Coursehero >> Illinois >> Loyola Chicago >> LAW 1

Course Hero has millions of student submitted documents similar to the one below including study guides, homework solutions, papers, exam answer keys and textbook solutions.

We When May Balance Evils: Agent-Relative Restrictions On Consequentialist Justifications by Michael S. Moore Loyola University Law Faculty Workshop Chicago, December 13, 2005 Walgreen University Chair, Professor of Law, Professor of Philosophy, Professor in the Center for Advanced Studies, and Co-Director of the Program in Law and Philosophy, University of Illinois. Table of Contents I. II. Introduction: Natural Law and Substantive Ethics Some General Considerations About Obligation and Justification in Ethics A. The Three Level Analysis 1. 2. 3. Level One: The Background of Consequentialist Reasons Level Two: An Overlay of Agent-Relative Reasons Level Three: The Reappearance of Consequentialist Reasons Over Some Threshold of Moral Horror B. III. Meta-Ethical Intrusions Intending vs. Foreseeing: The Doctrine of Double Effect A. B. C Introduction The Psychological Distinction Between Intending and Foreseeing The Moral Force of the Psychological Distinction IV. Doing versus Allowing I: Omissive Allowings A. B. C. Introduction The Metaphysical Distinction Between Acts and Omissions The Moral Force of These Metaphysical Distinctions V. Doing versus Allowing II: Non-Omissive Allowings A. B. Introduction Non-Omissive Allowings 1: Removing a Defense One Was Entitled not to Have Provided 1. 2. Introduction The Metaphysics of Removing a Defense to a Death While Not Causing That Death 3. C. D. E. The Moral Significance of Removing Defenses Non-Omissive Allowings 2: Accellerating a Death About to Happen Anyway Non-Omissive Allowings 3: Aiding Another Person to Cause Non-Omissive Allowings 4: Being a Minor Cause of Harm vis- -vis Nature 1. 2. Extraordinary Natural Events as Big Causes Redirecting Presently Existing Threats VI. Conclusion 2 I. Introduction: Natural Law and Substantive Ethics In a sense my topic in this article is within the natural law and not about it. A full natural law theory as I understand it has two theses to it: (1) the metaethical thesis that morality is objective, and (2) the relational thesis that human law is intimately related to that objective morality.1 I come not to defend either thesis here, although I believe each to be fully defensible.2 Rather, I wish to explore some general issues about the content of the objective morality that makes up the natural law. (It is in that sense that my topic is within the natural law and not about it.) I do this moral exploration freely mixing legal with moral examples. This is not because I think the connection between law and morality to be so tight that if something is legal it must also be moral. Rather, I find the law instructive in ethics because it contains examples we armchair types would have trouble dreaming up. Moreover, the thought experiments these legal examples provide come with proposed resolutions by judges, who are both uncontaminated by acquaintance with the moral theories we wish to test, and whose resolutions are serious in the sense that real world consequences turned on what they decided. I thus join J.L. Austin in thinking (for rather different reasons) that it is a perpetual and salutary surprise to discover how much is to be learned from the law when one is doing ethics.3 Aquinas told us that the first precept of the natural law is to do good and to avoid evil.4 Very broadly speaking, my topic is to make that precept a bit more precise. More specifically, I am interested in the interplay between the good and bad states of affairs that our actions produce, on the one hand, and what we are obligated to do, on the other. Put popularly, when does the end justify the means? Consider the contemporary debate about whether torture can ever be justified when the stakes are high, as in the current war on terror. Few disagree that torture is an evil and that See Moore, Law as a Functional Kind, in R. George, ed., Natural Law Theory (Oxford: Oxford University Press, 1992). 2 See the articles collected in Moore, Objectivity in Law and Morals (Aldershott, U.K.: Ashgate Press, 2004). 3 J.L. Austin, A Plea for Excuses, Proceedings of the Aristotelean Society, Vol. 57 (1957), pp. 1-30. 4 More exactly, good is to be done and ensued, and evil is to be avoided. Summa Theologica, First Part of the Second Part, Q. 94, Art. 2. 1 3 morality prima facie forbids it.5 My question is whether something like torture may nonetheless be justified sometimes by its good consequences. When such good consequences are in the offing, can we justify: sending terrorist suspects to other countries when we know the regimes of those countries will torture but we do no torture ourselves nor do we intend that they torture?6 Failing to save terrorists from torture, when we could do so?7 Allowing nature to take its course in causing torture-like pain to terrorist suspects?8 Redirect torture from one suspect to others more likely possessed of life-saving information?9 Etc. It is interesting that in the context of torture people often find these distinctions to the pretexts and evasions. Perhaps we are just used to being lied to in this context.10 Yet even such evasions may be the proverbial compliments vice pays to virtue. The distinctions themselves may be more plausible than their use in this context. In any event, such distinctions are my topic. My interest in the topic is threefold. First, this part of ethics does have real and important payoffs, both in public policy and in how we lead our individual lives. It matters, in a way few things in the academy do. Secondly, whether the morality of our most stringent obligations can be free of conflict between such obligations depends on how we draw the kinds of distinctions above alluded to. It may not be, as Kant famously proclaimed, that a conflict of such obligations is literally inconceivable. 11 Yet it would be unfortunate for us in the extreme if morality often confronted us with choices where we On torture being evil, see David Sussman, What Is Wrong With Torture?, Philosophy and Public Affairs, Vol. 33 (2005), pp. 1-33. On morality prima facie obligating us not to torture, see Moore, Torture and the Balance of Evils, Israel Law Review, Vol. 23 (1989), pp. 280-344, revised and reprinted as chapter 17 of Moore, Placing Blame: A General Theory of the Criminal Law (Oxford: Oxford University Press, 1997). 6 The CIA s extraordinary renditions are described in some detail in ____________, New Yorker. 7 Some of the pain used to extract information from terrorist suspects has been pain that interrogators have omitted to prevent, not pain that they have caused by torture. Al Qaeda s one-time chief of security, Abu Zubeida, for example, was wounded in the chest, thigh, and groin when captured by Pakistani forces; CIA interrogators withheld pain medication for some time to induce him to talk. Los Angeles Times, March 6, 2003. 8 Sister Dianna Ortiz, a Catholic nun tortured in Guatemala, terms these techniques torture light. Adopted from British interrogative techniques used in the campaign against the IRA, they consist of sensory deprivation, sleep deprivation, distraction by loud lights and sounds, and the like. See the description of these techniques in Los Angeles Times, March 3, 2005, p. A-1. 9 Arguably the extraordinary rendition of the Albanian Al Qaeda cell in 1998 was partly a redirection of Egyptian secret police from less knowledgeable to more knowledgeable members of the Egyptian Jihad. See the detailed story of this operation under President Clinton in The Wall Street Journal, Nov. 20, 2001, p. A-1. 10 See, for example, the plainly false statements by President Bush and his then Attorney General, John Ashcroft, that we did not even know that countries like Egypt were torturing the terrorist suspects we sent to them. 11 Kant, The Metaphysical Elements of Justice (J. Ladd trans., 1965), at p. 25. 5 4 will be damned if we do and damned if we don t. The distinctions we shall examine hold out the possibility of so limiting our stringent obligations as to minimize or even eliminate such situations of moral conflict.12 Thirdly, my interest here is like that of George Mallory. When asked during a visit to New York in the 1920 s why he wanted to climb Mt. Everest, he is reported to have replied, Because it is there. 13 For us natural lawyers, morality is there just as solidly as is Mt. Everest. We are understandably as curious about the moral features of our world as about its physical characteristics. II. Some General Considerations About Obligation and Justification in Ethics A. The Three Level Analysis Although my topic is agent-relative obligation, many things I say will presuppose a certain view of more general matters in ethics. So I shall begin by describing this view briefly (which means, with little argumentative support). Such baring of presuppositions will hopefully make things clear, so that if I am wrong later I will at least be wrong clearly (which is not to say, clearly wrong). 1. Level One: The Background of Consequentialist Reasons I take consequentialism to be a true, general principle of practical rationality. I understand consequentialism to be the idea that we generally have reason to maximize good states of affairs and minimize bad ones in our actions and institutions. Usually this idea is broken down into two parts. First, it presupposes some theory of the good. More precisely, we need a theory about what states of affairs are intrinsically good. Welfare of persons (in the form of pleasure, happiness, preference-satisfaction, etc.) is one sort of answer, a utilitarian sort of 12 As many deontologists, beginning with Aquinas, have recognized. See the discussion in Anscombe, War and Murder, in Walter Stein, ed., Nuclear Weapons: A Catholic Response, (New York: Sheed and Ward, 1962), reprinted in P.A. Woodward, ed., The Doctrine of Double Effect (Notre Dame: University of Notre Dame Press, 2001). See particularly p. 256 (in the reprinted version), where Anscombe finds some of these distinctions to be absolutely essential to deontological ethics. 13 Mallory? 5 answer if such welfare is the only intrinsic good thought to exist. But the keeping of moral duties, the non-violation of moral rights, the possession of moral virtue, the justness of an institution, are alternative theories of what is intrinsically good. Second, consequentialism takes a position in the theory of the right: right action (or right institutions) consists of those actions (institutions) that maximally produce intrinsically good states of affairs. One can put this as a slogan: if something is good, surely more of it is better, and this betterness gives us reason to do acts that produce it. Surely many decisions in daily life are properly governed by consequentialist reasons. Many years ago a prominent law school was reconsidering the start date for that year s classes. Consequentialist arguments about shortened late summers versus earlier and longer holiday periods, summer employment opportunities for students versus faculty preferences, abounded. One well-known deontologist, however, spoke up: all these trade-offs between good states of affairs was by-the-by, he said, because by right he was entitled to the same start date as the school had when he began his employment. Needless to say, the argument did not carry the day. Isn t that because none of us live by conformity to the stringent duties of deontology alone? Much of our reasoning is consequentialist, and rightly so. I take consequentialism not only to provide a principle of practical rationality but also a principle of morality. Maximizing the good, that is, not only gives us an objective reason so to act. It sometimes gives us that pressing kind of reason we call a moral obligation. That there are purely consequentialist obligations in morality is important to what follows, even though my topic is agent-relative (or deontological ) obligations. So consider one example by way of making plausible the idea. The proverbial baby is drowning face down in the puddle. You, eating your lunch on a park bench, could tip it over with your foot and thereby save it from drowning. But you munch away and the baby drowns. Even though this is a stranger-rescue case the baby is no relation to you I am not so libertarian as to believe you lack an obligation to save the baby. And if you agree with me that you are so obligated, then also agree with me that your obligation is consequentialist through and through. You can appreciate this latter fact by imagining two other babies drowning in another 6 puddle, and time being sufficiently short that you cannot save all three babies. In this variation, surely you can justify not saving the one baby in order to save the other two. In which case, notice, your obligation to save the first baby is gone the moment good consequences appear not to be maximized by doing what was obligatory. Obligations easily justified in such a consequentialist way are consequentalist obligations: it is because the act of saving a baby produces a good state of affairs that the act is obligatory. 2. Level Two: An Overlay of Agent-Relative Reasons My (true) story of the law school start date was intended to show that no one can live by deontology alone. But the same is true of consequentialism. Two sorts of supplements are needed. First, we need agent-relative permissions not to be maximizing good consequences all the time. We might call this the good dinner point: as you head into the restaurant for a fine (and expensive) meal, several street people point out to you that the money you are about to spend on that meal could keep all five of them from being hungry for the next twenty-four hours and that (unless you have particularly gluttonous capacities) more good would be produced by you eating more modestly and giving them the rest of the money you are about to spend.14 The familiar point is that we need relief from the saintly pressures of consequentialist reasons. We need permission on many occasions not to do acts maximizing of good consequences. Secondly, we need agent-relative obligations that both prohibit actions even when those actions would produce the best outcomes, and require actions even when those actions would not produce the best outcomes. Intuitively, I may not torture one person even if my doing so will prevent the (equally awful) torture of two others by either me or someone else at some other time. Intuitively (at least to me), I must save my child from drowning even if that means I cannot save your two children nearby. 14 Not purely a hypothetical. Several blocks down from the University, on Telegraph Avenue in Berkeley, was at one time a fine restaurant. Often the entrance to it was lined (gauntlet-like) with Berkeley street people, some of them educated enough to make pretty good diminishing marginal utility arguments. Perhaps UC-Berkeley s Sam Scheffler ate there too. See his The Rejection of Consequentialism (Oxford: Oxford University Press, 1982), which argues for the existence of agent-centered prerogatives to go have a good dinner once in a while. Tom Nagel s example of a similar point is his permission to satisfy his desire to climb Kilimanjaro in Africa. See his The View from Nowhere (New York: Oxford University Press, 1986), pp. 166-175. 7 I like many others see these agent-relative permissions and obligations as a kind of overlay on a consequentialist background: we generally are governed by consequentialist reasons, save when we are permitted or obligated by agent-relative reasons.15 This gets me to my topic in this paper. I shall focus on agent-relative obligations, and usually negative rather than positive versions of those. I shall not directly be concerned with agent-relative permissions. With respect to such obligations, my topic is not what I take to be the major question about them, namely, their rationality. (Roughly: how can it be rational not to torture one person when that would minimize torture? If torture is so bad, why is not more of it even worse, and why doesn t that guide our action?)16 My topic is more preliminary, more clarificatory: what sorts of things form the content of our agent-relative obligations? Are we forbidden to do the action of torturing? To cause torture to be done? To form the intention to torture? To try or attempt to torture? To intentionally torture? To even consider torture as a possible action? Each of these possible objects of our deontological obligations has had its proponents within ethics.17 My question is which is correct. One way to argue for some position here is through what I shall call the front door. This is to give positive argument for why intentions to torture, for example, should be taken to be what we are obligated not to have or form.18 I shall proceed in a more indirect, if time-honored way, however, through what might be called the back door. I shall ask about the proper use of consequentialist justifications vis- -vis a variety of paired examples, hoping to tease out an answer to the question of what deontological obligations must prohibit, from considering when consequentialist justifications are permitted. There is nothing mysterious about such a method. On the three level view of ethics I have been describing, anything not governed by an agent-relative obligation or permission is For me, this way of thinking about these agent-relative reasons began with Robert Nozick s discussion of side constraints in his Anarchy, State and Utopia (________, 1974). 