I. INTRODUCTION
In her classic article ‘The Problem of Abortion and the Doctrine of the Double Effect’ Philippa Foot distinguished between two kinds of allowing:
There is firstly the allowing which is forbearing to prevent. For this we need a sequence thought of as somehow already in train, and something that an agent could do to intervene . . . So, for instance, he could warn someone, but allows him to walk into a trap. [There is secondly the allowing: KLR] which is roughly equivalent to enabling; the root idea being the removal of some obstacle which is, as it were, holding back a train of events. So someone may remove a plug and allow water to flow; open a door and allow an animal to get out; or give someone money and allow him to get back on his feet.Footnote 1
Both kinds of conduct contrast with doing harm, paradigm cases of which consist in ‘initiating or sustaining a harmful causal sequence’. In what follows I shall say that the second kind of conduct ‘enables’, while the first merely ‘allows’, an outcome.Footnote 2 My question is: where the outcome is harm to another, are these cases of non-doing morally equivalent?
Foot thinks that, on some occasions, allowing harm is ‘just as bad as’ doing harm. However, on other occasions it is clear that doing harm is worse than allowing harm, and on no occasions is allowing harm worse than doing harm. Ceteris paribus, then, doing harm is morally worse than allowing harm. Because this claim is not my focus in this article, and because common-sense moral intuitions clearly support it, I shall take it for granted.
This summary of Foot's views points to two questions about enabling harm: is it as bad as doing harm? Is it as bad as allowing harm? The simplest responses here either demote enabling to the moral status of doing (both being worse than allowing) or promote enabling to the moral status of allowing (both being better than doing). A more complicated position locates enabling midway: it is less bad than doing harm but worse than allowing harm. An even more complicated position would be that enabling harm is sometimes worse, sometimes less bad, than doing harm, and similarly so with allowing harm – it has no fixed, comparative moral status of this sort per se.
No philosopher who has addressed these questions has opted for the most complicated answer.Footnote 3 Some – e.g. Vihvelin and TomkowFootnote 4 – have defended the claim that enabling harm is, sometimes at least, as bad as doing harm, while others have argued that enabling harm is intermediate between doing and allowing harm.Footnote 5 Yet, others have claimed that enabling harm is (i) morally on a par with allowing harm, and (ii) less bad than doing harm.Footnote 6 I shall refer to (i) as the moral equivalence thesis. This thesis is the focus of my article. It implies that cases of allowing harm and enabling harm have the same moral status (e.g. one cannot be morally permissible when the other is not) when all other factors, such as the badness of their consequences, are held constant across the relevant cases. For instance, the moral equivalence thesis implies that allowing and enabling harm, even if they are not intrinsically morally different, do not interact differently with other moral factors, such as beliefs.Footnote 7 This means that, as long as all other morally relevant factors are constant, a case of allowing harm in which the agent has a certain belief about the consequences of his conduct and a case of enabling harm in which the agent has the same belief about the consequences of his conduct must have the same moral status.
The main conclusion of this article is that the moral equivalence thesis is false. Section II presents a counterexample to the conclusion of an argument given by Samuel Rickless in favour of the thesis. The counterexample involves self-defence. Also, it raises a doubt about Rickless's premise that enabling harm is morally on a par with allowing harm in the example he offers. Section III considers variants of the case Rickless's argument builds on. It argues that the moral significance of allowing and enabling harm can differ in situations where an agent acts on reasonably formed expectations about how to avoid harm. Section IV shows that certain cases of enabling harm, where it is permissible to enable one person to be harmed to save others, are best explained, not by the truth of the moral equivalence thesis, but by the similarities between these cases and standard trolley scenarios where it is even permissible to harm one to save more. Section V rebuts the way in which friends of the moral equivalence thesis have tried to accommodate recalcitrant moral intuitions by appealing to the distinction between a stable situation and one where causal sequences in train are set to continue. Section VI concludes.
