Public discourse and its problems

Abstract

It is widely believed that open and public speech is at the heart of the democratic ideal. Public discourse is instrumentally epistemically valuable for identifying good policies, as well as necessary for resisting domination (e.g. by vocally challenging decision-makers, demanding public justifications, and using democratic speech to hold leaders accountable). But in our highly polarized and socially fragmented political environment, an increasingly pressing question is: Do actual democratic societies live up to the ideal of inclusive public speech? In this essay, I explore Maxime Lepoutre's defense of discursive democracy from the challenge of defective public discourse. I argue that political ignorance, dogmatism, and social fragmentation present more formidable challenges to discursive democracy than Lepoutre acknowledges.

Introduction

Public deliberation, debate, and dispute are cornerstones of a democracy. When we think of democracy, we often think of open forums, passionate arguments, plural perspectives, and public speech. This reflects a discursive conception of democracy. According to this view, inclusive public speech is at the heart of the democratic ideal.

There are two common justifications for why open and public speech is vital for democracy. ¹ The first is epistemic: public discourse is instrumental for identifying good policies. Roughly, we need to pool knowledge from dispersed groups to identify social problems and counteract them, and the best way to pool such information is via inclusive public speech. The second justification is moral: a key mechanism for resisting domination—the subjection to arbitrary power—is by vocally challenging decision-makers, demanding public justifications, and using democratic speech to hold leaders accountable.

In our highly polarized and socially fragmented political environment, an increasingly pressing question is: Do actual democratic societies live up to the ideal of inclusive public speech?

In Democratic Speech in Divided Times, Maxime Lepoutre scrutinizes the mismatch between the reality of political speech and the ideal of discursive democracy. He acknowledges that the discursive ideal seems “tragically out of touch” with the reality of political discourse:

we find a public sphere saturated with emotional appeals, including intensely negative emotions such as rage and resentment. Instead of mutual respect, public speakers routinely use their airtime to ridicule, demean, or vilify others. And where sincerity should reign, campaigns of misinformation instead proliferate unimpeded. (Lepoutre 2021: 2)

This mismatch between the ideal and the reality of democratic speech presents a challenge. It is often said that public discourse cannot effectively address social problems and injustices unless it is governed by norms of orderly argument, mutual respect, and sincerity. ² When citizens are poorly informed, prejudiced, and treat their political opponents with antipathy, they fail to abide by the discursive norms that are allegedly required for a well-functioning democracy. As a result, political speech is unable to serve its epistemic and anti-dominating functions. Let's call this the challenge of defective public discourse.

In his excellent, timely, and important book, Lepoutre defends a discursive conception of democracy from the challenge of defective public discourse. His book provides a nuanced investigation of what democratic public speech would have to be like in non-idealized conditions in order to effectively address social problems. There are two main parts to Lepoutre's argument. Part 1 of the book develops an account of the norms that should govern democratic speech in contexts characterized by deep social divisions. Lepoutre argues that “some features of public discourse that are symptomatic of social divisions [e.g., angry speech] can nevertheless play a positive and important democratic function” (2021: 6). In Part 2 of the book, Lepoutre examines whether the recommended norms of inclusive public speech (which he develops in the first half of the book) succumb to the challenge of defective public discourse. More specifically, he scrutinizes whether his normative account of democratic speech is undermined by widespread affective polarization, political ignorance, and social fragmentation. Lepoutre argues that his account can largely withstand the challenge of defective public discourse; thus, we should not jettison the ideal of discursive democracy, even in highly polarized societies.

There is much to agree with in this well-argued book. For example, I agree with Lepoutre that open and public speech is central to democracy and is valuable for both epistemic and moral reasons. I also agree with him that discursive democratic ideals often seem out of touch with reality, and that these ideals must not be too distant from real-life political discourse to be plausible. Moreover, I agree with Lepoutre that pessimism about the value of public discourse is sometimes overstated. As he reminds us, “History is replete with examples of inclusive public speech being used, with great effect, to tackle pressing social problems such as gender oppression and racial injustice” (2021: 3). Finally, I take no issue with the norms of democratic speech that Lepoutre articulates and defends in the first part of his book.

In this essay, I want to examine Lepoutre's defence of discursive democracy from the challenge of defective public discourse. I will argue that political ignorance, dogmatism, and social fragmentation present more formidable challenges to discursive democracy than Lepoutre acknowledges. As a result, his account occasionally veers from warranted optimism to unwarranted idealism.

Pessimism, optimism, and idealism

To better understand Lepoutre's argument, it will be helpful to identify three possible responses to the challenge of defective public discourse. The three options are: pessimism, optimism, and idealism.

