Electoral representation is the double-edged sword of democracy. The ability of millions of citizens to choose proxies for their own will has made modern, nationwide democracy possible, yet it has also reduced the democratic citizen to a passive agent in her nation’s political life. In The Social Contract, Rousseau remarks caustically that the English public “is free only during the election of the members of Parliament. As soon as they are elected, it is a slave, it is nothing.” That implicit understanding—less pejoratively phrased—has governed democratic life ever since; yet even that wobbly contract may have expired. Democratic citizens have rarely, if ever, been more disgusted with those who govern them, and with democracy itself. An ancient deference no longer applies; entrepreneurs of mass resentment have gained power by turning voters against political elites. The global populist contagion forces us to ask whether we need less democracy or more. Do we, that is, need more elite control or a more engaged—and more responsible—public?

It was the frail hope that the answer might be the latter that brought me this past June to the online “deliberative poll” known as America In One Room. The deliberative poll is a practice invented, and more or less owned and operated, by James Fishkin, a political scientist who teaches at Stanford and runs the Deliberative Democracy Lab there. A deliberative poll assembles what Fishkin calls a “mini-public”—a group of people, usually numbering around 500, who have been rigorously selected to represent a nation’s full demographic and attitudinal range. The group receives carefully balanced briefing materials that prepare it to discuss a series of policy ideas—in the case of the exercise I witnessed, proposals to reform democracy, including deliberative polling itself. Then the participants talk. They are given an opportunity to ask questions of experts. At the end they do not vote, since no political authority is waiting for their judgment, but they fill out extensive surveys that show how the deliberative process has changed their views.

The partisan hatred that increasingly poisons our politics and endangers our democracy would seem to doom in advance any hopes for a more participatory politics. Yet Fishkin’s experiments, and kindred ones, have shown that, under the right conditions, deliberation causes people to listen to one another, and even respect one another, despite profound differences of opinion. Deliberation blunts polarization. A more deliberative politics might be the precondition for a more participatory democracy—if, that is, we can figure out how to scale up the exercise from 500 people to 300 million.

In his 1991 book Democracy and Deliberation: New Directions for Democratic Reform, Fishkin explained that “a deliberative poll models what the public would think if it had a more adequate chance to assess the questions at issue.” Unlike a conventional poll—which simply asks people what they think about a given subject, and thus repackages as “public opinion” the welter of bias, ignorance, and media frenzy that forms much of our thinking—Fishkin envisioned a forum that would give citizens access to the same process of education and discussion available to legislative bodies. The deliberative poll was designed not to supplant parliaments, but to send a signal to them: This is what the public really thinks. No less, Fishkin wished to dignify the democratic public by demonstrating that it could think if given the chance. He claims to have been the first political scientist to bring the idea of deliberation down from legislatures, where Montesquieu and Madison had put it, to the level of the ordinary citizen. Certainly he was the first to put the idea into practice. The deliberative poll was a thought experiment brought to life.

There was something charming about the idea of deploying a proposed democratic reform to assess the popularity of proposed democratic reforms. Yet the imperative of halting democratic erosion has given a fierce urgency to concepts that had once been the private darlings of political scientists—ranked-choice voting, nonpartisan primaries, multimember districts, the National Popular Vote Interstate Compact (designed to circumvent the Electoral College), and the like. The implicit subject of the exercise was thus: How can we save democracy? Fishkin and his colleagues had impaneled a group of political scientists, liberal and conservative, to compile the list of proposals. (Of course, political scientists are hardly the final authority on how to save democracy.) Once the subject matter had been determined, the Stanford lab turned to NORC, a social science research and polling organization at the University of Chicago, to do “sortition”—the selection of a representative sample. Researchers there consulted the U.S. Postal Service address file to randomly select census tracts, then blocks, then actual addresses. They then contacted 30,000 individuals and surveyed the 40 percent or so who said they would consider participating to determine demographic and ideological characteristics. That produced a much smaller sample, which was reduced yet further as people dropped out. At the end of this process of distillation, which took several months, NORC delivered to Stanford what was, in effect, America in one room.