16 See the classic discussions of this topic collected in S. Scheffler, ed., Consequentialism and Its Critics (Oxford: Oxford University Press, 1988). 17 See the very nice survey of possibilities in Heidi Hurd, What In the World is Wrong?, Journal of Contemporary Legal Issues, Vol. 5 (1994), pp. 157-216. 18 See, for example, Christine Korsgaard, (Illinois Philosophy Dept. paper, 2005). 15 8 governed by consequentialist reasons. (That is a consequence of the view of ethics as a consequentialist background with agent-relative overlays.) So if an omission, say, to rescue a stranger baby is something I am not deontologically obligated not to do, then I am permitted to justify such an omission on consequentialist grounds. One can thus seek the content of deontological obligation through the seeking of the range of permissible consequentialist justification. One way to describe my topic thus is to call it the scope of permissible consequentialist justification of prima facie prohibited actions. Notice, however, that we are here referring to weak permissions: such permissions are simply the absence of deontological obligations. Such weak permissions do no more than land one back into the realm where consequentialism holds sway, so that one may yet be subject to consequentialist obligations. Contrast this with the strong, agent-relative permissions I mentioned earlier. These permit one to do some action A even if A does not maximize good consequences. Such strong permissions are thus not at all like weak permissions, which subject weakly permitted actions to the dictates of consequentialist reasons. The distinction between weak and strong permissions is no mere technical distraction. On the contrary, it importantly bounds my topic. Consider the right of self-defense. Often in discussions of the kind I am about to begin, self-defense examples are presented to make one point or another about the bounds of permissible consequentialist justification. Thomas Aquinas is an early sinner here, using self-defense cases to illustrate the doctrine of double effect.19 Puzzles then get raised about how it can be correct for you to defend your life by killing any number of attackers no matter what the discount placed on attackers lives, surely your (innocent) life at some point gets outweighed by their discounted but numerous lives?20 Such puzzles would be real if we were only weakly permitted to defend our lives with deadly force. For then our justification for self-defense would be consequentialist, and we d have to defend the view that our life is a greater good than the lives of any number of culpable 19 20 Aquinas, Summa, IIa, IIae, Q.64, Art. 7. See, e.g., the discussion of discounts and forfeitures in the self-defense context, in Judith Thomson, Self-Defense and Rights, Lindley Lecture, University of Kansas, 1976. 9 aggressors. Yet (on my view) we are strongly permitted to defend ourselves no matter what the consequentialist balance might say.21 Culpable aggressor self-defense examples are thus out of place in the discussion to follow, focused as it is on agent-relative obligation (and thus, weak permissions) but not on agent-relative (or strong) permissions. I similarly think (although here I am more tentative) that innocent aggressor self-defense cases are instances of strong permissions as well.22 When two innocents threaten your life they are children who do not understand the guns they are about to fire at you or they are two fat people falling on you through no fault of their own I like Nozick23 think you may kill them if necessary to save your own life. This may is not the language of excuse: you are justified in killing them. Moreover, you are justified in killing them even though more good would be produced if you did not kill them and you died instead. You are, in other words, strongly permitted to kill them. If you agree with all of this, then you will agree with what I (in any event) intend to do for the remainder of this paper: I shall also put aside innocent threat cases as not germane to my discussion. 3. Level Three: The Reappearance of Consequentialist Reasons Over Some Threshold of Moral Horror In this sketch of ethics so far we have an agent-relative overlay on a consequentialist background. Now I want to add a consequentialist override of the agent-relative overlay. I confess it: I am only a threshold deontologist. This means that over some threshold of truly awful consequences, I will potentially do virtually anything to avert such consequences. If I can locate and defuse a nuclear device at 42nd street only by torturing the innocent child of the terrorist who planted it there, I torture. For reasons explored by Phil Montague, in his Self-Defense and Choosing Between Lives, Philosophical Studies, Vol. 40 (1981), pp. 207-217. As Montague recognizes (p. 211), in a standard self-defense situation one is not simply permitted to kill in self-defense, but one has a right to do so, and to kill as many aggressors as is necessary to save himself 22 An early discussant of innocent threat cases was George Fletcher. See his Proportionality and the Psychotic Aggressor: A Vignette in Comparative Criminal Theory, Israel Law Review, Vol 16 (1973), pp. 367-390. 23 Nozick, Anarchy, supra note 15. 21 10 It would require a lengthy paper of its own to defend this (very un-Kantian) kind of deontology.24 Other than a recent, detailed critique of my version of threshold deontology by Larry Alexander,25 there is little argument in the ethics literature about this level. Assumptions and assertions abound, however, from Kant s not even though the Heavens may fall, 26 through Elizabeth Anscombe s it is immoral to even think about such examples, 27 through Bernard Williams this is beyond morality, 28 to brief, undefended confessionals by other threshold deontologists such as Tom Nagel,29 Charles Fried,30 and many others. But defending threshold deontology is not my present topic. Distinguishing its licensing of consequentialist justification, from the licensing of consequentialist justification given by weak permissions, is important to my topic, however. Examples of catastrophic consequences will not do to make any point about the scope of deontological obligations/weak permissions. If persuasive, such examples make a very different point, viz, how and when deontological obligation can be overridden. My topic is what the content of those obligations may be in the first place. B. Meta-Ethical Intrusions One other glance at the forest before we start looking at the individual trees. I am not only a deontologist in my ethics, but (what some I know find peculiar to combine with that) I am also a naturalist-realist in my meta-ethics.31 Many find it unsurprising to combine realism in meta-ethics with a deontological approach to ethics, but the naturalist version of realism is something else. Most naturalists are not realists to start with, and of those like Nick Sturgeon who are both,32 they tend to be consequentialists (as is Sturgeon) in their substantive ethics. 24 25 I make a few gestures in this direction in my Torture and the Balance of Evils, supra note 5. Larry Alexander, Deontology at the Threshold, San Diego Law Review, Vol. 37 (2000), pp. 893-912. 26 Kant 27 Heidi Hurd discusses Anscombe s (and Peter Geach s) refusal to discuss such cases in What in the World is Wrong?, supra at pp. 170-174. 28 Williams, A Critique of Utilitarianism, in J.J.C. Smart and B. Williams, Utilitarianism: For and Against (Cambridge: Cambridge University Press, 1973), at pp. 90-93, 118. 29 Nagel, War and Massacre, in S. Scheffler, ed., Consequentialism and Its Critics, supra. 30 Charles Fried, Right and Wrong (Cambridge, Mass.: Harvard University Press, 1978), at p. 10. 31 See the essays in Part I of my Objectivity, supra note 2. 32 See the citation to Sturgeon and other contemporary naturalist realists in my Objectivity, supra note 2, pp. 105106 n. 27. 11 Naturalism, as I have defined it in ethics, is the view that moral properties supervene on natural properties. Supervenience means asymmetrical co-variance: any difference in moral properties must be underlain by a difference in natural properties, but not vice-versa. There can be alternative, multiple realizations (in natural properties) of one and the same moral property. This has the following implications for the present paper: if there are moral differences between intending versus foreseeing, doing versus allowing, acting versus omitting, etc., then these moral differences must be underwritten by natural differences between these pairs. Put more simply, we need these distinctions to be based on real distinctions in psychology, in the nature of events, in the metaphysics of causation, etc. Giving an analyses of such distinctions in terms of further moral properties omittings are failures to act where there was a moral duty to have acted, for example will thus be unacceptably incomplete. The worry here is not circularity. Sometimes there is admittedly that worry too, as when someone defines intending as that which one cannot justify consequentially, or defines allowings so that some very blameworthy omissions (such as Philippa Foot s starving of a beggar to death for his organs) are defined to be active killings on the ground that they are as blameworthy as truly active killings.33 Rather, the worry is naturalism: for every moral difference, find a natural difference. Of course, one can be a patient naturalist: we don t need to reach immediately to natural properties. Just so long as we eventually get there. This suggests something important about doing this kind of ethics, which is that one cannot do it without simultaneously doing the philosophy of mind, the philosophy of action and of events, the metaphysics of causation, and the like. Such I have always thought in seeking to use metaphysics to moral advantage in my lifetime preoccupation with moral responsibility and blameworthiness. 33 Philippa Foot, The Problem of Abortion and the Doctrine of Double Effect, originally published in The Oxford Review (1967), reprinted in Woodward, Double Effect, supra note 12, at p. 150. Foot reconsiders this in her Morality, Action, and Outcome, in T. Henderich, ed., Morality and Objectivity (London: Routledge, 1985), reprinted in Woodward, Double Effect, supra note 12 pp. 81-82 n. 6. 12 As J.L. Austin used to say,34 enough preliminary cackle. Now to the trees, considered one by one. III. Intending vs. Foreseeing: The Doctrine of Double Effect A. Introduction The oldest and the best known of the distinctions relevant here is that between intending a harm by one s action and merely foreseeing that a harm will be produced by one s action. The idea is that one may justify causing something bad with good enough consequences only if one did not intend to cause the bad thing when one acted. If one merely foresaw that one s act would produce the bad effect, then one is outside the agent-relative prohibition and may justify the action if it is productive of good enough consequences. A standard example is bombing in war-time. Killing non-combatants is bad. Suppose an act of bombing causes a number of such deaths. Suppose further the bombing takes place during a just war so that the bomber s side winning such a war is a very good consequence, and that the bombing will produce just that consequence. Can the act that causes one bad thing (the killing of non-combatants) be justified by the fact that that act also causes a very good thing (the winning of a just war)? The doctrine of double effect ( DDE ) holds that the bombing can be justified if the killing of the non-combatants was only foreseen as a side effect, not if it was intended as either a means or an end. So a bomber who is trying to hit military targets, but knows he will hit some civilians, is eligible to justify his bombing by its beneficial consequences. But the bomber who seeks to so dispirit the enemy by killing civilians, that the enemy will surrender, and thus aims his bombs at civilian targets, is not eligible to justify his bombing by its good effects. He has intended the death of the civilians as a means to his end (winning the just war); he has not merely foreseen that the death of the civilians will be produced by the means he has chosen to end the war. 34 J.L. Austin, 13 As with each of the distinctions we shall examine, there are two questions to ask here. First, what more precisely is the distinction marked by the DDE? What (ultimately natural) properties are being invoked when we distinguish intendings from foreseeings? Second, can this natural (psychological) distinction carry the moral freight with which the DDE loads it? How plausible is it that foreseen deaths can be justified by good consequences but that intended deaths cannot be justified by those very same good consequences? These two questions are not unrelated we should adjust the natural distinction in any way that makes more plausible some moral significance for it. Still, like a pair of trousers, it helps to do one leg at a time. B. The Psychological Distinction Between Intending and Foreseeing Start with the apparent psychological distinction. On the surface, the psychological distinction is clear enough. For over a thousand years the folk psychology we all share has posited the existence of three (or two, depending on who you like here)35 separate kinds of representational states that figure in practical rationality: there are motivational states of desire or of ultimate intention; there are cognitive states of belief; and there are executory ( conational ) states of mediate intending or willing. I want to be rich, I believe that the world is so arranged that if I work hard, I will be rich, so I intend to work hard, and I do just that. The DDE distinguishes cognitive states of belief from both motivational states of ultimate intention and from executory states of mediate intention.36 The strategic bomber who only believes his bombings will kill civilians is to be distinguished both from the malicious bomber who is ultimately motivated by his desire to kill civilians and from the terror bomber who aims to kill civilians as his means to terrorizing the enemy into surrendering. The DDE allows the strategic bomber the chance to justify his acts; it does not allow the malicious or terror bomber this chance, no matter how good the consequences may be of their bombings. So far so good. Belief, even predictive belief, is a distinct kind of mental state, different from both ultimate and mediate intentions. (Let us ignore Bentham s odd stipulation that such I am thinking of the Davidson/Bratman debate about the independence of intention from both belief and desire. See the summary in Moore, Act and Crime (Oxford: Clarendon Press, 1993), Chap. 6. 36 Ultimate/mediate are Betham s Terms. Bentham, An Introduction to the Principles of Morals and Legislation (Oxford: Basil Blackwell, 1948), pp. 201-202. 35 14 predictive beliefs are a kind of intention, which he dubbed oblique intention. )37 Things get sticky when we move from the nature of the mental states involved, to their representational content. When I was less of a fan of the DDE that I am now,38 I presented the following kind of example as showing the doctrine s absurdity: Herod did not desire the death of John the Baptist for its own sake; he preferred that John live. Still, Herod did want to please Salome, and Salome wanted John s head on a plate. So Herod cut off John s head. Yet he did not intend John s death, not as a motivating end nor even as a means. He merely foresaw that John was going to have a hard time living without his head. So if pleasing Salome was a really good thing, Herod could justify cutting off John s head under the DDE. (Like examples were developed by Philippa Foot,39 Warren Quinn,40 Jonathan Bennett,41 and others.) A defender of the DDE cannot accept this kind of example. It is not that the psychology is bad, although it might be. Rather, it makes the DDE morally counterintuitive. If the correct psychology makes individuation of the content of representational states as fine-grained as this example suggests, then few of us ever intend (as opposed to foresee) anything bad flowing from our actions. Bennett s example: terror bomber doesn t intend the death of the civilians in order to dispirit the enemy and end the war; no, he sees he only needs the enemy to believe the civilians were killed, because it is that belief that does the dispiriting; so terror bomber explodes the bomb over the civilians because that is what gives such an appearance that they are killed.42 Of course, exploding it over them also kills them, but this further effect he only foresees; he strictly intended only to create the appearance of their deaths, not their actual deaths. I shall not rehearse the various attempts of defenders of the DDE to get around this problem. (These include H.L.A. Hart s idea of inevitable causal connection,43 Foot s idea of 37 38 Id. Moore, Intention and Mens Rea, in R. Gavison, ed., Issues in Contemporary Legal Philosophy (Oxford: Oxford University Press, 1987), reprinted as chapter 11 of Moore, Placing Blame, supra note 5. 