Abortion Bans and Cruelty

  • F. M. Kamm (Rutgers University)


Abortion bans have been characterized as cruel especially in not allowing exceptions for rape or incest. The article first examines one approach to morally justifying bans based on the Doctrine of Double Effect (DDE) which distinguishes morally between killing or letting die intending death versus doing so only foreseeing death. It then presents some criticisms of the implications of the DDE but also argues that what the doctrine permits helps provide a ground for the permissibility of abortions even if the fetus is a person. Whether there are limits on applying this argument is also considered. In conclusion, the article considers whether and why not permitting exceptions to bans for rape and incest is cruel and whether cruelty is a ground for opposition to bans.

Keywords: abortion bans, fetus, Doctrine of Double Effect, intentional killing, duty to aid, time limits on abortion, cruelty, rape

How to Cite:

Kamm, F. M., (2023) “Abortion Bans and Cruelty”, Journal of Practical Ethics 11(1). doi:



Published on
22 Dec 2023
Peer Reviewed

The United States Supreme Court announced its decision in Dobbs v. Jackson Women’s Health Organization on 24 June 2022. The ruling overturned the court’s decision nearly fifty years prior in Roe v. Wade that women in the United States had a constitutional right to abortion, and several states immediately introduced legislation to ban abortions from conception on or beyond a few weeks into pregnancy. Occurring as it did only a few months before the 2022 midterm elections, Dobbs featured heavily in political discourse. Some Democratic party ads described a ban on abortions even in cases of rape and incest as ‘cruel’.

Legislation that bans abortions from conception onward is reminiscent of the Roman Catholic position on abortion as a moral matter, and so it could be useful to consider the possible rationale of such a ban by considering the Catholic position. It might also be useful to consider why opponents of such a ban find it cruel to not allow exceptions for pregnancies that result from rape and incest.

It is my hope that this article will help each side understand the point of view of the other, see where they agree and where they disagree, and facilitate respectful mutual engagement. I will sometimes employ unrealistic hypothetical cases. Investigating the moral judgments we make about such cases can help us uncover what factors underlie our moral judgments in real life cases.

In section 1, I examine the moral doctrine that underlies one approach to banning abortion yet permits other acts that would also result in the death of a fetus (even when it is considered to have the moral status of a person, a status that confers ‘maximum rights’ on an entity). In section 2, I consider criticisms of the doctrine based on some of its problematic implications. Then, in section 3, I argue that based on the acts it seems to permit, the doctrine in fact provides a ground for the permissibility of abortions. In section 4, I consider whether there are limits on applying this argument, and in section 5, I consider whether and why not permitting exceptions to abortion bans (for rape and incest) is cruel and whether cruelty is a ground for opposition to such bans.

1. The Doctrine of Double Effect and the Morality of Abortion

The ban on abortion in Catholic moral theology can be largely explained by one part of a principle that is known as the doctrine of double effect (DDE). It claims that killing someone intentionally (or intentionally refraining from action in order to let them die), whether as an end in itself or even as a means to achieve a greater good, is strictly prohibited. The DDE is a general philosophical view not explicitly tied to Catholicism and is well known for its role in traditional just-war theory, which condemns intentionally killing civilians as a means to win a war. The doctrine supports a ban on abortion only if the conceptus at all stages of development, including the embryo and fetus,1 is considered to have the moral status of a person and only if abortion involves intentionally killing the conceptus as means to removing it from the woman.2 It would not support a ban on, for example, intentionally killing a bird to remove it if (in a hypothetical case) it were lodged in a woman’s uterus. This suggests that abortion bans from conception attribute the moral status of persons to the fetus from conception.

The prohibition in Catholic moral theology on intentionally killing or letting die has been interpreted to imply that even if both the woman and the fetus would die unless an abortion is performed, one may not intentionally kill the fetus to save the woman. One must let them both die because foreseeing their deaths, even with certainty, is not the same as intending their deaths, and intending death would make an act or omission morally wrong on this view.

