COULD KANT HAVE BEEN A UTILITARIAN?

RM Hare

[An extract from Sorting out ethics, ©1997 RM Hare, ISBN 0-19-823727-8 Published by Oxford University Press.]

 

… the supreme end, the happiness of all mankind

 

The law concerning punishment is a Categorical Imperative; and woe to him who rummages around in the winding paths of a theory of happiness, looking for some advantage to be gained by releasing the criminal from punishment or by reducing the amount of it …

(Rl A196 = B226 = 331)

8.1. MY aim in this chapter is to ask a question, not to answer it. To answer it with confidence would require more concentrated study of Kant’s text than I have yet had time for. I have read his main ethical works, and formed some tentative conclusions which I shall diffidently state. I have also read some of his English-speaking disciples and would-be disciples, but not, I must admit, any of his German expositors except Leonard Nelson. But my purpose in raising the question is to enlist the help of others in answering it.

To many the answer will seem obvious: for it is an accepted dogma that Kant and the utilitarians stand at opposite poles of moral philosophy. This idea has been the current orthodoxy at least since, in the early twentieth century, Prichard and Ross, deontologists themselves, thought they had found a father in Kant. John Rawls, in turn, has been deeply influenced by these intuitionist philosophers, and does not think it necessary to document very fully the Kantian parentage of their views. As a result, the story that Kant and utilitarians have to be at odds is now regularly told to all beginner students of moral philosophy.

But is it true? My own hesitant answer would be that it is not. The position is more complicated. Kant, I shall argue, could have been a utilitarian, though he was not. His formal theory can certainly be interpreted in a way that allows him – perhaps even requires him – to be one kind of utilitarian. To that extent what J. S. Mill says about the consistency of his own views with Kant’s Categorical Imperative is well founded (1861: ch. 5, middle). But Kant’s rigorous puritanical upbringing had imbued him with some moral views which no utilitarian – indeed, which few modern thinkers of any persuasion – would be likely to endorse: about capital punishment, for example, and about suicide, and even about lying. These rigoristic views he does his best (unsuccessfully in the view of most expositors) to justify by appeal to his theory.

I shall be looking at some of these arguments. To deontologists who seek to shelter under Kant’s wing they give small comfort; for if his theory is consistent with one kind of utilitarianism (what kind, I shall be explaining), it does not do them much good if some of his arguments which most people would now reject are anti-utilitarian in tendency. Kant was, indeed, a deontologist, in the sense that he assigned a primary place to duty in his account of moral thinking. But he was not an intuitionist of the stamp of Prichard and Ross. He did not believe, with Prichard, that ‘If we do doubt whether there is really an obligation to originate A in a situation B, the remedy lies not in any process of general thinking but in getting face to face with a particular instance of the situation B, and then directly appreciating the obligation to originate A in that situation’ (1912: s.f.). Kant would have called this ‘fumbling about with the aid of examples’ (Tappen vermittelst der Beispiele, Gr BA36 = 412).

On the contrary, though in the Groundwork he respects what he calls ‘ordinary rational knowledge of morality’, and throughout his writings is happy when common moral convictions support his views, the title of the first chapter shows that he is engaged in a ‘transition’ from this to ‘philosophical knowledge’. The second chapter is called, likewise, ‘Transition from Popular Moral Philosophy to Metaphysic of Morals’. Kant would not have been content, as Prichard was and as many of our contemporaries are, and as Rawls almost is, to rely on our ordinary moral convictions as data, even after reflecting on them. Instead, he developed a highly complex and sophisticated account of moral reasoning: the ‘Metaphysic of Morals’.

In this he was right. Moral philosophy, which Prichard thought rested on a mistake (1912: title), began when Socrates and Plato, faced with a collapse of popular morality because of the inability of its adherents to provide reasons for thinking as they did, set out in the search for these reasons. Kant is in this tradition; Prichard and Ross are not, and Rawls, in some respects their follower, is half in and half out of it. He is only half a rationalist, and half an intuitionist, in that he relies on intuitions altogether too much (H 1973a). This chapter is the beginning of an attempt to rescue Kant from some of his modern ‘disciples’.

