Can We Improve Political Discourse?

Seven years ago, I wrote The Three Languages of Politics because I was concerned with the state of political discourse. It struck me that both professional and amateur political commentators were expressing opinions in ways that were inflammatory and unconvincing. Subsequently, two more editions of the book have appeared, most recently this past August. Meanwhile, other observers have become interested in political polarization and tribalism, in part because the problem has gotten worse.

New analyses of polarization keep appearing, and new signs of the severity our political fault lines keep emerging. As a result, I sense that the latest edition of The Three Languages manages to be both timely and out of date. This essay will sketch some of the book’s key points, and then I will offer my current thinking, particularly concerning cultural forces that I think are behind the surge in polarization, and what we might do to try to counter those forces.

The basic problem with contemporary political discourse is that we aim to demonize rather than to persuade. In Persuasion Mode (think of a high school debate team), we treat people on the other side with respect, we listen to their presentations of facts and logic, and we try to assemble facts and logic that would influence them to change their minds. In Demonization Mode (think of Paul Krugman describing Republicans or Rush Limbaugh describing Democrats), we tell anyone who will listen that people on the other side are awful human beings.

Observe the political opinions that you see expressed on Facebook or Twitter, and you will find that nearly all are in Demonization Mode. The posts that are widely shared are not oriented to changing the mind of someone who disagrees. Instead, posts are written to appeal to people who agree, with the effect of hardening their opinions and closing their minds.

The Three-Axes Model

Consider three words: oppression, barbarism, coercion. Oppression occurs when one group in society is abused by a more powerful group. Examples include slavery in the United States and Jim Crow segregation. Barbarism occurs when people engage in engage in acts that were once common but now seem abhorrent. Examples include torture, rape, and pillage. . Coercion occurs when people do something not voluntarily but under threat of violence. Examples include paying taxes. All three words have negative connotations. We are against all of them.

But progressives, conservatives, and libertarians differ in their preferences for framing issues. Progressives often describe political life as a moral struggle against oppression, in which people with other points of view are on the side of the oppressors. Conservatives often describe political life as a moral struggle to preserve civilization, in which people with other points of view are on the side of barbarism. Libertarians often describe political life as a moral struggle to preserve liberty, in which people with other points of view are on the side of coercive government power.

These different axes can be used to frame many issues. Progressives, conservatives, and libertarians can express divergent opinions, with each convinced that other points of view are morally flawed.

For example, consider the story that emerged last year concerning professional football players who knelt during the national anthem. For progressives, the oppressor-oppressed framing suggested that since African-Americans are a historically oppressed class, those who were against the kneeling were oppressors. For conservatives, the civilization-barbarism framing suggested that since the American flag and the national anthem are symbols of our civilization, showing disrespect represented support for barbarism. For libertarians, the ritual of playing the national anthem and standing at attention amounts to state-worship, and if it went away altogether that would be for the best.

Framing an issue in terms of your preferred axis is useful for demonization, but not for persuasion. As a progressive, you can score points with other progressives by labeling others as oppressors, but that is unlikely to prove persuasive to conservatives or libertarians. Similarly with conservatives labeling others as barbarians or libertarians labeling others as statists.

Psychological Foundations for Political Tribalism

The three axes are heuristics or shortcuts for framing political issues. We use these heuristics because it takes less effort to have a closed mind than an open mind.

We seem to experience political disagreement as a threat, like seeing a snake or a tiger. We react against this threat.

One reaction is called confirmation bias. Suppose that we oppose raising the minimum wage. When we see a study that purports to show that raising the minimum wage would be socially beneficial, we search for methodological flaws in the study. If instead the study purported to show that raising the minimum wage would be harmful, we do not worry about the methods and just tout the study as providing important evidence.

Another reaction is called Asymmetric Insight. We claim to know the true motives of our opponents better than they know themselves. Moreover, those motives are bad. (“They just want to serve big corporations.” “They just want to tear down America.” “They just want more power to direct our lives.”)

If we treat those with whom we disagree as having bad motives, then that saves us the trouble of having to pay attention to the substance of what they are saying. What I call Demonization Mode usually serves this purpose.

In addition to individual psychology, social psychology also impels us toward demonization. Our prehistoric ancestors feared being excommunicated from the tribe and being left in the wilderness to die. On the other hand, they were rewarded for acts of tribal loyalty. If you are a Republican who finds yourself in a group of associates consisting of Democrats, you will be inclined to mute your own views in their presence, rather than risk excommunication. But if you find yourself among fellow Republicans, your inclination will be to state your hatred of Democrats in extreme terms, because doing so has some upside and little downside.