It may be difficult to see the difference between foreseeing something (even with certainty) and intending it. Philosophers use what is known as the counterfactual test to distinguish between these concepts (even if imperfectly): Suppose that, contrary to fact, some deadly effect of your act (including omission to act) would not occur holding other things constant. Would this provide a reason (even if not a definitive one) not to do your act? If not, then you didn’t intend the effect. If the fact that the effect would no longer occur would provide a reason not to do your act, then you intend the effect. For example, consider a case where a procedure would save a woman’s life. An effect of the procedure is that her fetus will die. Is the death of the fetus intended? Apply the counterfactual test, imagining that the fetus would not die as a result of the procedure and could be raised to viability outside the woman’s body. Knowing that the fetus would not die, would this give the woman a reason not to pursue the procedure? If she would not pursue the procedure, then she intends the death of the fetus. If the fetus not dying would give her no reason to refrain from the procedure, she does not intend the fetal death.3

Now consider a second part of the DDE, according to which if one acts with the intention to produce a greater good and this results in the lesser bad of someone dying as a side effect, that death need not make the act morally impermissible. Yet the very same death would make an act impermissible if it were brought about intentionally as an end in itself or even as a means to the same greater good. On this ground, Catholic moral theology permits the removal of a pregnant woman’s cancerous uterus though it is foreseen that doing this will result in the fetus’s death, even if the fetus would be born alive before the woman dies of cancer were the uterus not removed. (Similarly, war causing foreseen collateral deaths of civilians can sometimes be permissible when intentional deaths are not. This can include bombing military sites for a greater good when it is foreseen but not intended that some of the bombs will also hit people nearby, killing them as a proportionate side effect.) One view of the cancerous uterus case is that removing the uterus kills the fetus as a side effect. Another construal is that, since there is no direct attack (even unintended) on the fetus, removing the uterus ends life support the woman is providing and so lets the fetus die (rather than killing it) as an unintended foreseen side effect. Actively terminating life support one is providing is typically thought, as a conceptual matter, to involve letting something die (whether or not it is permissible) rather than killing something. Just as killing may involve either a foreseen or intended death, letting die may also involve either a foreseen or intended death.4

Some revisionists have attempted to construe the death of the fetus in some abortions as an unintended side effect and so possibly consistent with the DDE. For example, in a craniotomy, crushing the fetus’s skull involves an intended attack on the fetus, but the fact that it consequently dies is said by revisionists to be only a side effect of a necessary means to its removal (the craniotomy). One could use the counterfactual test to try to show this: there would still be a reason to crush the skull to remove the fetus if, counterfactually, the fetus survived after removal from the woman’s womb. This understanding of a craniotomy as not itself involving death seems implausible. Another approach has been to revise the DDE so that it also rules out intending the involvement of the fetus (as in crushing its skull) in a way that foreseeably results in its death. Some philosophers make more fundamental criticisms of the DDE, denying that intention determines an act’s permissibility and offering other grounds to account for the impermissibility of many of the acts that the DDE condemns (e.g., Thomson 1999; Scanlon 2010; Kamm 2008). In this article, I will just focus on the DDE as traditionally understood.

State bans on abortion may depart in some ways from the uses of the DDE we have considered. For example, we have looked at the Catholic view that the DDE prohibits abortions involving a fetus that would otherwise die anyway. However, notably, some abortion bans in public policy—departing from the Catholic view—do allow intentional killing of the fetus in this circumstance. But what about cases in which it is not certain that the fetus would otherwise die, and yet the pregnant woman’s life is at stake? For example, it is not clear whether the state laws that would ban all abortions would permit the removal of a pregnant woman’s cancerous uterus on the grounds that the death of a fetus that might survive were the cancer to remain untreated is only foreseen and not intended. Insofar as state laws, unlike the Catholic view, permit intentional abortions to save the pregnant woman from a threat to her life caused by the pregnancy, they presumably would also permit removal of the uterus when the threat to her life is caused by uterine cancer.

2. Problematic Implications of the DDE

So far, we have considered the DDE in conjunction with the view that the fetus has the status of a person. We’ve seen how the principle can be used as a defense for prohibiting abortion and, at the same time, how it can permit other procedures that also result in the death of a fetus. Now let’s take a closer look at how it may be difficult to draw conclusions about the morality of abortion based on the DDE.

