Donald Trump has said so many deeply offensive and demonstrably false things during the last year that one especially revealing sentence went virtually unnoticed. At a rally in May he declared, “The only important thing is the unification of the people, because the other people don’t mean anything.”

Some people mean something; others don’t. This is a politics predicated on identity, defined mainly by exclusion. Who exactly gets excluded and how—whether Mexicans by way of a wall or Muslims by way of a religious test—can vary from day to day, but the fundamental logic is always the same. There is a real America, Trump knows what it is, and whoever does not agree does not belong—and is implicitly opposed to making America great again.

What is telling about this thinking is that it exposes Trump as a populist, in the strict sense. Many observers have spoken of a resurgence of populism in America and in Europe, but their characterizations of the phenomenon often get it wrong. For all the talk, it is not clear that we know what we are talking about when we talk about populism. The morning after the Brexit vote, for example, Tony Blair diagnosed a “convergence of the far left and far right.” He explained that “the right attacks immigrants while the left rails at bankers, but the spirit of insurgency, the venting of anger at those in power and the addiction to simple, demagogic answers to complex problems are the same for both extremes.” But demagoguery is not the same thing as populism. The former is a matter of false promises or manipulating citizens’ emotions; the latter is about claiming a moral monopoly on representing the so-called “real people.”

The defining feature of populism is not anti-elitism but anti-pluralism. Populists claim that they, and they alone, represent the “real” people.

Getting populism wrong can distort our understanding of political reality in at least two crucial ways. For one thing, there is the danger of equating populism—which justifiably has a bad name across Europe and in parts of the Americas—with legitimate protest against specific policies. At the same time, it underestimates the perils posed by actual populism, such as that of Trump and Marine Le Pen. Populists don’t just criticize elites or play on the prejudices of people using emotionally charged rhetoric. They posit that there is one true, unified people and that they alone are its legitimate representatives. In a pluralist democracy comprised of diverse interests and identities, this claim opens the path to excluding entire groups—and in the worst cases, to authoritarianism. Opponents of populism must be explicit about this danger: they should not pretend that the populist is just another type of politician, if a bit more emotional or uncouth. But they also have to avoid a trap: excluding the supporters of the excluders. This is the tactical error Hillary Clinton made when she criticized Trump voters as “deplorables.”

Populists always claim that they, and they alone, represent the people. Consider Trump’s salvific boast in his speech at the Republic National Convention: “I am your voice,” with the corollary: “I alone can fix it.” This rhetoric condemns other political contenders as entirely illegitimate. The populist does not just disagree about policy; disputes are always matters of character. Other politicians are corrupt, or they put elite institutions before the people. In a word, the defining orientation of the populist is not anti-elitist or even antiestablishment (for populists are perfectly content with the establishment when they are in power) but anti-pluralist. This attitude translates into exclusions both at the level of party politics—other politicians are crooked—and at the level of citizens: opponents are likened to traitors to the nation.

Of course, this is not to say that every moral claim in politics is populist or inherently dangerous. All political discourse is shot through with more or less direct appeals to values and principles. The point is that populists claim a privileged understanding of what the real people—by definition morally irreproachable—are like and wish for.

• • •

Every populist operates with a symbolic and ultimately moral distinction between the real people and those who don’t belong. Indeed U.S. populists—think of George C. Wallace—have often deployed the phrase “real Americans.” Wallace is mostly remembered, of course, for his infamous statement upon his inauguration as governor of Alabama in 1963: “In the name of the greatest people that have ever trod this earth, I draw the line in the dust and toss the gauntlet before the feet of tyranny . . . and I say . . . segregation today . . . segregation tomorrow . . . segregation forever.” It may be hard to hear amidst the chant of segregation, but note the appeal to “the greatest people that have ever trod this earth” and the claim exclusively to speak in its name. This is the signature move of the populist.

All populists do identity politics, then—which is not to say that all identity politics is populist. Trump’s leadership of the birthers is profoundly revealing of his particular brand of populist identity politics, not a historical accident. In the eyes of right-wing populists, Obama embodies two groups that don’t belong to the real America: bicoastal liberal elites and African Americans. It is no surprise that Trump took such a prominent role in trying to prove that Obama’s presidency was not just symbolically illegitimate but plainly illegal; a non-American had infiltrated the nation’s highest office. Trump now deflects the issue and accuses Hillary Clinton of instigating the whole thing. Yet birtherism is at the very heart of his political worldview.

