September 27, 2017

From Political Ignorance to Political Polarization

In recent weeks I’ve argued in this space that the current political order—here and in most of the rest of the world—has the following features:

(1) It is egalitarian: The interests of each member of the polity are supposed to count equally.

(2) It is nationalistic: Its egalitarianism stops at the borders of each nation-state.

(3) It is democratic: Within those borders, not only are each person’s interests supposed to count equally, but each person is supposed to have an equal share of political power.

Each of these three components of the status quo is unremarkable on its own. Combined, however, they may shed new light on our present predicament.

Combining features (1) and (2), I’ve argued, produces “sociotropic nationalism.” (Sociotropic means “dedicated to the common good.”) The sociotropic nationalist aims to elect officials who will use public policy to further the common good. More specifically, the sociotropic nationalist tries to elect officials who will use public policy to solve the social and economic problems besetting the people of the country.

Combining sociotropic nationalism with feature (3) produces the sociotropic nationalist citizen. The sociotropic nationalist citizen participates in politics, mainly (but not exclusively) by voting, so as to advance the interests of his or her conationals. But that raises a serious epistemological problem: How can citizens know which policies, or parties, or politicians will advance their conationals’ interests? The role played by the sociotropic-nationalist citizen brings with it tremendous epistemic demands, and these may lead citizens to make grave mistakes.

The Epistemic Challenges of Modern Politics

Consider the types of knowledge a sociotropic-nationalist citizen would need if he or she were to avoid mistakes. (This is not intended as a complete list.)

  1. Knowledge of which purported social and economic problems are real and actionable at a given time; and knowledge of which real, actionable problems are likely to crop up during the tenure of a potential office holder.
  2. Knowledge of which real, actionable problems are or will be significant enough (in scope and severity) to have a claim on scarce public resources.
  1. Knowledge of the sources of the significant, real, actionable problems. Arguably, at least, this requires the citizen to know what causes what in a modern society. The wider the relationship of societal causes and social and economic problems, the wider the citizen’s causal knowledge will have to be.
  1. Knowledge of which policies will address the causes of real, actionable, significant social problems without imposing costs—anticipated and unanticipated—that outweigh the policies’ benefits. This type of knowledge is particularly important. It’s not enough to have reliable causal diagnoses of real, significant, actionable social problems, or even reliably proven treatments; one also needs to know that the treatments will not cause worse side effects than the diseases. Thus, the citizen needs to know about unintended costs, not just costs that are budgeted or forecast.
  1. Knowledge of which politicians or parties have the ability to push through net-beneficial policies to address the causes of significant, actionable social and economic problems.


Not all of these types of knowledge are equally difficult to acquire. The fifth type is at least relatively easy to get, if we bracket the first four. But acquiring the first two is harder than it may seem (consider controversies about which unemployment rate best measures that problem). The third type of knowledge is clearly quite difficult to arrive at, and knowledge of the fourth type may be unattainable.

This explains the fact that on any given policy issue, full-time, life-long specialists frequently find themselves divided. Is a claimed social problem real? Is it significant? Is it actionable? What is its cause? Will the benefits of action outweigh the costs? Will this or that party or politician take the needed action? Typically, each of these topics is the subject of heated debate among technocratic experts. 

When experts disagree with each other, it follows that the experts on one or the other side (or on both sides) must lack the necessary knowledge. (Even if the experts agree, of course, they may still lack the necessary knowledge; an expert consensus can be wrong.) If even experts can lack the necessary knowledge, should we really expect ordinary citizens with limited time to acquire it?

Logically speaking, we shouldn’t. Yet we do.

We’re taught to vote, and we’re taught that each vote counts. It follows that we should exercise our right to vote responsibly, by making ourselves “well informed.” Obviously, though, we aren’t expected to master the five types of knowledge listed above. That would be unreasonable. So in practice, being “well informed” means, roughly speaking, that one has a passing familiarity with political controversies that are prominent in the news media at a given time. If you watch a half hour of news a day, you are, by most accounts, “moderately” informed. If you read a good newspaper, you are “well” informed. Yet the best newspaper in the world doesn’t convey a fraction of the knowledge listed above.

Epistemological Utopianism

Democratic, sociotropic nationalism asks the impossible of us. It asks us, in effect, to stand above the din of technocratic disagreement and discern which experts are right. This would require that we, the people, adjudicate the substance of policy and political debate, which in turn would require that we, the people, become omnicompetent (omniscient?) super-technocrats ourselves. 

“Democratic, sociotropic nationalism” can therefore be abbreviated as citizen technocracy: a polity that puts each of us in the untenable position of being a citizen-technocrat. A citizen-technocrat must be wiser, somehow, than technocrats of the usual sort—the experts whom political theorists call “epistocrats.”

