Category Archives: deliberation

important findings about the persuasive power of facts

Leave a reply

There is a huge body of research that suggests that people are not very susceptible to good arguments. Apparently, we believe things for unexamined reasons, cherry-pick evidence to support our intuitive beliefs, and minimize the significance of inconvenient evidence.

These findings contribute to a general skepticism about people’s capacity for democracy, and I fear that this skepticism is self-reinforcing. If we presume that humans cannot reason well, why would we try to build institutions that promote reasoning? Only half jokingly, I sometimes say that the theme of current social science is: people are stupid and they hate each other.

But I also argue that at least some of this research employs methods that are biased against discovering rational thought. In particular, if you ask random samples of people disconnected survey questions that interest you (not them) and then use techniques such as factor analysis to find latent patterns, you will, indeed, often discover that people are stupid and hate each other. More prosaically, you will develop scales for latent variables like knowledge or tolerance that yield poor scores. But such methods may overlook the idiosyncratic ways that reasons influence individuals on the topics that matter to them.

Of all people, those who believe in false conspiracy theories are generally seen as the least susceptible to good reasons; and previous efforts to convince them have often failed. However, in a 2024 Science article, Thomas H. Costello, Gordon Pennycook, David G. Rand report results of an intervention that substantially reduced people’s commitment to conspiracy theories, not only in the short term, but also two months later.

In this study, holders of conspiracy theories wrote about why they held their beliefs, and then an AI bot held a conversation with them in which it supplied reliable information directly relevant to the specific factual premises of each respondent. For instance, if a person believed that 9/11 was an “inside job” because Building 7 collapsed even though no plane hit it (see Wood and Douglas 2013), the AI might provide engineering information about Building 7. Many people were persuaded.

These results are consistent with a study of conversations with canvassers who succeeded in persuading many voters “by listening for individual voters’ … moral values and then tailoring their appeals to those moral values” (Kalla, Levine, A. S., & Broockman 2022). The two studies differ in that one used people and the other, an AI bot; and one emphasized facts while the other focused on values. But both results point to a model in which each person holds various beliefs that are more-or-less connected to other beliefs as reasons, forming a network. Beliefs may be normative or empirical–they function very similarly. Discourse involves stating one’s beliefs and their connections to other beliefs that serve as premises or implications.

People actually do a lot of this and are relatively good at assessing the rigor of such conversations when they observe them (Mercier and Sperber 2017). However, many of our methods are biased against discovering such reasoning (Levine 2024a and Levine 2024b), leaving us with the mistaken impression that we are a bunch of idiots incapable of self-governance.

Sources: Costello, T. H., Pennycook, G., & Rand, D. G. (2024). Durably reducing conspiracy beliefs through dialogues with AI. Science, 385(6714); Wood MJ, Douglas KM. “What about building 7?” A social psychological study of online discussion of 9/11 conspiracy theories. Front Psychol. 2013 Jul 8;4:409; Kalla, J. L., Levine, A. S., & Broockman, D. E. (2022). Personalizing moral reframing in interpersonal conversation: A field experiment. The Journal of Politics, 84(2), 1239-1243; Mercier, H. & Sperber D, The Enigma of Reason (Harvard University Press 2017; Levine, P. (2024a). People are not Points in Space: Network Models of Beliefs and Discussions. Critical Review, 1–27 (2024a), and Levine, P. (2024v). Mapping ideologies as networks of ideas. Journal of Political Ideologies, 29(3), 464-491.

listeners, not speakers, are the main reasoners

Leave a reply

Robert Brandom offers an influential and respected account of reasoning, which I find intuitive (see Brandom 2000 and other works). At the same time, a large body of psychological research suggests that reasoning–as he defines it–is rare.

That could be a valid conclusion. Starting with Socrates, philosophers who have proposed various accounts of reason have drawn the conclusion that most people don’t reason. Just for example, the great American pragmatist Charles Sanders Peirce defines reason as fearless experimentation and doubts that most people are open to it (Peirce 1877).