The pessimist denies that public discourse is an effective tool for addressing social problems in polarized societies, such as the United States. The pessimist may acknowledge that inclusive democratic speech is in principle a good thing; for example, it may be instrumental for identifying good policies and resisting domination when it is governed by norms of orderly conduct, mutual respect, and sincerity. But, according to the pessimist, this gives us no reason to value public speech in contexts where social groups are divided by mutually antagonistic attitudes, by patterns of social and spatial fragmentation, and by unjust inequalities. When citizens are polarized, ignorant, biased, and unwilling to listen to each other, public speech cannot serve the epistemic and moral functions that allegedly justify discursive democracy—or so the pessimist claims. ³

The optimist, by contrast, maintains that public discourse is an effective tool for addressing social problems. They acknowledge that citizens are deeply divided, fragmented, misinformed, ignorant, angry, prejudiced, and unwilling to trust their political opponents. But the optimist nevertheless defends the value of democratic public discourse even in highly non-ideal conditions. Instead of ditching discursive democracy for political ideals that de-emphasize inclusive public speech (as some pessimists do), the optimist aims to develop an account of the norms that should govern democratic speech in polarized societies. Once we have a more realistic account of these norms, we will see that inclusive political speech is indeed an effective tool for addressing pressing social problems, according to the optimist

Lepoutre is an optimist, but not an idealist More accurately, he's not an out-of-touch idealist. His book does have elements of idealism; for example, he suggests that political discourse isn't where it needs to be, and he provides an outline for what would need to be in place for it to work well (e.g. the chapter on integration, which I’ll discuss later). But even those elements are supposed to outline conditions that aren't too far from reality. The danger with ideal theory, as Lepoutre observes, is that it risks becoming too distant from real life. A very distant ideal may not provide us with adequate action-guidance; moreover, it may tell us very little about the value of democratic speech in existing human societies (Lepoutre 2021: 2–3). Lepoutre finds these worries compelling, so he aims to defend discursive democracy without retreating to the realm of distant ideal theory.

With this taxonomy in place, we can now sharpen the challenge of defective public discourse. The challenge can be formulated as follows:

Discursive democracy is valuable (or justified) only if it serves important epistemic and/or moral functions.

To effectively serve these epistemic and/or moral functions, political discourse must be governed by specific discursive norms.

In deeply divided settings, the relevant norms of democratic public discourse do not (and likely will not) properly govern political speech.

Thus, political speech is unable to serve its epistemic and/or moral functions.

Thus, discursive democracy is not valuable (or justified).

Both pessimists and optimists are willing to accept premises 1 and 2. They may disagree about what specific rules should govern the content and form of public discourse, but they agree that effective democratic speech must be governed by some such norms. The disagreement is over premise 3. The pessimist accepts premise 3 and thus doubts the value of democratic public discourse (hence, they accept 4 and 5). ⁴ The optimist rejects premise 3 and seeks to defend democratic public speech as an effective tool for addressing pressing social problems, even in non-ideal conditions. In what follows, I will investigate Lepoutre's defense of optimism and suggest there are still good grounds for pessimism.

Political ignorance and public discourse

It is widely believed that a well-functioning democracy requires an informed citizenry. As James Madison once said, a popular government without an informed public “is but a prologue to a farce or a tragedy, or perhaps both” (1822). Yet one of the most consistent findings in political science over the past 60 years is the jaw-dropping extent to which citizens are ignorant about politics. ⁵ The average voter does not know even basic political information, such as who their elected officials are, what their opponents believe, and which important laws or policies were passed in recent years.

Political ignorance is a problem for many theories of democracy, but it seems especially problematic for discursive conceptions of democracy. ⁶ If open forums, passionate arguments, and public speech are to help us identify good policies and resolve social problems, then presumably citizens must know basic political facts. After all, it would be pointless for citizens to come together to advance ideas, weigh the pros and cons, and criticize each other's views, if they lack the relevant knowledge to do so competently. Citizens cannot exchange reasons when they have none. Political ignorance is thus a significant threat to democratic public discourse. Even if citizens are willing to engage with one another in good faith, fruitful public communication is likely impossible when the electorate is poorly informed, dogmatic, and epistemically irrational.

A common response to this worry is that people can use ‘information shortcuts’ to circumvent their political ignorance. ⁷ The basic idea is that we can turn to more informed acquaintances, opinion leaders, and experts for advice regarding which politicians or policies we should support. For example, we might rely on rules of thumb like, “Is this candidate liked by people I respect?” or “Does my preferred political party support this policy?” If these shortcuts are reliable, then voters can cast well-informed ballots despite having little information.

Lepoutre appeals to information shortcuts to defend the value of democratic public discourse. In particular, he argues that ‘group cognition’ is an epistemically valuable practice for uninformed people to go about forming political beliefs. His book provides a number of examples to illustrate this point, such as the influx of Boston Jews to the Democratic Party during the New Deal Era, and the increasing mobilization of black Americans within the Democratic Party from the 1950s onwards. ⁸ I will have more to say about the epistemic value of group cognition shortly. For the moment, I simply want to highlight one of Lepoutre's main claims: that group membership can be a useful heuristic to determine what to believe, at least under the right conditions.

Let's suppose this claim is correct. Does it salvage the value of democratic public discourse? I suggest not.

It is important to recognize, first, that information shortcuts do little to reduce levels of ignorance. Information shortcuts are valuable precisely because they allow citizens to make good decisions in the absence of knowledge. (If information shortcuts were useful tools to increase citizen knowledge, we wouldn't be confronted with an ocean of data on political ignorance.) Instead of requiring citizens to learn more, proxies like group membership and opinion leaders can be used to figure out how to advance one's interests and concerns without significantly increasing one's knowledge. To illustrate, I might know nothing about the pros and cons of a flat tax system, but I might come to know that I shouldn't support it because someone who knows a lot about taxation (and shares my values) tells me it's bad. I do not thereby learn anything about the relative merits of a flat tax versus progressive taxation. But I am nevertheless able to vote competently, assuming my information shortcut is reliable. In other words, I am able to circumvent the problem of political ignorance without overcoming my ignorance.

