Political Realism: An Essay on the Politics of Value Conflict

The Ithiel de Sola Pool Award is for work throughout my career—work that I have done with many and with the help of so many more. I thank the committee for this honor and, with it, the opportunity to honor another: Merrill Shanks.

My aim is to be a spokesperson for an idea: political realism. 2 The root intuition is that to understand the choices that citizens make, the starting point should be the circumstances in politics to which they are responding. This may seem self-evident. Would that it were, for what I show is that political scientists in the same line of work as I am, have gotten caught up in analyzing the psychology of politics at the expense of politics.
What work is "realism" in political realism doing? My guide is G. H. Hardy. Mathematics is so, Hardy declared, "not because we think so, or because our minds are shaped in one way rather than another, but because it is so, because reality is built that way" (1940, 130; italics in original). I make the same claim for politics twice. Liberals who respond to the Democratic Party as the party of liberalism on civil liberties and conservatives who respond to the Republican Party as the party of conservatism are not engaging in motivated reasoning. Neither are liberals who believe that traditional Muslim values regarding the rights and responsibilities of women conflict with liberal values. On the contrary, in these quite different circumstances, liberals respond as they do "because [political] reality is built that way" (Hardy 1940, 130). Reality bites-that is my claim. I begin with a motivating example: the "Big Sort."

A MOTIVATING EXAMPLE
American party politics is intensely polarized, all agree. Republicans fear and loathe Democrats, Democrats fear and loathe Republicans, if not more than ever before-at any rate, far more so than 30 years ago (Iyengar, Sood, and Lelkes 2012;Iyengar and Westwood 2014). A small library of studies reports that political feelings spill over into their personal lives-who they have as friends, prefer to date, and would like their children to marry and what they elect to listen to and read. 3 If Democrats like to live their lives with Democrats and Republicans with Republicans, it is a small step to suppose that each would like to live next to-or, at any rate, nearby -one another. So, when they move, they sort themselves into politically agreeable communities-or so it is claimed.
Why do people make the choice of where to live that they make? Because they prefer what they prefer-in this case, to live in politically like-minded communities. This is how the narrativeacademic as well as popular-of the Big Sort goes. But people searching for a new home, Mummolo and Nall (2017) caution, have many preferences. 4 Living with politically like-minded neighbors may be one, but many other considerations matter more: affordability, a safe neighborhood, good schools, and living close to work among them. Democrats and Republicans alike prioritize neighborhood quality and affordability. 5 People do not have endless time to find a new home; they must choose from what is available. As Mummolo and Nall (2017, 57, italics added) rightly noted, "[P]otential movers with a stated propensity to move to more copartisan neighborhoods, regardless of their underlying motivations, are more likely to have few choices [to live in copartisan communities] if they select first on affordability and quality." The Big Sort narrative fails to reckon with reality-grounded constraints on the alternatives available.