Brandom’s theory could support a similarly pessimistic conclusion. But that doesn’t sit well with me, because I believe that I observe many people reasoning. Instead, I suggest a modest tweak in his theory that would allow us to predict that reasoning is fairly common.

Brandom argues that any claim (any thought that can be expressed in a sentence) has both antecedents and consequences: “upstream” and “downstream” links “in a network of inferences.” To use my example, if you say, “It is morning,” you must have reasons for that claim (e.g., the alarm bell rang or the sun is low in the eastern sky) and you can draw inferences from it, such as, “It is time for breakfast.” In this respect, you are different from an app. that notifies you when it’s morning or a parrot that has been reliably trained to say “It is morning” at sunrise. You can answer the questions, “Why do you believe that?” and “What does that imply?” by offering additional sentences.

(By the way, an alarm clock app. cannot reason, but an artificial neural network might. As of 2019, Brandom considered it an open question whether computers will “participate as full–fledged members of our discursive communities or … form their own communities which would confer content” [Frápolli & Wischin 2019].)

Whenever we make a claim, we propose that others can also use it “as a premise in their reasoning.” That means that we implicitly promise to divulge our own reasons and implications. “Thus one essential aspect of this model of discursive practice is communication: the interpersonal, intra-content inheritance of entitlement to commitments.” In sum, “The game of giving and asking for reasons is an essentially social practice.” Reasoning in your own head is a special case, in which you basically simulate a discussion with real other people.

The challenge comes from a lot of psychological research that finds that beliefs are intuitive, in the specific sense that we don’t know why we think them. They just come to us. One seminal work is Nisbett and Wilson (1977), which has been cited nearly 18,000 times, often in studies that add empirical support to their view.

According to this theory, when you are asked why you believe what you just said, you make up a reason–better called a “rationalization”–for your intuition. Regardless of what you intuit, you can always come up with upstream and downstream connections that make it sound good. In that sense, you are not really reasoning, in Brandom’s sense. You are justifying yourself.

Indeed, the kinds of discussions that tend to be watched by spectators or recorded for posterity often reflect sequences of self-justifications rather than reasoning. I recently wrote about the scarcity of examples of real reasoning in transcripts and recordings of official meetings. As Martin Buber wrote in The Knowledge of Man (as pointed out to me by my friend Eric Gordon):

By far the greater part of what is called conversation among men would be more properly and precisely described as speechifying. In general, people do not really speak to one another, but each, although turned to the other, really speaks to a fictitious court of appeal where life consists of nothing but listening to him.

Some grounds for optimism come from Mercier and Sperber (2017). They argue that people are pretty good at assessing the inferences that other people make in discussions. Although we may invent rationalizations for what we have intuited, we can test other people’s rationalizations and decide whether they are persuasive.

Furthermore, our intuitions are not random or rooted only in fixed characteristics, such as demographic identities and personality. Our intuitions have been influenced by the previous conversations that we have heard and assessed. For instance, if we hold an invidious prejudice, it did not spring up automatically but resulted from our endorsing lots of prejudiced thoughts that other people linked together into webs of belief. And it is possible–although difficult and not common–for us to change our intuitions when we decide that some inferences are invalid. Forming and revising opinions requires attentive listening, critical but also generous.

The modest tweak I suggest in Brandom’s view involves how we understand the “game of giving and asking for reasons.” We might assume that the main player is the person who gives a reason: the speaker. The other parties are waiting for their turns to play. But I would reverse that model. Giving reasons is somewhat arbitrary and problematic. The main player is the one who listens and judges reasons. A speaker is basically waiting for a turn to do the most important task, which is listening.

This view also suggests some tolerance for events dominated by “speechifying.” To be sure, we should prize genuine conversations in which people jointly try to decide what is right, and in which one person’s reasons cause other people to change their minds. This kind of relationship is the heart of Buber’s thought, and I concur. But it is unreasonable to put accountable leaders on a public stage and expect them to have a genuine conversation. None of the incentives push them in that direction. They are pretty much bound to justify positions they already held. Although theirs is not a conversation that would satisfy Buber, it does have two important functions: it allows us to judge people with authority, and it gives us arguments that we can evaluate as we form our own views.

