It’s been a rough ten years for democracy. It isn’t just that Donald Trump was elected against the wishes of the majority of American voters, and that he then tried to overturn an election that he lost. Nor is it even that democracy is challenged elsewhere in the world, as China doubles down on single-leader rule and fragile democracies like Turkey and Hungary become more despotic. It’s that democracy’s sense of manifest destiny is gone.
Three decades ago, it seemed plausible that the despots were embattled holdouts, desperately trying to stem the inevitable tide of democracy. Now, they appear stronger. Many democratically elected politicians and their supporters seem to long for a future without democratic competition, in which the right people (i.e., they) are permanently in charge, and their enemies are marginalized or eliminated. Some right-wing intellectuals provide ammunition for the anti-democrats, claiming that democracy can’t work because citizens are just too biased and ignorant. They argue that democracy should be shrunk down or even replaced by new systems of rule, where the intelligent and knowledgeable (i.e., those who believe in neoclassical economics and efficient markets) would be privileged over those too foolish and uninformed to understand their own best interests.
If democracy is to do more than survive—if it is to flourish—it needs to change. The period of its apparent greatest success was also when the rot set in. When the citizens and leaders of seemingly stable democracies took that stability for granted, they mostly ignored democracy’s suppurating underbelly: the systematic economic inequalities, the groups that consistently lost out under it, and the many opportunities that it offered to game the system. Many social scientists took its benefits for granted, too. Some offered abstract justifications for democracy, which tended to be based on unrealistic claims about how human beings think and act. Most just assumed that democracy would somehow keep itself on track.
Fixing democracy will require a myriad of reforms. Just in the United States, this includes preventing gerrymandering, getting rid of the filibuster, guaranteeing voting rights, and constraining the power of an anti-democratic Supreme Court. But to ensure that such reforms add up to a healthier and more sustainable form of democracy, we need something more. Instead of democratic triumphalism, we need to understand when collective decision-making works well and when it works badly. We can’t base this understanding on idealistic theories of how democracy is always wonderful, or how human beings are always responsible, knowledgeable citizens. Instead, we need to start from the psychology of real humans, and build up toward institutions that actually work and are likely to be stable over time, even if citizens are flawed. In short, we need to move away from the assumption that democracy is best, to an understanding of how to make it better.
What is surprising—even paradoxical—is that such an understanding can be built on the kinds of scientific results that self-described intellectual opponents of democracy say ought to make democracy impossible. New research in cognitive science shows that even if individual human beings are biased in favor of their own side and against others, they can still make good collective decisions. It is an unfortunate truth of the human condition that we believe our own bullshit, even when it is ridiculous. But we are also pretty good at detecting other people’s bullshit and enjoy telling them so, especially when we disagree with them. That is why group argument can lead to better outcomes than you might expect.
Under the right circumstances, people can actually correct for each other’s biases. Figuring out what those circumstances are, and how to reproduce them, can help us move past democratic complacency and pessimism. We can start to understand when democracy works and how, guiding reforms so that they work better and keep working. What might be called “no-bullshit democracy” would be a new way of structuring democratic disagreement that would use human argumentativeness as a rapid-growth fertilizer. But first we need to sluice away the bullshit that is being liberally spread around by anti-democratic thinkers.
A few prominent libertarian critics say they have scientific evidence that democracy fails on its own terms. Jason Brennan, a philosopher at Georgetown’s business school, believes that nearly all citizens are either naïfs or knaves, ignorant “hobbits” who don’t really know or care about politics, or fanatical “hooligans” who care about politics all too much. He deplores Americans’ ignorance of basic political facts but suggests that it would be a mistake to educate them more by getting them more engaged with politics. In his description, human beings fall victim to a whole host of biases. They don’t think well, ignore or misinterpret inconvenient evidence, become emotionally attached to political parties, succumb to peer pressure, and in general make terrible citizens. When hobbits start caring about politics, they’re likely to become hooligans, transforming dull apathy into ferocious partisan zeal.
The solution, Brennan believes, is to replace democracy with “epistocracy”—the rule of those who know better. Those who are more knowledgeable might have their votes count for more, or those who fail a qualifying exam (which might test their knowledge of “introductory microeconomics and introductory political science”) would be barred from voting. Perhaps, he says, political knowledge is “negatively correlated” with being Black and “strongly negatively correlated” with being a woman. Still, that is a bullet that Brennan is prepared to bite. He thinks that the disadvantaged are much more likely than others to be mistaken about what they really need.