THE ROAD TAKEN
The dominant approach in the study of support for civil liberties is the "least-liked" method of John Sullivan, George Marcus, and their colleagues. Political Tolerance and American Democracy (Sullivan, Piereson, and Marcus 1982) and With Malice Toward Some: How People Make Civil Liberties Judgments (Marcus et al. 1995) are their two landmark works. Because I advocate for an alternative approach, it is even more important to make plain why their work was field defining.
Political tolerance, Sullivan and his colleagues observed, implies a willingness to "put up" with those ideas and people that one intensely dislikes or with whom one intensely disagrees. It is necessary, therefore, to first ensure that people are being asked about groups and ideas that they intensely dislike or with which they intensely disagree. This was simple but brilliant. Like the best ideas, it appears to be just right-after someone has had the creativity to point it out. It revolutionized the study of political tolerance because the least-liked approach is simultaneously a theoretical insight and a measurement procedure. From a list of controversial groups, respondents are asked to choose the group that they dislike most (or like least). Then they are asked whether this group is entitled to a series of rights-for example, whether they should be allowed to teach in public schools, or hold public rallies, or have their phones tapped by the government. The leastliked measurement method has become the preeminent-indeed, in the view of many, the mandatory-method of measuring political tolerance.
Measurement only gets a research project off the ground. Methods must be substantively enlightening. Why has the leastliked method become the dominant way to think about and measure political tolerance? What do we now know that we did not know thanks to the least-liked method? Two standout findings are claimed for it.
The first is the illusion of progress. In his pioneering study of political intolerance at the height of McCarthyism, Samuel Stouffer showed that popular support for civil liberties-although impressively high at the level of abstract principle-melted away in the heat of actual controversies. Even so, reflecting on the changes underway, Stouffer (1992) offered a famous prophecy: "Great social, economic, and technological forces are working on the side of exposing ever larger proportions of our population to the idea that 'people are different from me, with different systems of values, and they can be good people, too.'" Subsequent studies during the 1960s and 1970s showed that levels of political tolerance markedly increased (Davis 1975;Nunn, Crockett, and Williams 1978).
Still better times ahead was the forecast, as increasing proportions of Americans became better educated, more geographically mobile, and more exposed to Americans of diverse backgrounds and outlooks. However, Sullivan, Piereson, and Marcus (1982) discovered "a fly in the ointment." Three decades after Joseph McCarthy's heyday, communists were no longer the agreed-on threat. Ask Americans about the particular groups that each fears now and they prove to be as politically intolerant as they were in the 1950s. This is a devastating result. The appearance of progress is an illusion. The optimism of the first generation of researchers is a hollow hope. The only thing saving us, Sullivan and his colleagues argued, is the happenstance that we disagree on who we dislike most.
Similarly striking is the second result. It had been supposed that the liberal tradition in American politics provides a measure of support for democratic rights-and, indeed, using traditional measures of support for democratic rights, liberals are more committed than conservatives to an open society. However, if everyone is asked about the group they dislike most, liberals choose groups on the right, conservatives choose groups on the left, and both come up short in supporting the rights of those on the other side of the political fence. Threat perception, it follows, is the key to the politics of tolerance. 6

THE ROAD NOT TAKEN
In 1960, Herbert McClosky, Paul J. Hoffmann, and Rosemary O'Hara published "Issue Conflict and Consensus Among Party Leaders and Followers." It was the first large-scale study of the political beliefs and ideological impulses of politically active and influential people. Capitalizing on an innovative survey about the policy beliefs of convention delegates to the presidential nominating conventions in 1956, McClosky and his colleagues brought into the open for the first time the ideological cleavage between the two parties. Democratic delegates were systematically liberal and Republican delegates were systematically conservative over an array of issues, including public ownership of natural resources, government regulation of the economy, equalitarianism and social welfare, tax policy, and foreign policy. 7 Summarily stated, McClosky et al.'s (1960) first landmark discovery was that the politics of the American party system at the elite level is ideological politics. 8 Reality bites. For voters, the choice is between Democrat/ liberal and Republican/conservative. Of course, because the My aim is to be a spokesperson for an idea: political realism. The root intuition is that to understand the choices that citizens make, the starting point should be the circumstances in politics to which they are responding. This may seem self-evident. Would that it were, for what I show is that political scientists in the same line of work as I am, have gotten caught up in analyzing the psychology of politics at the expense of politics.
political choices are on offer on these terms, it does not follow that these are the terms on which citizens make their choices. They may mistake or misunderstand what the terms are, supposing that the liberal alternative is the conservative, the conservative is the liberal; or fail to perceive that the liberal-conservative divide runs across issues of civil liberties. Instead, they may respond to them, one by one, as if each is separate and distinct; or take the side that their party advocates simply because their party advocates it; or, even, choose at random a side on an issue because they have not troubled themselves to form a genuine attitude about it. These possibilities are not merely possible but probable according to the mainstream view that citizens are not capable of engaging politics on ideological terms. They give too little attention to politics, have too little information about it, and are at sea when it comes to knowing what goes with what-let alone why. Their political ideas often are jumbled-if, indeed, they qualify as genuine ideas, given how many citizens appear to choose one side of an issue rather than the other by flipping a coin (Achen and Bartels 2016;Berelson, Lazarsfeld, and McPhee 1954;Campbell et al. 1960;Converse 1964;Kinder and Kalmoe 2017;Zaller 1992). Generations of PS101 students have been taught that it is a rookie mistake to suppose that ordinary citizens can engage politics on an ideologically coherent basis-that is, consistently taking liberal or conservative positions on the issues of the day. That, as the saying goes, is the conventional wisdom.
Hence the importance of McClosky and colleagues' subsequent discoveries McClosky and Zaller 1982). Purely as a matter of fact, a large portion of the general public has ideological convictions. Those on the left systematically evince an openness to social change, confidence in human nature, and concern for the disadvantaged and distressed. Conversely, those on the right stress the risks of change, the weaknesses and selfindulgence of people in general, and the necessity of individuals to take responsibility for their own lives and overcome their own problems rather than striving for a handout from government. 9 Moreover, liberal and conservative convictions pervade the political thinking of ordinary citizens, systematically showing up in ideologically coherent stances even on issues of civil liberties and civil rights.
How this is possible I (prudently) leave until the conclusion of this essay. What is vital here are evidentiary credentials for the claim that ordinary citizens take ideologically coherent stances on issues of civil liberties. Figure 1 reproduces a finding from McClosky and Brill's (1982) study of civil liberties-one of a mass of comparable results. 10 The top panel shows the relationship between ideology (self-designation) 11 and a multi-indicator measure of support for civil liberties for a large sample of community leaders; the bottom panel shows the same relationship for ordinary citizens. 12 Consider first the impact of ideology on support for civil liberties among community leaders. The result is arresting: 87% of liberals score high on the multi-item Index of Civil Liberties, compared to only 35% of conservatives. Conversely, 24% of conservatives score low, compared to only 4% of liberals. It is not a surprise, of course, that the judgments of the politically engaged about civil liberties and civil rights are ideologically grounded. The

