Self-Defence and the Conflict of Rights
The clearest example of justified homicide recognized in law and in morality is that of killing in self-defence.1 It is surely permissible for me to kill someone if that is what is necessary to defend myself from her potentially fatal attack on me. Selfdefence features to some extent in Thomson’s reasoning when she remarks that ‘if anything in the world is true, it is that you do not commit murder, you do not do what is impermissible, if you reach around your back to unplug yourself from that violinist to save your life’.2 Likewise, Thomson is sure that a pregnant woman is permitted to kill a fetus-person to save herself from an otherwise fatal pregnancy.
But legitimate self-defence killing is subject to some very important conditions. The English law of self-defence sets down two main tests for justified
1 See, in English law, the Criminal Law Act 1967, s 3, and the Criminal Justice and Immigration
Act 2008, s 76.
2 Judith Jarvis Thomson, A Defense of Abortion’ (1971) Philosophy and Public Affairs 1, 52.
Arguments about Abortion: Personhood, Morality, and Law. First Edition. Kate Greasley. © K. Greasley 2017. Published 2017 by Oxford University Press.
self-defence: necessity and proportionality. Harming another person in selfdefence must be necessary in order to avoid physical harm threatened by him, and the harm inflicted in self-defence must be reasonable, meaning proportionate to the harm threatened. I am not justified in stabbing someone to death in selfdefence merely to prevent him from slapping me or pulling my hair. Importantly, though, proportionate need not mean equal. As Jane English highlights in support of the self-defence analysis of abortion, it is not the case that you may inflict death only to avoid death, or inflict a black eye only to avoid a black eye.  Rather, as she puts it, ‘.. . our laws and customs seem to say that you can create an injury somewhat, but not enormously greater than the injury to be avoided’.5 To fend off an attack as serious as rape, you may inflict something as serious as the loss of a finger, and to avoid having your clothes torn, as English says, you can blacken an eye. Other than this, the basic rule in English law is that the injury one inflicts must be the minimum necessary to deter or incapacitate the attacker, with a little leeway to account for calculations made in the heat of the moment.
Despite a certain degree of latitude in the proportionality condition, it still seems that a pregnant woman who aborts will struggle to meet it in any scenario where her life is not put at risk by the pregnancy. Some might doubt that killing another person can ever be proportionate self-defence for any reason other than to avoid risk of death. At the very least, the injury to be avoided must be extremely serious. There is a question, then, over whether the risks associated with normal pregnancy are serious enough to qualify.
Furthermore, there might be particular problems meeting the necessity requirement in the case of late-term abortions. Before the fetus is viable, it is true that aborting and thus killing the fetus is the only way of avoiding whatever physical harms the pregnant woman would be risking through continued pregnancy. Postviability, however, the necessity part of the test is in doubt if the pregnancy can instead be brought to an end by safely extracting the fetus alive. Killing the fetus is not strictly necessary in such circumstances, so an important limb of the selfdefence test will not be met.
Responding to the proportionality issue, English acknowledges that the threat posed by the fetus-person would need to be sufficiently serious before defensive killing were justified. As she sees it, though, the threats to a woman’s wellbeing, life prospects, and physical and mental health that can unexceptionally accompany pregnancy are comparably serious to other harms the avoidance of which, on her view, warrants killing in self-defence, such as losing a finger, or a serious beating.7 Whether the normal physical burdens of pregnancy are truly comparable to these sorts of harms is certainly debatable. Pregnancy and childbirth can no doubt be perilous, but in countries where standards of maternal care are good, few pregnancies involve serious threats to the pregnant woman’s health or life.
This said, the physical burdens of pregnancy and childbirth that arise in perfectly typical pregnancies are by no means trivial. They include, among other things: nausea and vomiting, bloating and indigestion, fatigue, weight gain, itching, swelling, sleeplessness, difficulty moving (in the later stages), and, finally, the severe pain of labour, which not infrequently results in tissue damage, and the recuperation process. It is difficult to say whether these burdens are truly tantamount to losing a finger or to a ‘serious’ beating, as English supposes. They are, I think, less serious in the respect that they do not entail the permanent loss of any function or lasting, debilitating damage. This is a far cry from their being negligible. But the question, remember, is whether they are comparable to the sorts of threats which legitimate defensive killing, the most extreme self-defensive measure.