Brennan wants epistocracy to rely on the judgment of people who can come closer to thinking like “Vulcans,” displaying a Spock-like ability to reason clearly and dispassionately on the basis of good evidence. Brennan implies, as you might expect, that he himself is among that elite minority. As he generously informs his readers, he is a “named professor…at an elite research university, with a Ph.D. from the top-ranked political philosophy program in the English-speaking world.” Hence, it should be no surprise that he has “superior political judgment on a great many political matters compared to many of [his] fellow citizens.”
Brennan’s ideas have received lengthy discussion in publications like The New Yorker and Vox. His most prominent book, Against Democracy, has been translated into multiple languages. Bryan Caplan, a George Mason University economics professor, is somewhat less well known, and lacks Brennan’s penchant for obnoxious trolling: While he sometimes makes startling and implausible claims, he appears to be reasoning sincerely from libertarian first principles, rather than deliberately setting out to provoke. Caplan maintains that “rule by demagogues…is the natural condition of democracy.” Like Brennan, he retails endless research findings, all intended to show that voters are ignorant, that bias is endemic, and that democracy, unlike markets, provides little incentive to people to get it right.
Brennan and Caplan have worked to revive old conservative and libertarian arguments against democracy—and to invent new ones. When Brennan argues that the democratic electorate is “systematically incompetent,” and says that if “the facts turn out the right way, some people ought not have the right to vote, or ought to have weaker voting rights than others,” he is scoring the tune for an enthusiastic partisan choir. National Review commentators have used closely similar arguments to defend Republican state legislatures that are trying to make voting harder. After all, if most people are biased fools, isn’t it better that they not vote?
At first glance, Brennan and Caplan look as though they want to replace democracy with very different things. Brennan wants the rule of the clever, while Caplan wants free markets. But this is more a difference of emphasis than a true disagreement. In Against Democracy, Brennan suggests that Marxists and social democrats could be Vulcans (and that most libertarians are hooligans). Still, he elsewhere advocates the rule of those who share libertarians’ enthusiasm for free-market mechanisms such as Uber and Lyft’s jacking up prices when demand is high. He abruptly ended his lively career as a public Twitter commentator when people got outraged at his apparently sincere argument that “[a]nyone who opposes surge pricing should be disenfranchised. That’s how we should decide who decides in epistocracy.”
Caplan wants the rule of markets themselves, on the notion that economists want them and they know best. He does acknowledge that economists aren’t infallible, but he reckons we can measure citizens’ bias by seeing how much they disagree with expert economists. This assumption allows him to invent cognitive blind spots hitherto unknown to psychological science, such as “antimarket bias.” Such biases make it plainly impossible to convince ordinary people of the (supposedly) commonsense truth that markets do far better than democracies, because markets force people to think clearly about costs and tradeoffs.
But just as Brennan shades from epistocracy to markets, Caplan extrapolates from markets to epistocracy. He believes that there is “much to be said” for giving extra voting power to educated voters, “since [they] think more like economists.” Both Brennan and Caplan agree that they would greatly prefer a world in which democracy was limited in favor of the rule of markets and the economically enlightened.
Brennan and Caplan aren’t wrong to say that human bias is ubiquitous. When Brennan, for example, quotes the claim that “[r]easoning was not designed to pursue the truth. Reasoning was designed by evolution to help us win arguments,” we partly agree. It would be weird if we didn’t, since (as he notes) those two sentences are a limited summary of research one of us (Hugo Mercier) has done with Dan Sperber. What is terribly wrong is Brennan’s assumption that this and other findings from cognitive psychology mean that it’s a bad idea to give everyone the vote.
The truth is that it isn’t just ordinary voters who are biased. Experts, including Brennan and Caplan (and for that matter ourselves), can be at least as enthusiastic as ordinary citizens to grab at ideologically convenient factoids and ignore or explain away inconvenient evidence. That, unfortunately, is why Brennan and Caplan’s books do a better job displaying the faults of human reasoning than explaining them.
