In The American Political Science Review, Henry Farrell, Hugo Mercier, and Melissa Schwartzberg (2022) challenge two influential views.

One view paints a “despairing picture” of democratic reasoning. It assembles evidence that individuals demonstrate “ignorance and incompetence” about political matters, while groups “invariably” suffer from “conformity,” “affective polarization,” “the rejection of countervailing arguments from nongroup members, and backfire effects.”

The other view holds that deliberating groups are wiser than individuals because they can pool intelligence and combine perspectives.

Farrell, Mercier & Schwartzberg argue that both theories generalize too much. Some democratic processes work out well; some do not. They cite recent “interactionist” research in psychology. “Instead of looking to the (supposedly invariant) cognitive limitations of ordinary citizens as skeptics do, an interactionist approach suggests that we should investigate the social context of decisions—how groups are structured—to understand when group identity and social pressure can distort or swamp problem solving [or not].”

Farrell and his colleagues use Elinor Ostrom as a model or inspiration for this agenda. Ostrom did not emphasize the main question of deliberative democratic theory, which is something like this: How should we debate and reach conclusions about contested matters, such as public policies? Instead, she asked how people should coordinate their individual behavior to achieve outcomes that they all endorse. Whereas democratic processes involve reasoning and discussion, many of the examples that interested Ostrom were about quiet work, e.g., digging irrigation canals or editing Wikipedia articles. Still, she confronted a similar situation to the one that Farrell et al. describe in democratic theory today.

When Ostrom got started, a dominant view held that individuals cannot coordinate effectively without external compulsion: the “tragedy of the commons” problem, articulated most famously by Garrett Hardin. Mancur Olson (1971) argued that sometimes voluntary coordination succeeds, and the key variable is the size of the group. Small voluntary groups can function; large ones cannot. Ostrom absorbed that claim as part of a much more ambitious research agenda. She strove to identify aspects of groups that vary and then explored which variables affect the quality of outcomes. The size of groups turned out to be relatively insignificant, not even appearing on her list of “design principles.”

If I may say so, I adopt a very similar position to Farrell and colleagues in my new book, What Should We Do? A Theory of Civic Life (2022). I summarize Ostrom’s agenda in chapter 3 and then turn to deliberative processes. Like Farrell and colleagues, I argue that we should approach deliberative democracy much as Ostrom addressed coordination. We should experiment and test which specific conditions make discussions go well. I argue that Ostrom’s school of political economy should pay more attention to deliberation about values, because that is a necessary activity of groups. (In many of Ostrom’s cases, there is no dispute about values.) I also discuss nonviolent social movements, but that is a different topic.

When the argument is stated as in Farrell, Mercier & Schwartzberg, I anticipate two difficult but fruitful questions.

First, how should we identify successes and failures so that we can decide which democratic processes work? After all, people often disagree about what constitutes a success. One option is to use obvious cases of failure, when everyone would regard the outcome as suboptimal. We can ask why that happened. But this approach shifts the research away from debates about principles and goals toward purely instrumental problems, like how to preserve common resources, which Ostrom and her many colleagues already studied. The more controversial the topic becomes, the less we already know about how to structure conversations about it–although I would cite literature that Farrell et al. don’t mention, such as work by John Gastil, Archon Fung, Tina Nabatchi and Matt Leighninger, Celina Su, and others. These authors have identified variables that affect conversational quality, such as the method of recruiting or admitting participants, the nature of the facilitation, and the stakes of the discussion. Still, the problem of identifying success remains. (They do mention essential work by Michael Neblo et al.)

Second, what should be done with the findings? This research has implications for the design of governments and other big institutions. The US Constitution already requires jury trials; the Constitution of India requires every village to have an empowered annual meeting. Knowing more about the conditions that make discussions go well can help us to understand whether such provisions are optimal and what other rights and institutions we should add.

But I think the main audience will be civil society. Voluntary groups–including those that seek to influence governments–are best positioned to experiment with formats for discussion. They have more flexibility than governments, they can change more easily, and their sheer number and variety creates more opportunities.

Sources: Farrell, H., Mercier, H., & Schwartzberg, M. (2022). Analytical Democratic Theory: A Microfoundational Approach. American Political Science Review, 1-6; Ostrom, Elinor (1990). Governing the commons: The evolution of institutions for collective action. Cambridge University Press, Olson, Mancur (1991), The Logic of Collective Action, Harvard University Press. See also: don’t let the behavioral revolution make you fatalistic; how political talk relates to its context; this is what deliberative democracy looks like; why study real-life deliberation?; the legacy of Elinor Ostrom and the Bloomington School; etc.