However, it may be thought that it is far more so the threats to wellbeing, future happiness, and life fulfilment presented by the prospect of unwanted motherhood (or that of surrendering one’s child for adoption) that are the truly ominous threats posed by unwanted pregnancy. The problem is that these are not at all the kinds of burdens that people are ever thought justified in using force, let alone lethal force, in order to avoid. A competitor applicant to my dream job who is more likely to prevail might pose an extremely serious threat to my future happiness and wellbeing, but I am not entitled to kill her in self-defence. Nor can romantic rivals claim that killing their competitor is reasonable self-defence because of the perceived threat they pose to life fulfilment. The main burdens of unwanted pregnancy, serious though they are, are simply not of the sort that ever justify killing in self-defence. This is brought out even more sharply when considering the fact that born children can pose exactly the same kinds of threats to their parents. But a parent is not taken to act in legitimate self-defence if she kills her children to live a less encumbered and better life, or because the alternative of having her children adopted entails too much emotional distress.
So far, then, only an abortion carried out to prevent death or serious injury to the pregnant woman—what I will call ‘therapeutic abortion’—is even a candidate for justified killing in self-defence, if the fetus is a person and if abortion is homicide. But there are further possible conditions on justified self-defence which raise problems even for therapeutic abortion. First, unlike aggressors in typical self-defence scenarios, the fetus is an innocent party that cannot help its physical imposition on the pregnant woman. English claims that this is immaterial, since it is clearly permissible to kill even an innocent attacker in proportionate selfdefence. Suppose, she argues, that a mad scientist had hypnotized some innocent people to jump out of bushes and attack passers-by with knives.8 Surely, if you are attacked, you have the right to kill your attacker to save yourself from death or serious injury, though his actions are not blameworthy. It is largely thought,
moreover, that English criminal law recognizes the right to defend oneself against an innocent attacker, with fatal force if necessary.
But there may be something amiss with English’s analogizing the fetus in the therapeutic abortion to the hypnotized knife attackers. For all their innocence, the attackers are still in the role of aggressors. They are assaulting passers by, and their attacks are patently unjust. But to describe unwanted pregnancy as an ‘attack’ by the fetus upon the pregnant woman, even a faultless or unwitting one, seems somewhat strained. As Joel Feinberg argued, the bezerk assailant may be innocent, but he is still an assailant.  The fetus, on the other hand, is not just innocent; it is not clearly an aggressor. He writes:
It did not start the trouble in any fashion. Thus, it would seem that while we are justified in killing an innocent assailant if this were the only way to prevent him from killing us, it does not follow that we are similarly justified in killing a fetal person, since, unlike the innocent aggressor, the fetus is not an aggressor at all.11
One possible reason for our difficulty with viewing the pregnancy situation as an attacker-victim scenario is that the fetus is not inflicting, or trying to inflict, any force on the pregnant woman, a paradigmatic feature of justified harming in selfdefence. To be sure, its presence affects her body quite dramatically, but these effects arise only as a result of its situation, not its exercise of any agency, faultless or otherwise. Although pregnancy has serious physical effects, Feinberg’s claim is that the fetus is simply too passive to be in the role of an innocent aggressor. It cannot do anything other than merely exist, so it surely cannot aggress.
Even if the fetus is not in the role of an attacker, some will still be inclined to accept Thomson’s assertion that a pregnant woman does not have to wait passively while it unwittingly kills her. However, therapeutic abortions are rarely performed by the pregnant woman herself; they are performed by third party medical practitioners. The fetus’s presence is not threatening them. How, then, can they avail themselves of the self-defence justification? This is another possible point of disan- alogy with the violinist situation, where the question is whether you may unplug yourself from him, not whether someone else may do it for you.
English dismisses this problem by arguing that when a doctor performs an abortion she is acting as the agent of the pregnant woman. If the woman has the right to deadly self-defence, she can just as readily transfer this right to her agent if she cannot safely abort herself. In English law, certainly, actions justified by self-defence can extend to those carried out in the defence of others.12 But this is an unsatisfactory answer to the third party problem if one is presuming that the right to life of the fetus and of the pregnant woman are equally strong, and that the fetus is not an aggressor. If the lives are pitted against each other, why should the third party not choose the life of the fetus instead and act on its behalf? Brute hiring power cannot be the tiebreak. Moreover, if there is nothing to choose between them at all, and if, all other things being equal, it is better to allow death than to kill, surely the practitioner should sooner allow the woman to die, thus saving the fetus (where that is possible), than kill the fetus to save the woman.
