The Ethics of Abortion

Abortion is the intentional termination of a pregnancy, either via surgery or via the taking of medication. Ordinary people disagree about abortion: many people think abortion is deeply morally wrong, while many others think abortion is morally permissible. Philosophy has much to contribute to this discussion, by distinguishing and clarifying different arguments against abortion, distinguishing and clarifying different responses to those arguments, offering novel arguments against abortion, offering novel defenses of abortion, and offering novel views about the relevant issues at stake.

This entry’s central question is: is abortion morally wrong? This question has important legal and political implications. If abortion is morally wrong, it is nevertheless an open and further question whether abortion should be outlawed or legally restricted in some way. It is also a substantive question how a society should decide whether to outlaw or restrict abortion. This entry focuses on the moral question and does not delve into these important legal and political questions.

The first six parts of this entry present six arguments against abortion. Subsections of those parts present responses to these arguments – that is, defenses of abortion. The seventh part discusses why the ethics of abortion matters.

1. The argument from a right to life

Let’s start by considering the following argument:

1. The fetus is a person.
2. All persons have a right to life.
3. Therefore, the fetus has a right to life.
4. It is morally wrong to kill something that has a right to life.
5. Abortion kills the fetus.

Therefore, abortion is morally wrong.

This argument has some real intuitive appeal. Abortion is the killing of a human being, so it does seem to be the killing of a person. In general, it is wrong to kill persons; it does seem that all persons have a right to life.

We’ll consider five lines of objection to this argument.

1.1 The equivocation objection

We might object to the argument from a right to life by arguing that the argument equivocates. The word “person” could have two different meanings for the purposes of the argument. Suppose that “person” means any animal that is a human being; on this reading, saying a being is a person tells us what species the being is. It is on this reading that claim 1 is obviously true: it says that the fetus is human. But suppose that “person” means “moral person” or something with full moral status – something that counts fully for moral purposes, the way that an ordinary adult human being counts. It is on this reading that claim 2 is obviously true: it says that all moral persons have a right to life. The objection is that the argument is only plausible if claim 1 is interpreted with one meaning of “person” and claim 2 is interpreted with another meaning of “person”.

What this objection shows is that the argument from a right to life, as we have presented it, has two different versions. One version of the argument uses “person” to mean “human being” throughout. For this version of the argument, premise 2 is a controversial assumption: it says that all human beings, even early human fetuses, have a right to life. Another version of the argument uses “person” to mean “moral person” or “being with full moral status.” For this version of the argument, premise 1 is a controversial assumption: it says that all fetuses, even the early human fetuses that die in first-trimester abortions, have full moral status.

This objection shows that the argument from a right to life has (at least) one controversial assumption: either premise 1 or premise 2 is controversial.

1.2 The violinist case objection

Suppose that a famous violinist is very sick with a serious kidney disease. All medical records have been searched, and you are contacted with the following surprising request: would you please come stay in the hospital for nine months, lying in a bed next to the violinist, so that your kidneys can cycle his blood? You are the only person who is the right match to help him, and after nine months he will have made a full recovery. You think it over, and reluctantly decline. While you would love to help another person, spending nine months in the hospital is a huge sacrifice that you are not prepared to make. The next day after you decline, you wake up in the hospital, attached to the violinist. It turns out that the Society of Music Lovers kidnapped you. The doctor tells you that, while it was wrong to kidnap you, it would also be wrong for you to unplug yourself: the violinist needs to stay attached to you for nine months, or he will die. If you unplug him, you would be killing him, and he has a right to life. While you have a right to control what happens in and to your body, his right to life is surely stronger.

What may you do in this case? Surely you may unplug yourself. (So claims Thomson 1971, which introduced this famous case.) But this shows that the argument we have been considering, the argument from a right to life, does not succeed. The violinist does have a right to life; yet it is permissible to kill him. Thus, it is not always true that one may not kill another human, just because that human has a right to life. The violinist case appears to show that claim 4 of the argument from a right to life is false. (The violinist case also undermines another claim, the claim that it’s not okay to violate one person’s right to life for the sake of a weaker right; this is the claim that Thomson aims to undermine, when she offers the case.)