8.2. I want first to draw attention to some passages in the Groundwork which bear on my question. I will start with the famous passage, beloved of anti-utilitarians, about treating humanity as an end. In full it runs: ‘Act in such a way that you always treat humanity, whether in your own person or in the person of any other, never simply as a means, but always at the same time as an end’ (Gr BA66 f. = 429). To understand this we have to know what Kant means by ‘treat as an end’. He gives us some important clues to this in the succeeding passage, but unfortunately he seems to be using the expression in at least two different senses. Broadly speaking, the first and third of his examples, those concerned with duties to oneself, are inconsistent with a utilitarian interpretation, but the second and fourth, those concerned with duties to others, are consistent with it. As we shall see, this difference is no accident.

I will take the second and fourth examples first. The second concerns false promises. He combines this with similar examples about ‘attempts on the freedom and property of others’. The fault in all such acts lies, he says, in ‘intending to make use of another man merely as means to an end he does not share (in sich enthalte). For the man whom I seek to use for my own purposes by such a promise cannot possibly agree with my way of behaving to him, and so cannot himself share the end of the action’. Other people ‘ought always at the same time to be treated as ends – that is, only as beings who must themselves be able to share in the end of the very same action’.

The fourth example I will quote in full:

Fourth, as regards meritorious duties to others, the natural end which all men seek is their own happiness. Now humanity could no doubt subsist, if everybody contributed nothing to the happiness of others but at the same time refrained from deliberately impairing their happiness. This is, however, merely to agree negatively and not positively with humanity as an end in itself unless every one endeavours also, so far as in him lies, to further the ends of others. For the ends of a subject who is an end in himself must, if this conception is to have its full effect in me, be also, as far as possible, my ends.

I interpret this as meaning that, in order to fulfil this version of the Categorical Imperative, I have to treat other people’s ends (i.e. what they will for its own sake) as my ends. They must be able to do the same, i.e. share the end. In the Tugendlehre Kant explains the relation between an end and the will as follows: ‘An end is an object of the power of choice (Willkür) (of a rational being), through the thought of which choice is determined to an action to produce this object’ (Tgl A4 = 381). We shall be examining later the distinction between ‘Wille’ and ‘Willkür’, and the alleged distinction between will and desire. On this, see esp. Tgl A 49 = 407, where Wille is both distinguished from Willkür, and identified with a kind of desire: ‘nicht der Willkür, sondern des Willens, der ein mit der Regel, die er annimmt, zugleich allgemeingesetzgebendes Begehrungsvermögen ist, und eine solche allein kann zur Tugend gezählt werden’ (‘not a quality of the power of choice, but of the will which is one with the rule it adopts and which is also the appetitive power as it gives universal law. Only such an aptitude can be called virtue’).

Elsewhere Kant qualifies this explanation of what it is to treat others as ends, by saying that the ends of others which we are to treat as our own ends have to be not immoral (Tgl A119 = 450: ‘die Pflicht, anderer ihre Zwecke (so fern diese nur nicht unsittlich sind) zu den meinen zu machen)’. Some utilitarians, for example Harsanyi, take a similar line and rule out immoral or anti-social ends from consideration (1998c: 96). I am tempted to say, in the light of the similarity between the views of these utilitarians and Kant, and of the passages we have been discussing, that he was a sort of utilitarian, namely a rational-will utilitarian. For a utilitarian too can prescribe that we should do what will conduce to satisfying people’s rational preferences or wills-for-ends – ends of which happiness is the sum.

We may notice in passing that this same passage in Kant (Gr BA69 = 430) provides an answer to self-styled Kantians who use what has been one of their favourite objections to utilitarianism, that utilitarians do not ‘take seriously the distinction between persons’ (Rawls 1971: 27; see Mackie and Hare in H 1984g: 106, Richards and Hare in H 1988c: 256). It is hard to understand precisely what the objection is. Clearly utilitarians are as aware as anybody else that different and distinct persons are involved in most situations about which we have to make moral judgements. Probably what the objectors are attacking is the idea that we have, when making a moral decision about a situation, to treat the interests, ends, or preferences of different people affected by our actions as of equal importance, strength for strength. This is the same as to show equal concern and respect for all (another slogan of the objectors, which seems inconsistent with the one we are considering). In other words, I am to treat the interests of the others on a par with my own. This, according to utilitarians, is what is involved in being fair to all those affected. It is to obey Bentham’s injunction ‘Everybody to count for one, nobody for more than one’ (ap. Mill 1861: last chapter). And if we treat equal preferences as of equal weight, utilitarianism is the result.

