So, Bill Bennett’s taken a bunch of heat for a recent radio appearance in which he objected to instrumentalist defenses of abortion by arguing that one could lower the crime rate by aborting every black child, and we’d all regard that as morally monstrous—and has also picked up some unlikely defenders. The defenders are mostly pointing out that Bennett was making a reductio argument, that he wasn’t actually suggesting we abort black babies wholesale. Well, yeah, obviously. That’s not really the problem with what he said.
Now, it’s certainly uncontested that, currently, a disproportionate number of crimes really are committed by African Americans. But to assume that crime would drop dramatically as a result of the kind of mass-abortions he was imagining, it seems to me that you have to posit that this isn’t a highly contingent fact having to do with a whole range of potentially alterable sociological and economic circumstances, but a kind of timeless truth. That is, you have to suppose not just that 2005 America is such that blacks commit more crimes (which is true), but that they’re intrinsically more likely to commit crimes—that this will hold true 20 years from now or 30 or whenever. Now, you might think that’s likely to be true just because you’re pessimistic about the prospects for the relevant social and economic conditions changing. But there’s at least a whiff in Bennett’s comments of the idea that there’s some kind of genetic predisposition to criminality. Maybe Bennett doesn’t think that, but if not, he displayed severe tone-deafness in picking an example that seems to imply it.
4 responses so far ↓
1 Caliban // Oct 3, 2005 at 1:13 pm
Wasn’t he using the “abort all black babies” comment as an example of the stupidity of generalizing statistics?
I don’t like Bennett much, but that was what I got from the full context.
2 Julian Sanchez // Oct 3, 2005 at 1:55 pm
No. He was using the comment to illustrate the problems with cost/benefit sorts of arguments about abortion; I don’t think the point had anything to do with statistical generalizations.
3 J. Goard // Oct 3, 2005 at 4:16 pm
That is, you have to suppose not just that 2005 America is such that blacks commit more crimes (which is true), but that they’re intrinsically more likely to commit crimesââ?¬â?that this will hold true 20 years from now or 30 or whenever.
And also that crime is such that is primarly results from “crime-proneness” factors internal to certain individuals or their subcultures, as opposed to having a major basis in hawk-dove game theory such that others in the general population would pick up the slack.
4 delong // Oct 8, 2005 at 1:19 am
Well I, for one, am pessimistic about sociological conditions changing sufficiently in the next thirty years…. But you are right, there is a tinge in Bennett… a guy who has never been known to approve of sociological causation… who blames criminality on a lack of moral virtue (either one’s own or one’s parents)…
So touche.
Which reeducation camp do I report to?
Brad DeLong