Resisting the Fatalism of the Behavioral Revolution

I love Peter Levine’s latest post, “don’t let the behavioral revolution make you fatalistic.”

“Tversky’s and Kahneman’s revolutionary program spread across the behavioral sciences and constantly reveals new biases that are predictable enough to bear their own names. […] These phenomena are held to be deeply rooted in the cognitive limitations of human beings as creatures who evolved to hunt-and-gather in small bands on African plains. Not only has the burgeoning literature on cognitive biases challenged rational market models in economics, but it undermines the “folk theory” of democracy taught in civics textbooks and widely believed by citizens and pundits.”

I think Levine captures something important about the literature on cognitive biases and heuristics: that they tend to put people in labs and poke them in such a way as to show the ways in which individuals are prone to mistakes. Yet this is widely known, and many of the worst mistakes to which individuals are prone are things we have developed solutions for in ordinary life.

“Behavioral science would have predicted the demise of the independent newspaper–but about a century too soon. In fact, “the press” (reporters, editors, journalism educators, and others) sustained the newspaper as a tool for overcoming human cognitive limitations for decades.”

As such, the lab work undermines methodological individualism but doesn’t actually help us understand communities of inquiry or institutions of knowledge-production. We are extended minds, always dependent on cognitive “prosthetics.” We depend on watches and newspapers and Google and our friends to remember and process information. And yet I think Levine is perhaps too optimistic about the possibilities of “prosthetics.” (One of Levine’s finer qualities is that he regularly make “too optimistic” seem realistic in retrospect.)

I think we should especially push on the idea that journalism is or has been a solution to cognitive limitations. The golden age of journalism was a short period of time marked by very low elite disagreement on major issues as they joined forces against communism and to first suppress–and then manage–the Civil Rights revolution for women and Black people. This cynical potted history of the trustworthy news ignores much–but so does the optimistic one.

I’ve always thought that the main power of the “behavioral” revolution was to give scientific precision and credence to insights from earlier philosophy, political theory and psychology, as well as parsing the size of effects and the disagreements between cliches that would often emerge. So sure: you can find a lot of Tversky and Kahneman in Francis Bacon, Adam Smith, David Hume, and Friedrich Nietzsche, but you can also find a lot in those authors that has been overturned or rendered more carefully in later work.

And the big insights about democracy’s weaknesses–the ones that go back to Plato and Aristotle–those didn’t go away in the middle of the 20th Century. They were perhaps suppressed by the Cold War abroad and the race war here at home, but something big happened when the demographic models for redistricting got an order of magnitude better in 2010 than they had been in 2000. And those models are just getting better.

What’s more, the behavioral revolution can also be used for good: I’ve repeatedly defended these insights when applied to criminal justice, for instance. And one of the most famous “cognitive bias” studies come not from the lab but from the real world. Danziger, Levav, and Avnaim-Pesso showed that:

“experienced parole judges in Israel granted freedom about 65 percent of the time to the first prisoner who appeared before them on a given day. By the end of a morning session, the chance of release had dropped almost to zero.

After the same judge returned from a lunch break, the first prisoner once again had about a 65 percent chance at freedom. And once again the odds declined steadily.”

This is the kind of thing that we might have suspected before. Any professor with a stack of papers to grade might have suspected they were more lenient after dinner, for instance. But this is definitive, real world proof of a problematic bias, a result of the behavioral revolution.

Ironically, it doesn’t actually make me very fatalistic: it gives me hope. I hope that Israeli judges are reading this and worrying about it. I hope they are taking snacks to work. I hope that parole lawyers everywhere are taking note of these facts and acting to protect their clients from these biases. New information about our cognitive limitations doesn’t have to make us hopeless. And really, that’s Levine’s point.

The Problem with Honor: Cold Wars and Hard Hearts

Dr. J responds to my criticism of her position on plagiarism detection. I am, she accuses, guilty of Cold War paranoia and preemptive warfare with my students:

That is to say, the classroom is a millieu in which everyone is suspicious and noone can be trusted, so every preemptive security mechanism one can employ should be employed. This is, in my opinion (and all due respect to AnPan), an awful way to conduct a classroom, especially a Philosophy classroom. That kind of CYA attitude is exactly what motivates students to cheat in the first place, convinced as they tend to be that grade-grubbing and ladder-climbing is more important than actually learning. I cannot stretch my imagination far enough to imagine a scenario in which condoning and confirming this world-view is beneficial to anyone. I am not my students’ antagonist. I’m not their enemy. We are not at war with one another.