If we want to question the legitimacy of citizen technocracy, we might begin with the gap between the epistemic demands placed on citizen-technocrats and their limited capacity to meet these demands. [1] But if we want to understand how citizen technocracy works in practice, we need to begin with a slightly different point: the fact that, as a rule, citizen-technocrats aren’t aware of this epistemic gap. To the extent that we, the people, hold opinions about technocratic policies (policies designed to address our conationals’ social and economic problems), and to the extent that we vote on the basis of those opinions, we cannot be fully aware of the unreliability of our opinions and our votes. If we gained this awareness, we would lose the confidence needed to hold the opinions and cast the votes. We would start answering survey questions “Don’t Know” and would withdraw from political participation. As Walter Lippmann wrote in Public Opinion,

There are two kinds of uninstructed voter. There is the man who does not know and knows that he does not know. He is generally an enlightened person. He is the man who waives his right to vote. But there is also the man who is uninstructed and does not know that he is, or care. [2]

To function as citizen technocrats, we have to be like the second man. To carry out our political responsibilities (no matter how well or badly we carry them out), we must fail to recognize the extent of the knowledge we’d need if we were to carry them out well. That is, we must be “radically ignorant” of the knowledge we need—unaware not only that we lack it, but that we need it. [3]

The inadequacy of our knowledge must be an unknown unknown to us. Otherwise we wouldn’t be able to form political opinions or vote, let alone engage in more intensive forms of political action.

The Other Kind of Technocracy

Lippmann published Public Opinion in 1922, when the Progressive Era had peaked. He had been an ardent Progressive himself: a believer in citizen-technocracy. A socialist in college, Lippmann was hailed by his classmate John Reed (later the semi-official chronicler of the Russian Revolution) as the “all-unchallenged Chief” of the Harvard radicals. Later he helped to found The New Republic. But World War I gave many Progressives pause about the wisdom of the people. By the time Public Opinion appeared, Lippmann had come to think that we should substitute, for a technocracy run by citizen-technocrats, a technocracy in its ordinary usage: one run by epistocrats. 

Yet in order to go along with epistocracy, we’d have to believe that experts, unlike the rest of us, have reliable access to the five types of knowledge. What could justify this conclusion?

If “experts” are defined as those who are “knowledgeable” in an absolute sense, then by definition, epistocrats will have whatever knowledge they need. [4] But this definition begs the real-world question of how epistocrats could acquire the needed knowledge; and the question of why, if acquiring this knowledge is feasible, they disagree with one other about what constitutes true knowledge about a given topic of controversy. [5]

Consider the fourth type of knowledge—knowledge of the intended and unintended effects that policies will have. We can’t just assume that some identifiable class of people will reliably know what the intended and unintended effects of policies will be, merely by virtue of the fact that we label them “experts.” Such an assumption may be warranted, although I’m unaware of any philosophy of social science that would warrant it. There is as yet no good theory of unintended consequences. (I try to provide one in a forthcoming book, No Exit: The Problem with Technocracy.) But we shouldn’t complacently equate relative knowledgeability and sufficient knowledgeability. Experts know more than non-experts. But that doesn’t entail that they know what they need to know if they are to make competent technocratic decisions. 

The Self-Evidence Assumption

The need for the five types of knowledge suggests that social and economic problem solving is complex. Lippmann’s growing awareness of this complexity drove him away from the Progressive populism of his youth: the desire to place democratic “weapons in the hands of the people,” the better to force the government to solve the people’s problems. By 1922, he was advising withdrawal from what we would now call citizen activism, which he saw as epistemologically naïve.

The activist, however—even the “activist” whose sole political action is to vote—has to assume, in practice, that the political world is simple. In fact, the activist tends to assume that the relevant truths about the political world are self-evident. This assumption enables the activist to act in the face of what would otherwise be uncertainty. A citizen-technocrat who truly understands the difficulties in obtaining the five types of knowledge will be inclined to select him- or herself out of the electorate. The citizen-technocrats who remain in the electorate will be those who are insensible to these difficulties.

Take Donald Trump. As a candidate, he always treated his opinions as indisputable—even though they were based on such slender evidence that, once in office, he had to abandon them more quickly than had any president before him. After having repeatedly called Obamacare a “disaster” that he’d easily replace with something “beautiful,” he suddenly declared: “Nobody knew health care could be so complicated.” Likewise, “after listening [to Chinese President Xi Jinping] for 10 minutes,” he told The Wall Street Journal, “I realized that it’s not so easy. [Before,] I felt pretty strongly that they had a tremendous power [over] North Korea. . . . But it’s not what you would think.”

