Friday, April 17, 2009


I’m a Second Amendment fan. It’s the First Amendment continued by other means. Every country inhabited by responsible adults should permit people to have weapons to defend themselves. Better safe than sorry.

Of course, there’s a downside. Some very bad people use guns to do some very bad things. They use guns like assault rifles. It’s hard to know what to do about that. It’s not enough to say “no one needs an assault rifle to hunt or defend his home against burglars.” The Second Amendment is not about deer hunting. It’s about killing people, including not only burglars but storm troopers. Yet, we seem all right with the ban on Tommy guns, not to mention SAMs and other stuff that a serious insurrection would require. The fact is that we can’t let very bad people have insurrection-class weaponry, so we all have to do without it.

That we allow any kind of weapon makes one think “We know what kind of girl you are – now we’re just talking price.” It’s a slippery slope, youbetcha. But there are lots of slippery slopes, and we don’t slide down all of them. We let the government set the highway speed limit at 55 without fear that they would lower it to 5, or 0.

The gun-control slippery slope is interesting, though: it’s more a sliding scale than a slippery slope. Because we don’t let bad guys own machine guns and RPG launchers, the assault weapon is the worst thing we face in quantity. But if enough criminals use assault weapons to commit enough crimes, we’re back where we were with the Tommy gun. It’s not that our standards of freedom are slipping down the slope but that the downside of allowing certain weapons is becoming intolerable where once it wasn’t.

Likewise, the non-dealer exception to the background check law is proving quite porous. Too many non-dealers are dealing, at gun shows primarily, which is why people call the exemption the “gun show loophole.” There seems to be some disagreement about whether that exemption is making a difference in the flow of weaponry to Mexico. Something tells me that if the Mexican cartels had to get their guns somewhere else, they’d get their guns somewhere else. It’s one thing to show that the exemption enables bad guys to get guns here; it’s another to show that ending the exemption will prevent bad guys from getting guns somewhere.

Wackos, on the other hand, are another story. The non-dealer exemption allows people who would fail a background check to evade a background check. That shift in who buys where makes the “insignificant” number of non-dealer sales, well, insignificant. What matters is what percentage of dangerous gun buyers buy through non-dealers. If that number is too high, the exemption has to go.

Thursday, April 9, 2009

The Prisoners' Dilemma and the Role of Government

You can read the metaphoric version of this game here. Simplified, these are the rules:

You and another guy are playing a game. You each have two buttons in front of you, a red one and a green one, and you each get to push one button once with no opportunity to know or influence the other's choice. The choices you each make will determine the pay-off.

  • If you both push Green, you each get $2
  • If you both push Red, you each lose $5
  • If you each push different buttons, the one who pushes Red gets $10 and the other loses $12.
What do you do?

First, I hope, you compare the possible outcomes for each choice you can make:
  • If you push Green, you get $2 if the other guy pushes Green or you lose $12 if he pushes Red.
  • If you push Red, you get $10 if the other guy pushes Green or you lose $5 if he pushes Red.
Note that your pay-off for pushing Red is better than the pay-off for pushing Green, whatever the other guy does. But note also that the combined pay-off for both players is either zero or negative unless you both push Green. So, your best strategy is to push Red, but the strategy that is best for “the team,” if you are a team, is for you both to push Green. But you are not a team under these rules.

So let's change the rules. Under the new rules, you don't play only once. You play repeatedly. Now there's something in it for you to cooperate by pressing Green together. You could reliably earn $2 a round. Yes, you could earn $10 by defecting, but if you do, trust is lost, and the opportunity to continue to earn those $2 bills ad infinitum goes away. That's a heavy price to pay for one quick score.

Well, zillions of pages have been written about this little game, because virtually all of civilization runs on it. The Prisoners' dilemma is a “plus-sum” game: if the players cooperate, they come away with a combined win. Those $4 victories make peace an asset and prosperity possible. Those victories are why, if we cooperate, our lives need not be solitary, poor, nasty, brutish, or short.

Of course, real life is messier than parlor games, but the simple game described above can be restructured to map some important real world situations. Suppose that if the players push different buttons, the guy who pushes Red gets $100,000,000, and the guy who pushes Green is killed (and so the game does not repeat), the women he loves are raped, and his children are sold into slavery. Not much point in pressing Green in that game. But now add that one player can arbitrarily force the other to press Green while the former pushes Red. We can call this game “Darfur.”