“An awful way to conduct a classroom”? Ouch! But if Dr. J is right, I want to change my conduct. Is she? At this point, I feel like I’m no longer a good judge: some of the things she says in that post make me feel less receptive to her arguments, and indeed she suggests that my own use of the phrase “turning a blind eye” to describe her attitude towards cheating may have ratcheted up her own feelings of offense. In such a rhetorically hardened atmosphere, how can either of us expect to make accurate judgments? Phrases like “awful conduct” and “blind eye” are affronts to our honor: to accept that Dr. J is right would be to accept that I am an awful teacher, just as for her to accept that her policy towards plagiarism detection software is “turning a blind eye” would be an affront to her own participation in the Rhodes Honor Code, which requires signatories to report infractions. On this matter, we have hardened our hearts and our reputations are at stake: how can we have open minds?

“Honor Talk”

It’s a conundrum, certainly. I believe it is a conundrum generated by the question of honor itself, which is one of the reasons we ought to avoid “honor talk.” One objection to the language of honor and deference is that honor cannot operate without the structure of the affront to honor which demands satisfaction. Any insult, slight, or impugnment threatens an honorable person’s status, and thus it calls for a response: a demand for retraction or satisfaction… the duel. For me, the problem with honor is that it shuts down (reasonable, friendly, cooperative) conversation. It renders some things out of bounds of polite, honorable discussion. As a person grows in status, this penumbra of the unsayable becomes larger: just think of the things you’d discuss with your friends that you wouldn’t dare mention in front of the President. But this expanded aura of respectability is merely a customary privilege. It can’t be due to every person, and so it is at odds with equality. And so, honor produces error.

I think this is one part of my objection to an honor approach to plagiarism that Dr. J does not really pick up in her response. In my last post I wrote:

The commitment to merit seems especially to operate among elites, who will not bear any questioning of their integrity and use this language to justify their privileges. […] Though we may not like this power, we cannot ignore our role as gatekeepers, giving grades and writing recommendations that will serve, in future, as the justification for claims of merit and honor. If those grades and letters are based on false premises, the subsequent merit will be similarly false. But oh, how its bearers will pretend otherwise!

Dr. J doesn’t bring this concern up, and when she does mention the passage from which I’m quoting, she doesn’t perform her standard charitable reading of my position (I think because my position was not clearly stated) and so merely reiterates her own:

For my part, I think we’re morally obligated to respect persons, and morally prohibited from holding persons in contempt, qua “persons.” Things like Honor Codes, on my view, enforce these principles. Honor Codes operate on the presumption that it is not required in advance that one “prove” oneself “deserving” of respect, though the implicit caveat of those same codes is that if one does violate the presumptive trust, then one runs the risk of being expelled from the community that the Honor Code governs.

Let me try to clarify what I meant. “Respect” is one of the classic examples of Wittgenstein’s “family resemblance term,” insofar as it might mean many things in different contexts. When we speak of equality of respect, we generally mean something like the duty to treat each person as an autonomous end-in-themselves, both subject and legislator of the moral law. This kind of respect is intrinsic and indefeasible. It’s always a violation of a person’s equality to withdraw or refuse this sort of respect.  It’s precisely this kind of respect that rules contempt for persons out, on my view.

But “respect” can also be an evaluative term, used as an analogy for esteem: we hold some in more esteem than others, they are “highly respected.” To my mind, there is a tension between the evaluative and the egalitarian conceptions of respect. We cannot both hold each other simultaneously as equal ends-in-ourselves and as better and worse. Conceptually, of course, we can distinguish the equality of persons and the inequality of their actions or characters, but we have a strong tendency to let the one bleed into the other. In metaphysical terms, we mistake a person’s predicates for their substance.

Even if it weren’t at odd with equality, it would still be at odds with epistemic virtue. The fact that it is more difficult to admit that we are wrong and to change our position when our status is under threat seems like ample evidence of that!