Before we turn to my critique of the moral equivalence thesis, there are three preliminaries. First, the equivalence thesis is significant for the two reasons expounded in this and the following paragraph. The moral inequivalence of doing and allowing harm has become something of a hallmark of deontological thought. This is so, even if, conceptually speaking, deontologists and non-deontologists need not disagree over this issue.Footnote 8 Within deontology there is pressure therefore to explain why doing and allowing differ. By far the most plausible explanation attaches moral significance to the distinction between what the victim of harm has independently of the person who countenances harm (either by doing, allowing, or for that matter enabling harm) and what the victim of harm has only through the person who countenances harm. By way of illustration, consider Judith Jarvis Thomson's famous case of the violinist in need of life support who has been hooked up with another person without this person's consent.Footnote 9 On Thomson's view, the supporter would kill the violinist if she were to disconnect him, yet this would be morally permissible. Frances Kamm suggests that the explanation of killing being permissible in this case is that it is an untypical killing in that the violinist would only lose something that she has through the killer/supporter, i.e. continued life. As Kamm puts it: the baseline relative to which one must not harm the violinist is ‘the condition he would have been in if he had never been attached to his supporter’.Footnote 10 This motivates the following argument: because countenancing harm is less wrong when the person harmed loses only what he has through the agent of harm – as Thomson's case illustrates – and because it is true by definition in cases of allowing harm that the victim loses only what he could have had through the agent who does not assist, but not similarly true by definition in cases of killing, it follows that allowing harm is less wrong per se than doing harm.Footnote 11 However, if enabling and allowing harm are morally equivalent, this argument for the moral significance of the distinction between doing and allowing harm fails. This is so, because in cases of enabling harm, where the enabler of harm is different from the one who provides the obstacle to harm if such an agent exists, the victim is deprived of a good she would have had if the enabler of harm had not been present at the scene. Given this, it seems that in order to offer a plausible rationale for deontology one would have to defeat the moral equivalence thesis.Footnote 12 Hence, the present issue is theoretically important.
A further reason why the equivalence thesis is significant relates not so much to moral theory as to practical, real-life moral concerns. Thomas Pogge has influentially argued that citizens of rich countries do not merely allow global poverty. Rather, they bring it about through the imposition of an international regime skewed in their favour at the expense of the global poor, and accordingly they have much stronger moral obligations to eliminate poverty than it is normally assumed. In Pogge's view we must ‘stop thinking about world poverty in terms of helping the poor’.Footnote 13 In response, Barry and Øverland have submitted that once we substitute the tripartite distinction between doing/allowing/enabling harm for the bipartite distinction between doing/allowing harm with which Pogge operates we see that, typically, citizens in rich countries enable the global poor to be harmed (e.g. through the erection of trade barriers) rather than doing harm to them (e.g. through military interventions). Although they reject the moral equivalence thesis, Barry and Øverland allow that it ‘would undermine quite significantly Pogge's argument that affluent people have particularly stringent moral reasons to address global poverty’.Footnote 14 They also hold that rejection of the thesis does not rescue Pogge's position. They take enabling harm to be a distinct moral category intermediate between doing and allowing harm; and in their view ‘moral reasons associated with contributing to harm through enabling it seem to be less constraining than reasons associated with doing harm’ though more constraining than those associated with allowing harm.Footnote 15 On this basis, with certain additional assumptions, they conclude that Pogge exaggerates the stringency of the duties of affluent citizens to eliminate global poverty.Footnote 16
Second, I have formulated the moral equivalence thesis in comparative terms that require us to see actions as better (less bad) or worse (more bad). However, the bald assertion that one act is worse than another is bound to convey rather different things to different people. One person might detect a Kantian theme: conduct deteriorates, morally, as it fails to track duty and/or is done from motives other than the sense of duty. Another person will simply assume that an action of the worse kind has the following feature: it can be morally impermissible when an otherwise comparable action of the less bad kind is not.Footnote 17 It is this latter sense of badness that I and, so I believe, other people who have written on the topic have in mind. For instance, Rickless appears to have something like this sense in mind when he writes: ‘it is more difficult to justify doing harm than it is to justify allowing harm’.Footnote 18
Third, it is interesting that some enablings of lethal harm are best viewed as killings, while others are naturally described as cases of ‘allowing someone to die’. What criteria guide us here is a very good question,Footnote 19 but not one that I can address in this article.
