How many bizarre electoral outcomes does it take to shake our faith in democracy? Apparently, one is enough. Even before the presidential election last November, New York magazine’s Andrew Sullivan was fretting about so-called “hyperdemocracies,” in which people have an unquenchable thirst for equality and refuse to accept limits on the popular will. This summer, writing in the Los Angeles Times, James Kirchik concluded from the Brexit vote and the recent snap election in the United Kingdom that “our duly elected representatives” should have the courage to ignore “the uninformed opinion of the masses.”

Social scientists as well as political philosophers have been ready to second that opinion: Ever since Philip Converse’s pioneering studies in the 1950s, American political scientists have amassed a wealth of evidence confirming just how little voters know—and just how incoherent or plain illogical their political choices can be. This empirical work has run in tandem with that of political theorists less worried about voters’ ignorance than about their intolerance. John Rawls, still the most influential liberal philosopher in the United States today, argued that for a liberal polity to be stable, “unreasonable” citizens would have to be “contained” just like “war and disease.”

One might think that the obvious answer to voter ignorance is education, and the answer to the more specific quandary of voter unreasonableness is perhaps some sort of civic reeducation. But the political philosopher Jason Brennan is having none of this argument. In his book Against Democracy, Brennan points to evidence that the generally rising education levels in the United States have not made citizens more knowledgeable about politics. Like many social scientists, he thinks there’s a simple explanation for why Americans remain so clueless: Ignorance is a rational choice. Since one’s individual vote has an infinitesimally small chance of actually deciding the outcome of an election, it simply isn’t worth the time and effort to bone up on policy basics—or even read the Constitution. As Brennan argues in another of his writings on the subject, democracy’s “essential flaw” is that it spreads power out widely, thereby removing any incentive for individual voters to use their own, more diffuse power wisely.

Of course, some voters seem happy to participate in the process nevertheless; they still display a passionate interest in political, and even constitutional, matters. But most of them, according to Brennan, treat politics like a spectator sport or, even worse, a brutal contact sport. The completely ignorant are what he calls “hobbits”; by contrast, those who root for one team and hate the other are “hooligans.” For hooligans, a little knowledge is a dangerous thing: They understand enough to be deeply convinced that their team is on the side of the angels and that the other side are devils (witness how 40 percent of Trump supporters in Florida thought that Hillary Clinton had literally emerged from hell). But they are incapable of rationally weighing policy options or even comprehending their own basic interests. For the hooligans, it’s all about identity.

In Brennan’s peculiar typology, there is a third species of voter, which he calls “vulcans.” Vulcans coolly examine the evidence and then form their political judgments accordingly. Needless to say, they’re a minuscule minority, and, less obviously, they cannot be upheld as anything resembling role models: After all, most people simply don’t have the leisure to become vulcans—as Oscar Wilde once said of socialism, it takes too many evenings. More worryingly still, hobbits are so ignorant and ill-informed that they can’t recognize the superior reasoning of the vulcans and take their cues from them. For these low-information voters, Brennan asserts, certified experts are more or less on the same level as the far-right radio host and conspiracy theorist Alex Jones when it comes to professional reputation and credibility.

Brennan’s answer to this “essential flaw” of democracy is as drastic as it is seemingly logical: Restrict the franchise on the grounds of some basic test of knowledge. Following the philosopher David Estlund, Brennan dubs this “epistocracy”—rule by the knowledgeable—which has a long and disturbingly distinguished history in Western political thought: Plato advocated it, as did, in much attenuated form, a 19th-century liberal like John Stuart Mill, who wanted university graduates to have additional votes. (He got his wish: In the UK, “university constituencies”—which allowed Oxbridge alumni to cast two ballots—were only abolished in 1950, by a Labour government.)

Although Republicans remain busy in the United States restricting the franchise on the basis of essentially fraudulent claims about “voter fraud,” neither the Republican nor the Democratic party openly advocates for the exclusion of voters on the grounds of incompetence—a notion that is still taboo in contemporary democracies. Even so, excluding children and the mentally incapacitated from casting a ballot is a largely uncontroversial practice across these same democracies; and in many American states, felons are disenfranchised for the rest of their lives.