Again, if we focus on the listener rather than the speaker, we may see more value in an event that is mostly a series of speeches.

Sources: Robert R. Brandom, Articulating Reasons: An Introduction to Inferentialism. (Harvard 2000); Charles S. Peirce, “The Fixation of Belief,” Popular Science Monthly 12 (November 1877), 1-15; María José Frápolli and Kurt Wischin, “From Conceptual Content in Big Apes and AI, to the Classical Principle of Explosion: An Interview with Robert B. Brandom” (2019); Richard E. Nisbett and Timothy D. Wilson. “Telling more than we can know: Verbal reports on mental processes,” Psychological review 84.3 (1977); and Hugo Mercier and Dan Sperber, The Enigma of Reason (Harvard University Press 2017. See also: looking for deliberative moments; Generous Listening Symposium; how intuitions relate to reasons: a social approach and how the structure of ideas affects a conversation

looking for deliberative moments

Leave a reply

It’s not so unusual to enter a discussion without having a firm view of what should be done (perhaps without even having focused on the agenda), to listen and speak for a while, and finally to reach a collective decision that wasn’t predictable in advance. Before the discussion begins, some of the participants may know what they want, but others are unsure or flexible, and the conversation forms or shifts the consensus.

However, it is not easy to find recordings or transcriptions of such processes. I spent part of the last long weekend reading the minutes of US school boards and city councils, scanning the translated transcripts of 50 village assemblies in southern India, dipping into legislative records from several countries, and even searching for archives of email exchanges within organizations that have come to light due to leaks or lawsuits. I found brief moments in which a few people seemed to be listening to make up their minds, but no sustained examples of that phenomenon.

My belief in the existence of deliberation might seem naive, except that I have experienced it many times–even in the past few months–and nobody seems surprised when it happens. I think we are used to it, but recorded examples are scarce.

Two reasons occur to me, and they may have some significance for political theory.

One is a kind of Heisenberg Uncertainty Principle for deliberation. Once you record a conversation and make it publicly available for public review, the tenor shifts. Participants know that they are speaking to an audience composed of people who are not part of the conversation, not accountable for the decisions, and not able to reply in time to affect the outcome. Participants begin articulating positions and reasons that strangers can assess later, instead of trying to decide what to do. Thus the presence of an observer shifts the phenomenon, especially if the observer is an anonymous audience in the future.

The other reason involves the typical function of official meetings. As I scan town council minutes, legislative records, and transcripts of Indian village assemblies, I see people talking under tight time pressure in order to convey specific positions. For instance:

[Member of the] Public: You have read that street lights are going to be fixed. When are you going to install?

Male (Officer): We have passed resolution, made a budget and allotted the work. Very soon we will do it. 

Public: We need ration shop at Pappanaickenpatti. There are more than 300 cards in our town. Every time we have to come all the way from our village. By the time we reach the shop, all the items will be sold out. We request you to consider our petition. We are very badly in need of separate ration shop at our village. We are ready to construct a building for this purpose.

Male (Officer): We will discuss with the civil supplies officer and let you know. (source)

Or:

[Board of Trustees] President Raymond remarked she was impressed by the commitment to distance and blended learning by the District through the creation of the Department because it illustrated the learning models were ones the District was interested in fully developing and not just using during the current school year.

Trustee Taylor wondered if the vision was to have a single platform for all distance learning in the District.

Ms. Anderson cautioned it was important the District not make any major shifts at the present time related to the changing of learning platforms since the current model was very fragile. She would like to focus more on developing best practices for the platforms currently being used, due to the current strain teachers and students were under, and then look at creating more of a sustainable model in the future.

(source)

I think these people are doing valid public work. They are putting concerns, questions, directives, and commitments on the record. Before and after these recorded moments, the same individuals may have had many conversations in which they tried to learn about the situation and about other people’s views. The public event may also spur later discussions in their community.

These transcripts do not make me cynical about deliberative democracy. I imagine that the same people listen and learn. But deliberation is very rare during the moments that are captured on the public record.

