Jul 27, 2013

Understanding the Zimmerman verdict

An awful lot has been written and said about the Zimmerman verdict since my last post. Some of that has added to our understanding of Florida law, racial bias, and other factors that may have played a role, and some of it has muddled the issues. I've tried to understand what happened as best I could, and in this post I'll try to sort out what the best arguments are, what we know, and what we don't know.

Let's start where I left off. I mentioned that 2006 and 2011 court cases changed the burden of proof in self-defense cases in Florida, and that they made Florida unusual because the prosecution now has to prove beyond reasonable doubt that the defendant did not act in self-defense. This statement has been challenged. Eugene Volokh claims that all states except Ohio have the same rules for burden of proof as Florida. That is incorrect according to a criminal defense lawyer from Wisconsin and to lawyers.com, but Florida may not be that unusual, either:
The burden of going forward with a case varies in different jurisdictions. For example, in New York, the defendant has to prove an affirmative defense by a preponderance of the evidence. Compare this with Massachusetts where, after the defendant has satisfactorily raised an affirmative defense, the prosecution must disprove it beyond a reasonable doubt. 
The source for Volokh's claim seems to be a self-help book for those who plan to use the self-defense defense (which is, conceptually, rather creepy and oxymoronic), which does not give serious thought to the burden of proof:
This is of more concern to your lawyer than it is to you, so you won't see it discussed extensively in this book. But we mention it because it's something that you will eventually wonder about while studying the subject of self-defense. In all states except Ohio, the state has the burden of proving beyond a reasonable doubt that the defendant did not act in self-defense.
It is interesting—and troubling—that it is hard to find this information about state laws. But perhaps it doesn't matter so much what the rules for burden of proof are. We hear a lot of strange stories about jury verdicts in Florida, and I don't recall any similar stories from Massachusetts, which seems to have similar rules. After all, juries don't learn the rules by studying law, but from the instructions that the judge gives them. The details of jury instructions matter a lot.

Speaking of the jury instructions, many people have claimed that "stand your ground" (SYG) did play a major role after all: although the defense didn't make it a SYG case, the wording of the jury instructions was crucially influenced by SYG. This does not seem persuasive to me. The defense cited a pre-SYG court case during the negotiations about the jury instructions, and that seemed to influence the judge (and thus the instructions, and thus the jury) in an important way. Jonathan Turley explains in more detail why SYG had little or nothing to do with the verdict. What I find interesting, however, is the "lesser crime" part of the instructions, where the jury is instructed what it would need to convict Zimmerman of manslaughter.
To prove the crime of Manslaughter, the State must prove the following two elements beyond a reasonable doubt:
  1. Trayvon Martin is dead.
  2. George Zimmerman intentionally committed an act or acts that caused the death of Trayvon Martin.
George Zimmerman cannot be guilty of manslaughter by committing a merely negligent act or if the killing was either justifiable or excusable homicide
(Emphasis mine.) But this is how the relevant Florida statute defines manslaughter:

(1) The killing of a human being by the act, procurement, or culpable negligence of another, without lawful justification according to the provisions of chapter 776 and in cases in which such killing shall not be excusable homicide or murder, according to the provisions of this chapter, is manslaughter, a felony of the second degree, punishable as provided in s. 775.082, s. 775.083, or s. 775.084.
Note the absence of "culpable negligence" in the jury instructions and the explicit instruction that finding of "mere negligence" is not sufficient for a manslaughter conviction. This, in my opinion, is the main reason Zimmerman is free.

There is a well-documented history of revisions to the "culpable negligence" instructions in Florida case law, but I see nothing in that report about omitting such instructions completely. Once negligence was excluded as a possible basis for manslaughter, it is not surprising that the jury couldn't reconcile what they thought Zimmerman did with the elements of the crime as explained in the instructions, especially given the evidence actually shown in the trial. On the other hand, based on the infamous Juror B37's spilling of the beans and the later counter-spilling by Juror B29, it seems that several jurors at least believed Zimmerman had acted recklessly and may well have found him "culpably negligent" if that possibility had been presented to them.

There were several factors that contributed to the verdict: Florida laws are crazy, the prosecution overreached (murder was pretty much impossible to prove under Florida law) and then handled the case poorly (from jury selection to introducing the self-defense claim to ineptly selecting evidence to introduce), but most importantly, the judge's instructions to the jury favored the defendant by way of two significant omissions: they did not mention initiating the confrontation as an exception to self-defense or culpable negligence as a basis for manslaughter.

