Category Archives: philosophy

what makes conversation go well (a network model)

Leave a reply

I’m looking forward to presenting later today at NULab’s first annual conference, on the theme: “Keeping the Public Sphere Open.”

I think of the “public sphere” as all the venues where people come together to share experiences, emotions, and reasons in order to form public opinion. In turn, public opinion should then influence institutions; that makes the society democratic.

An open public sphere, as in the title of the conference, is one that permits and appropriately responds to every person’s ideas; no idea or person is blocked. The state can threaten the openness of the public sphere by censoring ideas or blocking individuals from participating. The marketplace can threaten the openness of the public sphere when, for instance, ISPs charge more money for some content, or when private donors flood the airwaves with campaign commercials. Thus, to preserve an open public sphere, we need policies like a strong First Amendment, net neutrality, and campaign finance reform.

But openness is not enough. The conversations within any public sphere can go well or badly. Along with several colleagues, I have been thinking about deliberation in the following way:

  1. People hold ideas prior to a conversation that we can think of as networks. Each idea may be connected to each other idea by reasons. The person’s network has content (what the ideas say) and also a form. For instance, someone might arrange all of her ideas around one central node, or might hold a set of disconnected principles.
  2. When we talk, we share portions of our existing networks, one node or one reason at a time.
  3. Interaction with other people may cause us to change our network. We can adopt ideas that other people disclose, see new connections or doubt that connections really hold, think of new ideas on our own, or even adopt contrary ideas. In any case, our personal networks are subject to change.
  4. The discussion itself can be modeled as one network to which the various participants have contributed nodes and links.

If we could develop a valid and reliable way of modeling an individual’s private network with respect to a given topic before a conversation, and then we put individuals in dialogue and modeled their interactions, I would predict that: 1) the formal properties of their networks before the discussion would influence the quality of the discussion, 2) the quality of the discussion would be related to changes in their personal networks, 3) an individual’s networks would tend to look formally similar even when the topic changed (e.g., some people would be prone to thinking about most topics in a centralized or in a scattered way), and 4) a given issue would tend to produce formally similar networks for diverse individuals (e.g., the abortion debate and a budget discussion would generate different-looking networks regardless of the participants).

There then follow a whole set of questions about what a good conversation looks like and how people should structure and change their thoughts.

See also: it’s not just what you think, but how your thoughts are organized; tracking change in a group that discusses issues; network dynamics in conversation; and assessing a discussion.

on the relationship between ethics and politics

The basic ethical question is “What should I do?” Three prevalent ways of addressing that question are: 1) to universalize, asking what you’d want anyone to do who was similarly situated, 2) to maximize, asking how you can do the most good for the most people, given your resources and options, or 3) to exhibit and develop virtues, such as courage, generosity, and truthfulness. Philosophers love scenarios in which these methods yield conflicting answers, but in a vast range of ordinary circumstances, they concur.

The basic political question is “What should we do?” The verb is plural because politics exists once people belong to groups of any kind, from small voluntary associations to nation-states. To be sure, the ethical question never vanishes, because you can ask whether you should belong to a given group and what you personally should do in relation to it. But the plural question raises a new set of issues that are not directly addressed in individual ethics.

For one thing, we decide what we should do together—not necessarily democratically or equitably, but as a result of several people’s influence. Since each of us is fallible, and other perspectives have value, it may be wise to yield to a group’s judgment even if you would have done something different on your own. You may be especially inclined to go along with a group’s decisions if its processes were equitable and deliberative. The virtues of intellectual humility and civility argue for supporting the group’s decision. But that is the wrong choice if the group is misguided, and you retain the options of resistance or exit.

This means that issues of complicity arise in politics that are not salient in individual ethics. A group to which I belong acts in my name. Am I therefore complicit in the harm that it does? On the other hand, how do I know that what I would have decided alone is really better than what the group has decided by discussing?

The group has potential value. It can accomplish more than an individual can—whether for good or evil—as long as it holds together. To form and maintain a functioning group is an achievement, requiring individuals to coordinate their behaviors and often to sacrifice for the whole. Because groups have potential and are vulnerable, it can be wise to support less-than-ideal behavior in order to maintain the group for another day. In Talking to Strangers, Danielle Allen emphasizes that democracy always involves sacrifice, and the sacrifice is usually unequal. So the question “What should we do?” implies that all have given—and some may have given much more than others—to create the “we” and to act together. There comes a point when the sacrifice is too high or too unequal to sustain, but some sacrifice is necessary to create the conditions for politics in the first place.

