Scalia, right? Nope. Scalia barely cracks the top ten, behind Alito, Kennedy, Thomas, and even Breyer. The actual measure is “friendliness” rather than meanness, and these guys have among the lowest friendliness scores, which is the percentage of positive words used by justices in their opinions minus the percentage of negative words. (Negative and positive words taken from here.)

The friendliness score comes from A Quantitative Analysis of Trends in Writing Style on the U.S. Supreme Court, a new paper by Keith Carlson, Michael Livermore, and Daniel Rockmore, and it contains all kinds of other fun stuff, like the influence of law clerks on judicial writing style. The authors are pioneers in the use of textual analysis to analyze supreme court opinions. One of their findings is that opinions of modern justices are a lot less friendly than the opinions of earlier justices. (They are also written at a lower grade level.)

The friendliest justice–by a long shot–is John Jay, reflecting perhaps his experience as a diplomat. But he wrote very few opinions. I’m therefore handing the title to #2, Oliver Ellsworth. And the meanest? An obscure, one-term justice named Thomas Johnson. [N.B.: an earlier version of this post confused him with William Johnson. The ABA Journal correctly identified him.]

Guest post by Ryan Doerfler, Bigelow Fellow and Lecturer in Law, The University of Chicago Law School:

I too find the position Eric discusses (absolute prohibition against torture plus judicial leniency for justified instances) puzzling, or at least frustrating.

My sense is that there are two explanations for the position.  The first, which Eric discussed, has to do with incentives or, as I would put it, epistemic reliability (maybe these are the same at the end of the day).  The argument here is an application of the more general argument for rule utilitarianism: Because individuals will overestimate systematically the considerations speaking in favor of particular sorts of action (e.g., torture, lying) if allowed to reason on a case-by-case basis, better to adhere to absolute prohibitions as a bulwark against bad reasoning.  As is obvious, one would have to do the math to determine whether an absolutist regime is preferable to a case-by-case regime in a given instance since there will be errors under both.  The suggestion of judicial leniency in the case of torture indicates that even those advocating an absolute prohibition do not think the math comes out favorably if the prohibition is really absolute.

Thinking about non-repeat players, I guess I do not think of the suggestion as arbitrary.  Plausibly, the tendency to overestimate the considerations speaking in favor of torture are greater in the heat of the interrogation chamber than in the cool of the courtroom.  And, so long as would-be torturers are unaware of the prospect of judicial leniency (plausible, in the case of non-repeat players), one might get decent results under this regime (e.g., one would torture only when the apparent need to torture was so great as to warrant personal sacrifice) without human sacrifice.  This is all speculative, of course, but at least not implausible.  The problem is that, in the real world, would-be torturers are almost all repeat players (or at least members of repeat-play institutions).  Hence, the prospect of judicial leniency would be well known.

The other explanation for the position, I think, has to do with the impulse to preserve both absolutist (e.g., ‘Thou shalt not kill’) and non-absolutist (‘Thou shalt not kill, unless …’) moral intuitions.  My sense is that this impulse has not to do with accuracy or expected outcomes but instead with bedrock intuitions about moral decency or something like that (e.g., a common attitude is that one should cringe at images of killing or torture, even if the killing or torture in question is justified).  That impulse manifests in various places in moral philosophy.  Where it is plainest, though, is in discussions of so-called “dirty hands” cases, i.e. cases in which a particular action is justified but somehow morally problematic, regrettable, etc.

I have always found this idea hard to understand (e.g., If an action is justified, how could it be regrettable?).  But, for whatever reason, it has real popular appeal.  One high-visibility, non-scholarly example is the television show 24.  On the one hand, 24 is written in such a way that the audience can be expected to think of Jack Bauer’s actions as justified (as Justice Scalia said, what jury is going to convict him?).  On the other hand, the show is written such that (and the showrunners are expressly of the view that) Jack Bauer must suffer so that the rest of us can be safe.  I suppose one could interpret this as a metaphor for the psychological costs torturers must incur, which are real.  More plausible, though, is that the underlying attitude is that Jack Bauer in some way should suffer for what he has done, even though what he has done is justified.  Again, I think this is confused.  But it is a pervasive sentiment.  What Jack Bauer does is right … but also wrong.  Good … but also bad.  Alas.

