In recent months, since September 11th 2001, David Fisher and Brian Wicker have been exchanging correspondence on the question whether it is ever morally licit intentionally to kill an innocent person, whether in war or in any other context. This correspondence began with a letter from Brian Wicker to David Fisher following the bombing of the World Trade Towers and the start of the subsequent war in Afghanistan. It was further prompted by the talk given by Nicholas Denyer to the CCADD lunchtime meeting on December 5th 2001. The authors both think that this debate may be of interest to other CCADD members, and have therefore agreed to put it on the CCADD website. The letters are in chronological order in documents numbered 1 to 14. Others are invited to join in. Contributions should be sent by email attachment to Liza Hamilton, who will then put them on the website. A copy should also be sent to Sir Hugh Beach, as our ‘webmaster’. Of course, copies would be welcome to each of us as well.

David Fisher

Brian Wicker

March 2002





Dear David,


Reflection on the current military campaign in Afghanistan, and the deaths of innocents, has led me to go back to some arguments put forward by yourself concerning the principle of double effect. I myself think that despite those arguments, Elizabeth Anscombe was right (in War and Murder) about double effect and the morally crucial distinction between intention and mere foresight of consequences. This is partly because I share her belief in a few moral absolutes, of which the immunity of the innocent from being killed is one. In my opinion there are no circumstances in which it would be licit for us (whoever we are) intentionally to kill someone who, in the given circumstances, had done us no harm. Consequently, like her, I want to maintain the validity of double effect (when properly stated, and not twisted by spurious concepts of ‘intention’) even against the sorts of argument given by Paskins and Dockrill in The Ethics of War and by yourself in Morality and the Bomb. So I wonder if you would like to respond to the following considerations?


  1. In order to reject double effect, both books quote the silly example from Bernard Williams about Pedro, the Latin American thug and the unfortunate Jim who is told to shoot one of Pedro’s hostages. Morality and the Bomb, pages 39,40. Now Paskins and Dockrill (but not you, though it is clearly presupposed in that case too) make it a specific axiom of the theorem that, despite his obvious wickedness, Pedro is a man of his word. When he promises to release the other hostages once Jim has killed the first of them, he is to be trusted to keep his promise. This axiom is necessary in order to put Jim in the dilemma which (so the argument goes) leads to the rightful killing one of the hostages. But I maintain that the whole scenario is self-contradictory. The idea that Pedro can be trusted to keep his part of the bargain is absurd. A man as wicked as Pedro simply cannot be trusted, and Jim ought not to trust him. Whatever Jim does, he is morally obliged not to take Pedro at his word. In other words, Jim is a fool in accepting the bargain, and his folly is a moral mistake. Hence the alleged conclusion, that here is a case of rightful killing of the innocent, collapses, and the example has no detrimental effect upon the double effect principle.


  1. In Morality and the Bomb, p. 37 you discuss another example, drawn from Frank Richards’s experience in World War I of throwing grenades into enemy cellars where civilians may be sheltering. You claim that when ‘Richards has good reason to believe that civilians are sheltering in the cellars (and) he nonetheless lobs in the bomb without warning (and) the young lady and her family (who are hiding there) are killed’ the deaths of the lady and her family are still unintended (just as they are when he shouts to warn anyone in the cellars to come out, before he throws the bomb, but they do not respond in time). I do not accept this argument. It seems to me clear that in this case the deaths are intended, and that this distinguishes it from the others. That Richards does not want the young lady and her family to be killed, and would prefer them not to be, is beside the point. He still intends to kill whoever is in the cellar. So once again, double effect reasoning is unaffected.

  3. You discuss Strawson’s examples of somebody crushing my hand (Morality and the Bomb pp. 33-36). Three cases are mentioned: (i) the agent intended to crush my hand; (ii) the agent foresaw that my hand would inevitably be crushed as the result of what he was doing and was, nonetheless, willing that it should be crushed; (iii) the agent crushed my hand accidentally (he was, for example, pushed or slipped). You claim that there is little moral difference between (i) and (ii) because ‘in either case I should feel a not dissimilar degree of resentment’. But the degree of resentment is not the point. Suppose I am trapped in the ruins of the World Trade Towers, with my hand sticking out of the rubble. The difference between (i) and (ii) might then be between (say) a) a fireman with a grudge who wanted to hurt me and had an opportunity to do so as he was clambering round the ruins of the World Trade Towers, and b) a fireman who was trying to rescue someone else from the World Trade Towers and could not avoid crushing my hand in the process of doing so. In such a case, even if my resentment were equal in both cases, the moral difference is quite clear. Double effect remains valid.

  5. One of the great advantages of double effect reasoning is that it can have the effect of compelling people to try to find ways round an otherwise insoluble problem. You argue that in the numerous abortion cases, and given existing limits on surgical techniques, there is little difference between a) the surgeon who, in order to save the mother, removes her cancerous womb, and thereby unintentionally but inevitably causes the death of the baby inside it, and b) the surgeon who deliberately kills the baby by crushing its skull in order to save the life of the mother. Now these are treated by yourself as parallel by making the supposition that ‘if the operation is carried out the mother will live but the foetus will die; if the operation is not carried out, both mother and foetus will die’. This supposition is made ‘to avoid difficult questions about the relative value of the lives of mother and foetus’. But these questions cannot be avoided. The truth is that both lives are equally valuable: if they were not, then the whole problem which these examples raise would vanish. Granting that both lives are equally valuable, the only way of avoiding a negative judgement about (i) is to raise certain questions. For in the absence of answers to these, it would simply be the case that killing the baby becomes part and parcel of what the surgeon intends to do, and thus the baby’s death would itself be intended. To avoid this conclusion, questions have to be raised such as: could the mother survive long enough for the baby to be born before the cancer finally kills her (if it does)? How likely is it that she will die of the cancer before the baby can be born? If the cancer is allowed to grow until the baby can be born, how much life is the mother likely to lose as a consequence of postponing the hysterectomy? How likely is it that the cancer will kill the baby in the womb anyway before it comes to term? If, as is almost certain to be the case, there is no absolute clarity about the answers to these questions, I think a good surgeon and a good mother would at least consider postponing the surgery in the hope of avoiding killing the baby.

    In short, examples of this kind can only be forced into producing clear-cut dilemmas by being shorn of many of their irreducible uncertainties. Taking these into account commonly makes apparently stark choices less stark than they appear, and thus the apparent implausibility of double effect reasoning disappears. ‘Real’ situations are seldom as clear-cut as those who advocate ‘situation ethics’ think they are. What is more, the advantage of maintaining that b) is absolutely illicit is that this offers a huge incentive to researchers to find ways of saving the mother without deliberately killing the baby. (Of course, it is all to the good if researchers find ways of avoiding a) as well: but the absolute illicitness of b) makes the matter far more urgent. Without such absolute prohibitions I suspect that some medical research would be judged too expensive or too ‘unproductive’ to be worth pursuing).


  7. However, while I maintain that double effect arguments remain valid, there are difficulties. These come out in the case of military activity. Distinguishing between intended and merely foreseen consequences is often difficult (as Anscombe acknowledges). And it would seem that the more likely the foreseen consequences are, the harder it is to say that they are unintended, as I have argued in the abortion case.


While the key moral gap between merely foreseeing a certain consequence and intending it remains, the more probable that consequence is, the harder it is to say that it is not intended. This is why it is tempting to make no moral distinction between the abortion cases (i) and (ii), even though the temptation ought to be resisted. And this point clearly applies in the case of bombing in wartime. To be able truthfully to say that civilian casualties were not intended (i.e. were truly side effects, or ‘collateral’) their deaths must have been less than inevitable. For if they were certain to occur, then they would become part and parcel of what we intended to happen, and we would be guilty of murder. This is a key point at which I think we have to dispute the thesis put forward by the founder of the humanitarian laws of war, Francisco Vitoria, in the sixteenth century, that it may be licit to deliberately attack a city which we know contains innocent civilians as well as soldiers, in order that the justice of our war be not balked. This thesis amounts to saying that in such a case it is all right deliberately to kill the civilians, together with the soldiers, since they can all be lumped together as part and parcel of what we must, in all justice, do. I do not see how to reconcile this thesis with Vitoria’s much stronger judgement that intentionally killing the innocent is always illicit. At this point I think it has to be admitted that Vitoria is not wholly consistent in his argument.


But how uncertain do the consequences have to be before we can be let off the hook? I think there has to be something the victims can do to avoid being killed (and if necessary we must make sure that this is so) if we are not to be guilty of intentionally killing the innocent. This is why it makes all the difference, in the Richards case discussed above, whether or not the family is given the chance to announce their presence in the dugout. Without this, it seems to me, we cannot avoid the conclusion that what we did, i.e. kill the innocents, was intended.


Maintaining the validity of double effect reasoning is a necessary condition for holding a strong doctrine on the just war criterion of discrimination. It is no accident that your abandonment of double effect arguments leads you to drastic revision of discrimination itself. In your hands it becomes no more than the demand to minimise non-combatant casualties (Morality and the Bomb p. 50). And even then, the key demand is diluted: ‘exceptionally the consequences of not breaching the rule may be so bad as to override the presumption and that in such cases it would be wrong to focus attention on the internal quality of the act (e.g. intentionally killing the innocent) to the neglect of the consequences (e.g. saving the lives of many more innocents)’. The dilution evident here comes with a dubious assumption that the intention involved is an ‘internal quality of the act’. I would dispute this: I don’t think this is the right way to think about intentions. An intentional killing is a different act from an unintentional one. As Anscombe says, in Intention #19, ‘we do not mention any extra feature attaching to an action at the time it is done by calling it intentional’.


