Killing is justifiable if it is necessary. If killing is unnecessary, it is unjustifiable and therefore wrong. Then, a question is: When is killing necessary or unnecessary? Oberman addresses this question in relation to the act of other-defence: When is killing someone necessary or unnecessary to defend others against death?
A standard answer is this: killing a person is necessary if there is no alternative way to save others from death besides killing that person, whereas killing is unnecessary if there is such an alternative. However, in fact, in some cases, whether killing someone is necessary or unnecessary to save others is more difficult to answer than we normally think. Compare the following two cases:
Case (A): There is Attacker. Attacker will kill Victim. You can save Victim now by killing Attacker. But you can also save Victim by waiting a few minutes until Attacker turns her back.
Oberman calls this a ‘standard case’ because the conclusion is the same no matter which of the views to be discussed below is adopted: killing Attacker is unnecessary in this case.
Case (B): There is Attacker. Attacker will kill Victim. You can save Victim only by killing Attacker. But there is also Hiker bitten by a snake. You can save Hiker only by driving her to hospital immediately. You can save only one of them.
Oberman calls this a ‘problem case’ because whether killing Attacker is necessary or unnecessary in this case depends upon which of the views to be discussed below is adopted.
In case (A), it seems clear that killing Attacker is unnecessary to save Victim from death. You could wait a few minutes. But in case (B), whether killing Attacker is necessary or unnecessary depends upon which of the following views we adopt. On the standard view, killing Attacker is necessary because there is no alternative way to save the specific individual – Victim – from the specific threat – the one posed by Attacker – other than killing Attacker. But Oberman defends a different view: what he calls the ‘revisionary view’ of necessity. On this view, the just end for other-defensive killing is ‘saving a life’, not ‘saving a specific person’s life (e.g. Victim’s life)’. You could achieve this end by saving Hiker. Therefore, there is an alternative to killing Attacker (i.e. saving Hiker). And therefore, killing Attacker is unnecessary.
In short, there are two different views regarding the necessity of killing. On the standard view, the just end for other-defensive killing is saving a specific person’s life from a specific threat. Therefore, on this view, killing is unnecessary if there is an alternative that the rescuer can choose to save that specific person from that specific threat. Meanwhile, on the revisionary view, the just end for other-defensive killing is saving a life from a threat, not saving a specific person’s life from a specific threat. Therefore, on this view, killing is unnecessary if there is an alternative that the rescuer can choose to save a specific person from a different threat, or to save a different person from a different threat.
1. What is a just end for other-defensive killing?
To defend the revisionary view, Oberman first identifies the end for other-defensive killing that can be considered just. The choice is among (a) ‘saving a specific person’s life’, (b) ‘saving a specific person’s life immediately’, and (c) ‘saving a life’.
According to Oberman, it is inappropriate to choose (b). Consider case (A) above again. In that case, the rescuer could either kill Attacker to save Victim now or wait a few minutes to save Victim later. A just end for other-defensive killing must be the one that is, from the rescuer’s standpoint, so important that it is worth killing for. Is ‘saving Victim now’ in case (A) worth killing Attacker for? The answer is ‘No’ because you could wait a few minutes to save Victim later.
The choice is, then, between (a) and (c). Consider case (B) above. In that case, should the rescuer prioritise Victim’s life over Hiker’s? The rescuer may have a reason to prioritise Victim’s life if she has a special connection with the latter (e.g. if Victim is the rescuer’s daughter, etc.). But if Victim and Hiker are both strangers to the rescuer, there is no reason for the rescuer to prioritise one over the other. Both deserve equal respect in the rescuer’s moral judgement.
Taking these points as well as several exemplary cases into consideration, Oberman argues that it is (c) ‘saving a life’ that can be regarded as the just end for other-defensive killing.
2. Is ‘saving Victim from the specific threat posed by Attacker’ a just end for defensive killing?
On the standard view, the just end for other-defensive killing is (a) saving a specific person (e.g. Victim) (b) from the specific threat that the person is subject to (e.g. the one posed by Attacker). We have already seen why (a) is untenable: the just end for other-defensive killing is not ‘saving a specific person’s life’ but ‘saving a life’. What about (b)? Oberman argues that there is a case that indicates (b) is also untenable. Consider the following case:
Victim has had a stroke that affects the left side of her body. She will be paralysed on her left side unless you inject her with chemical L immediately. You are about to inject L when Attacker comes in to inject R – another chemical that will paralyse Victim on her right side. You can prevent Attacker’s injection of R only by killing her. Now, suppose you have only two options: you can either inject L now, allowing Attacker to inject R, or prevent Attacker’s injection of R now, giving up your injection of L. If you take the first option, Victim’s left side will be saved from paralysis by your injection of L. But her right side will be paralysed by Attacker’s injection of R. Meanwhile, if you take the second option, Victim’s right side will not be paralysed, because the stroke has no effect upon that side, and because Attacker is killed and prevented from injecting R. But her left side will be paralysed due to the stroke.
On the standard view, killing Attacker in this case is necessary to save Victim from the specific threat posed by Attacker – the injection of R which will paralyse the right side of her body. There is no way to prevent that threat besides killing Attacker. However, saving Victim from being paralysed on her right side is not what we really care about – in particular, if Victim does not care whichever side will be paralysed. The just end for defensive killing in this case is broader than that: what we really care about is saving Victim from being paralysed on one half of her body. The rescuer can achieve this end without killing Attacker if she chooses the first option; so, killing Attacker in this case is unnecessary. This also indicates that, in certain cases, we can achieve a just end without preventing the specific threat posed by Attacker. Therefore, (b) fails.
3. Is there any plausible argument for the standard view of necessity?
So far, Oberman has presented a case for the revisionary view of necessity. But is there any plausible case for the standard view? Oberman addresses several lines of argument for the standard view and concludes that none of these is persuasive. For example, he refutes the following argument in the following way.