Each of the prevalent methods for addressing individual ethical questions can be applied in politics, but with important modulations.

First, instead of universalizing in a hypothetical mode, we can create actual covenants that bind all. In ethics, a person asks, “What would I want anyone to do if she faced my situation?” In a group, however, we can ask, “What must everyone actually do in situations like this, and how will we set and enforce penalties for those who fail to do it?” Sometimes, actual covenants should differ from ethical norms, because it can be wise to overlook or even accept non-ideal behavior in order to preserve liberty or to maintain a group whose members would quit if the rules were too strict. That means that the logic of real covenants differs from the logic of hypotheticals.

Second, instead of maximizing the benefits of individual actions, we can maximize the benefits of what a group does together. The main difference is that we must consider the group’s future capacity to act effectively. In many cases, a group that maximizes net benefits for the world would dissolve, because the level of sacrifice expected of its members would be too great, and they would exit. Since the existence of a group permits deliberation and coordinated action, which are impossible for individuals, dissolution may be too high a price to pay.

Christopher Winship acknowledges that justice demands raising the quality of the schooling available to the least advantaged American students. However, he argues, “the best way to approach serving the interests of the least well off [may be] to avoid policies that decisively pit the interests of the less advantaged families against those of the more advantaged families.” He cites evidence that Scandinavian countries have achieved the highest levels of shared prosperity and economic equality in the world today not by directly pursuing equality but by negotiating policies that are attractive to business as well as labor. These compromises have created durable and accountable states that have been able to deliver high-quality services for all.* This is an example of how preserving the group (in this case, a Nordic democracy) can do more good than maximizing the benefits of the group’s actions at any given moment.

Third, we can consider the virtues of a group—virtues understood, in an Aristotelian way, as dispositions that are reflected in, and reinforced by, actions. In other words, virtues are habits that can be deliberately shaped. Groups as well as people can have virtues, such as courage, temperance, magnanimity, etc. Developing and maintaining virtues requires different strategies when a group instead of an individual is the thing that is virtuous or vicious.

This discussion has assumed a simple dichotomy of individuals and groups. That scheme must be complicated in two fundamental ways.

On one hand, individuals do not really precede groups. Anyone who thinks in a language is already part of a linguistic community. Anyone who asks of her nation “What should we do?” probably developed her opinions under the influence of that already-existing nation. These are examples of the ontological dependence of individuals on groups.

On the other hand, groups are rather like individuals in their interactions with one another. Robert O. Keohane and Elinor Ostrom co-edited a book that explored the close parallels between collective-action problems in small communities and among states. In both contexts, there is typically no single enforcer who can determine the behavior of the parties. There is plenty of room for disaster, yet sometimes the parties work out solutions, from rules for pasturing goats on common land to international arms treaties.

Furthermore, governments do not merely work “within their jurisdictions by imposing authoritative rules on their subjects” (p. 11) Even dictatorships cannot do that, because they cannot police and control their populations without a great deal of voluntary cooperation. A government is not a single actor that stands apart from society and directs it, but rather as a whole set of human actors (politicians, civil servants, front-line workers) who constantly interact with each other and with people outside the government. Not much is accomplished unless they are able to motivate voluntary compliance with agreements.

Elinor and Vincent Ostrom and their Bloomington School see governance as “polycentric.” At the local level, we are constantly interacting in game-like situations with other people who may either cooperate or not. There are islands of command-and-control in which some individuals tell others what to do, but their capacity to control usually depends on norms of willing compliance. Nation-states exist in a global anarchy, without any power above them, but they have managed to work out some arrangements for cooperation. And between nation-states and local communities are complex webs of arrangements involving intermediary organizations such as municipalities and regional governments, parties, interest groups, and media organizations. Cooperation, competition, and mutual destruction are all possible in all of these contexts.

I think that the categories of the ethical and the political constantly recur at all scales, and which one is most salient depends mainly on the perspective that seems most appropriate in the situation, that of an “I” or a “we.”