Note that the violinist case objection defends abortion while granting the claim that the fetus has a right to life. The objector claims that even if the fetus has a right to life, it simply does not follow that abortion is morally wrong.

Consideration of this case may adjust our moral framing of the choice whether to abort a fetus. While it is natural to see this simply as a choice whether to kill a living being, reflection on the violinist case may lead us to see it rather as a choice whether to provide life-sustaining aid to a living being – with the further fact that one must kill that being to avoid providing the aid.

1.2.1 Response from consensual sex

The violinist case might appear to defend abortion only in the case of rape. It may seem to be important to the case that you were taken into the hospital against your will.

Thus, one might respond to the violinist case by offering a different argument against the permissibility of abortion:

1. A pregnant woman who has consensual sex is responsible for the fetus’s dependence on her body.
2. If you are responsible for someone’s dependence on something of yours, then it is morally wrong to deprive him of that thing.

Therefore, abortion is morally wrong.

To respond to this new argument, we might consider the following variant of the violinist case. Suppose that after you declined to help the violinist, you heard that the Society of Music Lovers was planning to kidnap you. You could stay home, in which case you would be safe. Or you could go about your normal life, with some hired bodyguards, hoping that any kidnapping attempt would fail. You choose to go about your normal life, with the bodyguards, and yet you are kidnapped and hooked up to the violinist. Surely in this case (the argument goes), you may still detach from the violinist, thereby killing him. This is true even though your choice to go about your normal life was a cause of your kidnapping, and in that sense you are responsible for the situation in which the violinist is dependent on you. Thus, claim 2 is shown to be false: although you are responsible for his dependence, you may deprive him of your aid. Surely a woman who has sex with contraception is no more responsible for the fetus’s dependence on her than you are in this variant of the violinist case; thus, claim 2 is false as applied to her case as well. (For further discussion of the responsibility objection, see Kamm 1992 and Boonin 2003.)

Perhaps surprisingly, one thing that emerges from considering this worry is that it may matter whether having sex is a central and important life activity, such that abstaining from sex would be a significant burden. In the variant of the violinist case, it may be important that staying at home to avoid kidnappers would be onerous. Would refraining from sex similarly be onerous?

(A different argument regarding responsibility holds that a pregnant woman is a mother to the fetus and has the role-based duties associated with that relationship. But that assumes that there is already, and will be in the future, a meaningful relationship between the woman and the fetus. See Little 1999.)

1.2.2 Does the violinist case provide its own positive argument?

So far we’ve discussed the way that the violinist case can be used to defend the permissibility of abortion in the face of the argument from a right to life. We might also consider using the violinist case in an argument to establish that abortion is morally permissible:

1. If abortion is morally wrong, then it’s morally wrong to detach oneself from the violinist.
2. It’s not morally wrong to detach oneself from the violinist.

Therefore, abortion is not morally wrong.

This argument isn’t made in Thomson 1971, but it is worth considering. Much of the discussion about whether the violinist case provides an effective response to the argument from a right to life will also shed light on whether this is a sound argument. What are the differences between the abortion case and the violinist case? Do they make a moral difference?

1.3 What is a right to life?

One might object to the argument from a right to life by probing the question of what a right to life is, and what kinds of duties it creates in others. This probing may provide another way of rejecting premise 4.

We can begin by noting that a right to life is not a right to be given whatever is necessary to sustain one’s life. If a person is dying, and all that she needs to save her life is the touch of (famous actor) George Clooney’s hand on her forehead, then nevertheless she has no right that he come to her bedside to cure her with his touch. (See Thomson 1971, page 55.) Similarly, the famous violinist mentioned above has no right to the use of your kidneys to cycle his blood. And he has no right that you remain attached to him – even though detaching yourself deprives him of what is necessary to sustain his life.

