Brad defended Bennett (while calling him a fungus). Brad partially recants here
IIRC, Steve Levitt's argument--which I don't know if I believe--was that the children of unplanned pregnancies have a much higher propensity to grow up to commit crimes than the children of planned pregnancies. Hence the coming of abortion-on-demand--with its enormous reduction in the number of unplanned pregnancies brought to term--carried with it a large reduction in the crime rate a generation later. When Bennett tries to summarize this argument on the fly and on the air as part of constructing his reductio ad absurdum argument, he replaces unplanned with African-American. That is not a good thing to do.
You could try to save Bennett by imposing a non-racist interpretation on his statement. You could interpret him as meaning: "society puts enormous pressures on African-American males as they grow up that greatly increase the odds that they commit crimes." Indeed, that's how I first interpreted him. But on reflection that cannot be what Bennett meant. Bennett does not believe that society creates criminals: Bennett believes that if you commit crimes it is because you are a bad, flawed, weak, irresponsible, immoral person--that there is something wrong with your essence and with your choices, because it is your responsibility to obey the law.
I agree with Brad's orignal post and disagree with his partial recantation.
Bennett was making a reductio ad absurdam, it would make no sense to simply say "by the same logic, one could argue pro choice since fewer unwanted pregnancies carried to term implies less crime 18 years later" The reason is that this argument is not absurd. I don't know about Levitt but I bet many people consider his empirical work to be an (additional un necessary) sound reason to support abortion choice (I do). Bennett doesn't like consequentialism, but he had to describe likely good effects of a clearly appalling policy to make his argument.
Further not only did Bennett not advocate genocide by abortion, he also didn't consider it. He chose as his example a policy which is obviously unthinkable just because it is unthinkable to him too.
Brad argues that Bennett is racist because "Bennett does not believe that society creates criminals" and yet thinks that Blacks are statistically more likely to commit crimes. His recent gaffe adds nothing to this (valid) argument that he is a racist. As a matter of fact, according to official statistics, African Americans are even more likely to commit crimes than European Americans (and much much more likely than Asian Americans or European Europeans). Thus Brad's proof that Bennett is a racist was just exactly as strong last week as this week.
Now I agree with Brad on this too. I am sure that all people who disapprove of sociological explanations of behavior on the grounds that they undermine the principle of personal responsibility are ipso facto racist. This has nothing to do with the gaffe, since the new information is just that Bennett is aware of official crime statistics (and believes them). Sure, his high sounded claims about how you shouldn't look for the roots of crime in poverty are, on inspection, vile and despicable. However they have been publicly available for inspection for decades. The gaffe added nothing new to the case against William Bennett (which required no new evidence as Brad noted in your original post).
If Bennett's beliefs about personal responsibility are inconsistent with believing that expected crime conditional on childhood poverty is greater than expected crime conditional on childhood non poverty, then he must believe that Blacks are born bad. I doubt that he believes any such thing. I think he understands that accurate conditional predictions of criminality conditional on childhood poverty are based on the recognition that poverty causes crime. However, I think he also thinks that one should not conclude from this that crime is fundamentally or essentially caused by poverty. The key bits are the words "fundamentally" and "essentially". These words have nothing to do with conditional probabilities and concerns about what is fundamental and essential can not be addressed empirically.
Roughly, I think Bennett thinks that, of course, there would be less crime if there were fewer unwanted births and less poverty etc but that it is wrong to state this fact. This is madness caused by ideology. The sense that there are true statements of positive fact which it is wrong wrong to state is pure ideology its own self.
Bennett's critics (and I) are similarly unwilling to face facts. It is very reasonable to believe that if you know that Mr Smith and Mr Jones are 40 year old Americans and the only other thing you know is that Mr Smith is African American and Mr Jones is European American, a glance at official statistics would lead a rational person to assign a higher probability to the statement "Smith has committed a felony" than to the statement "Jones has committed a falony." If one believes offical statistics are not massively massively biased by racism, then this is just noting that given a draw from a population probabilities are equal to frequencies in that population.
Note I said massively massively. Even if you are sure (as I am) that official statistics are massively biased by racism, the evidently unacceptable statement follows, since the differential in official statistics is very large.
Bennett was not recently condemned for denying facts in a way which supports racism. He was condemned for applying the logic of the caller to known facts and, thereby stating an evidently unmentionable fact.
I'd say I see no reason for Brad to retract anything given the argument in Brad's partial retraction. Bennett is a fungus but his recent performance on radio provides no relevant evidence of his fungal nature.