That said, scale matters, because it influences how we should think about agency and responsibility. We shoulder the most responsibility at the smallest scales, especially when we act alone. Agency is also most tangible at that scale: we can see what we accomplish by ourselves. However, we cannot accomplish much. At very large scales, agency is hard to detect because millions or billions of others are also at work, and it is unreasonable to expect the whole population to shift at anyone’s will. In the middle range (which I think is under-theorized), we can take part in effective action. That is politics. Politics is an ethical matter, in the broadest sense–there is a difference between right and wrong–but the ethical principles appropriate for individual action no longer suffice. A new set of considerations becomes important when we move from I to we.

*Christopher Winship, “From Principles to Practice and the Problem of Unintended Consequences,” in Meira Levinson and Jacob Fay, eds., Dilemmas of Educational Ethics: Cases and Commentaries (Cambridge: Harvard Education Press, 2016), pp. 177-8.

See also: against methodological individualism; is social science too anthropocentric? and two basic categories of problems.

don’t confuse bias and judgment

Leave a reply

“Even good and, at bottom, worthy people have, in our time, the most extraordinary fear about making judgments. The confusion about judgment can go hand in hand with fine and strong intelligence, just as good judgment can be found in those not remarkable for their intelligence.” — Hannah Arendt to Karl Jaspers, Dec. 29, 1963, translated by Elisabeth Young-Bruehl.

The best lack all conviction, while the worst
Are full of passionate intensity.
–W. B. Yeats

(Dayton, OH) A bias is a manifestation of cognitive limitations or weaknesses. It reveals that a person has failed to see the whole picture, to weigh evidence appropriately, to revise prior assumptions, to counter the influence of self-interest, or to grasp other perspectives. We are all subject to bias, constantly. It is wise to scrutinize our own and other people’s positions and opinions for signs of it. Certain professionals–reporters, researchers, teachers–are taught to acknowledge bias and to counter it. They are held accountable for any bias they appear to show. Anyone who writes in a public forum and purports to be a scholar or a teacher will sooner or later see the comment: “You’re biased!” (Or, just as likely, “Your bias!”, since grammar is not always the strong suit of the affronted.)

At the same time, each of us must make good judgments. Faced with almost any situation of any consequence, we can judge wisely and well, or judge badly, or fail to judge at all and thereby display negligence. Even journalists, scholars, and educators who advertise themselves as minimally biased must make constant judgments: what to cover, study, and teach, how to present information, whom to address, whom to take seriously, what counts as a legitimate position in a debate, and so on.

It’s essential to separate the language of bias from the language of judgment. They have different grammars: for instance, we say “good judgment” but not “good bias.” We accuse people of bias but not of judgment. A person can make the right judgment despite being biased; in fact, her bias may alert her to what really is the right conclusion. Or a person can somehow counter his own biases and yet make poor judgments. A typical example of the latter is a person who decides that taking a position would evidence bias and therefore fails to act–e.g., in response to a presidential candidate who is violating fundamental norms. Bias is empirically demonstrable, but demonstrating it does not prove that the speaker has reached the wrong conclusion. Judgment is wise or poor, but the difference is not empirically demonstrable in a straightforward way.

I think confusion between bias and judgment is one of the reasons that “even good and, at bottom, worthy people have, in our time, the most extraordinary fear about making judgments.” And that leaves the worst of us to display the most passionate intensity.

evolution, game theory, and the morality of modern human beings

Leave a reply

It’s valuable to model the development of phenomena like altruism and spite (harming someone else at a cost to oneself) by combining game theory with evolutionary theory. The results should be seen as predictions to be tested against empirical evidence about actual organisms. My question is what this combination tells us about our situation as human beings in historical time.

The basic assumptions are:

  1. Organisms interact with each other so that each one can win or lose. For instance, a parent feeds its child, a predator eats its prey, a mite hitches a ride on an insect. These interactions can be modeled as games in which each player makes a choice (e.g., kill or don’t kill), and positive or negative outcomes result for each as a joint result of their decisions. (“Choice” is a metaphor, because completely non-sentient organisms can be modeled as players in a game. For instance, a plant can release a chemical or not.)
  2. Changes in how organisms interact in game-like situations arise more or less randomly. As a result of a genetic mutation, an organism may begin to mimic another species. Or, as a result of a change in climate, an organism’s prey may become scarce.
  3. If playing a game in a particular way increases the odds that a species will reproduce, that behavior will become more common. For instance, if mimicking works, it will spread.
  4. This means that the payoffs that matter from game-like interactions among organisms are best measured in terms of evolutionary fitness. Whatever an organism may want or think or feel, what matters is whether its chances of reproducing increase or decrease.
  5. Given the first four assumptions, under certain conditions, behaviors that we might consider proto-moral, such as helping offspring, helping others outside the family lineage, or even sacrificing oneself to punish another for violating a norm, predictably arise. By being altruistic (or punitive) in game-like interactions, an organism may gain evolutionary fitness.
  6. Thus we can explain proto-moral behavior through a combination of game theory and evolutionary theory. The behavior is a consequence of background conditions. This form of explanation applies to homo sapiens, who have cognitive capacities and instinctive drives for things like fairness and punishment because of the conditions that pertained before historical time when we evolved into our current form.

