Even before Charlottesville, critical analysis of the Trump phenomenon had tended to take the easy path: attributing the racism, xenophobia, and ethnocentrism of his most extreme supporters (and, arguably, of Trump himself) to his millions of other supporters. This equation is likely to become even more popular post-Charlottesville, and there’s doubtless some truth to it—but how much?

How Xenophobic Were Trump Voters?

In this post I’ll examine the xenophobia theory.[1] It makes sense on its face: In Trump’s announcement speech he slammed illegal Mexican immigrants, and building a wall on the Mexican border was the most prominently reported and controversial promise of his campaign. Yet Trump’s victory depended heavily on the votes of self-identified Republicans, 89 percent of whom supported him over Hillary Clinton. Given that more than half of all Republican primary voters had opposed Trump, the effect of xenophobia on his victory is anything but self-evident. Nor is it self-evident how many of even Trump’s hard-core supporters were xenophobic.

Still, a recent study could be seen as evidence for the xenophobia theory. In the New York Times, UCLA’s Lynn Vavreck writes of the study that it shows that “an exclusionary notion of American identity” had a significant grip on Trump supporters, and that such restrictive views had played “a major role in the Trump campaign.” Vavreck cites the study’s finding that while only 6 percent of John Kasich’s supporters agreed with the statement that “to be of European heritage or descent” is “very” or “fairly” important “to being truly American,” 30 percent of Trump’s Republican supporters agreed.

However, Republicans amounted to 27 percent of the electorate at the time of the study. Trump supporters accounted for 46 percent of that 27 percent. Thirty percent of that 46 percent comes to 3.7 percent of the electorate—barely more than voted for Libertarian Gary Johnson. This might have been decisive in a razor-thin election, but it does not seem very significant numerically (as opposed to morally).

Equally important, it is well established that opinion surveys may measure not real beliefs but off-the-cuff reactions to a question, or off-the-cuff endorsements of a proposed answer to it. Standard surveys, such as the one cited by Vavreck, do not allow respondents to answer in their own words: they provide answers in a multiple-choice format. Therefore, the respondents may be thinking about both question and answer for the first time while they are weighing the multiple choices. Not surprisingly, then, standard survey responses are frequently based on nothing but recently heard media chatter that the respondents pull from the top of their heads. So one has to wonder if the responses to the survey’s “European heritage” option really indicate a deep-rooted cause of Trump’s support. The other possibility is that the respondents supported Trump for other reasons and then, in response to the multiple-choice prompt, repeated what they’d heard people say about him—that he is Eurocentric.

We can operationalize the point by asking whether survey respondents’ willingness to check off the “European heritage” response indicates a desire to prevent non-Europeans from immigrating to the United States—which is where xenophobia might cause someone to support Trump’s candidacy. While there is no direct research on the question, there is suggestive evidence about immigration that uses more reliable research designs than a multiple-choice opinion survey.

For example, in an experimental study published in 2015, Jens Hainmueller of Stanford and Daniel J. Hopkins of Georgetown found that while the level of education, profession, employment prospects, and English-language proficiency of prospective immigrants made tremendous differences in the willingness of Americans to allow prospective immigrants admission into the United States, prospective immigrants’ country of origin—Mexico, France, Germany, the Philippines, Poland, India, China, the Sudan, or Somalia—made almost no difference. The one exception was with potential immigrants from Iraq, who prompted significantly greater resistance than potential immigrants from the other countries. Presumably, however, this reluctance was caused by fears of terrorism, not xenophobia.

However, the Hainmueller and Hopkins study concerned legal immigrants, while Trump’s Wall was intended to exclude illegal immigrants. A study by Morris Levy of USC and Matthew Wright of American University provides the best available evidence about attitudes toward illegal Mexican immigrants. In June 2015, Levy and Wright presented non-Hispanic California voters with two scenarios about one of three illegal immigrants: a Mexican national (“Juan”), a Chinese national (“Yuan”), and a German national (“Johan”). In the first of the two scenarios, half of the respondents were told only that the immigrant in question (whether Juan, Yuan, or Johan) had been in the United States for two years. In the second scenario, the other respondents were told, in addition, that the immigrant in question spoke English and had been working steadily as a waiter. In the first case, knowing only that the immigrant in question was illegal, 67 percent of the respondents favored allowing Juan into a legalization program, while 73 percent favored allowing Yuan or Johan into such a program—a 5-point gap that might be attributable to anti-Mexican bias. But when told that Juan, Yuan, or Johan spoke English and was employed, this relative bias against Juan disappeared: 83 percent favored allowing him in, compared to 82 percent for Johan and 81 percent for Yuan. It appears, then, that while 5 percent of the respondents were initially using a stereotype of illegal Mexican immigrants as non-English-speaking and/or non-working, in the absence of the stereotype they had no more of a problem with illegal Mexican immigrants than with illegals from other countries.