When populists clamor for “the people themselves” to speak, they do not really call for more democratic participation. Rather, they rely on a symbolic representation of the “real people”—a fictional entity, a homogeneous and morally unified body excluded from democratic procedures whose alleged will can be played off against actual election results. It is not an accident that Richard Nixon’s notion of a “silent majority” has had such an illustrious career among populists: if the majority were not silent, it would already have a government that truly represented it. If the populist politician fails at the polls, it is not because he did not convince a sufficient number of voters, but because the majority has not yet dared to speak—or, even worse, because it has been prevented from doing so by elites.

But, one might object, don’t populists often advocate for more direct forms of democracy? Don’t they demand referenda? Yes, but we must be clear about what a referendum means to the populist. Populism has no particular interest in getting people continuously to participate in politics. A referendum is not meant to catalyze open deliberation to generate a range of well-considered judgments about policy. Rather, it serves to ratify what the populist leader has already discerned to be the genuine popular interest as a matter of identity, not as a matter of empirically verifiable ideas and interests. Populism without participation is an entirely coherent proposition.

In fact, populists are not even inherently anti-elitist, if one takes the term to mean that power should be as widely dispersed as possible. Populists have no problem with the principle of representation, so long as they are the representatives. They are likewise fine with elites, so long as these elites speak for what populists take to be the people. One does not score a decisive point against Trump by pointing out that he is part of the economic elite; the same is true of businessmen-turned-politicians in Europe such as the Swiss populist Christoph Blocher. The supporters of these men know very well that they are members of the elite. What matters is their promise that they will not betray the people’s trust, that they will faithfully execute the people’s mandate. It is no accident that populists in power often adopt a kind of caretaker attitude toward an essentially passive people. Think of Silvio Berlusconi’s reign in Italy: matters of state were to be left to Il Cavaliere, who would successfully govern the country like a very large business corporation. There was no need to enter the piazza to participate.

Populists do not really call for greater democracy: populism without participation is an entirely coherent proposition.

Political scientists draw a distinction between imperative mandates and free mandates in political representation. An imperative mandate requires that representatives must act according to the express policy wishes of the voters; free mandates permit representatives to use their own judgment. Populists assume that the entire people can speak with one voice and issue something like an imperative mandate. There is therefore no need for debate, let alone the messy back-and-forth of deliberation in national assemblies. Yet the imperative mandate does not really derive from citizens at all; at best it comes from those who shout the loudest at rallies. Its supposedly detailed instructions are ultimately based on an interpretation by populist politicians. Many political scientists have long argued that a completely coherent, single “popular will” is a fantasy, and that no one can credibly claim, as Juan Perón used to, that “the political leader is the one who does what the people want.” Pretending that there is such a will also weakens democratic accountability. Populists can always turn back to the people and declare: “we implemented exactly want you wanted, you authorized us; if anything went wrong, it’s not our fault.” By contrast, a free mandate puts the burden on representatives to justify how they used their political judgment.

Principled, moralized anti-pluralism and the reliance on a non-institutionalized notion of “the people” also helps to explain why populists so often reject unfavorable empirical outcomes of elections in favor of what they deem the morally correct one. Think of the Hungarian populist Victor Orbán claiming after unexpectedly losing the 2002 parliamentary elections that “the nation cannot be in opposition” (thus equating his political party with the nation). Or of Andrés Manuel López Obrador arguing, after his failed bid for the Mexican presidency in 2006, that “the victory of the right is morally impossible” (and declaring himself “the legitimate president of Mexico”). Or of Tea Party Patriots claiming that Obama, even though he won a majority of the vote, is “governing against the majority.” Then there is the example of right-wing populist Geert Wilders, who has called the Dutch House of Representatives a “fake parliament” with “fake politicians.” And finally there is Trump again, reacting to every loss in the primaries with the charge that his opponents committed fraud and proleptically claiming that a Clinton win in November must be due to a rigged system—or even that she is fundamentlly ineligible for the presidency because she is a criminal and ought to be in jail. The problem is never that the populist failed to capture a majority of voters; it is always that corrupt institutions produce the wrong outcomes. Conspiracy theories are thus not a curious accident of populist rhetoric; they are rooted in the very logic of populism itself.

• • •

Populism, then, is not just a matter of style—being folksy or lauding the wisdom of ordinary people. It is also not a matter of irresponsible policies or political pandering. This latter view is widespread in Europe; as the German-British sociologist Ralf Dahrendorf once put it, populism is simple, while democracy is complex.

To reject this characterization of populism is not to deny that some policies embraced by populists are irresponsible—for misrepresenting the evidence, say, or ignoring long-term consequences for short-term electoral gains. One does not have to be a neoliberal technocrat to judge some policies plainly irrational. Think of Hugo Chávez’s hapless successor as president of Venezuela, Nicolás Maduro, who sought to fight inflation by sending soldiers into electronics stores to put stickers with lower prices on products.