Political Tribalism Driven by Recent Cultural Trends

The psychological impetus for political tribalism is always present. To explain why the problem seems to have gotten worse in recent years, we need find some other factors. Two trends that I suspect are important are cultural sorting and social media.

Cultural sorting is the phenomenon in which we associate less with people of different backgrounds, especially with respect to education. Today people with college degrees interact professionally and socially with people without college degrees much less than they used to. General Motors had a large cadre of workers who were not college educated. Google does not. Fifty years ago middle managers worked with secretaries. Today they do not. Fifty years ago recent college graduates befriended neighbors without college educations. Today they live in separate enclaves.

Prior to the 2018 congressional elections, the Wall Street Journal reported on a poll showing that white women with a college degree differed sharply from white men without a college degree.[1] The mutual dislike between these two demographic groups is illustrated by democratic presidential candidate Elizabeth Warren’s joke that a man who believes that marriage should be between a man and a woman “just marry one woman…assuming you can find one.”[2]

Indeed, American literature and movies have long treated the subject of the sexual tension between the highbrow woman and the rough-hewn man. The classic Bogart-Hepburn film The African Queen comes to mind as one example. More recently, Curtis Sittenfeld’s latest collection of short stories included a tale of a fraught sexual encounter between a feminist academic and the Trump-supporting driver who takes her from the airport to the hotel where she is participating in a conference.[3]

One consequence of what Bill Bishop termed The Big Sort is that in many congressional districts an incumbent is more vulnerable to being unseated by a primary challenger than by a member of the opposite party. This reinforces polarization.

The other major factor that I believe has worsened political discourse is social media. People process social media in a rapid-fire, emotional way. Facebook and Twitter are not conducive to reflective, nuanced thinking.

Both traditional media and social media face incentives to maximize consumer engagement, which in practice means stoking outrage. This favors Demonization Mode, not Persuasion Mode.

I recall that during the Vietnam War and the Watergate scandal, public opinion gradually shifted as the stories developed and were reported in the media. Today, the media lack the credibility to create a major shift in public opinion, and instead the major outlets just act to harden opinions on their respective sides. This can be seen in the current impeachment controversy, in which there appear to be two completely incompatible narratives, with left-leaning media saying that the president’s behavior with respect to Ukraine “crossed the line” and right-leaning media muttering about an “endless attempted coup.”

As a result of these trends, the perceived psychological stakes in politics have gone up dramatically, even though polarization on substantive issues has not increased nearly as much. In The Three Languages, I cite the work of political scientist Lilliana Mason. She found that although Democrats and Republicans have moved only slightly farther apart on policy questions in recent decades, but hostile feelings toward the other party have risen dramatically.

Populism: Is There a Fourth Axis?

I am not sure what to make of populism, a worldwide phenomenon that is represented in the United States by the election of President Trump. In the latest edition of The Three Languages, I suggest that we might need a fourth axis, one of cosmopolitan elites on one side and their opponents on the other.

But on further reflection, I doubt that the elite/anti-elite divide is comparable to the three axes of my original model. For one thing, we can agree that oppression, barbarism, and coercion are all bad. But there is no consensus that it is bad to be elite or that it is bad to be anti-elitist.

Another difficulty with putting populism on an axis is that populism is not very coherent. As Martin Gurri points out in The Revolt of the Public, populists express what they are against with great fervor, but they are far from clear on what they are for.

As libertarians, we might be happy with a movement that is highly critical of the politically powerful. Indeed, if populist anger could be directed toward supporting checks and balances along with constitutional limits on government, we would welcome it. But if the history of Latin America is any guide, populism will instead be channeled into support for a charismatic demagogue.

President Trump: A Negative Role Model

Regarding the Trump Presidency, I am one of the few people I know who is ambivalent. Everyone else seems to be either much more favorable or much more unfavorable.

But in terms of political discourse, I see President Trump as a negative role model. He personalizes politics, and he tries to demean his opponents.

If I were teaching students to engage in Persuasion Mode, I would tell them not to personalize class discussions. Once an idea is raised , let it float away from the individual who raised it. Do not judge the idea based on the person who proposed it, and do not judge an individual based on the idea that the person proposed. Treating ideas as impersonal helps to encourage rational discussion and to attenuate emotional reactions.

Suppose that a student makes an offensive statement. I would invite the student to explain and clarify his or her remarks. My assumption is that the offensive statement is an attempt to test the teacher or cause trouble. By asking the student for clarification, I would be showing the student that I expect participation to be thoughtful and constructive.

