Hostname: page-component-cd9895bd7-7cvxr Total loading time: 0 Render date: 2024-12-26T04:48:51.262Z Has data issue: false hasContentIssue false
  • English
  • Français

Killing, Letting Die and Preventing People From Being Saved

Published online by Cambridge University Press:  26 January 2009

Matthew Hanser
Matthew Hanser
Affiliation:
University of California, Santa Barbara, [email protected]
Get access Rights & Permissions [Opens in a new window]

Abstract

The distinction between killing and letting die is too simple. A third category – preventing people from being saved – must also be recognized. Like killing, preventing a person from being saved is a species of doing harm; like killing, it infringes one of the victim's negative rights. Yet preventing a person from being saved is morally on a par with letting die, which infringes one of the victim's positive rights. It follows that we cannot explain the moral inequivalence of killing and letting die by saying, as so many have, that negative rights are more stringent than positive rights. A more promising strategy is suggested at the end of the article.

Type
Articles
Information
Utilitas , Volume 11 , Issue 3 , November 1999 , pp. 277 - 295
Copyright
Copyright © Cambridge University Press 1999

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Drawing the distinction in accordance with this assumption might require departing from ordinary usage, but then ordinary usage does not always carve up the world exactly along analytically significant lines. It is for precisely this reason that many writers prefer to introduce artificial names for the categories at issue. Thus Warren Quinn, who defends the distinction's significance, prefers to speak of a person's agency as being either positive or negative in respect of a harm; and Jonathan Bennett, who denies the distinction's significance, speaks of a person's being either positively or negatively instrumental in the production of a state of affairs. (See Quinn, Warren, ‘Actions, Intentions, and Consequences: The Doctrine of Doing and Allowing’, Philosophical Review, xcviii (1989), 291Google Scholar; and Bennett, Jonathan, ‘Morality and Consequences’, The Tanner Lectures on Human Values, vol. 2, ed. McMurrin, S., Salt Lake City, 1981, pp. 47–52Google Scholar.) That Quinn is implicitly guided by the exhaustiveness assumption is evident from the fact that without comment he draws his distinction between positive and negative agency so that the two categories jointly cover every case in which an agent's behaviour helps explain someone's suffering a harm (Quinn, 301 f); and Bennett explicitly states that he wants his distinction ‘to include every case where a person's conduct makes him in some way and to some degree responsible for a given state of affairs’ (Bennett, p. 48). Indeed, one of the reasons Bennett thinks the ordinary language distinction between doing and allowing is unsuitable for purposes of moral theorizing is that ‘the two are not jointly exhaustive of the genus “prima facie responsibility”’ (ibid., p. 49). (It should be noted that Bennett's distinction between positive and negative instrumentality allows for the theoretical possibility of agency falling precisely on the line dividing the two. But such cases would never arise in practice.)

2 Note, however, that since death is not always a harm, killing is not always an instance of doing harm, nor letting die an instance of allowing harm to occur.

3 See, for example, Quinn, 306.

4 Perhaps a more natural example of a case in which an agent prevents someone from having his life saved would be one in which one agent interferes with another agent's attempt to save someone's life. In such cases the agent who prevents the person from being saved stands in morally significant relations to two people: the person who dies, and the person whose beneficent agency he thwarts. In paradigm cases of killing and letting die, by contrast, the agent stands in a morally significant relation to just one person: the person who dies. This, however, is not the difference that makes me want to treat preventing people from having their lives saved as a distinct, third category. To highlight what I think is the more fundamental difference between killing and preventing people from being saved, then, I focus on cases in which an agent prevents someone from being saved without thereby thwarting anyone else's beneficent agency. (Readers who find my way of describing these cases unnatural should feel free to regard me as giving the words an artificial sense. My interest is in how various acts can be usefully grouped for theoretical purposes, not in how ordinary language would describe these acts.) I should also stress that, as I am using the expression, an agent does not ‘prevent someone from being saved’ unless that person's being saved was actually in the offing – unless he would have been saved had not something (such as the agent's act) prevented this from happening.

5 Bennett, pp. 52 f.

6 Suppose X poisons Y with a poison that will cause him to die in six months. When Z hears that Y has been poisoned, he abandons his own plan to shoot and kill Y, which he had meant to put into effect that very day. As a result of being poisoned by X, then, Y is prevented from being killed by Z, and indeed from dying at all for another six months. Does it follow that Y receives a benefit as a result of being poisoned by X? I think it does. But his death from being poisoned still comes to him as a harm. By poisoning Y, X kills Y, and so in my view harms him; and the fact that Y also receives a benefit as a result of this act does not diminish the act's moral objectionability. (If poisoning Y had been the only way to prevent Z from killing Y, and if X had done this in order to prolong Y”s life another six months, then perhaps his act of harming Y would have been justified. But that is another issue.)