As already explained, the DDE permits killing as a side effect only if one’s act is intended to produce a greater good that outweighs that bad effect. Hence, if it were permissible to remove the cancerous uterus and in doing so kill the fetus (that would otherwise survive to birth) as a side effect, this would be because saving the woman is seen as a greater good and the death of the fetus a lesser bad. However, it is not clear why this should be so if both the woman and the fetus have the status of persons, and if in leaving the cancer untreated the woman would die while the fetus would survive. To see this, imagine that the removal of a woman’s cancerous uterus (hypothetically) had the unintended side effect of killing a person outside her womb who was in no way imposing on and dependent on her.5 The DDE would not permit the removal of the uterus that causes the other person’s death given that the woman’s living is not a greater good than the other person’s living. Furthermore, if there were twin fetuses who had the status of persons in the woman’s cancerous uterus, the greater good would seem to be the survival of two people rather than one person.

Suppose the removal of the cancerous uterus would nevertheless be permitted even when it causes two side effect deaths of fetuses who would otherwise survive. This might be because the fetuses do not have the moral status of persons or because, unlike the person wholly external to the woman, the fetuses are imposing on her and they stand to lose life that they only have due to this imposition. The permissibility of removing the uterus in this case might also be because the deaths, while a foreseen side effect of the procedure, would be a letting die rather than a killing since there is no direct attack (intended, as in the case of the craniotomy, or merely foreseen) on the fetus. After all, it would be implausible if the DDE permitted a letting die with foreseen death only if a greater good is intended. If it restricted letting die in that way, I couldn’t permissibly let someone die rather than sacrifice my arm to save them since intending to retain my arm wouldn’t be to intend a good greater than the bad effect of that person dying. But while it could be an admirable act beyond the call of duty (known as supererogation) to sacrifice my arm, I have no such moral obligation.

Any or all of these factors could morally distinguish a fetus (that dies as a side effect when the uterus sustaining it is removed from the woman’s body) from an external person (unconnected to the woman’s body who would be killed as a side effect of her uterus being removed) and could make removal of the uterus permissible in one case and not the other.

Some of these factors may also open the door to permissible abortions. For example, traditional Catholic moral theology permits lethal self-defense against an aggressor so long as one does not intend their death as an end or means (as when one pushes an aggressor away, merely foreseeing they will fall over a cliff and die). Although the fetus is not a willing aggressor, it might be seen as what is called a ‘morally innocent threat’ because through no fault of its own it threatens a woman’s bodily autonomy and possibly also her well-being or life. This is in contrast to a person outside her womb. The threatened woman might therefore, by this reasoning, be entitled to defend herself (or seek another’s help to defend her) against the threat—whether the threat is a cancerous uterus (where nullifying the threat results in a fetal death as a side effect) or the fetus itself against which one may need to act. Further, it is an implausible implication of the DDE that one may not intentionally kill a willing aggressor if that is necessary to protect oneself or others. Hence, it may even be permissible to intentionally kill a morally innocent threat, especially if the life it loses is one it only has due to imposing on the person who intends the killing to end the imposition. I’ll discuss how this can be so further in section 3.

3. How the DDE Might Bear on the Permissibility of Abortions

In section 2 we considered the possibility that some factors (such as stopping an aggressive threat) may sometimes make even intentional killing morally permissible. Now we’ll consider in more detail whether intentionally killing a morally innocent fetus in an abortion could be permissible.