Do information shortcuts, such as shared group membership, do anything to improve democratic public discourse? If these shortcuts do little to reduce political ignorance (as I’ve argued above), then they will likely do little to improve political discourse. This is because people will remain insufficiently knowledgeable to engage in meaningful public debate. In order to competently exchange reasons with others, we must be adequately informed. But shortcuts do not increase our knowledge of the relevant issues; they simply allow us to make political choices that advance our interests in the absence of such knowledge.

In reply, Lepoutre claims that “we can readily use information shortcuts to get the information we need to then engage fruitfully with others” (2021: 157). But it is unclear how we can fruitfully engage in political discourse when shortcuts do little to increase our knowledge. Recall the flat tax example: I am able to vote competently, but the information shortcut does not improve my ability to make any meaningful contribution to public discussion about the issue. This point is acknowledged by even the most enthusiastic supporters of information shortcuts (see Lupia 2006). Lepoutre seems to assume that people who start out knowing little about politics will be able to subsequently engage in fruitful public discussions merely by relying on information shortcuts. But it is unclear what the argument for this assumption is supposed to be. As a result, it is unclear how these shortcuts would strengthen the case for open and public speech.

In fact, Lepoutre's defense of information shortcuts might backfire against his defense of discursive democracy. The successful use of information shortcuts might reveal that public discussion, deliberation, and debate do not play a significant role in enabling citizens to make good political choices. Instead, citizens can unthinkingly adopt whatever preferences are endorsed by representatives of their own social group without the need to engage in political speech. In other words, the more emphasis we place on the value of information shortcuts, the less need there is for inclusive public speech.

Perhaps this is too quick. Let's consider another way to salvage the value of democratic public speech. According to Lepoutre, “Information shortcuts often operate by means of public speech” (2021: 157–8). The public testimony or advice given by experts and opinion leaders, for instance, are paradigmatic examples of information shortcuts. Thus, Lepoutre might argue that public discourse plays an important role in circumventing political ignorance because, without public speech, citizens would lack the means to reliably advance their interests.

I acknowledge that information shortcuts often operate by means of public speech. But we should ask: Is this a solid foundation on which to rest the value of discursive democracy? It seems a far cry from the ideal of public discourse as involving the mutual giving, understanding, and criticizing of reasons. The marketplace of ideas would not be an arena in which reflective opinion on public matters can develop; it would rather be a domain in which we listen to the announcements of likeminded individuals and passively receive their opinions. Thus, it is unclear to me why the availability of information shortcuts provides much support for discursive democracy.

Perhaps the primary contribution that ordinary citizens make to public discourse is in terms of revealing and expressing their concerns about the world, highlighting the problems that affect them, and identifying which issues they think should be prioritized given their own situation and values. This type of discourse might not suggest immediate solutions to those problems, but it seems valuable nonetheless. Moreover, this kind of discourse doesn't seem obviously threatened by the familiar worries about factual political ignorance. ⁹ What should we say about this alternative view?

Lepoutre does not consider this idea, but it is worth taking seriously. If we accept this proposal, it would imply that discursive democracy does not require citizens to be very knowledgeable about political facts or data relating to policy questions. It would also suggest a distinctive view of the appropriate political discursive norms, how public discussion should take place, and what kinds of contributions should be encouraged. While I have sympathy for this idea, it raises at least three problems for Lepoutre's argument. First, it would not help Lepoutre to achieve his goal of defending the value of information shortcuts. (Instead, this view seems to deny that information shortcuts are needed for public discourse to serve its primary function.) Second, this proposal is quite distant from the ideal of public discourse as involving the mutual giving, understanding, and criticizing of reasons (a view that Lepoutre and other discursive democrats tend to advocate). If expressing one's concerns about the world were all there is to citizens’ contributions, then citizens would not need to participate in collective deliberation in any robust sense. ¹⁰ Third, we might worry that elite manipulation can affect citizens’ interpretations of their interests in quite profound ways, such that it corrupts even their ability to express their concerns about the world in a way that reliably tracks their actual interests. For example, rural inhabitants might be led to believe their concerns are getting fewer tax dollars per capita, which is false. Or followers of Trump might think COVID-19 is a hoax, which illustrates how deeply elite manipulation can influence perceptions of what problems one is facing. That said, it is still worth distinguishing between these different kinds of contributions to public discourse because it suggests that citizens may not need much factual political knowledge to make some meaningful contributions.

Group loyalty and political judgment

Let's turn to a deeper problem. When people seek out information, they are often strongly influenced by group affiliations such as their race, gender, ethnicity, religion, and political party. As a result, one's group identity strongly influences what one believes. Achen and Bartels defend this idea in their book, Democracy for Realists. They write, “for most people, most of the time, party and group loyalties are the primary drivers of vote choice” (2016: 272). Lepoutre dubs this phenomenon ‘group cognition’.

An increasingly common worry is that group cognition is an epistemically dubious basis for political judgment. ¹¹ There are two justifications for this view. First, social group memberships tend to originate in arbitrary ways, not from reasoned choices that are based on one's prior preferences. The typical voter becomes attached to a particular “team” largely due to unchosen characteristics (like race or ethnicity) or accidental historical circumstances (such as where one grew up). Second, we develop emotional attachments to social groups that impair rational thought. Abundant evidence shows that ingroup favoritism has a pervasive influence on political judgment. As a result, people are not motivated by the desire to acquire accurate information; rather, they are motivated to believe whatever supports their own team. Does this present a problem for democratic public discourse?