[N.B.: there is an ancient literary and artistic theme that the person who saves the community by breaking its norms must himself be expelled from the community, or otherwise suffer and be made an outcast. This person must be a hero who follows a higher morality and accepts the sacrifice. I think philosophers like McMahan unconconsciously reproduce this logic, not realizing that institutions cannot themselves be designed to permit the exceptional act. In real-life institutions, if you tell agents they will be punished for doing X, and they believe you, they won’t do X. — EP]

An interesting article here in the NYT about the “Uber model.” Uber drivers enjoy flexibility–they can drive whenever they want–partly because the app connects them to customers but mainly because there are so many drivers. People who want rides can get them because of the high probability of a nearby riderless Uber car. The author argues that this model can be applied to many other settings, including legal services and medicine. A doc with a bit of spare time can make himself available via app and you might consult him if you happen to be nearby.

The relevant law here is not the law of taxi or doctor licensing but the law of large numbers. It’s what ensures that someone is nearby when you need him, even though drivers and doctors have all kinds of other unpredictable commitments, given a large enough pool. I tell my students that the most important law in banking regulation is the law of large numbers. It’s what makes it possible for a bank to offer money in a steady way to borrowers when the bank’s own lenders–short-term depositors–might need their money on a moment’s notice. The Uber model is, at bottom, the bank model.

In the NYT, the philosopher Jeff McMahan argues that torture is almost always morally wrong, but he believes that in certain cases–when it is used to prevent a greater evil like killing or mass killing–it is morally permissible or even morally obligatory. Yet he believes that torture should be banned even then. That doesn’t seem to follow. If McMahan’s moral position is correct, shouldn’t the law permit torture just in those conditions when it is morally permissible, and otherwise ban it?

As far as I can tell, his answer is that morally justified torture may be so rare that it can be safely ruled out by an absolute ban. That may well be right; maybe that is the lesson of the Bush torture debacle. But it does seem puzzling. Consider shooting-to-kill. Shooting-to-kill is also a horrible thing to do to people (worse than all but the most extreme forms of torture), and it is rarely justified. Yet police officers are permitted to shoot-to-kill in order to prevent greater evils. Even after recent events suggesting abuse  and discrimination in police shootings, no one wants to impose an absolute ban on them.

McMahan, like other philosophers (such as Henry Shue, who is mentioned in the piece) who both want to ban torture but believe that it is sometimes morally justified, can’t bring himself to enforce the absolute ban with absolute strictness. Instead, courts should be allowed to exercise leniency. Yet McMahan appears to believe that the agent cannot be excused. Some punishment (a very light punishment?) must be imposed.

But why exactly? Why would you punish (even lightly, as in this famous German case) someone for engaging in a (by hypothesis) morally permissible or even morally obligatory act of torture? The answer seems to be that if you don’t, then other agents will engage in torture that is morally wrong. I confess I have never understood this argument. Doesn’t the logic of it suggest that we should prohibit the police from morally justified shootings because if we don’t, police will engage in morally unjustified shootings?

Or taken from the other side, if leniency is permitted, why shouldn’t we worry that the prospect of leniency will encourage agents to engage in wrongful acts of torture? After all, if non-punishment of morally justified torture will encourage wrongful torture, as McMahan claims, why wouldn’t lenient punishment of morally justified torture also encourage wrongful torture? The effort to split the difference by banning torture but providing for leniency seems arbitrary.

Let’s go back to police shootings. Possibly, torture can be distinguished. One distinction is that police shootings are usually (almost always?) morally justified, whereas acts of torture are almost never morally justified. But do we actually know this? The reason that police shootings are usually morally justified, or seem to be, is that the police are given training, and police shootings are always investigated carefully. There is no comparable institutional infrastructure for torture. Maybe if there were, acts of torture would seem as morally justified as police shootings (although no doubt much rarer).

Arguments like McMahan’s, which are scattered throughout the philosophy literature, always seem to be based on psychological (about how people respond to incentives) and institutional assumptions (about how organizations operate) that are not articulated.

At Slate. This is based on the paper I’ve written with Cass Sunstein. The Slate piece discusses some surveys that we did (the paper provides more detail). If you have any comments on them (or the paper), please email me. We could do some more surveys if you have ideas.