That the DDE permits removing a pregnant woman’s cancerous uterus is of significance in this regard. It shows that abortion bans that use DDE reasoning and consider the fetus to have the moral status of a person do not rely on two other distinct views: (1) a pregnant woman has a duty to provide the use of her body to the fetus at great cost to herself or, on the flip side, (2) a fetus has a right to use the woman’s body at great cost to her when it needs it to survive. If it did have such a right or the woman had such a duty, the woman would not be permitted to have her cancerous uterus removed when doing so leads to the death of a fetus that would otherwise survive. Imagine that a woman’s fetus was initially growing outside her body and would die if not transferred to her body. A ban on abortion based on the DDE does not imply that the woman has a duty to transfer it to her body rather than let it die. And while the cancerous uterus case does not imply that the woman may refuse to let the fetus use her body when the cost to her would be much less than death, Catholic moral theology does not claim that the woman has a general duty to let the fetus use her body and it does not base an abortion ban on there being such a duty.6 Suppose actively terminating life-saving aid that one is providing (e.g., one removes a person from one’s breathing machine) is a case of letting die rather than killing. Then if an abortion involved only preventing or terminating assistance to a fetus to avoid costs less than death to the woman, it too might be permissible according to the DDE even if it foreseeably leads to the fetus’s death.

On 23 December 2022, an announcement was made that the contraceptive pill Plan B prevents ovulation and does not prevent a fertilized egg (a zygote) from implanting in the uterus (a process that some referred to as an abortion). However, suppose a pill did prevent a fertilized egg from implanting in the uterus by, for example, blocking the route of the zygote to the uterus or making it inhospitable to the zygote. Using this revised Plan B pill would be comparable to blocking the route to the uterus of the fetus growing externally (in our example) without any attack on the fetus. If the latter were permissible according to the DDE, it is not clear why using the revised Plan B pill should not be permissible as well according to the DDE. This also bears on Judge Matthew Kacsmaryk’s description of the abortion medication mifepristone as ‘starving the unborn human to death’ in his 7 April 2023 decision aiming to ban use of the drug in the United States. Starving involves refusing sustenance (in this case by the woman’s body) foreseeing death and so could also be consistent with the DDE.

In these respects, at least some supporters of a ban on abortion who rely on the DDE agree (surprisingly) with some claims emphasized by Judith Jarvis Thomson in her groundbreaking 1971 article ‘A Defense of Abortion’. In that article, Thomson assumes only for the sake of argument that the fetus is a person. She then argues that a person has no right to use another’s body, even to save their life. This is consistent with the view that one does not have a duty to always bear a lesser loss to save someone else from a greater loss (e.g., one is not morally obligated to give up one’s arm to save someone’s life). Thomson thinks this implies that one may detach another person if they are already using one’s body, even when one’s life or health is not at stake and when doing so will foreseeably lead to this other person’s death. She sees such detachment as a permissible killing. We have already seen that detachment may be an active form of letting die rather than a killing—it involves neither a deliberate attack on the fetus nor an intention to kill it as a means of removal. In a follow-up article, ‘Rights and Deaths’, Thomson makes clear that she wants to justify intentionally killing even an innocent person by either the woman or someone who chooses to help her if this is a necessary means to stop the person’s use of the woman’s body against her will (Thomson 1971, 1973).7

By contrast, as we have already seen, some argue for a ban on abortion because in their view it involves intentionally killing a fetus as a means to remove it from a woman’s body. This is so, even while they agree that she doesn’t have a duty in general to provide such support to save the life of someone outside her body. The question is whether these two positions are in tension with each other. In what follows, I aim to show that it is possible to make the grounds for the permissibility of intentionally killing the fetus in an abortion more precise if bans are not based on women having a duty to sacrifice themselves for the fetus’s life or the fetus having a right to use a woman’s body to save its life per se.

Suppose the woman has no duty to provide the fetus with use of her uterus merely to save its life. Then when she is pregnant, she is providing life support to the fetus to which it has no right even to save its life. This means that any right the fetus has to continue receiving life support would have to be based on the possibility that intentionally killing it to remove it is impermissible (since not being killed is not the same as having one’s life actively saved).8 But (1) it is a mistake of the DDE to imply that it is never permissible to intentionally kill (as discussed earlier), and (2) since in being killed the fetus would lose only life it is getting from imposition on the woman to which it has no right merely to get that life, (3) intentionally killing it (which deprives it of that life) in order to end that imposition can be permissible. This argument (which I first presented in Kamm [1992]9) does not merely point to the fact that the fetus is imposing on the woman’s body in a way that she need not permit, even to save someone’s life, as this could be true even when the imposition is not in fact saving someone’s life.10 It emphasizes that she is in fact providing it with life support and so killing the fetus deprives it of no more than the benefit of an imposition that is not justified by the benefit it produces, in order to end the imposition.

