The short answer is that one has to ask whether the analogy is a good one, and my immediate intuition is that it is not a good one in the case of AIDS, at least not as you are using it.

To what, in that case, is the expanding baby supposed to be analogous? The AIDS virus? If so, then what the thought experiment suggests is that one woudl be justified in killing the AIDS virus. But of course we already knew that. If the thought experiment is supposed to "justify killing everyone who had AIDS", then, the expanding baby would have to be analogous to a person who had contracted the AIDS virus. But it isn't: People who have the AIDS virus do not pose an imminent risk of death to those who do not, and killing all such people is (fortunately!) not the only way to save oneself, if one has not contracted the virus.

Whether the analogy is a good one in the other case is a different question, and one's answer to that question will turn upon one's understanding of the relationship between a woman and the fetus she is carrying. The question how we should understand that relationship is, to my mind, the most profound one raised by Thomson's paper.

Agreed, the bumper sticker has very little persuasive appeal to those it purports to be addressing. But something could be said against your more general point, as follows.

Suppose someone believes that abortion is morally (roughly) on a par with infanticide and murder, and also that infanticide and murder are terrible crimes that ought to be criminalized and punished severely. Would it be incoherent for such a person also to hold that the criminal law should not interfere with any woman's decision about whether to have an abortion or not?

One could hold these two views together -- in fact, Mario Cuomo held them together when he was governor of New York State. One could rationalize the combination like this: "I am convinced that abortion is murder. But the grounds of my conviction are rooted in a religion that many of my fellow citizens do not share. In fact, my country is deeply divided on the issue of abortion, with many reasonable persons on either side. Under such circumstances, it would be wrong to use the criminal law to suppress conduct that many reasonable people conscientiously believe to be morally permissible. It is not wrong, by contrast, to use the criminal law to suppress infanticide and murder, because their status as terrible crimes is widely recognized by citizens from all the diverse walks of life."

Hard to fit on a bumper sticker, this point is nonetheless worth a thought. A pro-choice person can fully respect others' conviction that abortion is murder and still make a reasoned appeal to these pro-life people that they should not use their political power to conform the criminal law to their largely religious convictions.

You are absolutely right about the bumpersticker, and your analogy brings out precisely what's wrong with it. I am in favor liberal abortion laws. But I cringe every time I see that bit of rhetoric emblazoned on a car or button.

Here, perhaps, is what the bumpersticker would say if one had a much, much larger bumper: "Reasonble people can disagree as to whether a zygote, embryo, or fetus ought to be accorded the same moral status as a mature human being. If they are not properly accorded that status, then killing them is not murder. Given that reasonable people can disagree about all this, the state ought not to legislate against abortion, but should rather leave it up to the conscience of each idividual whether to have one or not."

It is a sign of the general degradation of our public political discourse that so much of it involves the mere trading of elliptical, misleading, and inflammatory bumpersticker-size slogans. The general trend against sober, reasonable discussion in the public realm is exacerbated in this case by the two main opposing advocacy agencies involved in the abortion debate: NARAL, and Right-to-Life, who each have a vested interest in keeping the debate polarized. Polls cosistently show that U.S. citizens favor liberal abortion laws for at least the first trimester of pregnancy, that they want abortion available in certain circumstances during the second, and that they do not want abortion permissible in the third trimester except to protect the life or health of the mother. This compromise position would have very general support, and if implemented, might end the rancorous controversy for good. But you’ll never see either group propose it – for both groups, it must be all-or-nothing.

By the same reasoning, your mother would have done a terrible thing to you if she had abstained from sex on this occasions or if she had insisted that your father use birth control. Such conduct, too, would have denied a good life to a person in the future. I assume that you would not want to classify such conduct as worse than murder, but rather as perfectly permissible -- even though sometimes, when you use birth control, a person who otherwise would have come to exist will not.

The problem with your analogy is this: In the case of murder, there is a person who is harmed, a person who is deprived of further years of life, prevented from realizing her plans, projects, and ambitions. In the case of abstention or birth control, by contrast, there is no person who is harmed, no person whose life is cut short, no person disabled from realizing her plans, projects, and ambitions. Murder harms an existing person, birth control merely prevents the coming into existence of a person. This explains why it is not wrong to abstain from sex or to use birth control.

Now the case of abortion is different from both others. It is like the murder case in that something exists whose life is cut short. And it is like the abstention case in that there is no person suffering a harm. Those who believe abortion to be morally permissible liken it to the abstention case: An abortion does not kill a person, but merely prevents the coming into existence of a person.

Proponents of the Doctrine of Double Effect draw a distinction between two sorts of cases: (1) you intend to achieve a particular result R through your action (i.e., this result is the purpose of your action) and (2) you intend to achieve a different result T through your action but foresee that your action will have an additional (though non-intended) result R. What’s crucial, according to this doctrine, is the object of one’s intentions. In the case of abortion, the intended purpose of the procedure is to terminate a pregnancy. In most cases of ectopic pregnancy, the purpose of the surgery is to save the life of the pregnant woman, and a foreseeable, though not intended, result of the procedure is the termination of a pregnancy. (How can you tell whether the result is intended, rather than merely foreseen? See whether you would still perform the action if you were to learn that the result would not occur.)