Score on Civil Liberties Scale:
Middle Low surprise is shown on the bottom panel: 66% of liberals score high on the Civil Liberties Index, compared to only 23% of conservatives. Conversely, almost half of conservatives (42%) score low, compared to only 11% of liberals. The relationship between identifying as a liberal or as a conservative and consistently taking liberal or conservative positions on issues of civil liberties is not as strong among ordinary citizens as it is among community activists. Nevertheless, it is unmistakable. This should have been a breakthrough discovery-but no. Like Bishop Berkeley's tree falling unseen in the forest, in research, esse est percipi-to be is to be perceived. McClosky's demonstrations of the role of ideology in the political thinking of ordinary citizens have remained unremarked. Even now, in the most ideologically polarized era in American politics in more than a half-century, the mainstream view is that Converse's (1964) analysis of the 1950s remains the best mirror of the political thinking of Americans in the 2020s. "Most Americans," Kinder and Kalmoe (2017, 41) concluded in their recent review of research, "are indifferent to or mystified by liberalism and conservatism as political ideas." Wrong and a pity. The naked clash of conservative and liberal values is the key to the ferocity of contemporary American politics. Suppose you are a liberal and I am a conservative. If you are truly a liberal, you are not liberal on only one issue but rather on a whole array of issues-and very likely a Democrat to boot. Ditto for me. If I am truly a conservative, I am conservative not on only one issue but on a whole array of issues-and very likely a Republican to boot. We disagree about many things, so we disagree many times-and just so far as our disagreement is grounded in opposing political orientations, our quarrel goes all the way down.
There is an irony here. According to many political scientistsand still more public pundits-the polarization of American politics is a political fever, the product of raw emotion. This is true for some, but as great a risk for as many or more is the satisfaction of being true to their principles. They know and agree with what their party stands for. Their political faith imprisons them just because it is not blind. So they stand with their side, come what may. Ideological politics is passionate politics because core values are at stake.

INCLUSION OF MUSLIMS AND CONFLICTS BETWEEN LIBERAL VALUES
To understand the choices that citizens make, the starting point should be the choices open to them. In the American party system, the conflict between liberal versus conservative values delimits the choices open for citizens to make on issues of civil liberties. A second illustration of why-to account for the choices that citizens make, the starting point should be the circumstances to which they are responding-is the conflict not between conservative and liberal values but instead between liberal values themselves.