Suppose this argument for the permissibility of intentionally killing the fetus (even if it is a person) in order to stop providing it with some types of life-saving aid is correct. Killing in order to stop all aid that is supererogatory to save a life (i.e., aid we don’t have a duty to provide to save a life) might still be impermissible. For example, suppose one needn’t give someone a large sum of money to save their life. It can still be impermissible to deliberately kill the person whose life is being saved by the money even if that is the only way to get the money back. (This is consistent with it being permissible to remove the money even though this has the foreseen effect that the person dies.) Hence, it is the seriousness of imposition on the woman’s body (by contrast to losing money) that also plays a role in the justification of killing as a means to stop the imposition that provides life support.

4. Possible Limits on this Argument for Permitting Abortion?

In section 3, I offered an argument for the permissibility of abortion even if the fetus is a person. Now I’ll consider whether there might be limits to its applicability. Some argue that abortion can only be permissible if the fetus is not a person. Thomson, by contrast, argues that abortion is permissible even if the fetus is a person from conception onward (although her own view is that the fetus becomes a person at a late stage of pregnancy). It might be thought that if abortion is permissible when the fetus does not have the status of a person and it is permissible even if it is always a person, then abortion is permissible whether or not the fetus is a person. However, this might not be true in virtue of the following issue (that Thomson does not discuss): If a fetus is not a person from conception onward but achieves personhood at a certain later point of development, there may be a moral difference between aborting it before versus after it reaches that point. This is because it is presumably a less serious moral matter to kill a nonperson than a person. Suppose one is morally responsible for failing to do the less serious act within a certain time frame and for thus putting oneself in the position of having to do the more serious act of killing a person. Is it still permissible to abort a person or should there be a time limit on abortion?11 The possible ground for a time limit suggested by these considerations does not depend on the moral significance of a fetus’s stage of development per se. This is because if the fetus were a person from conception onward, there would have been no time at which one could have aborted a nonperson, so killing a person in an abortion would still be permitted by the arguments discussed earlier.12 The ground for a time limit we are now discussing depends on whether or not abortion at an earlier, nonperson stage is possible.

One possible response to this concern is to consider whether a time limit would similarly apply to letting a fetus die (rather than killing it). That is, suppose a woman had to take a harmless pill every month to keep a pregnancy going. If she was able to stop taking the pill early on before the fetus was a person but didn’t, is it permissible for her to stop taking the pill when the fetus is a person? Suppose (for the sake of argument) it’s still permissible to stop taking the pill even though taking it is so easy that doing so in order to save a person would ordinarily be her moral duty (since in general the easier it is to do an act that saves a person the more likely one has a duty to perform it).In that case, omitting to take the pill could be permissible because the fetus still has no right to the use of her body despite her not having stopped aid earlier. This would imply that in the case of delayed abortion in which the fetus has become a person, it too would still have no right to the use of the woman’s body to save its life. This would make its abortion still permissible since in being killed it would only lose life it gets from use of her body to which it has no right for purposes of retaining its life, and killing the fetus is the only way to stop such use (as argued in section 3).13

5. What Constitutes Cruelty in Abortion Bans and Can It Be Grounds for Opposing Them?

Now let us consider why some think it is especially cruel not to have a rape exception to a ban on abortion.14 The impression given by the ads in which this claim was made is that the cruelty is a reason why a complete ban is wrong. In what follows, I will use hypothetical cases to help tease apart factors that are typically combined in discussion of pregnancies that result from rape. This will enable us to see which factors (alone or in some combination) might be sufficient or necessary to ground an exception to an abortion ban and how cruelty bears on this.