This is a challenge for me because I am helping to develop methods for modeling deliberative conversations, and I am not finding a lot of material to analyze. I am especially interested in the regular decision-making processes of governments and other organizations. Yes, I also like citizens’ forums or “minipublics”–venues where representative people are convened to discuss public questions. I even co-edited a Handbook about them! But I think they are atypical, which means that evidence derived from them would not generalize well. (Besides, Indian village councils are minipublics, and they sound a lot like US public meetings.)

If you have suggestions for transcripts or videos that I could model, I would be grateful. In any case, I think our theories of democracy should take account of this pattern. Although we imagine that deliberation should occur in legislatures and town meetings, this is unlikely. A public meeting is a moment in a larger stream of discourse. Its function is to memorialize a set of positions and reasons, while the learning and the change take place elsewhere.

The Vuslat Foundation and Generous Listening

Leave a reply

The Vuslat Foundation has opened a public website as generouslistening.org. At the Tisch College of Civic Life, we are one of their partners, as you can tell from the description of a conference that we co-organized and held at Tufts last year (a symposium on “Generous Listening in Organizations“); a blog post by my colleague James Fisher about Quaker dialogues in West Africa; and other references on their site.

The Foundation also does much work on their own or with other partners, including remarkable support-groups for women displaced by the earthquakes in Southeast Turkey in 2023.

I come to this partnership as someone who has studied political deliberation–for instance, as a co-editor (with John Gastil) of the Deliberative Democracy Handbook. SInce the late 1900s, public deliberation has been a movement of theorists and practitioners, but it is rooted in much older ideas about politics that have typically emphasized speech, communication, persuasion, and rhetoric–as both virtues and threats.

The Vuslat Foundation has helped me to shift my focus from one side of the exchange to the other–from speaking to listening. Of course, these acts always go together (even when they are metaphors for written speech, signs, or gestures). It is hardly a novel insight that communication requires at least two people. But I have benefitted from thinking more about the listening side.

First, there’s an ethical imperative. Listening well (“generously,” in the language that the Vuslat Foundation has developed) is an important virtue. Using one’s voice well is also virtuous, and sometimes even obligatory, but the need to be a good listener seems especially compelling.

Second, we can think about listening holistically. One aspect is listening to other people in deliberations, but we also listen to ourselves, to animals, waves, or the wind, to human soundscapes, to near-silence, perhaps to the divine, and to those who are long dead. I have found it useful to think of civic listening as just one kind of listening.

Third, I am taken by the “interactionist” theory of Hugo Mercier and Dan Sperber, which has helped me make sense of some of my own data–in forthcoming articles. To summarize their model crudely, imagine two or more people discussing what to do. When individuals speak, they tend to use motivated reasoning: inventing justifications for what they already want to believe, sometimes for bad reasons, such as self-interested bias. But when they listen to other people offer reasons, they are relatively good at assessing whether these points are valid, and they may change their minds. Mercier and Sperber offer an evolutionary explanation that suggests that highly social and verbal primates would develop the ability to make arguments to advance their own interests, but also the ability to assess others’ arguments in order to make good collective judgments.

Mercier and Sperber never suggest that listening always goes well. We can certainly listen selectively and exhibit bad motives when we select whom and what to listen to. But their theory suggests that we can improve individual skills and conditions for listening–perhaps more easily than we can improve speaking.

Finally, listening has spiritual (or at least psychotherapeutic) benefits that have been recognized and developed in many traditions. Although we can also gain spiritually from communicating well, the listening side is especially relevant to meditative practices of all kinds.

See also: how intuitions relate to reasons: a social approach; an agenda for R&D for democracy; “you should be the pupil of everyone all the time”; ‘every thing that lives is holy’: Blake’s radical relativism; “The Listeners“; “Midlife“.

lessons from the Virginia social studies controversy

Leave a reply

In Politico, James Traub offers a deeply reported account of the recent conflict over standards in Virginia, entitled “Virginia Went to War Over History and Students Actually Came Out on Top.”

Standards are official guidelines about what must be taught in public schools. They may influence enforceable policies, such as which textbooks are purchased and what is covered on exams, and hence the experience of students and teachers. Standards for history and civics often provoke the most intense debates, because they address the nature of our society. Although I had no involvement in the Virginia episode, I have been deeply engaged in other efforts to write frameworks and model standards for social studies, and Traub’s account rings true to me.