But that is precisely what Kant is telling us to do in this passage, as Mill observes (ibid.). For if I make the ends of others my ends, I shall, in adjudicating between them when they conflict, treat them in the same way as I would my own ends. In so doing I am not failing to distinguish between different people, but, as justice demands, giving equal weight to their and my equal interests (the ends which they and I seek with equal strength of will), just, as I give equal weight to my own equal interests. So, if the objection did undermine utilitarianism, it would undermine Kant too.

8.3. But now we have to turn to Kant’s first and third examples. In the first, he is against suicide because it involves ‘making use of a person merely as a means to maintain a tolerable state of affairs till the end of his life’. But this is not the same sense of ‘use as a means’ as that which contrasts with ‘treat as an end’ in the second and fourth examples. I might have as an end the saving myself from intolerable pain. Obviously there is no difficulty in my sharing this end with myself, or agreeing with my way of behaving to myself. Kant must therefore be here using ‘use as means’ and ‘treat as an end’ in some different sense. I shall not here investigate what it is; but it seems to be something like ‘regard (or not regard) a human being (myself) as at my own disposal to do what I like with for my own purposes’.

But this objection to suicide, if valid at, all, is different from those to promise-breaking and non-beneficence. To treat myself as at my own disposal is not to frustrate the ends that I will. Perhaps Kant is here harking back to something he heard when young, that man is created as a human being to fulfil an end ordained by God, and therefore ought not to act contrary to God’s will by not fulfilling God’s ends. But to argue thus would be to follow a principle of heteronomy such as he later rejects (Gr BA92 = 443). It cannot be turned into an autonomous principle by simply substituting ‘myself’ for ‘God’. For if it is not God’s will but my will that is in command, then it can, within a consistent set of ends, choose suicide in these special circumstances.

The same could be said about the third example concerning the cultivation of one’s talents. For a full statement of the example we have to refer back to Gr BA55 = 423. I shall discuss this earlier use of the example shortly. Here it is to be noted that Kant speaks of ‘nature’s purpose for humanity in our person’ (Gr BA69 = 430), thus again betraying the theological and heteronomous source of his argument here. A person could certainly with consistency will as his end (whatever nature intended) to live like the South Sea Islanders of whom Kant has earlier spoken slightingly; and he could certainly share this end with himself, and agree to it. So the sense of ‘treat as an end’ used in the second and fourth examples would provide no argument at all against his ‘devoting his life solely to idleness, indulgence, procreation, and in a word, to enjoyment’ (Gr BA55 = 423). In the sense used in the second and fourth examples, treating humanity in myself as an end would not preclude my lotus-eating, any more than it would preclude suicide.

I should like to mention here that in my own adaptation of the Kantian form of argument in FR ch. 8 I specifically excluded from its scope personal ideals not affecting other people, and said that about these one could not argue in this way. So my view on these first and third examples of Kant is that he is going astray through trying (in order to buttress his inbred convictions) to use arguments from universalizability outside their proper field, which is duties to other people.

There is a possible objection to the assimilation of wills to preferences that I have just made: that a preference, being something empirical, is not the same as a will, which is, in the pure Kantian doctrine, something noumenal (cf. KpV A74 f. = 43). To this objection I shall return (8.8).

8.4. But now we must turn to another famous passage, the formulation of the Categorical Imperative which runs: ‘Act only on that maxim through which you can at the same time will that it should become a universal law’ (Gr BA52 = 421).

This version too is consistent with utilitarianism. If we are going to will the maxim of our action to be a universal law, it must be, to use the jargon, universalizable. I have, that is, to will it not only for the present situation, in which I occupy the role that I do, but also for all situations resembling this in their universal properties, including those in which I occupy all the other possible roles. But I cannot will this unless I am willing to undergo what I should suffer in all those roles, and of course also get the good things that I should enjoy in others of the roles. The upshot is that I shall be able to will only such maxims as do the best, all in all, impartially, for all those affected by my action. And this, again, is utilitarianism. To link it up with the other formula about treating people as ends: if I am to universalize my maxim, it must be consistent with seeking the ends of all the other people on equal terms with my own.