The possibility that the jury instructions are responsible for the verdict reveals a flaw in the US legal system. In the US, it is (as far as I know) not possible for the state to appeal a "not guilty" verdict. Although the defendant can appeal a guilty verdict on grounds that the judge's instructions to the jury were prejudicial against him or her, a symmetric appeal is not available to the prosecution. I understand and agree that asymmetry is necessary in criminal trials since the stakes are fundamentally asymmetric. The burden of proof is one area where asymmetry is obviously necessary. Asymmetry is also obviously warranted with regard to appeals based on ineffective counsel: it would be ridiculous for the state to appeal because its prosecutor did a lousy job. But it is by no means clear that the judge's conduct of the trial is an area that calls for a one-sided availability of appeal.

Jul 14, 2013

Zimmerman, self-defense, Florida law, and negligent pundits

First, the obvious and sensible: If an armed person (a civilian, not a police officer in the course of duty) pursues another and ends up killing them, that's at least manslaughter. The pursuit is the initiation of any conflict that may arise, so it is against every logic and moral intuition that the pursuer could be entitled to a claim of self-defense.

We are told, however, that there is something weird about the law in Florida regarding self-defense, and that the jury acted properly within that strange law when it acquitted George Zimmerman. But no article I've seen so far has stated properly what it is that is strange about Florida.

You hear a lot of people, mostly casual observers, blaming Florida's "stand your ground" (SYG) law. SYG is a sick and warped self-defense doctrine that removes the duty to retreat, which is normally part of the concept of self-defense. But SYG was not invoked in the Zimmerman trial. It's a bad law, but irrelevant for this case.

More careful pundits get closer to the real issue, but still miss the mark. They tell us the prosecution failed to prove beyond reasonable doubt that Zimmerman did not act in self-defense. They tell us, correctly, that it has to do with the peculiar institution that is Florida law. Here is Cohen:
But this curious result says as much about Florida's judicial and legislative sensibilities as it does about Zimmerman's conduct that night. This verdict would not have occurred in every state. It might not even have occurred in any other state. But it occurred here, a tragic confluence that leaves a young man's untimely death unrequited under state law. Don't like it? Lobby to change Florida's laws.
And here is Coates:
6.) I think the message of this episode is unfortunate. By Florida law, in any violent confrontation ending in a disputed act of lethal self-defense, without eye-witnesses, the advantage goes to the living.
That all rings true, but the explanations do not add up. Self-defense is an affirmative defense, which means it is not incumbent on the prosecution to prove beyond reasonable doubt that the defendant did not act in self-defense. The burden of proof, albeit with a lower standard (generally preponderance of the evidence, or "clear and convincing", depending on the jurisdiction), is on the defendant.

The fact that Zimmerman shot Martin was never disputed, and Zimmerman's 911 call is clear evidence that Zimmerman started pursuing Martin without Martin doing anything to him (or even being aware of his existence). I don't know about murder, but the prima facie case for manslaughter (which the judge instructed the jury to consider as an alternative verdict) is complete right there. The prosecution has no burden, under standard rules, to prove beyond a reasonable doubt that self-defense does not apply.

But, it turns out, the rules in Florida—or some part of it—are anything but standard, and that that goes back to a 2006 court case, Murray v. State:
Six years ago in Murray v. State, 937 So.2d 277, 279 (Fla. 4th Dist. 2006), the Fourth District Court of Appeal in Florida ruled that once a defendant in a criminal case has introduced proof that he acted in self-defense the jury is entitled to consider the defense, and the jury may not convict the defendant unless it finds beyond a reasonable that he did not act in self-defense.
(Bolded emphasis mine.) And now what makes this relevant to the Zimmerman trial:
Last year the Fifth District Court of Appeal quoted this language from  Murray and followed the same rule in the case of  Montijo v. State, 61 So.3d 424 (Fla. 5th Dist, 2011). (...) Seminole County, where Trayvon Martin was killed, is in the  is in the Fifth Appellate District, (sic) so the rule in  Montijo is controlling unless and until the law is changed.
That's what's rotten in the state of Florida! The nature of self-defense as an affirmative defense has been turned on its head by activist judges. In effect, in Florida (or rather, its 4th and 5th Appellate Districts, but the article also points to a proposed law change that would affect the entire state), self-defense has been turned into an ordinary defense.