But that was only the beginning. Scholars at Stanford prepared a précis and background description of the agenda items, which Norman Bradburn, the former director of NORC and an expert in the drafting of survey questions, used to draw up a polling document asking participants to state their current views on the 40-odd reform proposals. This baseline document would allow researchers to later determine how dramatically participants’ views changed as a result of the polling exercise. Stanford officials also drew up a draft of the background materials that would be sent to each participant; this was submitted to an advisory group of scholars to ensure the highest possible level of accuracy and fairness. Though expensive and immensely time-consuming, the process would have been yet more elaborate had the participants been convening in person, as they had in the past. At a 2019 event, NORC used a “concierge approach” to help participants get leave from jobs and find them child care.

America In One Room convened virtually at noon Eastern time on June 3. I followed a group of 12 or so who agreed to let an outsider eavesdrop. They began by introducing themselves: George, a retired police sergeant in Galveston; Brian, a nurse in Albuquerque; Larry, a retiree in San Diego; Amanda, a mother of three; Ean, an Afghan War veteran; Diana, who worked for FedEx. Despite the fine-grained sortition process, the group skewed old and white. The discussion was guided by Stanford’s Online Deliberation Platform, which features an AI-assisted virtual moderator that offers prompts to stimulate discussion, nudge a shy participant, curb an abusive speaker, move on to the next subject, and so on. The system, which makes it possible for thousands of people to deliberate simultaneously, is so persuasively human that Brian, the nurse, later told me that he had assumed the voice belonged to “Alice”—Alice Siu, a colleague of Fishkin’s who helped manage the event.

After the introductions, the group was shown a short video that recapitulated information they had received earlier about a series of proposed reforms. The video, like the literature, was exactingly balanced, with advocacy introduced by locutions like “some say….” Many of the participants were initially reluctant to speak, but soon about half a dozen were regularly using their one minute of allotted time and then responding to further comments. At first they rambled over topics close to their heart, like the Electoral College, which no one liked. Then they focused on the first topic, ranked-choice voting, which allows voters to list their top three or more candidates, eliminates the last-place finisher, and then redistributes votes to the remaining candidates until one emerges with a majority. The idea struck Amanda as dumb. “What if you don’t like that least-ranked candidate?” she asked. “Who would remarry their least-favorite person?” It wasn’t clear if she or many of the others had read the briefing literature. Larry, the San Diego retiree, who very clearly had, explained that voters could just choose one candidate if they wished. As they talked their way through the proposals, the group gravitated toward those that seemed most likely to reduce polarization, like nonpartisan primaries. “Let’s just vote the human being,” Brian said. They were flummoxed by unfamiliar ideas like fusion voting, which allows a candidate to run on more than one ticket and make common cause with other parties, and proportional representation, in which voters choose multiple winners for large, multimember districts. Both help third parties and at least aim to reduce polarization, and so should have been popular, but this wasn’t clear to the group.

After an hour of discussion, the virtual moderator signaled that time was up. Now the participants were to draw up questions they would like to pose to the experts and then to vote on which they preferred. After a 60-minute break, our very human moderator, public radio personality Ray Suarez, turned to political scientist Lee Drutman to talk about the growing polarization that the reforms were intended to address. Drutman is on record as a skeptic of some of the reforms under discussion, including ranked-choice voting, and a devotee of one of them—proportional representation. “We need five or six parties,” Drutman explained. “Break the binary.” Afterward Brian complained, not unreasonably, that he had looked for “teaching, not preaching.” Other experts followed. They were temperate and thoughtful. I did, however, note my own reaction: “Asleep by 2:10,” 40 minutes into the hour-long session. Deliberation asks a lot of the citizen.