  1. The question of the woman’s moral responsibility for causing her pregnancy, and thus for the fetus’s existence, is different for victims of rape than for participants in consensual sex. Arguably, a rape victim does not bear any moral responsibility for causing her pregnancy since (1) she did not consent to the sexual act, and (2) she is not required to use contraception merely because it is possible that she could be raped. We could imagine cases besides rape in which the pregnant woman would not be morally responsible for the presence of the fetus. For example, suppose a woman with a cognitive disability consents to sex without realizing that she could become pregnant as a result.15 Then if the absence of moral responsibility for pregnancy grounded the rape exception, abortion might also be permissible in this case. However, that someone is not morally responsible for being in a situation that is bad for them (such as unwanted pregnancy) does not yet show that there is a permissible way of getting them out of that situation (nor that abortion is such a way).

Emphasizing the nonresponsibility ground for the rape exception might suggest that engaging in voluntary sex (even with contraception), which gives a woman at least partial responsibility for her pregnancy, makes it harder to justify having an abortion. If this were so, it would probably also imply that responsibility for pregnancy gives the woman a duty to provide the fetus with the use of her womb independently of having a duty not to kill it to remove it. Unlike the grounds for the ban on abortion we have already considered, this view about the role of moral responsibility for pregnancy connects (1) a duty not to intentionally kill the fetus to remove it from the uterus with (2) a duty to provide costly aid to the fetus rather than let it die. Thus, it might imply, where the DDE does not, that it is wrong to remove a fetus from a woman’s body when doing so would involve its merely foreseen death.16 That a ban based on the DDE does not permit a rape exception—because an abortion in those circumstances would still intentionally kill a fetus as a means of removal—itself suggests that a ban based on the DDE need not be connected with judging the morality of women’s responsibility for sexual activity.

  1. One thing that focusing only on nonresponsibility for pregnancy omits is the use of force against a woman in rape. It also omits that another person is acting on her body against her will or without her consent, which could occur even without force (e.g., when she is drugged so as to be unable to resist sex). All these factors could also be present in a case of nonvoluntary artificial insemination; the absence of sexual intercourse in this case would not reduce the strength of the ground (if there is any) for an exception based on nonresponsibility for pregnancy. All these cases involve wrongful involvement by another person in addition to the woman’s nonresponsibility for the pregnancy. This means that the woman who is raped is the victim of wrongdoing and has been wronged (which need not be true of the woman with the cognitive disability who becomes pregnant in our earlier example).17 The consequences of being wronged should be corrected when it is possible and permissible to do so. However, that abortion would correct the consequences of being wronged is not enough to show that it is a permissible means of doing so.

  2. When the rape exception is raised, some refer to how horrible it is to be required to bear the biological offspring of one’s rapist. (The emphasis is placed on ‘bearing’ it, being complicit in its development, rather than the fact of such an offspring existing.) However, suppose the rapist used the semen of an innocent party to impregnate the woman. His doing so would not undermine the case for a rape exception. The fetus would still be developing in her body due to the rapist’s act. Hence, the horribleness of having to bear the rapist’s offspring would not be a necessary condition for a rape exception to the ban.18

Opposition to the rape exception even to eliminate carrying the rapist’s offspring might again be based on the DDE claim that it is always morally wrong to intentionally kill a person, especially one that is morally innocent (assuming the fetus has the moral status of a person), even to correct the wrong that was done to another morally innocent person (the woman impregnated by her rapist).

Even if this view about abortion is incorrect (as was argued earlier), another hypothetical example can help make clear that ending pregnancy due to rape is not always sufficient to justify killing. Suppose that removing a fetus produced by rape from the woman’s womb would (somehow) require intentionally killing a second fetus of that rape that is outside of and not dependent on the woman’s body. If fetuses had the status of persons, it would be clearly wrong to kill the outside nondependent fetus—and it would be wrong to characterize the choice not to kill it as being cruel to the pregnant woman, even if refraining from killing it denied the rape victim her only chance to have the other, dependent fetus removed from her body. This is so despite the nonresponsibility of the pregnant woman, the attack on her involved in rape, her having to bear a rapist’s fetus, and that the nondependent fetus to be killed would also be the rapist’s. If it made no difference to the permissibility of killing the fetus in an ordinary abortion that the fetus is imposing and dependent for life on the woman’s body, then the impermissibility of killing the outside nondependent fetus resulting from rape would undermine the rape exception to the ban on abortion.