What if you don't know that the game is on or where the buttons are, or that someone has surreptitiously placed a green button where you will inadvertently press it as he lies in wait to press Red and abscond with his winnings. You may not even know your $12 is missing, or if you do, you don't know who has the $10 prize. We can call that game “Burglary.” Or maybe “Madoff.”

But there are plus-sum games to be played, where pressing Green means nothing more than not doing any of the things that we generally treat as crimes. It turns out that merely respecting the person and property of others is all it takes for the $2 bills to start rolling in, as people organize enterprises on the assumption that they will be able to bring them to production, sell their output, and keep their profits.

What we need is to be sufficiently confident that no one will play Darfur with us and that too few people will play Burglary or Madoff to matter. How do we do that? The answer, in a word, is government. Not necessarily “Leviathan,” but government – an organization with a monopoly on the legitimate use of force, with the mandate to use that force to assure, and only to assure, that external enemies do not play Darfur with us, and internal enemies do not play Burglary.
That's what governments are constituted to do. How else do we secure the blessings of liberty to ourselves and our posterity?

I am not arguing that national defense and law enforcement are the only activities in which government can engage. Public education, for example, creates a populace capable of earning those $2 bills and thus not tempted by some less salutary game. Public health makes the playing field safe for the players. These are just two examples of the things government can do to protect the standard iterated plus-sum game that we call civilization. Obviously, there are others, the merits of which are sometimes apparent and sometimes not.

So, when analyzing public policy, I suggest that the question to ask is simply “Is this good for the game?” If it helps the game, it's a good idea; if it doesn't, it isn't.

Wednesday, April 1, 2009

The Fifth Axiom

You geometers know that what makes a plane a plane is Euclid's fifth axiom. The first four axioms describe elemental shapes that can work in more than one surface. But of those surfaces, only on a plane, through a point outside a line, there is one and only one line that can be drawn parallel to that line. Without that fifth axiom, you could be talking about who knows what kind of place. But with it, you're talking about something real, or at least something that maps to something real. With that axiom, you can build.

I'm an unabashed reductionist. I like to think that things follow from things. I know it gets dicey at the very beginning, but Euclid didn't ask why his fifth axiom was true, he just asserted it. In the non-material realm, what people build is called civilization. To build civilizations, we need a human geometry, and to be useful, the human geometry must have a fifth axiom – a single fact about humans that we must know to build a civilization. My vote goes to this: people have competing interests.

It would be nice to imagine how different the world would be if people could be assumed to be uninfluenced by self-interest in their dealings with others. But how could that be? Evolution favors the self-interested. We can define self-interest broadly enough to encompass the enlightened variety, in which we cooperate to prosper, and we can allow for altruism via Dawkins's selfish genes, but there are just too many situations in which narrowly defined self interest matters for us to have evolved without the ability to perceive and serve our own advantage, sometimes to the detriment of others.

The animal that is capable of self-interest recognizes that power is a good thing to have, because in power one can serve one's self interest. So it hardly surprises not only that people who attain power tend to serve themselves, but that the most self-consciously self-interested people actually scheme to get power, especially if power is unchecked. Consequently, the first sign of a successful civilization is that those with power are meaningfully prevented from using it to their own advantage, so that good people are not corrupted by high office, and corrupt people do not seek it.

This principle is obviously embodied in a polity's organic law. Much of The Federalist is devoted to diffusing power, to checks and balances, intended to keep high office from being too powerful or profitable. But the issue of self-interest, and especially corruptibility, is less obviously, but nevertheless always, present in public law, too, sometimes buried so deeply that the law doesn't seem to make sense. Not, that is, until one imagines implementing the law without the offending nod to corruptibility.

Take, for example, what some call “legal luck.” Two men drive drunk. One gets stopped by the police before doing any damage; the other kills a teenager. By what calculus does the former deserve to be treated any more leniently than the latter? Both indulged in precisely the same behavior. Both have the same measurable blood alcohol level. The only difference is that one did damage and one did not. And yet the law treats them differently. Why?

Legal luck is often discussed alongside “moral luck,” the phenomenon whereby the community judges the drunk driver who has an accident more harshly than the one who doesn't. It might seem that the punishment meted out to these two drivers is different because their moral status is different – that their legal luck is different for the same reason that their moral luck is different, indeed that their legal luck is a reflection of their moral luck. But I want to look through the other end of that telescope, adopting as my key principle the corruptibility of all involved.