In the afternoon, the group learned about and discussed some extremely wonky subjects, such as the vote compact, and some hot-button issues, like term limits for the Supreme Court. The participants appeared to be of one mind on almost everything—get rid of the Electoral College, impose term limits on the Court—though it was hard to be sure since several of them remained silent. One extremely weird thing happened: A new member, Doug, suddenly loomed up gigantically in his camera, his hair plastered across his forehead, his words slurred. “We have a President who doesn’t know his asshole from a hole in the ground,” Doug blurted out, apropos of nothing. The platform registered Doug’s use of a forbidden word and prompted the others to decide if he should be cut off. He should not, they voted. After another outburst, George wrote, “I believe Doug is intoxicated.” But the collective commitment to civility was so deep that Doug was tolerated as if he were the village eccentric.

Indeed, the most striking feature of the six hours of group deliberation was the palpable wish to maintain among themselves the bipartisan comity that so many of them said they yearned for in our public life. Many of the proposals felt so technocratic that the participants weren’t clear about which side, if either, they were likely to benefit. At times, however, the discussion veered toward the third rail of voter fraud. On day one Susan, a West Coast liberal, said, “I do believe our system is safe.” Amanda, a Southern conservative, retorted: “There’s no way. Our votes are not safe.” The following day, Larry, a long-ago political science major and now a careful student of politics, noted that “research has shown the incidence of fraud is minimal,” at which Brian asked how, then, it could be that 7,500 people had voted in a New Mexico city of 5,000, as he had heard. Nobody persuaded anyone, but neither did anyone lose their temper. The very fact of coming together to discuss weighty questions had apparently induced the group to speak softly and listen courteously. “Voter fraud is possible,” Amanda said at one point, “but I respect y’all’s opinion.”

The extensive questionnaire that all participants filled out afterward showed how deeply the discussions had changed their views, albeit on issues where few of them may have had deep-rooted convictions. Self-described Republicans found themselves far more receptive than they had been before to reforms designed to remove impediments to voting and registration, such as allowing online or Election Day registration. Republicans went from opposing to supporting restoring the right to vote to felons who had served their time. On the highly charged issue of transferring power over redistricting from state legislatures to nonpartisan commissions, members of both parties switched from negative to positive, but the change among Republicans, whose local political leaders have perfected the art of gerrymandering, was much larger. On other issues, like ranked-choice voting and proportional representation, partisans remained divided, with Democrats generally pro and Republicans con—but the latter moved significantly closer to the former.

Though the summary of the outcome drawn up by Stanford scholars emphasized that on several issues Democrats also drew closer to Republicans, conservatives would have good reason to fear that calm and reasonable deliberation has the effect of clarifying the merits of good-government reforms advanced by mainstream figures and of dissipating suspicions designed to demonize the other side. In short, bad for them. Astonishingly, the percentage of Republicans who pronounced themselves satisfied with American democracy rose from 19 percent to 50 percent. That is good news for democracy, but not necessarily for Donald Trump’s Republican Party.

Deliberative democracy is itself a reform, if a carefully circumscribed one, of the classical model of representative democracy. The early Federalists, above all John Adams and Alexander Hamilton, feared “the mob” and regarded electoral representation as a bulwark against the irrationality of the average citizen. The mass movements that cheered the rise of fascist governments in the 1930s confirmed those dark premonitions. In 1942, the Austrian economist Joseph Schumpeter, who had left Germany for the United States a decade earlier, argued in his immensely influential Capitalism, Socialism, and Democracy that democracy operates like a marketplace, with aspiring leaders competing for voters’ favor as companies do with their products. Voters choose, and then withdraw. Schumpeter effectively resurrected Rousseau’s gibe as a virtue: “The electoral mass,” he claimed, “is incapable of action other than a stampede.” Postwar political scientists characterized the disengagement of the average voter as an act of “rational ignorance.” Having agreed to cede active control over the political process, voters had little incentive to educate themselves on the subject.