Bayesian Reasoning in Light of New Facts

If we are now at odds over status, there might still be a way to return our attention to substantive question. Here’s one place to start: Dr. J did not always feel so confident of her views on cheating and plagiarism. Back in 2009, Dr. J wrote this about her attitude towards cheating:

I’m not absolutely committed to my position on this. I would say that I’m about 80% committed. I can be dissuaded, but it’s going to take some convincing. That’s what the comments section is for.

In the spirit of finding a common ground, let me reiterate two facts that seem material, in hopes that they will reach the principled disagreement between us.

1. I have not prosecuted any cases of plagiarism since I began using the detection services. Prior to that, I prosecuted one or two cases each semester.

2. Here’s why I no longer need to prosecute plagiarism: students submit their papers for grading electronically through SafeAssign, and they can revise those papers to provide proper citation for any material marked unoriginal. I only see the revised paper, not the original with potential plagiarism concerns.

Here’s how I see these facts reflecting a principle upon which Dr. J and I might agree: this swift, sure, and minor punishment (non-judgmental revision) seems preferable to delayed, uncertain, and severe punishment (catching an obvious plagiarist and failing him or her for the assignment or class.) The principle, for me, is that prevention is always better than punishment. Prevention does sometimes involve preemption and precaution, but so what? Not every precaution is reducible to post-9/11 paranoia.

Unfortunately, I don’t think Dr. J would put her commitment at only 80% today, in part because in our discussions of the matter we’ve both taken a harder line. But what’s really changed? Back then, she cited three reasons: (1) detection isn’t worth the costs of time or lost trust, (2) philosophy professors are not police, and (3) “constant and ubiquitous surveillance actually weakens students’ ethical sensibilities.” Though obviously electronic plagiarism detection partially eliminates (1), these are basically the same objections she has repeated in this exchange. (Plagiarism detection no longer takes up our time, but it may still lead to lost trust.)

Dr. J also suggests that one of her motivations to attend less rigorously to plagiarism concerns than her colleagues is that the harm accrues to the student, that the cheater “suffers a paucity of reflection, a diminishing of his or her deliberative capacity, an incomplete and crippled world-view.” I agree with this latter, of course, though I suspect that our true disagreement is what our obligations are vis-a-vis those with “impoverished reflection” and “crippled world-view.” What are the reasonable accommodations for such disabilities? What costs will those who are fully able and trustworthy bear in the name of those who are trust-challenged? I think plagiarism detection software is a reasonable accommodation, and Dr. J does not. In large part, this is an empirical question: how much does plagiarism detection actually reduce trust between professors and students? It certainly makes me feel more trusting. How much does it “weaken ethical sensibilities”? How would we measure that?

Policing the Classroom?

Thankfully, we can escape these empirical questions for a while, because for Dr. J, the primary question is the question of “policing”:

I’m surprised that AnPan didn’t state more explicitly his objection to my objection to (preemptive) “policing,” because it seems to me that it is on this point that our positions fundamentally diverge. I don’t see anything in AnPan’s articulation of his position that indicates he has any problem whatsoever with being a policeman.

Do I have any problem (“whatsoever”) with being a policeman? Well, I call myself a pacifist, but in a weak way: I would use force in self-defense myself, and I believe it is often necessary to protect the vulnerable. I just don’t like guns or violence, and I tend to believe that there are few (if any) just wars. But that’s not Dr. J’s issue, which is why it seems like hyperbole to call plagiarism detection “policing.” One thing that worries me about the metaphor of policing is that there’s a big difference between the kinds of problems that come with the state’s monopoly of “legitimate” force and the kinds of issues that we’re discussing here. Both Dr. J and I are reaching towards higher-order conceptions of justice and political community to discuss this classroom policy. But while this does seem to require some discussion of the role of honor and respect, it doesn’t necessarily require us to invoke the language of warfare or policing. One of the pernicious effects of Foucault’s discussion of disciplinary power is that it has become tempting to reduce every form of control to the coercive methods of the prison or panopticon. But I think this is hyperbolic. There’s no threat of force, no mixture of juridical power and some specialized knowledge.