II. RESCUING AND SELF-DEFENCE
I start by considering an important argument recently given in support of the moral equivalence thesis by Samuel Rickless. Rickless sketches a case in which enabling and allowing harm seem to be morally alike – in fact, the case is a modification of one of a series of cases developed by Jeff McMahan in the course of arguing that at least some cases of lethal withdrawal of aid are morally equivalent to killing (hence the ‘3’):Footnote 20
Burning Building 3
A person trapped atop a high building that is on fire leaps off. Seeing this, a firefighter quickly stations a self-standing net underneath. He then quickly leaves the scene to assist those of his comrades who are battling the blaze. At the same time, Sally, a passer-by, then immediately notices that two other persons have jumped from a window several yards away. She therefore repositions the net so that it catches the two. The first jumper hits the grounds and dies.Footnote 21
According to Rickless, Sally acts permissibly. He claims this supports the moral equivalence thesis.Footnote 22 For while we might ordinarily describe the case as one in which Sally kills the single individual, it is a case of her enabling harm to the single person trapped in the building in Foot's sense. After all, Sally simply removes an obstacle to the process in which the lone individual falls to his death. The largely uncontroversial fact that she can permissibly enable harm to one person to save two (each) from the same harm suggests that the enabling is morally equivalent to allowing harm. It does so because, as participants in the debate acknowledge, one is permitted to allow harm to one person to save two from it. If enabling harm were morally equivalent to doing harm, however, one would expect to conclude that Sally acted impermissibly. Normally it is impermissible to do harm to one person to save two from it. This, I take it, is Rickless's argument.Footnote 23
There are at least two reasons to be sceptical about this argument, one of which is internal to Rickless's own example and one of which is not. The first is this. It is doubtful whether, in Rickless's example, enabling harm really is morally equivalent in all respects to allowing the same harm. If the single victim could somehow paralyse Sally before Sally repositions the net below a second burning building, where there is also one innocent person trapped, this might be permissible. In any case, intuitively, his doing this would be less wrong than the act would be in a slightly different case in which he paralyses Sally, who is at that moment considering whether to position the net below the first or the second building, and this somehow results in the net being placed below the first building.Footnote 24 In this case, he would clearly act impermissibly.Footnote 25 This does not immediately contradict anything Rickless says, but it suggests that it might not be enough, for the purpose of testing the moral equivalence thesis against the case of Burning Building 3, to look merely at the permissibility of repositioning of the net.Footnote 26
Second, this suspicion is strengthened when we consider simpler cases of self-defence not involving a third party trying to maximize the number of people saved. Suppose you and I are both trapped in unowned cages. A hungry bear is outside punching the cage doors. Eventually, it will succeed in opening one of them, eat the person inside, and then, no longer being hungry, depart leaving the other person unharmed. In the first case to consider I can reinforce the door to my cage, thereby allowing the other person to be eaten. In a second case, I can weaken the door of the other person's cage with my metal saw, thereby enabling the other person to be eaten. Intuitively, my conduct in the first case is less wrong than it is in the second.Footnote 27 As Barry, Lindauer and Øverland put it: ‘reasons against enabling harm appear to constrain conduct more with regard to shifting costs from ourselves to others than do reasons against allowing harm’.Footnote 28 Indeed, all things considered, my conduct in the first cage case seems morally permissible, unlike my conduct in the second case. Hence, further support for the case against the moral equivalence thesis can be derived from cases of self-defence.Footnote 29 I conclude that Rickless's argument appeals to a false premise and has a conclusion which, on other grounds, we have reason to think is false. Below I shall offer further reasons for scepticism about both it and the moral equivalence thesis.