7 There has been much uncertainty as to whether actively disconnecting a patient's life support machine should count as killing him or as letting him die. I would classify it as preventing him from having his life saved. It does not follow, of course, that all such acts are equally objectionable – it matters greatly who disconnects the machine, for what reason, and by what authority. When an agent's intention in disconnecting a life support machine is to produce the patient's death, and when the agent's motive for producing the patient's death is bad, it is natural to describe the agent as killing the patient. We need not put too much weight on this fact, however. I suspect the point of calling such acts killings is to call attention to their extreme moral badness; it does not follow that we must classify them as killings for purposes of moral analysis. I shall discuss why some cases of preventing people from having their lives saved are just as objectionable as paradigm cases of killing, and why others are no more objectionable than paradigm cases of letting die, in section II. One caveat: if a patient's condition has deteriorated to the point that death will no longer come to him as a harm, then the prolongation of his life will no longer come to him as a benefit; and if having his life saved will not come to him as a benefit, then an agent who disconnects his life support machine does not do him the harm of preventing him from receiving the benefit of having his life saved. The act could of course still be objectionable on other grounds.

8 Contrast A Car, a Rock, and a Bomb with a case in which someone else would have moved the rock from the path of Y”s car if X had not. As before, Y would have died anyway even if X had not acted as he did. This time, however, the objection to X's behaviour is not simply that he risks harming Y. The difference is that in this case, having his car stopped by the rock would have come to Y as a benefit: it would have prevented his car from going off the cliff, and so would have saved his life. By moving the rock, then, X prevents Y from receiving this benefit, and so harms him. This is so no matter how certain it was that someone else would move the rock if X were not to do so first.

9 Here and later in this article I speak both of the moral equivalence of individual actions and of the moral equivalence of types of action. These usages are connected. When I say that two types of action are morally equivalent, I mean that, other things being equal, individual acts of these types are equally objectionable. When I say that two individual acts are morally equivalent, I mean both that they are of equivalent types and that the acts are in fact equally objectionable (i.e., that other things are in fact equal).

10 This example is structurally analogous to the much discussed ‘trolley problem’. The main difference is that in my example the agent is in a position to redistribute a benefit rather than a harm. (I use the word ‘redistribute’ broadly, to encompass any case in which an agent, in providing a benefit to one party, prevents another party from receiving a comparable benefit. If bringing life-saving medicine to five people were to necessitate dislodging the rock from the path of y's car, I do not think this would alter the moral structure of the example.) The trolley problem was introduced by Philippa Foot and has generated a substantial literature. See Foot, , ‘The Problem of Abortion and the Doctrine of the Double Effect’, originally in Oxford Review, v (1967)Google Scholar, repr. in her Virtues and Vices, Berkeley and Los Angeles, 1978, p. 23 of the reprinted version. See also two papers by Thomson, Judith Jarvis: ‘Killing, Letting Die, and the Trolley Problem’, Monist, lix (1976)Google Scholar; and ‘The Trolley Problem’, The Yale Law Journal, xciv (1985)Google Scholar.

11 The moral inequivalence of killing and preventing people from having their lives saved may in turn help explain why there is a moral difference between ducking harm and using someone as a shield. If an agent who is being shot at in a crowd wants to avoid being killed, it is permissible for him to duck, thereby exposing the person behind him to the gunman's fire, but it is not permissible for him to grab the person behind him to use as a shield. Perhaps the moral difference between these acts can be explained by the fact that the former involves preventing the bystander from having his life saved (the bullet's hitting the agent would have prevented it from hitting the bystander), whereas the latter involves killing the bystander. This puzzle was introduced by Boorse, Christopher and Sorensen, Roy in ‘Ducking Harm’, Journal of Philosophy, lxxxv (1988)Google Scholar.

12 Even if it turns out that I am wrong about the moral equivalence of preventing people from being saved and letting people die, there is still reason to think that the moral difference between killing and letting die cannot be accounted for entirely by means of an appeal to the negative/positive distinction. I shall return to this point at the end of the present section.

13 Some might consider this moral equivalence a ground for classifying instances of preventing people from having their lives saved as instances of letting die, but there are powerful reasons for not doing this. Letting die is clearly a species of allowing harm to occur, and I have argued that preventing people from being saved is a species of doing harm. Furthermore, it is central to the concept of letting die that it involves failing to prevent death, and it is simply not plausible to describe every agent who actively prevents someone from being saved as thereby failing to prevent that person's death. It may be permissible to stretch our ordinary concepts a little when constructing a moral theory, but the concept of letting die will not stretch this far.

14 I say that the negative right not to be killed comprises by itself an unvalenced right to continued life because I do not believe there is a positive right to be provided with continued life; and I do not believe there is such a right because I do not see how any agent could provide an already extant person with continued life. Agents have the power to deprive people of continued life, or to help prevent them from being deprived of continued life, but strictly speaking no agent has the power to give someone continued life. Of course, speaking loosely, we might say that an agent who tosses a life preserver to a drowning swimmer thereby ‘provides’ that swimmer with the rest of his life. But strictly speaking we must say that the agent does no such thing. The agent (merely) provides the swimmer with something he needs in order to be prevented from losing his life. (It could be argued that a person's parents provide him with the good of life by causing him to exist, but even if this is right, the offspring could not have had an antecedent right to be provided with this good, since necessarily, prior to his receiving this good, the offspring did not exist.)

15 It should be noted that there is nothing consequentialist about this hypothesis. I do not claim that the outcome of violating someone's right to continued life is worse than the outcome of violating someone's right to life-saving aid.

16 I would like to thank Jeff McMahan and Christopher McMahon for their helpful comments on an earlier version of this paper.