These findings held across Democrats and Republicans,[2] but non-Californians were not surveyed. Even if the findings are nationally representative, however, not only did 5 percent of the respondents have a damaging stereotype of Mexican immigrants, but 17, 18, or 19 percent (depending on whether the respondents were assigned Juan, Johan, or Yuan, respectively) were opposed to illegal immigration under any circumstances, regardless of the employment history or language skills of the immigrants. This might suggest undifferentiated xenophobia rather than xenophobia targeted at Mexicans alone. Another possibility, though, is opposition to the sheer illegality of undocumented immigrants. Nearly 100 percent of Republicans checked off the “respect American political institutions and laws” option as defining someone as “truly American.

Nationalism, Not Xenophobia

However, a Morning Consult/Politico survey shows 33 percent of the respondents opposed to legal immigration—nearly double the 17-19 percent who, in Levy and Wright’s study, opposed illegal immigration.

In explaining opposition to legal immigration, the Hainmueller and Hopkins study suggests that we should consider the perceived effects of immigration on the economic or other interests of current American citizens. This seems to be what Hainmueller and Hopkins’s respondents did when they chose to allow highly educated professionals into the country rather than poorly educated laborers. Rather than opposing immigration because of a dislike of immigrants in general, or a dislike of immigrants from a particular country, Americans may oppose immigration out of a desire to serve or defend the interests of their fellow Americans (defined not as those of European descent, but those who qualify as American citizens already).

I will call this “nationalist” opposition to immigration, as opposed to “xenophobic” opposition. By nationalism, I mean political favoritism toward one’s fellow citizens based on nothing but shared citizenship with them.

The decisive question, for a nationalist, is whether immigrants will or will not cause harm to current American citizens. Those nationalists who think immigration will harm current American citizens would tend to oppose it, while those who think it will help Americans, or will not hurt them, would tend to support it.

For example, an August 2016 Pew survey showed that while only 15 percent of Clinton supporters agreed that undocumented immigrants mostly fill jobs that U.S. citizens would like to have, 35 percent of Trump supporters agreed. Similarly, while only 13 percent of Clinton supporters agreed that undocumented immigrants are likely to commit serious crimes, 50 percent of Trump supporters agreed. Trump supporters, like the opponents of immigration in the Hainmueller and Hopkins study, seem to be worried that immigrants will harm the respondents’ fellow Americans.

Along these lines, Karthick Ramakrishnan, Kevin Esterling, and Michael Neblo find through experimental manipulation that American citizens tend to think of both immigrants and illegal immigrants as impoverished, and that both legal and illegal immigrants tend to be seen as “hurting the economy by driving down wages or increasing unemployment.”[3] In conjunction with the Levy and Wright and Hainmueller and Hopkins studies, this suggests that the main division over immigration is based on citizens’ differing judgments about the effects immigrants will have on the social and economic well-being of current American citizens.

To summarize: the Levy and Wright study suggests that opposition to illegal immigration is not based on country of origin but the effects of illegals on Americans (although it may also be based on opposition to illegality). In turn, the Hainmueller and Hopkins and Ramakrishnan, Esterling, and Neblo studies suggest that opposition to legal immigration is also based on perceptions about the effects of immigrants on the well-being of respondents’ fellow Americans. Thus, opposition to immigration seems to be based primarily on nationalism, not xenophobia.

The Ubiquity of Nationalism

Nationalism, however, is just as irrational as xenophobia. One’s physical residence vis-à-vis a line on a map has no bearing on one’s humanity. If xenophobia is antipathy toward others whom one has never met based on their living outside the borders of one’s country, nationalism is sympathy toward others whom one has never met based on their living inside those borders. To the nationalist, people within these arbitrary geographical boundaries deserve to have their interests protected by the national state—even at the expense of the interests of those outside those boundaries. Press a nationalist to justify this distinction and one is likely to be met with puzzlement.

The irrationality of nationalism is nearly invisible, however, because nationalism itself is so much a part of our cultural background that it’s hard to notice. We are raised watching news that is primarily about “our” country; we learn “its” history in school; even our weather maps show only the weather in the United States, stopping at the borders. It therefore comes to seem natural or commonsensical that “our” government should serve “our” people. In reality, though, these ubiquitous sentiments are anything but natural or commonsensical. They have to be pounded into people’s heads. State control of education seems to have been particularly important to the development of nationalism almost everywhere.[4]

The most baleful effects of nationalism are found in the political arena. Virtually every politician—not just in the United States but around the world—displays nationalist biases that are considered wholly unremarkable. Consider Bernie Sanders, who in July 2015 said: “I frankly do not believe we should be bringing in significant numbers of unskilled workers to compete with [American] kids. I want to see these kids get jobs.” In other words, American kids deserve jobs while their unskilled competitors, being non-Americans, do not. Sentiments such as these pass without condemnation because they are not xenophobic. But they are nationalistic, which is equally unjustifiable.

