September 6, 2017


Morgan Marietta’s and Will Wilkinson’s replies to my essay on nationalism could hardly be more different. The differences bear not only on how we account for Donald Trump’s surprising political success, but on the purposes and procedures of social science.

Like my essay, Marietta’s extends the principle of interpretive charity to Trump supporters. In my view, interpretive charity should be the first principle of social science. If anything, however, I think that Marietta takes interpretive charity a bit too far by endorsing nationalist sentiments as morally legitimate.

It’s true that if we use interpretive charity to try to understand ideas with which we disagree, we might become proponents of those ideas. Indeed, one of the advantages of interpretive charity is that it can change our minds. Conversely, if a scholar of nationalism, for example, fails to agree with the ideas of nationalists, it seems at least possible that this is because he has failed to understand those ideas. If I really comprehended why you believe X, I would have to understand all the considerations that led you to believe X, and I’d have to understand them in the same way you do. But, having achieved this mind meld, shouldn’t I, too, believe X?

Not necessarily. I may know of counter-arguments against X that you don’t know about, and these may lead me to disagree with X even when I completely understand your reasons for agreeing with it.

I think this is the case with nationalist beliefs. One can be fully charitable toward these beliefs even while noticing that they tend to be inculcated very early, among children, through symbols (such as the national flag) and biased information samples (such as the media’s massive overweighting of attention to the citizens of one’s own cfdountry rather than people who live elsewhere). By the time someone is capable of thinking critically about her own nationalist assumptions, she may find it hard even to identify them—and unnecessary, too, since everyone around her will tend to take the same assumptions for granted.

Thus, while I argued, in the spirit of interpretive charity, that nationalism is distinct from xenophobia, I also maintained, and continue to maintain, that nationalism is morally indefensible: most nationalists have simply failed to think about the arbitrariness of the group loyalties that were pre-rationally constructed for them long ago. 

Interpretive charity is not Wilkinson’s project. He argues that a significant proportion of Trump supporters are xenophobes beholden to an irrational hostility to foreigners. I’m suspicious of such explanations because they tend to demonize “the other”—in Wilkinson’s case, Trump supporters—which is exactly what Wilkinson accuses Trump supporters of doing (when it comes to foreigners).

Demonization amounts to a confession that one has failed to understand the other on his or her own terms. This usually means that social science has failed. Not always, though. It’s possible, in a given case, that people’s behavior or their ideas are so irrational that they can be explained only by appealing to influences of which they are unaware, and which they would deny if they were asked about them. In such cases, we may need “deep” psychological theories to explain the other. Like the authoritarian-personality theory of Trump’s support, which I discussed in my last post, the xenophobia theory posits a subterranean psychological force that has erupted in the form of Trumpism.

This isn’t inherently unbelievable, but the evidence for it is weak; and there is strong evidence against it. So instead of xenophobia, my essay advanced nationalism as a partial explanation of Trump’s support. Nationalism fits the available evidence better, and it’s interpretively charitable.

A good test of interpretive charity is whether those whose actions or beliefs you’re trying to explain would accept your explanation. It’s doubtful that many Trump supporters would accept Wilkinson’s irrationalist explanation of their actions and beliefs, but they would probably accept the nationalism explanation. For this explanation suggests that support for Trump is consistent with the “commonsensical” nationalist presuppositions of everyday politics. The xenophobia theory, in contrast, is not only interpretively uncharitable and weak in evidentiary terms, but it conveniently locates Trumpism far away from the “liberal” traditions of everyday politics—where it can safely be vilified without threatening the status quo.

Trump as Deep Nationalist

Marietta’s essay begins by underscoring the fact that Trump constantly and unreflectively appeals to the interests of “America” as the supreme good. Marietta is willing to call the basis of these appeals an ideology: “deep nationalism.” By this, Marietta means that Trump’s nationalism is the prism through which he seems to view nearly every policy issue—at least those issues in which he takes an interest.

The four most important of these are immigration, U.S. relationships with foreign allies, military policy, and international trade. All one need do is listen to what Trump says, as Marietta has done, to discover a connective ideological thread among these issues: nationalism. At the same time, Trump’s “deep nationalism” explains his lack of interest in a host of policy issues that preoccupy conservative and liberal ideologues, such as Obamacare, the minimum wage, regulatory policy, global warming, income inequality, tax rates, etc., ad infinitum. These latter issues do not easily lend themselves to “America-first” analyses, and thus are not clarified by the deep nationalist lens through which Trump (and, both Marietta and I suggest, many of his supporters) view politics.

I think Marietta is making an excellent point. Nationalism does seem to function for Trump in an ideological manner, at least in the sense in which political scientists tend to use this term: as a master heuristic that orients the ideologue politically, organizing most or all of her political ideas. However, at the risk of quibbling, I think Marietta dilutes the power of this analysis of Trumpian nationalism by describing it not only as an ideology, but also as a branch of conservatism, as a symbol, as an identity, and as a value. Let me say a few things about each.

