Economist Steven Levitt has been more cautious about presenting his conclusions after the publication of his new book, "Freakonomics," than he was the first time around, in 1999, when he began reporting the then-unpublished results of the study he and Stanford law professor John J. Donohue III conducted examining legalized abortion's impact on crime.

Back then, Levitt was more forthright about what we should take away from his study. He told the San Jose Mercury News that his study "reinforces the fact that allowing women to choose has a benefit to society." The reported benefit of these choices was a good chunk of the dramatic fall-off in crime in the latter half of the 1990s.

Levitt hasn't changed his basic finding but he has hedged a bit. He and "Freakonomics" co-author Stephen Dubner add a flood of context to blunt charges of racism, classism, and all the other isms that are likely to arise from such explosive subject matter.

For example, in the chapter dealing with abortion, they begin with a cute story about the role of non-aborted youth in the overthrow [and execution] of former Romanian strongman Nicolae Ceausescu. They allow that "even for someone who considers a fetus to be worth only one-hundredth of a human being, the trade-off between higher abortion and lower crime is, by an economist's reckoning, terribly inefficient."

And yet, here is the nut of what Levitt and his scribe have to say on the subject: "[W]hen the government gives a woman the opportunity to make her own decision about abortion, she generally does a good job of figuring out if she is in a good position to raise the baby well. If she decides she can't, she often chooses the abortion." And this sorting has led to more "wanted" children, fewer neglected, troubled kids and, thus fewer crimes.

There are a number of objections one could make here, but I will concentrate on the ethical dimension of this proposition. That is, suppose that economists and social scientists from other disciplines subject Levitt's conclusions to a battery of tests and find he has proved not only loose correlation but ironclad causation. In other words, suppose that more abortions do translate into lower incidence of crime, and go from there. Should that affect how we think about abortion?

Short answer: no.

If the issue was other than abortion, most reasonable people would agree with this answer with little protest. They know that there are many things that could be done to cut down on crime but that policymakers flinch from these remedies with good reason.

Case in point: People who are found guilty of murder could be executed quickly, without the possibility of appeal. Yes, this would result in the death of the occasional innocent party, but most economists who have looked into the issue agree that higher execution levels are likely to raise the stakes for people contemplating a capital crime, and thus lower their incidence.

But legislatures do not do this horrible thing because efficiency is not their only concern. There is also justice.

Like a slender and ever-wavering majority of Americans, I am in favor of the slow, methodical, and, one would hope, just application of the death penalty that currently prevails. But many of us would become screaming death-penalty opponents if the legal safeguards against summary executions were removed.

We would change our position because we recognize that not all methods of deterring or punishing crime are legitimate. At some point, the crackdowns on crime become at least unjust, if not outright criminal.

Take torture. More crimes could be solved if cops were allowed to beat confessions out of people, but police are barred from doing this for two reasons: respect for the dignity of our fellow persons and the intuition that torture corrupts the torturer. In order to inflict damage on suspects, torturers must treat them as less than human, and this process has a tendency to distort how one looks at the world.

As with torture, so with abortion. There have always been, and will always be, abortions. But giving it the sanction of law corrupts society all the way down the line.

In addition to data from the United States, Levitt and Dubner cite a few maddeningly vague "studies" in the non-Romanian parts of Eastern Europe, "and in Scandinavia," to buttress their case that abortion curbs crime. In most of these nations, they explain, "abortion was not forbidden outright, but a woman had to receive permission from a judge in order to obtain one." The authors report that the composite everywoman who was denied an abortion "often resented her baby and failed to provide it with a good home."

The fact pattern of the scenario rings true, but I would draw different conclusions than the authors. Here you have a woman who is told by her government that abortions are OK as long as the law is willing to go along. Then the law refuses to do so, thus encouraging the notion that the child is a burden that the state has ordered her to bear.

Of course there is going to be some resentment against the child. And of course the child's home will be less than ideal. At the state's urging, the mother came to think of her child as a thing that could be disposed of rather than as a person who deserved her respect and protection.

Some say that the law is a teacher. In the case of legal abortion, it teaches us to treat the least among us as less than human. To say that access to abortion cuts down on crime is well and good--and I'm all for free and open inquiry on the subject--but it is far too narrow a way of taking in the picture. It's like saying that the Tiananmen Square massacre led to a remarkable decline in Chinese shoplifting.