The very different mass movements of the 1960s, deeply idealistic and less violent than those of the 1930s, seemed to belie the assumption that citizens were content simply to designate their proxies. Their disengagement from partisan politics could be construed as a sign of alienation from a nonresponsive system, as the radical students who drew up the 1962 Port Huron Statement argued. Younger, left-wing scholars trusted leaders less and ordinary people more than Schumpeter’s generation had. In her 1970 Participation and Democratic Theory, the British political theorist Carole Pateman argued—as de Tocqueville had—that citizenship in democracy was a learned and active skill. “Participation develops and fosters the very qualities necessary for it,” she wrote; “the more individuals participate the better able they become to do so.” In Strong Democracy, political scientist Benjamin Barber argued that participation made the people worthy of democracy. “Without participating in the common life that defines them and in the decision-making that shapes their social habitat,” Barber wrote, “women and men cannot become individuals.”

Yet the idea of replacing representation with direct participation presupposed a faith in the wisdom of the average citizen that struck more mainstream liberals as wildly utopian. In Political Liberalism, John Rawls argued that democratic legitimacy depended not on a transfer of power from representatives to citizens, but on acts of deliberation that ensured that policy decisions were based on what he called “an ideal of public reason.” Rawls, in turn, was building on the work of the German political philosopher Jürgen Habermas, who argued that free and widespread debate—in cafes as well as parliaments—ultimately produced outcomes whose legitimacy all could accept. Writing at the same moment, James Fishkin proposed a practical way of institutionalizing these informal modes of deliberation based on the model of the ancient republics. He reminded readers that while the Athenian Assembly had been open to all male citizens—that is, fully participatory—the Athenians had deployed a lottery system to choose the agenda-setting Council of 500 as well as juries that judged the work of the assembly.

Fishkin first got to put his theories into practice in 1996, when he was teaching at the University of Texas. State electric utilities asked him to hold a series of deliberative polls to determine how to provide electric power in various regions. Between 1996 and 1998, Fishkin convened eight groups, whose members strongly endorsed increasing the use of wind power. A subsequent study found that state utilities “changed their level of interest in and commitment to renewables and efficiency as a result of what they heard from customers.” Texas is now first in the country on wind power production. In 1996 Fishkin also held a National Issues Convention in which several of the Republican candidates for President addressed the participants. The event was co-sponsored and broadcast by PBS. The convention offered a modest proof of concept, since participants reported having changed their views on a number of subjects, such as the efficacy of foreign aid.

As the years passed, however, the Texas poll remained the only example of a political entity in the United States authorizing one of Fishkin’s deliberative polls and accepting its outcome. In What’s Next California, in 2011, and in the first iteration of America In One Room, in 2019, political scientists and issue experts set the agenda and drew their own inferences from the outcomes of the discussions. The exercise seemed to have stalled somewhere between social science experiment and democratic reform. Deliberative polling has nothing like the popular following that some of the reforms the participants debated in June enjoy—a fact that perplexes its inventor. Earlier this year, when I paid a visit to Fishkin, a thickly bearded and bespectacled 75-year-old, he said, “I’m amazed at how many people haven’t heard about it.” That might even include some readers of this article.

Even before the advent of deliberative polling, more rough-and-ready forms of deliberative or participatory democracy struck deep roots with citizens, though chiefly outside the United States. Since 1990, Brazil has held several hundred national and municipal “public policy conferences” that focus on social issues like the rights of minorities. The discussions begin in municipalities, where they are open to all citizens. Those participants choose delegates to state events who in turn elect representatives for a national deliberation. The public policy conferences completely bypass the stage of sortition; anyone can participate. Fishkin told me bluntly that self-selected processes offer no meaningful insight into what the larger public believes. But the idea speaks to something ordinary people in countries long under the thumb of authoritarian leaders may care about much more deeply: having a voice. The Brazilian model has spread to other nations in Latin America, including Mexico and Ecuador.