One of the many persuasive and novel aspects of Lepoutre's book is his attempt to clarify the conditions under which group cognition functions well. In keeping with his optimistic stance, Lepoutre argues that group cognition need not be a problem for democratic public discourse. Drawing on work by Mansbridge (1999), Young (2000), and others, he demonstrates that the testimony, advice, and positions of members of one's social group often constitute useful heuristics for determining what one should think about political issues. In other words, the fact that someone belongs to our social group can provide us with an epistemic reason to defer to them about political matters.

What is it about social groups that makes them an adequate epistemic indicator of whose political judgment to trust and of what policies or parties to support? The explanation has three steps, outlined below.

First, members of social groups typically have distinctive experiences in virtue of their group membership, and this provides them with a shared perspective that non-members often lack. In particular, the members of a particular social group will often experience different “constraints and enablements by the laws, norms, and physical infrastructure that constitute the social context” (Lepoutre 2021: 164; see also Young 2000: 95–100). For example, women are more likely to experience sexual harassment as well as the expectation to be primary caregiver to children, while black men are more likely to be harassed by the police and followed by suspicious store workers. ¹² These experiences culminate in what Young (2000) calls a ‘social perspective’.

Second, this social perspective provides a strong reason for voters to be represented by members of their social groups. These representatives will have “knowledge of particular facts about the workings of society that one gleans from one's group specific experiences of constraints and enablements” (Lepoutre 2021: 165); moreover, these constraints “are more salient to members of a particular social group, [so] they are more likely to be objects of concern to those members” (ibid). For example, women are not only likely to know more about sexual harassment in virtue of their greater personal experience of it, they are also more likely to take it seriously as a problem to be addressed (Young 2000: 140). Mutatis mutandis for other social groups with different constraints. The upshot is: it is epistemically reasonable to give preference to representatives who come from one's social group because they are more likely to identify and counteract the social problems that are in the foreground of one's own thinking.

Third, the fact that members of one's social group disproportionately belong to or support party X also provides a pro tanto epistemic reason to support party X (Lepoutre 2021: 166). Thus, we not only have an epistemic reason to support candidates that belong to our social group, we also have an epistemic reason to support political parties that are supported by members of our group. On these grounds, Lepoutre claims that group cognition is epistemically valuable.

I’ll now highlight some potential limitations of this argument.

Are voters epistemically rational?

Assume that group cognition is epistemically valuable for the reasons outlined above. This would support two claims. First, group cognition helps voters to advance their interests and concerns: when group identity influences our political judgment, we are more likely to support policies, politicians, and parties that will tackle issues we care about. Second, it is epistemically reasonable for voters to use group identity as a proxy for trustworthiness, assuming they know it is a reliable proxy.

Does Lepoutre's argument also show that voters are being epistemically rational when group identity influences their political judgment? I find this doubtful. To illustrate why, consider the following thought experiment:

Infallible Irena. Imagine that I am married to a remarkable woman named Irena. What makes her remarkable is that she is infallible and completely benevolent. It would thus be epistemically rational to defer to Irena's judgment on all things. Unfortunately, I do not know (or even believe) that she is infallible. Luckily, I do defer entirely to her judgment, but only because I am so hopelessly in love with her that I am willing to irrationally believe whatever she says.

It is obviously epistemically valuable for me to defer to Irena, as it leads me to form accurate beliefs. Moreover, it would be epistemically rational for me to defer to her judgement if I knew that she was infallible. But the fact that I am deferring to her (and deriving epistemic benefits) does not thereby make my behavior epistemically rational. In fact, it is epistemically irrational to believe everything someone says simply because one is hopelessly in love with them.

Let's now consider Lepoutre's defense of group cognition. To be epistemically rational, individual voters must not only engage in group cognition and derive epistemic benefits from this process; they must engage in it because they believe it is epistemically valuable. That is, they must give preference to representatives who come from their social group (or to parties that are supported by members of their group) because the relevant political agents are more likely to identify and counteract the social problems that matter to these voters. However, Lepoutre does not provide any evidence for this psychological claim. In the absence of such evidence, the psychological basis for group cognition is epistemically dubious.

To be clear, I am not denying that group cognition may provide an epistemic reason for voters to be represented by members of their social groups. I am simply questioning whether voters rely on group identity for this reason. In other words, I doubt that voters are making use of cognitive shortcuts in the way Lepoutre describes. Their loyalty to a group will often be due to arbitrary and irrational emotional attachments. In fact, this is precisely what the evidence on ‘minimal group paradigms’ suggests: people are willing to give preference to group members who share even the most arbitrary and temporary characteristics. ¹³ Thus, when Southern whites massively migrated to the Republican Party from the 1950s onwards, for example, their behavior may not have been epistemically rational. Even if this realignment was primarily due to a sense of affinity with (predominantly white) Southern identity, and even if group affinity is a reliable basis for political judgment, the basis for their loyalty may have had nothing to do with the epistemic value of group cognition. ¹⁴

Perhaps we shouldn't fuss over the distinction between epistemically rational behavior and epistemically valuable behavior. What really matters, one might argue, is that people are capable of making good political choices (i.e. choices that would advance their interests and concerns) in non-ideal conditions. If group cognition is epistemically valuable, then people are capable of voting reliably even when they are largely ignorant and motivated by ingroup favoritism or dogmatic loyalty. This seems like good news. Nevertheless, this fact does little to shield voters from the charge of epistemic irrationality, bias, or group dogmatism. It would be a matter of epistemic luck that group cognition functions well. Voters would be selecting information shortcuts (even reliable ones) for reasons that have nothing to do with getting at the truth.