III. EXPECTATIONS AND THE ALLOWING/ENABLING HARM DISTINCTION
In this section I will argue, pace the moral equivalence thesis, that allowing and enabling harm interact differently with conduct based on expectations about harm. To do this I shall consider some variants of Burning Building 3.
Consider first the no-enabling harm belief case. In this case the person trapped at the top of the blazing building acts on the reasonable belief that a self-standing net has been positioned next to his building. That is, unlike his counterpart in Rickless's Burning Building 3, he jumps after forming the reasonable belief that the firefighter has placed the net under the building and, as Sally is aware, he has no reason to suppose that Sally might move it. Compare this with the no-allowing harm belief case. Here the firefighter unfolds the net, but then dashes off to attend to something else. The trapped man notices Sally passing by and, noting that Sally has picked up the net, reasonably forms the belief that Sally will position the net under his window. He then leaps, believing that the net will save his life. Again, we can suppose Sally is aware that he holds this belief.Footnote 30
Intuitively, these cases differ. The fact, in the no-allowing harm belief case, that the trapped man reasonably believes Sally will not allow him to die and acts on the basis of this belief clearly does not render it impermissible for Sally to reposition the net when doing so will save two people. By contrast, the fact, in the no-enabling harm belief case, that the trapped man reasonably believes he will save his life by jumping, and that Sally will not enable him to be harmed, and acts on these beliefs, makes it intuitively impermissible for Sally to reposition the net. Sally would not be permitted to remove the net if it were only one person jumping from the other window. Also, it would be permissible for her to allow, say, the wind to move the net to the other window. In this variant of Burning Building 3, then, the enabling of harm does not appear to be morally equivalent to its being allowed – the two sorts of conduct interact differently with reasonably formed beliefs upon which the man threatened with harm acts.Footnote 31
It might be replied that, even if this is so, it remains true that in the basic Burning Building 3 case Rickless presents repositioning the net as permissible. However, in order to defeat the moral equivalence thesis I need not deny – even if, as indicated in the previous section, I think there is a case for denying – that enabling harm is equivalent to allowing harm in cases where the agents involved have not acted on the assumption that a certain obstacle will be in place preventing them from being harmed. All I need claim is that the two are morally inequivalent in cases where the agents involved have acted on such an assumption.
Rickless's appeal to Burning Building 3 could be buttressed, alternatively, by saying the following: my argument draws illegitimate support from the contingent truth that, generally, people in a position to be saved by an observed obstacle to harm that is already in place expect to be saved, while people who might be saved by an obstacle to harm if a person, whom they can see, intervenes and repositions the obstacle do not expect to be saved. It could be said that it is, or at least feels, morally worse to countenance harm to people who do not expect it than to people who do. Our sense that removing the net to save two is impermissible in the no-enabling harm belief case might reflect this moral difference – we tacitly assume that the two people at the other window believe they are doomed, while the lone man thinks he is safe.