If Brennan had read Mercier and Sperber’s research carefully, rather than relying on secondhand summaries, he might have come to different conclusions (he also might not have—as he himself notes, bias may lead people to wave away contradictory facts and findings). What Mercier and Sperber show is that under the right collective circumstances, humans can do a pretty good job countering individual bias—and even harnessing it to useful purpose. Groups can make better decisions than individuals on their own, by allowing humans to correct each other’s bullshit and counter each other’s biases. We explain this in a new open access article in the American Political Science Review, where we discuss other ways in which democracy skeptics either miss or misunderstand much of the psychological research that they rely on to discredit democracy.
For example: Brennan makes a big deal out of Solomon Asch’s famous “conformity experiments” from the early 1950s, which put individuals into a large group of people who had been instructed to say that obviously false things were true. Some of these individuals then gave into the majority, leading Brennan to argue that “Human beings are wired not to seek truth and justice but to seek consensus” and “cower before uniform opinion.”
That is not, however, what Asch himself concluded when he carried out the experiment. He found that people yielded only one-third of the time to the group consensus, sticking to their opinion in two-thirds of the trials, and concluded that people were mostly able to think independently, even under heavy social pressure. Moreover, people practically never yielded when their opinions were kept private, suggesting that they still had the ability to discern the true from the false. As other psychologists have noted, it is peculiar that “many accounts of Asch’s work draw from it the very assertions he was intending to refute.”
Brennan and Caplan also cite Brendan Nyhan and Jason Reifler’s political science research on the “backfire effect.” This effect posits that people don’t just fail to change their minds when they encounter contradictory evidence, but double down even more strongly on their beliefs. It’s not surprising that Brennan and Caplan like scientific research that proves what they already wanted to believe. In fairness, they couldn’t have predicted that evidence for the backfire effect would turn out to be the very rare exception, rather than the rule, when further and bigger studies were done.
More broadly, skeptics like Brennan and Caplan have a sweet tooth for scientific findings that seem to support Team Libertarian. That means that their withering skepticism of democracy goes hand in hand with a benign and avuncular indulgence toward the alternative systems that they would like to displace it. For example, Brennan’s arguments for epistocracy fall flat unless highly educated individuals, who know how to reason, are less prone to bias than ordinary citizens. But the sad truth is that self-proclaimed Vulcans aren’t any better able to think clearly than the rest of us. The Mercier and Sperber research that Brennan carelessly invokes says that humans reason in order to win arguments with each other, rather than to understand the world. If that is true, then people who are better at reasoning don’t necessarily understand the world any more clearly than those who reason less well. Instead, they are just better at crafting superficially plausible arguments for why everyone ought to agree with them. People who are enormously impressed by their own cleverness risk getting high on their own supply.
It isn’t just Mercier and Sperber who say this. Brennan enthuses at length about the research of two well-known political scientists, Chris Achen and Larry Bartels, who show how the textbook arguments for democracy don’t work. He seems to have missed the implications of the part where Achen and Bartels stress that “the highly educated…have gone astray in their moral and political judgments as often as anyone else,” such that they aren’t writing about “the political misjudgments of people with modest educations…[but] the conceptual limitations of human beings—including the authors of this book and its readers.”
Caplan, like Brennan, displays what critics might see as notable biases of his own. He thinks that people who accuse economists like himself of “market fundamentalism” are just blind to economists’ “open mindedness.” Indeed, they are accusing economists of the “cognitive misdeeds” they themselves are guilty of. Like creationists attacking evolutionary theory, “democracy fundamentalists” hate on economists because they feel scientifically inferior to their opponents. Those who are “staunchly pro-market,” in contrast, aren’t dogmatic. They have simply noticed that the other side’s arguments mistake “theoretical possibility for empirical likelihood.”
Maybe Caplan is right to think that the people who disagree with his views are just jealous might-as-well-be creationists. Who knows? Still, there is a notable shift of standards and evidence when Caplan briefly stops talking about democratic bias to ask whether there’s bias in market thinking too. He is convinced that the scientific evidence comprehensively demonstrates democratic irrationality, but when it comes to research on market irrationality, Caplan is all in favor of “nuanced” interpretations that acknowledge “experiments’ limitations.” Psychological experiments don’t disprove the benefits of markets, because “[i]t just takes time for incentives to work their magic.” In Caplan’s view, we can rely on “everyday experience” to tell us that pro-market economists are right in their understanding of human behavior.