(1) It is more important to save a life from wrongful killing (e.g. Victim’s life) than it is to save a life from natural misfortune (e.g. Hiker’s).
(2) If premise (1) is true, then ‘saving a life from wrongful killing’ can be considered to be a distinct end that justifies defensive killing, separate from ‘saving a life from natural misfortune’.
(3) In problem cases like case (B) above, there is no way to save a life from wrongful killing other than defensive killing.
Premises (2) and (3) together lead to Conclusion: In problem cases, killing is necessary.
In response, Oberman argues that, even if premise (1) is true, it does not necessarily support premise (2). There are at least two ways to defend premise (1). First, wrongful killing may be bad for Attacker. It may make Attacker’s life worse. Second, wrongful killing may be bad for Victim. It not only deprives Victim of her life, but may also undermine her honour. However, premise (1) does not necessarily support premise (2). ‘Saving a life from wrongful killing’ can be more important than ‘saving a life from natural misfortune’; but it can be important without being so important as to justify defensive killing. Oberman writes: ‘Only things that are extremely valuable will be worth killing for’.
The question, then, is: Is ‘saving a life from wrongful killing’ so important as to justify defensive killing? Let’s consider the first argument first. Wrongful killing may be bad for Attacker; so, preventing it (by saving Victim’s life from it) may be important. If so, it may make Attacker better off if the rescuer punches or kicks Attacker to prevent her from killing Victim. However, it seems hard to say that Attacker would be better off if she were preventatively killed. Therefore, ‘saving a life from wrongful killing’ is not so important as to justify defensive killing by the rescuer.
Let’s consider the second argument. Wrongful killing may be bad in the light of Victim’s honour. Can this be a reason to justify defensive killing by the rescuer? ‘Perhaps’, Oberman writes, ‘honour is something that a victim can only defend for herself; others cannot defend it for her’. This is because defending one’s honour is about asserting oneself as an active agent, as opposed to a passive object. The beneficiaries of other-defence are not asserting their agency (but, instead, displaying their dependence upon others for survival). If so, defensive killing by the rescuer does not serve Victim’s honour. Also, perhaps, ‘Victim has already resisted Attacker with considerable force. Although Victim will die, no one will doubt she has asserted herself as an agent’. Or, perhaps, Victim and everyone who matters to her believe that Victim should not be treated in the way Attacker treats her; that the way Attacker treats Victim undermines Victim’s honour irrespective of whether she fights back or not. ‘She knows that and everyone who matters knows that. She has nothing to prove’. If so, defensive killing by the rescuer is unnecessary in the light of Victim’s honour.
4. Is the revisionary view of necessity overly demanding?
The revisionary view requires that the rescuer should not kill if there is an alternative to killing that rescues a life. But what if choosing such an alternative is too demanding either because it is too costly or because it conflicts with one’s professional duty?
If saving lives without killing is too costly, the rescuer is permitted to see it as an irrelevant alternative. She is permitted to kill in order to save lives. But is it too costly to save lives without killing? Empirically, Oberman argues, it isn’t. Oberman indicates this point by comparing the cost of anti-poverty with the cost of military intervention intended to save lives:
‘One does not have to scour the world to save people to rescue; one can donate to anti-poverty programs at a click of a button. The most cost effective anti-poverty programs can save lives from poverty for as little as $3000 to $4000. Killing to save people, by contrast, is often dangerous (since it involves confronting aggressors) and can be financially costly as well (since it can require weapons and other equipment). There is no reason, then, to think that cost can save the standard view.’
Also, one’s professional duty is a relevant consideration too. Some professionals, such as soldiers and the police, may have no relevant alternative to killing because they have duties to perform the roles they have assumed. In this case, their other-defensive killing can be seen as necessary killing. (But we should also note that one could have chosen a career that does not require one to kill to save lives in the first place.)
5. What if one can save both Victim and Hiker?
Consider a variation of case (B) above:
There is Attacker. Attacker will kill Victim. You can save Victim only by killing Attacker. There is also Hiker bitten by a snake. You can save Hiker only by driving her to hospital. You can easily save both.
In this case, killing Attacker is necessary in so far as you will actually save both. If you will save Victim by killing Attacker, but will not save Hiker, then killing Attacker is unnecessary. If you will save both, no matter which you will save first, then killing Attacker is necessary.
6. When is killing necessary?
Finally, Oberman notes that the revisionary view should not be confused with pacifism. He identifies 5 types of case in which, even on the revisionary view, it would be necessary to kill Attacker:
‘1. Partiality. The rescuer may have strong reason to be partial to the victim over other people the rescuer could rescue.
2. Impartial necessity. Even from an impartial standpoint, there is no equivalent threat that could be prevented without killing.
3. Costs. Pursuing alternatives would be so costly to the rescuer that it is permissible for the rescuer to disregard them.
4. Duties. The rescuer has duties that prevent her pursuing alternatives.
5. Reasons beside defence. There are reasons to kill that are not reasons of defence, for instance prevention, deterrence or retribution.’
Our discussion covered questions on: disanalogies that may complicate the application of Oberman’s argument to the practical issues Oberman also discusses in the paper (e.g. the choice between anti-poverty programmes and military interventions to save lives); empirical evidence that may question whether military interventions in the real world are actually intended to ‘save lives’; justifiability v. permissibility of killing (permissible killing and justifiable killing may come apart); and many more. Oberman’s paper and his responses were full of interesting points supported by analytical rigour. The paper will be a great contribution to the relevant literature, and to the way we think about the means to save people’s lives.
Written by Yuki Iwaki
Kieran Oberman is Lecturer in Politics at Edinburgh University.