It would have been nice if the pundits who have written about the verdict did their homework and explained this key fact to us.

Mar 2, 2013

A Catholic in an Atheist's Body?

I really like Roger Ebert. He may not be the most insightful movie critic around (in fact, the most insightful reviews are often written by amateurs because they have the luxury of writing only when they really have something to say), but he has a consistency of taste that usually allows me to infer from his review whether I would like the movie, even when I predict I would disagree with him. He is also a prolific humanist writer: both his reviews and other writing generally promotes humanist values and a sense of decency. Besides, I admire how he has kept his vigorous productivity despite his illness and disfigurement.

But I really don't know what to think about his recent blog post, "How I am a Roman Catholic". I think he is terribly confused, and his persuasive writing spreads the confusion to his readers.

The centerpiece of confusion is the following statement:
I consider myself Catholic, lock, stock and barrel, with this technical loophole: I cannot believe in God. I refuse to call myself a atheist however, because that indicates too great a certainty about the unknowable.
Sorry, but this makes no sense. If you don't believe in God (or gods), you are an atheist. That is by definition. "I don't believe in God" and "I am an atheist" are factual statements with exactly the same meaning. You cannot claim that one is true and the other false. The "certainty" argument is a fallacy of which, I am sure, Ebert is well aware. (And yes, I am aware that he said "I cannot" rather than "I don't" believe in God. But the former implies the latter.)

A commenter suggested that he should perhaps call himself a "secular Catholic" or "cultural Catholic" like some people call themselves secular Jews or cultural Jews. I agree, but I also know people who call themselves Jewish atheists and even some Catholic atheists, and that makes complete sense because they acknowledge their cultural background and the default religious milieu in reference to which they developed their atheistic views. There are various ways in which one can embrace one's cultural background (which often includes religion) as well as one's unbelief in the supernatural.

That said, I think Ebert's article is worth reading, precisely because it honestly lays bare the mind of a religiously confused modern, liberal, humanistic person. His confusion is, I suspect, fairly typical among educated people in today's Western civilization. Many people I like and respect are probably similar to Ebert in this respect, and I can also use an occasional reminder to examine what I believe and why. Even the most comfortable atheists among us developed amidst all kinds of religious influences.

But, let anyone speak honestly for long enough, and they'll end up saying something really troubling. Here is where Ebert's confusion becomes harmful:
Birth control? Here I subscribe to an unofficial "double" loophole often applied in practice by Catholics faced with perplexing choices: Do that which results in the greater good and the lesser evil. I support freedom of choice. My choice is to not support abortion, except in cases of a clear-cut choice between the lives of the mother and child. A child conceived through incest or rape is innocent and deserves the right to be born.
Be careful not to step on the shards of the broken logic here, it could be really painful. I didn't know Ebert had such strongly Roman Catholic views on abortion, and I am surprised, given his generally liberal positions. And he seems to try to reconcile it with his liberalism. His language about choice, though unclear, suggests that he is speaking of his personal belief, but that he does not want to impose it on others. However, that makes sense only when a woman says it. A woman can hold the position "I would not choose an abortion, but others can make the choice for themselves." For a man to say that is obviously ridiculous. A man cannot consistently have a personal view on abortion distinct from a public policy position. Sure, a man can be a hypocrite and choose a laxer standard for his family or people of his social status than for the rest of the society (Remember Dan Quayle's "I would trust my daughter's choice"?), but it simply makes no sense for a man to have a personal opposition to abortion that he is not imposing on others.

So the confusion is not just about semantics. It is also about ethics and liberty. As religious confusion often turns out to be.

Feb 23, 2013

Is Happiness a Warm Gun?

This is the text (slightly edited and with added links) of the talk on gun violence and gun control I gave at the Northern Virginia Ethical Society on February 10, 2013.

On Jan 8, 2011, US Representative Gabrielle Giffords was holding a constituent meeting at a supermarket in a suburb of Tucson, when Jared Laughner drew a 9mm Glock pistol with a high-capacity magazine, shot Giffords in the head, and proceeded to shoot at other people gathered. Six people died, including a federal judge and a 9-year-old girl, and 13 more were injured.