Like English, Thomson thinks that this problem of choice is a false dilemma. For her, the deadlock between the two innocents is resolved quite quickly once we appreciate that the woman has a prior claim to her own body. As she says several times, ‘the body that houses the child is the mother’s bodyl13 Thomson thinks it plain that a ‘third party who says [of the fetus and woman] “I cannot choose between you” is fooling himself if he thinks this is impartiality’.!4 She compares the scenario in which Smith and Jones are both fighting over a coat they need to keep themselves from freezing, even though the coat in fact belongs to Smith. It is not impartiality, she claims, to say: ‘I cannot choose between you’ if Smith owns the coat! It makes all the difference for Thomson that the body belongs to the woman, not to the fetus. But how exactly is the woman’s ownership of her own body supposed to establish that the fetus can be killed on her behalf?
The Smith and Jones example provides an interesting insight into Thomson’s thinking here. Like Jones’s possession of Smith’s coat, the fetus’s unwanted presence in the woman is, in her view, a bare violation of the woman’s rights, whether or not it has yet cost her anything, equivalent to trespass on her property. Since the fetus has no right to be in her body in any case, it is legitimate for the third party to choose the woman over the fetus and help her evict it. Thomson’s reliance on the pregnant woman’s ‘just, prior claim’!6 to her own body is a foundational premise which is subsequently used to underpin her further claims that not only can the pregnant woman and the third party choose her life over that of the fetus, but that she and her agents are also entitled to kill the fetus for a ‘less weighty reason than preserving her own life’.!7 In fact, since she owns the house (her body), Thomson believes that the woman needs no special reason at all to eject the fetus, or to have it removed. To do so is straightforwardly within her rights, given that the fetus’s presence there is an unjust invasion of her property.
The problem here is that Thomson seems to be relying on a prior ordering of the respective rights of woman and fetus when characterizing the fetus’s unwanted presence in the womb as, to begin with, inherently unjust. Why is the fetus behaving unjustly simply by being where it is? Why is it trespassing? The answer to this cannot be ‘because the woman in whose body it resides has the right to have it detached, even if this spells its death’—whether this is so is what is being debated. Thomson answers the question whether the fetus has the right not to be killed by pointing to the woman’s prior title to her own body. But we might equally direct the inquiry the other way. We might ask whether the fetus, in all the circumstances, has the right not to be killed, and from that deduce whether the woman’s title to her body entails the right to kill the fetus in order to be free of it. Just as, as Thomson asserts, a person’s
- 13 ibid. My emphasis.
- 14 ibid 53.
- 15 ibid.
- 16 ibid 54.
- 17 ibid.
right to life might not include or entail the right to be supported by another’s body at all costs, it is not a foregone conclusion that the moral content of the right to one’s own body includes the right to kill a fetus-person inhabiting it in any given circumstance. Whether the fetus (if a person) has the right not to be killed must surely take account of what it will cost the woman to let it live and what is proportionate.
John Finnis has argued that, in likening the fetus to a trespasser, Thomson implies something much stronger than that the woman ‘is not under a strict duty to allow it to stay under all circumstances’ (which would follow from the simple self-defence argument).  She relies, he says, on a different proposition: that the fetus is under a duty not to enter or stay in the woman’s womb. This he takes to be implied by her claim that the fetus’s mere presence in the womb is unjust, the same way that it is unjust for Jones to take Smith’s coat. Finnis thinks this analysis comfortably fits the violinist scenario, even though the violinist is innocent. As he describes it, ‘our whole view of the violinist’s situation is coloured by this burglarious and persisting wrongfulness of his presence plugged into his victim.’i® But he maintains that it cannot be reasonably said of the fetus that it was under a strict duty not to be where it is and is in breach of such a duty by remaining. It is indeed difficult to believe that the fetus is in breach of a duty simply by coming into existence and remaining alive in the womb. For how could it help coming into existence and where else could it conceivably be? In one clear sense, Finnis thinks that the fetus does have a right to be where it is. This is the sense that it is not under a strict duty not to be there. If this is correct, its very existence is not a bare violation of the woman’s proprietary right to her body. Indeed, if the source of the woman’s proprietary right in her own body is merely that she exists in it, Finnis asks why the fetus might not lay claim to the same proprietary right for exactly the same reason.
One challenge Finnis faces here if he wishes to preserve the possibility that it is justified for you or your agent to kill the violinist (if unplugging him does amount to killing him) is that the violinist is no more at fault in being in his situation of bodily dependency than the fetus is in being in the womb. The violinist was, after all, plugged into you whilst unconscious, and cannot help his rare ailment. Finnis acknowledges the violinist’s faultlessness, but maintains in response that his duty not to be in his current situation—hooked up to your kidneys—is not a fault-based duty. Why, then, not so of the fetus? As we saw, Finnis relies at least partly here on our likely judgement that the violinist’s dependency is of an essentially ‘burglarious’ nature, regardless of fault, whereas the fetus’s presence in the womb is not. But what underpins that differing judgement?