However accurate these points are, they cannot explain the moral difference between the no-allowing and the no-enabling harm belief cases. And in any case the built-in bias can be removed. Thus consider a slightly different case in which Sally acts as the enabler of harm: the equalized no-enabling harm belief case. Here the two people at the second window jump again on the basis of a reasonable expectation that Sally will reposition the net and, thus, not allow them to be harmed. They know that Sally has often been in such situations before, and that she has always sought to maximize the number of lives saved. Like the lone man, then, they now have a reasonable expectation that in jumping from the building they will be saved. If our intuition that Sally's removal of the net in the no-enabling harm belief case is impermissible was determined by our belief that, in cases where harm is not actively done to another, it is morally worse to allow/enable harm to be suffered by people who believe they are safe than it is to allow/enable harm to be suffered by people who think they are doomed, we should find that our intuition changes now that this factor has been neutralized. (Everyone now thinks it is safe to jump.) However, this does not happen – at least, in my intuitive judgement. Hence, reasonable expectations formed on the basis of an obstacle's actually having been positioned are different, morally speaking, from reasonable expectations formed on the basis of an agent's disposition to reposition obstacles. This suggests the following: enabling harm to a victim who acts on a reasonable expectation of not being harmed that was formed on the basis of a belief about an obstacle to harm having been positioned by someone other than the agent enabling the harm is worse than an otherwise comparable case of allowing harm.Footnote 32 Given this, we must conclude that at least some forms of enabling harm are worse than otherwise comparable cases of allowing harm: allowing and enabling harm are not morally on a par if they interact differently with expectations. Hence the moral equivalence thesis is false.
I will now try to provide further support for this claim by looking at cases of burning buildings in which the persons in question face different options that each will give them some chances of survival and choose among them on the basis of their beliefs about the presence or non-presence of obstacles to harm:
Burning Building 4
A person is trapped atop a high building. He can either jump, which in the absence of a net located underneath will give him a 1 per cent chance of survival, or try to exit the building using the stairs, which will give him a 25 per cent chance of survival. Seeing this, a firefighter quickly stations a self-standing net below, which means a 95 per cent chance of survival for the person if the net remains in place, and he jumps. The firefighter then leaves the scene to assist those of his comrades who are battling the blaze. Seeing that the net has been placed below the building the person jumps, reasonably believing this to be, by now, his best chance of surviving. At the same time, Sally, a passer-by, then immediately notices that two other persons are in an identical predicament. They have jumped from a window several yards away acting on the reasonable belief, formed on the basis of Sally's behaviour on previous occasions, that Sally will reposition the net, and, like the lone man, they have not tried to exit the building via the stairs. Sally therefore repositions the net so that it catches the two. The first jumper hits the grounds and dies.
Intuitively, Sally acts impermissibly in this case. (Uncontroversially, she would do so in a one versus one case, and in such a case the single person might permissibly use force to prevent the repositioning of the net.) In fact, it seems to me that even if Sally were another firefighter, not a passer-by, it would be impermissible for her to reposition the net.Footnote 33 Now compare Burning Building 4 to 5:
Burning Building 5
A person is trapped atop a high building. He can either jump, which in the absence of a net located underneath will give him a 1 per cent chance of survival, or try to exit the building using the stairs, which will give him a 25 per cent chance of survival. Believing that it is very probable that the firefighter will place a net below, he leaps off the building. Seeing him jump, a firefighter quickly stations a self-standing net underneath, which means a 95 per cent chance of survival for the jumping man if the net remains in place. The firefighter then leaves the scene to assist those of his comrades who are battling the blaze. At the same time, Sally, a passer-by, then immediately notices that two other persons in an identical predicament have jumped from a window several yards away. She therefore repositions the net so that it catches the two. The first jumper hits the grounds and dies.