Furthermore, given the enormity of what is supposedly at stake, as in the debate on nuclear deterrence of Soviet aggression, this permitted minimum of civilian killing may be very large, as Michael Quinlan has admitted (in Thinking About Nuclear Weapons p. 84). A further watering-down of the criterion sometimes comes with the thesis that the discrimination criterion must be taken along with the criterion of proportionality. There is a good and a bad way of doing this. The bad way is to say simply that the amount of killing of the innocent must be no more than is proportionate to the good proposed, without reference to the question of intentions. Worst of all, W.V. O’Brien tries to balance non-combatant immunity against considerations of military effectiveness, with the result that discrimination becomes a merely relative principle which simply enjoins ‘the maximisation of non-combatant protection’. (On O’Brien see Richard B.Miller, Interpretation of Conflict, p. 215)


For Anscombe, on the other hand, discrimination forbids the intentional killing of the innocent as murder. The American Catholic bishops say the same thing in their letter on nuclear deterrence: ‘the lives of innocent people must never be taken directly, regardless of the purpose alleged for doing so’. (#104) Paul Ramsey agreed. Miller summarises his thesis thus: directly to attack innocent people, regardless of one’s cause, is to violate both the spirit and the letter of neighbour-love. (Interpretations of Conflict, p. 151. Cf. Ramsey, The Just War Chap 6). This is the first thing that has to be got clear. Only once it has become purely a question of the truly unintentional killing of the innocent, does the question of proportionality arise. That is to say, even though some unintended or genuinely collateral killing of the innocent is permissible, if it goes beyond a certain point this alone may vitiate the justice of the war. (It is a good question to ask, in the current Afghan war, how many innocent Afghans may we unintentionally kill before the whole conflict becomes illicit? Would killing as many as died in the World Trade Towers be proportionate or disproportionate?) But in any case, the intentional killing of the innocent is absolutely forbidden, come what may. If this means that we have to suffer defeat, so be it. As Anscombe says, this ‘is not a vague faith in the triumph of the spirit over force (there is little enough warrant for that) but a definite faith in the divine promises’.


Brian Wicker





Dear Brian




1.You criticise the use of examples with some simplifying conditions to help tease out ethical issues. If your criticism were valid it would call into question a great deal of philosophy from Plato onwards. It might also have ramifications elsewhere. Are the principles of geometry invalid because the contours of the physical world do not correspond to Euclid’s postulates? Let us consider some of the examples you discuss.


2. The simplifications made in the case of Jim and the Indians seem to me an entirely legitimate manoeuvre to pose the dilemma at to what you should do if faced with a choice between killing one person and saving nineteen. Your criticism would only have force if our moral life were so constructed that such appalling moral dilemmas never arose. But unfortunately they do: from Agamemnon at Aulis to the very close to real life abortion examples you quote later in your paper. Nor is the simplifying condition itself totally implausible. No human is wicked in every respect. The Kray twins loved their old Mum and did a lot of good work for local charities. But they were also murderers. Equally I find it not implausible that a killer like Pedro could be punctilious about keeping his word. It thus seems to me a cop-out to use the fact that we typically operate in conditions of some uncertainty to avoid answering the moral dilemma this example poses.


3. That the dilemma is real enough is, as I say, shown by the abortion examples. Again you try to evade responding to the parallel ethical dilemma these raise by latching onto my remark that we can avoid difficult questions about the relative value of the two lives. This is, however, to miss the point of this simplifying assumption which is precisely to show that the dilemma can arise if we share your view that the lives are of equal value or if we hold that the mothers’ life is of higher value than a foetus. Even on your view there is a difficult choice as to what should you do if you can save one life by operating or do nothing with the result that two lives are lost? Your answer may be different from mine but you cannot evade the dilemma by pretending such choices do not arise.


4.The weakness of your general onslaught on the use of examples with simplifying conditions to bring out ethical issues is well illustrated by your own inconsistent use of this technique with the signalling example. I have no problem with your use of this example but you should do! Why are you not refusing to answer the questions you pose because in real life there would be many more uncertainties? And isn’t it implausible to suppose two trains are shunted onto single track? What I do have a problem with is your supposition that in case b) he is guilty of deliberately causing the deaths and yet might rightfully get off a murder charge. Surely if the prohibition on intentional killing is absolute he is in your book guilty of murder? I do not see why on your ethical view there is any moral difference between b) and c). On your view both are intentional killing, both are equally wrong.


5.On the next page you say: "The more certain the consequences, the less the excuse. While the key moral gap between merely foreseeing a certain consequence and intending it remains, the more probable that consequence is, the harder it is to say that it is not intended." But this surely questions the very distinction that double effect seeks to make between a consequence foreseen but wanted neither as a means nor an end and one intended. (That incidentally is why you are wrong to dismiss the Richards example. If he does not want the deaths as a means nor an end then according to double effect he does not intend them, even if he foresees the deaths as likely.)


6. In sum I do not think you can evade addressing the ethical issues posed and hence the strains that the double-effect comes under by criticising the philosophical use of examples.


David Fisher





Dear David,


I was not intending to attack the general practice of using simplified examples in philosophy, but only to clarify what they can and can’t do. It seems to me we have to distinguish at least two kinds of ‘examples’ in moral philosophy: a) those which discuss hypothetical cases, i.e. situations which allegedly could happen and b) those which discuss actual cases, i.e. actions which are presented as having actually happened. The Williams and Smart example would come into a), whereas your Richards examples come under b). The abortion examples could be either, and this is their difficulty.


  1. Pedro and Jim: Morality and the Bomb pages 39-40. Of course, it is possible even for an unscrupulous thug like Pedro to have a habit of keeping promises. Unlikely but not impossible. There’s nothing so rum as folk! - as the Kray example shows. But my point is a different one. According to the Paskins/Dockrill version of the example, Jim ‘knows’ that Pedro is in the habit of keeping his promises. But how reliable is this knowledge? Who told him? In real life, Jim would be very anxious to know this. If the information came from one of Pedro’s friends, it is likely to be just as unreliable as Pedro is unscrupulous about murdering people. And even if the information were thought by Jim to be reliable, there is nothing to stop Pedro from breaking his habit next time. I submit that it is entirely right for Jim to distrust Pedro’s word simply from the evidence of what he sees staring him in the face. It is crazy to trust the word of a manifest would-be murderer just because somebody else has told you that he has up to now had a curious hang-up about not breaking promises. All of this is hidden from view by the omniscient word ‘knows’ in the Paskins/Dockrill version. Its use is to prevent the reader from asking the elementary question anyone in the real situation would want to ask: how come that Jim ‘knows’ this? Your own version (and presumably Williams and Smart’s) on the other hand, omits to assure us (or Jim) that Pedro’s promises are trustworthy. But this omission only makes my argument stronger. If Jim has not been given any information about Pedro’s character, it is only right for Jim to distrust Pedro on the basis of what is blatantly obvious, namely that he is a murderous thug.

  3. By turning Pedro’s punctiliousness into a ‘datum’ of a moral theorem, rather than treating it as the expression of an opinion from a real person whose reliability would have to be weighed critically by Jim, the whole example ceases to be a ‘real life’ case and becomes a mere ‘theorem’. It may have its uses to illustrate a certain kind of argument (perhaps a bit like the unavoidably misleading but yet helpful two-dimensional diagrams in Stephen Hawking’s book The Universe in a Nutshell): but not as a guide to real life dilemmas.

  5. This brings us to the Richards cases, which are apparently about real actions, not hypothetical ones. (see Morality and the Bomb p. 36) I agree that in the first two cases Richards clearly does not intend to kill the civilians. But case (iii) requires further clarification. You say that ‘Richard’s sole aim remains that of clearing the cellars of enemy soldiers’. Apart from a dangerous ambiguity about the meaning of ‘aim’ (see below), this anyhow looks to me suspiciously like another ‘datum’ in a moral theorem. We need to know where this information about Richards’s ‘sole aim’ comes from. But we are not told. In cases i) and ii) this does not matter because in i) we are given the source of the ‘real life information that there are no civilians in the dug-outs (albeit it turns out to be wrong, as real life information often is) and Richards can be excused for supposing it to be true; while in ii) Richards’s action shows that he had no intention of killing the civilians, because he gave them warning. As he says, he nearly murdered them. But in iii), what Richards does is intentionally to kill whoever is in the cellar, civilians along with the soldiers. This seems to me perfectly obvious simply from the account of what he does. Indeed, as he himself says, it would have amounted to murder. Of course, he claims that the murder would have been ‘innocent’: but I think he is wrong about this. In order to get Richards off this hook, you introduce a theoretical ‘datum’ into this example: the talk about his sole ‘aim’ being to kill only soldiers. But in real life (say, in a subsequent court-martial of Richards on a charge of murdering civilians) we would want to ask whether his claim about his ‘sole aim’ was consistent with what he did. Killing enemy soldiers may have been the commander’s ‘sole aim’: but was it Richards’s? He might, after all, have done what he did in order to get

out of a hot spot as quickly as possible. If that was his aim, notwithstanding his plea of ‘innocence’, then the discrimination principle, forbidding the killing of innocents for whatever putative end, would condemn him, even if it was not applicable to the commander. What is more, in saying that somebody’s sole ‘aim’ is to kill only the soldiers, we are in danger of misusing an ambiguity in English, in order to get people off the hook. The word ‘aim’ can either mean ‘point at’ or ‘objective’. In the ‘objective’ sense, Richards is alleged, by your theoretical datum, to have had only the soldiers in mind; but what he actually did shows that he aimed (in the other sense) to kill everyone, civilians and soldiers alike (for some purpose about which we are not informed).



Thus I cannot accept your conclusion that in these cases the distinction between intended and merely foreseen consequences is drawn in the wrong place. And hence I reject the further implications you later suggest, which are designed to undermine the absolute rule of not intentionally killing the innocent. Abandoning the absolute prohibition on intentionally killing the innocent amounts to an enormous divergence from the tradition of Augustine and Aquinas; and it raises huge theological as well as ethical questions. I feel that your book (along with a lot of other modern works on the subject) does not sufficiently recognise the enormity of the departure from the just war tradition which you advocate.