In Europe and East Asia, national governments have increasingly come to use versions of the deliberative poll as a democratic means of breaking political deadlock. In 2012 Ireland convened a 100-person “constitutional convention,” with two-thirds drawn randomly from the public and one-third from elected representatives. Of the eight issues the convention considered, by far the most explosive was same-sex marriage. After 79 percent of the group endorsed putting “marriage equality” on the ballot, Ireland’s right-of-center government came under pressure to hold a referendum. When the right to same-sex marriage easily passed in 2015—a virtual revolution of opinion in this once devoutly Catholic nation—the constitutional convention was widely credited with shaping public opinion. Virtually the same sequence of events occurred three years later on the yet more explosive subject of abortion—resounding endorsement of legalization by a “citizens’ assembly” modeled after the convention followed by approval of the nation in a referendum. Fishkin regards the citizens’ assembly as a “debilitated version” of his method, yet the Irish appear to have accorded it great legitimacy.

Political leaders elsewhere in Europe have impaneled assemblies to extricate themselves from trouble. In the aftermath of the “yellow vest” protests provoked by an abrupt increase in gasoline taxes to reduce gas consumption and pay for climate change legislation, French President Emmanuel Macron convened 150 randomly selected citizens to recommend policies to reduce carbon emissions. The group met seven times in 2019 and 2020 and forwarded 149 proposals to the president. In the end, however, Macron ignored the most far-reaching suggestions. Apparently he had regarded the poll as a means not so much to hear the voice of citizens as to demonstrate to them that he was doing so. In a postmortem, the participants awarded the government a 3 out of 10 for its response.

Fishkin’s Deliberative Democracy Lab has itself organized many mini-publics in Europe and Asia, always adhering strictly to Marquess of Queensberry rules. An initial poll in Mongolia in 2015 so persuaded the country’s leaders of the merits of the exercise that in 2017 the legislature passed a law requiring that a deliberative poll be held before the government considers a constitutional amendment. In 2003, President Roh Moo-hyun of South Korea proposed a deliberative poll on controversial plans to build a giant tunnel through a mountain. The effort failed, but the idea that deliberative polls could serve the interests of political leaders took hold. Chun Seok Kim, a member of the board of directors at Hankook Research, which organizes mini-publics in collaboration with Stanford, offered a telling explanation for the institutionalization of the practice: “All of these attempts were grounded in practical reasons rather than theoretical questions about democracy.”

One of the most remarkable vindications of the merits of the deliberative poll came in 2017, when South Korea’s president, Moon Jae-In, authorized one to determine whether his government should continue building two nuclear reactors. Moon had campaigned against nuclear power but agreed in advance to abide by the outcome. Political scientists and politicians publicly worried that a subject so technical would be beyond the capacity of the average citizen. Yet the deliberators, who had been selected according to Fishkin’s exacting standards, arrived at a complex split decision, voting both to approve the construction of two new reactors at the Shin-kori facility and to scale down nuclear power over time. Moon, true to his vow, accepted the outcome. Earlier this year, the South Korean parliament, deadlocked on a series of electoral reforms, authorized a poll on the subject. South Korea elects members of parliament through both single-member districts and proportional representation. After hearing arguments on both sides, the participants proved vastly more favorable to proportional representation than they had been in pre-deliberation surveys. The legislature has not yet acted, but Kim notes that the proceedings were broadcast live on Korea’s public TV service. “It is impossible,” he said, “for members of the National Assembly to just neglect the citizen voices and ignore the context.”

Why have political elites created space for citizen deliberation in Europe and East Asia, but not the United States? Larry Diamond, a scholar of democracy at Stanford and partner of Fishkin’s in the Deliberative Democracy Lab, points out that the polarization that deliberative polls have the capacity to help us overcome is precisely what is blocking the adoption of all sorts of democratic reforms, including this one. “There’s an appetite for reform,” he notes, “but no common vision of what that reform can be.” If Republicans think that making it easier to register or to vote, or putting an end to gerrymandering, or deferring to the outcome of a deliberative poll is bad for them, they’ll make sure it doesn’t happen. Some of the resistance, however, is bipartisan. American democracy, like baseball, is the very first, the most venerable, the most mythologized. Change feels like sacrilege. Baseball only modified some of its rules this past year when the game had become so slow and boring that it seemed to be shedding its mystique. The crisis in American democracy is just as grave, at least if you can accept that democracy matters as much as baseball; but so far, the growing sense of despair among fans hasn’t compelled management to act.