— Please don’t talk that way in school.

— It’s a free country; I can say what I want.

Both of these speakers describe the context in which they’re speaking in order to support their goals or values. Even if they’re in the same place, both could be making valid points, because we can operate within several contexts at once. For instance, a classroom can be located within the United States.

These speakers are not completely free to describe their contexts as they wish. Unless the first speaker is actually located inside a school in which certain norms are commonly observed, that statement is odd–perhaps a joke or an idiosyncratic remark rather than an effective intervention. The first statement assumes a real, bricks-and-mortar building that has prevalent norms.

However, these statements are not completely determined by their objective context. They reflect choices: speakers can select which contexts to highlight and can identify preferred features of the contexts.

If many speakers make the same choices, they can influence the context. For instance, if teachers consistently say, “You can’t curse here,” the school may become a place where public cursing is rare. Teachers could decide to begin or to stop describing the school’s norms in that way. They are more influential than their students; as in most cases, power in unequally distributed. However, we only get the speech-context we want to the extent that the norms we advocate are actually observed. If teachers say, “We don’t talk that way here,” but everyone does anyway, they will begin to look foolish. In that sense, everyone influences the context, albeit to unequal degrees.

We can sometimes even use speech to create the context for speech, as in performative utterances like these:

— I call the meeting to order.

— Let us bow our heads in prayer.

(The second statement might change a secular gathering into a spiritual one for a time.)

I’ve recently learned that John J. Gumperz (1922-2013), a founder of interactional sociolinguistics, pioneered the idea that language has a dynamic, two-way interaction with social contexts. I look forward to learning more, especially about the political implications.

After all politics requires good conversation. The definition of good political talk is itself a matter of debate. Who must be included in each discussion? Must the discourse be civil? Must it be public-spirited? Must it aim at consensus? Must it be secular? What counts as appropriate evidence for empirical claims? Which emotions are valuable and when?

Contexts influence what forms of speech actually occur and prove effective. Political speech uttered in a church during a faith-based social movement will inevitably be different from political speech uttered in a faculty meeting, a union hall, or a courtroom. I am skeptical that we need just one type of speech. Pluralism is good.

Speech contexts are shaped by:

1. The implicit norms reflected in typical speech within each context. For example, if it is common to criticize other participants by name, then that is the norm.
2. Explicit characterizations of the context. “You really shouldn’t keep citing scripture here–most of us are not Christian” would be such a move. It describes the local norm as secular, and if people accept this description, it may affect their speech.
3. Other aspects of the institution: Who is permitted and/or recruited to participate? What behavior is rewarded? Who makes key decisions? Even literal architecture may matter. For instance, a bricks-and-mortar school probably consists of many rooms that are designed to hold one adult with 15-30 children or youth. Discourse would be different in a stadium, a prison, or along a forest trail.

We should envision speakers as operating in contexts that they may or may not endorse. At one level, they make ordinary points about what they believe or advocate. How they talk either conforms to the norms of the speech-context or violates them to some degree. Widespread violation can change the norms.

At another level, individuals may seek to change the speech-context, either by moving to another context (exit) or by seeking to alter its norms (voice). They can use their voice to advocate directly for different speech-norms, as in statements like, “Everyone is being too politically correct here–we must tolerate uncomfortable opinions.” Or they may use their voice to support changes in the institution that would likely change the norms. For instance, changing the demographic composition of a school or the balance of power between teachers and students might change the frequency of various forms of discourse in the school.

Discourse ethics is then not exhausted by the question: What kind of arguments should individuals make about policies and issues? It also encompasses questions about how to design, create, choose, and influence the contexts of speech, both directly and indirectly.

This is a mild critique of the idea that one kind of speech is desirable in a liberal democracy and that institutions should enact rights, rules, and procedures that encourage such speech. Instead, I am suggesting that people are embedded in diverse speech-contexts, which they also influence; such pluralism is desirable as well as inevitable; and people need ethical forms of voice and exit that they can use to affect their various speech-contexts.

See also: what sustains free speech?; a civic approach to free speech; this is what deliberative democracy looks like; modus vivendi theory; and judgment in a world of power and institutions: outline of a view.