Political scientists have long worried about the political ignorance of most voters. In Donald Trump, their worst nightmare has come true. The personification of political ignorance has been parachuted into the Oval Office.

From Oversimplification to Hatred

Citizen-technocracy would be impossible without the widespread assumption that the truth about public policy issues is self-evident, or close to it—at least to the reader of a good newspaper. Only this assumption makes it possible for citizen-technocrats to so readily believe that they have reliable (if not perfect) knowledge of whom to vote for on sociotropic grounds.

But while citizen-technocrats fail to recognize the knowledge problems they face, they cannot avoid recognizing the political problem they face: their political opponents.

Like epistocrats, citizen-technocrats disagree with each other. They disagree about which social and economic problems are real, which ones are actionable, which ones are significant, what their causes are, what the effects of proposed policies will be, and who can be trusted to implement net-beneficial policies. Yet if, as each of them is inclined to think, their own views about these matters are self-evidently justified, then how is it possible for anyone to disagree with them? As Diana Mutz once put it, puzzled citizens say to themselves, “The answers are obvious and we all agree on them. So what is wrong with all of those other people?”—i.e., our political opponents.

There has to be something wrong with them insofar as they disagree with the self-evident truth. Maybe they’ve got psychological problems (they’re authoritarians, xenophobes, victims of “motivated reasoning,” snowflakes). Maybe they’re self-interested (not sociotropic). Maybe they’ve been bought off by special interests. [6] Maybe they’re liars, merely pretending to disagree even though they know they’re actually wrong. Maybe they’re evil.

Of course, such dark possibilities may truly explain the existence of a political opponent in a given case. But citizen-technocrats are inclined to leap to the conclusion that such things must be true of their political opponents—because citizen-technocrats have trouble acknowledging what Rawls called the “burdens of judgment”: the “hazards involved in the correct (and conscientious) exercise of our powers of reason and judgment in the ordinary course of political life.” [7] Citizen-technocracy aggravates these burdens, as it requires us to be knowledgeable about abstruse and highly contestable matters of empirical fact in the ordinary course of political life.

In Lippmann’s words, one’s political opponent

presents himself as the man who says, evil be thou my good. He is an annoyance who does not fit into the scheme of things. Nevertheless he interferes. And since that scheme is based in our minds on incontrovertible fact fortified by irresistible logic, some place has to be found for him in the scheme. . . . Thus . . . out of the opposition we make villains and conspiracies. [8]

The Paradox of Citizen-Technocracy

As a citizen-technocrat, I can participate in politics, whether by voting or through more persistent activism, only if I am first convinced that I know the truth about the social and economic problems facing millions of anonymous fellow citizens (or if I think I can learn the truth through political participation). But if citizen-technocrats took the full measure of the knowledge they need, either (a) they’d recognize that as a practical matter they are unlikely to have this knowledge, leading them to select themselves out of the electorate; or (b) their political participation would consist of handing power over to experts (or people who strike them as being experts). In the latter case, citizen-technocracy would turn itself into epistocracy. In the former case, citizen-technocracy would perpetuate itself, but only by weeding out the most sophisticated citizen-technocrats, leaving the most simplistic and thoughtless to make decisions. Among them, disagreement would tend to congeal into mutual hostility.

Political disagreement can be civil only if we understand our political opponents’ disagreement with us as a result not of villainy, but fallibility. This would require us to explain how our opponents might have innocently reached mistaken conclusions. The explanation would have to turn either on the inadequacy of their knowledge or the inadequacy of their reasoning. But either of these explanations could rebound against ourselves in a way that attributions of villainy cannot.

Each of us knows that we aren’t villains, or crazy, or liars, or evil. Uncharitable accusations work only when applied to the other side. But unlike uncharitable interpretations of the other side’s disagreement with our side, charitable interpretations of our opponents can apply to ourselves. Each of us knows that we can make logical errors and that we aren’t omniscient. Nobody’s perfect. So if logical error or inadequate information may explain the views of our political opponents, logical error or inadequate information may also explain our own views. It’s hard to hate one’s political opponents if they, like we, are merely victims of their humanity. But it’s also hard to continue viewing them as opponents, because a recognition that we suffer from the same human limits as our opponents would undermine our confidence in our own political judgments.