Fast-forward a year and a half. On July 20, 2012, James Holmes shot 70 people, 12 of whom fatally, in a movie theater in Aurora, CO. He used several firearms, which he had purchased within two months before the shooting.

Fast-forward a few more months. On Dec 14, 2012, Adam Lanza shot his gun-enthusiast mother to death in their home in Newtown, CT, then, using her semi-automatic rifle, shot his way into Sandy Hook elementary school, where he killed 20 first-grade students and 6 teachers and other employees before shooting himself when the police arrived.

Besides Aurora and Newtown, last year there were 5 more shootings in public places with 4 or more victims, which is the FBI criterion for mass killing. 70 people died in these incidents, including 5 of the 7 shooters, and 68 more were injured. This makes 2012 the year with the most mass shootings, most fatalities from mass shootings, and most total victims from mass shootings, in recent US history. The Newtown massacre was particularly shocking since most of the victims were 6 or 7 years old. Suddenly, the issues of gun violence and gun control, which had been on a political back burner for some years, leaped into the forefront of the public debate.

Mass shootings are scary. Imagine facing a madman armed with a rapid-fire weapon (or several of those), intent on killing as many people as possible. And anyone can be a victim, anytime. Keeping away from dangerous places or activities doesn't help: mass shootings have occurred in schools, workplaces, places of worship, shopping malls, theaters.

And yet, those incidents are extremely rare. Even in an outlier year like 2012, well under 1% of firearm homicides in the US resulted from mass shootings, and less than 1 in 4 million Americans died in those events. Averaged over the last 15 years—a period that includes Columbine and Virginia Tech and Fort Hood and Aurora and Newtown—the average death toll from mass shootings was 22 per year. About five times more people are killed by lightning. And on an average *day*, more than 30 people are killed with guns (and about twice as many kill themselves).

School shootings are the subset of gun violence that scares people the most. Columbine was shocking, but it was still teenagers killing teenagers, something not completely unfathomable. But Newtown, where an adult was killing children, triggered our deepest fears, much like the similar massacre in Dunblane, Scotland, in 1996. Suddenly people on both sides of the gun control issue feel there is a crisis and we have to do something about it, and the proposed things to do range from reasonable (like better background checks and banning high-capacity magazines) to dubious (like putting armed guards in every school) to bizarre (like arming teachers) to pathologically insane (like training children to rush a shooter).

But one tragedy doesn't make a crisis. Schools are safe. Very safe, at least as far as lethal violence is concerned. About as many high-school students die playing football as are killed in school, whether in a mass shooting or in an individually targeted homicide (including non-gun-related ones). More elementary-school students die riding their school bus—and even more as pedestrians hit by a school bus—than are killed by guns or other violent means. Schools have also gotten safer over time. In the 1992-93 school year, 34 K-12 students were killed in school; in the last decade, the annual average has been about half of that. And that's out of about 55 million students.

Rep. Donna Edwards said on ABC's "This Week" that “since Columbine, there have been 181 of these school shootings.” This information comes from a fact sheet on the Brady Campaign web site. FactCheck.org called it an "exaggeration". I often criticize fact-checking outlets for nitpicking liberals' statements to appear fair and balanced, but this time I fully agree with the characterization. Most school shootings on the list actually happened outside school, typically either after school or after a sports game. Most were non-fatal. Some were suicides. Some were teachers shooting principals. Also, the list includes shootings at colleges and universities. Yes, those are also schools, but a college campus is much more open than a high school, and "on-campus" events cover 24 hours a day. When I was a student at the U of C, somebody unconnected with the university was killed at the gas station next to the building where I lived. I think that was part of the campus, so that would probably be included on the list if it went back to 1990 or so.

So, schools are safe, but, as I said, more than 30 Americans are killed with guns on an average day—about 11,000 people every year.

There are good and bad news about our murder rate. The good news is that it has been declining. It is about one-half what it was at its peak in 1980s and early 90s, and about as low as it has been at any time in the last 50 years. The bad news is that it is still much higher than in any other highly developed country. Our murder rate is 4 times higher than the OECD average, and no other rich country has even half the US murder rate. The disparity in gun murders is even greater. Switzerland, in second place, has 1/4 the US gun murder rate, and some countries, like Japan, hardly have any gun murders.