Is group cognition epistemically disvaluable?

A more significant worry is that group cognition fosters epistemic distortions that outweigh its epistemic benefits. Consider the following examples: Should whites not have voted for Barack Obama? Should men have voted for Donald Trump rather than Hillary Clinton? (I have deliberately chosen examples that will ruffle the feathers of my liberal academic colleagues.)

In response, one might argue that group cognition sometimes misfires even though it is typically reliable. That might be true, but it is an open empirical issue that requires more investigation. Work in political anthropology suggests that group cognition is often an unreliable indicator of whose judgment to trust. In Strangers in their Own Land, Arlie Russell Hochschild (2016) describes her five-year-long ethnographic research amongst supporters of the Tea Party in Louisiana, where social group membership led individuals to adopt cultural values that went against their own self- and group-interest. She terms this the ‘great paradox’. Often the very same groups that would profit from state regulations (e.g. on pollution) or government benefits (e.g. Medicaid) are the ones opposing these measures. Likewise, Thomas Frank's What‘s the Matter with Kansas? describes the rise of political conservatism in Kansas, which he says espouses economic policies that do not benefit the majority of people in the state. According to Frank, many people continue to vote Republican, even though they are voting against their best interests, because of group cognition: politicians and pundits point their fingers to an out-group of “liberal elites,” a straw man representing everything that conservatism is not.

Lepoutre is aware of this objection (see 2021: 170–172). He recognizes that while social perspectives may facilitate distinctive knowledge of some important social constraints, these perspectives might also obscure other, likely more important, constraints. They foster blindspots, lead us to overlook urgent problems faced by other groups, and encourage us to give excessive weight to some political information at the exclusion of other information. Thus, even if social perspectives have some epistemic value, they may also foster important distortions. This threatens to undermine the epistemic value of group cognition. If the epistemic benefits are outweighed by epistemic distortions, then group membership does not provide a pro tanto epistemic reason to form judgments about political matters.

This leaves us with a challenge. We must determine whether the epistemic benefits of group cognition come at too high an epistemic cost. To address this challenge, Lepoutre distinguishes two arguments against the epistemic value of groups:

The Problem of Technical Group Cognition:

Group identity influences what we think on matters that are largely irrelevant to these social group perspectives, such as highly technical or scientific questions.

The Problem of Dogmatic Group Cognition:

Even when group-specific experiences are relevant, they may dogmatically be given too much weight at the exclusion of other politically-relevant information.

Lepoutre aims to defend the epistemic value of group cognition from each of these objections. He claims that pessimists have too quickly written-off group cognition as epistemically unreliable. Against the pessimists, Lepoutre argues that group cognition is an epistemically appropriate way to form political beliefs, even about highly technical questions and even when social group identity leads to dogmatic thinking. While I agree with much of Lepoutre's argument, I will continue to raise some concerns about whether he provides sufficient grounds for optimism rather than pessimism. I will focus on the problem of technical group cognition in the next section, and I will return to the problem of dogmatic group cognition in the section titled ‘The value of dogmatism and the costs of integration’.

Before turning to these issues, I would like to add an important wrinkle (perhaps more than a wrinkle) to the topic of democratic group cognition. In his book, Lepoutre neglects the importance of identity heterogeneity. We are all members of many social groups, not just one, and these group identities overlap and conflict in many contexts. In his discussion of group cognition, however, Lepoutre often flattens our identity to membership in one of a few social categories, such as ‘white’ or ‘black’ or ‘man’ or ‘woman’. Yet, we are also progressives or conservatives, college-educated professionals or working class, under 40 or over 40, rich or poor, gay or straight or bisexual, Christian or not, socially conservative or socially liberal, disabled or not, and so forth. To persuasively argue that our identities are useful heuristics for navigating the political world, we may need an account of how to reconcile these different group identities. Do we have some core identities that we privilege in all contexts? Should some identities matter more for political decision-making? Or do we somehow consider the issue from all our different social perspectives and group identities, and see how those weigh up? Without an account of how these overlapping identities interact, or which should win out in which context, it isn't clear how the group cognition story is supposed to work—or so one might argue. ¹⁵

I will say two things in defense of Lepoutre on this issue. First, those who have been the most influential in raising epistemic concerns about group cognition, such as Achen and Bartels (2016), Brennan (2016), and Somin (2016), are also relatively insensitive to the role of identity heterogeneity. These ‘identity-politics pessimists’ (as I’ll call them) have argued that race, gender, partisanship, and other group identities have a pernicious influence on political judgment, but they have made this argument without calling much attention to the multifaceted nature of our identities. To the extent that these pessimistic arguments are intelligible (as many people seem to find them), it follows that Lepoutre would incur no special burden to provide an account of how intersectional identities influence the epistemic outputs of group cognition. ¹⁶ So, if there is a more fundamental problem about our ability to evaluate the epistemic reliability of group cognition without factoring in identity heterogeneity, it is not just a problem for Lepoutre's argument—it is a problem for both optimists and pessimists about democratic group cognition.