Of course, Brennan is well aware that restricting the franchise on the basis of tests was long deployed in the United States for the purpose of racial discrimination. But he wants us to ignore its past uses and marshals a range of abstract arguments as to why epistocracy deserves serious consideration. For one thing, he claims, democratic citizenship is not like fandom for a sports club. Even if one’s personal vote is unlikely to make a difference, letting lots of ignorant people cast a ballot for their favorite “team” has dramatic consequences: their choices empower lawmakers to pass legislation which ultimately authorizes police officers to coerce anyone who is not willing to comply with the “team’s” ideals. Here “fandom” is sure to result in violence. Brennan also insists that the varying degrees of ignorance and prejudice displayed by different citizens don’t somehow end up neutralizing each other. Ignoramuses, he says, don’t vote randomly; instead, they will empower those who support irrational economic policies or seek to trample our civil liberties. And if all that weren’t bad enough, mass-democratic politics turns people into “civic enemies” of one another, in Brennan’s view. Rejecting the pious notion that political participation tends to “educate, enlighten, and ennoble,” Brennan argues that more political involvement is likely to turn hobbits into hooligans. One need only think about the polarization in the United States today to see his point that “politics gives us genuine grounds to hate each other.” Or so it might seem.

An obvious rejoinder to Brennan’s call for disenfranchisement is that any meaningful concept of democracy is predicated not just on an ideal of freedom, but on a notion of political equality. Epistocrats have to reckon with the fact that they are advocating for a basic inequality in our fundamental rights. Brennan thinks this isn’t really a problem: What are sometimes called the “expressive” functions of democracy, he argues, are massively overrated. If people want to express themselves, they should write a poem instead of heading to the ballot box; and if the state wants to communicate to its citizens that it cares about them, it should ensure decent policy outcomes—as opposed to formal political equality combined with enormous social injustice, as is the case in most democracies today. Epistocracy, in other words, would be paternalistic—but, so Brennan claims, even the worst-off today would benefit from seeing it instituted.

As other critics have pointed out, Brennan is long on identifying the flaws of the actually existing democracies today and short on the institutional details of even an ideal epistocracy. Who would create the test to establish who gets to vote and who doesn’t? Who would be truly competent to judge other people’s competency—and, for that matter, what are the measures of competence? How would the transition to epistocracy be engineered and justified? Should people actually have to vote to disenfranchise themselves? And if the resulting policies were still judged insufficiently rational, would the franchise have to be ever more restricted? After all, even if the ignorant weren’t able to vote anymore, they could still use their rights of free speech and free assembly to advocate for what the 44th president of the United States famously called “stupid shit.” Would the end point then be what social scientists have referred to as “bureaucratic authoritarianism”—which is to say, rule by the few, purportedly in the name of the collective well-being? Reading Brennan in Beijing, one would think, must provide a boost to the leaders of the self-declared People’s Republic.

Apart from practical questions, there is also the issue of whether one of the most basic conditions of democracy—the universal franchise—is quite so easily waved away. Can disenfranchisement really be done in a “clean” way that doesn’t jeopardize equal treatment and status in other spheres? In effect, epistocracy would amount to a kind of political quarantine: We, the knowledge-bearers, need to protect the country from “them.” It is hard to see how the people in the latter category would not be effectively stigmatized, even if the disenfranchisement were somehow accomplished benignly. Symbolic politics is never merely symbolic; perceptions have consequences. To treat disenfranchisement as if it were primarily a question of “self-esteem,” or hurting citizens’ feelings, is frivolous at best.

Brennan himself seems uneasy about this possibility and hence feels compelled to suggest a way for everyone to affirm epistocracy and thereby smuggle some notion of equality back into his system. He writes that “it’s not difficult to imagine an epistocratic society in which everyone regards one another as having equal status. Perhaps they endorse epistocracy because they think it tends to produce more equitable results, and for that reason think their commitment to epistocracy expresses a commitment to equality.” But voluntary consent to epistocracy would presumably undermine the very rationale for wanting such an elitist scheme in the first place: If citizens were sufficiently sophisticated to understand the reasons for their own exclusion from electoral politics and were willing to consent to such an indignity, they would likely also understand the basic pros and cons of candidates and policies, which would qualify them as voters in the first place.

So epistocracy doesn’t seem to square with democracy’s intrinsic value in affirming equality and giving everyone a chance to advance his or her interests, as idealized as that notion might be. But Brennan’s advocacy is also based on a curiously unrealistic way of understanding how democracies fail on an instrumental level. In his view, modern democracies are broken because they don’t achieve rational ends. But the democratic process isn’t really about individual voters making rational or irrational choices—a perspective that can only ever yield variations on H.L. Mencken’s quip that “democracy is a pathetic belief in the collective wisdom of individual ignorance.” Nor is democracy about rationality versus irrationality; only technocrats think that way about politics. Rather, it is a system that allows leaders to gain power on the basis of their claims to represent different ideas, interests, and identities. Obviously, the latter are ultimately individual, but they end up being combined into larger political forces. These are not objectively given; they are formed in a dynamic, open-ended process of struggle over who and what gets to be represented, how, and by whom. Democratic representation is therefore neither about finding the one right policy answer nor about the mechanical reproduction of already existing interests and identities. New blocs of interests and forms of identification are themselves an outcome of politics.