Thus, while we can maintain our technocratic spirit by attributing perfidy to our opponents, we sap our technocratic spirit if we attribute ignorance or error to them. Since we, too, might be the victims of ignorance or error, we must calmly accept that either our opponents’ technocratic opinions or our own (or both) may be tainted by ignorance or error. The end point of this way of thinking is scholarly distance from political controversy and a reluctance to take political positions: political open-mindedness. But that’s the opposite of what citizen technocracy requires: a perfervid conviction that the political truth is easy to identify, and a self-righteous hostility to those who make a different identification. The corrective to technocratic polarization, then, undercuts citizen-technocracy itself.

Political polarization waxes and wanes, and citizen-technocracy is not its only source. Moral disagreements lend themselves to the illusion of self-evidence even more readily than do technocratic disagreements about facts. Moral disagreements, however, are less frequent in a citizen technocracy than we are inclined to think. Our demonization of our technocratic opponents makes everything in a citizen technocracy seem like a moral issue, though. Thus, citizen technocracy contributes to a simmering sense of Manichean conflict that contribute to a legitimacy crisis—my topic next week.


Jeffrey Friedman, the Director of the Niskanen Center’s Institute for the Study of Politics and a Visiting Scholar in the Department of Political Science, University of California, Berkeley, is the editor of Critical Review and the author of No Exit: A Critique of Technocracy (Oxford University Press, forthcoming 2018).



  1. That is my goal in Part I of No Exit: A Critique of Technocracy (forthcoming).
  1. Walter Lippmann, Public Opinion. New Brunswick, N.J.: Transaction, 1997.
  1. I use “radical ignorance” (a phrase coined by Sanford Ikeda) as a contrast against “rational ignorance” and “rational irrationality”—the theories of voter incompetence developed, respectively, by Ilya Somin (building on a long tradition in economics and political science) and Bryan Caplan (building on economic theory alone). One problem with Somin’s and Caplan’s theories is logical: They assume that voter ignorance is a deliberate choice produced by voters’ knowledge that their votes are (individually) inconsequential. But if I deliberately choose to underinform myself because I recognize that my vote doesn’t really matter, why would I go ahead and vote? My decision to underinform myself would amount to a decision to deprive my vote of a basis in fact, leaving me without a motive for voting. Similarly but separately, why would voters bother to vote if they recognized that their votes are inconsequential? Even if one accepts Caplan’s empirically unjustified (and logically incoherent) suggestion that people take pleasure in believing opinions that they know are false (in reality, one cannot believe to be true what one knows is false), why would people take the trouble to vote on the basis of these opinions? A third problem is that there’s no evidence that people do, in fact, know that they are inadequately informed, as opposed to knowing that they’re poorly informed. (One can acknowledge that, in the abstract, one is poorly informed, in that one should really learn more about politics and government. Yet if one holds a political opinion or takes a political action, one must perforce believe that one knows enough to do so.) We can avoid all these problems, thereby making sense of underinformed voting, by noticing the simplistic understanding of the political world that most people have. If, as radical-ignorance theory suggests, peope assume that politics and policy are simple, then people won’t think they need much information in order to form an opinion or justify a vote. Similarly, the radical-ignorance theory avoids the so-called paradox of voting (which is a paradox only for rational-choice theorists). If you don’t know that the odds are against your vote making a difference—perhaps because, unlike economists and political scientists, you’ve never given the matter any thought—you won’t find the act of voting to be problematic. (I’ve published a full-fledged critique of Caplan’s theory here, and one of Somin’s theory here.)
  1. Thus, Jason Brennan has recently argued for an epistocracy consisting of people who score relatively high on tests of “political knowledge.” But the knowledge that voters need in a citizen-technocracy is contestable. The “right answers” can’t be looked up and then put into a voter-qualification quiz. The notion that the necessary knowledge is testable reflects a simplistic view of the problems facing the citizens of modern polities, producing a simplistic view of what political “expertise” consists in.
  1. Of course, natural scientists manage to cut through these interpretive difficulties. But (at least arguably) social scientists don’t, because people are more unpredictable than molecules. People are influenced by ideas, and the impact of ideas on human behavior cannot—as a practical matter—be predicted. (See chapter 3 of my forthcoming No Exit: A Critique of Technocracy.) This isn’t to say that individual social scientists or policy experts cannot have reached adequate conclusions about a given issue. But it does mean that we can’t tell which social scientist or policy expert is right about a given issue; nor can we expect that a consensus of social scientists or policy experts will tend to be right (as we can expect of a consensus of natural scientists).
  1. Notice that public choice theory is simply the “they’re selfish” explanation of one’s political opponents packaged into a grand theory of politics. As such, it stymies the investigation of what political actors actually believe (as opposed to the motives the public choice theorist imputes to them a priori).
  1. John Rawls, Political Liberalism, Columbia University Press, 1996, pp. 55–56.
  1. Lippmann 1997, 83.