Bayes Theorem, Guns, and Mental Illness

Since the recent Navy Yard shooting, the Left has been talking about gun control, and the Right about mental illness. The Left's arguments are mostly specious or a dead end (specious: that an AR-15 was involved - one wasn't; dead end: that we should get rid of all guns - we won't).

But the Right's case isn't so great, either. Yes, we should have better mental health policies. And it's pretty clear that the Navy Yard shooting, at least, could have been prevented with more attention to mental health. Evidently the shooter had displayed clear signs of mental illness when speaking to police just days before the shooting. But the police hadn't called mental health officials. Had the done so it's possible the shooting could have been prevented.

In general, though, would better mental health policy have much effect on gun violence? Not really. The problem is that most people with mental illness aren't extraordinarily violent, and most violence isn't perpetrated by the mentally ill. So not only is it difficult - nearly impossible - to determine who is likely to commit gun violence, even if we could keep guns out of their hands the effect would be only marginal.

This is a problem with no easy, obvious fix.

Gun Control

Inevitably, the recent shooting in Aurora, Colorado, has rekindled the debate about gun control. I have my opinions on this, but they're not especially interesting or unique and I'm not going to detail them here.

What I do want to address is a particular argument that's been making the rounds on the Right: that no conceivable gun-control law would have stopped this shooting, and that putting in place gun sales bans now would do little good considering the 200 million guns already in private hands in the United States.

This logic is faulty, and the Right would see this if they applied similar arguments they have made in the past. Here's an argument from the other side: we shouldn't open up new areas for offshore drilling because it won't make a difference in the high price of oil now; it'll take ten years before that new production is online. The counter, of course, is that if we'd opened up those areas ten years ago, we'd have the production now. If something should be done, then actually doing it should not depend on whether it will have an immediate impact.

Or consider this one: there's no point to tighter immigration controls because we already have 10 million illegals (or whatever you favorite number). Again, the counter is that illegals move across the border both ways, so if we stopped the flow coming in, eventually their numbers would diminish.

The argument that gun control is pointless because of the large number of guns already extant has a similar counter. Suppose we banned all private gun sales tomorrow. This would still leave a huge number of guns in private hands. But those guns are already aging. Over time - it will be decades, because guns are durable - those guns will wear out, and their numbers will dwindle.

Whether this is the right thing to do is an entirely separate issue. But the argument itself is specious: if we want to control guns, the fact that it's going to take a long time shouldn't stand in our way.