Current research on the inclusion of Muslims documents an array of instances in which majorities of non-Muslims in Western
Europe support stripping Muslims of democratic rights (Ivarsflaten and Sniderman 2022). 13 The majorities are one sided, often 75% or more refusing to take the Muslims' side. The countries are all liberal democracies-indeed, among the most supportive of an expansive understanding of the rights of citizens. Moreover, the rights at issue are bedrock democratic rights (i.e., freedom of religion, freedom of speech, and freedom of assembly) confronted not in the abstract but in real life-for example, being allowed to rent a meeting room in a public library.
This may seem an old story. Conflicts between liberal democratic values and some traditional Islamic tenets are well documented (e.g., Norris and Inglehart 2002;Sniderman and Hagendoorn 2007). Value conflicts on this order diminish support for the rights of Muslims; however, they do not uproot support for Muslims' right to hold a public event. Something more is needed to mobilize a one-sided majority against Muslims being allowed to rent a meeting room in a public library.
Conflict with religiously grounded values goes deeper than conflict with cultural values. The gateway question is why. Islamophobia is the "off-the-shelf " answer. Imams are a stereotype magnet-zealots, in the eyes of many non-Muslims, promoters of a crabbed, misogynistic, parochial way of life and religion. Hence, there is an immediate readiness of Islamophobes to oppose Muslims being allowed to rent a meeting room in a public library.
This reasoning is not wrong; it is elliptical. Islamophobes will oppose Muslims being allowed to rent a meeting room in a public library but not because conservative Islamic ideas about the rights and responsibilities of women is the topic. Islamophobes cannot stand Muslims. They oppose their being allowed to use a public library to hold a meeting pretty much whatever the topic. However, Islamophobes are a minority-an angry, agitated minority, but a minority all the same. What needs explanation is why onesided majorities can be mobilized in opposition to Muslims being allowed to hold a public meeting.
Reality bites is the motif. The fact (and it is a fact) that conservative Muslim ideas about women are religiously grounded values is the "something more." The insight follows from the theory of speech acts-specifically, the concept of performative utterances. Performative utterances are instances of when "the issue of the utterance is the performing of an action." Two iconic examples illustrate the idea: "(E. a) 'I do (sc. Take this woman to be my lawful wedded wife)'-as uttered in the course of the marriage ceremony. (E. b) 'I name this ship the Queen Elizabeth'-as uttered when smashing the bottle against the stern." (Austin 1962, 5) Speaking the words "I take this woman to be my lawful wedded wife" is not merely to say something; it is to do something. To "Most Americans," Kinder and Kalmoe concluded in their recent review of research, "are indifferent to or mystified by liberalism and conservatism as political ideas." Wrong and a pity. The naked clash of conservative and liberal values is the key to the ferocity of contemporary American politics. name a ship is not merely to say something; it is to do something. So, too, is a profession of faith. Both a rabbi declaring "Jews should keep kosher" and an Imam proclaiming "A Muslim should not shake hands with a member of the opposite sex" describe what should be done and call for action to be performed.
Professions of faith are necessarily performative utterancesinstances of action as well as expression. Conflicts with religious values thus have two dimensions: substantive and performative. To this point, all eyes have been fixed on the substantive dimension-on the clash between liberal ideals of the larger society and traditional ideals of conservative Muslim communities. What now deserves attention is the performative dimension. Just as affirmations of Islamic ideals that conflict with the liberal ideals of the larger society are calls for compliance, the intensity of the conflict escalates. The combination of what is being urged and the fact that it is being urged generates extreme value conflicts.
Rights are the concern: bedrock claims to freedom of speech and religion. In The Struggle for Inclusion, Ivarsflaten and Sniderman (2022) introduce the public rally experiment. The design is 2x2: one factor varies the substantive topic of a rally, uncontroversial or controversial; the other factor is its function, communicative or performative. A nationally representative sample of Norwegians was asked if Muslims should be allowed to hold a public event. The topic of the event was either "Islamic values" or "conservative ideas about women"; its function was either "to explain" or "to preach." The reasoning underpinning the design is straightforward: set as a baseline, a low-conflict condition between liberal ideals and traditional Islamic values-a meeting to explain Islamic values. Then, to determine whether the substantive and performative dimensions independently increase the intensity of conflict, there were two more conditions: a public event either to explain "conservative ideas about women" or "to preach" Islamic values. The fourth condition was to determine the combined effect of substantive and performative dimensions: a meeting the purpose of which is to preach rather than to explain, and the topic is "conservative ideas about women" rather than the vaguer "Islamic values." Figure 2 shows the results of the experiment. The fourth condition stands apart. In the other three conditions, only a minority opposes allowing Muslims to hold a public rally; in the fourth condition, a majority-indeed, a clear majority: two in every three are opposed. This is an extraordinary turnabout, a volte-face, from decisive support for Muslims being allowed to hold a public rally in the baseline condition to decisive opposition to their doing so in the combined performative/substantive condition.