This case seems morally different from Burning Building 4. For here it is not clear that Sally acts impermissibly when she repositions the net. It makes a difference that in Burning Building 5 the lone man does not jump as a result of having formed a belief induced by the firefighter's positioning of the net. It is harder to justify enabling harm to such an agent than it is merely to justify allowing it to an otherwise comparable agent. What makes the two cases different, morally, is not whether the lone person jumped believing that the net would be placed under his building or jumped believing it had already been placed there. This is morally irrelevant. The crucial difference is that in Burning Building 4, in contrast to Burning Building 5, the firefighters act in a way that leads the lone person to assess his situation in a particular way, and then to act in a way that will make him even worse off if the agent enables harm than he would have been had the agent not intervened at all. By way of support for this claim consider a case where the presence of a coast guard ship has led some sailors sitting on an adjacent shipwreck to believe reasonably that they will be picked up if they jump into the cold water and swim the short distance to the ship. Once the men are in the water and, let us suppose, unable to get back on the shipwreck, the coast guard ought not to abandon them simply because he can rescue more sailors elsewhere. The shipwrecked sailors gave up on another way of saving their lives – staying put on the shipwreck – in response to the presence of the coast guard ship and the more favourable opportunity this led them to believe they had. I conclude that this is a second kind of case (the first was the case of self-defence mentioned at the end of section II) which indicates that allowing and enabling harm are morally inequivalent. In both kinds of case, allowing and enabling interact differently with the choices people make on the basis of reasonable beliefs about whether others will enable or allow one to be harmed.Footnote 34
IV. THE BURNING BUILDING: A TROLLEY-LIKE CASE
In this section I want to consider a fourth reason to be sceptical about Rickless's argument as it is presented in section II. This reason applies to several of the other cases that have been offered in favour of the moral equivalence thesis.
A possible source of error weakening Rickless's appeal to the intuition that repositioning the net in Burning Building 3 is permissible is this: almost everyone acknowledges that there are cases where one can permissibly kill one person in an attempt to save more people. One case of this sort is the famous trolley example, in the Thomson variant, where a bystander can divert a trolley from a track on which five innocent people are trapped onto another track where only one innocent person is also trapped.Footnote 35 Quite how we should interpret this acceptable harming, and specifically what makes it an exception to the wider moral prohibition on killing one person when that will save five (if such prohibition there is), is, of course, highly controversial. But we need not enter that controversy for present purposes. The important point is that, if, and to the extent that, our assessment of Burning Building 3 is influenced by factors of the sort that persuade us to view the killing permissible in the trolley case, then it is an error to take the intuition that it is permissible for Sally to reposition the net to support the moral equivalence thesis – at least, without further argument. We already know there are some cases in which killing one to save more is permissible (just as it is permissible to allow one to die when the alternative is to let even more people die). Given this, enabling to die could be morally equivalent to killing despite the permissibility of Sally's repositioning of the net in Burning Building 3. In other words, it may be that the permissibility of enabling harm in this case is due to the factors that explain why, in the trolley case, it is permissible to kill one to save more.Footnote 36
Rickless comes even closer to drawing on trolley-like intuitions. He offers a case (reminiscent of cases discussed by Boorse and Sorenson and by Vihvelin and Tomkow) in which, in his view, enabling harm is equivalent to allowing harm:Footnote 37
Drive Away 2
A's car is parked between a rolling rock and one helpless child. A now sees a second rock rolling towards a group of five other helpless children. If A leaves his car where it is, the one helpless child will be saved but the five helpless children will die. But A can drive a short distance and stop, thereby interposing his car between the second rolling rock and the five helpless children. If A does this, the five helpless children will be saved but the one helpless child will die.Footnote 38
With a similar argumentative aim, Hanser appeals to a similar example.Footnote 39
Drive Away 2 looks like a circumstantially adjusted trolley case, inasmuch as all of the innocent children in it face the same kind of threat and the agent can save five innocents by killing one.Footnote 40 Admittedly, with the pair of rolling rocks it involves two distinct (but qualitatively identical) threats, but this does not look like a significant moral difference. Suppose all six children are trapped in a narrow canyon and the one child is closer to the rolling rock than the other five. A's car is parked in front of the one child, but if it remains there the rock will hit the car, bounce into the air, and, having passed the one child, crush the other five. A can park his car in front of the five instead, saving them with the side-effect that the one child is crushed. If A were to do so, this would be morally no different from the decision, made by A in Drive Away 2, to move his car. If this is right, it suggests that the special features of the trolley case do not include the fact that the case involves the redirection of just one threat as opposed to several qualitatively identical threats. Admittedly, some might deny that this is right, contending that it makes a moral difference whether we are dealing with the same or distinct threats. It is implausible to claim that the fact the agent is dealing with one and the same threat makes it harder to justify intervention, so the claim has to be that it makes it easier. But Rickless is in no position to push this line of argument, because he thinks intervening in Drive Away 2 is permissible. On this view, despite the fact that it involves two distinct threats, the intervention must be treated as sufficiently like a trolley case so far as permissibility is concerned.