It would be unfair to say that Brennan and Caplan are completely misguided. Some of the secondary evidence they have gathered compellingly demonstrates problems with democracy, and Caplan has some genuinely interesting ideas about how people’s unselfishness expands the range of democratic performance, making both good and bad outcomes more likely.
Moreover, democracy—in the United States and other countries—isn’t doing a particularly good job right now. Demagogues like Donald Trump are supported by many citizens, while online social media seem purposely designed to allow people’s political biases to feed into conflagration. But Brennan and Caplan’s own biases seem quite as serious as those they deplore in ordinary citizens. Would-be experts too are liable to overvalue evidence that supports their priors and to ignore or discount evidence that contradicts them.
Indeed, Brennan and Caplan’s preferred free-market technocracy has no obvious claim to be better than democracy, and could plausibly be worse. If this has been a bad decade for democracy, it has been a terrible 15 years for the notion that economic experts typically get it right, starting with the financial crisis and getting worse from there.
Trump’s political success is in part due to the immiserating consequences of open trade with China for the Rust Belt. In a recent path-breaking book, the sociologist Elizabeth Popp Berman has documented how the “economic style of thinking” that Brennan and Caplan emphasize reshaped politics to make it less democratic, and less responsive to people’s expressed needs. Policy debate has shriveled into a series of fights between hard-right anti-tax fundamentalists and centrist economic technocrats. We see the results around us today in the United States. What we have isn’t Singapore, that dream republic of epistocrats and free-marketers—it’s a system that is near political collapse. Brennan and Caplan’s apparent program for reform—we must market harder, comrades, if we are to be true epistocrats!—is a recommendation for disaster.
We need both better debates about democracy and better insights about how to fix it. The two should go together. Both fights over ideas and democratic fights over policy are competitive. People (and this article demonstrates that we are not immune) take glee in pointing out when those on the other side use evidence sloppily or make bad arguments. That can lead to vicious spirals of tu quoque. But results from cognitive psychology demonstrate that argumentative discussions can also be harnessed for a kind of excoriating mutual improvement.
That is the actual message behind the Mercier and Sperber results that Brennan gets wrong. Mercier and Sperber build on their own work and others’ to show how discussion can improve competing arguments despite, or sometimes even because of, the biases on both sides, arguing for what they call an “interactionist” view of human reasoning. A skeptical—indeed “no bullshit”—account of how humans think as individuals can lead to some cautiously optimistic conclusions about how we can better think together.
This works roughly as follows. People are hopelessly biased when it comes to their own views and how to defend them. We trust our intuitions and, when we start to reason, instead of objectively assessing the pros and cons of various views, we spontaneously pile up arguments supporting those intuitions—whether they involve what car to buy or whom to vote for. This confirmation bias (which psychologists more accurately call “myside” bias) has been demonstrated in hundreds of experiments. Even when people are given time and incentives to reason objectively, they fail to consider that they might be wrong, and to critically assess the often shallow arguments they accumulate to defend their views. When people produce arguments, they are both biased and lazy. The result is that reasoning on one’s own or with like-minded peers often leads nowhere good.
So far, this is similar to the claims that Brennan, Caplan, and their peers in the commentariat rely on. The crucial difference is that people are much sharper in evaluating the flaws and bullshit in other people’s reasoning. A host of experiments suggest that people are good at spotting the holes in what other people say. One experiment even found that people were less inclined to be convinced by their own arguments when they thought that these arguments had been made by someone else. And people aren’t just critical of others for the sake of disagreement. We can be harsh and demanding—but also quite objective. The evidence suggests that we can spot a good argument made by someone else, even when we disagree with it or loathe the viewpoint that it supports. It may be heartbreaking when the worst person you know makes a good point, but the experimental research suggests that you can still acknowledge it.
This means that bias can actually be useful up to a point. When people argue with even a minimum of good faith, they both produce arguments and evaluate the arguments of the other person. Myside bias means that everyone will produce arguments for their own side. These arguments will still be biased and lazy—but bad arguments will get shot down by the other side, while good arguments may prove annoyingly convincing.