This hypothetical case in which aborting the dependent fetus requires killing the nondependent fetus shows that we must demonstrate the permissibility of an intentional killing before we can argue that it is cruel to refuse to kill in a rape case. If a serious act is impermissible, refusal to perform it can’t be cruel. This case also suggests that we need to draw a distinction between (1) a state of affairs being cruel, (being pregnant as a result of rape) and (2) it being cruel not to stop that state of affairs. For another example, consider that the US Constitution does not permit ‘cruel and unusual punishment’. Suppose that someone being tortured by a mechanical device is a cruel punishment (and so constitutes a cruel state of affairs), but the only way to stop this happening is to kill several innocent people. That the punishment is cruel does not show that it is cruel not to stop it by the only means available. Similarly, it might be a cruel fate to be carrying the offspring of one’s rapist, but that’s not sufficient to show that refusal to terminate the pregnancy is cruel. It is only after abortion has been shown to be permissible that refusal to permit it could be considered cruel.

However, suppose one argues that abortion is permissible on the ground that the fetus is imposing on the woman, would lose only the benefit of that imposition, and its being killed is the only way to end that imposition. Then adding to that argument that having to bear the offspring of rape is a cruel state of affairs can help show that the imposition is significant enough to justify killing to end it.

After providing a correct argument for permissibility of an abortion the cruelty of not allowing an abortion comes to the fore. Denying permission for an impermissible act is not cruel, but denying permission for a permissible act can be. Hence, once it has been shown that abortion in a particular case is permissible, the decision not to allow the abortion (which would end what may be a cruel state of affairs) can be characterized as cruel. The cruelty of the state of affairs is not sufficient alone to ground the permissibility of abortion. Nevertheless, those who claim it is cruel not to permit abortion in rape cases could be correct, even if they are incorrect in claiming that it being cruel not to permit such abortions is an argument for the permissibility of abortion when in fact it being cruel not to permit abortions only follows from already established permissibility of abortion.


In discussing points of agreement between some opponents and proponents of the moral permissibility of abortion (e.g., about whether a woman has a duty to share her body in order to save a fetus) and also arguing that contrary to the DDE intentional killing is sometimes permissible, I introduced potential grounds for defending a woman’s right to seek an abortion. I also argued that the truth of claims about the cruelty of denying abortions (even to end cruel states of affairs) depends on first showing that those abortions are permissible. This is by contrast to the permissibility of abortions depending on the cruelty of not permitting them. If what I have said is correct, banning abortion is not a morally justified policy and so can be cruel even if the fetus is taken to have the moral status of a person.19