  1. The abortion issue raises the most difficult examples, because it is not always clear whether we are dealing with a Williams-style abstract moral ‘theorem’ or a real case. In real life, case a) can only be assessed in the light of the sorts of questions (and answers) which in my original paper I suggested a real doctor (plus mother, and preferably father) would have to consider. Given sufficient information, it may turn out be that the problem can be got round without intentionally killing the foetus. This is why dismissing this possibility from the start, for the sake of making the theoretical argument, won’t wash; for in doing so the example is turned into another moral ‘theorem’ and this lessens its relevance to real life. For in the real world we need to ask: who says the mother will die if the foetus is not killed? Why not get a second opinion? How reliable are medical opinions of this sort anyway? Are there any other hospitals which have better records of success in trying to saving both mother and foetus? etc etc. These are just the kinds of question that responsible parents would ask in the real world. They would also have to ask themselves (and answer) the fundamental philosophical question, namely: is their unborn offspring really a member of our species or something less than this? For that is the crucial issue in evaluating its life. (Talk of killing ‘foetuses’ rather than ‘babies’ here can all too easily be used to evade answering this question; just as talk of ‘babies’ can disguise the obvious difference between a fertilised ovum and a human person. That is why the best, neutral term to use here is simply ‘member of the human species’). True, even after all this it might still turn out that a surgeon of genuine integrity decides to remove the uterus with the foetus inside it. At this point, I think, we have a genuinely tragic story: that is, a dilemma to which there is no just solution. Now there are people who argue that a tragedy in this sense cannot arise within a world that is created by God. They argue that tragedy was essentially a product of a world of conflicting gods (as with Agamemnon), and that an insoluble dilemma cannot arise in a Christian scenario (nor perhaps a Muslim or Jewish one). But I don’t think this is true, although it needs teasing out in theological terms.
  2. On the other hand, what the surgeon does in case b), like the Richards case, on the face of it is intentionally to kill an innocent person (the word baby is your term here). Of course, if he (and the parents) have earlier asked such questions as: who says that the woman will die unless I crush this baby’s head? Or that she certainly will live if I do so? Is this information reliable? Has anybody tried to avoid this dilemma? If so, will the authorities pay for me to have a go at avoiding it myself? etc etc.; and it still turns out to be impossible to avoid deliberately killing the baby, I see no way in which such an action can be justified, although it may very well be forgiven. (So too, of course, may the equally difficult decision not to kill the baby). In other words, there is no solution to this tragedy either, only forgiveness. Forgiveness is the Christian answer to tragedy.
  3. In short, I am against trying to produce ‘solutions’ to tragedies of this kind by bending the ethical rules: it is better to leave them alone and admit that sometimes there is no satisfactory answer other than forgiveness. Nevertheless double effect still operates. I mean if, at the end of all the questioning, the surgeon were still faced by the stark choice of saving one human life at the expense of another, the dilemma would be agonising. (Because it would be agonising, there is an enormous incentive to find ways of avoiding it, by better methods, more research etc. For unless we confront the agony, we are unlikely to do the necessary work or pay for it to be done). But this wouldn’t solve his (or the parents’) immediate problem. Now whether the choice ever is so absolutely stark as this brings us to the difficulty of knowing which kind of example we are dealing with: a ‘real life’ one or a mere ‘moral theorem’? The danger is of mixing up the two. My point here is that if the choice were one that occurred in real life (I don’t know if it ever does or not), then the doctor (of course together with the parents) just has to do what seems best and take the consequences, praying and hoping not to be blamed for it. Whereas it seems to me your ‘solution’ seeks to evade this tragic consequence, by either making out that one life is less valuable than the other, or that one life may be sacrificed for the other. I think these propositions are both untrue, and thus not solutions at all.


8. None of this quite touches the problem with which I began all this, namely the problem Vitoria discussed, about besieging (or indeed bombarding) a community which contains a large number of innocents along with soldiers. (Konduz

seems to be a classic case here). He argued that it was just because otherwise the justice of the besieger’s cause would be balked. It was this which took me back to your and the Paskins/Dockrill arguments. The trouble is that I can see no way of avoiding the conclusion that such action amounts to the intentional killing of the innocents, something forbidden on Vitoria’s own basic premise. This makes things very difficult for people in real wars. Vitoria himself makes no attempt to show that his two theses are compatible. The problem is like the Richards third example writ large. I should be interested to know how you would deal with it, without resorting to either of the false propositions which I detect in your treatment of the abortion case.


Yours ever,









Dear Brian,

With apologies for the delay, I offer herewith a few more thoughts on your latest comments.
It may help first to underline the points on which we agree. I entirely accept that in real life dilemmas it may be appropriate to seek answers to the kind of questions that you pose and that this may affect how we act. (But typically, of course, we act in situations of uncertainty where, however hard we press such questions, we may not get answers and yet still have to act.) I equally agree that the dilemmas may involve tragic choices, where, whatever we do, may involve harm. Regret, remorse and forgiveness may then be appropriate responses.

Where I still disagree is in your treatment of philosophical examples. For it still seems to me entirely legitimate for philosophers deliberately to simplify situations -provided they do not stray beyond the bound of possibility - in order to pose starkly ethical dilemmas with which we could be faced. That is what Williams does with Jim and the Indians and it still seems to me that he presents an entirely realistic dilemma, however fanciful you may think some of the details are of his stage-setting. The dilemma he poses is: is it ever justified to take one life if that is the only way to save nineteen? Are you suggesting that we could never face such a dilemma? But if we could, we surely need to consider what our response might be. In my view taking a life could be justified in such circumstances. In yours I assume it could not.

That the dilemma posed, however fanciful the stage setting, is a real one is surely demonstrated by the events of September 11th. Would a USAF pilot be justified in shooting down a hijacked plane heading for the Twin Towers, killing thirty innocent passengers in order to save many thousands of innocents? In my view he would and the President of the USA would be justified in ordering him to do so. You would no doubt want to ask all sorts of questions. But the dilemma would still remain and it may be necessary to decide how to act in minutes or less. What would you then do? (Nor in my view can double effect help for the reasons argued in my book and particularly not in case like this where the deaths are a certain consequence of the pilot's action.)

I still, therefore, think that you are you are using the device of asking for more information as a way of trying to avoid answering the dilemmas posed by the examples you cite. But if the dilemmas are realistic - as I believe they are - an answer is still required as to how we should act. I am left very unclear what you think should be done e.g. in the craniotomy case. We can agree that it is a tragedy, to which the Christian response is forgiveness. We can agree that, whatever is done, the consequences are tragic. But the dilemma is a real one and the Jesuitical wriggling will have to stop at some point. What then would you do? My view is that in such a case it would be justified taking one life (the baby's) in order to avoid two deaths (mother and baby). What is your view?

I develop some of these themes further in the attached piece on euthanasia where again I believe the dilemmas posed are real and where I assume we would also reach different conclusions?










Dear Brian




We will all die. For the fortunate death may come painlessly. For others it may be preceded by a prolonged period of excruciating and degrading pain. If such a fate were to befall us, the rational choice would surely be to choose when to die in order to terminate the agony, assisted, as necessary, by medical help.


Such was the view of Lillian Boyes as she lay dying in August 1991 in the Royal Hampshire County Hospital. For twenty years she had suffered from an extremely painful form of rheumatoid arthritis. Now at seventy she had developed ulcers on her arms and legs, fractured vertebrae, internal bleeding, deformed hands and feet, swollen joints, and gangrene from steroid treatments. Simply to be touched caused intense pain. The pain was uncontrollable. She accordingly sought the assistance of her physician, Dr Cox, to help her die in order to end the agony. Her two sons endorsed this decision. As Peter Singer relates,


Dr Cox injected two ampoules of potassium chloride – a double dose of a lethal

drug that has no curative or pain-killing properties. Mrs Boyce relaxed, and for the

first time for months her son could clasp her hand. Then she died."1


A totally justified mercy killing? Not in the eyes of the British law. Dr Cox was

charged with attempted murder. He was found guilty and sentenced to twelve months suspended prison sentence. The judge told him that he had, "betrayed his unequivocal duty as a physician." 2


In Britain voluntary active euthanasia is illegal. There have been sporadic attempts in the intervening years since 1991 to mount legal challenges against that judgement. But all have failed. Most recently, a petition by a woman, suffering from motor neurone disease, to secure the right to a medically assisted death was rejected by the Law Lords in autumn 2001. Ten years on euthanasia remains illegal in Britain and there is no immediate prospect of any relief from this harsh judgement.


This is, moreover, the legal position in virtually every country. The only notable exception is the Netherlands where, although euthanasia is illegal, a defence is recognised of "divided loyalties". It is accepted that the physician’s duty to relieve unbearable suffering can on occasion over-ride his legal duty not to kill. The Royal Dutch Medical Association has drawn up guidelines for when euthanasia is permissible:



Why has the rest of the world been so reluctant to follow the Dutch lead? There are, I think, in the main three reasons.


The first is that outside of the Netherlands there has been very been very little rational debate about euthanasia. It is not a subject debated in the British Parliament, unlike abortion which poses similar moral difficulties. This in turn reflects the post-Enlightenment aversion from any mention of death in polite society. For modern man, death is the final and perhaps only taboo subject. Death has been dethroned from the honoured position in life that it held throughout the pre-Enlightenment period. We are promised eternal life but have for the most part forgotten what that means. Instead we fondly pretend that our present lives will last forever, unheedful of the unending, repetitive banality that this would involve. When death occurs we avert our eyes, change the subject and move on. There is after all no point in being morbid and dwelling on the topic. And the consequence of this prolonged mauvaise foie, this sustained self-deception is that, unlike our forbears, we are quite unprepared for death when it creeps up on us and, in particular, unready for the possibly protracted process of dying that may painfully precede death. Nor have we properly considered whether, when and how we might wish to exercise a choice to die.


There is thus a powerful force of inertia that keeps the law as it is. But there are also more respectable reasons for this. Chief amongst these are the general moral prohibition against killing the innocent and the specific duty placed on doctors by the Hippocratic oath:


I will neither give a deadly drug to anybody if asked for it,

Nor will I make a suggestion to this effect.