Still, one cannot generate a definition of populism this way. For there is no absolutely clear line between responsibility and irresponsibility, and charges of irresponsibility can themselves be highly partisan. Construing political debate as a matter of responsibility raises the question, responsible according to which criteria? Free trade agreements can be responsible in light of a commitment to maximizing GDP, say, and yet have unacceptable distributional consequences. Setting up a distinction between populism and responsible governance only obscures the issues at stake. It can also be an all-too-convenient way to discredit criticism of certain policies, as has regularly happened when the policies of Bernie Sanders and the leaders of the new left-wing parties Podemos in Spain and Syriza in Greece were attacked as populist. Of course, one can say that Sanders conforms to a historical understanding of the word “populism” in the United States—grass-roots, progressive, critical of Wall Street. But it would be wrong to say that we are dealing here with all-out anti-pluralists who posit a one-and-only true people. This is not to say that left-wing populism is an impossibility: Hugo Chávez was a populist; disagreeing with his “twenty-first century socialism” became tantamount to being an enemy of the people.

Populism is also not a matter of advancing particular policies. It is never so simple as “You’re against immigration? You must be a populist!” Or: “You want to end austerity in Southern Europe? You must be a populist!” Populism has a clearly identifiable structure, but the specific content of its policy commitments can vary significantly across historical and geographical contexts. What matters is that populists will find a way to communicate that only some people are the real people, and that only some are the authentic spokespersons of the people. These claims are bound to clash with a reality where the populists do not get unanimous support at the polls. The explanation can never be that the people made a mistake—for the people are always pure and unerring in their will. (This assumption distinguishes populists from Leninists or religious fanatics, those who know better than an unenlightened or sinful populace.)

Populism is based on a fiction but it is not fictional politics: there are real citizens supporting it.

Some, especially in the United States, might insist that populism must be characterized as a movement driven by those who feel marginalized by neoliberal globalization. Yet defining populism in terms of particular socioeconomic groups is empirically dubious: a number of studies have shown, for instance, that Trump’s supporters are not on average lower-income, nor are they disproportionately concentrated in areas most adversely affected by globalization. Moreover, such an argument often relies on a largely discredited set of assumptions from modernization theory. It is true that in Europe, those who vote for what are commonly referred to as right-wing populist parties make less and are less educated. They are also overwhelmingly male—a finding also borne out in the United States, but not in Latin America. Yet these patterns can change; the National Front has attracted many female voters in recent years (and not just because of the prominent roles played by Marine Le Pen and her niece Marion). As the German social scientist Karin Priester has shown, economicaly successful citizens often adopt an essentially Social Darwinist attitude, justifying their support for right-wing parties by asking, in effect, “I have made it—why can’t they?” (Think of the Tea Party placard demanding “Redistribute My Work Ethic!”) In some countries such as France and Austria, populist parties are effectively catchalls: they draw a large number of workers but also many from other walks of life.

Surveys have shown that socioeconomic status and support for right-wing populist parties often do not correlate at all. We cannot reduce perceptions of national decline or danger (“elites are robbing us of our own country!”) to personal fears or status anxiety. Many supporters of populist parties pride themselves for being freethinkers and vehemently deny that their views are merely personal or emotional. In any case, it is wrong to conflate the content of a set of political beliefs (populism is, after all, an “ism”) with the socioeconomic positions and the psychological states of its supporters. This is like saying that the best way to understand social democracy is to characterize its voters as workers envious of the rich. The sociological and demographic profile of supporters of populism does matter in how we think about the phenomenon. But it is both patronizing and counterproductive, without any real empirical basis, to call it an inarticulate political expression on the part of the supposed losers of modernization, as Blair once did when he said that many people fail to understand “how the modern world works.” Populism is based on a fiction—the one homogeneous real people with a unitary will—but it is not fictional politics all the way down: there are real citizens supporting it, and we cannot from the get-go exclude the possibility that they have real, rational reasons for doing so.

Why, then, do so many politicians, journalists, and academics keep resorting to the language of resentment? Because they continue to draw, whether knowingly or unwittingly, on assumptions derived from the heyday of modernization theory in the 1950s and 1960s. It was liberal intellectuals such as Daniel Bell, Edward Shils, and Seymour Martin Lipset who described populism as a helpless articulation of anxieties and anger by those longing for a simpler, premodern life. Lipset, for instance, claimed that populism was attractive for “the disgruntled and the psychologically homeless . . . the personal failures, the socially isolated, the economically insecure, the uneducated, unsophisticated, and authoritarian personalities.” The immediate targets of these social theorists were McCarthyism and the John Birch Society, but their diagnosis often extended to the original American populist revolt of the late nineteenth century. Victor C. Ferkiss saw the original followers of the Farmer’s Alliance and the People’s Party as nothing less than the precursors of a distinct American variety of fascism. This thesis came to be challenged, of course, but the background assumptions remain among many social and political commentators today.