For a yet more fanciful example, suppose you are trapped in a tiny house with a rapidly growing child. The house is very small, and the child is extremely large – if his growth continues unchecked, he will crush you to death. The child himself will be unscathed; eventually he will grow big enough that the house will burst apart and he will walk free. Does the child’s right to life mean that you cannot defend yourself from the threat he poses to you? Must you simply submit to die in this way? Surely not, Thomson 1971 argues.

It might seem that a right to life is at least always a right not to be killed. But it is not always wrong to kill another person, even though that person has a right to life. If someone attacks you, and the only way to defend yourself is to kill him, then you may kill him. And consider again the violinist: detaching yourself kills him, yet it is permissible to detach yourself. These observations might seem to show that if someone has a right to life, it is not simply wrong to kill them, but rather it is wrong to kill them unjustly. Perhaps a right to life is merely a right against being unjustly killed (Thomson 1971). But if that is correct, then the argument from a right to life needs a different premise:

5*. Abortion kills the fetus unjustly.

This premise is much more controversial than the original premise.

Can self-defense justify abortion?

Some of the cases we have just been considering involve innocent threats and the idea that a being with full moral status can become liable to be killed in self-defense even when they are morally innocent. For example, the growing gigantic child who will crush you in your house unless you kill him first is not blameworthy for posing a threat to you; but he poses a threat nonetheless.

The idea that the fetus has full moral status from the moment of conception can actually be used to argue for the permissibility of abortion, as follows. Pregnancy itself is caused by the embryo, which implants into the woman’s uterus. If the fetus is an independent person, then it is like a mentally incompetent agent who perpetuates an illicit bodily intrusion against a woman, and then starts redirecting vital resources and nutrition away from her. Understood in this way, ending pregnancy is simply defending oneself against attack from another independent being. (See McDonough 1996.)

But one might deny that morally innocent threats can be treated in the way that morally guilty threats can be treated – at least when it comes to what third parties may do. Consider the following view: a person who is threatened by an innocent threat may kill to save her own life, but a third party may not choose sides between the innocent threat and the innocent potential victim. On this view, the threatened person has a mere “agent-relative permission” to defend herself: while she can defend herself, that does not mean that a third party can also act to defend her. (By contrast, it is of course true that a third party can act to defend someone against a culpable assailant.) If this view is right, then we cannot use the idea of self-defense to defend ordinary surgical abortion in the face of the argument from a right to life. While granting that women would be morally permitted to perform abortions on themselves, it would not follow that doctors are permitted to perform abortions. (See Davis 1984. For the view that doctors can act as the pregnant woman’s agent in her self-defense, see English 1975.)

What should we think about this worry?

First of all, it is important to note that the reframing of abortion that the violinist case suggests need not be understood in terms of “self-defense”; rather, abortion is reframed as fundamentally about the withdrawal of aid.

Second, it is implausible that mere moral innocence renders a threat immune to attack from third parties. It does seem that a temporarily insane person who rushes at an innocent person can be killed by anyone, not just by the attacked person; this seems true no matter how blameless the attacker is. (For more on innocent threats, see Frowe and Parry 2024.)

1.4 Potentiality

The objections we considered in sections 1.2 and 1.3 granted the claims that fetuses are persons, and that persons have a right to life: these objections granted claims 1, 2, and 3 of the argument from a right to life. Let’s turn now to considering a challenge to claim 3, which holds that fetuses have a right to life. The objector holds that having the potential to become a being with a right to life (or with full moral status) is not sufficient to currently have a right to life (or to currently have full moral status).

(Note that this objector will also deny either claim 1 or claim 2. They will deny claim 1 if it is read as the claim that fetuses have full moral status, or they will deny claim 2 if it is read as the claim that all human beings have a right to life. See section 1.1.)

We’ll consider two different challenges to the idea that potentiality is sufficient for a right to life.