If the prohibition on killing the innocent is absolute, admitting of no exception then doctors are clearly barred from administering death – inducing drugs, however great the suffering that may thereby be relieved. Indeed, if the prohibition is absolute it would presumably also forbid suicide, so the patient would be banned from taking his or her own life, even without medical assistance. Such, indeed, was the view of Aquinas who additionally held that "life is God’s gift to man, and is subject to his power … Hence whoever takes his own life sins against God."4


Indeed, suicide was, as euthanasia is now, illegal, punishable, with a certain unconscious irony, by death. That is, however, no longer the case and we already concede suicide as an exception to the prohibition on killing, at least as far as the law is concerned. There are, moreover, other exceptions now generally conceded, such as military action, even if this involves civilian deaths, provided efforts have been made to minimise civilian casualties and those that are caused are not disproportionate to the military objective aimed at. The law also allows abortion even where the foetus is well developed and in the process of birth if this is the only way to save the mother’s life. The classic case is when the head of a foetus becomes stuck during labour and cannot be dislodged. The surgeon can then perform an operation known as craniotomy which involves crushing the skull of the baby: an act of taking life deemed justified because it is the only way to save the mother’s life. Without the operation both mother and baby would die.


Perhaps it might be argued that these cases are at the borderline and the general prohibition on killing still stands. So I believe it does in the sense that a very strong moral presumption against taking life remains. The question at issue is whether that presumption can exceptionally be overridden if that is the only way to prevent very much greater harm; whether, for example, one life can be taken to save many others.


Philosophers used to manufacture examples to put this issue to the test, such as Jim and the Indians discussed by Bernard Williams.5 The unfortunate Jim, visiting a South American town, is presented by the unscrupulous bandit Pedro with the chance to save nineteen Indians if he will kill one. What should he do? Post -September 11 2001, we no longer need to manufacture examples. We have all agonised whether it would it have been right to shoot down a hijacked plane with thirty innocent hostages on board if it were about to hit the Twin Towers in New York killing many thousands? Few in America or elsewhere dispute that it would have been right so to act on September 11th, had the opportunity arisen. US Air Force pilots are now under orders to shoot down hijacked planes posing similar threats and this publicised order has occasioned little controversy.


It thus seems clear that, however strong the presumption against taking life, we are prepared to concede exceptionally that it may be overridden if much greater harm can thereby be avoided. But would this license a doctor to assist the suicide of a terminally ill patient? Not necessarily. For even if the general prohibition on killing is not absolute, the Hippocratic oath imposes special responsibilities on doctors to preserve and not take life. These responsibilities, it may be claimed, are crucial to the maintenance of the trust on which the patient-doctor relationship is based.


Perhaps so. But the doctor also has a responsibility to protect a patient from unbearable suffering. In an ideal world these duties might never conflict and the doctor could always relieve pain without risking the patient’s life. Unfortunately, in the world we inhabit it may not be possible to control the pain except by terminating the patient’s life: the duties may conflict. What then should the doctor do?


Doctors are already treading on thin ice. For they may and routinely do administer drugs in order to relieve pain, even though they know that death will be a likely consequence. The law also already concedes that a doctor may legitimately terminate life by discontinuing treatment. In Arthur Hugh Clough’s words:

Thou shalt not kill; but need’st not strive

Officiously to keep alive.6


The Hippocratic oath is thus already deemed to admit exceptions, and exceptions that are by no means unusual. An American study of deaths in a geriatric nursing home found that 81 out of 190 patients who developed fever received no treatment and that of these 48 died.7 Death through medical omission thus appears acceptable but we

still flinch from licensing death though medical commission. But can the act/omission distinction bear such moral weight? I believe not.


There are important moral distinctions between some acts and omissions even where the consequences are identical; not least since omissions may, rather more of

ten than acts, be the result of negligence and thoughtlessness, with the resulting consequences neither foreseen nor consented to: "I hadn’t realised", "I didn’t know", "I didn’t think" are typical excuses in such cases. "But," as I have argued previously,


where an act or omission has identical consequences, equally within the agent’s

control and consented to by him, the moral distinction becomes tenuous. The

hospital orderly, who has inveigled an elderly, terminally ill patient to bequeath him

all her money, hardly seems less culpable if he deliberately fails to switch her life

support machine on (an omission) than if he deliberately switches it off (an act). In

such circumstances, it seems implausible to claim that the act is – of its intrinsic

nature – somehow worse that the omission.8


Even if, however, we concede that neither the general prohibition on taking life nor the specific provisions of the Hippocratic oath can be regarded as absolute, admitting no exception, it may be argued that we should still maintain the prohibition on voluntary active euthanasia because if this were licensed as a general practice it would open the flood gates to all sorts of dubious practices. We would be launched down a slippery slope from which there would be no going back, a slippery slope that would lead to the proliferation of calculating medical killers like Doctor Shipman, preying on the elderly and infirm. It was on such grounds that the Supreme Court of Canada maintained its ban on euthanasia since, as they put it, " a relaxation of the absolute prohibition takes us down ‘the slippery slope’".9


But is the slope really as irresistibly slippery as that? The Hippocratic prohibition, as we have seen, is in reality not absolute but already admits of exceptions. Doctors perform abortions even on infants in the process of birth. They regularly bring about patients’ deaths by withholding treatment. The prohibition has thus already been breached without dire consequences.


Second, it is surely fallacious to suppose that the line can be drawn and held only in one place. If we decide, as a society, that mercy requires us to shift the line a little we can then surely choose vigorously to defend the new position. It will, of course, be necessary to define the new position carefully so that only those cases that we wish to permit are permitted and others clearly remain prohibited. The guidelines laid down by the Royal Dutch Medical Association would appear to furnish a rigorous basis for such a definition, requiring, inter alia, an explicit request from the patient; that there is no other way of relieving unbearable pain; and that the doctor in charge should seek a second opinion from an independent doctor.


Indeed, it is arguable that the robust enforcement of such carefully thought out guidelines would be much easier to defend than the present position which is precariously based on maintaining the fiction that the Hippocratic oath is absolute, even when elsewhere we have admitted exceptions. Courts would surely be more prepared to impose real penalties on those breaching such guidelines than they are in support of present arrangements. For courts are showing themselves increasingly reluctant to impose the full penalties available against euthanasia. Courts may uphold the technical illegality of euthanasia but, as in the case of Dr Cox, sentences are typically suspended and in that case the General Medical Council took no further action, deciding in its words " to temper justice with mercy."10 Vainly attempting to defend an increasingly indefensible position seems more likely to lead to a slippery slope than shifting ground to a carefully chosen and more readily defensible position, from which any breaches can be robustly countered and punished.


My conclusion is, therefore, that we should have a legally recognised right to choose, when and how we die, calling on medical assistance where required, if the alternative is a prolongation of unbearable pain. In such cases the duty not to take life is overridden by the duty to relieve suffering since the harm resulting from prolonging the suffering would be much greater than the harm caused by the loss of a life that is – ex hypothesi - in agony and already about to expire.



David Fisher

January 1st 2002







  1. Peter Singer, Rethinking Life and Death, (Oxford University Press, 1995) p.140.
  2. Quoted Singer op. cit. p.140.
  3. J.K.M.Gevers, "Legal Developments Concerning Active Euthanasia on Request in the Netherlands" Bioethics, vol 1, 1987, pp156-62.
  4. St Thomas Aquinas, Summa Theologica II, ii, Q.64, A.5.
  5. J.J.C. Smart & Bernard Williams, Utilitarianism For and Against (Cambridge University Press, 1973) pp.98-9.
  6. Arthur Hugh Clough, "The Latest Decalogue".
  7. Quoted Singer op. cit. p.156.
  8. David Fisher, Morality and the Bomb ( London, Croom Helm, 1985) pp.38-9.
  9. Sue Rodriguez v The Attorney General of Canada, Canada Supreme Court Reports, Part 4, 1993, vol.3, p603.
  10. Quoted Singer op. cit. p. 140.







Dear David,


Many thanks for your interesting further contribution to our mutual exploration of philosophical examples. You say, rightly, that it is legitimate for philosophers to use simplified examples to make essential points about ethical dilemmas. I would not dream of objecting to this. But you also enter a caveat: ‘provided they do not stray beyond the bound of possibility’. But this is just where the Pedro and Jim example does go astray.


My point is that, while it is not impossible for Pedro to be in the habit of keeping his promises, it is impossible for Jim to know this. The very fact that the information about Pedro's promise-keeping habit has to be given to us, as readers, by a piece of ‘omniscient narration’, gives the game away; because it shows that there is no way in which Jim could come by this knowledge through his own experience. Indeed everything he knows of Pedro points the other way. Of course if, per impossible, Jim knew that Pedro would keep his promise, he would be in a very difficult ethical dilemma. But the point is that the dilemma cannot arise in the way it is alleged to arise: it is ‘beyond the bound of possibility’. In other words, this example does not present us with a useful, simplified example of an ethical quandary, since nobody in Jim’s position could (or should) trust Pedro to keep his word, and the dilemma alleged by this example thereby collapses. Of course, if a genuine example were presented, in which somebody were morally compelled intentionally to kill an innocent person, it would be up to me (or anybody else who thinks like me) to revise my opinion. (Incidentally, I have pondered the Richards examples a bit more, in the light of a report, or suggestion, that at one time Osama Bin Laden might be hiding in a cave with his children. This thought led me to some further reflections which are to be found in the attachment to this letter).


Your examples of the craniotomy case and of the airliner approaching a skyscraper are more troubling, and need deep examination. I’m sorry if you think such examination is ‘Jesuitical wriggling’. Surely the point is that, since the proposed actions in these examples are so horrific, we must exert every sinew to try to avoid having to consent to anybody committing them. If this involves a lot of preliminary questioning and probing, so be it: the examples deserve it.