39 Foot, Abortion, supra note 12. 40 Quinn, Action, Intentions, and Consequences: The Doctrine of Double Effect, Philosophy and Public Affairs, Vol. 18 (1989) pp. 334-351, reprinted in Woodward, Double Effect, supra note 12. This aspect of Quinn is critiqued in Fisher, Ravizza, and Copp, Quinn on Double Effect: The Problem of Closeness, Ethics, Vol. 103 (1993), pp. 707-725, reprinted in Woodward, Double Effect, supra note 12. 41 Jonathan Bennett, Morality and Consequences (Salt Lake City: University of Utah Pres, 1981), pp. 110-111. 42 Id. 43 Hart, Intention and Punishment, The Oxford Review, Vol. 14 (1967), pp. 5-22, reprinted in his Punishment and Responsibility (Oxford: Oxford University Press, 1968). 15 closeness 44 Quinn s idea of only intending the act that is aimed at the victim and which foreseeably, directly causes his death,45 the common idea that Herod actually intends John s death as the means to getting what he does intend, John s head on a platter;46 or the suggestion of Lord Hailsham, that some intentions are morally equivalent even if not actually identical.)47 I think the solution is close to several of these suggestions, but I will put it my own way. Notice that the problem is not one of individuating numerically distinct intention-tokens, as I once thought.48 Rather, the problem is a two-fold problem of classification that inevitably will have some slop in it, no matter how fine-grained may be the psychology of representational content. To explain. The problem of individuating intention tokens is this one: considering two differently described intention tokens, are these putatively distinct things really one and the same thing? If Herod intended to kill John, and Herod intended to cut off John s head, are these one intention or two? I still think what I thought over twenty years ago on this issue: these are two distinct representations, and as long as Herod has them both in his head (rather than an outside observer using these different descriptions to refer to Herod s intention(s)), then Herod has two distinct intention-tokens.49 Yet this is not dispositive of the issue at hand. Suppose morality like the law contains an agent-relative prohibition on maiming (disfiguring) another (common law mayhem). Suppose the DDE is right: what is prohibited is not just maiming, but intending to maim. If D swings a stick intending to put out V s left eye, and (because V moves at the last second) D puts out V s right eye, has D violated the prohibition? Stipulate (as is legally true) that putting out an eye is maiming (disfiguring). So the acttoken done instantiates the type of action prohibited. Now focus on the intention: did he intend to maim, and did he intentionally maim? For more than one reason, these are two different 44 45 46 47 48 Foot, Abortion, supra note 12. Quinn, Double Effect, supra note 40. I explore Lord Hailsham s views in my Intention and Mens Rea, supra note 38. Id. 49 Id. 16 questions. Take the second first. This asks whether there is a match between what defendant did and what he intended to do. What the defendant did was an act token; he put out V s right eye is one description of it. What the defendant intended to do is the representation of an act-type. We never intend or predictively believe about particular events being in the future, it is hard to refer to them as particulars.50 We always intend or predict the occurrence of some instance of a type of event. So the match question here is one of classification: is the act of putting out V s right eye an instance of the type of act that D intended? (If not, D is only an accidental maimer, not an intentional one.) So what did D intend? Did he represent the type of event he was trying to bring about as a putting out of V s left eye, as a putting out of V s right eye, as a putting out of an eye of V s (in the sense of any eye), as a putting out of an eye of V s (in the sense of one particular eye), as a harming V, a disfiguring V, a maiming V, etc. My supposition: any of these representations will suffice to match what D did to what he intended to do, to make D an intentional maimer, even though it is unlikely in the extreme he had more than one or two of them as the representation under which he acted. If so, notice how much slop there is in fixing what it was that D intended. Now look at the other classificatory question. Even though the intention D had takes a representation of a type as its content, the intention itself is a particular, an intention-token. To see whether V had the intention the norm against maiming prohibits, we need to see whether the intention-token D had instantiates the type of intention the norm prohibits. The norm prohibits (it s my hypo, so I can tell you authoritatively) intents to disfigure a person. Did D have such an intent? Suppose he represented what he was trying to do as, put out V s left eye is this an intent to disfigure? Well, not in one literal sense; D did not use the representation, disfigure, to refer to the type of thing he was trying to achieve. He used, put out V s left eye. But putting out an eye is within the extension of the act-type, to disfigure, you say. Yes, but an intent to put out an eye is not necessarily within the extension of the relevant intention-type, an intent to disfigure. As is well known, just because some particular act x is within the extension of some predicate F in the real world, is no guarantee that the intention to do x instantiates the type of 50 Id. at n. 27. 17 intention, to do F. The only intention tokens that are clearly intentions to do F are those that have F as their Intentional objects. Since very few wrongdoers use representations exactly matching the act-type descriptions of moral norms, we have to get sloppy again. And we do. I am confident that D s intentiontoken will and should be taken to be an instance of the type of intention morality prohibits. D intended to disfigure V no matter which of the representations earlier mentioned he had in mind. So let us reframe our question(s) about Herod s intention in these last directions. Did he intentionally kill John, and did he intend John s death, conceding that he did intend John s head on the platter? Surely the answers to these two classificatory questions are yes, and yes, despite the slippage between what he did and what he intended, as well as the slippage between what he intended and what is prohibited in the way of intentions. If we reframe Philippa Foot s closeness doctrine to be a doctrine about allowable slippage in the two dimensions just described, then we can get out of the closeness doctrine what she sought, a relief of the DDE from moral absurdity. Then if terror bomber intends to explode a bomb right over a person, he intends that person s being blown to bits and thus that person s death. Then a terrorist who places a bomb on a plane to explode it mid-air, in order to collect the insurance proceeds (having insured the plane and not the passengers), intends their death because he intends to destroy their only mid-air possibility of life. Then the surgeon who crushes the skull of a wedged fetus intends the death of the fetus in intending to crush its skull. Etc. One way to view my answer to the closeness problem about the DDE, is as a version of the familiar, broaden-the-discomfort strategy. That general argumentative strategy goes like this: if there is a problem about some particular conclusion, show that that same problem infects some other conclusion that you or your audience holds dear. The strategy tends to motivate critics of the original conclusion to become interested in finding a solution to it in order to save the second conclusion from their own prior criticisms. As applied here, the problem of closeness is not just a problem for the DDE. It is also a problem for the use of intention as a marker of culpability generally. To distinguish an intended from a known harm, or to distinguish a known 18 harm from a risked but unintended harm, one has to solve both of the classification problems adverted to earlier. One will need some doctrine of allowable slippage to do this. And the defender of the DDE needs no more. Moreover, no one wishes to give up the crucial distinction between intended harm and inadvertently caused harm. As Justice Holmes quipped, even a dog knows the difference between being stumbled over and being kicked,51 as children early recognize with their accusation, you did that on purpose. So we need to solve the closeness problem generally and stop pointing fingers at the DDE particularly. C The Moral Force of the Psychological Distinction There is undoubtedly a great deal more to be said here, but let me turn to the second query I raised about the DDE, that about the moral freight carried by the psychological distinction between intent and predictive belief. Assuming we can distinguish intended deaths from foreseen or risked deaths, should we do so in the context of permissible consequentialist justification? One way to get at this moral questions was Herbert Hart s way.52 He stripped the examples to be considered of their putative justifications and compared intent and foresight in terms of their relative culpability. His question was this one: if A intends to kill B, but C only foresees that the act he intends to do will certainly cause D s death, and both kill when neither killing produces any good consequences, is A more culpable than C? No, Hart thought, because both A and C knew that an innocent life would be lost if they acted and both chose to kill to get what they wanted. They both had control over the death of another, and chose to accept that death as an acceptable cost to getting what they wanted. To put some meat on these abstract bones, suppose with Hart that C is seeking to break his mates out of prison, and that he intends to blow up the prison wall to accomplish this.53 C knows that a guard is on the other side of the wall and that the explosion will certainly kill him; but the death of the guard is not necessary to anything C wants or needs. Is C less culpable than 51 52 O.W. Holmes, Jr., The Common Law (Boston: Little Brown, 1881), at p. ___. Hart, Intention and Punishment, supra note 43. 53 From the Nineteenth Century case of Regina v. Desmond, Barrett, and Others, The Times, April 28, 1868. 19 A, who intends the same end (escape of his mates) by the same means (blowing the prison wall), with this difference: A places the bomb at the point of the wall where the guard is stationed in order to kill the guard (who otherwise would prevent the break or identify the perpetrators)? Hart s answer was one of equivalent culpability. It is common to respond to arguments like Hart s by noting that even if there is no difference in culpability in cases where justification is not in issue that does not show anything about cases where justification is in issue. This is the same kind of response as is made in act/omission cases, where people who find act/omission pairs equally blameworthy when no justification is in issue yet find one to be justifiable by good consequences and the other not when such justification is in issue. Logically, of course, this is a possibility. Yet if there is no difference in culpability (in the justification-less cases) between intent and foresight, is it not something of a puzzle why there should be such a marked difference in their potential for consequentialist justification? I confess that I am one of those who thinks that there is a difference in culpability between intent and foresight in the no-justification-in-sight cases. Suppose that the death intended or foreseen does not occur; it was practically certain to occur, but miraculously it did not. Is the foreseeing defendant the one who does an act conscious that it will surely cause death as blameworthy as the one who tries to kill, when in both cases death does not come about? One, after all, tried to kill, whereas the other only (strongly) risked death. Grant me the conclusion I have argued for elsewhere, that the intender is more culpable.54 Then it would not be surprising that the more culpable one could not justify his act by its good consequences but that the less culpable killer could do so. One might object to this line of reasoning by thinking that the wrong (killing an innocent) each killer does is the same, even if their culpability differs, and that it is degrees of wrongness 54 Moore, Prima Facie Moral Culpability, Boston University Law Review, Vol. 75 (1995), pp. 319-333, reprinted in Placing Blame, supra note 5. 20 that track permissible consequentialist justification, not degrees of culpability.55 This objection sounds better than it is. To begin with, the objection rather begs the question. Why is wrongness limited to acts like killing? Why doesn t the wrongness relevant here include intentions as well as actions, intentional killings rather than killings as such? The objection pretty much stipulates its own answer here.56 Second, the objection trades on the action-guiding nature of moral norms. In their (prospective) role of guiding our behavior, moral norms may omit to mention intentions. They may say, do not kill, rather than do not murder (i.e., intentionally kill). But this is only because in their acting-guiding role, mention of intention would be superfluous. Don t kill could be telling us not to aim at killing, even though it doesn t mention intention. Whereas in their (retrospectively) role, moral norms are used to judge behavior already done. When we judge our own or others behaviors, we do care about intentions. Or at least, we could care even if the (prospective) version of the norm made no mention of them. Putting aside culpability comparisons when justification is not in issue, we can test the DDE more directly. Is it intuitive that strategic bomber can justify his bombing (in a just war) in a way that neither terror bomber not malicious bomber can? In World War II the Americans engaged in daylight bombing so that they could see to aim at military targets. (They did this in part for strategic reasons, namely, they thought if they could destroy certain key German industries they could end the war more quickly; but they did it also for moral reasons.) By contrast, the British engaged in nighttime bombing where no such aiming was possible; the aim was to dispirit the German people by destroying the civilian population of the cities. (A similar motivation lay behind the German bombing of Coventry, the Allied fire-bombing of Dresden, the American fire-bombing of Tokyo, and the nuclear bombings of Nagasaki and Hiroshima.) Were there real moral differences at stake here, or were the Americans simply deluded in thinking this, a delusion that cost them extraordinary losses in planes and men? It seems to me that there is something to be said for the DDE in these cases. You may not aim at evil, but you may tolerate it for good enough results. Yet this general conclusion is 55 Joseph Raz voiced this objection to my reliance on intention in the joint seminar we taught in 1989 on consequentialism versus deontology 56 Although one could bolster the objection by bringing in other considerations making it important to confine wrongdoing to action, leaving intention to be distinguished as culpability. See, e.g., Moore, Placing Blame, supra note 5 at pp. 192-193. 21 too crude. How strongly does this psychological distinction mark the borders of permissible consequentialist justificiation? One way the DDE plainly does not mark this border is as follows. In 2003 in Frankfurt, Germany, the Frankfurt Chief of Police faced the following problem. He had in custody the probable perpetrator of a kidnapping-for-ransom of the eleven year old son of a leading German industrialist. The Chief had reason to believe that the son might still be alive, but not for much longer, so to find out the victim s location he threatened the kidnapper with severe torture. (The threat was a colorful one a specialist in torture was being flown in who knew how to inflict pain like no one else.) Moreover, the threat was not empty; the Chief intended to torture the kidnapper to reveal the child s location. As it turned out the threat to torture was enough, the child s location was disclosed, although tragically the child was already dead by suffocation.57 An implausibly strong version of the DDE would say that intending torture, by itself, is absolutely forbidden, even if no torturing is done. The Police Chief thus was unjustified in forming the intention to torture. I find this implausibly strong because suppose one could form the intention to torture knowing that it is very likely someone will stop you from executing your intention.58 Surely that formation of an intention can be justified by good enough consequences. I think the Police Chief was justified in forming such an intention in the actual case, even without a belief that he would likely be prevented from torturing. The Chief was justified in forming the intention, even if one thinks that the Chief would not have been justified in actually torturing. (I am not one of those who thinks it rational to do what you intend when the balance of reasons for the one do not match the balance of reasons for the other, as in Kafka s well known toxin puzzle.)59 So the objects of agent-relative obligation are not intentions simplicatur. They are intentions plus something, presumably something connected to the wrongful actions intentionally done. Yet even this may be too strong to be plausible, for it suggests that intending is necessary to agent-relative prohibition. Yet there seem to be cases where merely foreseen badness will be 57 Recounted in Richard Bernstein, Kidnapping Has Germans Debating Police Torture, New York Times, International, Thursday, April 10, 2003, p. A-3. 