Imagine a place where driving under the influence carried the same penalty as manslaughter. Whether one goes to jail for a long time on a first offense will turn on whether one gets stopped by the cops and on just what side of an arbitrary line a fallible machine says one's blood alcohol level falls. The damage that a policeman with an agenda could do in such a place is frightening. “Framing” is always an issue in law enforcement, but it is constrained to a large extent by a requirement of a corpus delicti – an actual victim whose very existence proves that a crime has been committed. With no visible victim, we have to take the policeman's word not only that the defendant did it, but also that it even occurred.

In discussing moral luck, some philosophers refer to the requirement of a bad result as a precondition to strong condemnation as an epistemic issue. The reason, they say, that we assess actors differently on the basis of result of their behaviors is that we have far greater confidence that their behavior was dangerous if it in fact resulted in damage. I'm with those philosophers, but there is a flip side to this coin. Ideally, we would either condemn a person for his behavior or we would excuse it, and the damage it actually caused would be irrelevant to that assessment. And the judgment we would pass for creating risk would likely be harsher than it currently is for creating risk without harm, but less harsh than it is for creating risk that ends in harm. So, let's imagine a world in which that rule applied. If people could be trusted to behave decently (the occasional creation of risk aside), this scheme would work out fine. But if we allow for human corruption, not only might people be wrongly accused of creating risk (an easier crime to prove than those requiring a corpus delicti), but people who intend harm could easily make their actions look like the (mildly punished) creation of risk that resulted in (irrelevant) damage. Indeed, it seems to me that this latter concern is reason enough for allowing legal luck to exist as the lesser of evils.

The fact that a bad result has occurred makes us more certain that the behavior involved was at least risky, but at the same time, it makes us more skeptical that the behavior was merely risky. We can't be sure, but as a prudent strategy for acting in response to conduct, those postures seem well-chosen. It makes sense, therefore, that our emotional equipment that internalizes this sort of strategy so that we don't have to rethink it every time the need arises, runs in the same way. Thus, we tend to react less harshly to risk without injury than to risk with injury. And this response makes sense if one understands moral judgment not as an intellectual exercise, but as a basis for decisions regarding responsive action.

Although moral luck and legal luck are often discussed in tandem, they are very different. Moral luck is about judgments; legal luck is about responsive action. For the two ideas to be analogous, moral luck would be considered in terms of how we should act toward individuals who have behaved badly. The difference is important to the moral luck question because much of the study of moral luck turns on things that people would usually ignore in deciding how to respond to behavior. Thomas Nagel, in an important analysis of the subject, suggested that luck plays a part in almost all actions susceptible to moral judgment. Who we are, how we are raised, where we find ourselves, are all to some extent beyond our control. Thus, the only difference between a Nazi Collaborator and a US citizen in 1938 might well be where the individual found himself. To someone trying to assess the “fairness” of judging the collaborator more harshly than the individual who would have been a collaborator if he were in a position to be one, such issues of “constitutive” or “circumstantial” luck may matter. But in deciding whether to shun the collaborator, his status relative to people not tested is irrelevant, as is his upbringing, socialization, genetic endowments, etc. The strategy we adopt to protect ourselves from relationships with bad people does not require us to be just; it requires us to be prudent. In other words, moral judgment for the purpose of deciding how to act is an exercise in self-interest! And it is prudent to treat an actual harm-doer more warily than we treat a merely alleged risk-causer.

Our legal system is full of prudential provisos. The corpus delicti is just one. The right of cross-examination – with a corollary ban on hearsay evidence - is another. But why do we insist that defendants be allowed to cross-examine their accusers? Is it just that accusers may misremember? Or is it that accusers might lie if they didn't have to be cross-examined, and might not even exist if they didn't have to be identified. Is it not because people are self-interested and, therefore, corruptible?

The point of this post is not some misanthropic rant. I don't mind that people are corruptible – it seems to me entirely necessary that they be so if we as a species are to survive. If we did not compete, the strongest would not survive, and if the strongest don't survive, the chances of the species surviving are meaningfully reduced. No, I'm not against corruption as a necessary frailty. I just think people lose sight of it when they think about how things might be – like thinking we would each contribute according to our abilities for payment in accordance with our needs. I expect I'll return to self-interest and its firstborn son, corruptibility, again. Their importance is, after all, axiomatic.