Under the right circumstances, good ideas and reasonable suggestions are more likely to be adopted through collective disagreement, even when individuals are very biased. Deep group discussion is hard when the group gets big; it’s hard to have proper conversations with more than five people. Trust and some sense of common purpose are important—people should be committed to solving a shared problem, or discovering a common truth—but equally, there should be disagreement over how to do it, or else there would be no grounds for argument. Under the right conditions, discussion groups perform better than their average members at a wide range of tasks, including mathematical problems, medical diagnoses, trivia, and judicial decisions. If people’s answers are aggregated in the right way, discussion groups can even do better than the best individual member, and can beat other means of aggregating opinion, such as voting.
These ideas can have practical consequences. Daniel Kahneman, one of the great scholars of human bias, describes how he began “adversarial collaboration” on science with people he sharply disagreed with—beginning with his wife, Anne. He and colleagues would carry out research together on topics that they didn’t agree on, but they would try to agree in advance about how to test who was right. They were forced to sharpen their own arguments and, by Kahneman’s account, ended up being more friendly with one other than would otherwise have been likely. A lot of bullshit got eliminated as the adversaries were forced to acknowledge one other’s criticisms.
Adversarial collaboration works in small groups, but it requires a lot of time, resources, and commitment. Less demanding forms of argument can still have benefits. For starters, when both sides are willing to acknowledge—at least in principle—that their own positions have limits, and to engage with serious critics, argument is likely to proceed better.
Notably, there are libertarian skeptics of democracy who seem more willing than Brennan or Caplan to acknowledge limits to their claims and to entertain possible doubts. Ilya Somin, a law professor whom both Brennan and Caplan rely on (and who in turn is friendly to them), uses much the same kind of evidence, but he argues more moderately that “widespread public ignorance is a type of pollution,” so that “democracy might function better if its powers were more tightly limited.” While he says that voters aren’t good at understanding political issues, he agrees that they do “fairly well” at identifying which party is associated with which policy position. Somin argues that greater pessimism about democracy is warranted, and that we ought to move some decisions away from democracy, relying more on markets and individuals’ ability to vote with their feet. Still, he acknowledges that the benefits of markets and foot voting “are likely to vary from issue to issue, from nation to nation, and perhaps also from group to group.”
Somin’s apparent approach—continued conviction on perceived essentials, but openness to argument on the particulars—is a better basis for adversarial disagreement than trolling or ideological dogmatism. And sometimes people are willing to actually revise their opinions. When Nyhan and Reifler saw the evidence that contradicted their arguments about the backfire effect, they changed their minds. As Nyhan put it, “[i]t would be a terrible irony if evidence contradicting the backfire effect provoked me into doubling down on the backfire effect.” Both democracy skeptics and advocates of democracy would do well to acknowledge the evidence that complicates their views as well as the evidence that supports them. We are unlikely to be convinced that democracy should be tightly curtailed, any more than Somin is likely to start thinking that democracy should be radically extended across the economy. But we can learn from intelligent and honest skeptics, exactly because they are better able to see many of the weaknesses in our arguments than we can ourselves.
Counterintuitively, that means that the skeptics who want to dismantle democracy might be useful to democracy despite themselves, providing criticisms that could help improve it. Even those who are most trollish or dogmatic may have snouted up truffles of insight here or there. And those who maintain strong positions but are open to real debate may have trenchant criticisms that would never have occurred to democracy’s defenders but may be valuable to them.
Democracy takes place on a much wider scale than scholarly spats and small group discussion. But interactionist ideas may still highlight ways to help fix democracy itself. After all, democracy is a system for structuring, containing, and ideally extracting value from disagreements, in which people with sharply different values and understandings of the common good contend over who is to be in charge for a set period of time. It is also a system that various actors are attempting to pervert with an endless onslaught of bullshit, in the United States and many other parts of the world. Many people blame social media for this. Others blame media outlets like Fox News, which don’t encourage contrary arguments. But as cognitive psychologists like Mercier and Sperber show, other ways of shaping discussion can help to moderate bullshit and push us to grudgingly acknowledge criticisms.
It is challenging to do useful disagreement at scale. Still, one surprising example of how it might work is Wikipedia. Few pay attention to the inner workings of Wikipedia, which is run by legions of volunteer editors who work to write, update, and improve posts, without any financial reward. Those who do pay attention often point to the online encyclopedia’s many problems—the lack of women editors, the painful internal bureaucracy, and the sometimes ridiculous fights over what Wikipedia entries on controversial topics should say.