The craniotomy case is something on which the medical experts, especially those who think of themselves as ‘pro-lifers’, have spent much time and thought, and I would like to know what they say. I have no expertise on it, so my opinion is not worth a lot. But one thing seems clear: namely that the doctor alone is answerable for the baby’s death. (I’m assuming that the mother did not have the chance to agree to her baby being killed). There is no way that he can share the guilt with somebody else. (This makes it quite different from the skyscraper terrorist case: see below). When, after it is all over, and he has to confront the mother, the doctor ought to say to her (or at least to himself): ‘yes, I’m very sorry but, for the best of motives, I killed your baby’. Merely saying ‘I’m afraid we couldn’t save the baby’ would not be enough to get him off the moral hook, or assuage his conscience, even though it might be the most emollient thing to say in the circumstances. For the fact is that he, and he alone, chose to kill the baby, and he has to admit this to himself (which would be extremely difficult). We also have to assume something that, purely as a layman, I am unsure about: namely that there is no possible way of anticipating the eventuality and doing something about it by e.g. a caesarean section. That is, we have to be absolutely sure that there was no question of carelessness, negligence, or lack of foresight (or money) on the part of the doctor, or other medical staff. The more frequently such circumstances arise, the more likely it is that such an eventuality can, and ought, to be foreseen and pre-empted. My only other remark at this point is that, where a dilemma is genuinely tragic, that is, where there is no solution which avoids doing something forbidden, the concept of ‘justifying’ the action taken, whatever it is, seems out of place. Was Agamemnon justified in killing Iphigenia? is surely a question that would have made no sense to Aeschylus: it was simply something he ‘had to do’. Beyond this, I don’t think I can go at this stage, and have to reserve my judgement until I can find out more about the medical problem itself. I realise that this will seem an inadequate response, which it is: but I need to learn more before I can make up my own mind.


The case of an airliner approaching a skyscraper, after Sept 11, is equally troubling. But perhaps one clue to a response lies in the fact that, beyond a certain point, whatever the fighter pilot does, or does not do, the passengers are doomed anyway (just as the baby is allegedly doomed in the craniotomy case). The only question then is whether they die because the pilot shoots them down, or they die because he refrains from shooting and the airliner crashes into the tower. But the two cases are different in another way: a doctor (of all people!) killing a baby at close quarters by crushing its skull seems much more horrific than killing thirty air-passengers by mechanical means from a safe distance. Yet I think we will agree that the key ethical point is not about this. Nor is it about the difference between choosing one life or two (as in the craniotomy case), and choosing between thirty and several thousand (as in the airliner case). So here is a good philosophical example (unlike Jim and Pedro).


My thoughts here begin with the phrase: ‘beyond a certain point’. For at the outset it is likely to be unclear exactly what is going on. But the pilot should certainly not shoot the airliner down (nor should his commander give the order) until it is clear beyond reasonable doubt that the passengers are already doomed. This is likely to be a difficult point to establish, and mistakes made. How do you distinguish between a plane that is involved in a catastrophic navigational or mechanical accident (like the Dutch example where a jumbo jet crashed into a residential area some years ago), and an empty plane that is being flown by terrorists on their own? or a plane full of passengers that has been hi-jacked? It would take time to discover this. If, as a result, a mistake (either of omission or commission) is made, so be it: such is life. But then, whatever happens, the mistake cannot be justified, by comparing what was done against some rule about what should have been done. We can’t justify mistakes: we can only be sorry for them.


Once it is clear beyond reasonable doubt that the passengers are already doomed, the really difficult question arises: would shooting down the airliner by the fighter pilot in these circumstances constitute intentionally killing, i.e. murdering, the innocent passengers? (I am deliberately, and I think correctly, naming any intentional killing of the innocent as murder, whether or not this is what the law says). Here the key point to hang on to is that, whichever way the passengers meet their deaths, their murderers are the terrorist hijackers. So the question then is: is the fighter pilot an accessory to these terrorist murders? Is he also guilty of their murder? It seems to me not. First of all, it would clearly be a nonsense to say that the fighter pilot was in any way responsible for the murder of the people in the skyscraper, if the plane crashed into it because the pilot did nothing. The only question is whether he is morally responsible, as an accessory, for the murder by the terrorists of the passengers in the aeroplane, if he shoots the plane down. This seems implausible: how can you be an ‘accessory’ to a murder when all your efforts are being spent in trying to thwart the crime of the real murderers? It would be more accurate to say that what the pilot was doing was, as far as in his power, saving as many as possible of the terrorists’ victims, by licitly killing the terrorists before the rest of their victims were killed. The deaths of the passengers were genuinely incidental to what the pilot was doing, in the sense that only the manner of their deaths, not that they were killed, was his responsibility. This means he is not a murderer. Admittedly this line of reasoning is horribly close to Richards’s talk of having almost ‘innocently murdered’ the people in the dugout. Yet the case is not really like that of Richards, simply because here a third party (i.e. the terrorist gang) has already doomed the victims to death, irrespective of what the person does who decides the manner of their death.


My objection to your own alternative way of describing what happens, which presumably is that the pilot murders the passengers in order to save the office workers in the sky-scraper, and that this is morally OK, is that it is not a true description of what the pilot is doing. Throughout, it is the terrorists are doing the killing; and to say that the pilot is helping them to kill the passengers is a mis-description of what he is doing. For one thing, they don’t need any help, they are doing the job by themselves. So he is not an ‘accessory’. But neither is he intentionally killing them himself, since for all practical purposes the passengers are dead already. To suggest that, in the circumstances, he intentionally but rightfully murders the passengers is false; it leaves to one side the murderousness of the terrorists, and exaggerates the responsibility of the pilot by turning him into an ‘innocent murderer’ (which of course is a contradiction in terms).


Of course, the heroic solution to the dilemma is for the passengers themselves to overwhelm the terrorists, as they did on September 11th in one case. I think that this example brings us to a point at which theology helps to supplement a purely ethical argument. The sheer heroic self-sacrifice of those people ought to be recognised by a posthumous equivalent of the George Cross. I would have no difficulty if they were declared martyr-saints. For they provide the best, perhaps the only, genuine answer to the dilemma we are discussing. Their example shows that, beyond a certain point, heroic self-sacrifice is required of us if we are to remain fully human. I don’t find this a difficult truth to accept: it is simply part of what the Christian religion sometimes asks of people.


Brian Wicker 07.01.02





Dear David,




At the recent CCADD meeting with Nicholas Denyer on ‘Just War’ I raised a question which elicited some discussion. Suppose Osama bin Laden is spotted in a cave somewhere in the Hindu Kush, surrounded by his children. An anti-Taliban fighter has come near, armed with a bomb, and without warning has lobbed his bomb into the cave. The bomb was designed in such a way that it would kill everyone in a cave. It went off as it was designed to do, and everyone in the cave was killed. The question then is: did the soldier kill the children intentionally? Or to put the question another way: was what the soldier did an intentional act of killing the innocent? Or was it (as some people in the discussion suggested) just a ‘side-effect’ of the action, merely ‘collateral’?


To answer this question, we must assume several things, which are not in dispute among just war theorists: i) Osama himself was a legitimate target of lethal attack in self-defence because of overwhelming evidence of his crimes against humanity; ii) his children were innocents simply because they were children, doing no harm to anyone; iii) it is forbidden intentionally to kill any innocent person, even in war; iv) bringing about the deaths of innocents in war as side-effects of military operations, even when these deaths are foreseen as virtual certainties, can in certain circumstances be licit. So, were these circumstances applicable in this case?


I am supposing, purely for the sake of brevity, that the anti-Taliban fighter understands and accepts the Anscombian arguments about intention which Denyer referred to in his paper. Without this assumption the dialogue could go on round in circles forever, without producing any significant result. So to discover the fighter’s intention in acting in the way he did, it is necessary to ask:


‘Why did you lob the bomb?’


‘Because we had been told to do so if we found Osama.’


‘But you knew the children were there as well. You could have disobeyed orders and taken the consequences, so why didn’t you?’


‘Our orders were to kill Osama, if that was possible’.


‘But you killed the children as well. So you lobbed the bomb in order to kill the children too?’


‘No, I intended to kill only Osama. The deaths of the children were unintended side-effects of what I intentionally did.’


‘But you knew that if the bomb went off, it would kill everyone in the cave. So what you did was intentionally to bring about what the bomb was designed to bring about, namely the deaths of everybody in the cave, the children as well as Osama. So you did intend to kill the children.’


‘No, I intended to kill only Osama.’


‘You mean that you directed your intention only to the place where Osama was, avoiding directing your intention to the places where the children were? So you regard your intention as a kind of search light in your mind, by which you were able to pick out one person only, and leave the rest in the dark, even though the bomb itself could not make this distinction?’


‘No, this is an absurd theory of intention’.


‘Agreed. Nevertheless, you intentionally brought about the deaths of everybody in the cave, since this is what your bomb was designed to bring about, and you intentionally detonated it. So you intentionally brought about the deaths of the children, didn’t you?’


At this point, if the fighter says Yes, the argument is over: he admits he has intentionally killed the innocents. So to get off this hook he has to say No.


‘No. I intentionally brought about the death only of Osama. I did not intentionally bring about the deaths of the children. I had nothing against them! But throwing the bomb was the only way to get at Osama. I didn’t want to kill the children. I wished I had had a different sort of bomb, designed with a targeting mechanism which could pick out and kill just one individual in the cave and leave the rest alone. If I could, I would have used that’.


‘But you didn’t have such a bomb’.


‘No. So the only way to get at Osama was to throw in the one bomb I had.’


‘But your talk of its being the "only way" to get at Osama is ambiguous. Either you mean literally that the way to Osama was blocked by the children (perhaps they were sitting between you and him). If you had had a gun, you would then have had to shoot the children first in order to get at Osama later. But that would have meant that you would have intentionally shot the children. Alternatively, by your talk of "the only way" to get at Osama you may be meaning that killing everybody was the only method available for killing Osama. But saying that killing the children was part of the only available method for killing Osama, amounts to saying that killing them was part of the means you used for your end, namely killing Osama. And killing the innocent as a means even to a good end is itself forbidden’.


‘But killing the children was only part of what I did. That’s my point. I admit that the only means to the end of killing Osama included killing the children. But that does not necessarily mean that killing the children was the means to that end’.


‘So the killing of the children was included in the means you had to choose to kill Osama, but yet was somehow a separable part of that means? You have already admitted that killing the children was an unavoidable part of the means, or method. Now you seem to be contradicting yourself, by saying that killing the children was avoidable in some other sense of the word, a sense which gets you off the moral hook. What sense can this be, given that throwing the bomb and killing everybody was the only means you had? Are you saying that the means to killing Osama included several distinct actions, not all of which were intentional, so that killing the children could be described as a separate action from killing Osama: an action done with a different intention? If so, please describe this separate action’.