Trump and Conservatism

Marietta’s conception of conservatism strikes me as too schematic. I know many conservatives who do not see society as fundamentally fragile and in need of social glue, and many who do not care about anything like “ordered liberty” or a golden mean between freedom and authority. Marietta’s description fits certain conservatives, such as Straussians, but they are a tiny band of intellectuals without any discernible popular influence. At the mass level, standard journalistic depictions of three main groups of conservatives—Tea Partiers (small government/constitutional conservatives), cultural conservatives, and foreign-policy conservatives—do not seem to be in need of updating, at least not yet.

What Trump’s surprising popularity does show, I think, is that nationalism unites many conservatives of all three types—along with many non-conservatives, too. The transcendent appeal of nationalism makes considerable intuitive sense, as nationalism is more elemental than the ideologies that attract well-educated and politically literate adherents. It’s so basic that small children can understand it. Indeed, no matter how little you know about politics, it is likely that you were indoctrinated with nationalism when you were a small child. Trump, indoctrinated in the same way, and having learned little else about government, policy, or history in the meantime, is the ideal exponent of the most simplistic possible political ideology: that of “America first.”

This ideology offers its adherents a key to understanding the otherwise-confusing world of politics, even if they lack much interest in or knowledge of it. The key is to ask oneself whether a politician intends to put the interests of “Americans” before those of “foreigners.” The question of whether the mere intent to help Americans will accomplish the objective (let alone the question of whether Americans deserve priority over non-Americans) goes unasked. This immensely simplifies what would otherwise be an opaque world of public policy: the world of policy debate. In policy debate, what is at issue is usually whether a given proposal that sounds as if it will serve the interests of Americans (for example) will actually do so. The answer is rarely as straightforward as deep nationalists believe. But this is part of the appeal of deep nationalism, which has little to do, as far as I can tell, with conservatism.

Nationalism and Symbolism

Part of the way nationalism gets inculcated is through the apotheosis of symbols such as the American flag. These symbols acquire emotional resonance, and this emotional resonance may help to explain why people turn “naturally” to their nationality when they think about politics (especially insofar as they know relatively little about it). So a full analysis of the cognitive role of nationalism might very well need to explore the emotional power of nationalist symbols. Such an investigation would have to go beyond both hyper-rationalist (rational-choice- inspired) theories of political heuristics and the irrationalist, psychological understandings of politics that I’ve been trying to challenge in my ongoing series of essays.

For this very reason, however, it’s important to spell out carefully how the emotional or a-rational appeal of nationalism is connected to its cognitive function. Until we succeed in doing that, I worry about the confusion that might be created if we use the language of “symbolism” to describe nationalism, since this language currently connotes the groundlessly emotional. My position is that nationalism is illogical, but that the lapse in logic is not apparent to people who have been indoctrinated with nationalist presuppositions. It would be unfair—uncharitable—to say that they are irrationally clinging to nationalism, in the sense that they somehow know it is wrong. But to say that nationalism is symbolic may suggest something similar: that nationalism is merely an empty screen onto which people project their irrational desires. I don’t think Marietta is saying that, but it’s a connotation of calling nationalism “symbolic” that I think it’s best to avoid.

Nationalism and Identity

The same worry colors my reaction to using the language of “identity” to describe nationalism. There’s no denying that nationality is probably the central “identity” of most people in the modern world—at least in the bare sense that, if asked “who they are,” they are likely to answer “an American,” “a Mexican,” etc. Yet, without doubting the existence of national identities (in people’s heads), I wonder how important “identities” are to people who have not been influenced by academic discussion, where identity politics has been extremely important for several decades.

National identity can be very important in helping people to organize their thoughts about politics. Yet thinking about politics isn’t all that important to most people. So we may mischaracterize the situation if we project onto such people an obsession with their national identity. Similarly, identity itself may not be important to most people. Non-intellectuals’ answer to the question of “who I am”—namely, that I am the person behind my eyeballs—may feel so unproblematic that the very question of identity is a non-issue for them. The language of “identity” may inappropriately import the preoccupations of academics into our understanding of mass politics.

Nationalism and Values

Unquestionably the ideology of nationalism takes certain values for granted, but I don’t see the sense in calling nationalism a value, as Marietta does. Nationalism—the ideology—is not itself valued by nationalists (except perhaps by those who derive so much meaning and purpose from it that they love it for its own sake—members of the alt-right, for example). Nor, at least in the American context, does it seem right to say that “the nation” is valued, as such, by nationalists. There is no tradition of extolling “the American nation” as if it were an end in itself. But that’s what I take a value to be: an end in itself.