I am not a fan of the Doctrine of Double Effect. It seems to me to depend on a distinction that makes no moral difference. To see this, let’s imagine that I have a very important appointment across town and that I’m currently stuck in traffic. The only way that I can get to my appointment on time is by driving on the sidewalk and running over a lot of people. Of course, I don’t intend to kill anyone; killing is not the purpose of my action. If I could get across town without killing anyone, I would. Nonetheless, I can easily foresee that driving on the sidewalk will put many people in harm’s way. On my view, the fact that I don’t intend to kill the pedestrians whose death I could easily foresee makes no moral difference whatsoever.

Now, of course, proponents of the Doctrine of Double Effect also would not endorse my running over pedestrians in the circumstances that I imagine. They would say that in order for my action to be permissible, the harm that I’m trying to avoid must be “proportionately” as significant as the harm that I can foresee that I will cause. Since in the case of an ectopic pregnancy the life of the mother and the life of the fetus are equally significant, surgery that terminates the pregnancy is permissible.

So, let’s change the case of our impatient driver just a bit. Let’s say that I’m a very courageous and skilled firefighter called to a burning apartment building across town. Let’s also stipulate that my town is very poor, and though populous, very remote. I am the only firefighter within hundreds of miles. I believe that it is important to save the lives of the people who are trapped in the burning building, but the only way that I can get to them in time is by driving on the sidewalk (due to budget cuts, my broken siren and horn still have not been repaired, and due to an injury suffered in a previous daring rescue, I am mute). According to the Doctrine of Double Effect, so long as I don’t intend to kill the pedestrians (even if I can foresee their death) and so long as I intend to save the lives of the many trapped residents of the burning building, my sidewalk escapades would be morally permissible. While I myself think that it is actually quite difficult to explain why I should not drive on the sidewalk in these circumstances, I trust that most Catholics will agree with me that I should not.

Of course, the case of an ectopic pregnancy is importantly different from this last case. If a physician doesn’t save the life of the mother, both the fetus and the mother will die. But isn’t this the fact that should be most morally relevant to a person who believes that human life is sacred from the moment of conception? The fetus will die no matter what; the mother’s life can be saved. If human life is sacred, the mother’s life should be saved. There’s no need for an appeal to the Doctrine of Double Effect.

These are obviously questions about biology, not philosophy. I'd try a good biology text.

That said, it isn't plausible that personhood has to do with number of cells. Number of cells is roughly proportional to size. One would therefore expect that the number of cells in a large tree would dwarf the number of cells in a mature human being. Same for elephants and whales, let alone the strange fungus that is reported to be the world's largest living creature.

Exactly the difficult with the "life is sacred" slogan. If a choice has to be made between lives, how does one carry it out unless there is some way of balancing lives against each other? On the other hand, one can see the logic of leaving it to God, if one believes in him, since how can we play one life against another? As the Talmud puts it when considering whether one person should be sacrificed for someone else, how do we know whose blood is redder? I think that means it is not possible to say that one person's life is more significant than someone else's.

For a consequentialist, though, this just looks like a decision not to enter into a calculation at all when there are instances where this should be done. Triage is based on the idea of the efficient use of limited resources, and it seems right to treat a patient who looks as though she might recover as compared perhaps with someone who looks as though they will not. It might still be argued that life is sacred, but interpreted in such a way that we would allow intervention based on distinctions about whose life in a particular situation should be saved.

The answer to your question depends in part on whether and how the moral status of a fetus differs from the moral status of an infant. One might reasonably think, for example, that there is no significant difference in moral status between a fetus the day before its born and an infant the day after its born. The only difference seems to be one of residence: one resides inside a uterus, and the other resides in a hospital's maternity ward. And we don't normally think that where one lives is especially relevant to one's moral status. Thus, if it's wrong to kill one, it must be wrong to kill the other. That said, one might also reasonably think that there is a significant difference in moral status between an embryo that is only two weeks old--a cluster of cells that looks something like this--and a newborn baby. And if there is a significant difference in moral status, then it is perfectly consistent on one hand to defend the right to have early-term abortions but on the other to oppose infanticide.

Of course, questions about the moral status of an embryo or fetus are in many ways the locus of ethical questions about abortion in general. (They do not play as central a role in the legal debates about whether laws banning abortion are constitutional.) Other questions might also be relevant to the ethical debate, though. Some philosophers--most notably Judith Jarvis Thomson--have argued that abortion is not always wrong even if the fetus has the same moral status as an infant. These philosophers often appeal to a woman's reproductive rights--her right to control her own body and to control whether or not she will become a mother--in order to defend the right to abortion. Thomson's rights-based defense of abortion is available online here.