Game theory is part of my own toolkit. I believe it clarifies many situations that confront human beings as we interact with each other and helps us to devise solutions to collective problems. I also acknowledge that we are a biological species that evolved with certain capacities and drives, and that inheritance must be taken into consideration as we diagnose and try to address our problems as a species. However, I tend to believe that Darwinian evolution gave us certain capacities that now fundamentally change the premises described above (points 1-6):

  1. We can design games. The original Prisoner’s Dilemma, for example, is a situation intentionally created by a prosecutor within a legal system. The prosecutor could change the game, or he could be required to change it by a legal reform. A shared pasture is a very different game from a Prisoner’s Dilemma, but it’s also intentionally designed.
  2. We can choose goals. If natural selection determines change in a whole population, then it doesn’t matter what each organism wants; it matters what promotes survival and reproduction. But human beings can choose what we want in specific interactions. Sometimes we want things that reduce our chances to survive and reproduce, but we compensate with other strategies.
  3. We can change our identities. If a person’s main identity is a parent, his impact on his own offspring is central. But he could instead choose to identify primarily with a church, a community, a nation, or other grouping.
  4. We can design and change the groups within which our interactions occur. As an example, the size of a group influences how organisms interact. But we human beings can merge small groups to form vast nations, decentralize governance to small groups, nest communities within states, or place people in multiple overlapping groups. We can intentionally vary not only the size of groups but also their internal diversity, spatial extension, equality of influence, and cost of entry and exit.
  5. We can influence individuals’ predilections to play games in various ways, e.g., to be altruistic, trusting, selfish, spiteful, or punitive. We can influence children in lasting ways by raising and educating them to have certain character traits. We can also influence behavior in local and temporary ways by changing messages and contexts to encourage desired behaviors.
  6. We are influenced in all the above choices by norms, but we disagree about the best ones, and the available norms are the results of deliberate human creativity. In short, we invent and choose norms.

Game theory remains relevant–in fact, it is an especially useful toolkit for a creature that is capable of designing and redesigning its own interactions. I am less sure that evolutionary theory is relevant, except insofar as it explains certain proto-moral tendencies and limitations that now contribute to our challenges.

Notes: Points 1-6 are guided by my Tufts colleague Patrick Farber and specifically his excellent paper “Reciprocal Spite” (with Rory Smead). Points 7-12 are heavily influenced by Elinor Ostrom, who emphasized the diverse outcomes that result when people face collective action problems and the importance of their intentional choices about groups, rules, and norms.

John Stuart Mill, Stoic

Leave a reply

I sometimes envy my fellow academics in the humanities who regularly renew their acquaintance with fundamental works that have slipped pretty deep into the well of my own memory because my job is to conduct and administer empirical research about current politics. For just that reason, I am thoroughly enjoying reencountering some major works as I teach first-year undergraduates this semester.

For instance, I now see Mill’s Utilitarianism in an entirely new way thanks to re-reading it with students after our extensive discussions of authors like Epicurus, Buddha, and Emerson. It seems much less an explanation of the utilitarian principle of justice (maximize everyone’s happiness) than I had remembered, and more an exploration of how an individual should pursue happiness. It thus belongs to a genre that Mill knew very well, the tradition of therapeutic philosophy inaugurated by the Hellenistic schools and revived by Montaigne.

In the text of Utiliarianism, Mill refers several times to Epicureanism and Stoicism. For instance: “I do not, indeed, consider the Epicureans to have been by any means faultless in drawing out their scheme of consequences from the utilitarian principle. To do this in any sufficient manner, many Stoic, as well as Christian elements require to be included.” This passage suggests that Mill is interested in constructing the kind of “eclectic” view (drawing from multiple Hellenistic schools) that was popular from the time of Cicero and continued in early Christianity. Continue reading →