Nationalism and Egalitarianism

Even more important than Liah Greenfeld’s distinction between civic and ethnic nationalism, cited by Marietta, may be her observation, in Nationalism: Five Roads to Modernity, that nationalism is inherently egalitarian—within the borders of a given nation-state. Thus, the end in itself that is presupposed by nationalism is the equal worth of the lives of one’s conationals. In this view, rather than being a value, nationalism is premised on a value: the equal worth of one’s fellow citizens. (The causality may run from the establishment of a nation-state to the presumption of equality among its citizens, but this is probably because the idea of nationality implicitly contains the presumption of equality.)

Since I share Greenfeld’s view, I resist Marietta’s suggestion that nationalism embodies a value that competes with egalitarianism. It seems to me that nationalism is a form of egalitarianism; and that it is, in fact, the form that egalitarianism almost always takes in the modern world.

However, nationalist egalitarianism is self-contradictory in limiting itself to equality among the human beings who happen to live within historically arbitrary national borders, while treating the lives of those outside those borders as if they have no worth. This is what makes nationalism illogical. Its tacit definition of who should be treated equally is arbitrary.

Wilkinson’s reply illustrates the illogic. So let me analyze his response before turning, in conclusion, to Marietta’s qualms about the psychological practicability of cosmopolitanism.

The Nationalist Scapegoat: Xenophobia

Wilkinson is an egalitarian and extols the egalitarianism that nationalism makes possible within the borders of a nation-state. Yet, according to nationalism, equality stops at those (arbitrary) borders. Viewed from outside of those borders, nationalism is inescapably inegalitarian.

Wilkinson is aware of this problem but does not really address it. Instead, he presses hard on the distinction between nationalism and xenophobia, with xenophobia taking the rap for inegalitarianism. But even if xenophobia did not exist, nationalism would remain inegalitarian from the perspective of those outside a given nation-state’s borders.

This isn’t just a philosophical issue. Non-xenophobic nationalism can easily justify the immigration restrictions that Wilkinson opposes, as well as the trade restrictions and the foreign-policy isolationism that Trump advocates (or used to advocate, before being enlightened about its adverse consequences by his generals). Such policies can be seen—indeed, are seen, every day, in the normal state of our political discourse—as serving the interests of our fellow citizens, not as punitive exercises directed at despised outsiders. This is the political discourse that I’m trying to get us to examine critically. Insisting that Trump is set apart from this discourse, because instead of nationalism he appeals to xenophobia, inadvertently blocks an examination of the discourse itself. It entrenches the complacency with which nationalism—“good,” non-xenophobic nationalism—is typically viewed, because it contrasts this type of nationalism against its evil, xenophobic twin. Yet both types of nationalism can produce the same policies—the very policies Wilkinson opposes.

How Non-Xenophobic Nationalism Works

A nice example of how this works is suggested in the course of Wilkinson’s odd paean to the GDP growth that could be unleashed by open borders. It almost reads as if Wilkinson thinks that open borders are justified by the contribution of migrant laborers to the stock of global wealth. (There’s gold in them thar migrants—trillions and trillions of dollars of it!) But what if migrants contributed little to GDP? What if they reduced it? By Rawlsian standards at least, their contribution to or subtraction from GDP does not matter. What matters is that migrants, frequently among the least advantaged people in the world, would be helped by open borders. I think Wilkinson means to say this, because he asserts that most of the GDP gains would go to the poor. It is that fact that matters, not the sheer fact that “trillion-dollar bills” are allegedly being left on the proverbial sidewalk by closed borders.

Why, then, do developed countries close their borders? Wilkinson points out that closed borders are “constantly re-affirmed” by the democratic polities of the West. But what exactly is the political dynamic of “liberal-democratic institutions,” as he puts it, that accounts for this?

The answer, I believe, is nationalism, which is taken for granted in the politics of Western countries (and all other countries). From a nationalist perspective, the welfare of one’s conationals is what matters; the welfare of “foreigners” does not. To sustain the high wages of one’s conationals, then, closed borders, tariffs on manufactured goods, and trade wars are often thought to be justified—not because nationalists want to hurt workers or would-be workers in other countries, but because they want to help their fellow citizens at home.

Foreign workers, as such, are invisible in the political discourse of the nationalist status quo. This accounts for the fact that the policies to which Wilkinson objects are of long standing. Trump did not invent them; they are products of the political dynamics of liberal-democratic institutions.

These policies display a callous indifference to foreigners (nationalism) but not hostility toward them (xenophobia). The fact that closed borders hurt potential immigrants is never praised, as it surely would be if substantial numbers of voters supported closed borders out of hatred toward foreigners. Rather than being praised, the fact that closed borders hurt potential immigrants goes unnoticed—just as we would expect if non-xenophobic nationalism simply blinds people within a set of national borders to the interests of those outside those borders.