  1. Henceforth, I shall use ‘fetus’ to refer to all stages of the product of conception. [^]
  2. I leave it open that something other than an actual person could have the moral status of a person. Something without this status could still be valuable in itself or instrumentally. However, supporters of abortion bans focus on fetal personhood and rights are associated with this status rather than with value. [^]
  3. Here is a further example from war: Suppose (contrary to fact) that if one bombed a munitions facility no civilians would be killed. Would this give one a reason not to bomb? If not (it is said), their deaths are shown to be mere foreseen side effects whose absence would not interfere with the purpose of one’s bombing mission. But if their deaths were required as a means to terrorize the population, one would have a reason to call off one’s mission showing that in the actual bombing case with civilian deaths one intended the deaths as a means. [^]
  4. Hence, one can draw a two-by-two square:
    Kill (K) Let Die (LD)
    Intend (I) K/I LD/I
    Foresee (F) K/F LD/F
  5. Being outside someone’s body is not necessarily the same as not imposing on and being dependent on someone’s body. This is because we could imagine a hypothetical form of pregnancy in which the fetus imposed on and was dependent on the woman by growing around (i.e., being outside) rather than in her. [^]
  6. However, according to the DDE, it would be morally impermissible to let the fetus die because one intended its death rather than merely foreseeing its death as an unintended side effect. [^]
  7. Thomson’s second paper was a response to John Finnis (1973). Some do not see the need to complete the argument given in Thomson’s first paper so as to justify a deliberate attack on the fetus in abortion. Different ways of removing the fetus are discussed in Kamm (1992). [^]
  8. One does not save someone’s life merely because one doesn’t kill them. [^]
  9. For more on this particular argument, which focuses on the fetus losing only the benefit of an imposition to which it has no right merely to save its life, see Kamm (1992). [^]
  10. Thomson’s cases often involve someone receiving life support. However, her argument speaks of ending use of one’s body that one needn’t provide even to save a life. This leaves it open that one might end use of one’s body when it wasn’t actually saving someone’s life. She does note that when someone is receiving life support from someone else’s body, if one may not remove them, then they will wind up getting support it was agreed they had no right to have merely to save their life. But this is not sufficient to show that they do not have a right to keep the support as a way to avoid being killed rather than merely in order to save their life. My argument aims to show why their losing only the life they get from support can make it morally unnecessary to avoid killing them. [^]
  11. I raised this issue in Kamm (2013). [^]
  12. By contrast, the decision in Roe v. Wade placed emphasis on stages of development per se in altering requirements for legally permissible abortions in different trimesters. [^]
  13. There being other ways of removing the fetus that do not involve killing it and do not require considerably more imposition on the woman than an abortion could also limit the argument for abortion. For example, suppose the means of removing the fetus to an external gestation device imposed on the woman no more than removing the fetus in an abortion. Then the former means may be required if the fetus has the status of a person. I first discussed how real and hypothetical external gestation devices could affect the right to seek an abortion in Kamm (1992). [^]
  14. Cruelty is here not being considered as a person’s motivation but rather as a characterization of an act, a refusal to act, or even a state of affairs independent of motivation (as we shall see). Incest may usually involve rape, but its involving abuse by a parent or sibling may be thought to make it worse. I will here consider it a variation on rape. [^]
  15. Science fiction-type cases could also provide examples of nonresponsibility. For example, suppose a woman unavoidably became pregnant by cloning when she menstruated and this was an inexplicable, unusual occurrence. [^]
  16. For detailed consideration of how various forms of responsibility (including intentional pregnancy) might or might not affect the permissibility of abortion, see Kamm (1992 and 2013). [^]
  17. Notice that she would also be a victim of a wrong if (in especially odd hypothetical cases) impermissibly twisting her arm or cheating her out of money made her pregnant against her will. Presumably the worse the wrong that results in pregnancy, the more important the correction of the wrong would be (other things equal). [^]
  18. Nevertheless, having her body nourish an entity that conjoins the rapist’s sperm with her egg may be particularly repellent to the pregnant woman. Would it be worse or better if the rapist had managed to clone his embryo and inserted it into the unwilling woman so there would be no mixing in the embryo of the woman’s and rapist’s genetic material? In the latter case, the woman who seeks an abortion would not be killing her own offspring. [^]
  19. I thank Thomas Douglas and two anonymous reviewers for comments on earlier versions of this article. I also thank Margaret Collins for help in making the article more accessible. [^]


Finnis, J. 1973. ‘The Rights and Wrongs of Abortion: A Reply to Judith Thomson’, Philosophy & Public Affairs, 2: 117–45

Kamm, F. M. 1992. Creation and Abortion (New York: Oxford University Press)

Kamm, F. M. 2008. ‘Terrorism and Intending Evil’, Philosophy & Public Affairs, 36: 157–86

Kamm, F. M. 2013. ‘Creation and Abortion Short’ in Bioethical Prescriptions (New York: Oxford University Press)

Scanlon, T. M. 2010. Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Harvard University Press)

Thomson, J. J. 1971. ‘A Defense of Abortion’, Philosophy & Public Affairs, 1: 47–66

Thomson, J. J. 1973. ‘Rights and Deaths’, Philosophy & Public Affairs, 2: 146–59

Thomson, J. J. 1999. ‘Physician Assisted Suicide: Two Moral Arguments’, Ethics, 109: 497–518