People constantly accuse politicians, judges, and commentators of flip-flopping on institutional issues. Republicans who objected to filibusters of Bush’s nominees now defend the practice as applied to Obama’s–and the Democrats who defended filibustering then attack it now. Most of the liberal commentators who accused Bush of abusing executive power have now fallen silent, while the earlier Republican defenders of Bush have now, under Obama, discovered the dangers of the imperial presidency. Justices who appeal to the majesty of democratic rule in the course of upholding a statute today turn around and strike down a statute despite majority support for it tomorrow. And so on.

Many flip-flops reflect meaningless political posturing, but so do many of the accusations of flip-flopping. An apparent flip-flop can turn out to be nothing of the sort once one pierces the often sloppy rhetoric. Perhaps real flip-flops can be justified as the result of learning, at least to a limited extent. But beneath the surface, there is much of interest. Flip-flopping can result from an ambiguous or unsettled institutional norm. People are not just posturing but trying to get the norm settled in a way that advances their interests.

Much more can be said, and is said, in a new paper that I have written with Cass Sunstein, available at SSRN.

Terrorist attacks generate a familiar pattern in public debate. First, conservatives (and often middle-of-the-road types) argue that the government’s failure to stop the terrorist attack shows that counterterrorism policy is too weak. Then, liberals (and often other middle-of-the-road types) argue that we should not strengthen counterterrorism measures if doing so will sacrifice our civil liberties to security. This sets up a debate about security versus liberty. Typically, civil libertarians argue that there really is no tradeoff (an argument I have never understood), or (more plausibly but I think wrongly) that the government will inevitably put too much weight on security and not enough on liberty. An important subtheme, one that resonates with American historical experience and mythology, is that the people who put more weight on security are cowards who sell our liberties too cheaply.

Thus, the rhetoric. In truth, there is liberty on both sides of the equation. People who fear terrorist attacks lose some of their liberty as they avoid airplanes and public places; and the people who die in those attacks lose their liberty along with their lives. Nonetheless, it is undeniable that the civil libertarian position is understood to place greater weight on due process than on security, and that position has very powerful resonance in our society, perhaps because of distrust of the government.

The Charlie Hebdo attack has not followed this pattern for an interesting reason. The attack, both by design and in effect, was targeted at a liberty–freedom of expression. In this respect, the attack is unique among all the terrorist attacks since 9/11, none of which singled out freedom of expression as a target among all the western vices. The planned French crackdown on civil liberties thus sets up a clearer, harder-to-deny, liberty-liberty tradeoff: liberty from surveillance, arbitrary detention, and the like, versus liberty to speak one’s mind. It’s harder for a civil libertarian to argue that “mere” security is at stake, that principled people must oppose stricter counterterrorism measures.

This tradeoff has not yet received much attention, though it is implicit in the debate about whether Charlie Hebdo’s speech was really worth defending. Civil libertarians should ask themselves: if greater censorship in France made the French safer, with the result that they don’t need to give police greater surveillance and detention powers, would they be better off or worse off?

This is the most important policy question that has emerged from the attack. Why has no one asked it?

I make the case in Slate, which is that the market for short-term, on-demand car rides is inherently monopolistic. That is in fact why taxi regulation exists, and always has, virtually everywhere. The Slate piece arose from some initial thoughts in this blog post, further stimulated by Ilya Somin’s criticisms of that post. One point of disagreement centers around how to interpret people who consent to and then complain about surge pricing. Somin thinks they are irrational. I think they are reasonably concerned that they are being overcharged. The underlying problem is the high cost of search in this market, as explained in the Gallick & Sisk JLEO paper I cite. There is an interesting sense in which Uber’s disruption of the taxi market replays an earlier disruption in the 1920s when mass-produced automobiles threatened to unravel taxi pricing with the introduction of part-time drivers who skimmed off the best fares.

Everyone says it is. Here, for example, is a representative statement from Ilya Somin. The argument is just that unregulated markets are efficient and therefore price caps are inefficient. Somin concludes that everyone opposed to surge pricing is irrational or ignorant.

I don’t oppose surge pricing but I don’t think Somin is right.