This formulation of the Categorical Imperative is followed by another rather similar one: ‘Act as if the maxim of your action were to become through your will a universal law of nature’ (Gr BA 52 = 421). After this, Kant illustrates these two formulations with the same examples as we have been discussing in connection with the ‘humanity as an end’ formulation. Here again the promise-keeping and beneficence examples fit well with a utilitarian interpretation, but the suicide and cultivation-of-talents examples do not. In the promising case, he uses a form of argument usually now called by English-speaking writers utilitarian generalization; he asks ‘How would things stand if my maxim became a universal law?’, and answers that promises would become ’empty shams’. This is not a strong argument, because one might will as a universal law that people should break promises in precisely one’s own present situation, when one can get away with it and the institution of promising would survive. (Recent work on the difficulty of drawing a line between act- and rule-utilitarianism is relevant here; cf. FR 130 ff., Lyons 1965: ch. 3). The argument against promise-breaking we considered earlier, which says that the victim cannot share the end of the promise-breaker, is much stronger, and is similar to one I would myself, as a utilitarian, rely on (H 1964d: s.f.).

Kant’s argument here against non-beneficence comes to much the same as the one I discussed earlier, and one which I should myself, as a utilitarian, employ, and I have no time to analyse it further. The argument against suicide is again very weak. I could certainly without contradiction will universally that those who would otherwise have to endure intolerable pain should kill themselves. This could indeed become a universal law of nature, and I could act as if it were to become so through my will. Kant thinks it is a good argument only because he thinks (perhaps owing to his rigorist upbringing) that maxims have to be very simple. If we have a choice between the simple maxims ‘Always preserve human life’ and ‘Destroy human life whenever you please’, we shall probably opt for the former. But there are many less simple maxims in between these extremes which most of us would will in preference to either of them: for example ‘Preserve people’s lives when that is in their interests’ (and perhaps we would wish to add other qualifications). As we have seen (8.1) moral principles do not have to be as simple and general as Kant seems to have thought, and they can still be universal all the same (H 1972a, 1994b).

As regards cultivation of talents, Kant is also on shaky ground. It is perfectly possible to will that, those who are in the fortunate position of being able to live like the South Sea Islanders should do so; and this could become a law of nature if nature were as benign everywhere as it is said to be in Tahiti. The best argument, against lotus-eating is a utilitarian one, which Kant does not use though he could have; namely that one person’s indolence may, in the actual state of nature, harm others whom he might be helping if more industrious, and who therefore cannot share his ends.

8.5. The score at this point is that Kant’s theory, in the formulations of the Categorical Imperative we have considered, is compatible with utilitarianism, and so are some arguments that he uses, or could have used consistently with the theory, in some of his examples. By any reckoning the first example (suicide) is the only one that cannot be handled in a utilitarian way in accordance with the Categorical Imperative in these three formulations, although Kant himself does handle both this and the third example in a non-utilitarian way. So, as I said at the beginning, Kant could have been a utilitarian, in the sense that his theory is compatible with utilitarianism, but in some of his practical moral judgements his inbred rigorism leads him into bad arguments which his theory will not really support. I do not think that this score ought to give much comfort to modern anti-utilitarians who usurp Kant’s authority.

It does, however, emerge from his discussion of the examples in the Groundwork that there is a tension in Kant’s thought between utilitarian and non-utilitarian elements. How this tension is to be resolved becomes a little clearer in the Doctrine of Virtue. There, a main division is made between duties to oneself and duties to others. This distinction and other related ones are laid out, in Tgl A34 = 397, in the top half of a table headed ‘The Material Element of Duty of Virtue’. ‘My own end, which is also my duty’ is said to be ‘my own perfection’; and ‘the end of others, the promotion of which is also my duty’ is said to be ‘the happiness of others’.

The immediate impression we get from this is that there is a utilitarian part. of Kant’s theory, and a non-utilitarian part. The utilitarian part prescribes duties to others, and these are compatible with utilitarianism (qualified by the requirement, as above, that we have to advance others’ ends only in so far as they are consistent with morality). But the other part (duties to oneself) seems to be not utilitarian at all, but perfectionist. However, these impressions are too superficial. This becomes apparent if (taking a hint from what he says against perfectionism in Gr BA92 = 443) we ask, first, in what the perfection is supposed to consist; and secondly, what ‘consistent with morality’ is to mean. As we answer these questions we shall see that the tension between the utilitarian and non-utilitarian elements in Kant’s theory begins to ease.