Deliberative polling in the United States, to reverse Kim’s terms, has been driven more by academic theories than by practical expedients. James Fishkin is a political scientist by training and temperament, and the leader of a movement more or less by accident. He is extremely protective of his invention; his reluctance to release the pure strain into the wild, where less exacting practitioners might relax rules on sortition or deliberation in order to save time or money, has probably limited its propagation. Yet the world has changed in ways that make deliberative democracy much more important than it was a generation ago, or even five years ago. In 2018, when he wrote Democracy When the People Are Thinking: Revitalizing Our Politics Through Public Deliberation, Fishkin was still preoccupied with the perfecting of public will. He expressed confidence that deliberative polling would enjoy widespread legitimacy “provided the citizen has a normative commitment to democracy.” That is taking for granted precisely what is now in dispute. Many Americans don’t have that normative commitment and won’t feel heard in an experiment carried out by Stanford political scientists. Finding out what the public would think under ideal circumstances of deliberation feels a lot less urgent than persuading Americans to deliberate with one another at all.

In 2019, Fishkin and his colleagues stumbled on a finding that had been largely latent until then: Deliberative polling has a “depolarizing” effect on participants. As in the National Issues Convention in 1996, a representative group was convened to discuss the hot-button issues of the upcoming election—the economy, health care, climate change, and foreign policy. The 526 participants met in a hotel outside of Dallas; they spent the weekend not only deliberating together but eating and drinking and schmoozing together. The effect was remarkable. Brian and Matt, from the group I watched, had both participated in the Dallas event. (Fishkin intentionally includes repeat participants in order to explore the effect of multiple deliberations.) When we talked afterward, both spoke almost rapturously of the sense of camaraderie they felt across all lines of background and ideology. “What I found is that we’re human beings first,” Brian told me. “We all want the same thing.”

A study of the pre- and post-deliberation surveys by Fishkin, Diamond, and colleagues concluded that on immigration, for example, deliberation on the most intensely contentious issues produced “massive changes” in opinion. Fishkin had seen this pattern often before. But what was new was the precipitous drop in “affective polarization”—how you feel about the other. On a 1 to 100 scale, Democrats rated Republicans 13 points higher than they had beforehand; Republicans rated Democrats 14 points higher. Those at the extreme ends of partisanship reported the largest increase in favorability ratings of the other. This past June’s version of America In One Room would confirm these findings: The fraction of voters who said, “I respect their point of view though it is different from mine” went from 57 to 75 percent, with similar increases among Democrats, Republicans, and independents.

The cause was no mystery: As the authors noted, the mid-century psychologist Gordon Allport had posited that contact between antagonistic groups could reduce polarization so long as it was conducted under conditions that foster “equal status between groups, common goals, cooperation, and institutional support.” America In One Room did just that. Yet a good deal of recent research in social psychology would not have predicted this hopeful outcome. The legal scholar Cass Sunstein has cited studies of juries to argue for what he calls a law of “group polarization,” in which deliberation produces a convergence around more extreme views than the average speaker held beforehand. Extremists typically exercise outsize influence on moderates. Yet Sunstein recognized that “the circumstances and nature of deliberation” shape that dynamic, and conceded that the design of deliberative polls may, in fact, produce convergence around the middle.

What’s more, the depolarization effect persists over time. Norman Bradburn of NORC told me that while the policy views of participants tend to revert once they get back home, change in views of the other remains. What’s more, Bradburn says, the Dallas event “changed their self-perception as citizens. It made them a little bit more independent of their group ideas and somewhat more attentive to information.” Participants became more likely to “vote for a policy reason” rather than because of vague personal feelings. Bradburn added however, “That’s based on the maximum treatment.”