Duties not to be a certain kind of imposition, even faultlessly, are sometimes intelligible. English’s hypnotized knife attackers might be an example. Although the attackers are not in any way at fault, it makes sense to think of them as nevertheless being in breach of the negative duty not to threaten others. It could be argued that it is equally intelligible for the violinist to be under a strict duty not to impose himself on your body. It is at least conceivable that he could avoid being plugged into you, even though in reality he was not at fault. But it is far less easy to make sense of the claim that all human beings begin their existence in prima facie violation of a strict duty not to be where they are. Admittedly, part of what makes this analysis seem so strange is the sheer normalcy of pregnancy and its inextricability from human existence. Can it be possible that all human beings risk violating strict duties to be elsewhere at the very outset of their existence? Given that the only alternative entails their non-existence, this is a difficult proposition to accept.
Of course, one can instead take the position that there is no self-defence justification for killing the violinist unless he poses a threat to your very life, and that third parties may not assist with the unplugging even then, since they may not choose between two faultless antagonists. However, if we agree with Finnis that the very nature of the violinists imposition on you is ‘burglarious’, albeit innocent, it follows that a third party can favour you over the violinist if your life were endangered, and perhaps also that you or that third party can unplug him to avoid less harm to you than that. If, on the contrary, the fetus’s presence in the woman is not clearly in violation of any duty—is not an instance of trespass—then it remains far from obvious that a third party may, in his capacity as the woman’s agent, abort the fetus even to save her life.
But if abortion is presumed to be homicide, a greater problem remains for the self-defence justification as regards non-therapeutic (and especially non life-saving) abortions. This is that rights violations do not neutralize proportionality constraints on permissible defensive action. Let us assume that the pregnant woman’s ‘prior claim’ to her own body does mean that the title is hers exclusively, and that the fetus is in violation of that right by being where it is against her wishes. This is still not enough on its own to justify her killing the fetus-person regardless of proportionality, as Thomson seems to believes that it is. Mary Anne Warren put this point well:
Mere ownership does not give me the right to kill innocent people whom I find on my property, and indeed I am apt to be held responsible if such people injure themselves while on my property. It is equally unclear that I have any moral right to expel an innocent person from my property when I know that doing so will result in his death.
Owning a property does not give me the moral right to eject a trespasser from it if this means throwing him out to his certain, immediate death, unless his presence places a proportionately serious risk upon me. I am not, for instance, morally entitled to throw an intruder out of the safety of my home to a pack of ravenous wolves outside, simply because the title to the home is mine and I am not in the mood for guests. Indeed, when we look closer at the self-defence justification, it is entirely clear that the initial violation of the defender’s right cannot possibly preclude all proportionality constraints. If this were the case, proportionality constraints would never apply to self-defence at all, for the initial violation is a standing feature of the self-defence scenario. There always is an unjust threat to the justified defender, and still the law—and morality too—demands moderation. Given the usual proportionality requirements, it is questionable whether normal pregnancy would meet the conditions for self-defence killing, whether or not the third party problem could be surmounted.
-  Gladstone Williams  3 All ER 411; Beckford  3 All ER 425; Whyte  3 All ER416; Clegg  2 All ER 334; Owino (1996) 2 Cr App R 128.
-  Jane English, Abortion and the Concept of a Person’ (1975) 5 Canadian Journal of Philosophy233, 237.
-  ibid 237. 4 Palmer  AC 814 (Privy Council) (Lord Morris).
-  7 English (n 4) 237—8.
-  See George Fletcher, ‘Proportionality and the Psychotic Aggressor: A Vignette in the ComparativeCriminal Law Theory’ (1973) 8 Israel Law Review 376.
-  Joel Feinberg, ‘Abortion’, in Joel Feinberg, Freedom and Fulfillment: Philosophical Essays (PrincetonUniversity Press 1994) 62-3.
-  ibid. I2 Thomson (n 2) 54.
-  John Finnis, 'The Rights and Wrongs of Abortion: A Reply to Judith Thomson’ (1973) [Wiley]2 Philosophy and Public Affairs 141.
-  ibid.
-  Mary Anne Warren, ‘On the Moral and Legal Status ofAbortion’ (1973) 57 The Monist 43—61.