The fighter seems stumped by this request. The enquiry goes on:


‘Surely the fact is that, in killing Osama, you did nothing else, no other action, than throw the bomb into the cave! The bomb itself did the rest! Of course, the bomb’s own ‘action’ (if we can call it that), was not intentional, because bombs can’t have intentions. So we are back with the question: why did you do the one and only intentional action you performed, which was to throw the bomb? It was your intention in the one action of throwing the bomb that we need to elucidate. When you suggest that the means by which you killed bin Laden included killing the children, but in such a way that this included bit of what you did was not done with the same intention as the rest of the one and only action you did, how would you describe this included bit? Was the child-killing action just part of the larger action of killing everybody, in the way that (say) boiling the greens is part of the action of getting dinner ready? But then, boiling the greens is done with the intention of getting the dinner ready, so that example doesn’t help you. Or perhaps it’s like saying that pulling the trigger is ‘included’ in the action of shooting someone? But this example doesn’t help you either, since you can’t intentionally shoot anybody without intentionally pulling the trigger. Of course you can, in certain circumstances, describe the action of trigger pulling in such a way that it is not included in a larger action of shooting anybody. You could be intentionally fiddling with the trigger to find out how it worked, and were doing so with the gun pointing at somebody, who then got shot. But then killing that person would have been an accident. On the other hand, if your intention is to shoot somebody, then that is your intention in pulling the trigger. Pulling the trigger is not a side effect of the shooting: it is unavoidably ‘included’ in the act of shooting. Is this what you mean by ‘included’? Surely, when you threw the bomb into the cave, killing the children was not a side effect of killing Osama, but was ‘included’ in it, like pulling the trigger. But if this is the kind of thing you meant when you said that killing the children was unavoidably included in the killing of Osama, then you can’t say that killing the children was not intended.’


‘But killing the children was not like pulling the trigger in an act of shooting. The relation between the killing of the children and the killing of Osama was not like the relation between pulling the trigger and shooting the person’.


‘So you say: but what exactly is the difference? And how does this difference show that the intentional act of killing Osama did not ‘include’ your (allegedly unintentional) act of killing the children, even though everybody in the cave was killed by your one act of throwing the bomb?’


‘But as you say, I didn’t do more than one act, which was to throw the bomb. So my point is that I did not perform any action of killing the children at all. My only action was throwing the bomb, which resulted in the killing of Osama. The death of the children was not the result of any action of mine at all, but was merely the unintended side effect of my action.’


‘But their deaths were plainly the result of your one and only action, which was the throwing of the bomb.’


‘If it was a result, it was an unintended result.’


‘In what way was it unintended? Do you mean that you did not intend this result any more than you intended the falling of lumps of rock from the cave wall as a result of the bomb going off? But surely that can’t be so. You intended to bring about anything that the bomb was designed to bring about, and that included the falling of the rocks. It’s merely that the falling of the rocks was not in your mind when you threw the bomb: it didn’t matter to you. But surely you don’t claim that the deaths of the children didn’t matter to you in that way? If that were so, then you are certainly guilty of reckless disregard of human life, if not murder!’


The fighter seems stumped again. Has anybody got a good defence of him? Or is the fighter simply excusing himself by a piece of self-contradictory nonsense, like the excuse David Fisher quotes, in Morality and the Bomb

p. 36 from Frank Richards (who had nearly thrown a bomb into a cave containing some innocents, but shouted a warning just in time): ‘If the young lady had not cried out when she did, we would have innocently murdered them all’ (my italics).


Brian Wicker. 10.12.01







Dear David,


I think I can now add a word of comment on your paper on ‘A Right to Die?’ which I was not able to do in the note in sent you earlier.


First of all, the ‘craniotomy’ example. As I said earlier, I was not capable of giving you an adequate response to this example without better medical knowledge. But I now feel better able to reply, having just had a long and extremely interesting and informative telephone conversation with Virginia Griffin, who is a midwifery expert, and the fertility co-ordinator for the organisation called ‘Life’. As she explains it (I hope I am interpreting her correctly) the situation is as follows:


  1. ‘impaction’: this is the technical term for the case where a baby’s head gets stuck and cannot be got through the mother’s pelvis (the case you quote from Singer). In modern medicine this kind of case never presents the dilemma Singer poses, of having to kill the baby. The point is that, even at the stage he describes, a caesarean section is still capable of being carried out in such a way that, if the baby is not dead already (which it often is), both lives can expect to be saved. The situation, of course, is different in primitive circumstances where there is no adequate medical care available. In the past, both mother and child would normally die anyway. Nothing could be done about it. But even a ‘craniotomy’ is no answer to the ‘primitive’ case, because such an operation itself requires considerable medical equipment and expertise, and the mother needs an anaesthetic, and has to be in a good enough condition to undergo one (which often is not the case where the crisis occurs very late in pregnancy). Furthermore, there is always a considerable danger of killing both mother and baby in the course of a ‘craniotomy’ operation. In modern conditions it is not difficult to foresee any problems in this area, and deal with them satisfactorily e.g. by caesarean section. So, in short, the dilemma that Singer poses never needs to arise in modern medical circumstances, and where adequate facilities are absent both mother and baby are likely to die anyway. The conclusion seems to be plain, that no responsible doctor today would, or needs to recommend a craniotomy as a way of solving the alleged quandary. So this example does not present a valid example of a dilemma soluble only by killing the innocent.

  3. A much more difficult modern situation concerns a condition called ‘eclampsia’. This may happen to a mother with high blood pressure. It can mean that she will certainly die unless the pregnancy is terminated. If this done, there is a chance of the mother’s recovery. But the point here is that in this condition, where the termination has to be performed, the baby is doomed already: it has no chance of long-term survival anyway. Nevertheless, because terminating the pregnancy can save the mother at the cost of killing the innocent, even though the innocent is not going to live for long without it, this example is a genuine dilemma. Virginia Griffin tells me that she would not object to the termination of the pregnancy in such a case, because there is anyway no chance of the baby surviving to the stage of being born alive. Accepting what she says, I suppose my answer to this very difficult situation is essentially the same as my answer to the ‘shooting down of the airliner’ example. All that the doctor is doing in terminating the pregnancy is determining the manner of the innocent’s death, not its death as such. I think that if my argument is valid in the one case, then it is valid in the other. You may regard the distinction as a quibble, but I don’t think it is. Good examples like this one help us to make necessary distinctions. That is their value, as we agree. These distinctions may look like ‘jesuitical wriggling’ but they are not – any more than some of what may seem ‘jesuitical wriggling’ that I have seen in defence of nuclear deterrence!


As to the rest of your ‘Right to Die’ paper, I have no quarrel with most of it. The Dutch rules look pretty sound to me, once a right to die is granted: though I would want to add that as soon as medical advances make it possible to render the patient’s sufferings bearable (condition 4 of the list) the legal permission for euthanasia ought, logically, to be withdrawn. But it probably would not be, and this is where the ‘slippery slope’ argument begins to bite. The second point is to ask whether a person who has asked to be killed, for whatever reason, is any longer ‘innocent’ in the sense of having done, or intended, no harm. For being killed certainly does you harm! And it seems to me that consequently asking for this harm to be done to you makes you no longer ‘innocent’ in the relevant sense - even though your life is unbearable. And the doctor who kills you is surely an accessory to that harm-doing. A person who asks to be killed is clearly not ‘innocent’ in the way that a harmless child being bombed, or killed in the womb, is innocent. For this reason, even if the euthanasia were legitimate, this would be because it would not properly be a case of killing an innocent.


Of course, I accept that legalising a right to die might be justified if the ‘slippery slope’ argument could be met. But you try to justify this by saying that the harm done by the euthanasia is less than the harm of the unbearable pain. But, without in any way evading the inescapable unbearability of the pain, and the doctor’s unquestioned duty to relieve it, I maintain that this proportionality argument remains insoluble, and is no way forward for justifying a lethal procedure for relieving an appalling calamity for the sufferer. That is why I don't accept your final sentence, on the ground that the two harms are not comparable in kind. You can’t really compare a degree of pain with sheer extinction: they are in quite different categories, and a comparison of the amounts of suffering done by each is senseless (though doubtless the sufferer cannot accept this). If there is a good case for legalising a right to die this proportionality argument is not it!


Yours ever, Brian


PS. In my next, I offer a supplement to your Right to Die? piece







Dear David,


Here are some supplementary comments on the ‘Right to Die’


I think I have not quite answered your point about the ‘craniotomy’ example. If, at an extremity which lay beyond the reach of modern medical care, a doctor were somehow faced with trying to save a mother by attempting a craniotomy on the baby, this again would be a case of determining the manner of the baby’s death, but not its death as such.

For the baby would be doomed already. In other words, although the doctor would be killing the baby, this might not count as murder since he would only be hastening the inevitable. I wonder how such a defence would stand up in a trial for murder, if such were brought against the doctor? I admit I am not happy about this argument: I am interested to hear what you think of it. It looks like hair-splitting. But it just seems to me a better argument than having to live with the paradox of saying, as you seem to suggest, that it is murder but this kind of murder is morally ok.


An obvious objection to my way of dealing with the difficulty might be put thus: we are all doomed to die some time, so every murder is excusable as simply hastening the inevitable, i.e. determining the manner of death of the victim, but not his death as such. This is obviously an absurd and unacceptable reply. But what exactly is wrong with it? Well one point is that, both in the airliner terrorist case and the craniotomy case, there are only two ways of dying involved, not the whole range of ways of dying which lie open before us simply by our being alive. So the two examples are not comparable. My arguments holds, if at all, only where the person faced with dealing out death already knows what these alternatives are, and thus what sort of death the victim will undergo if he chooses to do nothing. Whereas an ordinary criminal murderer has no idea what manner of death the victim will die if he, the murderer, refrains from committing his crime. This surely makes all the difference. But why? That is the question that troubles me. Of course, this argument might also apply to euthanasia, at least under the Dutch rules. This is why the slippery slope argument is the decisive one in that case.