Finally, there is the claim that populism must have something to do with those who first called themselves populists. Think of the Russian intellectuals known as narodniki in the late nineteenth century and their ideology of Narodnichestvo—usually translated as “populism”—which idealized peasant life and saw the village commune as a political model for the country as a whole. Like many urban intellectuals, they found that the people did not welcome them in the ways they had hoped.

For many observers, there has to be a reason why something called populism emerged simultaneously in Russia and the United States toward the end of the nineteenth century. The fact that both movements had something to do with farmers and peasants gave rise to the notion, prevalent at least until the 1970s, that populism had a close connection to agrarianism—that it was, in essence, the revolt of reactionary, economically backward groups in rapidly modernizing societies. While we have by and large lost that association today, the origins of what has historically been called “populism” in the United States still suggest to many observers that populism must favor the least advantaged or bring the marginalized into politics. This sense is reinforced in Latin America, where advocates of populism have always stressed its inclusionary and emancipatory character in what remains the economically most unequal continent on the globe.

But political and social theory cannot root itself in one particular pattern of historical experience. We have to allow for the possibility that a plausible understanding of populism may exclude historical movements and actors who explicitly called themselves populists. Historians would not generally argue that a proper understanding of socialism needs to make room for National Socialism just because the Nazis called themselves socialists.

• • •

How should other politicians address this form of highly exclusionary identity politics? Not by turning the tactic of exclusion back on the populists themselves. In Europe it has been a common strategy to shut populists out of televised debates altogether, but this is a mistake. This dismissive attitude confirms populists in what they have been telling their supporters all along: elites are uncaring and ultimately alien, or at least alienated from the average citizen’s daily problems. Yet talking with populists is not the same as talking like populists. One can take problems seriously without accepting the way populists frame them.

Conspiracy theories are rooted in the very logic of populism itself.

On the other hand, it is naïve to think that populists can be defeated just by setting facts straight or lecturing citizens that their economic interests won’t be served by identity politics (or, in Blair’s parlance, explaining to them how the modern world really works). Hardly any facts speak for themselves; they are always part of a larger political narrative. Leaving the narrative to populists is again a mistake: just think of the way that Nigel Farage dominated debates about Brexit. His story about the referendum—that it was a matter of defending English freedom against the Brussels dictatorship—largely prevailed, not least because so many supposedly mainstream conservative politicians declared it essentially correct. When the Remain campaign pointed to the fact that leaving the EU would make British households £4300 worse off, it was already too late: the narrative had been lost. What is a few thousand pounds compared to the value of freedom?

So what to do? First, we must be very careful how we describe populists’ supporters. Words like “frustration,” “anger” and, especially, “resentment” are not neutral. In the abence of reliable data on what actually moves citizens to opt for populists, observers far too often resort to the clichés of nineteenth-century mass psychology: the masses are by defintion irrational and swayed by charismatic, demagogic leaders. Yet the very split between reasons and emotions on which such diagnoses tend to rely is misleading: nobody is just angry out of nothing. Emotions also have reasons, and citizens can usually give an account of them. This does not mean that we have to accept them, but suggesting all supporters of populists are driven by raw rage and resentment will ensure that we never get to the point of debating reasons.

This is not to deny that there really are racists among Trump supporters. But we cannot just write off all those who have supported him over the last year or so as simply irredeemable. Instead of demonizing populist supporters collectively, one should call out the populist politician directly and precisely for rejecting democratic pluralism—and, in the case of Trump in particular, for the way he incites hatred against minorities and systematically denigrates women.

Of course, when confronting populists, one should also hammer away at the facts whenever possible; normative and empirical lines of attack are not mutually exclusive. As Michael Kazin has reminded us, a significant part of Wallace’s support in his 1968 presidential campaign disappeared, for instance, after unions started to bombard their members with information about both the actual situation of “the working man” in Alabama and how little Wallace had done as governor to improve it. There is no guarantee that this strategy will work. But if such aggressive counterattacks are part of a narrative about shared democratic ideals, they just might. We should remember that a single question from Joseph Welch—“Have you left no sense of decency?”—helped instigate the undoing of Senator Joe McCarthy, a figure comparable to Trump in his cruelty and divisiveness.