58 An argument suggested in Hurd, What in the World is Wrong?, supra note 17. 59 Greg Kafka, 22 enough for prohibition, if that is combined with other things. Philippa Foot gives the example of a hospital in which you intend to manufacture a medicine for five patients, which will save their lives; unfortunately the only way to make the medicine in time is by use of a machine that will pump poisonous exhaust gas into another ward, killing one patient in that ward.60 My sense is the same as Foot s: you may not justify saving the five by using the machine, even though you would only be foreseeing the one s death and not intending it as either an end or as a means. (This suggests a structure that is more criteriological than it is criterial in the mode of combining these different factors.)61 0. One cannot take this elimination of intention too far, however. One cannot think that deontology could prohibit actions alone, with no intention or foresight requirement. This would be the view that, say, the act of killing is absolutely prohibited by itself, with no mental state qualifier. The problem with any such view is that we kill quite often, and think it to be unproblematically justified. We mine coal, build tall buildings, operate transportation systems, in ways that cause some deaths. E.g., our driving an automobile sometimes kills children who jump out in front of us. Yet we justify such activities because in doing them we do not either intentionally or knowingly kill the individuals whom we do admittedly kill in such accidents. We do not think such accidental killings violate agent-relative prohibitions for if they did, how could we justify driving? We need some mental state to be attached to actions as the object of deontological prohibition.62 This pair of examples kidnap, hospital are together suggestive of the role of intention within the content of agent-relative obligation. Kidnap suggests that agent-relative obligation does not simply prohibit intentions, even intentions to torture; rather, it prohibits intentions when conjoined to something else, something like the action of torturing. Hospital suggests that if the something conjoined is too strong a form of the prohibited action say a direct gassing of a patient to death then intention is not needed, foresight conjoined with such action will form the Foot, Abortion, supra note 12 at pp. 152-153. If the mode of combination of these factors were criterial, then we could say the presence of each such factor as intention was individually necessary, and only in combination with the other factors jointly sufficient, for deontological prohibiton. If the mode of combination were criteriological in the sense of the later Wittgenstein, then each such factor was neither individually necessary nor sufficient, nor can one say clearly what combination of factors are jointly sufficient short of all of them. 62 Hurd s conclusion in What in the World Is Wrong?, supra note 17. 61 60 23 content of the prohibition. Now both points together: something related to action must be added to intention to give a plausible content to deontological prohibition, but add too much of that something and intention drops out as unnecessary to the content of the prohibition. We next need to enquire about the nature of that something else in order to see if intention can maintain its balance on this narrow ground. IV. Doing versus Allowing I: Omissive Allowings A. Introduction We now need to pin down what the something else is that is the object of deontological obligation besides intention. Generally speaking, the most obvious candidate is action. Acttypes like killing, raping, torturing, maiming, etc., are what we are categorically prohibited from doing (when we intend them). As opposed to what? Well, one suggestion: as opposed to all the other act-types that are morally ok, such as visiting a sick friend, teaching a class, etc. But these contrasting cases are uninteresting; they depend on the particular content of morality s prohibitions. Wanted is something more general. That more general candidate is usually called an allowing. Killing is prohibited, letting someone die is not; more generally, doing something awful is prohibited, allowing that awful thing to occur is not. The general statement is often called the doctrine of doing and allowing (the DDA ).63 To quote Frank Jackson at a recent seminar of mine: the doctrine of doing and allowing is a mess. 64 Indeed. It will require the rest of this paper to untangle its many strands. Let me start with a distinction of Philippa Foot s,65 seconded explicitly by Warren Quinn66 and implicitly by Frances Kamm.67 There are two kinds of allowings, Foot told us. There are 63 Warren Quinn, Actions, Intentions, and Consequences: the Doctrine of Doing and Allowing, Philosophical Review, Vol. _____ (1989), pp. 287-312. 64 Colloquium on Moore on Causation, Research School of the Social Sciences, Australian National University, Canberra, May 2003. 65 Foot, Morality, Action, and Outcome, supra note. 66 Quinn, Doctrine of Doing and Allowing, supra note 67 Frances Kamm spends her time analyzing acts that remove defenses, but her examples accept omissions as kinds of allowings. See Kamm, Action, Omission, and the Stringency and Duties, University of Pennsylvania Law Review, Vol. 142 (1994), pp. 1493-1512. Kamm s detailed discussion of allowings is in her Morality, Mortality Volume II (New York: Oxford University Press, 1996), Chs. 1-5. 24 omissive allowings, and there are enabling allowings. In the first, the actor does nothing to prevent some mishap that is about to happen, from happening. E.g., he watches the baby drown rather than prevent it from drowning. In the second the actor does something; he enables nature to take its course by removing what was preventing it from doing so. E.g., he switches off the respirator, and the patient dies from lack of oxygen. Finding both a solid metaphysical difference and some moral significance to exist between these two kinds of allowings, I shall consider them separately. Here I focus on omissive allowings. Unlike Foot,68 I see no harm in calling these simply omission cases. The DDA in this context can be called the doctrine of acts and omissions, as it commonly is. As with the DDE, it will be helpful to separate the what is the difference? question from the what difference does it make? question. I shall again pursue the first question first. B. The Metaphysical Distinction Between Acts and Omissions Just what precisely an omission is, and how it differs from action, is a surprisingly contentious issue. Like intentions, this is a topic I have visited before,69 and rather than rehearse what has gone before I shall here simply pick up where I last left off with omissions. My conclusions were these. First, generically in omission is literally nothing at all. An omission to kill is not some ghostly kind of killing. It is like an absent elephant, which is no elephant at all. An omission to kill is an absent event of killing. Not the absence of any particular event of killing it is not an absent act-token for there are no negative events. There are negative propositions about events, such as, James omitted to kill Smith, meaning, it is not the case that James killed Smith. Such negative statements are negative existentially quantified ones: if there is an omission to kill, then what is true is that it is not the case that some instance of the type of event, killing, existed. 68 69 Foot, Morality, Action, and Outcome, supra Moore, Act and Crime, supra __________ at pp.22-34, 54-59, 267-278 Moore, More on Act and Crime, University of Pennsylvania Law Review, VOl. 142 (1994), pp. 1749-1840, reprinted in Placing Blame, supra______, at pp. 262-286. 25 Second, from this generic sense of omissions one can derive a more specific sense. Since omissions generically are absent actions, we make this derivation by plugging in our idea of what actions are. On my theory, defended elsewhere,70 a killing is: 1. 2. 3. 4. 5. A willing (or volition) having as its object some bodily movement, when That willing causes The bodily movement willed, and That bodily movement causes Death. An omission to kill is the absence of any one or more of the above five items. Either the victim doesn t die (no. 5), or if he does, not because of my acts (no. 4), or I don t move (no. 3), or if my body moves it is not because I willed it to do so (no. 2), or I don t even try to move my body (no. 1). Notice that on this view I can omit to kill while I am quite busy dancing a jig, writing an article, or doing anything else. Nothing requires that if I am to omit to kill I must remain motionless. Remaining motionless is one way I can omit to kill, but it is not the only nor even the most common way. Third: if we are to compare actions like killing with omissive allowings like letting die, we need to make sense of omissions to save from death, not omissions to cause death (i.e., kill). The positive moral duties we plausibly have require as not to omit to prevent deaths, not not to omit to cause deaths. So we need to make sense of omissions to save, that is, to prevent something from happening. This requires us to understand preventions, because the omissions we care about are absent preventions. A first temptation is to think that this is straightforward in that we can apply the above five-fold, causal analysis of killing to preventing death. A saving of Smith would then be a willed bodily movement that causes the survival of Smith. Yet Smith s survival is just the absence of his dying. We must penetrate the seemingly positive language ( surviving, starving, keeping of the secret, ignoring the pleas of the drowning man, etc.) to the 70 Act and Crime, supra 26 underlying reality: to survive is not to die.71 Now the problem: an absence of Smith dying at some particular time is not a particular thing, a kind of ghostly dying (a near death experience? ) What makes the negative proposition ( Smith did not die ) true is that there was no instance of the type, Smith dying, that occurred. If causation is a singular relation between either event-tokens or states of affairs as particulars, then how can there be causation of a nothing?72 Here my concern about the implications of meta-ethical naturalism for methodology intrudes. The late Warren Quinn once urged that ethical accounts like doing/allowing should be neutral between competing versions of causation, being framed so as to accommodate any plausible metaphysics of causation.73 I to the contrary think there is no help but to do the metaphysics, see what you think causation is, and use your metaphysical conclusions in your ethical argumentation. The conclusion of my forthcoming book on causation is that causation is a singular relation between token states of affairs.74 On this view of causation, you can t cause a nothing anymore than you can be caused by a nothing. Absent events, or absent states of affairs,75 can no more be effects than they can be causes. Causation does not relate absences. Plug that metaphysical hogchoker in here and preventions become a puzzle. Why do they look causal if they are not causal in structure? My answer76 (and Phil Dowe s)77 is that preventings are intimately related to a causal structure even if they are not themselves such a structure. When Jones saves Smith from drowning and thereby prevents Smith from dying, Jones does do something. He volitionally moves his body and that bodily movement causes something. If Jones hauls Smith out of the water with a rope, Jones movements cause Smith to exit the water. Since the water was the only place where Smith could have drowned right then, It is important in general to reject the seeming indifference of our language at labeling what is really a presence versus what is really only an absence. See Moore, Causal Relata, Annual Yearbook for Law and Ethics, forthcoming, 2005. 72 Id. 73 Quinn, Doctrine of Doing and Allowing, pp. 293-294. 74 Michael Moore, Causation and Responsibility, Oxford University Press, forthcoming. 75 Cogniscenti in metaphysics may raise their eyebrows at this because on one reading of states of affairs such things are propositional and may easily be negative. I intend the other reading of states of affairs, what Peter Menzies calls real situations, what David Mellor calls facta, and what David Armstrong calls simply, states of affairs. See Moore Causal Relata, supra note 71. Incidentally, in the text that follows I shall use events and states of affairs interchangeably, because for these purposes I don t need the distinction between them. 76 Id. 77 Dowe, Physical Causation (Cambridge: Cambridge University Press, 2000). 71 27 what Jones caused was a state of affairs incompatible with Smith drowning. Preventions are the causing of things incompatible with the occurrence of any instance of the type of thing prevented. And omissions are simply absent preventions. Fourth, and following on the discussion in third, above, omissions cause nothing. Nothing comes from nothing, and nothing ever can is good metaphysics, as well as catchy lyrics in musical productions.78 Absent elephants grow no grass by their absence; absent savings cause nothing, and certainly not the deaths they fail to prevent.79 To sum up these tersely stated, here undefended conclusions with respect to one act-type killing: An omissive letting die is not a kind of killing. Such an omission contains no act, not even a mental act of willing, and the absence of any act is not itself an event or a state of affairs; an omission likewise is not a cause of the death it fails to prevent. To be held responsible for a failure to prevent Smith s death is thus very different from an act that causes Smith s death, that is, a killing. The difference lies in the existence of real states of affairs and of real causal relations for killings but not for lettings die. C. The Moral Force of These Metaphysical Distinctions We now need to address whether these significant metaphysical distinctions make any moral difference. Again, the difference we are interested in is the difference in consequentialist license: can we consequentially justify omissions by their good consequences (read as, the states of affairs I cause by doing something else), whereas we cannot justify a paired action by its good consequences? Can I, for example, justify not throwing you the rope and thereby not saving you from drowning because I throw the rope to five others in the water and thereby save them from drowning? By contrast, presumably I cannot justify drowning you (to get you off the rope) in order to throw the rope you had to five others in the water (and thereby save them from drowning). I think the answer to the first question is very strongly yes. When it comes to saving the lives of strangers just about everybody I know thinks that you at least may omit to save one 78 79 Julie Andrews and Christopher Plummer, Something Good, in Rogers and Hammerstein, The Sound of Music. Moore, Causal Relata, supra note 71. 28 in order to save five. (I think in addition you must do so given the consequentialist background mentioned earlier). Consider a decision by Winston Churchill during World War II.80 The British had broken the German coding device called Ultra, and learned that the Germans planned to bomb Coventry to dispirit the British populace. Churchill could have prevented the killings of the citizens of Coventry by alerting them; but this also would have tipped off the Germans that the British had obtained the German coding-device. Churchill justified his not saving the citizens of Coventry by his thereby saving many more British lives in a shorter, more successful war effort. His Socialist War Cabinet member urged him not to play God with people s lives. Yet given the consequences, wasn t Churchill right? Whereas if Churchill had had to kill the citizens of Coventry in order to maintain the British advantage in intelligence say some Coventry citizen was a spy who learned this but it is unknown which citizen it is surely we may not kill to achieve the same good consequence. People often at least profess to find this moral difference puzzling. They point to cases where no justification is in sight and contrast killings with omissions to save from death, finding both to be equally blameworthy. Notice that this is to replicate Herbert Hart s assumption with respect to the DDE, where Hart assumed that justification-less culpability comparisons between intention and foresight were determinative the of truth of the DDE.81 The most famous example is probably that of James Rachels.82 Two uncles are each bathing a baby who is their nephew; each wants the baby dead so that he can inherit the family wealth. One uncle is motivated by greed to drown the baby in the bathtub by holding its head underwater. The other is luckier: the baby slips under the water and cannot save itself. This uncle, for the same greedy motive, does nothing, and the baby drowns. Rachels own sense was that the two uncles are equally blameworthy, and this was taken to show the moral equivalence of acts and omissions. 