1.4.1 The kitten serum case objection

Suppose that there was a serum we could inject into kittens that would make them into cognitively sophisticated cats; these cats would be as intelligent as human adults, and they would have the emotional complexity that human adults have. One of these cats would surely have full moral status; it would surely have a right to life. (So claims Tooley 1972, which introduced this famous case and offered the following argument.) While we would have strict moral duties concerning these cats, just as we have strict moral duties regarding other persons, we would not have any duty to turn kittens into these sophisticated cats: we would not be obligated to inject any kittens. Furthermore, if a kitten were injected, but the serum had not yet had any effect, it would be morally permissible to inject a neutralizing agent that would prevent the serum from working; this would be just as permissible as failing to inject. A kitten who had been injected with the serum plus a neutralizing agent would be just like any ordinary kitten. Kittens have no right to life: there is no moral requirement to refrain from painlessly killing them (so the argument goes). But now consider a kitten that has just been injected with the serum. It has the potential to have a right to life. But it lacks a right to life: it is morally permissible to inject the neutralizing agent and then to painlessly kill it; this would not be morally permissible if it had a right to life. So the kitten that has just been injected with the serum is a counterexample to the claim that a being with the potential to have a right to life must already have a right to life. (This is the argument of Tooley 1972.)

For those who believe we may painlessly kill kittens, this argument poses a powerful challenge to the idea that fetuses have a right to life. But the rhetorical effectiveness of this argument may be greatly diminished by the fact that nowadays, many people think that animals do have a right to life, or at least that there are strong moral reasons against killing animals even painlessly.

Perhaps a variant of the argument can nevertheless gain some traction, even with those who believe that kittens do have a right to life. Consider the view that all conscious animals have a right to life, but that some animals have only a low level of moral status, while other animals have a high level of moral status. On such a view, it is natural to think that kittens have a low level of moral status, compared to the full moral status that adult humans have. If we consider the kitten serum case, it does seem right that we are not morally obligated to inject any kittens with the serum, and also that it would be morally permissible to inject a neutralizing agent after injecting the serum. But this does seem to show something about potentiality – it seems to show that having the potential to have a certain level of moral status does not already give a being that level of moral status. Consider a kitten that has just been injected with the serum. It has the potential to become a being with full moral status. If it already had full moral status, surely it would be wrong to inject the neutralizing agent, thus depriving it of the rich experiences that greater cognitive and emotional sophistication would provide. But it is permissible to inject the neutralizing agent; thus it does not already have full moral status.

Thus, even without the assumption that kittens lack a right to life, the kitten serum case can provide something important: an argument that the potentiality to have a certain level of moral status is not sufficient for already having that level of moral status. (This argument may provide a good response to other arguments against abortion, even if it does not engage the argument from a right to life.)

1.4.2. The actual future principle

Another objection denies the significance of the fetus’s potential future by contrasting it with the significance of the fetus’s actual future. Arguments for the claim that potentiality is sufficient for moral status may include claims such as “the fetus is the beginning stage of a person.” But this is not true for all fetuses. Some fetuses are the beginning stages of persons; the fetus that became you was the beginning stage of a person. But many fetuses are not; for a fetus that dies early in pregnancy, there is no person such that they were the beginning stage of that person. On one view, it is not potential future personhood but actual future personhood that makes for moral status: according to the Actual Future Principle, pre-conscious fetuses that will become persons already have moral status, while pre-conscious fetuses that will die as pre-conscious fetuses lack moral status. (See Harman 1999, which introduces the Actual Future Principle and argues that the Actual Future Principle makes sense of various commonly-held but apparently conflicting claims, such as the following two claims: that early abortion requires no moral justification, while nevertheless it is appropriate to love (and wrong to harm) those early fetuses that are being carried to term.)

1.5 The self-consciousness objection

Another objection to the argument also targets claim 3, that fetuses have a right to life. This objection holds that for a being to have a right to something, the being must desire that thing. But to desire to continue to live, one must have a conception of oneself as a being that persists through time. Fetuses lack any such conception. Therefore, they lack a right to life. (See Tooley 1972, which also argues that human infants lack a conception of themselves and thus argues that infanticide is morally permissible.)

A distinct objection arises from the weaker claim that to have moral status or a right to life, a being must be conscious (if not necessarily self-conscious), because to have any interests at all a being must be conscious. Since early abortions kill fetuses that are not yet conscious, this view defends abortion in these cases. (See Steinbock 1992.)