What accounts for this volte-face, a clear majority in support of Muslims holding a public rally transformed into a clear majority in opposition? The impulse answer is intolerance beneath the surface appearance of tolerance. Figure 3 shows that it is the other way around. The dotted line traces the likelihood of opposing Muslims being allowed to hold a public rally in the baseline condition. The solid line is the likelihood of opposition in the high-conflict condition, conditional on level of tolerance. 14 The gap between the two lines increases as levels of tolerance increase. Those who are intolerant oppose Muslims being allowed to hold a public rally regardless of the topic or purpose. Those who are tolerant overwhelmingly support their being allowed to do so in the lowconflict condition and overwhelmingly oppose it in the highconflict condition.
The dictum of political realism: to explain why people respond as they do, first determine the circumstances to which they are responding. In the low-conflict condition, the tolerant must choose between allowing Muslims to hold a public event to explain Islamic values or opposing it. This is an easy choice for those who are tolerant: say yes, obviously. In contrast, in the highconflict condition, they must choose between supporting equal rights for Muslims and equal rights for women. Support for equal rights for Muslims and equal rights for women are mutually reinforcing in most circumstances. Moreover, prioritizing equal rights for women is not a manifestation of intolerance. The outcome nevertheless is illiberal, a denial of Muslims' right to practice their faith. An illiberal outcome, because of support for a liberal value, sounds like a contradiction in terms. In fact, it is testimony to the constraints of political reality.
That is the lesson I tentatively take away from these results. Why tentatively? Because much remains to be done. First, this is the result of one experiment, in one country, at one point in time. Replication is a requirement. Second, there is a self-evident objection. Perhaps those who are tolerant do believe in equal rights for Muslims but this belief is superficial, easily overcome by any rival consideration. Third, and still more concerning, those who are tolerant do not truly believe in equal rights and are only saying what they believe they are supposed to say.
Skepticism is easily aroused. It is not easily quieted. What might provide confidence that those who affirm liberal ideals such as tolerance, yet in particular circumstances oppose equal rights for Muslims, do so on principle? Islam is viewed in Western Europe against a swill of ill will and suspicion of bad faith (Ivarsflaten and Sniderman 2022, 116-24). It surely is reasonable to ask whether what we have observed are responses to the profession of a religious faith or to the profession of Islam in particular as a religious faith.
Research is underway to answer this question (Bjånesøy et al. 2023). The logic of a "sincerity" test is worth describing. The overarching claim is that conflicts with religiously grounded values can have two dimensions-performative as well as substantive-and the combination of what is being urged on religious grounds and the fact that it is being urged on religious grounds escalates the conflict. If this is correct, then the same claim on behalf of a religion other than Islam should generate a similar reaction-hence, the design of the public library experiment.
In the public library experiment, a group asks to rent a meeting room in a public library rather than to hold a public event. This provides a pretext. There is a right to hold a public rally; there is no right to rent a meeting room in a public library. The group asking to rent a meeting room is randomly varied. Half of the time, the purpose of the meeting is "to preach conservative Islamic values that homosexuality is a sin." The other half of the time, the purpose of the meeting is "to preach conservative Christian values that homosexuality is a sin." Islamic/Christian: that is the only difference.
It was a foregone conclusion that there will be massive opposition to Muslims being allowed to rent the library meeting room to preach that homosexuality is a sin. It is similarly obvious that in a culture as liberal as Norway's, there will be opposition to Christians preaching that homosexuality is a sin. The question is: "Is the opposition as one-sided and as massive to Christians doing the same thing as Muslims?" Figure 4 shows that the answer is yes. Only a small minority (no more than 20%) of Norwegians believe that Christians should be allowed to rent a meeting room in a public library to preach that homosexuality is a sin. To be sure, on conventional tests of statistical significance, a numerically smaller minority believes that Muslims should be allowed to do the same. This is a poster-prize example of the mistake of relying on conventional tests of statistical significance. Figure 4 shows that 19% (þ/-2.0) of Norwegians are willing to rent to Christian fundamentalists, compared to 14% (þ/-1.9) who are willing to rent to Muslims-a difference that is not a difference. The sticking point is not the profession of Islam as a faith. It is the profession of faith.