The view that Matthew articulates--that the moment of birth is not morally significant in a way affects deeply the moral status of the newborn infant--is a popular one, but it has been challenged by some. For example, the feminist philosopher Mary Anne Warren argues that birth is morally significant in virtue of the newborn's expanded social relationships.

To say that newborns have a different moral status than nearly-borns does not mean that late-term fetuses ought not to be protected from harm -- but Warren's line of thought might provide a philosophical basis for concluding that newborn infants and late-term fetuses are not entitled to exactly the same degree of legal protection.

To answer your question more directly: If there is a substantive question about whether newborns and late-term fetuses have significantly different moral status, there is even a stronger case to be made that there are big differences in the moral status between newborns and fetuses that can be legally aborted. If this is correct, killing a newborn and using abortion to kill, say, a first-trimester fetus may well be morally different.

Finally, for more on Warren's view see her essay "The Moral Signficance of Birth."

There need be nothing inconsistent about this position. The first view, that abortion is morally impermissible, is a moral or ethical view. The second view, that each woman should be permitted to choose for herself whether to have an abortion, is a political view, one about what laws a state ought to have. The combination is therefore consistent so long as one denies that, if it is morally impermissible to do A, then it ought to be illegal to do A.

Why might one deny that claim? One might well ask why one should endorse it, but there is a better answer. Suppose one believed the following:

1. A law must be justifiable on the basis of principles that cannot reasonably be rejected by any citizen, where a rejection is "unreasonable" if it is flatly irrational or, more interestingly, based upon too many particulars of one's actual situation.
2. In particular, religious doctrine cannot figure in the justification of any law, since one's finding the appeal to any religious doctrine convincing depends too strongly upon one's actual situation, namely, one's particular religious affiliation.
3. In short: The brute fact of religious pluralism invalidates appeals to religion in the "sphere of public reason".
   

That's roughly the sort of view that is defended in John Rawls's A Theory of Justice, though I've adopted some terminology from Tim Scanlon in formulating the view (and so brought it closer, probably, to Rawls's later view, in Political Liberalism). The underlying idea, as applied to religion, is traceable to John Locke's Letter Concerning Toleration and Two Treatises on Civil Government. The latter is essentially the basis for the United States Constitution. (Indeed, large parts of the Declaration of Independence are based, fairly directly, upon the Second Treatise.) Indeed, one might regard (3) as a way of understanding both the basis and the content of the Establishment Clause.

If one held this sort of political view and also based one's moral objection to abortion on some religious ground—say, one's argument depends essentially upon the claim that life begins at conception, and one takes the ground for this claim to be "revealed truth"—then the combination of views mentioned is not only consistent but required.

There are many excellent philosophical discussions of abortion, and many of these do tie the question to general moral issues. One classic article is Judith Jarvis Thomson's "A Defense of Abortion", Philosophy and Public Affairs 1 (1971), pp. 47--66. Thomson's argument begins, contrary to what public political discussion might lead one to expect, by granting that the fetus has all the rights of a person. She argues that abortion is nonetheless justified because it involves a conflict of rights.

Her central example goes like this. Suppose you were kidnapped by music lovers and connected via tubes and wires to a famous violinist whose life now depends upon your remaining connected to him. If you remove the tubes, he dies; if you remain connected for nine months, you both live. Do you have a moral duty not to remove the tubes? Thomson grants that it would be very nice of you not to do so, but, intuitively, you have no such moral duty. The example is meant to be analogous to cases of rape. Here's a variation that's no longer science fiction. Suppose a woman were kidnapped and, while she was asleep, embryos were implanted in her uterus. Does she have a moral duty to carry them to term? If you don't have such a duty in the case of the violinist, how does the woman have such a duty in this case?

To extend the argument beyond cases of rape, Thomson offers other examples. There are large questions about responsiblity that arise here. Many people think that, in cases that do not involve rape, the woman bears a certain responsibility for the life of the fetus that is morally relevant. It's not an unreasonable view that there is some moral difference between the cases. Whether it is enough of a difference to make abortion morally impermissible except in rape cases is another question.

Thomson's paper is collected in several different places, including in Joel Feinberg's collection The Problem of Abortion, where you will find some critical discussion.

A much cited, more recent paper is Don Marquis's "Why Abortion is Immoral", Journal of Philosophy 86 (1989), pp. 183--202. There are two replies in the May 1990 issue of the Journal of Philosophy. Walter Sinnot-Armstrong's paper "You Can't Lose What You Ain't Never Had", Philosophical Studies 96 (1999), pp. 59--72, seems to me a balanced discussion that ends up being critical.

Why hasn't the philosophical discussion had much effect on the public debate? I expect the reason is very simple: Abortion tends to generate strong emotions, and where there are strong emotions, reason stands little chance of a hearing.