Xenophobia and Trump’s Supporters

Even Donald Trump has never expressed a desire to hurt Chinese or Mexican workers. But why hasn’t he, if, as Wilkinson maintains, he is trying to stoke his followers’ rampant xenophobia? Perhaps hurting Mexicans and Chinese was a hidden subtext of Trump’s promises to “build a wall” on the Mexican border and slap tariffs on Chinese goods. But why would Trump, who feels none of the constraints of “political correctness,” keep this subtext hidden?

There are other glaring problems with the xenophobia thesis as an explanation of Trump’s support. Xenophobic voters, animated by out-group hostility, would be expected to demand a bellicose foreign policy designed to defeat the foreign enemies (real or imagined) to whom these voters are hostile. Yet Trump advocated precisely the opposite policy: foreign non-intervention. And he justified this policy as a defense of the interests of American taxpayers—the “in-group.” Thus, Trump’s non-interventionism should have gained him the support of non-xenophobic nationalists. Xenophobic voters in the Republican primaries should have supported Lindsey Graham, not Donald Trump.

So the question is: where are all the xenophobes?

Before reviewing Wilkinson’s answer to the question, let me point out what is and is not at stake in this discussion. I’m not claiming that xenophobia doesn’t exist, or that nationalism doesn’t lay the groundwork for it. I’m simply saying that there’s little reason to think that xenophobia explains Trump’s astonishing political success, and that there’s considerable (if not decisive) reason to think that nationalism does at least help to explain it. More broadly, I’m arguing that by ascribing xenophobia to large numbers of Trump supporters, one may unintentionally obscure the possibility that Trumpists are not anomalous others whose beliefs are at odds with the presuppositions of the status quo, but that their beliefs embody some of the foundational presuppositions of the status quo. Insofar as this is the case, and insofar as we disagree with Trump’s supporters about immigration, we should recognize that we are not defending the status quo. We are, instead, criticizing the status quo as insufficiently cosmopolitan—due to its thoughtless (rather than vicious) nationalism.

Possibly Spurious Survey Results

While Trump’s campaign posture on immigration was consistent with both nationalism and xenophobia, his failure to attack foreign workers wasn’t, and neither was his foreign policy. In addition, as my essay pointed out, most Americans’ opposition to immigration varies according to whether or not the immigrants have characteristics that can be expected to cause social and economic problems for their fellow citizens. Unemployed immigrants who might increase the social-welfare tax burdens on American citizens, and who otherwise might drive down wages; immigrants from terrorist breeding grounds who aren’t carefully vetted; immigrants whose English is too poor to allow them to find work and integrate into American society–these are the types of immigrants that, according to the research I cited, prompt resistance. Immigrants without these characteristics, who thus would not be expected to cause social and economic problems for “America,” are welcomed. This pattern, too, is consistent with nationalism, not xenophobia. If people were simply hostile to foreigners as such, they would not express a willingness to allow them to immigrate if they are well-educated professionals who speak fluent English.

What, then, is the evidence for xenophobia?

There isn’t much of it. It consists, primarily, of the recent finding that 30 percent of Trump’s supporters in the Republican primaries were willing to affirm that “to be of European heritage or descent” is either “fairly” or “very” important to “being truly American.” My essay asked whether this oddly worded answer (provided by the pollster) is a spurious finding. 

In response, Wilkinson gives a fair sketch of political scientists’ long-standing contention that most members of the public are too inattentive to politics to have stable or even “real” positions on many of the issues about which they are routinely polled. The body of research supporting this contention goes back to Walter Lippmann’s 1922 study of Public Opinion—perhaps the greatest work of political science ever published. Lippmann’s insights were confirmed by the “Columbia school” of opinion research in the 1940s and 1950s, and then by the “Michigan school,” which has been hegemonic since its rise in the 1960s.

Three representative high points of this body of research are:

  1. The repeatedly confirmed finding that many people are willing to give opinions about what turn out to be fictitious political issues. For example, about 30 percent opined about the non-existent “1975 Public Affairs Act.” More pertinently, in one of the first studies of opinions about fictitious issues, conducted in 1946, large majorities of the respondents were willing to express opinions about the non-existent “Wallonian” people.
  1. The 1986 argument that people answer questions about non-issues because they feel pressured to provide a response when confronted by a pollster.
  1. The 1992 challenge to the assumption that a given opinion survey is measuring anything more than fleeting, top-of-the-head considerations introduced into respondents’ minds by media messages.

In the abstract, Wilkinson acknowledges this hard-won knowledge about the vagaries of opinion surveys, but he brushes it aside when it contradicts his suspicion that Trump voters are xenophobes.