Second, Lepoutre's core account of epistemic value may be able to accommodate the fact that people belong to multiple social groups (although Lepoutre does not make this point himself). His main claim is that the social position one occupies in one's capacity as a member of a particular group comes with valuable descriptive and normative information. So, from the fact that one's group identity is influencing political judgment, it doesn't necessarily follow that this judgment is uninformed or irrational. We can scale this up to multiple identities: people get distinctive descriptive and normative information based on their overlapping identities. The ultimate account of how exactly this works is complicated by facts about intersectionality (so that the experiences one has as, say, a person of color might be modified by one's gender). But the basic point remains that identities are connected to social positions in a way that can make them rich in information.

That said, Lepoutre's argument would be stronger if it included a positive account of which identity to privilege, if any. I suspect he will say all these identities provide rich (though partial) information, and individuals would consider the issue from all of our different social perspectives and group identities (or at least the most salient ones in the context), then attempt to weigh these up. ¹⁷

Is technical group cognition reliable?

Is the Earth warming due to human activity? What are the health risks of a given vaccine? Is it safe to eat genetically modified foods? These are technical, scientific questions that should be answered independently of one's group identity. Yet, an increasing amount of evidence indicates that group loyalty profoundly influences people's judgments about technical descriptive matters. For example, conservatives are less likely than liberals to believe that anthropogenic climate change is real, while liberals tend to deny the overwhelming consensus in psychological science that intelligence is moderately heritable. In a number of cases, it seems as though politics is driving our beliefs about the facts instead of the facts driving politics.

Lepoutre calls this the problem of technical group cognition. Technical group cognition is epistemically troubling because group attachments are allegedly irrelevant to assessing the truth of technical facts. ¹⁸ Why would one's race, gender, or political ideology have any bearing on the truth of purely descriptive matters? As Lepoutre remarks, “the experiences that characterize rural life in America—e.g., rising joblessness—do not give rural Americans special insight into the reality of anthropogenic climate change” (2021: 173).

We can formulate the problem of technical group cognition as an argument:

Group attachments strongly influence people's judgments about technical descriptive matters.

However, group attachments are irrelevant to assessing the truth of these technical facts.

Thus, group attachments strongly influence people's judgments about technical matters in ways that are irrelevant to truth.

If group attachments lead people to form judgments on technical matters independently of truth, then group cognition is epistemically unreliable.

Thus, group cognition is epistemically unreliable.

In response to this argument, Lepoutre maintains that group cognition is—or at least can be—an epistemically appropriate way to form beliefs, even about politically relevant technical matters. More specifically, he argues that normative concerns (e.g. joblessness is a pressing problem that needs addressing’) can be relevant to assessing the truth of technical descriptive claims (e.g. anthropogenic climate change is real’). Thus, he rejects premise 2 in the above argument.

To illustrate why group-based normative concerns can yield good epistemic grounds for accepting or rejecting testimony about technical matters, Lepoutre provides the following example. ¹⁹ Scientist Sam is deciding whether to assert ‘anthropogenic climate change is real’. Call this proposition p. When determining whether her evidence is strong enough to assert p, Sam must give significant weight to epistemic factors such as empirical adequacy, explanatory power, and predictive precision. According to Lepoutre, Sam should also consider non-epistemic factors such as “the potential environmental damage that would result if climate change is real and yet we fail to act against it, as well as the potential harmful effects of environmental regulation on certain jobs if climate change is false and we do act against it” (2021: 174). These are relevant factors because “it would be morally irresponsible for scientists not to consider this potential impact when deciding whether or not to say something” (ibid).

This is where moral and political judgments enter the equation. To determine whether Sam's evidence for p is sufficient to warrant asserting p, she must weigh the social costs of false positives (that is, asserting p when p is false) against the social costs of false negatives (that is, not asserting p when it is true). However, this weighting necessarily relies on Sam's normative judgments about the value or disvalue of certain outcomes. If she values environmental protection more strongly than job protection, Sam might judge a false positive to be preferable to a false negative. Consequently, she will adopt a lower epistemic threshold for asserting p then she would if she held the opposite values. Thus, non-epistemic considerations (such as moral and political concerns) will have an important role to play in determining the epistemic standard for assertion. ²⁰

This shows that scientists should rely on both epistemic and non-epistemic values when deciding whether to assert p. In other words, there is pragmatic encroachment on the epistemic standard for warranted scientific testimony. ²¹ This makes sense. If there is nothing to lose and everything to gain from publicly asserting p, the epistemic standard for assertion will be low. If much is at stake, the standard will be higher. ²²

As Lepoutre points out, this argument can be used to defend the epistemic value of group cognition. Here's why. Imagine a scientist who is more concerned about the potential environmental damage of climate change than the harms of joblessness. They will be more willing to publicly assert that climate change is occurring because their epistemic standard for warranted assertion is lower than someone with the opposite value priorities. ²³ Now suppose that some listener disagrees with this scientist's normative judgments. Lepoutre provides the following illustration:

Suppose, for example, that B comes from a rural community where jobs are scarce and, consequently, believes that one of the state's normative priorities should be creating jobs in economically deprived areas. Suppose, moreover, that environmental regulation would threaten key job-providing industries in rural areas. Since the stakes of environmental regulation are very high for B (given her normative concerns) it makes sense for her to adopt a relatively high epistemic standard for accepting claims asserting the reality of climate change. Suppose, finally, that scientist A is insensitive to the economic threat rural areas face—perhaps because this threat is less visible to A's social group than to B's. If so, then B has a reason to believe that, because of this normative blindspot, A may have an insufficiently high epistemic standard for asserting ‘anthropogenic climate change is real’. (2021: 175)

This illustrates that normative considerations, including those grounded in group-based perspectives, are in principle relevant to whether, epistemically speaking, one should accept claims about technical descriptive matters. If A has incorrect normative commitments, then B has a pro tanto reason to doubt A's testimony. Thus, Lepoutre successfully demonstrates that technical group cognition is not always epistemically dubious.