Obviously the perfection that Kant is after is moral perfection. It consists in the acquisition of virtue. Part of this virtue will clearly consist in the disposition to fulfil the duties to others laid down on the utilitarian side of the table. But what is the other part? That is, what content does moral perfection have, for Kant, over and above the utilitarian content consisting in practical love for other people. (For the notion of ‘practical love’ see Gr BA13 = 399 and Tgl A118 f. = 448 f.). It begins to look as if moral perfection, if it sought anything beyond this practical love, would be chasing its own tail. As he says in Gr BA92 = 443, ‘[the ontological concept of perfection] shows an inevitable tendency to go round in a circle and is unable to avoid presupposing the morality it has to explain’. There would be nothing else in the duty to make ourselves perfect, except the duty to make ourselves disposed to make ourselves perfect. It would still not have been determined what the perfection, or the performance of the duty to promote it, would consist, in.

But we must be careful here to distinguish between form and content. It, could be that Kant’s view is this: the perfection we are after is one of form, not of content. To explain this: a morally perfect character, or good will, as he sees it, is one formed by its own framing of universal laws in accordance with the Categorical Imperative. In seeking moral perfection, we are seeking to make our wills good in this sense. If this is what Kant means, then the utilitarian and the non-utilitarian part of his morality at once come together again. For a will that wills universally must, as we have seen, be a will that treats the ends of other people’s wills on equal terms with its own ends; and this is another way of expressing the practical love that we have already found to be required by our duties to others. In other words, the moral perfection of a good will is a perfection of form, and the form is the form of practical love, which is utilitarian, in that it seeks to advance the ends of all impartially. The ‘material element’, referred to in the title of the table, all comes either directly or indirectly from this source.

The same happens when we ask what it means to say that the ends of others which we seek impartially to advance have to be consistent with morality. Here we have to look in passing at what Kant says later in the Groundwork about the Kingdom (or Realm) of Ends. A good will has to be one that can be a lawmaking member of such a realm (Gr BA77-9 = 435 f.). This is Kant’s way of ensuring that the moralities of all rational beings will be consistent with one another. The lawmakers in the Realm of Ends will legislate unanimously, because each is constrained by the universal form of the legislation.

The effect of this is that the ends of others, which we have a duty to advance impartially, are those only which are moral, i.e. which they would retain if they were legislating universally, or forming universal maxims in accordance with the earlier formulations of the Categorical Imperative. But if these maxims, as they must, express practical love, they too will be consistent with utilitarianism. For utilitarianism is, simply, the morality which seeks the ends of all in so far as all can seek them consistently in accordance with universal maxims. If a utilitarian tried to promote ends which were not consistent with such a morality, he would run up against the obstacle that the ends he was promoting would be such as others could not ‘share’, as Kant puts it (see above); and so his entire moral system would come apart. It is part of the requirements for a consistent utilitarian morality that it should be able to be shared by all.

We thus see that even the apparently non-utilitarian part of Kant’s doctrine of virtue, and of his entire system, turns into utilitarianism at one remove. It does so because even the apparently non-utilitarian virtue of perfection requires aspirants to it to perfect themselves in practical love.

8.6. The objection might be made that, whereas for Kant human perfection is an end in itself, for the utilitarian it is a mediate end, the ultimate end being the furtherance of the ends of all. This objection is analogous to one which has been made against my own theory, that by dividing moral thinking into two levels I have demoted our ordinary intuitive convictions and prima facie principles into a merely instrumental role. For me, it is said, the real moral thinking takes place at the critical level and is utilitarian; what goes on at the intuitive level is only a means to help us fulfil, maximally and on the whole, our utilitarian duties as determined by critical thinking. We are to make ourselves into good people, and fulfil our duties, not for its own sake but because that will conduce to the greatest good. It is further alleged (e.g. by Bernard Williams, 1988: 189 ff.) that if we took such an attitude to our common moral convictions, they would soon ‘erode’; if they are to retain their force for us, we have to treat them as ultimate.