Brian Wicker






Dear Brian,

Many thanks for your interesting and thought provoking letters. In response let me offer a few more thoughts.

I am still not persuaded by your arguments on the Jim and the Indians example. I accept that Williams has perhaps cut the story- telling a little too short and perhaps owes us a little more explanation as to why Pedro can be trusted to keep his promise. Perhaps he has sworn an oath on his mother's life and, like Reg Kray whom I cited earlier, is inordinately fond of his dear old Mum as well as being a serial killer. Nothing implausible about that and, more crucially, nothing implausible about the fact that out ethical life can throw up real and worrying dilemmas of the kind helpfully illuminated by the Jim and the Indians example. Hence my view remains that it is a legitimate example, even if there may be a few shortcomings in the story telling.

But let us turn to the airline example. You attach great importance to the fact that in this example the passengers are doomed anyway. So let us vary the example slightly. The terrorists are not suicide bombers but bombers. They are not going to crash the plane into the Twin Towers but to drop highly effective explosives onto the Towers with similarly devastating impact and many thousands of lives lost. Their plan, once the deadly munitions have been dropped, is to fly the plane on to a pre-arranged field and make off in a getaway car. The hijacked passengers will be left, safe but frightened. You are the pilot of a fighter plane faced with the choice whether to shoot the plane down, killing thirty innocent passengers as the only way to save many thousands of lives. What do you do? As before, it seems to me licit for the pilot to shoot down the plane in order to save the thousands of lives that will be otherwise lost.

Let me now broaden out the difficulty I have with your approach. You want to hold absolute the prohibition on killing the innocent. A perfectly respectable philosophical position. But the way you defend it does not seem to me legitimate. For each time an example is produced showing that we can be faced with difficult ethical choices, you pick holes in the example in order to try to show that the absolute principle never comes under strain. But that is surely an elaborate form of self-deception. For in the post-lapsarian world we inhabit it is a fact, however regrettable, that we are all too frequently faced with such agonising dilemmas. To pretend we are not seems in Sartrean terms mauvaise foie.

The dilemmas can arise in dealing with criminals, terrorists, in war, in peace, in our homes, in our schools, and in hospitals. In hospitals, I certainly do not share your sanguine belief that impaction can always be dealt with by Caesarean section and we may never be faced with a choice of whether or not to perform a craniotomy. And we may have ageing relatives (or may be ageing relatives!) who choose to die in order to end unbearable suffering. And, if that is their choice, I doubt they will be very impressed to be told they cannot so choose because you consider death and suffering to be incommensurables. If they are Christians they may also be puzzled by your suggestion that death is extinction! I can quite understand an absolutist who argues in the face of such ethical dilemmas that the principle must be adhered to even when it causes suffering . He may perhaps holds that its breach would bring even greater suffering (the slippery slope argument). I would argue against this but would regard it as a perfectly legitimate manoeuvre. What does not seem to me legitimate is to defend absolutism by pretending we never face such dilemmas. Because we do. Hence the potential for tragedy in human life.







Dear David,


Your latest contribution to our discussion is most welcome, and I am happy to respond, at several levels.


You defend the Pedro and Jim example, despite my criticism of it, by saying that it represents a legitimate simplification which is used to make an essential point. But I think this is a mistaken way of looking at it. What is wrong with it has nothing to do with its being a simplification (I have nothing against that), but with its muddling up two quite distinct categories: on the one hand what I have called ‘moral theorems’ (which can be written out as valid syllogisms), and on the other hand ‘real life’ examples (of which the key attribute is that they can always be cast in the form of a narrative). What may be valid in the one may be invalid in the other. The trouble with the example as given by Bernard Williams is that it falls between the two, and is not valid in either. I think an essentially parallel point can be made about the World Trade Towers example you interestingly introduced into our debate. To explain this more clearly, I attach a statement of my argument about this, called Narrative Examples in Moral Philosophy.


You accuse me, understandably but I think wrongly, of ‘wriggling’ (just as I am tempted to think of the arguments in defence of cold war nuclear deterrence as ‘wriggling’). But this accusation is idle in both cases: all that matters is whether the argument is valid, and stands up to the test of achieving rationally what it sets out to do. If it is valid, no accusation of ‘wriggling’ can shake it. But I do think that your arguments fail to appreciate the enormity of any intentional killing of the innocent. Here ethics cannot avoid becoming theology. The reason why killing the innocent is so appalling is that it is always a participation in, a kind of sacrament of, the pivotal event in human history: the killing of God on Good Friday. The killing of Jesus, the quintessentially innocent, as Pilate saw (Matt. 27:23: ‘what harm has he done?’) reveals, as no other event ever could, what ‘the sin of the world’ really is; and thereby overcomes it. Good Friday dramatises how blasphemous the killing of the innocent must always be. This is surely why murder of the innocent is so fundamental a crime.


Nevertheless, justice sometimes demands that people be killed for the good of others, in self-defence etc. For that very reason, it is right, when we are presented with plausible cases purporting to justify murder, to use all available resources avoid being entangled in the trap. This is why some attempts to refute such tempting examples are likely to appear as ‘wriggling’. For in the nature of the case they have to use reason and human understanding at, sometimes beyond, full stretch. But if such refutations are valid, I think they should be welcomed. Of course, if they are invalid, they have to be instantly jettisoned. But not until then.


Yours ever,


Brian 29.01.02







Earlier in this debate I distinguished ‘moral theorems’ from ‘real life cases’. I want to explain this point further, by referring to our old friends Jim and Pedro.


This is how Smart and Williams stated the Jim and Pedro dilemma (J.J.C.Smart and Bernard Williams, Utilitarianism For and Against, CUP, 1973, pp. 98-9, quoted in David Fisher, Morality and the Bomb, Croom Helm, 1985 pp. 39-40):


Jim finds himself in the central square of a small South American town. Tied up against the wall are a row of twenty Indians, most terrified, a few defiant, in front of them several men in uniform. A heavy man in a sweat-stained khaki shirt turns out to be the captain in charge, and after a good deal of questioning of Jim, which establishes that he got there by accident while on a botanical expedition, explains that the Indians are a random group of the inhabitants who, after recent acts of protest against the government, are just about to be killed to remind other possible protestors of the advantages of not protesting. However, since Jim is an honoured visitor from another land, the captain is happy to offer him a guest’s privilege of killing one of the Indians himself. If Jim accepts, then as a special mark of the occasion, the other Indians will be let off. Of course, if Jim refuses, then there is no special occasion and Pedro here will do what he was about to do when Jim arrived, and kill them all..The men against the wall, and the other villagers, understand the situation, and are obviously begging him to accept.


The question is: what should Jim do? David Fisher says that, faced with such an appalling moral choice, Jim should shoot one of the Indians in order to save the lives of 19.


I maintain that this conclusion is a mistake, because it fails to analyse sufficiently the nature of this example.


The situation depicted can be analysed either as a ‘moral theorem’ or as a plausible narrative of a real life dilemma. But the story as presented above hovers between the two, and cannot partake of the validity of either.


As a moral theorem, the essentials of the argument can be summarised as follows (let’s call which ever man Jim decides to shoot, Paulo):


  1. Jim is invited to shoot Paulo in order to prevent Pedro from killing all the other prisoners.

  3. Jim knows that if he kills Paulo, Pedro will let all the other prisoners go free.

  5. Ergo, Jim ought to shoot Paulo.


Given certain assumptions, this seems to me to be a valid syllogism. The conclusion follows from the premises: i.e. Jim ought to shoot Paulo. Of course, the validity of the syllogism depends on certain basic assumptions being true. One is that the ‘greatest good’ (?happiness?pleasure?) of the greatest number’ principle is the key criterion for making true judgements in ethics. Another is that arguments purporting to show that ’ought’ can never be derived from ‘is’ are false. I accept the second, and anyhow ‘ought’ has no clear meaning in the absence of some more basic assumptions being true. There are grave objections to the first, but these don’t affect my point.


The trouble is that the second premise of the syllogism represents an impossibility. Knowledge of the kind required to create the dilemma for Jim cannot be had. Consider a logically parallel syllogism, to make this point clear:


  1. Fred is an astronaut on the moon who is the only man able to pilot his team home safely to earth. But he is starving to death. To get home safely, he must eat something, such as green cheese.

  3. The moon is made of green cheese.

  5. Ergo Fred ought to eat a piece of the moon.


This too is a valid syllogism, given the truth of certain assumptions with which I have no quarrel, such as that Fred owes it to his men to get them safely home. But here too, the second premise represents an impossibility, for the moon cannot be made of green cheese. So the syllogism does not really ‘prove’ that Fred ought to eat part of the moon. Similarly Jim cannot know that Pedro will release his remaining prisoners.


Now let’s turn the essentials of the Jim and Pedro example into a narrative, leaving out the unnecessary details which do not affect the key issue:


One hot day, Jim walked into a South American town square. He was interrogated there by an army captain Pedro, who had an armalite rifle over his shoulder, and who was holding twenty prisoners shackled by a wall. Pedro said to Jim: ‘If you shoot one of these men I will let all the others go free, as the mark of a special occasion’. But Jim protested: ‘How do I know you will keep that promise? After all, you are a would-be mass-murderer. Surely you wouldn’t stop at breaking the odd promise, if it helped you to get your way?’ To this Pedro replied: ‘I love my mother, and she always told me never to break promises, so I won’t break this one, because I still love her so much’. So Jim shot one of the men, Paulo, in the hope that thereby the other prisoners would be spared.


{Of course this is not the end of the story. We want to know what happened after Jim shot Paulo. I suggest three plausible endings: doubtless there are lots of other possibilities}:


a) After Jim had shot Paulo, Pedro laughed very loudly, and promptly shot all the other prisoners himself. Then he bought himself a beer, and offered Jim one as well.


b) After Jim had shot Paulo, Pedro was as good as his word and unshackled all the other prisoners and let them go, but with a warning to them not to join the local rebels.


c) As b) but with the further addition: Pedro then turned round and shot Jim, to prevent the evidence, of his having killed an innocent person on Pedro’s invitation, being published in the local papers.