The finding about “European heritage,” if it is not spurious, does suggest that there is a significant body of potentially xenophobic Trump supporters. But as I pointed out, low-information voters, as Trump voters tended to be, are more likely to produce answers to non-issues, and they may formulate these answers by drawing on recently heard media chatter. If the chatter has it that Trump is biased against “non-Europeans,” this may have been reflected in the survey responses of some of his supporters, regardless of whether it represented their deep-seated agreement with the responses. Abundant caution, then, is in order. A lot more research would have to be done to move this finding from the “interesting” into the “dire threat to liberal democracy” column.

I also pointed out that 30 percent of Trump’s primary supporters voters amounts to 3.7 percent of the American electorate. But Wilkinson is convinced that this sliver of the public is representative of Trump’s overall support, and that the finding itself is not spurious. Why? Because in a different part of the same survey, Trump’s GOP primary supporters’ self-assessments on a feeling thermometer were warmest toward “white people” and coolest toward “immigrants.” This comports with Wilkinson’s suspicion that Trump voters are all-around bigots—irrationally hostile to “elites, globalists, members of the mainstream media, despised minorities, immigrants, Muslims, city people, pointy-headed liberal intellectuals, etc.,” whom the bigots consign “to the liminal out-group of less-than-fully-American Americans.” In short, the feeling-thermometer finding, like the one about “European heritage and descent,” is consistent with the stereotype of Trump voters as xenophobic; so both findings, for Wilkinson, “reinforce the sense that those responses are picking up on real, stable attitudes.” This seems to me to be a circular argument.

The question is whether Trump supporters’ opposition to immigration is caused by xenophobic hostility to them or, alternatively, by a desire to protect the interests of their fellow citizens from potential problems (such as the threat of wage competition). Non-xenophobic nationalism—Marietta’s “deep nationalism”—would account for the latter type of opposition to immigration. The finding about “European heritage or descent” does suggest that up to a third of Trump’s GOP primary supporters might have held xenophobic attitudes toward non-European immigrants—assuming that the finding is not spurious. But in determining whether the finding is spurious, one cannot legitimately appeal to the fact that Trump supporters are cool toward immigrants and warm toward “white people,” since such attitudes are consistent not only with the xenophobia hypothesis but the nationalism hypothesis. If immigrants are perceived as threatening the jobs or wages of one’s fellow Americans, then this perception alone might account for “cool” thermometer assessments of immigrants. (As for “white people,” I aim to explore the question of Trumpists as racists in a future essay.) To legitimize the “European heritage” response by means of the feeling-thermometer response about immigrants begs the question at issue: is opposition to immigration caused by xenophobia or by nationalism?

The “Group Theory” of Politics

Wilkinson has a larger reason for endorsing the xenophobic stereotype of Trump voters: its fit with the “group theory” of politics—which, he asserts, is “the best empirical account of political behavior going.” 

“Group theory” is the view, derived from social psychology, that we have a hard-wired propensity to attach ourselves to arbitrarily delineated groups. I don’t dispute this view, but I question whether it can explain much political behavior in general; and in particular, whether it gives us a plausible explanation of the behavior, and the attitudes, of very many Trump voters.

Group theory is not the best empirical account of political behavior going. It is one of many accounts, each of which may apply to some types of political behavior and not others.

For example, political scientists have repeatedly found that partisanship exerts a strong influence on public opinion and political behavior. Does this mean that people “identify as Democrats or as Republicans because people like us identify as Democrats or Republicans”? Maybe, but a different explanation, which in my view is more plausible, is the cognitivist, non-groupish view that party labels offer a convenient way to make sense of politics, just as ideologies do. Thus, partisans may behave in herdlike manner because they say to themselves, “people who tend to agree with me are Republicans [or Democrats], so I will follow their cue.” 

Whatever the general merits of group theory as an explanation of politics, however, it may or may not account for Trump’s surprising degree of popular support. Wilkinson argues that it does by adducing a book-length presentation of a variant of the theory: Donald Kinder and Cindy Kam’s Us Against Them: Ethnocentric Foundations of American Opinion (2010).