REALISM: THE MISSING STEP
Mine is a modest proposal: to acknowledge that reality bites. What might it mean to accept it? At the end of a characteristically penetrating and self-assured analysis of the psychology of choice, Simon (1985, 302) upset the apple cart. "We need to understand," he declared, "not only how people reason about alternatives, but where the alternatives come from in the first place." The take-away point is that not one but rather two theories are needed-one to account for choice that is conditional on the alternatives available; the other to account for the alternatives available at the moment of choice.
Political realism can take us only so far. One of its virtues, however, is to call out a key mechanism: substantive coordination of alternatives. Consider this old chestnut: the "principle-policy" Motivation to control prejudice (0=lo; 1-hi)

Opposition to Muslim Rally by IMCP
Opposition to the public rally by internal motivation to control the prejudice index. puzzle. Racial equality at the level of principle is overwhelmingly supported; however, policies to achieve it are sharply contested. This gap between principle and policy, it is asserted, is evidence of the vacuousness of declarations of support for racial equality. White Americans say what they believe they ought to say-but when push comes to shove and an actual effort must be made to accomplish racial equality, they reveal their true colors. True for some, no doubt. 15 Still, realism counsels, ask first: What are the alternatives that citizens can choose, when they choose? The policies advertised to achieve racial equality are liberal policies. Why is support for equality at the level of policy rather than at the level of principle? Primarily because conservatives balk at supporting liberal policies (Sniderman, Brody, and Kuklinski 1992)-which is lamentable supposing that liberals are right. But why should conservatives presume that liberals are right? What is telling is the low probability that political psychologists assign a claim that conservatives reject liberal policies on race because they are conservative. 16 Politics, for some political psychologists, is substantively content free. Opposition to liberal policies is driven by racism of one variety or another. For a political realist, a relevant point is that the politics of race is a province of the politics of the American party system, and the polarized politics of the American party system is ideological politics. The Democratic party campaigns under the banner of liberalism and the Republican party campaigns under the banner of conservatism, and both are competing for support across the contested issues of the day. Racism is far from extinct-but in an ideological politics, liberals support liberal policies and conservatives support conservative politics on issues of race as well as social welfare. Substantive coordination across as well as within policy domains is the point. This is how far I have gotten and no farther, without progress in accounting for the dynamics of coordination, still less the generation of alternatives. After a career elaborating a theory accounting for choices conditional on the alternatives on offer, I find it difficult to escape the irony that in an essay arguing for political realism, I have made a case for the need of a theory that can account for what alternatives are on offer. 17

CODA
The lecture given in honor of Ithiel de Sola Pool is for a career of work. The lecture itself is a case study of how, as a result of collaboration, one thing has led to another. The focus of my current work is the conflicts between liberal values. But there would be no point in studying how liberals react to conflicts of liberal values if I were not convinced that liberal values have taken hold. I owe this conviction to participating in a study of the Cartoon Crisis of 2005 and 2006 in Denmark. I joined this study only after the creative work-hitting on the very idea of the project, designing the key experiments, superintending the data collection, and discovering the key result-had been done.
The result that my colleagues-Michael Bang Petersen, Rune Slothuus, and Rune Stubager-had discovered was the reason that I joined the project. The Cartoon Crisis had transfixed Danish politics. Even at the most intense moments of the crisis, ordinary Danes, they revealed, were as supportive of the civil liberties of Muslims as of fellow Danes (Sniderman et al. 2014). This result astonished me, and I owe my conviction that liberal values are taking hold to it. Standing up for the rights of Muslims, in the heat of the firestorm over the cartoons, is evidence of a "stand-your-ground" commitment to inclusive tolerance.
The democratic experiment remains a bet. For it to continue to pay off sometimes calls for faith that it will continue to pay off. Faith that citizens can keep their moral balance even under stress has underpinned my research since the Cartoon Crisis study-a satisfactorily paradoxical conclusion, I think, for a lecture titled "Political Realism." This is how far I have gotten and no farther, without progress in accounting for the dynamics of coordination, still less the generation of alternatives. After a career elaborating a theory accounting for choices conditional on the alternatives on offer, I find it difficult to escape the irony that in an essay arguing for political realism, I have made a case for the need of a theory that can account for what alternatives are on offer.