1. Although a practice may be efficient, it doesn’t follow that everyone is made better off by it. People rationally oppose surge pricing as long as they value the dollar savings resulting from a price cap more than the extra time they spend waiting for an Uber car or taxi to show up. These people oppose an efficient practice that happens to harm them. What’s so irrational about that? In fact, the contrary view would be irrational.

2. When surge pricing comes into effect, there is an undersupply of cars. This means that Uber has market power. Taxis can’t raise their prices, and Lyft apparently won’t match Uber’s price above a threshold. This means that for the class of people who will pay the surge price, there is no meaningful competition. It doesn’t take much imagination to believe that Uber–which takes infantile pride in its ruthlessness–will charge a price that eats up as much of the consumer surplus as possible, in the process pricing some passengers out of the market.

This is not an argument for banning surge pricing or even regulating it. It’s possible that Uber’s monopoly profits will bring in additional competition, or that the threat of additional competition keeps Uber in line, and that’s what we usually assume in antitrust law, so we let the Ubers of the world charge supracompetitive prices except in egregious circumstances. But people’s efforts to shame Uber into lowering its surge prices may not only be in their self-interest; they may serve social welfare as well.

3. Everyone thinks of Uber as an app that allows drivers and passengers to match. But also think of Uber as an efficient way of cartelizing drivers and obtaining and analyzing the data of passengers so as to maximize revenues. (Can Uber determine your price sensitivity based on past trips? Probably. Does it exploit that information by varying prices by passenger type? Maybe not. Yet.) True, Uber is a better cartel than the taxis. True also, many Uber drivers did not offer services before Uber organized them. But not everything Uber does is by definition a good thing.

In the spirit of the season, I speculate in Slate about the stagnation of democracy. I have already received an email from someone berating me for omitting Barack Obama from my list of dictators. “Your agenda shows,” she warned.

As I noted a few days ago, Christopher Kutz argues that the anti-torture norm is (or might be) dead. Another way of putting this claim is that the longstanding taboo against torture has lapsed. A practice is taboo if not only it is forbidden but open debate about it is forbidden. Anyone who challenges the taboo will be regarded as tainted or contaminated, as outside the community. Contrary to what we like to think, hundreds of taboos flourish in American society, as many I’m sure as in any of the tribal societies studied by early anthropologists from which the term was adopted. Our taboos surround not only religion, but also race, gender relations, and the treatment of children. Free speech is firmly entrenched in the law but anyone who thinks that one can speak freely about these topics without risking significant social sanctions hasn’t been paying attention. Taboos constantly change (many sexual taboos have lapsed, just in the last few decades), but while they prevail they are extremely powerful.

The process by which taboos break down is mysterious; Kutz doesn’t really explain why the torture taboo has eroded if it has. At least part of the explanation must lie with technological change that causes people to question traditional prohibitions. The invention of modern forms of birth control made many of the taboos surrounding sex, which may at some earlier time have been broadly functional (in the sense of protecting people from the burdens of unwanted children or quelling social conflict), seem nonsensical. Yet the erosion of those taboos (not yet complete) was complicated. People had to be motivated to challenge the taboos and endure social sanctions. Sexual desire is a potent motivation, and eventually the arguments could not be ignored. But if there is no strong incentive to challenge taboos–as may be the case with taboos that don’t ban behavioral anyone really wants to engage in (like cannibalism)–then they are likely to persist.

If the torture taboo is eroding, then the explanation must be different from change in technology. The torture technologies used by the CIA are decades, even hundreds of years, old. And as is common with many taboos, the prohibition was never complete–U.S. government has committed torture before (just as incest takes place despite the incest taboo); what’s new is that torture is openly discussed as a legitimate policy option, by some people. In the 1990s and earlier, the U.S. engaged in torture through proxies, and no one talked about torture used by American combat soldiers in wartime. What seems to have happened is that an unusual configuration of events–the 9/11 attack, the earlier enactment of torture laws that forced the CIA to seek legal cover through a Justice Department opinion, relatively new norms of government openness, and so on–forced torture out into the open, where it could no longer be ignored.