At this point, one can go either of two ways. One might argue that, in the trolley case, the agent who diverts the trolley does not, in the relevant sense, kill the person on the single-person track, but only enables him to die. Despite the fact that diverting the trolley to save the greater number is normally seen as a case of killing, the positioning of the tracks can be seen as an obstacle standing between the lone person trapped on the track and the threat.Footnote 41 However, this suggestion strikes me as highly revisionist. Suppose a gun will eventually fire and kill someone, say, fifty years in the future, unless someone pulls the trigger. I do so and as a result someone is hit by the bullet fired from the gun and dies. We would resist describing this as a case of enabling harm, so we should also resist the idea that the trolley case involves enabling harm.
Alternatively, one could infer – and this is the general lesson that I want to draw from the present section – that the permissibility of countenancing harm to the one helpless child in Rickless's case is explained, not by the fact that the case is one of enabling harm, but rather by the situation's trolley-case character. Accordingly, and in contrast with what Rickless contends, the fact, in this case, that you may enable one to die when you thereby save five from dying does not support the moral equivalence thesis.
V. WHEN DOES ONE ENABLE HARM?
The last three sections criticized Rickless's positive arguments for the moral equivalence thesis. They also presented reasons – relating to self-defence, and to the interaction between the allowing/enabling distinction and expectations of harm, plans of actions based thereon, and intentions (see note 35) – to reject it. In this section I will assess Rickless's attempt to fortify the moral equivalence thesis by showing that he can accommodate a case which seems to amount to a decisive counterexample to it.
When assessing the moral equivalence claim, or for that matter its denial, it is crucial to clarify the conditions under which one can be said to remove ‘some obstacle which is, as it were, holding back a train of events’ rather than initiating a new ‘train of events’. Rickless acknowledges this need and although he does not ‘attempt to provide a complete account of the metaphysics of causal sequences’, he believes that certain otherwise intractable cases of enabling death that seem morally like killing can be dismissed as counterexamples to the moral equivalence thesis on the ground that, unlike cases that are similar to the standard trolley case, such as the Drive Away 2 case discussed in section IV, they do not involve any threatening ‘pre-existing causal sequence’.Footnote 42 In this connection he considers a case in which a bystander can pull a lever. If the bystander does so a trap door will open under an innocent person with a noose around his neck. If he refrains from doing so, another trap door on which five innocent persons are standing with nooses around their necks will open. Intuitively, it is impermissible to pull the lever. Since the bystander is only enabling the innocent person to die,Footnote 43 the fact that it is impermissible to pull the lever suggests that ‘at least one kind of harmful enabling is morally equivalent to harmful doing’.Footnote 44 Call this example The Hanging.
In response to this putative counterexample to the moral equivalence thesis, Rickless maintains that, despite what one may initially think, The Hanging is not a case of enabling harm: ‘The opening of the trap door does not make it possible for a causal sequence that is already in train to continue, for there is no such causal sequence. In particular, the one is not in the process of falling. The situation is, rather, completely stable.’Footnote 45 Accordingly, the fact that it is impermissible to pull the lever does not even speak to, let alone support, the moral equivalence thesis.
For three reasons, Rickless's treatment of this putative counterexample to the moral equivalence thesis is unsatisfactory. First, without any explanation it employs a notion of enabling harm that deviates from Foot's own conception. When one removes a plug allowing water to flow, presumably, the situation prior to unplugging might have been ‘completely stable’ in Rickless's sense, with no water flowing. Yet this is one of the examples that she uses to illustrate her second notion of allowing. Admittedly, Foot's notion of enabling harm might be unusual here, since it might be more natural to describe the pulling of the trap-door releasing lever as a case of harming rather than simply enabling harm.