The maximum treatment—that’s a problem. A single exemplary deliberative event, with all the elaborate trappings of sortition and minutely balanced policy input, can reveal what the public really thinks under ideal conditions. But if the purpose of the reform is not to produce a certain outcome but rather to give people access to the process, because doing so will change them, then you don’t need and can’t afford the maximum treatment. You need the clash of antagonistic views, but you don’t need representativeness. You need to release the mini-public into the wild. The automated platform makes it possible to do so, to extend a milder version of the treatment, online rather than in person, to a vast number of people, though inevitably under less rigidly controlled conditions. When I mentioned this paradox to Fishkin, he agreed that his invention had turned out to be “a potential cure” for polarization, but he wasn’t sure how to “deliver the medication.” Diamond, a more worldly figure who spent time in Iraq in 2004 seeking to nurture the seeds of democracy there, now sees in retrospect that deliberative polling “began with one purpose, quite noble and laudable, but it’s taken on a second purpose, in regard to the future of democracy, that may be more important than the first.”

Deliberative polling is like a drug that was developed to treat, say, migraines, but then turned out to slow the growth of cancer. Of course, the migraine treatment—showing what the public would believe under ideal conditions—still matters, at least in countries that have begun to institutionalize the mini-public. But in order to make inroads against the cancer of polarization, the medicine must be administered very differently. The only value in depolarizing the views of 526 carefully selected Americans is to demonstrate that it can work for everyone; otherwise, it’s a niche treatment. Everyone needs to have the opportunity to meet the other under the conditions Allport stipulated.

Fishkin actually dreamed of universalizing the experience long before he thought about polarization. In 2004, he and the legal thinker Bruce Ackerman proposed a “Deliberation Day,” to be held two weeks before major national elections, in which all voters would listen to a debate between the candidates, gather in groups of 15 to come up with questions, and then convene in 500-person assemblies to hear local proxies for the adversaries respond. It’s the kind of grassroots exercise of which de Tocqueville surely would have approved, but it presupposed an act of political will by the two parties so unlikely that even the authors described it as “realistic utopianism.” Indeed, if we wait for American political elites to embrace deliberative democracy, it will remain every bit as remote as proposals to abolish the Electoral College.

Fishkin now lists “scaling up” as his third objective, after perfecting the democratic will and countering polarization. The Deliberative Democracy Lab recently secured a grant from the Helena Foundation to work on such initiatives. The efforts to date have been modest. Fishkin and his colleagues have begun working with Close Up, a Washington, D.C.-based organization that promotes civic education in schools. Introducing deliberative polling into civics classes is a promising idea, but Mia Charity, president of Close Up, says that so far she has held only a few scattered events. Alice Siu told me that she has tried to enlist adults for deliberative events, but the effort hasn’t gotten off the ground owing to a low “take-up rate.” That may prove, as critics of participatory efforts typically argue, that most voters really don’t want to be more engaged in their democracy; they just want it to produce outcomes they like more. But what the experience in other countries shows is that citizens are prepared to set aside the time if they feel that the stakes matter. People might be quicker to volunteer if the subject were, say, whether to build a homeless shelter in their neighborhood. It also helps to pay people, as the Athenians did.

Larry Diamond has some practical suggestions. Diamond saw more evidence than I did of heated discussion and then rapprochement during the latest iteration of America In One Room. “It’s kind of stunning that this keeps happening,” he said to me. “We get more and more polarized, we get more and more embittered, but if you bring Americans together under good conditions, they don’t necessarily all agree, but they leave with almost a sense of relief that they’ve been able to have the conversation.” How, then, to diffuse that experience? Diamond envisions a series of local events that build on America’s unique network of civil society organizations. “You could,” he mused, “go into any community in the United States and bring together a diverse group to deliberate” under the aegis of the local Rotary Club or a coalition of churches. “You would ask, `What’s dividing us?’ They would define their own agenda. It could be abortion, pornography, minimum wage.” The organizers would dispense with the machinery of sortition but assure enough diversity to generate sufficient friction. Diamond has played a central role in preparing the briefing papers for Stanford’s deliberative polls and has practically killed himself with the effort to boil down the social science into bite-sized, ideology-free passages. He imagines a national clearinghouse that would take over the job by compiling background material on a vast range of subjects.