You can see the persistent taboo-like character of torture in the debates surrounding the CIA’s interrogation practices. Many of the critics feel compelled both to argue against torture (“it doesn’t work,” “it violates our values”), and to argue that this argument is unnecessary because torture is plainly wrong or off the table (“it’s not who we are”). But the mere making of the first argument, which often requires elaborate claims about how institutions work, contradicts the second. Torture (unlike, say, cannibalism or incest) then becomes a matter of debate, perhaps like any other policy. The real force of the much-derided ticking time-bomb hypothetical is not that it provides a policy justification for institutionalized torture, but that it explodes the taboo. If you agree that torture may be acceptable in this setting, then you can argue against its expansion to less extreme scenarios only by making complicated empirical and institutional arguments that can be debated by people who have different intuitions.

I wonder whether the prosecutions that the CIA’s critics desire would have the perverse effect, even if they are successful, of further unraveling the taboo. In a court of law, defense lawyers will argue that their clients acted reasonably, and to do so, they will elicit testimony that some interrogation practices that amount to torture are actually effective. Whether or not this testimony is persuasive, the mere fact that it is introduced and debated will help remove torture from the realm of the taboo. Like so many (actually nearly all) police practices, there is just no reliable evidence of efficacy, in one direction or the other, and in such cases courts tend to defer to the judgment of experts. Whatever the outcome of the prosecutions, the efficacy of torture becomes merely an empirical question, deserving of further study perhaps, one about which reasonable people may differ–in which case it can’t be taboo.

This is, I think, what happened in the gay marriage cases, which helped destroy another taboo that until recently was extremely powerful. The importance of the evidence introduced in those cases was not so much that it supported the case for same-sex marriage but that it showed the question of same-sex marriage is an empirical one. Once empirical doubts are recognized, they cannot displace powerful equality norms.

From Adrian Vermeule:

Thanks to Will Baude for his thoughts on our paper project (see here and here for our puzzles and conjectures). It’s interesting that the proposal for judges to take into account the votes of other judges provokes a kind of instinctive resistance. But it’s not clear what exactly the objection is. Some possibilities:

(1) Will seems implicitly to assume that “textualists” and “purposivists” inhabit different methodological universes, so that judges in one camp would obtain no information from considering the views of judges in the other. That’s not how interpretation works, however. Purposivist judges are certainly interested in text and canons, in part because those things supply evidence of the purposes that a reasonable legislator might have. Conversely, many textualist judges, like Holmes, have been willing to examine legislative history and other extra-textual sources as evidence that might shed light on the ordinary meaning of text.

But even when textualist judges eschew legislative history altogether, that does not mean there is no overlap between their approach to interpretation and that of purposivist judges. Schematically, it is not the case that textualist judges consider sources or arguments {A, B, C} while purposivist judges consider sources or arguments {D, E, F}. Rather closer to the truth is a schema in which textualists consider {A, B, C} while purposivists consider {B, C, D}, or even {A, B, C, D}. This implies that judges in both camps will often gain relevant information — relevant even on their own theories — from observing the votes of other judges, even judges in other camps. And, of course, most judges are not theoretical at all, and just consider all sources and arguments in a sort of promiscuous jumble.

(2) Will also seems to think it important that judges in each camp think their own theory “correct” (Will’s italics). Under the Chevron framework, however, even if I think I am correct, the question I have to answer is whether I think the other person’s view is not only wrong, all things considered, but is actually unreasonable. The whole point of Chevron is to create space for that distinction. It is a symptom of a pre-Chevron mind (sub-Chevron mind?) to conflate these two questions, assuming that if my view is correct, yours must be beyond the pale. There is an interesting, under-explored question whether Chevron implies that agencies should have a kind of meta-discretion to choose among reasonable theories of interpretation. But the fact that proponents of competing views think their views correct will not help us figure that out.

(3) Yet another separate question, which we flagged in our opening posts, is whether and under what conditions it is systemically desirable for a given judge to take any of this information into account. We think that is the critical question for the paper, which will attempt to sift out the conditions under which it is or is not desirable. Will points out that sometimes it is better for decisionmakers not to attempt to consider all available information; certainly that is true. But he seems to assume that throwing away this particular category of information is necessarily desirable in all settings. His confidence in that approach seems to outrun the available evidence and theory, as far as we can see. It’s an interesting puzzle why our proposal provokes such a reaction.