As Wilkinson notes, the “ethnocentrism” explored by Kinder and Kam is not equivalent to xenophobia. “Ethnocentrism” is consistent with both nationalism and xenophobia. It all depends on whether one is merely partial to one’s self-identified “us,” or is hostile to one’s self-identified “them.” Either way, one counts as “ethnocentric” under Kinder and Kam’s definition. Indeed, they conclude that Americans are only “mild[ly] ethnocentric” even when the non-xenophobic sense of ethnocentrism (in-group partiality) is counted.[1] “No one claims categorical superiority” for their ethnic group, Kinder and Kam observe; “what we have here is a sense of perceptible but subtle superiority, widely shared” among all American ethnic groups.[2]  Kinder and Kam sharply contrast their position against that of William Graham Sumner, the nineteenth-century Yale sociologist, for whom “the typical manifestations of ethnocentrism included contempt, abomination, plunder, and war. . . . No doubt ethnocentrism can take extreme form,” they write, “but we do not insist on it; and in any case, it is not what we generally find.”[3]

I wonder, however, if Kinder and Kam distance themselves enough from Sumner’s red-in-tooth-and-claw version of ethnocentrism. Kinder and Kam measure out-group “animosity” by people’s agreement with stereotypes about out-group members being either “lazy,” “unintelligent,” or “untrustworthy.”[4] (They equate in-group favoritism with the view that one’s own in-group tends to be hard-working, intelligent, and trustworthy. The measures of in-group favoritism plus out-group animosity yield an ethnocentrism score.) Yet the endorsement of these negative stereotypes does not suggest animosity, let alone hostility. We all know plenty of people whom we perceive as dumb, lazy, or shady, yet whom we do not view with animus (as opposed, in the case of untrustworthiness, to viewing cautiously). To view people as dumb, lazy, or shady is to have a negative view of them. But it is not to have a primal emotional aversion to them. 

Wilkinson puts in boldface type Kinder and Kam’s conclusion that ethnocentrism (measured by the endorsement of the three in-group and out-group stereotypes) “emerges as the single most important determinant of American opposition to immigration.”[5] Yet “single most important” merely means that it is more important than any other single factor considered by Kinder and Kam. As they say, “views on immigration are a reflection of not one thing but several: parochialism, moral traditionalism, ambivalence about equal opportunity, skepticism about the capacity of the American economy, and last, but far from least, ethnocentrism.”[6] Ethnocentrism is not more highly correlated with opposition to immigration than all these other factors combined; it’s just more highly correlated than any one of them. And out-group hostility is just half of the ethnocentrism measure. 

More important, the other factors that Kinder and Kam consider do not include negative streams of information about immigrants. As the authors note, the activation of ethnocentrism “in particular instances requires resonance: a good fit between ethnocentrism, on the one hand, and how the particular issue is framed and understood, on the other.”[7] Therefore, studying the instruments of this framing and understanding—such as the news media—might produce an even bigger correlation with restrictive immigration attitudes than does out-group “animosity.” One does not need a “deep human predisposition to reduce all of social life to in-groups and out-groups”[8] in order to worry about immigration—not if one is constantly told that immigrants are prone to terrorism or crime or that they take away American citizens’ jobs. “On balance,” Kinder and Kam write, Americans “deny that Hispanic or Asian immigrants have anything positive to add to American culture; they worry that increasing immigration will lead to more demand for public services, which in turn will drive up taxes; and they believe that increasing numbers of Asian and Hispanic immigrants will take jobs away from Americans already here.”[9] These are the messages one would get from listening to conservative talk radio or watching Fox News Channel. But Kinder and Kam did not test for exposure to such messages.

Evidence against Xenophobia as Being Decisive

Thus, the evidence for xenophobia as an important determinant of Americans’ attitudes toward immigration, and thus (potentially) of Trump voters’ choice of presidential candidate, is weak. On the other side of the coin, the Hainmueller and Hopkins study that I cited shows that opposition to (legal) immigration varies widely, depending on the prospective immigrants’ profession, employment prospects, education, and fluency in English—precisely the type of traits that would concern a nationalist attempting to pre-empt social problems for her fellow Americans, not a xenophobe trying to keep all foreigners out just because they are foreigners. I also noted that the country of prospective immigrants’ origin makes little difference to Hainmueller and Hopkins’s respondents. This suggests that they are not hostile to immigrants because of an irrational animus against people from particular countries.

Wilkinson replies by quoting the authors’ statement that respondents who are high in ethnocentrism pay more attention to country of origin than do others, but he overlooks the authors’ gigantic qualification of this statement:

Yet such effects are relatively small in magnitude and limited in scope. Once we provide information on education, language, and other factors, Mexican immigrants appear to suffer little penalty as compared to German immigrants, a finding that . . . suggests the limits of ethnocentrism-based explanations.”[10]

These limits are even more sharply exposed by the experimental study I cited from Morris Levy and Matthew Wright, who completely eliminated what had been a 5-percent bias against illegal Mexican immigrants (compared to illegal Chinese or German immigrants) by stipulating that the Mexicans had a job and spoke English. Under those conditions, the respondents became slightly more willing to allow Mexicans than Germans or Chinese with jobs and English skills to enter a program offering them legal status.

So again: where are the hordes of xenophobes?

