. Having defined the public as all “those indirectly and seriously affected for good or evil” by the “human collective action” of some particular group (Dewey 1954:27, 34–35), Dewey understood the centrality of deliberating to define both public matters and the public: “An inchoate public is capable of organization only when indirect consequences are perceived, and when it is possible to project agencies which order their occurrence” (ibid. 131). In keeping with Dewey’s insight, actual public deliberations usually spend a great deal of time developing a public picture of what a problem is and how it affects those in the room and others throughout the political community. As deliberators gain an understanding of the nature and the many aspects of the problem at hand, they also begin to see themselves as a public.

Unlike the second model, which expects deliberators to act according to the Enlightenment, universalizing ideal, this model aims for integration of multiple, heterogeneous views, accepting and making use of citizens’ particular perspectives. Because each participant starts out with a limited picture of how a policy under consideration might affect others, they deliberate in order to learn. They seek information, not so much about facts but about the consequences of various policies. In this model, citizens’ partial perspectives can be integrated into a viable, sound policy choice that is always provisional and subject to change.

When people come together to deliberate on matters that affect their polities, they seem to transform personal concerns and interests into public ones. To understand this phenomenon, observers and political theorists need to move beyond the tired dichotomy between egoism and altruism. It is not that public deliberations turn participants into altruists. Rather, the deliberations help forge an immediate interest in public matters, conjuring up the history of the term “interest” itself, inter-esse, a way of being between and with others. Participants develop an interest in the welfare of their political communities.

Moreover, this model attends to the problem over which both the first and the second models stumble: how to set policy direction when there is not full, or even much, agreement. Participants use their disagreements as productive constraints that help them identify in which possible directions, albeit few, the polity might move. In the many deliberations that my colleagues at the Kettering Foundation and I have observed, participants left saying that even when they did not agree with other participants, they came to see why the others held the views they did and changed their views of others’ views. Even in the face of trenchant disagreement, participants would focus on coming up with a direction that would accommodate the plural concerns in the room. Unlike the first model, which leaves the aggregation problem to social planners, this integrative model understands that deliberators want to have a hand in shaping policy, that indeed, this shaping is central to deliberation itself. People do not want to just be preference inputs into some social utility function. They want to help decide what the policy should be in the very process of trying to understand what the problem is, how it affects all concerned, and what kind of polity they want to forge.

Of the three, this view is the least idealistic. It has the most “communal” understanding of human psychology, seeing people and publics as constituted through their common language, customs, norms, relationships, and communities. It does not call on participants to imagine themselves stripped of affective and social associations in order to deliberate well. People do, and should, bring their particular concerns to the table when they deliberate with others. This model of deliberation shares the same kind of philosophical frame as social and material theories of history and community as well as Gadamer’s hermeneutics and poststructural accounts of meaning making.

In practice, moderators try to ensure that all participants have an equal opportunity to speak, that no speakers dominate the deliberations, and that other factors in keeping with Habermasian speech-setting ideals are in place. But moderators also try to elicit stories from deliberators, using prompts such as “Tell us how you came to hold the view that you have.” (This usually brings forth a story that helps explain why someone would hold a view that others might find objectionable or unreasonable, in the process showing another aspect of an issue that others might not have considered.) As stories emerge, participants begin to weave them together into a larger narrative of the meaning of events, in a way not unlike what occurs in the process of testifying to traumatic pasts. Working through is evidenced by new shared accounts of what the circumstances are and what is politically possible.

This model aims at getting participants to take into consideration other participants’ concerns, aiming for a choice that reflects a considered, public judgment on the issue. Through their discussions, deliberators come to see possible outlines for public action. And they come to see themselves as part of a public, as public actors with considered judgments and purposes who can help shape public policy. The goal is not rationality per se, but the possibility of understanding the public dimensions of problems and identifying what, if any, sound and sustainable directions toward which the public might move. At the very least, it sets the boundaries of what Yankelovich calls public permission.

These three definitions certainly do not exhaust the alternatives, but they do capture what seem to be the three most prevalent views. All have their virtues. But unless the different emphases among these views are made clear—and unless their tensions are addressed—deliberative practice can falter. With this in mind, I close this chapter with some thoughts on how these themes manifested themselves in the National Issues Convention, sometimes undercutting the potential of the event. Deliberative polling and National Issues Conventions are tremendously valuable political events. I simply want to show why the proponents should move closer to an integrative model instead.