To treat democracy only in terms of the individual voter and his or her competence is like having microeconomics without any macro. Such an approach fails to see that the quality of democracy depends crucially on the space between individual voters and the policy decisions that bind them together. Different institutions—from the courts to the media to the rules of election campaigns—make an enormous difference here. As the American political scientist E.E. Schattschneider put it, “the problem is not how 180 million Aristotles can run a democracy, but how can we organize a community of 180 million ordinary people so that it remains sensitive to our needs. This is a problem of leadership, organization, alternatives, and systems of responsibility and confidence.”

The terms “leadership” and “responsibility”—as horribly elitist as they may sound to some—should remind us that ignorance and misinformation are not just facts of life; they are often also the result of fully conscious decisions by political elites who would like to protect and extend their interests. Americans are polarized and often treat their fellow citizens as “civic enemies” not because democracy has an inherently “gladiatorial” nature and brings out the worst in us; nor is it because the country is naturally divided into red states and blue states. Polarization is a project that confers great political and economic benefits; unreasonableness can be big business. Gerrymandering, doing away with the Fairness Doctrine in the media, and inviting Alex Jones on prime-time TV are not choices dictated by democracy as such.

Instead of blaming the people for their irrationality, we might ask instead about the “leadership” of a figure like David Cameron, who called a referendum on Britain’s membership in the European Union purely to pacify critics in his own party. Rather than citing endless statistics about how shockingly little Americans know about politics, we might also wonder about the “responsibility” of a figure like Senate majority leader Mitch McConnell, who can reshape the entire political game by deciding that his own and his party’s interests would be best served by maintaining a state of perpetual conflict, and never cooperation, with the Democrats.

True, what pundits have described as a recent, supposedly unstoppable “wave” of populism might confirm the kinds of worries about democracy that Brennan and plenty of journalists articulate. But it’s important to remember that a figure like Nigel Farage, the former leader of the populist UK Independence Party, did not bring about Brexit all by himself; he needed the backing of very established figures from the Conservative Party like Michael Gove. It was Gove, after all, who, in the face of warnings about Brexit by many experts, announced that “the people of this country have had enough of experts.” The irony was that Gove himself clearly spoke with the authority of an expert: He has always been seen as one of the Tories’ most prominent “intellectuals.” It took nothing less than an expert to convince people that claims of expertise are overrated.

Trump’s victory, in turn, is not best understood as a “revolt of the masses” driven primarily by an undereducated, racist white working class. Rather, it was a triumph of hyperpartisanship: Trump needed the blessing of not-exactly-grassroots populist figures like New Jersey Governor Chris Christie, former New York City mayor Rudy Giuliani, and former House speaker Newt Gingrich. When the latter told a CNN anchor last summer that he didn’t trust statistics on crime but believed instead in what people felt, he did the same trick that Gove had performed in the UK: Whatever else one thinks about Gingrich, he is considered an intellectual of sorts among American conservatives. So, as in Britain, it took an established expert to devalue the importance of expertise.

Once Trump had the blessing of Republican heavyweights, he became, above all else, the candidate of a very established party—and 90 percent of self-identified Republicans gave him their vote. Moreover, many of them explicitly registered their doubts about the candidate’s competence—but in a Mitch McConnell–esque world in which one side refuses to recognize the other side’s legitimacy, it was inconceivable for these voters to cast a ballot for Hillary Clinton. The issue is not so much that, as Brennan puts it, hating is fun, and more that once the deed of demonization has been done, it is difficult to undo in citizens’ minds.

Democratic politics can be an unsightly spectacle. Even idealists will sometimes be tempted to agree that the best argument against it is a five-minute conversation with the average voter. But it isn’t an accident that democracy is also the system that provides the most room for freedom, however imperfectly, and does the best job of protecting basic rights. As the economist Amartya Sen famously demonstrated, democracy is the regime that avoids famines; and as the political philosopher Thomas Christiano has emphasized, democratic institutions still offer the best means for the most vulnerable members of the population to secure and advance their interests.

Above all, when things really go wrong, democracy is the system that allows people to throw the bastards out—and contrary to Brennan’s suggestion, “knowing whether the bastards are doing a bad job” doesn’t require “a tremendous amount of social scientific knowledge.” As the political scientist Martin Gilens has observed, many voters have a good enough sense of how specific politicians have performed and, in fact, are able to pay attention to the cues that elected officials—as well as pundits and rival candidates—offer them. Of course, this doesn’t mean that the problems identified by Brennan are entirely fictitious; but many of them have more to do with the conduct of particular elites than they do with an ignorant or uneducated demos. Giving already powerful people even more power by restricting the influence of the most vulnerable seems an odd way of addressing them.