How to Argue with an Anti-Utilitarian
Excerpted from Moral Thinking, Oxford, 1981, pp. 131-135

Your first move should be to ask him what level of thinking [sc. critical or intuitive] he is talking about; for on this will depend what moves in the game are permissible, and in particular what examples can properly be adduced. In any case, if he has never heard of the difference between levels, mention of it will put him off his stroke.

Briefly, if he is talking about the critical level, he is allowed to bring up any examples he pleases however fantastic; but at that level no appeals to received intuitions are allowed, because the function of critical thinking is to judge the acceptability of intuitions, and therefore it cannot without circularity invoke intuitions as premises. If, on the other hand, he is talking about the intuitive level, he is allowed to appeal to any intuitions he thinks the audience will agree with, pending their examination by critical thinking, but must be very careful what examples he uses. For his audience's intuitions are the product of their moral upbringings, and, however good these may have been, they were designed to prepare them to deal with moral situations which are likely to be encountered, there is no guarantee at all that they will be appropriate to unusual cases. Even in the unusual cases, no doubt, the usual moral feelings will be in evidence; but they provide no argument.

The dispute is likely to resolve itself, therefore, into one about the admissibility and the treatment of examples. To illustrate this, suppose that your opponent's case is the following: there are in a hospital two patients, one needing for survival a new heart and the other new kidneys, a down- and-out who is known to nobody and who happens to have the same tissue-type as both the patients strays in out of the cold. Ought they not to kill him, give his heart and kidneys to the patients, and thus save two lives at the expense of one? The utilitarian is supposed to have to say that they ought; the audience is supposed to say that they ought not, because it would be murder. Let us suppose that your opponent manages to get the audience to treat murder as a descriptive, not a secondarily evaluative, word, and thus to call the statement that it would be murder a purely factual one, which can be established prior to any judgment that it would be wrong. I have simplified the case a little from the way it has sometimes been presented in the literature, because the complications will not affect the moves we shall be making.

On this example you have to mount a two-pronged attack. If we are to do intuitive thinking, the matter is fairly simple. It is murder, and would therefore be wrong. A utilitarian does not have to dissent from this verdict at the intuitive level. If he has been well brought up (and in particular if he has been brought up by a sound critical utilitarian thinker) he will have that intuition, and it is a very good thing, from the utilitarian point of view, that he will have it. For just think what would be the consequences of a moral education which contained no prohibition on murder!

Your opponent will now object that although on the utilitarian view it is a good thing for people to have these intuitions or feelings, it also follows from that view that they ought to overcome and act contrary to them in cases like this, in which, ex hypothesis it is for the best to do so. Let us ask, then, whether the doctors in the hospital ought to do this if they are utilitarians. It will turn upon their estimate of the probability of hitting off the act which is for the best by so doing. The crucial words are, of course, 'ex hypothesi'; for your opponent has constructed his example with the express purpose of making the murder the act which will have the best consequences. You must not allow him simply to assume that this is so; he has to convince the audience, not just that it really could be so in a real-life situation, but that it could be known to be so by the doctors I with a high degree of probability. For utilitarianism, as a method of choosing the most rational action (the best bet for a utilitarian) in a moral dilemma of this sort, requires them to maximize the expectation of utility (i.e. preference-satisfaction); and since, if they get it wrong, the consequences will be pretty catastrophic, the doctors have to be very sure that they are not getting it wrong. There is perhaps no need to go into any technicalities of games-theory to establish this point, though a full account would need to contain a method of weighing combinations of probabilities with utilities against one another, by asking which combination one would prefer, after exposure to logic and the facts.

It is fairly obvious that this high degree of probability will not be forthcoming in many actual situations, if any at all. Have the doctors checked on the down-and-out's connexions or lack of them? (How? By consulting the police records, perhaps! But a colleague of my psychiatrist sister once wrote in his notes, about a dishevelled individual brought in off the streets very late at night by the police, 'Has delusion that he is a high-ranking civil servant', and it turned out that he was in fact a very high-ranking civil servant.) Have they absolute confidence in the discretion and support of all the nurses, porters, mortuarists, etc., who will know what has happened? Add to this the extreme unlikelihood of there being no other way of saving these patients, if they can be saved at all, and it will be evident that your opponent is not going to get much help out of this example, once it is insisted that it has to be fleshed out and given verisimilitude.

But perhaps that was not his intention. Perhaps, that is to say, he all along intended the example to be a dummy. Is he not allowed, after a brief introduction of the example to give its general shape, to skip the details and simply posit that it is a case where murder would be for the best? The audience, which is probably prejudiced against utilitarians anyway, will have no difficulty in imagining that the details could be filled in a way that would suit his case and damage yours. But you must not let him get away with this. If we are talking about intuitive thinking in a real-life sort of situation, the example needs to be a real-life sort of example.

If, on the other hand, he claims the right to introduce any logically possible example, then he is exposed to the other prong of your attack. For then he has put himself beyond the range of intuition and cannot appeal to it. Critical thinking can certainly deal with such cases, and will give a utilitarian answer. If he tailors the case so that the utilitarian answer is that murder is the right solution, then that is the answer he will get. What you have to say to the audience is that this does not in the least matter, because such cases are not going to occur. This has two important consequences for the argument. The first is that allowing that in such a case murder would be justified commits us to no prescription to murder in the actual world, in which we have to live our moral lives. The second is a generalization of the first: the prima facie principles which the critical thinker will select for use in this world can also, and will, include a ban on murder, because for the selection of these principles this peculiar case, since it will not occur, is irrelevant.

Your opponent may say 'Are there not some cases occurring in real life, albeit rarely, in which murder is justified on utilitarian grounds?' To which you should reply that he has not produced any, but that if he really did find one, we should have to do some critical thinking on it because it would be clearly so unusual as to be beyond the range of our intuitions. If we then found that murder really was justified in that case, we still should not have shown that the rational moral agent would commit the murder; for he would be unlikely to have sufficient evidential grounds for saying that it was the right act. But, giving your opponent everything that he asks for, if he did actually have sufficient evidence (a very unlikely contingency), murder would in that case be justified; though even then the agent in question, if he had been well brought up, might not do it, because it would go so much against all his moral feelings, which in a good man are powerful. So, owing to being a good man, he might fail to do the right act. If he did bring himself to do it, it would haunt him for the rest of his life. But until your opponent produces actual cases, you should not let yourself be troubled overmuch with fictional ones. If the actual cases are produced, you will probably find that the critical discussion of them will leave you and the audience at one, provided that the discussion is serious.

Utilitarian Philosophers :: R. M. Hare :: 'How to Argue with an Anti-Utilitarian'