It has always seemed to me that this objection, whether to my own theory or to Kant as I have interpreted him, will not be sustained by anyone who has experience even of trying to live a morally good life. It is perfectly possible at the intuitive level to treat moral duty or virtue as ultimate and give them the ‘reverence’ that Kant demands, while at the same time to recognize that to establish that those traits of character really do constitute virtue, and that those intuitive moral principles really are the ones we should observe, requires more thought than the mere intuition that this is so. I am sure that Kant would have agreed, although he makes his account of the relation between virtue and duty much more obscure by failing to clarify the distinction between levels of moral thinking (see below). It is in this sense that we should understand passages such as Tgl A32 = 396: ‘that virtue should be its own end and also, because of the merit it has among men, its own reward’, and Tgl A33 = 397: ‘the worth of virtue itself, as its own end, far exceeds the value of any utility and any empirical ends and advantages that virtue may, after all, bring about.’

8.7. Why is the suggestion that Kant could have been a utilitarian thought so bizarre? It has been held that he could not have been for, in the main, two inadequate reasons. The first is that he often stresses that the Groundwork of the Metaphysic of Morals, as he calls his book, cannot appeal to anything contingent and empirical; and desires and preferences are of this sort. But here we have to be very careful to distinguish, as Kant insists on our doing, between the empirical and the rational parts of moral philosophy. He certainly thinks that it has both these parts. He says, about those who fail to distinguish the two roles, ‘What (such a procedure) turns out is a disgusting hotch-potch (Mischmasch) of second-hand observations and semi-rational principles on which the empty-headed regale themselves, because this is something that can be used in the chit-chat of daily life. Men of insight, on the other hand, feel confused by it and avert their eyes with a dissatisfaction which, however, they are unable to cure’ (Gr BA31 = 409, cf. BAiv = 388).

The important point to get hold of is that his strictures on bringing in empirical considerations apply only to what he is doing in this book: only, that is, to the Metaphysic of Morals, and indeed only to its Groundwork. I think it is legitimate to regard the Groundwork as a purely logical enquiry into the nature of moral reasoning, and as such it of course must not contain appeals to empirical facts, any more than any other kind of logic. This is the chief thing, as I said, that distinguishes Kant from some of his modern self-styled disciples.

Let us then look at the Kantian programme, or at this interpretation of it, in more detail. It rests on a metaphysical or logical enquiry into the nature of the moral concepts. This has to be the basis of any system of moral reasoning. We have to do it by considering the nature of the concepts only, not anything empirical. Kant believed in the synthetic a priori, and indeed calls his Categorical Imperative the ‘practical synthetic a priori’ (Gr BA50 = 420). But he explains later that the question how such a synthetic a priori proposition is possible and necessary lies outside the bounds of a metaphysic of morals (Gr BA95 = 440). The first two chapters of the Groundwork (those we have been concerned with), are ‘merely analytic’ (Gr BA96 = 445); he has been ‘developing the concept of morality as generally in vogue’. At any rate be would, I am sure, have rightly excluded from this part of his enquiry any empirical data, whether about what actually goes on in people’s minds or about anything else, including any antecedently held substantial moral judgements; for the only source of these could be something that goes on in people’s minds, that is, intuitions. That we have a certain intuition is an empirical fact, and as such is excluded from this part of the enquiry, for the same reason as desires that we contingently have are excluded. Kant explicitly rejects moral sense theories (Gr BA91 f. = 442), and would equally have rejected intuitionism of the sort expressed in the quotation from Prichard that I gave earlier. Ordinary people understand, indeed, the concepts of morality, but this is no moral sense apprehending the substance of morality.

8.8. The elements of Kant’s metaphysic of morals that I find most central are its reliance on the pure will, and its insistence that in moral reasoning we have to will universally. What does ‘pure’ mean, and what does ‘reliance’ mean? To understand this we have to consider Kant’s doctrine of the autonomy of the will. This, he says, is ‘the property the will has of being a law to itself (independently of every property belonging to the objects of volition)’ (Gr BA87 = 440).

Here it is very easy to go astray in one’s interpretation of Kant, and attribute to him a nonsense. One way of taking this doctrine would be to say that to be autonomous the will has to have no regard to what in particular it is willing. So, for example, when I am deciding whether to will to tell an untruth, I have to have no regard to the property of this proposed object of my volition, namely that what I should be saying would be untrue. Or, if I am contemplating killing someone, I am not to pay attention to the property of my action that it would consist in bringing about his death. I cannot believe that this is what Kant meant, because he certainly thought it relevant to the morality of actions that they were lies or murders.