Why is the US murder rate so high? Opinions differ, but relative to other developed countries, the US has more poverty, more organized crime, a culture generally supportive of violence, and, most significantly, much more firearms.

Poverty increases the incentive to engage in crime in general. Aggressive policy of prohibition, like the so-called War on Drugs, generates organized crime activity. It is hard to deny that the US has a culture of violence, from the kinds of sports popular here to the fact that killing someone was the most popular thing our President did in his first term. But nothing makes us as different from other countries as the extremely high rate of gun ownership. There are about 300 million firearms in the US, almost one per person. No country comes even close.

What you typically hear from NRA or libertarians is the cliché that "guns don't kill people, people kill people." But, in the real world, it is very easy to kill someone with a gun and quite difficult without. So the argument that someone who wants to kill would find other means if he didn't have a gun is complete hogwash. Also, the vast majority of homicides are not deeply premeditated and planned, but typically occur in gang wars, during robberies and other crimes, or in bursts of anger. If available weapons were limited to knives, baseball bats, and fists, fatalities would obviously be far lower.

International comparisons support this very strongly. Countries culturally most similar to the US - Canada, UK, Australia, and New Zealand - all have homicide rates between 1/4 and 1/3 of the US rate. But if you look only at non-gun homicides, they are all at about 2/3 of the US rate. So, without guns, we still have more homicides, but not egregiously more. But gun murders increase predictably with gun ownership rates: Australia has about twice the rate of the UK, Canada 3-4 times the Australian rate, and US 5-6 times the Canadian rate.

Somehow, Switzerland has become a fashionable example among gun advocates. They argue that Switzerland has even more guns than the US—their government mandates that every household have a gun!—yet their murder rate is among the lowest in the world. So there, that kills all your gun control arguments! Except that everything in their story is false. Switzerland indeed has a relatively high rate of firearm ownership, but it is still only about half the US rate. Because of their traditional concept of national defense, almost all adult men are considered military reservists and typically keep their military guns at home. Not surprisingly, those guns are highly regulated. But it is also notable that Switzerland has the second-highest rate of gun homicides in the developed world, about 1/4 of the US rate. It is true that the total homicide rate in Switzerland is still low, but it is significantly higher than in the neighboring Austria, probably culturally the most similar country.

Another argument against gun control is that private gun ownership is the source of liberty. Beside the obvious fact that countries with very low gun ownership, like England or Israel, or even near-zero ownership like Japan, are hardly dictatorships, it is also notable that Saudi Arabia, one of the most despotic regimes, has a higher gun ownership rate than any liberal democracy other than US and Switzerland. On the other hand, the people of Tunisia recently successfully overthrew a dictatorship despite having virtually no privately-owned firearms. There is empirically no correlation at all between liberty and gun ownership.

The guns-liberty connection might be a benign myth if it existed only in the heads of Cato Institute scholars. Unfortunately, it may be one of the most dangerous pro-gun ideas around. More and more people seem to believe either that the US government will become tyrannical and they will need guns to defend themselves from tyranny, or that the government will collapse and they'll need guns to survive in the ensuing anarchy, communist takeover, or whatnot. This has allegedly been a major shift in the mainstream gun culture in the US. Until some 30 years ago, a typical NRA member was a hunter, while people stockpiling military-style weapons for apocalyptic scenarios were a fringe, perhaps living in isolated places in Idaho. But in the last 2-3 decades, the fringe has become mainstream. Such people can now be found among seemingly normal, middle-class, suburban types. Nancy Lanza, the Newtown shooter's mother and his first victim, was apparently one.

How those people believe they could fight tyranny with their guns is hard to understand. The US military has more power than the rest of the world combined. If somehow the US government became a tyrannical regime that would turn even a fraction of that power on its own people, what could some amateurs with guns do against it? Perhaps more seriously disturbing is that they even think it likely that the US could turn into a tyranny against which taking up arms would be justified. Actions those people contemplate would be considered treason or terrorism were they ever to take them. But this is an increasingly influential strain of gun advocates, and apparently the dominant group in the NRA leadership.