One could, in short, envision a world in which deliberative exercises were both more politically salient and far more widespread than they are now. The salience may, in fact, depend on the proliferation. A recent report from Carnegie Europe notes that while “more or less any strategy, document, or debate about the state of European democracy now makes a routine call for more citizen participation,” these initiatives have not yet proven “capable of reshaping democratic politics in a more far-reaching fashion.” The author, Richard Youngs, argues that deliberative democracy can never become a mass phenomenon so long as it is handed down from above, its legitimacy assumed rather than earned. Rather, he writes, political parties, social movements, civil society organizations, and the like must come to embrace deliberative exercises as important elements of their own decision-making. Only when the process has been embedded in politics and society, Youngs argues, will it begin to systematically shape national legislative outcomes.

In such a world, the distinction between “participatory” and “deliberative” democracy would dissolve, because widespread participation in deliberative exercises would compel political actors to take their outcomes seriously. In The Oxford Handbook of Deliberative Democracy, a 2018 compendium of essays by many of the leading scholars in the field as well as some of its chief critics, Stephen Elstub, a British political scientist, makes the case for “participatory deliberative democracy.” Elstub believes in the legitimating function of sortition-based assemblies but thinks that all citizens should have access to some form of deliberative exercise. Like Youngs, he suggests that we can let a thousand flowers bloom—mini-publics; referenda; deliberation in political parties, workplaces, and parliaments. Elstub observes that some leading theorists of deliberative democracy believe, like Schumpeter and his followers, that ordinary citizens don’t really want the burden of decision-making, and that democratic legitimacy does not rest on universal or even widespread political engagement.

Many of us, in fact, believe that. The spectacle of millions of apparently rational citizens endorsing bizarre conspiracy theories and questioning the outcome of obviously legitimate elections seems to have strengthened the argument for more elite control—more “professional vetting” of political candidates, as the scholars Jonathan Rauch and Ray La Raja put it—rather than more direct participation. After all, is it really true, as Carole Pateman argued half a century ago, that more participation would make people better able to participate? Is it even true that people yearn for more political agency? In his 1819 essay “The Liberty of Ancients Compared with That of Moderns,” the French political philosopher Benjamin Constant offered the classic formulation of the right to be left alone—a right cherished by liberals ever since. Constant observed that while for the citizens of the Roman republic “liberty” meant the right of full engagement in the political life of the state, “the aim of the moderns is the enjoyment of security in private pleasures, and they call liberty the guarantees accorded by institutions to these pleasures.” Brian and Matt and Larry would have concurred: Everyone in my group regarded a proposal to make voting mandatory as an intolerable infringement on personal autonomy.

Yet everybody in the group also liked deliberative polling. People want to be left alone—but they also want to have a voice. Angry people are angry about not being heard. The era of deference—to experts, to teachers, to elites of every kind—is over, for better and for worse. In our commercial life, participation has become the norm; we can hardly engage in a transaction without having our opinion of it solicited. Can we really wall off our political life from that demand to be listened to? If not, then we need to determine the conditions under which people would both volunteer for a more active role and behave in ways that do more good than harm. That is just the question James Fishkin has spent the last three decades thinking about. In 2018 he wrote that we need to find a third way between populism (deferring to the loudest voice) and technocracy (deferring to expertise). The Fishkin-style mini-public does that, but it isn’t the only way to foster civilized discourse and participatory decision-making. Let’s try everything.