When ideas are de-personalized, people tend to moderate their expression of opinions. Group discussions, even concerning difficult issues, usually can proceed respectfully.

President Trump does not de-personalize. . He frames political arguments not as abstract disagreements but as conflicts between himself and his opponents. He doesn’t think of you as having a position on an issue. Instead, you are either with him or against him.

Another point that I would make to students is that the object of a political discussion is not to defeat or humiliate those who disagree with you. The object should be to understand those who disagree with you, to get inside their heads and see where they are coming from. I would say that the winner of a classroom discussion is the person who can best articulate the point of view of those with whom that person disagrees.

I understand that in the real world, politics is not as gentle as that. But in the past, the President tried to be somewhat “above the fray.” Presidents usually spoke in Persuasion Mode, and they left demonization to their surrogates. Instead, President Trump regularly insults and belittles those with whom he disagrees.

College Administrators as Negative Role Models

If any institution in America ought to be a model for Persuasion Mode, it should be the university. Colleges should be a bastion of Persuasion Mode, and Demonization Mode should be treated as intellectually inferior and educationally inappropriate.

Instead, we have seen the success of ideological suppression of free speech and free inquiry. When Lawrence Summers suggested fifteen years ago that the phenomenon of male-dominated math departments might reflect the statistical distribution of ability, he may or may not have made a convincing case. But he was certainly speaking in Persuasion Mode, and he was ousted in Demonization Mode. That has become a trend on campuses throughout the country, and few college administrators have the clarity of purpose or the strength of character to resist.

Students are being shown that demonization works, and that persuasion is neither necessary nor sufficient to participate in disputes. As I see it, the college administrator is currently the most damaging negative role model in America today.

Those Who Know Better Must Do Better

The current state of political discourse is not good. We observe too much demonization and not enough persuasion. Incentives are strong for politicians and pundits to take one-sided, closed-minded stances.

It is hard to come up with a straightforward way to get from where we are now to a state in which Persuasion Mode plays a larger role in political discourse. The best I can offer is a slogan: Those who know better must do better.

Are you one of those people who takes to Twitter or Facebook to read about or comment on politics? Stop. Just stop.

Are you convinced that the other side is so evil that you cannot associate with them in any way? If you think that the problem with political discourse is caused entirely by the other side, then look in the mirror instead.

Do you avoid people who have different political viewpoints? Find an activity that you enjoy in which they would be likely to participate. Get involved in that activity. Get to like those other people. Appreciate that their human qualities transcend your political differences.

Are you involved with a college as a parent, active alumnus, professor, or administrator? Take a stand for intellectual rigor at that college.

You care about the sorry state of political discourse. Other people care as well. I suspect that there are many people who long to see less Demonization Mode and more Persuasion Mode. If the people who know better will do better, perhaps we can achieve a transformation.




[3] Curtis Sittenfeld, “Gender Studies” in You Think it, I’ll Say it: Stories.

Also from this issue

Lead Essay

  • Arnold Kling has devoted years to observing and participating in American political discourse. Such as it is. He writes that all too often, Americans seek to demonize those with whom they disagree, rather than trying first to understand what they are saying in their own terms. Social media does not help us to do better, he argues, but one thing that could is surprisingly simple: “If you think that the problem with political discourse is caused entirely by the other side,” he writes, “then look in the mirror instead.”

Response Essays

  • Jonathan Rauch offers a “friendly amendment” to Arnold Kling’s argument: We must amend our weakening civic institutions. He recommends direct social action, including but not limited to various organizations that seek to start conversation across political divides; strengthening institutions that don’t line up neatly with partisan politics, but that serve good purposes anyway; and—perhaps controversially—re-engineering our social media environment.

  • Nikki Usher turns her attention to American journalism, which often sets the tone for our political conversations. She finds that with the decline of local journalism, journalists have an increasingly cohesive and uniform view—one that sees itself as cosmopolitan, but that is really parochial to Washington, DC. Journalists nowadays often stumble in covering other areas of the country, which increases the importance of Arnold Kling’s populist/cosmopolitan divide.

  • Donald Downs considers that free speech requires a concern for the personhood and well-being of others—and the concern to avoid offense can also be stifling. He characterizes this as a tension between chutzpah and “due doubt” of one’s own beliefs, and he argues for the importance of civility understood in a special way: Civility is not mere politeness in politics, he argues; it is a key part of the role that we must all play as citizens, regardless of political disagreements.