Several points are worth making here:


  1. all three endings are plausible, given the character-evidence given in the original, although I think b) is somewhat less likely than c).
  2. in the case of a) since Pedro has now shot the prisoners, the implication in the original was false: Jim did not ‘know’ they would be let free.
  3. Jim is clearly not only a fool but a knave to kill Paulo on the flimsy evidence before him.
  4. c) demonstrates just how far Pedro’s word can be relied on, and how much he loves his mother. HHe keeps his promise not to kill the other prisoners. He just kills Jim instead.


The big point here is that, in this narrative, the necessary knowledge required to put Jim in the alleged dilemma simply cannot be had. There can be no justification for the assumption implied in the original that Jim knew Pedro would keep his word. From what he has seen and heard, Jim can have had no such knowledge. This means that the statement that Pedro would certainly keep his promise cannot come out of the story as it stands. It can only be introduced into it by an ‘omniscient narrator’ (which is what the original covertly contains). Let me clarify what I mean.


Let us first try to introduce the required knowledge into the story without the insertion of an omniscient narrator, by adding the following (after Jim’s protestation of doubt):


..at this point Pedro called a couple of his subordinates over. Pedro said them: ‘This bloke Jim has cast doubt on my reliability in keeping promises. Can you vouch for me, please?’ They both replied: ‘Of course Pedro, anything you say. We know you love your old mother to bits, and we know you would never go against her advice. So, Jim, we promise you that if you shoot one of these men, Pedro here will let the others go free. We’ll put our bottom dollar on it’. So with this assurance Jim shot Paulo..(Any of the above-suggested endings then follow).


It seems obvious that this (or any other reference by other characters in the story) cannot provide Jim with the knowledge he needs. For he still has to assess the subordinates’ reliability, just as he has had to do with Pedro. He cannot know that what they say is true. So let’s try another tack:


..at this point Jim began to feel a bit dizzy. He fell into a trance, and saw God coming down to him in the shape of a dove. God said to him: ‘It’s OK for you to shoot Paulo. I give you my full permission, because I know that Pedro will keep this particular promise’. Thus reassured, Jim went ahead and shot Paulo..


Does this version offer any more certainty than before? Can we trust it? - as we are invited to trust the dream given to the three kings in Matthew 2:12? or the baptismal manifestation in Matthew 3:16-17? How should we, as readers of the tale, take this stuff about God appearing to Jim in a trance? Does it mean that Jim now knows that Pedro will keep his word? Only if we can believe in the reliability of the tale about the trance; and this means relying on the integrity and skill of the ‘omniscient narrator’ who apparently is privy, not only to Jim’s unexpressed thoughts and feelings, but to God’s mind as well. If this omniscient narrator hasn’t convinced us, we have good reason to distrust him too. (Unreliable ‘omniscient narrators’ are common enough, not just in bad, but also in very sophisticated novels).


The conclusion seems plain. To get Jim to know what will happen, we have to rely on a further character whose reliability has itself to be assessed: namely that of the ‘omniscient narrator’. Thus the problem is only repeated further along the line. We still have to decide whether the information given is to be trusted. Even bringing God into the picture hasn’t solved the difficulty.


What this argument shows is that the example quite fails to prove what it is alleged to prove: namely that Jim ought to shoot Paulo. I am not picking holes in the example simply in order to evade its conclusion. I’m just thinking-through the case. My argument is that when you think sufficiently hard about the way this narrative example is being used, the argument from it collapses. I maintain that whichever way it is analysed, whether as a moral ‘theorem’ or as a ‘real life’ narrative, the result is the same: not proven.


I want to make essentially the same point about the interesting variant on the World Trade Towers example (see Fisher, 27th January 2002). It would be quite easy to analyse this as another valid moral theorem:


  1. The fighter-pilot is told by operation control that he has their permission (or perhaps command) to shoot down the airliner, i.e. kill the innocent passengers, in order to save the people in the towers from the inevitable fate planned for them by the hijackers.

  3. The pilot knows that, unless he stops them, the hijackers will kill the people in the towers.

  5. Ergo: the fighter pilot ought to kill the passengers in order to save the people in the towers.


Here again, the second premise is an impossibility: short of an omniscient author, the pilot in this situation cannot have the required knowledge, without which the argument collapses. Consider the real-life scenario:


Operation control have overheard the hijackers talking on their intercom. They have been giving each other instructions about their planned bombing of the towers. These messages have also indicated that they will let the passengers go free after it is all over. At the same time (as in the real case of the airliner that was crashed into a field on September 11) operation control have been hearing some of the passengers talking about a plan to overwhelm the hijackers, even at the risk of their own lives. In the light of all this information operation control has given the pilot an order to shoot the plane down, i.e. kill the passengers. But throughout, both the pilot and operation control have been asking themselves such questions as: are these messages we hear from the airliner genuine? Do the hijackers mean exactly what they say? Is it all a gigantic trick? What chance do the passengers in the plane have of overwhelming the hijackers? what chance is there of the hijackers’ bombs not going off, or missing their target? How long can we wait before making a decision. etc. etc. The pilot has also been wondering how reliable operation control’s information is.


When the order comes, what should he do? Well, he may well decide to obey orders and kill the passengers. And it may well turn out afterwards that the whole thing was genuine, and he is congratulated for having gallantly killed them. It is possible that he saved the people in the towers. But equally it may turn out that the hi-jacking was all a ruse to create panic. The hijackers may turn out to have been simply asylum-seekers, like the hijackers at Stansted Airport. Or, having killed the people in the towers, they may then turn round and kill the passengers themselves. In short, all kinds of tragic circumstances may arise, and people as a consequence may do all kinds of brave or appalling things. What happens may excuse what was done.


On the other hand, while the pilot calculated that he was doing the right thing, his calculation may prove to have been wrong. It was not based on knowledge. Yet surely the point of the story is to show that anybody, faced with this scenario, should do what the pilot calculated it was best to do. But since the calculation could be mistaken, another pilot could well, given the same evidence, calculate differently. And it is possible that his calculation was the correct one. So the example does not prove that, in such a situation, the pilot ought to kill the passengers.


Of course, if I were asked: what would you do in the circumstances? my honest answer is that I might well obey orders and kill the passengers. But this wouldn’t prove I was right. And if it turns out afterwards that the hijackers, having bombed the towers, killed the passengers too, this would make the idea that the pilot ought to have killed them himself absurd. Who is to say that an outcome in which the hijackers kill the passengers is ruled out? It can only be ruled out by an omniscient narrator who has complete control over what happens in the story, and who can finish it any way he likes.


The point then is that in real life there is no character who is an ‘omniscient narrator’: such a character exists only where stories are, or can be read as, fiction.


Brian 29.01.02






Dear Brian,

I enjoyed reading your latest piece and am happy to concede, as I have before, that there may be shortcomings with Williams' artificial example of Jim and the Indians. That is why I have introduced into the debate real world examples such as craniotomy and the attack on the Twin Towers.

Your complaint about the latter is that the pilot cannot know that, unless he stops them, the hijackers will kill the people in the Towers. This seems to me an example of the sceptic's illegitimate manoeuvre of setting impossibly high standards to fulfil and then complaining that they cannot be met.

Do we ever have certain knowledge of the kind you require about the future? I would say very seldom, if ever. We may know on the basis of induction that the sun will rise tomorrow but such evidence is seldom available to predict the outcome of individual actions. Even where the future events are intended by us, fate can intervene to prevent what we intend occurring and, where the future events are outside our control, knowledge is most unlikely to be available.

It is, however, a feature of the moral life that we have to decide how to act or not act in conditions of uncertainty, where knowledge of the future is simply not available and it is unrealistic to suppose it could be. In such circumstances we have to act on the basis of the best judgements about what is likely happen that we can; we have to ensure that we have good justification for our beliefs.

If we apply this more reasonable test to the pilot's actions, it seems very clear that the pilot could have good reason to believe that, unless he acts, thousands of innocents will die, on the basis of the intercepted conversations, the proximity of the plane to the Twin Towers etc. The ethical dilemma which you seek to evade is: if he does have such good reason, what should he do? The post-lapsarian world we live in neither guarantees us certainty about the future nor that ethical principles may never conflict, giving rise to appalling ethical dilemmas. But we still have to act or not act, either of which may have appalling consequences. (You do not evade moral responsibility by inaction). These are the ineluctable and central facts about our moral life that you seek to evade.







Evading moral dilemmas


Dear David,


I am replying to your latest letter in a form which means the whole correspondence can be put on the CCADD website.


On the substance of what you say: of course, it is undeniable that much of the time we do not have complete certainty about the situation we face, and have to act on what we do know, or on the calculations we can make from the evidence available. To take the Jim and Pedro example, nothing I have said makes it any easier for Jim to do the best thing in his circumstances. He still has to judge what to do, given that Pedro is probably an unreliable liar. My own view, which I have already expressed, is that, since Pedro cannot be trusted, Jim should not shoot Paulo - come what may. If this means that Pedro shoots him instead, so be it. Being properly human often entails being killed. That is the central truth of Christianity. But ‘humankind cannot bear very much reality’, so we try to evade this truth by making out that there is some less drastic ‘solution’, like killing the innocent as the best means to a good end. My argument, however, has simply been to show that morally speaking we are never obliged, that is, it can never truly be our duty, to do this. But this does not touch the other examples under discussion, such as the Sept 11th case, or the medical cases, since refusing to kill the innocent in these cases is not likely to result in our being killed ourselves.


Here, I still think that we must first ask a lot of serious questions, not in order to evade the duty of acting, but in order to find out as clearly as possible exactly what we are being asked to do, and on what grounds. Having done that, all we can do is act according to our conscience in the hope that if we do wrong, we shall be forgiven. I suppose my main point has been that it is a mistake to soften the tragic dilemmas we may have to face, by making out that sometimes intentionally killing the innocent is some sort of virtuous, or permissible act which it is our moral duty to perform. Of course, this whole discussion raises immense questions as to what ethics is about, how ethical truths are arrived at, and the relation of ethics to law. But let’s leave that to another day!


Yours ever,

Brian 1.3.02