The role of experts: Drawing primarily on the preference-based model of social choice, Jim Fishkin and his colleagues take one of the central tasks of deliberation to be informing participants’ discretion, giving them the means and the opportunity to develop opinions “worth” listening to. As Luskin, Fishkin, and Jowell write, “The scientific value of the Deliberative Poll is that it provides a way of addressing the effects of information (and thought and involvement) on policy preferences” (2002). To this end, deliberative polling relies heavily on panels of experts and policy makers to answer questions that arise during deliberations. During the last NIC, participants worked through two policy choices and then stopped to select questions for experts. They then went into plenary sessions to listen to how the experts answered. Afterward, they returned to their small groups, deliberated a bit more, and worked on developing more questions. They ended the afternoon with another plenary session with the experts. Before dinner, they met again to come up with questions for policy makers. The Sunday morning session was devoted to policy makers taking citizens’ questions. By the end of the weekend, much more time had been devoted to formulating questions or putting them to experts, and listening to the responses, than to deliberating. Observers noted that participants took much care in how they worded the questions, more, it seemed, to get the panelists’ attention than to get information. Despite the organizers’ intentions, the experts did not seem to be there in service to the deliberators, and the participants felt frustrated when their own questions could not be posed due to limited time.

Expertise plays a much smaller role in the NIF or integrative style of deliberation. NIF does use “issue books,” balanced, informative guides that offer three or four policy choices, discussions of pros and cons, trade-offs, and other data. Deliberations tend to focus more on how various proposals will affect participants’ and their communities’ ends and purposes. Questions of fact arise far less often than questions of value and consequence. In my observations, deliberations proceed quite differently depending on whether they see expertise as a resource for (as NIF tends to do) or as something to aim for (which deliberative polling inadvertently does). Deliberative polls could improve by lowering the profile of “the experts,” treating them as interested parties (which they usually are) who have some knowledge of how proposed policies fit into the larger political picture. Instead of speaking on panels that take up large portions of the program, these parties could be available in the background to answer any questions that spontaneously arise.

The meaning of politics: The three models I have described here have radically different conceptions of politics. The first holds that politics is the practice, exogenous to public deliberation, of turning deliberative preferences into public policy. Politics is the province of government. The second and third models expand the arena of politics beyond government to include the deliberations that go on within civil society (see Dryzek 2000). The second hopes that the deliberative public can engage in policy making, to the extent that it is able to reach rational agreement. The third is more forgiving of disagreement, recognizing that politics begins because there is disagreement, and it puts what is obvious first: that people enter into politics to solve problems. Of course, the first two models might want to lay claim to that purpose as well. But they limit public problem solving to, respectively, the use of individual preferences and rationality. In the integrative model, participants are motivated by their sociality to meet with others they may neither like nor understand in order to find solutions to problems concerning what they care about dearly—the public world that they inhabit, the world they will leave for their children and future generations. They are motivated to fashion a new public world, which is like putting together a puzzle: trying to see what all the pieces are, especially those held by other participants, and then seeing how they might fit together, however imperfectly and provisionally.

None of the above is meant to suggest that there is anything intrinsically wrong with deliberative polling. To the contrary, I think it is a tremendously important advance in democratic practice. My concern is that deliberative polling has been too informed by a preference-based model of democratic deliberation and not informed (nor, as a result, formed) enough by an integrative model. In its concern to help individuals deliberate and refine their opinions, it has overlooked the public task of politics. Yet while many of the theorists behind the scenes might think they are culling well-formed individual preferences, deliberators steadily set about integrating their many perspectives, experiences, and purposes into potential policies that are decidedly public. Each deliberative polling experiment ends with a survey of individual views. Participants each retreat to a secluded spot in the room to take the “after” survey, without consulting anyone else. The survey researchers gather and compare the pre- and postdeliberation responses. Then they hold a press conference—and public television airs a program—revealing the extent to which individual opinions changed. It is thought that the results, when aggregated, will point to what public opinion would look like if people were to think about the issues. But back in that same room, at the end of the deliberative poll, just before the surveys are distributed, there is another public voice to be heard. It is the voice of the people comparing notes, trying to piece together all the moving and conflicting and unsettling matters they have deliberated on together; it is the voice of people trying to take account of and integrate their own and their fellow deliberators’ perspectives, concerns, and desires. The question they return to is, on this matter at hand now, what shall we do? In these rooms the people know that in politics, at the end of the day, our task is not to decide what each of us wants, but to decide what we as a polity should do.