What then did he mean? I think that what he meant was this. Our will is initially free to will whatever we will. We are not constrained to will this or that because of what this or that is. The will, is constrained only by what Kant calls ‘the fitness of its maxims for its own making of universal law’ (Gr BA88 = 441). This is what is implied in the ‘autonomy’ formulation of the Categorical Imperative. That is, it is only the universal form of what we are going to will that constrains us, and not any content. The content gets put in by the will itself. The will can accept only such contents or objects of its volition as can be willed universally. This is the same doctrine as I have myself expressed by saying that moral judgements have to be universal prescriptions.

So interpreted, the doctrine of autonomy would exclude as heteronomous many of the principles advocated by some modern so-called Kantians: for they do seek to constrain the will not just formally but substantially by saying that it has to have certain objects. Such intuitionists not only appeal, though they do not call it that, to something empirical, namely the contingent fact that we have certain intuitions or convictions, but seek to constrain the will and bind it to the substantial content of these convictions. This is most un-Kantian.

Returning, then, to the objection we are considering to calling Kant a utilitarian: the objection says that this cannot be so, because utilitarians appeal to desires or preferences, which are something empirical, and therefore excluded by Kant. To this the answer is first, that they are excluded only from the formal part of his enquiry, but have to be admitted into any application to concrete situations of the form of moral reasoning which the enquiry generates; and secondly, that there is nothing to prevent a utilitarian from dividing up his enquiry in the same Kantian way, as for clarity he should, and as I do myself. A utilitarian system also has a pure formal part, which (in my view) needs to rely only on the logical properties of the moral concepts. It operates, indeed, with the concept of preference (and whether this is a different concept from that of will needs further discussion); but it does not assume that preferences have any particular content. What people prefer is an empirical matter; it has to be ascertained once we start to apply our system of reasoning, but in order to set up the system we do not need to assume that people prefer one thing or another; that is, in setting up the system we look merely at the form of people’s preferences, not at their content.

It has to be asked whether Kant’s wills are any different in this respect. Gr BA64 = 427 would suggest that they are not: ‘Practical principles are formal if they abstract from all subjective ends’; and this is equally true of the ‘Principle of Utility’ in those utilitarians who have one, especially if it is expressed in terms of the formal notion of preference-satisfaction. It is an empirical fact that a person wills this or that, just as it is an empirical fact that he prefers this or that. But the form of the will or preference can be the same whatever he wills or prefers, provided that for categorical or moral imperatives, as both the utilitarians and Kant can agree, the form is universal.

That, for both Kant and the utilitarians, is the only formal constraint on the will. However, for both there are material constraints, in the concrete situation in which we are doing the willing. Such constraints are, for example, that if I were to say what I am proposing to say, I should be speaking falsely, or that if I were to pull the trigger I should be killing someone. I have to be able to will this universally for all similar cases, and this constrains me because of the empirical fact that in that situation the person I should be lying to does not want, or will, to be deceived (as Kant might put it, he and I cannot ‘share’ the will that he should be), and the person I should be killing does not want, or will, to be killed. Given that this is the will or preference of the other party, I am constrained by this, and by the form of the reasoning, to treat him as an end by making what he wills my end, or in other words to treat his preference as if it were my own. Otherwise I shall not be able to universalize my maxim.

It may be objected that for Kant the distinction between will and mere preference or desire is fundamental. To this there are three replies. The first is that for Kant there is an important distinction between the will which is ‘nothing but practical reason’ (Gr BA36 = 412) – i.e. the rational will – and the will that is the source of maxims whether good or bad, rational or irrational. He calls the latter ‘Willkür’ (sometimes translated ‘choice’). His Latin equivalent for this is liberum arbitrium, and it is the possession of this that gives us free will or autonomy. But this distinction is not much relevant to our present problem; for utilitarianism could easily be expressed in terms of rational will.

Secondly, when Kant draws, as he often does, a contrast between rational will and inclination (Neigung), it is often, though not always, selfish inclination that he has in mind. An example is Gr BA8 = 496. We are not to follow our desires in so far as they are desires for our own advantage; that would not be to treat others’ ends as our own ends. But of course a utilitarian could agree with this insistence that the desires that determine our moral judgement have to be universal and impartial.