The only thing that smacks of coercion is the evaluative function of university courses, the way they are graded and those grades help future employers sift applicants. But we’re all implicated, there, whether or not we use plagiarism detection software! So yes, insofar as we enforce rules, we professors take on something like a policeman’s law enforcing function. Here are some things that some professors “police” in their classrooms: the use of electronics and cell phones, interruptions, sexist and racist language. In my classroom, I only police the last of those, as I welcome interruptions and don’t care about electronics or cell phones. Plus, of course, we all use our unilateral authority to set the length, subject, due dates, and standards used to evaluate students’ performance. Even Dr. J does that.

Do these things constitute a failure to respect our students’ autonomy or honor? They might, but I don’t think they must. We can treat electronics scans for plagiarism as a matter of honor, or we can treat it as a kind of technical requirement, like a due date or page length requirement. If we don’t treat it as a matter of trust and honor, then perhaps it will not be experienced as a matter of trust and honor. In any case, I am not sure that we ought to view an offended students’ honor as truth-conducive, given that I’ve already suggested Dr. J and I should ignore slights to our esteem or dignity to ferret out the truth.

Perpetual Peace

The Enlightenment project was, if not exactly founded upon, at least encouraged and made international, by the challenge of Saint-Pierre’s A Project for Settling an Everlasting Peace in Europe. All of the eighteenth century’s philosophers took it up, and while they disagreed on the exact means, all felt that reason could lead the way. Saint-Pierre and Rousseau were persuaded that only an international federation, which brought together various European nations and restricted their sovereigns in military matters, could overcome the amour-propre (overweening self-regard) of monarchs. Voltaire challenged the notion that the rule of law would be sufficient to eliminate colonial violence, since he argued that the worst barbarities were performed by Christians against those whose religions they could not tolerate. In this, Voltaire demonstrates a keen grasp of the growing exportation of violence to the empires of the various European states, and argues that toleration for difference, inculcated through the unprejudiced use of reason, is the only solution. (“Peace, without toleration, is a chimera.”) Yet Kant did him one better, arguing that understanding and logic alone could not enforce toleration, but that specifically moral reason must be cultivated: he eventually recognized that this would require a cabal of reason, a sort of secret Masonry that would attempt to change religious and political institutions from within by exerting slow, but constant, rational pressure. Neither rules nor education alone could accomplish world peace: it would be necessary to change both the institutions and the culture simultaneously, which could only happen over time.

In the twentieth century, we’ve largely given up on the association of reason with pacifism. It has become popular to show that Enlightenment sensibilities bring their own, much more deeply embedded reasons for intolerace and barbarity, such that Voltaire’s hoped-for transition from religion to reason is the primary obstacle to peace. Perhaps the most famous argument for this view is Foucault’s book Madness and Civilization, where he argues that our pathologization of difference has gained the respect of medical experts, who allow their prejudices to become diagnoses, and then torture their subjects in an attempt to make them ‘well.’ Foucault’s work sparked a major shift in psychiatric practice, and his general concerns were popularized by novels like One Flew Over the Cuckoo’s Nest and Catch-22: today, we seem unwilling to electrocute those who make us uncomfortable.

Yet after two world wars, and in the midst of the Cold War, there did not seem to be any hope for a cessation of violence as such: just a softening of the domestic injustices that were close enough and small enough for a citizen to intervene. We have washed our hands of reason, since it seems only to supply firmer resolve in war and more dangerous weapons with which to fight it. What happened to any hope that an international federation like the UN might suppress hostilities? Obviously, the UN can’t accomplish anything without abridging the sovereignty of its member-states, just as Saint-Pierre initially proposed. What about education? Well, with such ambivalence amongst the world’s educators regarding the desirability of violence, it’s no surprise that our children come out as divided as their parents and teachers. What about the cabal of reasonable men and women, committed to ending violence a little bit at a time? In this case, I think the pacifists are losing ground to the neo-conservative, fundamentalist, and totalitarian cabals, since the major problem with secrecy is that it always confounds the means of reasonable discourse.

The fact that reasonable people (libertarians, egalitarians, and thoughtful conservatives) are more concerned with marginal tax rates, identity politics, and electoral mishaps than with sharing their freedom from domination with the rest of the world, means that they’ve abandoned the most important part of their participation in reason. They’ve lost track of which goals are worth striving for and devoting your life to, and which ones are simply amusing or interesting diversions. The fact that many Americans think that freedom can be shared at the business end of a rifle means that they’ve misunderstood the entailment relationship between means and ends. We need, I think, a new course of study in teleological reasoning.