Considerably more rational than owning guns for defense against a tyrannical government is owning them for protection against criminals. Nevertheless, it is mainly based on myths and delusions. Statistically, a gun kept at home is by far the most likely to kill or injure its owner or a member of the owner's family. Killing in self-defense is in fact exceedingly rare. In a year, less than 100 civilians kill defending themselves or someone else from an attack, and about 60 of those self-defense killings are with guns. Killing a spouse is over 10 times more common, as are deaths from firearm accidents. People even murder their children and parents more often than they kill in self-defense. Not to mention that people overall shoot themselves almost twice as often as anyone else.

Of course, the most relevant statistic is not how many times people kill in self-defense, but how many crimes are averted by confronting the criminal with a gun. The protective function is served most successfully if no one is shot. Unfortunately, such data aren't available, partly because pro-gun activists successfully lobbied Congress to forbid CDC to conduct any research in gun violence. But it is hard to imagine that there are so many cases of successful (and justified) protection by merely brandishing a gun to outweigh the risks of gun ownership for most people.

That said, not everybody is "most people". When Sam Harris wrote that he owns a gun for self-protection and that he regularly practices shooting, some commentators ridiculed him for living in an NRA fantasy world. However, Sam Harris is a public figure whom many people—particularly many fanatics—hate, and who regularly receives death threats. If I were Sam Harris, I would probably own a gun. At least, I would seriously consider it. And if someone owns a gun, the responsible thing is to regularly practice using it, as a gun in unskilled hands is a danger in a 360 degree angle. Although I suspect that Harris has a bit too much confidence in his skills, judgment, and safety precautions, criticizing his decisions based on general-population statistics is about as ludicrous as the NRA CEO Wayne LaPierre's insistence that the President's children are exposed to the same risks as anyone else and need the same level of protection.

As the wisdom of keeping guns for self-protection needs to be questioned, so does the ethics of it. When is it OK to use a gun? Many actions that pass as self-defense in the US would be illegal in other countries. The majority of Americans seem to think it is OK to shoot an intruder in the house. But is it? Is it morally permissible to shoot a burglar? Suppose breaks into your house to steal a TV, thinking  nobody is home. Say the TV is worth $1,000. Almost no one would say it is morally permissible to kill somebody to gain $1,000; why then would it be OK to do it to prevent a loss of $1,000? The argument obviously extends to any amount. Is it permissible to threaten to shoot? I would certainly question the wisdom of making such threats without at least some willingness to follow through.

A pro-gun argument would be that the mere awareness that many people have guns deters criminals and that it is one of the main reasons few home intrusions occur when families are at home. But the flip side is that the prospect of facing an armed homeowner is an incentive for a burglar to carry a gun. Additionally, guns are attractive targets for burglars, as there is a steady demand for them on the black market. Again, data are lacking, but I suspect that more guns are stolen from households than are used by those households in deterrence of crime.

Another self-defense myth is that guns are women's friends, as they level the playing field, enabling women to defend themselves from the naturally stronger men. Well, sorry ladies, but men kill their wives or girlfriends 4-5 times more often than women kill their husbands or boyfriends, and generally, about half of those murders, regardless of sex, are committed with a gun. As for self-defense, men kill in self-defense with a gun 10 times more often than women (and without a gun, just 3 times more). A woman somewhere in the US shoots someone to death in self-defense about once every 2 months. It's that rare.

We can debunk myths and present facts, but it's hard to make anyone change their mind. There do seem to be two tribes, the gun people and the non-gun people. In a recent article in The New Republic, Walter Kirn tried to explain how a gun person feels about guns and why. Josh Marshall, the publisher of Talking Points Memo, articulated the feelings of non-gun people. They seem to agree that growing up with guns has a lot to do with liking guns later in life. While the personal stories are interesting, I am not sure that generalizations are useful. Nevertheless, some people really like guns, and some are uncomfortable even at the sight of one. Any reasonable public debate needs to take this reality into account.

In this classification, I am a non-gun person, and I fit in the narrative as I generally did not grow up around guns. I shot a firearm on two occasions in my life. The last time when I was 15, as every 10th grader in the country where I grew up had to learn to shoot a military rifle. I later avoided military service, so that was also the end of my training. The time before that was when I was 5 or 6 and I pulled the trigger of my grandfather's hunting rifle that was stored in a rack in his home. It turned out to be loaded and it shot into the ceiling. Adults were in the other room and were scared to death until they saw me uninjured. Now my grandfather was a doctor and a very smart man, and he often told other hunters how important it was to keep guns safely… except he didn't quite follow his own advice. That taught me to be skeptical about everyone who says they are perfectly safe with their guns.