Stereotyping, Xenophobia, and Immorality

Wilkinson offers two responses. First, he says that the initial preference of 5 percent of the Levy and Wright sample for Chinese or German immigrants bespeaks stereotyping against Mexicans—that is, stereotyping them as Spanish-speaking and jobless. This seems undeniable, as both Levy and Wright and I pointed out. But it does not show, as Wilkinson goes on to say, that these respondents displayed out-group animus toward Mexican immigrants. Such animus, which Kinder and Kam describe as a “deep habit and a stable predisposition,”[11] would surely persist after the arrival of a bit of mildly flattering information about its targets, since the animus is not supposed to be grounded in unflattering information. According to group theory, out-group hostility is a psychological disposition, not an ideational effect. If xenophobic opponents of immigration can be turned into supporters of it by telling them that immigrants will integrate non-disruptively into American society, then they were never xenophobes to begin with.

Wilkinson’s second reaction to the Levy and Wright study is to change the subject from empirics to ethics: specifically, the ethics of stereotyping. He asks, “What does it say about your view of Mexicans if you think keeping them out of the country protects your fellow Americans from crime or from paying more in taxes to finance the dole?” “What does it say about your attitude toward Muslims if you think keeping them out of the country protects your fellow Americans from terrorism?” Well, it may say that some nationalists have inaccurate stereotypes. But it does not say that these stereotypes stem from xenophobia. {Addition, September 9, 2017: A letter to the editor of the Wall Street Journal on September 7 asks, “How is wanting an orderly immigration system hateful? How is wanting to provide jobs for all Americans hateful? How is wanting to provide for the defense of our country hateful? How is wanting all Americans to have the opportunity to build a life for themselves hateful?”}

Like Kinder and Kam, Wilkinson seems to assume that only a psychologically rooted animus could produce inaccurate stereotypes. But they might also be produced by the mediated messages about immigration that are (presumably) most people’s main source of information about them.

Wilkinson constructs a thought experiment designed to show that the stereotyping suggested by the Levy and Wright study (and the Hainmueller and Hopkins study) is, to put it plainly, evil. “Laura,” a white mother, thinks that blacks are “less intelligent, less hard-working, and more unruly and violent than whites,” such that her attitudes would fit (roughly speaking) under Kinder and Kam’s rubric for out-group “animosity” toward blacks. If Laura maintained these attitudes while being willing to send her children to school with “the children of very wealthy, highly educated, high-achieving black doctors, lawyers, and executives,” Wilkinson continues, she would nonetheless remain a racist.

Perhaps so, but the important point—not just for group theory and political science, but for the analysis of whether Trump voters were motivated by nationalism or by xenophobia—is the empirical significance (or insignificance) of Laura’s racist, or groupish, or putatively deep-rooted psychological dispositions, not its ethical significance.

“Laura” expresses prejudicial beliefs, but in Wilkinson’s example, she isn’t willing to act upon them. Similarly, the fact that we can establish the presence of negative stereotypes among members of the general public answering an opinion survey does not show that these stereotypes are driving their political actions. Indeed, even if negative stereotypes (or cool thermometer readings) are rooted in psychology rather than mediated messages, this does not establish that the stereotypes are strong enough to overcome cultural sanctions against prejudicial actions, or that they are heavy enough to outweigh competing political considerations. Going from a survey response to, say, voting behavior is not as straightforward as Wilkinson assumes.

Wilkinson provides another thought experiment. “You certainly wouldn’t say that somebody who doesn’t want his kids to hang around Jewish people because he thinks Jews kidnap and murder Christian children is just looking after his kids.” But it’s a far cry from thinking that all Jews tend to be murderers to thinking that some immigrants tend to be unintelligent, lazy, or untrustworthy. If we had reason to think that significant numbers of Trump supporters, Republicans, or Americans believed that immigrants tend to be child-killers, then we’d also have a solid reason to think that these beliefs would drive their votes. But we don’t have such a reason. All we have is Kinder and Kam’s evidence that everyone tends to think that members of all ethnic out-groups tend to be less intelligent, hard working, and trustworthy than members of their own ethnic in-group. If non-Jewish Germans had merely thought such things about Jews, there would have been no Holocaust. Nazi propaganda portrayed Jews as, literally, vermin. That’s animosity, and it’s easy to see how it could lead to hateful actions.

When the schools of the South were integrated, white parents resisted because they could not tolerate the idea of their children brushing against the skin of black children. For similar reasons, Southern whites required blacks to use separate drinking fountains, to sit apart from whites on buses, and so on. That, too, is animosity. Such parents would never have consented to send their children to school with black children just because their parents were economically successful, as Linda does. We have no reason to think that Trump voters are more like those Southern parents, or like the victims of Nazi propaganda, than they are like Linda.