But Wikipedia, like democracy, can produce value despite its gross imperfections, and sometimes indeed because of them. In 2013, the sociologist James Evans began to look for places where people could come together despite their political differences. He and his colleagues hoped that this would be true of science, but they discovered that Republicans and Democrats both seemed to gravitate toward the kinds of science that supported what they already wanted to believe. To their surprise, they found that Wikipedia, with its endless edit wars, was a much more productive space, and that the most politically disputed entries were, on average, the highest-quality ones. This wasn’t because bitter adversaries came to agree on the fundamental issues. It was because Wikipedia’s internal rules forced them to recognize each other’s best criticisms if they were to get anything written. In Evans’s words:
Editors working on a social issues page said, “We have to admit that the position that was echoed at the end of the argument was much stronger and balanced.” Did they begrudgingly come to that? They did, and that’s the key. If they too easily updated their opinion, then they wouldn’t have been motivated to find counter-factual and counter-data arguments that fuel that conversation.
Wikipedia, like adversarial collaboration, seems able to harness people’s desire to beat the other side; it helps staunch opponents reduce their bullshit and arrive, however grudgingly, at a better overall outcome.
We need to start experimenting with new institutions that might better harness disagreement, as adversarial collaboration does on a smaller scale and Wikipedia does on a wider one. Other countries show how people can come together over politically contentious issues. Ireland, for example, only decriminalized homosexuality in 1993. In 2015, it became the first country in the world to introduce marriage equality by popular referendum. Part of the reason for this remarkable and rapid transformation was a Constitutional Convention, where randomly selected citizens (and a smaller number of elected politicians) were appointed to debate how the Irish constitution should be changed. When the convention proposed that the constitution should protect marriage equality, it provided cover for politicians to propose a referendum on what had been until recently a wildly controversial proposal. Voters saw that other citizens like them had heard the arguments for and against, and they were swayed. Over 62 percent voted in favor of marriage equality.
And even in the United States, there are experiments that demonstrate how democracy can work better if it involves ordinary citizens. Between 2006 and 2008, political scientists Michael Neblo, Kevin Esterling, David Lazer, and William Minozzi worked with Democratic and Republican members of Congress to set up electronic town halls where they could talk to randomly selected citizens about controversial topics. The results were striking. Politicians and their constituents found greater agreement and more useful disagreement in a setting that emphasized straightforward back-and-forth communication over the retailing of talking points. Both the members of Congress and the citizens ended up happier (and the citizens who participated became more willing to vote). This team of political scientists is working with new people and in other countries to figure out how far this model can be pushed.
Such cases don’t mean that we should replace representative democracy and voting with wholesale direct democracy or rely on randomly chosen citizens to serve as legislators. Instead, they show how representative democracy can work better when politicians are able to learn from their constituents’ disagreement, both with their elected officials and among themselves. New institutions on their own may not be enough: While new forms of democracy, such as large-scale deliberation and referendums, might indeed make better use of citizens’ dispersed knowledge, they might also fall victim to the sorts of biases the libertarian critics caution against, or be vulnerable to manipulation by politicians. Healthy democratic politics requires not only institutional change, but healthy political movements and parties too. What we need is more experiments—many more—involving not just new institutions, but new ways of building groups too.
In one sense, the libertarian critics are right. Ordinary citizens are hopelessly biased. In another, they are terribly wrong. They themselves are biased, as are we, and as is everyone else. This does not doom democracy. Instead, it can help make it better. Biases—if harnessed properly rather than allowed to dominate—can be useful and even valuable. If we can figure out how to harness human psychology, with all its limitations, taking lessons from institutions like Wikipedia and experiments all over the world, we can start figuring out how to rebuild American democracy, and democracy elsewhere, on foundations that are deeper and more sure.
Instead of trusting that democracy’s manifest destiny will allow it to prevail, we can put our trust, as Franklin Delano Roosevelt did 90 years ago, in “bold and persistent experimentation,” and “enthusiasm, imagination and the ability to face facts, even unpleasant ones, bravely.” It has been a bad decade for democracy, and it may be that things will become still worse in the coming years, after a temporary period of respite. But the notion that human bias dooms democracy to failure is wrong. It is exactly on our biases that a better and stronger democracy might be founded.