Defending the Undefendable

The key point is that the borders we take for granted as the starting points of our politics are artificial. They are products of historical contingency whose normative significance is nil, and must be conveyed implicitly, through the culture’s obsessive preoccupation with people who live within them. Once the arbitrary nature of this way of dividing up humanity is made explicit, it becomes morally indefensible.

This does not stop people from trying to defend it, though. Writing in the Trump-friendly Claremont Review of Books, William Voegeli castigates philosopher Martha Nussbaum for encouraging us to think of ourselves as citizens of “the worldwide community of human beings,” and political theorist Joseph Carens for pointing out that it is “hard to justify” the nation-state’s conferral of “great advantages on the basis of birth”—especially inasmuch as national borders “entrench these advantages by legally restricting mobility.” But how does Voegeli rebut Nussbaum, Carens, and other critics of nationalism? His only actual argument appeals to John Stuart Mill’s description of a nation as a people united, most importantly, by “identity of political antecedents; the possession of a national history, and consequent community of recollections; collective pride and humiliation, pleasure and regret, connected with the same incidents in the past.”

Mill, however, took for granted the point at issue: the normative significance of the history of people who happen to live between arbitrary borders, in constituting what we would now call their “identity.” The question is whether one should identify oneself with people in the past who happened to live in the same geographical area in which one now lives.

Liberals—Mill being one of the most prominent—have long had a problem with whether “identity” is immutable and given or open to change. (The same is true of communitarians and conservatives, as Voegeli demonstrates.) Mill treated national identity as if it is an unalterable “preference,” but the very existence of people such as Nussbaum and Carens, who reject their national identities as arbitrary and partial, shows that such preferences can be altered. The question is whether they should be altered; Mill and Voegeli beg this question.

The Difference Nationalism Makes

If nationalism is as arbitrary and unjustifiable as xenophobia, does it matter that the term nationalists might better characterize Trump supporters than the term xenophobes? It seems to me to matter for two reasons.

First, if we understand Trump supporters as nationalists, we can notice their similarities with Trump opponents. They, too, are nationalists, as almost everyone is. That’s why Hillary Clinton came no closer than Bernie Sanders or any Republican opponent of Trump[5] to treating the interests of “Americans” and “non-Americans” as equivalent. The question they debated in the 2016 campaign, as had others before them, is whether or not immigrants cost Americans their jobs or safety—not whether even such a cost might be worth it to help some of the world’s least-advantaged people, who happen to live outside “our” borders. On this score, Trump supporters are not so very different from their opponents. If they are wrong, so are a lot of other Americans. And they are.

Second, having noticed the commonalities between Trump supporters and their opponents, we can begin thinking about whether Trumpism is a product of the status quo rather than being a deviation from it. This, in turn, can lead us to question what the status quo of which nationalism is a part (evidently a central part) is: how it functions, how it is sustained, and whether its continuation makes sense.

Before we get to that point, however, we need to examine the other extant theories about Trump’s support: as a product of authoritarianism, racism, ethnocentrism, a backlash against political correctness, or economic grievances. Stay tuned.

[1] I plan to tackle other theories in this space over the coming weeks.

[2] Morris Levy and Matthew Wright, “White Americans’ Opinions about Immigration Policy: The Role of Attitudes toward Latinos,” presented at the Research Workshop in American Politics, University of California, Berkeley, September 14, 2016, p. 8.

[3] Karthick Ramakrishnan, Kevin Esterling, and Michael Neblo, “Illegality, National Origins, and Public Opinion on Immigration,” p. 11.

[4] See, e.g., Eugen Weber, Peasants into Frenchmen. England, arguably the first national state (see Liah Greenfeld, Nationalism: Five Roads to Modernity), and the United States prior to the advent of mass public education—due to its unique origins in a revolution against England—are prominent exceptions to the general rule that state-controlled education is a prerequisite for universally shared national allegiance.

[5] Except, perhaps, Jeb Bush, who said that many illegal immigrants immigrate as an “act of love” for their families, whom they support with the income they earn in the United States.

Jeffrey Friedman, a Visiting Scholar in the Charles and Louise Travers Department of Political Science, University of California, Berkeley, is the editor of Critical Review and of The Nature and Origins of Mass Opinion Reconsidered (Routledge, 2014). His Technocracy: A Critique is forthcoming from Oxford University Press (2018).