None of the evidence reviewed by Wilkinson justifies his assertions that there is “a drive to exclude non-white, non-Christian, non-English-speaking Americans from full and equal membership in the political community” (except by the reviled members of the alt-right); that Trump won “by whipping ethnocentric voters into a xenophobic froth”; or that “populist nationalists” deny the Americanness of “elites, globalists, members of the mainstream media, despised minorities, immigrants, Muslims, city people, pointy-headed liberal intellectuals, etc.” These assertions constitute just the type of prejudicial stereotyping, and display just the type of out-group animosity, that Wilkinson opposes.

In the meantime, by scapegoating xenophobia, such assertions distract us from the possibility that “good” nationalists may have more in common with the Trumpian other than we are prepared to admit.

Nationalism Is a Form of Group Attachment We Can Do Without

Let me conclude by addressing (at inadequate length) Marietta’s qualms about the feasibility of cosmopolitanism. He argues that in condemning Bernie Sanders for advocating protectionist policies that would hurt “foreign” workers at the expense of Americans, I miss the point: Sanders is concerned about the needs of people he can see with his own eyes; and he identifies with people closer to him than those farther away. In part, however, that is my point: morally speaking, there is no justification for discrimination between the seen and the unseen or the near and the far-away. My main reason for pointing out the arbitrariness of these distinctions is to estrange us from the “common sense” that accepts them as natural—just as it accepts modern nation-states, deliberately created during the last two centuries, as natural.

However, another part of my point is that nobody in the modern world can actually see their conationals with their own eyes: there are far too many conationals. We have the illusion that we see them and understand their needs because we are shown snippets of a tiny fraction of them on TV, and encounter a tiny fraction of them in airports and grocery stores, and read about a tiny fraction of them in news stories (and opinion surveys). Despite the illusion, however, they are still, on the whole, anonymous strangers. That some of these strangers live a mile away from me does not give them much more in common with me than strangers who live on the other side of the world. Whether a national border does or does not lie in the space between them and me does not make them any the less strangers.

Friends, family members, and neighbors (those neighbors whom we know) are in an entirely different category than these anonymous others. We actually know them and we interact with them. We form emotional attachments to them by virtue of this knowledge and interaction, and by virtue of the biological drives that bring us together. Now if one accepts psychological group theory (as I do), one will acknowledge that there are also biological drives at work when we attach ourselves to nation-states. But the same is true when we attach ourselves to sports teams and other arbitrary groups. These attachments, like other biological drives, can be overcome or displaced if we view them as undesirable. Again, the question is the strength and permanence of these drives, none of which is established by the psychological literature of which I’m aware.

The foundational study in the “groupish” literature, Muzafir Sherif’s “Robbers Cave” experiment, found that boys brought to a summer camp and separated arbitrarily into two groups spontaneously developed both in-group attachments and hostility toward the other group. But they presumably went on to live full lives that were not dedicated to the eradication of their summer-camp foes. Had they later run into one of those foes, they surely would have laughed at their youthful foolishness. It would be crazy to suggest that because their group attachments were enabled biologically, they were somehow rendered impervious to rational restraint.

Everything people do is enabled biologically, but that doesn’t license us to do everything and it doesn’t force us to do everything we’re capable of doing. We can drop our arbitrary group attachments for rational reasons. That I grew up in Chicago was arbitrary, and so is my attachment to the Chicago Cubs. But this attachment brings me pleasure (occasionally) and it hurts nobody, so I allow it to continue. My more intimate attachments bring me much greater pleasures and, to the best of my ability, I allow them to continue only as long as nobody is getting hurt. Nationalism does hurt people—usually, the least advantaged. We can choose therefore to treat it as an attachment worth shedding. We can choose to get our groupish jollies from sports instead of politics.

Only the confusion created by analogies between total strangers and neighbors and friends stops us from doing so—along with the cultural indoctrination we’ve all received at the hands of schoolteachers, the media, and so on. Once we see clear of this confusion we may recognize that Trumpism represents not a monstrous perversion of modern politics, but an expression of some of its most blandly familiar features.

Kinder and Kam do not discuss the in-group of “Americans” versus “foreigners”; the groups with which they are concerned are ethnic groups within America.
[2] Donald R. Kinder and Cindy D. Kam, Us Against Them: Ethnocentric Foundations of American Opinion, University of Chicago Press, 2010; p. 57.
[3] Ibid.
[4] Ibid., p. 45.
[5] Ibid., p. 139.
[6] Ibid., p. 131.
[7] Ibid., p. 222.
[8] Ibid., p. vii.
[9] Ibid., p. 133.
[10] Jens Hainmueller and Daniel J. Hopkins, “The Hidden American Immigration Consensus: A Conjoint Analysis of Attitudes toward Immigrants,” American Journal of Political Science 59(3), p. 530.
[11] Kinder and Kam 2010, p. 35.

Jeffrey Friedman, the Director of the Niskanen Center’s Institute for the Study of Politics and 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).