80 Churchill s decision is described by William Stevenson, A Man Called Intrepid (1976). I owe this example to Peer Pedersen. 81 Hart, Punishment and Intention, supra 82 James Rachels, Active and Passive Euthanasia, New England Journal of Medicine, Vol. 292 (1975), pp. 78-80; Rachels, Killing and Starving to Death, Philosophy, Vol. 54 (1979), pp. 159-171. 29 There are three responses to Rachel s intuition here. One response is to deny the relevance of justification-less culpability comparisons to the question that interests us here, viz, when putative consequentialist justifications are in the comparison cases, are such justifications available for omission but not for actions? A second response has been defended at length by Francis Kamm.83 This response concedes the equivalent blameworthiness of the two uncles but refuses the general conclusion about acts and omissions. (This latter refusal is based on Kamm s very nice insight that acts and omissions have properties constituting them as acts or omissions and that those properties do make a moral difference; yet in particular cases such properties may be possessed by both of an act/omission pair, so that that pair of cases is morally equivalent.) There is much to be said for both of these responses. The first response strikes many people as very intuitive: even conceding equivalent culpability between the two uncles when no justification is present, when justifications are present they are much more available to justify omissions to save than actions of killing. The second response also is developed by Frances Kamm with great ingenuity. Still, I shall only pursue a third response to Rachels: I don t share his intuition of equivalent blameworthiness. To see why not, alter Rachels case a bit. We need to rid the cases of relationships between the uncles and the babies. Suppose no familial relationship exists in either case, nor is either person in charge of the baby s bath. So: A and B are each having lunch on respective park benches. Each have an unattended stranger baby toddle over to a puddle next to them. The baby next to A falls face-forward into the puddle and cannot get up. The noise it makes irritates A who therefore does nothing and watches it drown (this returns quiet to the park). The baby next to B is playing happily but noisily in the puddle. To return quiet to the park, B drowns it in the puddle by holding it face down in the water. Surely we agree that A and B are both dreadful human beings. Our aretaic judgments of their character, as revealed in these choices, is extremely negative. Indeed, if we suppose (as did Rachels in his original case) that A would have drowned a noisy baby if by luck it wasn t already 83 Kamm, Morality, Mortality II, supra 30 drowning, we should think A s and B s characters are equally awful. But these aretaic judgments mislead us about the deontic comparison we are supposed to be making: both are guilty of wrongful choices at a razor s point of time, but are they equally wrongful? One of them chose to kill, and killed, whereas the other chose not to save, and didn t. To be sure if one regards such deontic judgments to be elliptically expressed aretaic judgments the character theory of responsibility of Hume and his intellectual descendents then one won t accept my distinction. But I am not one of those.84 Our responsibility for character is real enough, but it is a different form of responsibility from our choices and what those choices cause on particular occasions. Those who persist in thinking that people like A and B are not just equivalently awful in their characters but are also equivalently wrong in their choices face a puzzle. Typically such people also share Churchill s intuition: we can justify failing to save people a lot more easily than we can justify killing them. If the not-saving were just as wrong as the killing, why is there such a difference in the justifiability between these two, supposedly equivalent wrongs? So think again about the conclusion of equivalent wrongness. Perhaps it helps (as Kamm suggests) to consider what are our secondary duties of prevention with respect to acts versus omissions.85 If we could prevent ourselves from killing versus prevent ourselves from failing to save, are such secondary duties to prevent equally stringent in both cases? There are two kinds of cases here, what Kamm calls pre- and post- effort cases. (1) Pre-effort. A has joined a gang. A comes to realize that if he stays in the gang some of their various activities will require him to kill a completely innocent person; A knows himself well enough to know that when such occasions arise, he will yield to the group pressures by killing such innocents. B by contrast has joined a gang made up hard-hearted libertarians. B comes to realize that if he stays in the gang he will yield to the group pressures never to rescue anyone in peril. Supposing the probabilities (of the occasions arising in the future, and of A s See Moore, Choice Character, and Responsibility, Social Policy and Philosophy, Vol. 7 (1990), pp. 29-58, reprinted in Placing Blame, supra 85 First used in her comment on my Act and Crime, The Stringency of Duties, supra ; developed at length in Morality, Mortality II, supra 84 31 and B s deadly responses to those occasions) to be equal, do A and B have equivalently stringent duties to quit the gangs? It seems to me much more obvious that A should quit his gang of killers than that B should quit his band of hard-hearted libertarians but then it seems to me much more obvious that A shouldn t kill than that B should not omit to save.86 (2) Post-effort. At t1 A starts a boulder down a hillside to kill V, his old enemy. The boulder will kill V unless A stops it, which he can do by rushing in front of the boulder at t2 and diverting it. Such diversion will probably kill A, however, if that diversion is successful in saving V. B by contrast happens upon a boulder rolling down a mountain toward V. The boulder just happened to break loose through no act of B s. At t1 B could have stopped the boulder early in its trajectory at no risk to himself; B didn t do so, however, because V is his old enemy who he wanted to die. At t2 B has another chance to stop the boulder. He can stop it by rushing in front of the boulder and diverting it. Such diversion will probably kill B, however, if that diversion is successful in saving V. Does A have more of a duty to place his life on the line than B? My own sense is that he does but then of course I think A was under a more stringent duty, at t1 not to kill than B was at t1 to save. It is difficult to argue for conclusions one finds intuitively obvious, and I must confess that I find the moral non-equivalence of acts and omissions to be obvious. So I shall cease beating (what for me is) a dead horse. We should clarify the intuitive conclusion, however. I am not reaching the libertarian conclusion that the only duties we have are negative duties not to cause harm by our actions. We do have consequentialist-based duties to prevent harm, even to strangers. I thus reject the idea that we have strong permissions to omit, across the board (the libertarian position). Rather, the conclusion I reach here is only that we have weak permissions to omit in many omission cases (not in all we do have positive, deontological obligations to the near and dear). That is, in omission-to-rescue-strangers cases we are not bound by agent-relative obligations and thus we are back in the land where consequentialist reasons hold sway. I have thus eschewed arguments that illicitly trade on intuition about strongly permitted omissions. I refer to arguments based on the triggers of our positive duties to strangers. Such 86 The qualifier has kept me from placing much weight on these arguments. See my More on Act and Crime, supra , at pp. 285-286 (in Placing Blame). 32 triggers consist in not only the danger facing some victim V from which we can save him; but also in the burden we must undertake if we are to save V. Lord Macaulay s old example:87 a surgeon need not take a train from Calcutta to Meerut in order to save someone not his patient, even though unless the doctor takes the train that person will die. Whereas, if the doctor can prevent some act of his from becoming the cause of that person s death (so that the doctor would then have actively killed the person), then the doctor is obligated to take the train. I am with Lord Macaulay: when the risk and even inconvenience are too great, the doctor is permitted not to take the train. And this is true even though that same level of inconvenience would hardly justify one is not completing a killing in which one is engaged. But notice the doctor is strongly permitted: the doctor does not have to justify his inaction on consequentialist grounds (nor can that inaction he so justified if one is balancing inconvenience versus life in a consequentialist balance, life wins). Such strong permissions not to save of course negate the existence of any obligation to save (and thus a fortiori negate the existence of any deontological obligation to save); but this shows little about the shape of our deontological obligations from the availability of consequentialist justifications. Strong permissions negate deontological obligations on their own, not by virtue of the shape of those obligations. V. Doing versus Allowing II: Non-Omissive Allowings A. Introduction If it were plausible that the objects of our deontological obligations were executed intentions, then we could stop where we are. Intending to kill, and then executing that intention in action would be enough, irrespective of success in achieving the object of the prohibited intention. Yet notice that unsuccessfully executed intentions are no more plausible than intentions alone as marking the boundaries of permissible consequentialist justifications.88 Revert to the example of the Frankfurt police chief and the kidnapper of the child: if the Chief knew that his attempt to torture the kidnapper would probably be thwarted, he might well be justified in trying to torture even if he would not have been justified in intentionally torturing. 87 88 Macaulay, Notes on the Indian Penal Code, 1837, in his Works, Vol. vii (New York, 1897), at p. 494. See Hurd, What in the World Is Wrong?, supra 33 The object of deontological prohibitions need to include whatever it is that makes something not only an act but an act of torture or of killing, rape, mayhem, theft, etc. We thus need to add something beyond intention plus acting to capture what we are forbidden to do. If we examine the structure of the complex verbs naming the kinds of actions deontological morality seems to prohibit, there are a variety of feature we must add.89 One feature that is common to all act-types morality forbids, however, is that of causation. The willed bodily movement (that is at the core of all actions) must cause death (for killing), penetration (for rape), asportation (for theft), disfigurement (for mayhem), etc. Although the point is contentious, I have elsewhere argued at length that whatever else is required for an act to be, e.g., a rape (such as lack of consent), also required is causation of something.90 Secondly (and even more contentiously) the connection between prohibited act-types like killing and causation is a tight one. It is uncontroversial that if x kills y, then x causes the death of y. Much more controversial is the converse: if x causes the death of y, then x kills y. The reason that many philosophers such as David Lewis, Judy Thomson, Jonathan Bennett, and Donald Davidson disagree with this latter implication is because they think causation to be a much less discriminating relation then I think it to be.91 They thus think, for example, that if I suggest to you where to find your victim, and you find him and kill him, then I have caused the victim s death but I haven t killed him you did that. Whereas I think that although you didn t do much to kill the victim in such a case you didn t do much to cause his death either. It will not ultimately matter much here whether one agrees with this second point of mine. Whether one agrees does, however, make a difference to how we should describe the distinctions I wish to draw shortly. I think they are all causal distinctions, and will discuss them in those terms. Those with less discriminating notions of causation will have to reframe the same distinctions in terms of the attributes of action verbs like killing 92 etc. I think that such a 89 90 I explore these in Act and Crime, supra , chapter 8. Id. 91 See citations, id. 92 Thus, David Lewis extremely non-restrictive notion of causation (as counterfactual dependence) does not track at all the limitations I explore; but then he rediscovers many such limits in the semantics of causally complex verbs of action like kill, See Moore, For What Must We Pay? Causation and Counterfactual Baselines, San Diego Law Review, Vol. 40 (2003), pp. 181-1272, at p. 1221 n. 117. 34 translation of the distinctions (and arguments for them) can be done, but not being very ecumenically disposed towards other people s theories of causation, that is not my concern. I thus begin a discussion of the second kind of allowing distinguished earlier, the enabling rather than the omissive kind. Consider the familiar example of medically justified, passive euthanasia. P is hooked up to a respirator which is the only thing keeping her alive. Her quality of life is very poor because she is in a coma, her condition is incurable, and the machine she is on can save another, higher quality, life. Under the doctrine of passive euthanasia, her doctor may unhook her, allowing her to die. But in the exact same circumstances, giving the exact same justification, her doctor may not give her a lethal injection. That would be active, not passive euthanasia, forbidden by the deontological prohibition against killing. Notice in the passive-euthanasia kind of allowing, the doctor is not an omitter. Turning off the respirator is an action. Although some courts have said that these are omission cases,93 they are surely stretching the concept of omission beyond what it can tolerate. Some critics of these courts, however such as my former self were wrong to think that such passive euthanasia cases, being act and not omission cases, must simply be strong permission cases.94 Both myself and the courts I was criticizing overlooked the possibility we need here to explore: that these are neither omissions nor active causings requiring justification, but are non-omissive allowings. I shall begin with the most intuitive kind of non-omissive allowings, a kind illustrated by the passive euthanasia example above. B. Non-Omissive Allowings 1: Removing a Defense One Was Entitled not to Have Provided 1. Introduction Barber v. Superior Court, 147 Cal. App. 3d 1006, 195 Cal. Rptr. 484 (1983); Airedale NHS Trust v. Bland, [1993] All Eng. Rep. 821. 94 See my Act and Crime, supra , at p. 27. I repeat the mistake in my More on Act and Crime, p. 276 (in Placing Blame). 93 35 There are seemingly many other well known examples of non-omissive allowings besides passive euthanasia. Judith Thomson s violinist is one.95 In the course of defending the right to abortion, Thomson imagined that you go to the symphony, and fall asleep. You awake to find the lead violinist hooked up to you in a life support system that only you could provide. After nine months he can be unhooked, but if he is removed before then he will die. May you unhook him? Thomson intends that you agree with her in saying yes. As well known are the Chris Boorse/Roy Sorenson tales of ducking.96 Take the one of the grizzly bear, a tale that has descended to the level of cocktail party conversation: you and I are camping in the woods. A ferocious and hungry grizzly bear charges out of the woods towards both of us, although I am closer to him. I jump into my running shoes. You shout at me, fool, you can t outrun a grizzly. I shout back as I take off, I only have to outrun you. May I duck by outrunning you, even though that means the grizzly will kill you? Boorse and Sorenson intend that you will agree with them in saying yes. The violinist and ducking cases share a common structure with the passive euthanasia cases. It is a two part structure. First, the acts are all done against the backdrop of an imminent threat to the victim: a death from disease or grizzly bear. All I do by my action is enable that threat to be realized by removing the victim s defense to it: I remove the respirator, remove myself as a life-support system, or remove myself from between you and the bear. Second, the defense you had was one I was providing. The doctor(or her surrogate) hooked up the respirator to start with; your body is the only defense the violinist has; my body is your only defense against the bear. Call this the baseline point: by doing what he did the actor was only returning the victim to the state she would have been in had the defense never been provided to start with. Yet despite this common structure I will focus on the passive euthanasia cases alone. This, because the doing/(non-omissive) allowing distinction operates too strongly in the violinist and ducking cases. In the latter cases, if you agree with the intended resolutions, that will be because the actors are strongly permitted to unhook or to run. They are not merely weakly permitted, meaning they must have good consequentialist reasons to do what they do. My sense 95 96 Judith Jarvis Thomson, A Defense of Abortion, Philosophy and Public Affairs, Vol. 1 (1971), pp. 47-66. Christopher Boorse and Roy Sorensen, Ducking Harm, Journal of Philosophy, Vol. 115 (1988), pp. 115-134. 