Thirdly, Kant, though he makes a clear distinction between will and inclination (Neigung), does not in fact always distinguish desire (Begierde) in the relevant sense from will, though he does in Gr BA124 = 461. In more than one place he identifies them. In the preface and the introduction to the second Critique there are two definitions, one of the faculty of desire (Begehrungsvermögen), and the other of will, which are in almost identical terms (KpV A17 n. = 9 n., A29 = 15). Later in the same work he speaks of ‘the faculty of desire which is therefore called the will, or the pure will in so far as the pure understanding (which in such a case is called reason) is practical through the mere conception of a law’ (A96 = 55). From KU BAxxiii = 178 n. (different versions in different editions) and Rl ABl ff. = 211 ff., it looks as if Kant came to see that there are different things that could be called ‘desire’, ‘inclination’, etc. (as indeed there are). If so, it may be that what modern utilitarians call ‘preference’ might be excluded from his ban on the empirical, and assimilated more to his Willkür or, if rationally universalizing, to his Wille.

8.9. Once we have distinguished pure from applied ethics, this first objection to enrolling Kant as a kind of utilitarian collapses. But now we are able to deal with the second objection, that Kant cannot have been a consequentialist, but utilitarians have to be. Once consequentialism is properly formulated, it is hard to see how anyone, Kant included, could fail to be a consequentialist. The doctrine gets a bad name only because its opponents, through their own confusions, formulate it incorrectly (1.8, 7.8, H 1993c: 123, 1998b).

Let us confine ourselves for the present to moral judgements which are on, or about, acts: for these are the judgements about which consequentialists and anti-consequentialists are supposed to be disagreeing. To act is to make a difference to the course of events, and what the act is, is determined by what difference. To revert to my previous examples (hackneyed ones, I am afraid): if I am wondering whether to pull the trigger, the main morally relevant consideration is that, if I did, the man that my gun is pointing at would die. Killing, which is the morally wrong act, is causing death, that is, doing something which has death as a consequence. Similarly, what is wrong about lying is that it is causing someone else to be deceived (to hold a false opinion) by oneself saying something false. The intended consequence is what makes it wrong. It would not be lying if it were not intended to have this consequence.

I am not saying that all the consequences of acts are morally relevant. Nor does any utilitarian have to say this. Many will be irrelevant. Which are relevant depends on what moral principles apply to the situation (the relevant consequences are those which the principles forbid or require one to bring about). So what the anti-consequentialists ought to be saying is something that consequentialists who understand the issue can also say: that there are some, consequences which are morally relevant, and that we ought to bring about, or not bring about, those consequences regardless of the other consequences which are morally irrelevant. Thus I ought to speak the truth and so inform the other party of it, even though there will also be the consequence that I am disadvantaged thereby. It is still the intention to bring about the consequence that he is misinformed which makes telling a lie wrong. Kant could not have disagreed.

A further point of objection is related but slightly different. Some of the consequences of actions are intended and some not. When we are speaking of the ‘moral worth of the agent’, or wondering whether to blame him, it is of course relevant whether he intended the consequences or not. We can say, with Kant, that the only good thing without qualification is a good will (Gr BAl = 393), meaning that people are judged by their intentions and not by the actual consequences.

But let us for the present leave aside these post eventum judgements and consider the situation of someone who is trying to decide what to do. He is trying to decide what to do intentionally, i.e. what intention to form; for we cannot decide to do something unintentionally (if it were unintentional, we could not speak of our having decided to do it). When we are wondering what intention to form, the intentions that are the possible candidates are all intentions to bring about certain consequences: that is, to do certain actions or to make the course of events different in certain ways. So the will itself, which is being formed in this deliberative process, is a will to bring about certain consequences. They are what is willed – the objects of volition, as Kant calls them. So, although the only good thing without qualification is a good will, what makes it a good will is what is willed (autonomously, universally, rationally, and impartially), and that is the consequences that are intended. Clearly I have been able only to scratch the surface of my question. There are many further points of difficulty in interpreting Kant, that I have not had room to raise, let alone discuss. The limit of my ambition has been to get intuitionists, deontologists, and contractualists, who are so sure that Kant was on their side against utilitarianism, to look more carefully at his (admittedly obscure) text. I am confident that, like me, they will at least find many utilitarian elements in it.