Second, it is unclear that complete stability of this kind is morally relevant. Suppose the innocent person happens to be in the process of falling: the trap door is opening by a centimetre a year, and before the innocent person's death from natural causes it will eventually open entirely. In this case there is a ‘causal sequence that is already in train to continue’. Hence, on Rickless's account, we have a case of enabling harm. If this is so, and if the moral equivalence thesis is true, the act of pulling the lever should differ morally (other things being equal) from the lever-pulling in the case Rickless discusses. Yet there seems to be no such difference. In my view this speaks against equivalence.
It might be said that the mere fact that the trap door is opening by a centimetre a year until the innocent person is hanged is irrelevant. The crucial point is that there is no causal sequence in train that is being held back by some obstacle in a way that prevents the innocent person from being killed in the immediate future. The causal sequence that will lead to the victim's death thirty years later is irrelevant to a moral assessment of pulling the lever.
This way of defending the moral equivalence thesis arguably involves a double-edged sword. Consider McMahan's Dutch boy, who has stuck his finger in a crack in a dyke to prevent it from crumbling.Footnote 46 The boy need not do anything to keep his finger there, or so we shall suppose. It is stuck in the crack, and to withdraw it would involve effort and time.Footnote 47 Suppose that even with the finger in the crack the dyke is still crumbling, however, at a reduced pace. If at some point the Dutch boy decides to free his finger, the dyke will crumble much faster. Intuitively, in this case, in withdrawing his finger the Dutch boy enables the harm that ensues. Yet, if we say that the fact that the dyke is slowly crumbling is a different causal sequence from the one that will lead to the destruction of the dyke if the boy withdraws his finger – which we must, if we offer the reply sketched in the previous paragraph – we can no longer say the boy is enabling harm.Footnote 48
Third, whether a situation presents as a ‘completely stable’ one depends on the way (or the level at which) it is described, and it is far from easy to see how a descriptive decision could reasonably determine moral permissibility. At one level, for instance, it might be quite correct to say that someone's standing on a mountain peak is a completely stable situation. Yet, at another level, it might be incorrect, because continents are moving, and mountains are being eroded, and so on.Footnote 49 Note also that there is a sense in which a situation in which a person is placed in a respirator is completely stable, although most of us believe that doctors simply enable a death to occur when they remove it.Footnote 50 For the three reasons given, I suggest that Rickless fails to explain away the counterexample to the moral equivalence thesis presented by The Hanging.
VI. CONCLUSION
This article has taken a closer look at some cases offered in response to the claim that enabling harm is morally equivalent to allowing harm. It has argued that these are not compelling. Cases of enabling harm to one or more people to save even more people from comparable harm are similar in important respects to the archetypal trolley case. As a result the permissibility of enabling harm seems to tell us little about the moral status of enabling harm relative to the moral equivalence thesis. It has also been argued that the countenancing of harm is harder to justify when people act on the basis of a reasonable belief to the effect that certain obstacles to harm will remain in place than it is when they act on the basis of a reasonable belief to the effect that others will not allow harm to them. Cases involving self-defence or malicious intentions also appear to rebut the moral equivalence thesis. Finally, I have raised some doubts about the moral relevance of a situation being completely stable, and some doubts about the serviceability of the notion of stability as a determinant of whether an act counts as an enabling of harm. Twenty years ago, reflecting on the moral significance of the distinction between killing and letting die in the light of cases involving enabling death, Jeff McMahan conjectured that ‘an ethical theory that aims to unify and systematize the intuitive elements of common moral thought will have to be quite complex’.Footnote 51 This article supports that conjecture to the extent, at any rate, of showing that common moral thinking on the enabling of harm is more complex than the moral equivalence thesis allows.Footnote 52