36 is that they are entitled not to balance the consequences in preserving their freedom or their lives. If I am correct in this last conclusion, then these cases are not apt ones with which to test the line of permissible consequentialist justification, the line which marks the borders of our deontological obligations. Focusing on passive euthanasia-like cases only, I shall again proceed by first asking after the natural properties involved in this kind of doing/allowing, and then asking after the moral significance of those properties. 2. The Metaphysics of Removing a Defense to a Death While Not Causing That Death We saw earlier that preventings are not themselves causal; a saving someone from death is not the causing of a non-death. Similarly, I argued that omissions are not causal; an omission to save someone from death is not a causing of death (even when a death occurs that the saving omitted would have prevented). Here we need one more piece of the puzzle. When the doctor does something like pull the plug on the respirator, the moral philosophers call this a removing of a defense. The metaphysicians call this a double prevention. 97 What the doctor literally does is prevent something (the respirator) from doing what it was doing, namely, preventing the patient s death. The doctor prevented a prevention of death. Thus the label. Double preventions cannot be causal if preventions and omissions are not causal.98 Consider an old example of mine:99 DP sees his old enemy V drowning in the ocean, which makes DP happy. Then DP spies a lifeguard, L, getting ready to save V. So DP ties up the lifeguard, and V drowns. DP has prevented the preventer, L, from preventing V s death. But DP has not caused V s death. What DP caused was L to be tied up, which was incompatible with L rescuing V. L therefore omitted to save V. If L s omission is not a cause of V s death, how can DP s act be a cause of that death? DP only caused something incompatible with L saving V; but 97 See, e.g., John Collins, Preemptive Prevention, Journal of Philosophy, Vol. 97 (2000), pp. 223; Michael McDermott, Redundant Causation, British Journal of the Philosophy of Science, Vol. 46 (1995), pp. 523- ; Ned Hall, Causation and the Price of Transitivity, Journal of Philosophy, Vol. 97 (2000), pp. 198; Phil Dowe, Physical Causation, supra 98 Moore, Causal Relata, supra 99 Moore, Act and Crime, supra at p. 278 n. 42. 37 DP no more caused an absence of saving by L than L caused the death of V by his omission to save. One can see this even more straightforwardly if preventions were causal but omissions were not. Then we could say, DP caused L s omission to save; but since L s omission to save cannot be a cause of V s death, and since there is no other way DP caused V s death but through L s omission, then DP did not cause V s death. Many people find this troubling because they perceive, correctly, that DP is morally responsible for V s death. Indeed, in criminal law DP is plainly guilty of murder. But we don t want this moral conclusion to drive the metaphysics here. Perhaps there are non-causal bases for responsibility, and unjustified double preventions is one of them. After all, it is true that V s death counterfactually depended upon DP s act, and perhaps this is enough for a kind of moral responsibility for V s death.100 Recall also that the lifeguard scenario lacks the second requirement for an allowing of this type. The defense V had against drowning was L, and that defense was not one DP had been providing. Suppose L was not a lifeguard but just a private citizen with excellent life-saving skills in the water. He is initially disposed to save V and even takes a few steps towards the water with that intention. But he suddenly realizes that saving V would clearly violate the code of the Hard Hearted Libertarian Gang of which he is a member, so L ties himself up to prevent his yielding to any more warm and fuzzy temptations. Make the defense one L himself is providing, as immediately above, and his removal of that defense his preventing himself from preventing now looks no more causal than a simple omission. Indeed, this kind of double prevention looks very much like an omission (because of the shortness of time during which the preventer a person disposed to rescue existed). So repair to the example of the doctor turning off the respirator. His act of turning it off looks not at all like an omission because his earlier act of turning on the machine was complete the machine was preventing the patient s death and would continue to do so until he turned it off. Still, the 100 Argued for in Moore, Causation and Counterfactual Baselines, supra 38 doctor merely allowed the patient to die because his act simply let nature take the course it would have taken had not he, the doctor, temporarily prevented it from doing so. The patient simply died the death she would have died had not the doctor intervened to start with. There is obviously a large counterfactual element at work in this second requirement for this kind of an allowing: is the harm allowed similar to the harm the victim would have suffered had no defense been supplied initially? Yet also at work is a moral element here, about the moral acceptability of the baseline to which return is made. If the good consequences justifying turning off the respirator at t2 had been present and known to be present at t1 when the respirator was turned on, would the doctor have been justified in not turning on the respirator? If the answer to this question is yes, then that baseline (when the patient was dying) is an appropriate one to which to return. One sees the moral element clearly in the violinist and the ducking cases as well. It is because I have a strong right to bodily integrity (that was violated by the hooking up of the violinist to me) that makes it so appropriate that the baseline to which return is made be the violinist unhooked (rather than hooked) up to me. It is because I have a strong privilege to defend myself against bears and other deadly perils that I may remove my body even though it is your only defense against the bear. (Notice it is because the baselines here are so strong I am strongly permitted not to hook up to needy violinists or to be eaten by bears that my returning to these baselines is also strongly permitted.) The appropriateness of the baseline is of course a moral question. Yet a patient naturalism will not be offended by this. We will need to give some naturalist account of why there is no duty to provide a respirator to permanently comatose, incurable patients when other patients with better life prospects are in need of that respirator. But this we can give, in pretty straightforward consequentialist terms. And the same is true for our rights to bodily integrity. 3. The Moral Significance of Removing Defenses So allowings metaphysically are double preventions when the preventer prevented was something the actor provided but was (in retrospect) entitled not to provide. Such allowers do 39 not cause the harm that they prevented something (they supplied) from preventing; why should this make a moral difference? Specifically, are such allowings eligible for consequentialist justification whereas causings of those same harms are not? Plainly the moral appropriateness of the baseline, together with the counterfactual judgment that that is where we are returning to, does a lot of the work here. We can see this by recalling how responsible was DP in tying up the lifeguard so that V drowns. That is a true double prevention case too, where DP does not cause V s death. Yet DP was not returning V to an appropriate baseline; DP had not provided the lifeguard and so was not entitled to tie him up. The appropriate baseline was thus not, V drowning with no lifeguard rescue in sight. Does the causal/double prevention distinction do any moral work here? Suppose the only way the doctor could unhook the respirator or that Thomson could unhook herself from the violinist was to directly cause death, by lethal injection, knife through the heart, or whatever. Would such killings-that-return-to-the-baseline also be justified? Your answer may be yes, as I believe it is for Frances Kamm,101 but is it so easily or obviously yes? Doesn t the addition of such strong, direct causation of death make us hesitate here?102 There is something morally relevant about the difference between causing death, and allowing nature to cause death by double prevention. Consider the stress and duress techniques used by the British during interrogations of IRA suspects in Northern Ireland, techniques adopted by U.S. interrogators post 9/11 in Afghanistan.103 These are largely double prevention techniques e.g., sleep deprivation, where the interrogator prevents one from sleeping. We know from no less an authority than Shakespeare that sleep is what knits up the raveled sleave of care, and that without such knitting we fall apart and eventually die.104 These techniques make no pretense of returning the detainee to some baseline that is morally appropriate (if the interrogator was the one making sleep possible to start with, then this might be such a baseline). Nonetheless, such techniques are regarded as torture light rather than true torture in part because of the role of nature in making things unpleasant for the detainee. 101 102 Kamm, Morality, Mortality II, supra See my Causation and Counterfactual Baselines, supra at pp. 1261-1263. 103 See note 8 supra. 104 Ernest L. Hartmann, The Function of Sleep, (New Haven: Yale University Press, 1973). 40 In any event, both elements together give us one place where consequentialist justifications are appropriate: we can justify acting in such a way as to prevent something from preventing some harm, so long as we supplied the thing preventing the harm and we were entitled in light of what we now know) not to have done so. We now need to see what else might be included as allowings of the non-omissive kind. C. Non-Omissive Allowings 2: Accellerating a Death About to Happen Anyway Once one sees the distinctive part played by each of the two elements of a removal-ofdefense-allowing, it is tempting to pry the two elements apart to see what we get. First I shall eliminate the double-prevention/cause distinction. So we will consider clear causing-death cases. But we will retain the second element above, the counterfactual judgment about returning the victim to pretty close to where he would have been had we not intervened to start with. A narrow interpretation of this second element would yield the following kind of case. I heroically saved you a year ago; it was so risky, what I did, that I had no obligation to do it, under anyone s view of positive duties. But I did it and now I need your organs to save a number of my relatives. Can I now kill you and take them? Surely not. That I gave you life when I didn t have to does not free me from my deontological obligation not to kill you a year later. So this narrow interpretation of the second element is not what is wanted here. Broaden the interpretation of that second element: what is really doing the work is the counterfactual judgment. When I remove a defense I was providing I return you to where you would have been without me, which is dead. True, the death you die when I remove the defense is not literally the death you would have died had I not provided the defense to begin with. (Recognizing that transworld identity claims are notoriously difficult, this one seems true.) But the difference may seem immaterial. All I did really was delay your death (by providing you a defense against it), but not by as much as I could have delayed it (because I later removed that defense). What if we reverse this somewhat sloppy identification of deaths. Suppose what I do accelerates slightly the death you were about to come to anyway? Could that too be chalked up 41 to the nature that was killing you (even though I finished the job for nature)? It is clearly widely intuitive to give an affirmative answer to this question. In the lifeboat cases, where all will die unless some are thrown out, we should throw some out because they are about to die anyway.105 Or, when all in the lifeboat will soon starve to death unless one is killed and eaten, one should be killed and eaten since the one killed would soon die anyway.106 Or in the mountain climbing cases, where all on a rope will be pulled to their deaths unless the rope is cut, the rope should be cut those dangling down rope are dead no matter what.107 In the wedged fetus cases, if both mother and fetus will die unless the fetus skull is crushed and the fetus is removed (saving the mother but not of course the fetus), we should kill the fetus.108 Or in Sophie s Choice-like situations, when the Nazis will kill all of a group unless you choose one to be the example, better to choose the one rather than have that one killed anyway along with all the others.109 I call these the almost dead cases, because the one we kill is almost dead anyway at the hands of someone else or of nature. Billy Crystal tells us in the film, The Princess Bride, that there is a big difference between being almost dead and being all dead. Well, not here Billy. Although it is far from uncontroversial, I find it very intuitive that killing is justified in all of the above cases. What is guiding that judgment is not a causal distinction. All of these are cases of pre-emptive overdetermination.110 You remember: two fires are approaching victim s house, each sufficient to burn it to the ground. Yet Fire #1 reaches the house first, burning it to the ground; Fire #2 gets there later and does not burn the house, although Fire #2 would have burned it to the ground had United States v. Holmes, 26 Fed. Cas. 360 (Cir. Ct. F. D. Pa. 1842). The ability to jettison some from the lifeboat to save the rest was dictum for the court, inasmuch at in the court s view one needed to have a lottery or other fair procedure to select who would be sacrificed. Justice Cardozo s intuition about Holmes was that even if a fair procedure had been adopted to select those who must die, no one could throw them overboard although they could jump in themselves and they would be supererogatory to do so. Cardozo, Law and Literature. (1930), pp. 110-114. 106 Regina v. Dudley and Stephens, 14 Q.B.D. 273 (1884) I obviously disagree with the court s holding in Dudley and Stephens, as did Glanville Williams. Criminal Law The General Part. 107 This is one of the Model Penal Code examples of when it is permissible to kill. Commentaries to the Model Penal Code (Philadelphia: American Law Institute, 1985). For a real life example, see the well told tale (recently made into a film) by Joe Simpson, Touching the Void (New York: Harper Collins, 1988). 108 The example of Pope _______ in his 1957 Encyclical. See Hart, Punishment and Intention, supra 109 Aside from the film Sophie s Choice, see Bernard Williams, A Critique of Utilitarianism, supra , where Williams famous example of Jim having to choose who in a group of villagers should be shot (else all will be), is another example of this sort. 110 On the idea of pre-emptive overdetermination cases, see Moore, Causation and Counterfactual Baselines, supra , at pp. 1238-1243, 1249-1257. 