Acknowledging that there are gun people and non-gun people, what to do about guns? There seems to be a consensus that we must do something… but what?

It was recently reported that over 90% of Americans support universal background checks for gun sales. Of course this is commonsense: who would want to allow criminals and lunatics to buy guns. But I suspect the consensus will quickly melt away when a specific proposal is made. It is easy to support the general idea, but what exactly should be checked, and what should make someone ineligible to buy a gun? A conviction for a violent felony, I suppose, is a fairly non-controversial disqualifier. But what about someone who's been in jail for marijuana possession? That doesn't strike me as a valid reason to be considered dangerous and, no, "it would have flagged Jared Loughner" is not a good counter-argument. And what about psychiatric records? What conditions should make one ineligible to own a gun? In what circumstances should psychiatric records be available to the background check? There are privacy concerns. There are concerns about discouraging people from getting psychiatric help. It is complicated. And if we let every jurisdiction experiment with its own criteria, background checks will be easily gamed by buying guns where the standards are lax. For the system to be effective, it has to be uniform, one-size-fits-all, with all the inflexibility and unfairness that necessarily goes with it.

Mental illness is a particularly sensitive issue in this context. Most mass shooters were mentally ill or have shown signs of possible mental illness. But, through a common fallacy, that creates a public perception that mentally ill people are especially dangerous. However, mentally ill people are statistically no more likely to be violent than the general population. Of course, some mentally ill individuals are dangerous, and some mental disorders are more likely to make the person dangerous than others. But generalizations are dangerous and one has carefully weigh potential violation of privacy of millions just because a small number of people actually pose a danger.

There is no doubt that our mental health system is performing poorly, and too many people who need help are not getting it. Improving mental health would probably reduce the number of mass shootings, and a proper treatment of substance abuse would reduce other crime and homicides that go with it. But seeing a potential mass murderer in every person with a mental illness is not a constructive approach.

Banning so-called assault weapons is another popular proposal: a majority of Americans support it, although it has little chance of being enacted. We already had such a ban, but it expired. The rational is clear. A semi-automatic weapon with a large-capacity magazine enables killing a lot of people before having to reload. Nobody needs a weapon like that for hunting or self-protection, but that's what mass shooters often use. If perfectly enforced, such a ban would reduce the fatalities from mass shootings. But it would be hard to enforce and any such law with a realistic chance of passing would not touch some very dangerous weapons, such as the pistol used by Loughner. In fact, handguns, which kill by far the most people, would be very difficult to ban, especially after the recent Supreme Court decisions.

Putting an armed guard in every school is a bad NRA idea that the Obama administration appears somewhat willing to consider. Not only would it be expensive, but in every school there would be an armed, likely authoritarian-minded, person with nothing to do. That seems like the recipe for trouble. Besides, some schools, particularly in high-crime areas, already have armed police or guards. Given how safe the schools already are, it is hard to imagine any benefit from putting more guards in schools. Besides, would an armed guard have stopped Adam Lanza? I think the most likely outcome would be a dead guard.

Better enforcement of existing laws would help. Too many people pass existing background checks who clearly shouldn't, and too many gun dealers are willing to sell against the rules, or "lose" parts of their inventory. Guns are sold at gun shows with no background checks at all. Getting serious about enforcement and closing loopholes would be fairly uncontroversial.

To me, it would make sense to treat guns the way we treat cars. They should be registered and one shouldn't be able to sell them, lose them, or abandon them, without notifying authorities. And people should be required to pass safety courses before being licensed to handle a gun. I don't think this would infringe on anybody's right to own a gun, but the NRA keeps fighting against such requirements and I haven't heard of serious attempts to enact them at the federal level.

In the end, every proposal politicians are currently fighting about would at best chip away a few murders at the margin. But if we were really serious about reducing our homicide rate, we would first end the War on Drugs and switch to prevention and treatment of addiction. Then we would focus on reducing poverty. And finally, we would drastically reduce the number of handguns. But those things are difficult and unpopular, so I don't think we'll see any of them in the foreseeable future.

And, of course, we should work on making our culture less violent. But that goes beyond even difficult policy.