But this conclusion faces an important obstacle: it over-intellectualizes the process by which ordinary voters decide what to believe about technical matters. Notice that several conditions must be met for group-based resistance to technical testimony to be epistemically reliable:

the influence of normative considerations on the testifier's epistemic standard is not trivial relative to that of epistemic values;

the testifier's normative commitments differ substantially from the listeners;

the listener is not mistaken about the normative concerns of the testifier;

the listener knows the influence of those normative concerns on the testifier's epistemic standard (i.e. they make appropriate inferences about the normative blindspots that result from the testifier's moral and political commitments);

the listener decides whom to trust based on the considerations listed in (a) through (d);

the listener's group-based normative concerns are not deeply misguided. ²⁴

It is highly unlikely that ordinary citizens will typically meet these conditions. How often will listeners be correct about the content of the testifier's normative commitments? How often will listeners know the stringency of the testifier's epistemic values? In all likelihood, listeners will often be too resistant to some technical evidence and insufficiently resistant to others. As revealed by the current attitudes toward global warming in the US, individuals are often mistaken not to defer to the testimony of experts about specific technical matters. Even when listeners are epistemically required to defer, group attachments may influence their assessments of scientific testimony in epistemically irrelevant ways. Thus, we cannot reasonably expect citizens to (reliably) identify the cases in which group-based normative commitments yield valid grounds to not defer to testimony about technical matters.

The value of dogmatism and the costs of integration

So far, we have explored the idea that group perspectives are epistemically relevant to technical and non-technical political judgments. But we have also seen that group cognition can foster significant blindspots that make us dogmatic. In particular, there are two ways in which relying on group perspectives might lead to dogmatic cognition: first, group-specific experiences might be given too much weight in some evidential reasoning; second, individuals might inappropriately ignore or discount evidence that is contrary to one's ingroup perspective (or which originates from an outgroup source). Let's reconsider the example above:

If rural communities witness job losses resulting from the regulation of polluting industries, but have little exposure to the environmental consequences of those industries, they may overprioritize job creation, relative to environmental preservation. This, in turn, might lead them to embrace excessively demanding standards for accepting reports that climate change is real. The reverse might be said for urban communities. (Lepoutre 2021: 179)

Even if group cognition provides important descriptive and normative information about the constraints encountered by one's own group, it may provide inadequate information about outgroups. This evidential bias may give rise to a form of group dogmatism. If dogmatic group cognition is common (or affects important political issues), then group cognition may still undermine the epistemic value of democratic public discourse.

According to Lepoutre, this worry is misplaced. He argues that dogmatism “can in fact be epistemically fruitful when considered from a broader, systemic, perspective” (2021: 171). In particular, Lepoutre draws an analogy with scientific progress to argue that it can be epistemically beneficial for groups to stubbornly stick to their own theories. For example, a system with dogmatic agents is less likely to prematurely discard superior theories. The balance of existing evidence is not always a good indicator of which scientific theories are best; thus, science is most productive when many rival research groups (traditions, programs) dogmatically defend their core hypotheses. ²⁵ On these grounds, Lepoutre argues that “a democratic community composed of many social groups, each of which defends its starting commitments with some (non-absolute) degree of dogmatism, is liable to generate more varied information that is relevant to forming sound political judgments” (2021: 181).

This argument raises a number of questions. For example, we might wonder whether the analogy between science and politics is apt. ²⁶ Are there important differences between the ways in which expert scientists conduct their inquiries, on the one hand, and ordinary voters, on the other? If so, do these disanalogies threaten the epistemic value of political dogmatism? Also, does Lepoutre's defense of dogmatism license a worrying form of reasoning in politics? In particular, does it let voters and politicians off the evidential hook for their views, since they can reason as follows: “I believe that anthropogenic climate change is a hoax. I realize most of the evidence conflicts with this hypothesis, but I will nevertheless continue to believe it is a hoax because the balance of evidence is not always a good indicator of which hypothesis is correct.” Should we encourage behavior that is epistemically irrational at an individual level on the grounds that it might be fruitful at the broader, systemic level?

Lepoutre does not discuss these issues, but he does consider an important objection: the systemic benefits of dogmatism will occur only if the information yielded by different inquirers circulates widely between them (see Lepoutre 2021: 181; Zollman 2010: 33). In a system with low circulation of information between agents, dogmatism tends to impair overall epistemic productivity. Thus, it becomes imperative to determine whether information does (or reasonably could) circulate widely across the political community even in highly divided democratic societies. If the information does not circulate widely, then the epistemic case for group cognition collapses into idealism.