105 42 the house still existed to be burnt. In no sense of the word does Fire #2 cause the destruction of the house; in a full sense of the word Fire #1 causes the destruction of the house. Mutatis mutandis for all of the above examples. We kill in each of them. The other factors would have killed but we didn t give them a chance. So what guides our judgment here is purely the counterfactual judgment: if we didn t do what we did, the victim would soon have died anyway. We thus accelerated his death a little, just as in the removal of defense cases we delayed it a little. But we didn t in either case change the world very much from where it was heading without our help. These are not of course true allowing cases, except by honorific designation. These are killing cases. Yet what we have done is export (Kamm s term)111 a property of true allowings (of the defense removal kind) to killings. Such a property can be possessed by both true allowings and by true killings. Not surprisingly, that property makes as much moral difference for killings as it did for allowings. We are justified in killing by the good consequences of doing so in these cases. How one accommodates this clear, weak permission into one s description of the objects of categorical obligation is a puzzle. We can t do what we did to delayings (removal of a defense allowings): there we fastened onto the double/prevention versus cause distinction to say one was not really killing in those cases. But in the acceleration of death cases we are clearly causing death, and clearly killing. I once opined that we should accommodate accelerations as exceptions to categorical norms, like self-defense, defense of others, etc.112 Perhaps that is right, although such exceptions tend to be strong permissions whereas this one seems clearly consequentialist in character. More tempting is to liken these almost dead cases to the mostly dead cases. Part of what is at work with permanently comatose patients in passive euthanasia cases is that they have already lost a large part of what makes a member of the human species into a person (just as 111 112 Kamm, Morality, Mortality II, supra Moore, Causation and Counterfactual Baselines, supra at pp. 1266-1267, and at p. 1267 n. 205. 43 fetuses are not yet fully persons). The almost dead typically are not physically deteriorated to the point that rationality, autonomy, emotionally, and the other attributes of personhood are lost.113 So the almost dead are not literally the mostly dead. Nonetheless, what the almost dead have is a lot less than the rest of us have. That is why in tort law when we kill one of the almost dead we for example, electrocute a boy who is already falling to his death114 his life is valued at a fraction of the life of one with a normal life expectancy. In light of this, we might think that our categorical obligation not to kill a person is not very clearly violated when we (clearly enough) kill what is in reality but a thin time/slice of a person. D. Non-Omissive Allowings 3: Aiding Another Person to Cause Now let me turn to the other of the two elements in the allowings of the removal of a defense kind, which is the purely causal/double prevention point. Again, a narrow interpretation of this element produces cases like that we have seen before, such as tying up the lifeguard. The intuition of serious responsibility is quite strong in such cases, until we add in the other element (of having ourselves provided the defense removed when we didn t have to). Contrast the lifeguard case with this one.115 Judge Tally heard that the Skelton boys were riding to another town to kill one Ross because Ross had seduced the Skeltons sister. The Judge also wanted Ross dead but heard that a warning telegram had been sent to Ross s town to alert Ross to the Skeltons imminent threat. Tally sent a countermanding telegram to Ross town, instructing the telegraph operator not to deliver the warning telegram. He didn t, the Skeltons caught up with Ross and killed him. How should we adjudge Tally s responsibility for Ross death? This is a double prevention case: Tally by his act prevented the warning telegram from preventing Ross killing by the Skeltons. So Tally did not cause Ross death. Yet since Tally is not the one who sent the warning telegram, he is not a mere allower in the removal-of-a-defense I explore many attributes of personhood in Moore, Law and Psychiatry: Rethinking the Relationship (Cambridge: Cambridge University Press, 1984), ch. 2. 114 Dillon v. Twin State Gas and Electric Co., 85 N.H. 449, 163 A. 11 (1932). 115 State ex rel Attorney General v. Tally, 102 Ala. 25, 15 So. 722 (1894). 113 44 sense and he seems thus quite responsible for Ross death. Yet compare his responsibility to that of Ross actual killers, the Skeltons. Is Tally as much to blame as those who shot Ross to death? This is a familiar issue to students of the criminal law, because what is at issue is the degree of accomplice liability. Are accomplices who do no killing themselves as blameworthy as the principals who do? Despite the formal doctrines of modern Anglo-American criminal law equating accomplices to principals, I think the general answer is yes. Absent special facts about psychological domination or manipulation by the accomplice, accomplices like Tally are less blameworthy than those whom they aid. Does this difference in blameworthiness when justification is not in sight make the kind of difference in which we are interested here, viz, the licensing of consquentialist justifications for accomplices but not the principals they aid? Consider in this regard the British duress cases. In Lynch116 the defendant drove the IRA gunmen to where they could kill a British policeman, and they did so. Lynch was not an IRA member or sympathizer; had he not driven the gunman they would have shot his family to death. The British courts had held that such threats could not justify actual killings; had Lynch himself shot the policeman, he would have had no defense. But merely aiding the killing (by driving) was eligible to be justified by the threats of greater harm if he did not drive the IRA to its target. The lesser wrongness of aiding a killing was outside the categorical prohibition of killing and so could be consequentially justified. My own sense is that this is correct. But why this should be right requires investigation. Contrast the actual Tally case with this variation of it: there is no warning telegram to be countermanded but Tally helps the Skeltons in a different way. He throws extra ammunition bags on their horses in case they have a prolonged shoot-out with Ross. As it turns out they didn t need the extra ammo, but they could have. Notice that this hypothetical Judge Tally is not a double preventionist. He helps to kill Ross but not by removing a defense Ross had against an impending death. Still, the hypothetical Tally is still incrementally less blameworthy than the Skeltons. It remains true that they killed Ross and Tally only helped them. 116 Director of Public Prosecutions for Northern Ireland v. Lynch, [1975] A.C. 653. 45 Notice that this is also true of Lynch. Lynch also did not remove a defense the British policeman had against an impending death (unless distance is a defense, but then anything is). Lynch simply made it easier for the IRA to kill the policeman by supplying transportation. Yet Lynch too does less wrong than those who kill, in his case allowing him the benefit of consequentialist justification for his action. My conclusion is that the double prevention/cause distinction is morally idle here. Some other factor is making the difference between direct causers of death and those who aid them. Those who aid have a lesser responsibility, irrespective of whether their aid takes the form of a double prevention or not. My long-held hypothesis is that the factor that is doing the work here is what lawyers for centuries have called intervening causation.117 The IRA and the Skeltons in the cases just discussed were free, informed, voluntary actors whose actions both directly caused death and intervened between Tally s and Lynch s efforts in that regard. As the lawyers say, such intervening causes break the causal chains that seem to exist between Lynch s and Tally s efforts, and the deaths. In law, neither Lynch nor Tally proximately caused those deaths. I have elsewhere attempted to restate the law s intervening cause doctrine in detail118 and will not do so here. Hart and Honore devoted a long and masterful book to the topic,119 showing how the legal notion of intervening causation has deep roots in common, every day usage of causal idioms outside of the law. Sufficient for our purposes is to note that in law and everyday thought, when one party sets a chain of events in motion that increase the probability of some harm, but another party intervenes by intentionally causing the harm the first party was trying to cause or risking, we conclude that only the later party caused the harm. The first party may have enabled the second to kill, may have aided the killing, may have made the killing easier or even possible at all but only the later party caused the death, only the later party killed. If this ordinary and legal notion of causation tracks the causal relation in reality, then we have what we need to make sense of the moral difference between accomplices and perpetrators: only the latter cause death, whereas the former enable it. We then have a metaphysical 117 118 So argued in Moore, Torture and the Balance of Evils, supra Moore, The Metaphysics of Causal Intervention, California Law Review, Vol. 88 (2000), pp. 827-877, at pp. 831-852. 119 H.L.A. Hart and Tony Honore, Causation in the Law (Oxford: Oxford University Press, 2d. edit. 1985). 46 distinction with which to ground the perceived moral difference. I long was hopeful that one could make out such a metaphysical distinction. But after a diligent exploration of the possibilities here things like a libertarian metaphysics that makes persons free choices literal uncaused causers and thus fresh causal starts my conclusion is that both law and ordinary thought are wrong here.120 In any acceptably literal sense of the words, there was some kind of causal relationship between Tally s act and Ross death, as well as between Lunch s act and the death of the British policeman. So there is no crisp metaphysical distinction here, between causers and aiders. There is, however, a vague by degree distinction here that may be serviceable. This is suggested by our willingness to say that Lynch and Tally causally contributed to the deaths, were a causal factor in the deaths, or even were a cause of the deaths; while being at the same time uncomfortable with saying that Lynch and Tally caused the deaths, or that they were the cause of the respective deaths, or that they killed those people. What I think these linguistic intuitions are capturing is the ultimately de minimus size of Tally s and Lynch s causal contributions. Compared to the causal contributions of the Skeltons or the IRA, Lynch and Tally were small potatoes. The suggestion is that what the law and ordinary thought have sought to capture with their bivalent notion of intervening causation, is in fact a rule of thumb about the degree of causal contribution: when there is a much larger cause of the harm that succeeds in time the accused s act, then that relegates the accused s act to a sufficiently small causal role that we think of it as not a cause at all. It is of course really a cause, but size matters here, so that moral blameworthiness is diminished as the degree of causal contribution is diminished. Such a suggestion has to answer two rather big questions. The first is a metaphysical question: are causal relations scalar, that is, a matter of degree? Not everything is. Being a little bit pregnant is usually thought to be only a joke; not being much of a father is not to be taken biologically, for in biology one either is or isn t the father. And causation could be like 120 Moore, Metaphysics of Causal Intervention, supra 47 these relations, not like being middle-aged. Secondly, there is a moral question: does moral responsibility diminish as the degree of causal contribution diminishes? I have argued elsewhere that causation is a scaler relation,121 so let me here assume that that leg of the trousers is nicely on and move to the other leg. Does it matter morally to be a small cause as opposed to a big cause of some harm? Is it less wrong to be a small contributor to a death rather than a big one, holding all else equal? There are two kinds of test cases. First, consider the at-a-time cases, where the accused s causal contribution occurs at the same time as other factors that also contribute to a single, indivisible harm. Suppose first the causal contribution of each of the factors is equal. Joint fire, joint noise, joint flood, joint pollution cases, for example.122 Is there more moral blameworthiness if you are one of three producers of some polluting harm, as opposed to one of a hundred? Does it matter to your conclusion here whether: (a) each polluter s contribution was necessary for the harm to occur, so that all together were only jointly sufficient; or (b) each polluter s contribution was sufficient by itself, so that all together were only jointly necessary; or (most commonly) (c) no polluter s contribution was either necessary or sufficient, since there was more pollution than was needed to produce the harm?123 My sense of these equal contribution cases is that it matters when you increase the numbers (and thus decrease the size of each contribution). Consider an example of Alvin Goldman s concerning the voter s paradox.124 Why is it rational to vote, when the chance of your single vote making any difference in an election is virtually nil? Goldman s answer: since causal relations can exist even when the cause is not a necessary condition for the outcome (surely right), each voter has some causal responsibility for the outcome of the election. Yet aren t we reluctant to think that this solves the voter s paradox, even if it is correct in its causal metaphysics? Elections have too many voters equally contributing to their outcome to think that Moore, Causation and Responsibility, Social Policy and Philosophy, Vol. 16 (1999), pp. 1-51, at pp. 13-17, reprinted in E.F. Paul et al., eds., Responsibility (Cambridge: Cambridge University Press, 1999). 122 For real world examples, see id. pp. 9-13. 123 What I call garden variety concurrent causes, overdetermination concurrent cause, and mixed concurrent cause cases, respectively. See id. 124 Alvin Goldman, Why Citizens Should Vote: A Causal Responsibility Approach, Social Policy and Philosophy, Vol. 16 (1999), pp. 201- . 121 48 my one vote contributes very much. I thus don t deserve much praise or blame for the election s outcome, given my de minimus (even if not zero) causal contribution to that outcome. Now vary the assumption of equal contribution. Suppose one cause is very much bigger than another in jointly producing an indivisible harm. E.g., Agatha Christie s Murder on the Orient Express, where the drugged victim dies of loss of blood from thirteen stab wounds, each of different intensity and each done by a different defendant. Did the slight stabbers do less wrong than the big stabbers? One can agree with the California Supreme Court when it got this kind of a case that drop by drop the life-blood flowed out of the victim from both wounds, and thus that each wounder is a causer of death125 and yet think that when there is more blood lost there is larger causal contribution, and when there is a larger causal contribution there is greater wrongdoing. American tort law apportions liability for multiply caused harms in proportion to degrees of causal contribution,126 and to the extent that tort law is based on moral fault that is some evidence that degrees of fault follow degrees of causation. In theory Anglo-American criminal law goes the other way here, holding each wounder fully liable for murder without regard to the degrees of causal contribution. Yet criminal law s official theory here is no more to be credited than it is in postulating equal blameworthiness between accomplices and perpetrators. My suspicion is that the law-in-action (as opposed to the law-in-books ) is quite different: more of a cause, more to blame. If causation itself matters to overall responsibility a big assumption for the Kantians amongst us, I know127 and if causation does come in degrees, then why wouldn t more of it be worse, less of it, better? Since causing more harm is worse, why isn t more causing of the same harm also worse? People v. Lewis, 124 Cal. 551, 57 P. 470 (1899). American Law Institute, Restatement (Second) of Torts 433A. 127 I argue that there is such a thing as result luck so that to succeed in causing what we try to cause or risk causing makes us more blameworthy. See Moore, The Independent Moral Significance of Wrongdoing, Journal of Contemporary Legal Issues, Vol. 5 (1994), pp. 1-45, reprinted as chapter 5 of Placing Blame, supra 126 125 49 If we turn to the more directly relevant over-time cases, we only need one assumption to apply here the results of what we did in the at-a-time cases. That assumption is that spatiotemporal distance is a pretty good proxy for degree of causal contribution. The closer some event c is to some event e where c causes e, the more causal contribution c makes to e. The assumption is supported by the thought that any event e is the product of numerous causes, and that each of them in turn are the products of numerous causes, and so on. Causation is then pictured as an inverted cone, not as a simple chain. The further up the cone from e is some c, the less causal contribution it makes to e (because it is joined by so many other causes). Thus later is usually greater, when it comes to degrees of causation.128 So I assume that accomplices like Lynch or Tally are small causes compared to perpetrators like the IRA or the Skeltons, and it is just this that explains diminishment of wrongdoing by the former to the point where consequentialist justifications are available. It bears stressing that this vague-by-degree difference cannot be the crisp distinction the legal doctrines of intervening causation and accomplice liability pretend it to be. Nor are those doctrines unerring in their capture of major versus minor causal contributions. Consider in this regard another of the British duress cases, Abbott.129 Abbott claimed the same justification for what he did in aiding a killing as had Lynch, namely, Abbott s family was threatened with death unless he participated. But what Abbott did was more: he held the victim while she was stabbed to death by the perpetrator (who had to take several stabs at it because he kept hitting bone with his saber). One might well think (as did the court), that the technical classification of Abbott as an accomplice is not what should govern the availability of duress as a defense, but rather, how much he helped. Abbott was too big a cause of the victim s death to avail himself of consequentialist justifications, no matter that he wasn t the actual killer but only the accomplice. Consider likewise the current practice of American intelligence agencies to do extraordinary renditions when it comes to the torture of terrorist suspects possessing lifesaving information. Extraordinary rendition is a euphenism for transporting suspected terrorists to countries such as Egypt who are less restrictive in their interrogative techniques than are we. In 1998, for example, the CIA flew five members of the Al Qaeda from Tirania, 128 129 See Moore, Causation and Responsibility, supra Abbott v. The Queen, [1976-3] All Eng. L. Rep. 140. 