A very brief summary: under former Gov. Ralph Northam, a Democrat, the Virginia state department of education drafted new state social studies standards. Before these standards could be reviewed by the state board, Northam was succeeded by Glenn Youngkin, a Republican, whose campaign emphasized his opposition to “woke culture” and “critical race theory.” Youngkin named a new superintendent of public instruction and a majority of members of the school board. With those appointees in place, the state paused and then dramatically rewrote the draft standards, with input from strong conservatives.

Then the board, despite its Youngkin majority, rejected the new draft as biased and error-prone. It stepped in and painstakingly revised the document in ways that satisfied all of its members (including those who had been appointed by Northam) and drew support from outside groups viewed as both liberal and conservative. Traub writes, “The six-month debate was an absolutely terrible experience for everyone involved, yet the standards the board finally approved achieved something almost miraculous: something close to unity.”

As an example of the results, the state board coalesced around this language in the new standards document:

The standards provide an unflinching and fact-based coverage of world, United States, and Virginia history. Students will study the horrors of wars and genocide, including the Holocaust and the ethnic cleansing campaigns that have occurred throughout history and continue today. They will better understand the abhorrent treatment of Indigenous peoples, the indelible stain of slavery, segregation, and racism in the United States and around the world, and the inhumanity and deprivations of totalitarian and communist regimes. Students also will study inspirational moments … 

For me, these are the most important general lessons from the controversy.

First, although people bring prior political views into debates about what should be taught, our opinions are highly diverse (not simply left or right), and most of us want students to encounter and assess ideas that we personally do not endorse. Philosophical diversity is valuable because even those of us who want students to encounter a wide range of views may have implicit biases that can be challenged in a discussion. When serious participants who are ideologically diverse try to write good standards or guidelines together, they need not polarize into two camps, or even take predictable positions as individuals.

Debates about content are nuanced and often involve the appropriate balance between social and political history, leaders and popular movements, compelling stories and complexities, and domestic and international affairs. These questions do not necessarily have liberal or conservative answers.

Second, the hot debates are not only about which topics and ideas should be “covered” but also about how to teach. Should all students be required to learn some information, whether it interests them or not? Or should students have a lot of choice about which topics to investigate? Should students encounter highly charged topics–at all ages, only as older teenagers, or at all? Specifically, should public schools confront students with ideas that challenge their sense that they belong and are valued in the school? Does it matter which students are so challenged? Should the emphasis be on skills or knowledge, on theory or practice, and on discourse or action?

Again, these debates do not line up so that there is a right and a left camp. For myself: I believe that all students should be required to confront some information about our past that many will find uncomfortable and that relatively few students would seek out if they could drive all the questions in their classrooms. This position would seem to align me with pedagogical conservatives, except that the same points are being made most forcefully by progressives. For example, The 1619 Project is all about conveying facts deemed essential.

As many have noted, the new Florida African American History Standards basically suggest that no one supported slavery. Florida students must learn “how the members of the Continental Congress made attempts to end or limit slavery” and “how slavery increased … in spite of the desire of the Continental Congress to end the importation of slaves.” Florida students will study white people who were abolitionists, but no one who actually defended slavery. John C. Calhoun is never mentioned, let alone assigned as an author to read. Florida students are supposed to “recognize” the title of Dred Scott as a “landmark Supreme Court case” but do not have to read that decision, which declared that people of African descent could never be US citizens.

I would require students to read racist texts (no “de-platforming” Sen. Calhoun or Chief Justice Taney) and learn specific information. Ron DeSantis defends omitting that information and has ordered that “A person should not be instructed that he or she must feel guilt, anguish, or other forms of psychological distress for actions, in which he or she played no part, committed in the past by other members of the same race or sex.” In partial contrast, the new Virginia standards say: “Students should be exposed to the facts of our past in a content-rich and engaging way, even when those facts are uncomfortable.”

Since these issues have many dimensions and nuances, it should not be surprising to find views shared across political differences. The Thomas B. Fordham Foundation is generally considered conservative. Commenting on the draft Virginia standards, their reviewers said, “The Dred Scott decision is not noted by name in any of the U.S. history course standards. Its enormous impact should at the least be mentioned here in what is (presumably) the high school course.” Likewise, they criticized the omission of McCarthyism, which “led to the violation of Americans’ rights.” I find myself perfectly aligned with this feedback despite being generally quite liberal as a voter.

Third, even when people’s views are diverse, nuanced, and unpredictable, there can be political advantages to presenting differences as polarized and defining the stakes so that a majority will agree with your own side. Glenn Youngkin waged a campaign against “woke” ideology in public schools. From the opposite end of the spectrum, someone went to a lot of trouble to create a popular meme about innocuous books that the DeSantis administration had allegedly banned, when the state had banned no books.

Actual misinformation is unacceptable, but I’ll mention a closer case. Florida did not pass a bill labeled “Don’t Say Gay.” That name was affixed by Democrats and liberals who criticized the law. The relevant provision says, “Classroom instruction by school personnel or third parties on sexual orientation or gender identity may not occur in kindergarten through grade 3 or in a manner that is not age-appropriate or developmentally appropriate.”

I am not sure that the label “Don’t Say Gay” is false, but it simplifies the law in order to drive opposition to it. This mode of political debate is not necessarily wrong or bad. I oppose the actual Florida law and understand why liberals would mobilize people against it.

Martin Luther King, Jr. and his colleagues chose Birmingham, AL as their target in 1963 because they knew they could draw a clear contrast with the racist outgoing police commissioner. King wrote that a nonviolent campaign

seeks so to dramatize the issue that it can no longer be ignored. … I must confess that I am not afraid of the word ‘tension.’ I have earnestly opposed violent tension, but there is a type of constructive, nonviolent tension which is necessary for growth. Just as Socrates felt that it was necessary to create a tension in the mind so that individuals could rise from the bondage of myths and half truths to the unfettered realm of creative analysis and objective appraisal, so must we see the need for nonviolent gadflies to create the kind of tension in society that will help men rise from the dark depths of prejudice and racism to the majestic heights of understanding and brotherhood.

In short, dramatizing differences with one’s political opponent is a legitimate move in a free society. However, onlookers should be aware when this strategy is being used and should assess whether the goals are appropriate, and whether any collateral damage is necessary to accomplish the goals. They should also ask whether rhetoric has strayed from divisiveness into downright falsehood.

Ron DeSantis does not have to wage a rhetorical war against liberal educators; he could choose to deliberate with them, as the Virginia board did. Voters should recognize the choice to polarize an issue for what it is. They should not assume that it is inevitable. The Virginia case shows that another outcome is possible (although not automatically preferable) — people with diverse opinions can come to agreement.

Although politicians can be tempted to polarize, official bodies such as state boards can be equally inclined to present consensus even when they have not quite accomplished it. Above, I quoted the Virginia standards’ aspiration to “provide an unflinching and fact-based coverage” of history, but anyone may each assess whether they offer that. In my personal opinion, the list of “principles” on p. 4 is mildly problematic, presenting the debate between socialism and market economies as closed when I would ask students to think about it for themselves. But I don’t believe that this list matters much. In my view, the presentation of slavery and Black American “accomplishments” in the body of the Virginia standards is appropriate. Overall, the standards seem to take a both/and approach, genuinely including both the crimes and the successes of US history.

The whole document is quite short and general, which is itself a choice, leaving a lot for teachers to decide (for better and worse). Any major commercial textbook series would be compatible with these standards, which means that in many classrooms, the textbook will determine the content. In fact, the most important policy question may be who should decide what is taught–students, teachers, parents, local authorities, state authorities, or publishers? Because of its generality, the Virginia document may actually represent a delegation to the publishers.

See also: two dimensions of debate about civics; “Teaching Honest History:” a conversation with Randi Weingarten and Marcia Chatelain; the relevance of American civil religion to K-12 education; what Americans think about teaching controversy in schools; a conversation with Danielle Allen about civic education; etc.