Are there meaningful networks of communication between dogmatic agents? It is here that the mismatch between the ideal and the reality of democratic speech is especially stark. As Lepoutre observes, “The public sphere of contemporary democracies is deeply fragmented, such that people predominantly talk to members of their own social group” (2021: 187). There are at least two related reasons for this lack of intergroup communication:

First, different groups often live segregated lives: they inhabit different regions or neighborhoods, occupy different jobs, belong to different social media enclaves, and consult different news outlets. Because of this fragmentation, different groups tend to have little contact with one another. Second, […] contemporary democracies involve substantial intergroup dislike and distrust (or ‘affective polarization’). Therefore, people from different groups may refuse to speak to each other even when their paths do meet. (Lepoutre 2021: 183–4)

This fragmentation of the public sphere threatens to undermine Lepoutre's argument. The strategies he outlines for resolving the problem of mutual dislike (in chapter 5 of his book) and the problem of ignorance (in chapter 6) both require a wide circulation of information in the system of public discourse. Yet, social fragmentation threatens this circulation. Lepoutre calls this the problem of fragmentation. ²⁷

To address this problem, Lepoutre recommends ‘integrative’ policies aimed at desegregating opposing social and political factions. This is another novel and rewarding aspect of his book. While many discursive democrats have lamented the threat posed by intergroup fragmentation, few have carefully examined the policies and costs of reversing this fragmentation. Lepoutre considers three broad strategies for reestablishing interaction between social groups across the various arenas where political talk occurs:

We should create socially mixed neighbourhoods to achieve spatial integration;

We should integrate other social arenas, such as the workplace, schools, and universities;

We should integrate online spaces.

To achieve such integration, Lepoutre recommends a battery of policies, including:

eliminating class-segregative zoning regulations; distributing housing vouchers that can only be used in certain districts; adopting aggressive affirmative action programmes in universities and the workplace; redefining the boundaries of school districts; abolishing group-based—particularly race-based—academic ‘tracking’ within schools; selecting for juries that have an integrated composition; and restructuring online spaces to ensure that different groups are exposed to one another. (Lepoutre 2021: 193; see also Anderson 2010)

Although integrative policies are deeply controversial, we have a strong reason to consider them because they are indispensable to fostering inclusive public discourse. In the remainder of this section, I will briefly summarize the main reasons why integrative policies have been controversial, then I will outline Lepoutre's replies.

One reason why integrative policies have been controversial is that they curtail people's associative freedom. The general idea is that people should be free to choose whom they do and do not associate with, so we shouldn't pressure members of different groups into sustained contact with each other. Another worry is that such policies are particularly unacceptable in the case of vulnerable minority groups. There are two reasons for this (see Lepoutre 2021: 198–9). First, there are significant costs to residing in neighbourhoods predominantly populated by members of the dominant majority, such as exposure to hostility and discrimination. Second, there are significant benefits to residing in a neighbourhood predominantly populated by members of one's own group, such as sharing heritage, culture, and life experiences.

In reply, Lepoutre argues that the value of associative freedom is not absolute: it can be outweighed by other factors, such as the importance of inclusive political communication. He also emphasizes that integration does not typically affect our intimate or close associations (the most valuable form of associative freedom), but rather only influences our ‘weak ties’, such as our colleagues, schoolmates, and neighbors. Further, Lepoutre challenges the alleged inconsistency between integration and associative freedom by arguing that integrative policies are able to advance freedom of association. ²⁸ So, even if such policies do curtail freedom of association in some respects, it does not follow that they are bad, overall, for associative freedom. In the case of vulnerable minorities, Lepoutre advocates thoroughgoing integration, not token forms of integration, which requires a ‘critical mass’ (20 to 50 percent) of group members to live in the neighborhood. This would allow members of vulnerable minorities to live alongside many who share their social perspective or cultural practices, whilst also providing some measure of refuge from hostility and discrimination.

We should pause, briefly, to appreciate the controversial nature of this proposal. In the first half of his book, Lepoutre develops an account of the norms that should govern democratic speech in contexts characterized by deep social divisions. In the second half of his book, he defends this normative account from the challenge of defective public discourse. Lepoutre aims to show that neither a lack of mutual trust nor political ignorance are adequate grounds to reject discursive democracy. However, his reply to these objections relies essentially on an integrated system of public discourse. Yet, there is abundant evidence that contemporary democracies are increasingly characterized by social and spatial fragmentation. Thus, the desirability of the norms he recommends presupposes background conditions that do not exist in actual divided democracies. This opens him up to the charge of idealism. To avoid this charge, Lepoutre recommends integrative policies to make public discourse less fragmented. As Lepoutre writes, “we need integrative policies to salvage inclusive public discourse” (2021: 189). Therefore, his defense of discursive democracy ultimately hinges on his argument for desegregating the democratic public sphere.

I suspect that many will find these anti-segregation strategies too controversial. Lepoutre acknowledges that “the associative costs to vulnerable minorities involved in residential integration remain serious” and “It is impossible to say in the abstract whether this cost is worth undertaking” (2021: 201). If the argument for integration is not successful, however, then the challenge of defective public discourse would remain unresolved. As a result, Lepoutre's defense of discursive democracy would fail. One of the main lessons we should draw from his argument, then, is that theorists should devote more attention to carefully examining integrative policies that focus on their value for inclusive public discourse, as well as evaluating their normative costs. Lepoutre's book is an important step in that direction.