50 Albania, to Cairo for interrogation.130 Assuming arguendo that good consequences could be obtained by such measures, are they eligible for consequentialist justification because the CIA only enables torture, it doesn t torture? From what has been said, that depends on the degree of causal involvement of the CIA in torture. If mere transportation is provided, it looks like Lynch; if in room suggestions are made during torture (as Sister Dianna Ortiz has said was true of the CIA involvement in her torture in Guatemala), then it looks more like Abbott. In principle, however, such extraordinary renditions could be eligible for consequentialist justifications (again, assuming there are such). I know that this conclusion troubles many people. Partly that is due to their reflecting on cases where no good consequentialist justification is in sight. And of course, if no justification is around, you are not free to give would-be killers ammunition, countermand warnings to their victims, drive the killers to their victim, or fly the victims to their killers. You are seriously blamable for such enablings of evil. The only point is that you are less blameable than those who do evil, enough so that you can justify helping someone else do an act that you couldn t justify yourself doing. Think of the Greek islanders in John Fowles novel, The Magus, who had to reveal the location of three Greek partisans in order to save their fifty children from being killed by the Nazis. I think that they would not be justified in shooting the three partisans themselves, but they are justified in telling the Nazis where they are hiding. E. Non-Omissive Allowings 4: Being a Minor Cause of Harm vis- -vis Nature If one takes the line I do about accomplices that although they do cause the harm they enable they do not do so strongly enough to rule out consequentialist justifications then in general any minor causal contribution should also be eligible for consequentialist justification. In the immediate prior subsection I considered other persons as the big cause in comparison to which your contribution is minor. Now consider cases in which nature plays the role of big cause. I shall consider two possibilities. The first is where the big natural cause intervenes between the accused s act and the harm; the second is where the big natural cause precedes the act of the accused in time, and thus is already a force in motion which he redirects. 130 See n. 10 supra. 51 1. Extraordinary Natural Events as Big Causes Again the legal doctrine of intervening causation is instructive. In law, a freakish hailstorm, an unprecedented wind, a bolt out of the blue, etc., will break causal chains almost as surely as will the intervening free, informed, voluntary act of another wrongdoer.131 There is even less metaphysical grounding for this bit of legal doctrine than there was for the intervening actor doctrine nobody thinks storms have free will, for example.132 So here it is even clearer we are simply making big cause/little cause comparisons when we say that some actor did not cause some harm because an intervening act of nature did. The law doesn t quite have the courage of its convictions here. If the original actor intentionally uses the intervening storm as the means to a harm he wants to bring about, then the storm doesn t break the causal chain so the actor is the cause of the harm after all. As I have argued elsewhere,133 this discrimination cannot be a causal one the storm either is big enough to be an intervening cause or it isn t, irrespective of the accused s intentions. Those intentions are not causal aphrodisiacs: they cannot make the accused s causal contribution bigger, nor the storm s smaller, all by themselves. So the reality is that if the accused s causal contribution (vis -vis some intervening storm) was small when the harm was not intended, that contribution is small when the harm is intended. If there is to be liability, it should be for aiding the storm, just as it is in the case of intervening human causers.134 If we can justify our small causal roles in the case of human intervenors, presumably we can do the same here. If I can justify telling the Nazis the location of the three partisans to save the fifty children, presumably I can order the three partisans to set out from the island in a storm that will surely kill them, if that is the only way to spare the fifty children from Nazi execution on the island. 2. 131 132 Redirecting Presently Existing Threats Moore, The Metaphysics of Causal Intervention, supra at pp. 844-846. Id. at pp. 852-877. 133 Moore, Causation and Responsibility, supra at pp. 18-19, 41-43. 134 Id. 52 Now reverse the temporal order of the examples: the big natural cause is already in play, and all you do is make a minor course correction for that natural cause. Philippa Foot s wellworn example was that of a runaway trolly:135 it is headed towards five workman who cannot get off the track. You cannot stop it but you can redirect to another track where there is only one workman trapped. May you turn the trolley if you are its driver? May you turn it if you are merely a bystander at the switch?136 Real world examples of such trolley switching abound. You are a pilot in the Air National Guard. Your jet fighter has just had a flame out and is going down, headed right at a crowded school. May you redirect it onto an occupied single family home instead? Or, you sit beside a dike with dynamite. A flash flood is approaching in the river that the dike holds, and the flood will shortly inundate an entire town, killing all of its residents. May you blow the dike, sending the flood waters onto a single family farm, killing all of the farm s inhabitants but saving those of the village?137 Or you and others are back in the woods with that charging grizzly. This time I don t duck by removing myself as your defense against the bear. Rather, I do a little bit more: I throw fresh meat in your direction so that the bear that was headed towards five others is now headed towards you. Like many people, I think the answer to these questions to be a pretty clear yes. You may (and I would add, must) redirect the trolley, plane, flood, and bear from where it will kill many to where it will kill few. My favorite example of this kind is from World War II. The V-1, V-2 rockets were targeted by burn time; this meant that the shots had to be called in by German spies in England so that the burn times could be adjusted. British counterintelligence had penetrated the German spy network calling in the targeting. The choice presented to Churchill: should British counterintelligence feed misinformation to the Germans, which would move the mean point of impact two miles per week to the east of central London? This killed a few people in Foot, Abortion, supra Thomson s variation. Judith Jarvis Thomson, The Trolley Problem, Yale Law Journal, Vol. 94 (1985), pp. 1395-1415. 137 This is another example used by the drafter of the Model Penal Code to illustrate when killing could be justified by its good consequences. See Conventions, supra 136 135 53 Kent and Essex rather than a lot of people in central London. According to William Stevenson s recollections: War is an evil thing, Churchill said in a Secret War Cabinet session. Do you wish us to surrender, Mr. Morrison? Morrison shook his head angrily. Then I greatly fear, sir, that in order to live, replied Churchill, lowering his head, we must play God. 138 I think Churchill got it right, here as at Coventry. Many philosophers who have examined these redirection cases find the conclusions compelling but the rationale puzzling. Some wish to fictionalize the threats, pretending that all are threatened,139 when in plain truth only some are (this is to mistakenly liken redirection cases to the almost dead cases). Others wish to rely on the DDE, saying that I only foresee the death of the few even though I intend to redirect deadly things into them.140 Still others, such as Judy Thomson, who has examined these cases more than anyone else, is just puzzled why redirection as such makes such killings justifiable by their good consequences.141 To me, redirecting cases seem plainly causal in their rationale. It was only a rule of thumb that later is usually greater when it comes to degrees of causal contribution. But rules of thumb are only rules of thumb because sometimes they prove to be bad heuristics. As here. Even though the trolley, plane, flood, bear, and rockets are all in motion when you act, your later act is a small contribution compared to theirs. You have limited decisional space, and you don t do much. You make, as I said, only minor course corrections. Make your causal contributions bigger by starting any of these things yourself and you may not justify your actions. So if a terrorist will kill many hostages unless you: start up a trolley to run down a few; steal a plane to William Stevenson, A Man Called Intrepid, supra, at p. 414. James Montmarquet, Doing Good: The Right Way and the Wrong Way, Journal of Philosophy, Vol. (1982), pp. 439-455. 140 Michael Costa, The Trolley Problem Revisited, and Another Trip on the Trolley, in John Martin Fischer and Mark Ravizza, eds., Ethics: Problems and Principles (New York: Harcourt, Brace, 1992). 141 Thus Thomson confesses that she does not find it clear why there should be an exemption for, and only for, making a burden which is descending onto five descend, instead, onto one .on the other hand, the exemption seems to allow those acts which intuition tells us are clearly permissible Trolley Problem, supra at p. ___. 139 138 54 crash it into a house; start a flood to flood a farm; attract a bear so that it kills someone; etc. you may not do these things even to produce good consequences, because such acts make a larger causal contribution to the deaths they cause. But keep the degree of causal involvement small, as in the redirection cases, then responsibility is lessened to the point that consequentialist justifications become available. VI. Conclusion It is time we again step back from the trees and look at the whole forest. From such a vantage point we can see four variables that have made a difference to the availability of consequentialist justifications: (1) acting, (2) intending, (3) causing, and (4) counterfactual dependence. When we act (rather than omit) because we intend to achieve something bad (rather foresee it or risk it), and when our act strongly causes that bad state of affairs to come about (rather than not cause it at all, or only weakly cause it), and when that bad state of affairs strongly depended for its existence upon the act done (rather than weakly so depended, either because that state of affairs would have occurred slightly earlier or later anyway in possible worlds very close to the actual world); then we may not justify what we do by its good consequences. Whereas when we only fail to prevent some bad state of affairs, or only foresee or risk that our acts will produce such a state of affairs, or our act only weakly causes (or prevents a preventer from preventing) that state of affairs, or that state of affairs could have rightly been brought about by us earlier, or it would have come about slightly later anyway, then we may justify our choices and behavior by their good consequences. If we are strongly permitted to justify these choices and behaviors, then, while we are outside of our deontological obligations, we are so by virtue of an exception to these norms of obligation. But if we are only weakly permitted to justify these choices and behaviors, that shows something about the kinds of things that form the content of these obligations, namely, that they are not directed toward our weakly permitted choices and behaviors. I have tried to limit the examples to those illustrating only weak permissions. If I have succeeded, then the shape of our deontological obligations is as I just described: we are obligated not to act so as to strongly cause something that was not about to happen later anyway, if such act is motivated by our intention to cause that thing. Even when we plug in all the somethings like deaths, 55 disfigurements, etc. in this formula, we will not have exhausted our obligations. But we will have exhausted our deontological obligations, leaving us free to navigate by our consequentialist lights alone. I said in the beginning that one of the reasons for being interested in the distinctions we have examined is the promise they hold out to reduce or eliminate conflict between our obligations. The system just outlined does in fact do a pretty good job of this. When one of our options is an intended, strong-cause killing (that is not a mere delay or acceleration of death), and the thing we must tolerate if we don t kill is a mere omission, or an allowing, a foreseeing, etc., then our path is clear: don t kill even though the number killed is less than the number not saved, etc. There is no conflict in our deontological obligations. It is interesting whether these distinctions completely eliminate potential conflict in our deontological obligations. Consider in closing two kinds of cases. In the at-a-time version, we are back on John Fowles Greek Island during the German occupation in World War II. The Nazi commandant has told you that unless you find and shoot the three partisans to death, or you shoot and kill all the children on the island, he will blow up Athens and all of its inhabitants. (He cannot himself cannot kill anyone on the island, but he can destroy Athens.) Suppose with me the destruction of Athens and the death of all of its citizens is a moral horror that is beyond the threshold of your threshold deontology, so the one thing you cannot do is nothing. You must act, and your only choice is to kill three partisans or to kill fifty children. In such cases we have a seeming conflict in our obligations. Now consider an over time version. I have begun a boulder rolling down a hill to kill five old enemies of mine, but as before I repent. Now, however, the only way to stop the boulder is with the corpulent body of one bystander: may I throw him before the boulder, killing him but preventing the deaths of the five? Doing so makes me a killer, but it prevents me from being a killer five times over. This, too, may seem to present us with a conflict in our stringent obligations not to kill. Kant would have an easy time with these examples. He would reject threshold deontology and thus refuse the threshold s deontologist s ruling out the option of inaction in the 56 first case. And he would regard the five killings as sunk costs in the second (because on Kant s view your wrong vis- -vis the five is already complete when you executed your intention to kill them; whether their deaths follow or not does not eliminate those five wrongs); so that you have only one obligation now, which is not to kill the fat bystander. For those deontologists who would reject both of these Kantian moves, as would I, things are not so simple. Still, I leave their resolution to another occasion.142 For those who cannot stand the suspense: it seems clear to me that one must abide by the more stringent obligation in the first case, which is not to kill the five, so that you do kill the one; and that one must not kill the one in the second case, on pain of collapsing deontology into a kind of duties-kept consequentialism. 142 57

Find millions of documents here - Study Guides, Homework Solutions, Papers, Exam Answer Keys and more. Course Hero has millions of course related materials that will enable you to learn better, faster and get an A in all your courses.
Below is a small sample set of documents: