4. UTILITARIANISM AND THE HARM OF KILLING
In this chapter I shall conclude my discussion of the implications of utilitarianism for the question of the moral standing of animals, by considering what a utilitarian should say about the value of animal life.
Dying, killing, and harming
I argued in Chapter 3 that utilitarians are committed to extending the principle of equal consideration of interests to animals, and that this would then imply that it is morally wrong to cause an animal to suffer, except in unusual circumstances. Some utilitarians, including Singer, have thought that the principle of equal consideration applies to animals very differently when it comes to the question of killing them. Some have argued, indeed, that there is no moral objection to killing an animal, provided that the death is unexpected and painless. We then get a moral position that does not entail moral vegetarianism, while it would rule out hunting and factory farming. Alternatively, some have argued that, while there are moral objections to killing animals, the value of animal life is much lower than that of persons. So although it is wrong to kill an animal for no reason, much less reason is required than is necessary to justify the killing of a person. The main arguments for these views will be considered in later sections. I shall begin by drawing some preliminary distinctions.
It is important to keep three different questions, that are often run together, separate from one another. The first is whether death is harmful to the one who dies, and if so, in what respect it is harmful. There is an ancient puzzle about this. For there is a problem, both about the subject who is harmed by death, and about the time at which the harm occurs. Before a person dies, there is presumably no harm, since death has not yet occurred. But as soon as the moment of death arrives, there is no longer anyone in existence to be harmed. (Since this book is written from a secular perspective, I shall assume throughout that death is the end of existence, for both persons and animals.) Many have therefore concluded that death is not, in itself, a harm at all, and that we suffer no evil by dying.
The second question is whether we have reason to be afraid of death. This is easily confused with the first, but is really quite distinct from it. Many who hold that death is not an evil believe that it follows from this that it is irrational to fear death. They argue that those who are afraid of death are mistakenly picturing the time after their death as a sort of positive, but empty, state - the state of existing in utter blankness. Whereas the reality is that those who are dead no longer exist at all. It is then argued that fear of death only arises because we mistakenly confuse the end of consciousness with an empty consciousness. But this is not so. Even if death is not a harm, it may still be rational to fear it. For our reasons for wanting to go on living are not that we wish to avoid the harm of death, but rather that continued life is a presupposition of most of our projects and desires. That I should not die first, is a necessary condition for satisfying almost any desire. (Exceptions would be desires for martyrdom and posthumous fame.) Therefore, in so far as I have desires for the future that require my continued existence, I shall also have reason to fear death. For in general, we have reason to fear anything that may prevent our desires from being satisfied.
The third question, that is closely related to, but distinct from, the other two, is why it is directly wrong to kill (ignoring side-effects, such as grief caused to loved ones and so on). If death is a harm, then this will receive an answer within utilitarianism - it is because killing causes harm. But even if death is not a harm, it may still be directly wrong to kill, at least from a contractualist perspective. This will be because killing infringes the agent’s autonomy - indeed, it is the ultimate infringement of autonomy. Since agents will generally have projects and desires that require their continued existence, they will not wish to be killed. In which case, killing them will infringe their right to pursue their projects without interference. If rational agents have reason to fear death, then it is obvious that rational contractors should agree not to kill one another, except in self-defence.
Our main question in this chapter, is whether the direct utilitarian objections to the killing of persons (ignoring side-effects) extend also to the killing of animals. I shall argue that they do. But first we must consider whether death is a harm, and if so in what respect.
The harm of death
The position of those who believe that death is not a harm may be summed up in the old adage ‘What you don’t know can’t hurt you’. On this view, since those who have died are no longer in existence to feel any deprivation, no harm has resulted to them from their deaths. The old adage is not strictly true, however. Something can harm me if it prevents me from enjoying things that I would otherwise have enjoyed, even if I never know and never feel the lack. Suppose that a rich uncle of whom I have never heard dies, leaving me millions in his will. But I never learn that this has happened, because a clever lawyer manages to defraud me of my inheritance. Here I may rightly be said to have been harmed by the lawyer’s action, even though I never feel the lack of the money. For there are many satisfactions that I would have enjoyed had the lawyer not interfered. Preventing someone’s satisfaction is just as much a way of harming them as causing them dissatisfaction.
In the light of this, it is obvious that there is one respect in which death is, normally, a harm to the one who dies. (Of course there are rare circumstances in which life is so awful that continued existence is no blessing - in which the person would be, as we say, better off dead. Recall the example of Anthony, the author, from Chapter 1.) For it will generally be true that if the individual had not died, then they would have continued to enjoy a satisfying existence. Death is then a harm, not because it causes us positive deprivation - not because it causes us felt dissatisfaction - but because it prevents us from enjoying satisfactions that we would otherwise have had. Death is a harm, not because of what it is, but for what it does - it cuts off future worthwhile existence.
Some have argued that there is quite another sense in which death is a harm to the one who dies, namely, that it causes many of that person’s desires to be objectively dissatisfied. But this will take some explaining. We first of all need to distinguish between objective and subjective satisfactions of desire. A desire is objectively satisfied if the event desired does in fact take place, whether the person knows of it or not. A desire is subjectively satisfied, in contrast, if the person comes to believe that the event desired has taken place, whether it really has or not. Suppose that I want the Washington Redskins to win the Superbowl, for example. Then imagine two scenarios. In the first, the Redskins really do win, but, as a result of some sort of misinformation, I come to believe that they have not. Then my desire is objectively, but not subjectively, satisfied. In the second scenario, the Redskins fail to win, but I somehow come to believe that they have. Then my desire has been subjectively, but not objectively, satisfied.
It might be claimed, against this distinction, that every desire aims at its own subjective satisfaction - in which case it is impossible for a desire to be objectively satisfied without being subjectively satisfied. So it might be said that when I want the Redskins to win the Superbowl, what I really want is the felt satisfaction of learning that they win. But this is plainly false, for two reasons. First, when I am pleased that the Redskins have won, I am pleased because I have got what I wanted. The pleasure is a result of my desire being satisfied, not what the desire was really a desire for. Felt satisfaction is a normal concomitant of the knowledge that a desire has been objectively satisfied, rather than being what every desire really aims at. The truth is certainly not that I wanted the Redskins to win because I guessed that their winning would please me! Second, many of our desires in any case aim at things that we know we shall never see happen. For example, many of us have desires for things about which we are quite clear-headed that they will only be realised after our deaths, such as the desire that our grand-children should live to a happy old age. Plainly we are not in this case wanting to see our grand-children live to a happy old age (though we may have that - impossible - desire as well).
With the distinction between objective and subjective satisfactions of desire in place, Nagel’s thesis can now be stated. It is that both objective and subjective frustrations of desire are a species of harm. Then death is a harm to the one who dies because all of that person’s desires for the future that involve their own continued existence are thereby objectively frustrated. For example, if I want to be rich and want to be famous, but die before I can become either, then those desires will have been objectively frustrated. Of course I will not feel any dissatisfaction, since I shall no longer exist. But it will be true that the things I wanted did not in fact take place, because death prevented them from happening. In which case death has harmed me, if objective frustrations of desire count as harms.
Is Nagel correct to claim, however, that objectively frustrated desires are a species of harm? The question is an important one for us, on the assumption that animals have many fewer, if any, desires for the future (an assumption we shall examine in Chapter 6). For then death will be much less of a harm for them, if a large part of the harm of death for humans consists in the objective frustration of our forward-looking, long-term, desires.
In order to test Nagel’s thesis, let us consider an example in some detail. Suppose that Kurt is married to Philippa, and wants very much that Philippa should be faithful to him. Philippa, however, has other ideas, and carries on a love affair with another man without Kurt discovering. Suppose that nothing in Philippa’s relationship with Kurt ever suffers as a result - so far as Kurt is concerned, things are just as they would have been if Philippa had in fact been faithful to him throughout. Is Kurt harmed by Philippa’s infidelity, merely because his desire that she should be faithful to him is objectively frustrated, and despite the fact it is subjectively satisfied? I do not believe that he is.
I admit to feeling a certain pull in the contrary direction. But I think that this can be explained. For we can distinguish at least two senses in which everyone, including utilitarians, might agree that what Philippa does is bad (quite apart from any question of breach of contract), but which neither of them amount to any misfortune for Kurt. First, we might allow that what Philippa does is bad, in the sense that she has taken a real risk of harming Kurt. For no matter how careful she may be, there is always a chance that Kurt will find out. Second, there is a sense in which it is bad for someone to be exposed to risk. Something bad has surely happened to an atomic power worker who becomes contaminated by radiation, for example, because of the risk that this will cause serious disease to develop in later life. But consider the situation after the fact. If the power worker lives to a ripe old age and dies of a heart attack, then no harm was in fact done by the exposure to radiation. Similarly, imagine yourself reviewing Kurt’s life shortly after his eventual death. He remained happily married to Philippa throughout, and never in fact discovered her infidelity. Then surely he was not harmed either. Although something he wanted not to happen did in fact happen, he was not harmed by it. This is because he never knew of it, and because (in the light of our earlier discussion of the fraudulent lawyer), it did not prevent him from enjoying satisfactions that he would otherwise have enjoyed.
For some people, the intuition that Kurt is harmed by Philippa’s action may survive the points made above. As someone might put it, ‘The harm done to Kurt is that his desire was for the real thing and what he got was fraudulent.’ But I think that this intuition derives from a wholly different perspective on ethics (namely, contractualism), and that there is no way in which it can be available to a utilitarian. I shall explain this briefly now, returning to the point, from a slightly different angle, in a later section.
It is true, of course, that what people generally want is the real thing, not a plausible substitute. (When Kurt wants Philippa to be faithful to him he wants just that - that she should be faithful - not that he should continue to believe that she has been faithful.) For this reason rational contractors cannot agree to principles that would prohibit them from subjectively frustrating the desires of others in certain circumstances, but allowing them objectively to frustrate those desires, even provided that there is no danger that the person in question should find out. For example, it might plausibly be held that marriage (or at least a certain kind of marriage) gives rise to an obligation to take the important desires and projects of our partners seriously, trying not to frustrate those desires if we can. Now, the important point is that this obligation, viewed from a contractualist perspective, has to be understood as dealing with objective satisfactions of desire. Since what we aim at is the real thing, it would be intolerable that we should agree to principles that would give equal credit to plausible substitutes. (This follows, I think, from the contractualist commitment to the ideal of publicity in moral principles.) So it may be that Philippa fails in her obligations to Kurt, even supposing that there is no real danger of him finding out.
I thus maintain that the intuition that Kurt has been harmed may derive, in the end, from the fact that wrong has been done him, understood from a contractualist perspective. So this is not a harm that a utilitarian can recognize. To see this clearly, we need an example where the putative harm is caused accidentally (so there is no question of wrong-doing), and where it is clear that there is no danger that objective dissatisfaction should ever become subjective (so there is no harm in the sense of risk, either). To this end, let me introduce the example of Astrid, the astronaut, variants of which will recur at various points through the remainder of this book.
Suppose that Astrid is a very rich woman, who has become tired of life on Earth with its squalor and constant violence. Accordingly, she buys herself a space-rocket, and takes off on a trajectory that is set irreversibly to carry her out of our solar system, and forever out of contact with her fellow humans. She does not even carry with her a radio with which she can be contacted. Now suppose that before leaving Earth she had erected a statue in memory of her beloved late husband, and one of her most cherished desires is that the statue should outlast her. But within months of her departure the statue is struck by lightning and destroyed. Is Astrid harmed? It seems to me plain that she is not, since she can never know. Yet her desire has been objectively frustrated. This confirms my suggestion that what really underpins the intuition that harm has been done, in cases such as Kurt’s, is that wrong has been done, involving either a risk of harm, or an infringement of principles that are only intelligible from a contractualist perspective.
I conclude that death is, indeed, a harm to the one who dies, but solely because death prevents future subjective satisfactions of desire (that is to say, continued worthwhile existence), not because it prevents many of the person’s desires for the future from being objectively satisfied. Let me stress again, however, that this thesis need not imply corresponding claims about the reasons we have for fearing death, or about the reasons why killing is wrong. To claim that the harm of death consists in preventing subjective satisfactions of desire certainly does not imply that our only reason for fearing death is to gain those satisfactions. On the contrary, almost any desire, whether it be a desire for a feeling of satisfaction or an objective state of affairs, can give one reason to fear death. Nor does that thesis imply that the only possible direct objection to killing is that it prevents future subjective satisfactions. On the contrary, contractualists, at least, will regard most killings as wrong because they infringe autonomy, quite apart from the harm that they do.
The wrongness of killing
If the conclusions of the previous section are correct, then it is clear that death is a harm to an animal in exactly the same way that it is a harm to a human being - in both cases death (normally) prevents future enjoyments and satisfactions that would otherwise have occurred. It follows that if killing humans is directly wrong, for a utilitarian, because of the harm that it causes - because it cuts off future worthwhile existence - then on precisely the same grounds it must be directly wrong to kill an animal. Since there can be no reason why an impartial observer should refuse to recognize the enjoyments of an animal as having moral standing, it would be mere speciesism to claim that it is wrong to prevent future enjoyments in the case of a human being without saying the same for an animal.
This is not yet to say that killing an animal would be as wrong as killing a human being, however. For as we shall see in a later section, some have claimed that the distinctive enjoyments of human beings have greater moral worth - are ‘higher’ - than those of an animal. In which case, although killing an animal would normally be directly wrong, there might still be no question of weighing up animal lives against the life of a human.
Even the conclusion that killing an animal is directly wrong may depend very much on what specific version of utilitarianism is endorsed, however. If utility is cashed in terms of happiness or pleasure, then these are apparently states that animals can enjoy just as much as we can. And then the fact that killing an animal would prevent future pleasure will be a reason against it, just as it is a reason against killing a human. But some utilitarians, including Singer, think that utility is better cashed in terms of the fulfilment of preferences. On such an account, the main reason against killing a human being is that most humans have a strong preference for going on living. But, it is claimed, most (perhaps all) animals are incapable of having such a desire. An animal has preferences for satisfaction as against suffering, but if animals are incapable of conceptualising their own future non-existence, then they cannot have a preference for their own future existence, as against non-existence. In Chapter 6 I shall consider to what extent these claims about the cognitive powers of animals are true. For the moment, let us see what follows on the supposition that they are.
A preference-utilitarian approach
How should preference-utilitarianism be understood? In particular, is it objective or subjective satisfactions of preference that are to enter into calculations of utility? Clearly, I think, the answer has to be that it is subjective satisfactions that matter, for at least two reasons. First, notice that if it was objective satisfactions of desire that were the basic utilitarian value, then in calculating utility we should be obliged to give as much weight to the preferences of those long dead as to those of the living. Suppose that all the inhabitants of Franksville in the year 1900, for example, wanted very much that the statue of their beloved founder Frank should stand in the town square for as long as the town lasted. None of those people is now living, and the present inhabitants of Franksville find the statue of Frank ugly, and wish to see it removed. Suppose that in the interim the population of the town has shrunk. Then if it is objective satisfactions that are to count, a preference-utilitarian might have to claim that we are morally obliged to keep the statue where it is, since this is the option that objectively satisfies the most desires. This seems intuitively absurd.
The second reason why preference-utilitarianism has to be understood in terms of subjective rather than objective satisfactions is more deeply theoretical. It is that it is impossible to see why an impartial benevolent observer should give any weight to satisfactions of preference that are merely objective. For why should such an observer count it as a good thing that people get what they want, as such, independently of whether or not they believe that they have got it? It is surely no part of benevolence to do something that satisfies someone’s desire in circumstances in which the person will never know what has happened.
To make this point vivid, consider a variant of the example of Astrid, the astronaut. As before, the statue of her late husband is destroyed soon after her departure from Earth. I argued previously that this cannot be regarded as harming her. Let us now ask whether, knowing her feeling on the matter, I would act benevolently if I were to arrange for the statue to be rebuilt. It is surely clear that I would not. Although such an action would objectively satisfy Astrid’s desire, and although it might serve to express my sense of mourning for her absence, it would not now be of any benefit to her. And benevolence surely has to do with the provision of benefit and the prevention of harm.
I conclude that preference-utilitarianism has to be understood in terms of subjective satisfactions of desire. We need to ask next, just which desires are to count. Suppose it is replied first, that only presently existing desires are to be considered. Then the desires of animals - for example, to avoid present suffering - may make it wrong to hunt them or factory farm them. But since animals do not, it is supposed, have desires for their own continued existence, it will not be wrong to bring about their deaths. For the future desires of an animal, that would be involved in its continued worthwhile existence if it were not killed, are not to be counted at all, on the present proposal. Since human beings, in contrast, do generally desire continued existence, we are obliged to respect that desire, and killing will normally be directly wrong in consequence.
Notice that the position we have reached here is in many ways a curious one. For so long as an animal has active preferences - so long as it is hungry, or thirsty, or is wanting to play - then the principle of equal consideration of interests will require that we should, other things being equal, try to satisfy those preferences. If there is nothing better that you can achieve with your time and resources, then a preference-utilitarian will have to claim that you are morally obliged to feed a hungry dog. But as soon as the animal no longer has any active preferences - is sitting contentedly after eating, for example, or has fallen asleep - then you would not be failing to fulfil any of its desires if you killed it. So you are obliged to feed the dog while it is hungry, but as soon as it is satisfied you may kill it. This combination of views seems strange, to say the least.
More importantly, if we are required to restrict attention to presently existing preferences, then we can give no moral weight to preferences that we know will exist in the future. For example, suppose that David is subject to temporary fits of severe depression, during which nothing seems worthwhile. At the moment he is suffering such a depression, and has no desire to go on living - he might kill himself if he could only find the energy. But I know perfectly well that by tomorrow he will be back to normal again. If it were only present desires that counted, for a preference-utilitarian, then it would seem that there is no direct moral objection to my killing David. But this is absurd. The fact that he will again have a strong desire for continued life in the future is surely sufficient to make such an action wrong.
Consider another example to reinforce the point. Adolescents and young adults commonly deny vigourously that they ever wish to have children. Indeed, there is no reason to think that they are insincere. But we know that, for most of them, the issue will strike them very differently in a few years time. Now other things being equal, a policy of offering free sterilisations to such people would surely be wrong, on the grounds that it prevents them from satisfying their future desire to have children. But if only present desires were to be counted, then there would be no direct moral objection to the policy.
I conclude that a preference-utilitarian should certainly give weight to the subjective satisfaction of both present and future desires. But then utilitarians will again be in the position of having to say that there are essentially the same reasons against killing animals as there are against killing human beings. It may be true that animals do not presently have desires for their own continued existence. But it is also normally true that they will have desires for satisfactions and for avoidance of suffering in the future, provided that they are not killed. And these desires should now be given equal weight with any others. Since you should try to ensure the satisfaction of the animal’s future desires, you are therefore normally obliged not to kill it.
A human being will have, in general, many more desires at any given time than will an animal. But this is not to the point. What matters, is the number (and intensity) of the desires that can or will be satisfied. And here there need be no difference between human and animal. So preserving the life of a human being will not necessarily lead to more desire satisfaction than would preserving the life of an animal. This will depend upon the details of the cases. The only difference between killing animals and humans thus far, is that by not bringing about the death of a human you will generally satisfy one more desire - namely, the presently existing desire for continued life. But in the case of depressed David there will not even be this difference.
If a preference-utilitarian should give weight to both present and future desires, then is the aim simply to maximise desire satisfaction (whether average or total) overall? This can seem counter-intuitive. For one way to comply with it would be to set about creating in people easily satisfied desires. Now, there may not be anything especially wrong with creating such desires. I do not particularly wish to commit myself to condemning consumer society at this point. But there can surely be no moral obligation to support such a society, merely on the grounds that with more and more desires continually being created in people by advertising, more and more desires are continually being satisfied.
There are difficult issues arising here for preference-utilitarianism. Some have attempted to overcome them by retreating to the notion of a rational desire, claiming that only present and future desires that qualify as rational ones are to be given moral weight. (Another issue arising, is whether it is only the future desires of actually existing creatures that are to be counted. This comes up especially in connection with population policy.) The notion of a rational desire is notoriously difficult to define. For our purposes it will be enough to distinguish two broad approaches to the problem. On the one hand, we could explain the notion of a rational desire in terms of the modes of desire-formation that are normal for the cognition of the creature involved. This would allow animals, and non-rational agents generally, to have rational desires. But then on the other hand, we could explain the notion of a rational desire in terms of the sorts of processes of thinking and reasoning that are distinctive of rational agents. Taking this option would exclude the future desires of animals from the moral domain once more.
To take this second option would be blatantly speciesist, however. It is impossible to see any reason why an impartial benevolent observer should discount a particular desire, merely because the creature in question had not subjected it to intellectual scrutiny. It is easy to see why such an observer might discount the present desires of depressed David, or desires produced by advertising, hypnosis, or drug addiction. For these desires have been created and sustained by processes that are disruptive of the normal cognitive lives of the agents involved. But there can be no reason why such an observer, if genuinely impartial, should discount or give less weight to the desires of an animal, merely because that animal had failed to engage in such activities as thinking carefully about alternatives, and checking through the presuppositions of its desire for false beliefs.
I conclude that utilitarians, of whatever variety, are committed to saying that the killing of an animal is almost always directly wrong, just as is the killing of a human. The question now, is whether a utilitarian must say that killing an animal is just as wrong, provided that the number and intensity of the desires or pleasures involved are roughly proportional.
The value of life
In addition to the appeal to preference-utilitarianism criticised above, Singer has quite another argument for saying that killing a rational agent - a person - is worse than killing an animal. It is that the lives of rational agents are intrinsically more valuable than the lives of at least most kinds of animal. Now this is not - and had better not be - an appeal to any form of moral intuitionism, which we considered and rejected in the opening chapter. The idea is not that the greater value of human lives is an objective fact, apprehended by us through a special faculty of moral intuition. Rather, Singer’s idea is to deploy a version of the classical utilitarian distinction between higher and lower pleasures, transformed now into a distinction between higher and lower modes of life.
(It is worth noting that Regan, too, feels obliged to deploy a variant of this distinction, although he is by no means a utilitarian. For he wishes to explain our intuition that, in a case where four men and a dog are adrift on a life-raft that can only safely support four creatures, it is right that the dog should be the one to be thrown off. Regan thinks that the basis for our intuition is that the distinctive enjoyments of a dog have less intrinsic value than those of human beings. Since Regan is not a utilitarian, this is presumably supposed to be one of those objective facts about the world that we are somehow - and mysteriously - to apprehend through the procedure of reflective equilibrium.)
The criterion for a pleasure, or a mode of living, being higher, for a utilitarian, is that anyone who has had experience of both would prefer it. Now, we have all had experience of animal pleasures - of a full stomach and a doze in the sun, for example. Yet no one would seriously wish for a life that contained only such pleasures, without the distinctively intellectual enjoyments of reading a novel, listening to music, or engaging in animated conversation with a friend. So it appears to follow that a mode of life that is distinctively human is more valuable than the life of an animal. As Mill famously maintained, it is better to be Socrates dissatisfied than a pig satisfied.
It is easy to see the rationale, for utilitarians, of the distinction between higher and lower pleasures in simple cases. For it provides them with a way of ranking pleasures, in terms other than intensity and duration, that all should rationally agree to. For example, suppose that everyone who has had experience of both, prefers the taste of pineapple to the taste of dry bread. This gives us reason to think that the same would hold even of those who have never tasted pineapple, were they to do so. Suppose Poppy is such a person. Then knowing these facts, she ought rationally to agree that, other things being equal, it is more important that someone should have the pleasure of eating a pineapple than that she herself should have the pleasure of eating dry bread. For she should acknowledge that she too, were she to experience it, would rank the former pleasure higher.
Problems begin to arise, however, in cases where the character of the subject must undergo substantial change in order to appreciate the new range of pleasures. For this may involve a corresponding inability to appreciate previous enjoyments. For example, many years of disciplined study may be required to appreciate certain intellectual pleasures, such as the pursuit of philosophy or of higher mathematics. But then it may be that the changes of character necessary to appreciate these pleasures unfit one for full enjoyment of singing, dancing, and spontaneity. In which case, the person who has had experience of both sorts of pleasure (that is, the intellectual) is no longer a competent judge of their relative values.
These problems become even more acute when we are trying to compare modes of life across different species. How are we, fairly and realistically, to compare the life of a horse with the life of a human being, given the vast changes in cognitive powers and dispositions that would be necessary to move from the one to the other? Singer attempts to circumvent this problem through the use of an imaginary device. He asks us to imagine a creature with the power to transform itself into each mode of life in turn - living first as a horse, then as a human, and then in some mode different from either, but retaining an exact memory of what each of the first two ways of living was like. Is it not plausible, he asks, that such a creature would judge the life of the human to be more valuable than the life of the horse?
Singer has clearly biased the issue here, however. For note that the creature in question is supposed to have articulate memories of its previous existences, and is supposed to be able to entertain sophisticated judgements about the relative values of those existences. In these respects the mode of existence of that creature is much closer to ours than to that of the horse. Small wonder, then, that such a creature should prefer the life of a human, since it might be expected to judge the life of the horse to be dull and unvaried by comparison. Yet these are, of course, distinctively human values, reflecting our relative cognitive complexity and sophistication.
What we have to do is, in fact, well nigh impossible. We have to imagine an impartial benevolent observer - a Martian, perhaps - with interests and mode of cognition no more similar to ours than to those of the horse, who nevertheless has full inside knowledge of what our respective modes of existence are like. In so far as I am able to form any conception of such an observer, I can see no reason why they should judge our human existence to be more valuable than that of the horse. I conclude that while the distinction between higher and lower pleasures may be intelligible and useful for a utilitarian in connection with simple cases, where it is merely lack of experience that prevents a direct comparison, that distinction is useless in attempting to rank pleasures and modes of life across species, with their differing modes of cognition.
The electrode workers
I have been arguing that there is no principled way in which utilitarians can show a human life to be more valuable than the life of an animal. It has to be admitted, however, that there is a powerful intuitive appeal behind the sorts of common-sense beliefs that utilitarians try to capture by means of the distinction between higher and lower pleasures. What I shall maintain in this section, is that this appeal is only really explicable from the stand-point of contractualism - thus driving one more nail into the coffin of the utilitarian approach to the animals issue.
Consider the following imaginary example, which is grounded in the well known fact that the brains of many mammals, including rats and monkeys, contain a so-called ‘pleasure centre’. If an electrode is inserted into this centre, then the animal in question will engage in an arbitrary activity - such as pressing a bar - for hours on end, in order to have its pleasure centre stimulated. Now imagine that such a centre were discovered in human beings. Suppose also, that some enterprising employers begin to offer implants to their manual workers, in such a way that those workers will have their pleasure centres stimulated every time they make one of the movements required for their work - say pulling a lever. Those who accept the offer soon come to live for their work, and the attendant pleasure it provides. They gladly work a sixteen hour day, eat on the job, and return home at night only to sleep. They say they cannot understand, now, how anyone else can choose to live differently. Yet there is, surely, a powerful common-sense intuition that their mode of life is impoverished, and that it might be morally wrong to opt for an implant-life, say by bringing one’s child into the factory to be wired up.
It is easy to understand how we, now, would have very good reason not to opt for an implant-life. For all of our current desires, interests, and projects would be lost sight of in such a life. We therefore have almost as much reason to fear an implant as we have to fear death - everything we presently care about would be lost. But at the same time we must recognize that someone who is already an electrode worker has just as much reason to fear the removal of their electrode. For in their case, too, the result would be the loss of everything that they presently care about, in exchange for a set of interests and concerns that they do not currently share. So there is nothing in this that can justify the claim that the one mode of life has, in itself, a greater moral value than the other. (Of course there are all sorts of secondary ways in which those who are not electrode workers may be more useful to other people.) Nor can a utilitarian provide a justification for such a claim through the distinction between higher and lower pleasures. For the positions of each of the two groups with respect to the other will be symmetrical - I presume that each would prefer their present mode of existence to their past one.
Yet the intuition that there is a moral difference here is surely very powerful. Suppose, for example, that the implant has to be made, and wired up regularly, quite early in life in order to be fully successful. You are now considering whether to have an implant made in the brain of your ten year old son, Imri. If you do so, you know that you will be more or less ensuring his future happiness! For he will then have just one overwhelming desire, that will be easily, and almost continually, satisfied. Yet it would be very wrong of you, surely, to commit Imri to such a future. For in the circumstances you know that, once wired up, he will never want to change - indeed, he will never again have a serious desire for anything else.
This intuition is only really explicable from a contractualist perspective, given that we believe that the electrode workers fail to retain the capacity for planning and choosing that is distinctive of rational agency. (For extended discussion of what is involved in the notion of a rational agent, see Chapter 6.) I think we are inclined to maintain that the desire for pleasure so dominates their cognition as to leave no place for the exercise of genuine autonomy. In fact, their situation is exactly like that of a willing drug addict, only without the debilitating effects of drug addiction. The electrode workers retain the potential for rational agency, of course, since if unplugged they would soon return to normal. Indeed, there is a real sense in which they continue to have the capacity for it too, since their minds presumably retain the cognitive structures necessary for autonomous action. It is merely that their implant prevents them from exercising this capacity - just as cotton wool packaging can prevent a brittle glass from exercising its capacity to be broken. (Seen in this light, there would then be a powerful case for rescuing the electrode workers from their plight, overriding their own vigourous resistance - though this is not the point I particularly wish to focus on.)
What this then means is that it will be us, rather than the electrode workers, who get to frame the terms of the moral contract. For moral rules, within contractualism, are created by rational, choosing, autonomous agents. To put the point slightly differently, the electrode workers cannot reasonably reject any proposed system of rules, since their situation is such that they are no longer capable of exercising their capacity for rational decision. So they can be allowed to have no objection to any rules that might be proposed, including those that would prevent Imri from becoming an implant boy. What would be more natural, then, than that we should decide to outlaw practices that undermine the exercise of genuine autonomy, as would becoming an electrode worker in childhood? For our status as autonomous agents is presupposed in almost everything that we care about. What therefore emerges, is that a contractualist should claim, in a way that a utilitarian cannot, that it would be directly wrong to take Imri into the factory to be wired up - which is just what our common-sense intuition tells us.
Life as a journey
Singer has, more recently, made yet another attempt to ground his view that the lives of rational agents are more valuable than the lives of animals, while retaining his utilitarian perspective. He argues that a human life may usefully be conceived of as a journey. If I am travelling on a journey, but am forced to abandon it, my disappointment will generally be in proportion to my nearness to the goal, and to the amount of effort that has gone in to the travelling, which now turns out to have been wasted. So, too, in life, Singer thinks. Much of early life is mere preparation for what follows, and many of us have long-term projects that give shape to, and help to make sense of, our lives. It is then less tragic if death should occur early in childhood, when the journey has barely begun, or late in life, when most goals have been attained, and most projects completed. There is also held to be less direct moral objection to killing human beings at these stages in their lives.
Most importantly for our purposes, Singer argues that the point at which the journey of life begins, from the perspective of those who travel, is the point at which they first begin to conceive of themselves as having a future and a past, and to think of some of their current activities as preparatory for the future. For suppose that this were so, and that most animals lack such a conception of themselves altogether (an assumption we shall examine in Chapter 6). Then such an animal will never have embarked on the journey, and death, for them, will be no tragedy, since it does not interrupt any journey. Nor will there be any direct moral objection to killing such a creature.
I can see no theoretical rationale for these views, however. Why should an impartial benevolent observer give less weight to (or discount altogether) the interests of someone who is standing still, as against someone who has embarked on a journey? Those who stay at home have desires, purposes, and feelings no less than those who go abroad. What may be true, is that an observer who is comparing two travellers will count as more serious an interruption in the journey of the one who is closest to completion, other things being equal. For, having invested more, that person will have more to lose. But this provides no ground for thinking that an impartial observer will only consider the interests of those who travel (that is, of those who have long term plans and projects). In which case, we have been given no reason why a utilitarian should maintain that the death of an animal, or a baby, is less serious than the death of a rational agent. On the contrary, the fact that death prevents future satisfaction of desires in all these cases provides the same (and the only) direct utilitarian rationale against killing.
It is true that the judgements that Singer seeks to explain through the metaphor of a journey have an intuitive appeal, for many people. Many do feel that the death of a baby, or an old person, is less of a tragedy, from the perspective of the one who dies, than the death of someone in the prime of life. But I think these intuitions are not utilitarian ones, and certainly provide no basis on which a utilitarian can claim that the death of an animal is less morally significant than the death of a rational agent. Let me explain.
Take the case of babies first. Anyone who has contractualist sympathies might be expected to share the thought that the death of a baby, by accident or from natural causes, is less of a tragedy, from the perspective of the baby, than the death of a normal adult. For such people may be expected to value rational agency above all else, and the baby is presumably not, as yet, a rational agent. (It is arguable, indeed, that all rational agents will value the possession of rational agency highly, quite apart from whatever moral views they may hold.) But it should be emphasised that this is not to claim that a contractualist will count the killing of a baby to be less serious than the killing of an adult, as we shall see in the next chapter. Nor is there any way in which utilitarians can motivate the claim that the death of a baby is less morally serious, unless they retreat to a form of indefensible intuitionism, claiming that it is an objective fact about the world that those who have a conception of their past and future are more valuable than those who do not. For, as we have seen, there is no reason why an impartial observer should discount the present and future desires of the baby, merely because they are not yet linked together by any overall life-plan.
Now consider the case of the very old. Again, many share the feeling that the death of such a person is less of a tragedy than the death of a young adult. But these judgements serve mainly to express a comparison with reasonable expectations, made to console the living (like saying ‘He had a good innings’) - which need not be a point of view shared by the one who has died. It is true that some old people may gradually wind down their activities and projects as they near the term of their expected span. And in some such cases it may rightly be said that the person had little more to live for by the time of their death. But others keep going as they have always done, as if death were only for others. (Interestingly, members of the latter group tend also to live longer.) There is nothing here to support the view that life, as such, may be thought of as a journey - only that some people may conceive of their lives in some such terms.
I conclude that utilitarians are committed to the view that there is the same sort of direct moral objection to killing animals as there is to killing humans. For there are essentially the same utilitarian reasons against such killings in both cases - that killing would prevent future enjoyment, and that not killing is necessary if the organism is to have its future preferences fulfilled. Moreover, there is no coherent way for a utilitarian to claim that the life of a human being has, in itself, greater moral value than the life of an animal, without degenerating into moral intuitionism. The only valid utilitarian reasons remaining, for why it will generally be worse to kill a human being than to kill an animal, are extrinsic ones. These are first, that human beings tend to live longer than most animals, so a greater extent of life will generally be cut off by death. And second, that the death of a human being will generally cause much suffering to friends and relatives, in a way that the death of an animal will rarely cause suffering to other animals.
Can such a position be acceptable under reflective equilibrium? I believe not. Consider the following development of the example of Kenneth, the kennel owner, first presented in Chapter 1. You arrive at a fire in his dogs’ home to find Kenneth unconscious on the floor, while the dogs are still locked in their cages. You judge that you have just enough time, either to drag Kenneth to safety, or to unlock the cages, but not both. Suppose you also know that Kenneth is quite old, and is something of a recluse who lives entirely for his work, without anyone to care for him. In these circumstances a utilitarian is clearly committed to the view that you should opt to rescue the dogs. For this is obviously the way to ensure the greatest future pleasure, and/or the greatest future desire satisfaction. Utilitarians cannot avoid this conclusion by discounting the interests of the dogs altogether, without engaging in a form of speciesism which must be unacceptable from their own perspective.
This conclusion is morally outrageous, however, as are its further consequences. Once it is accepted that the killing of an animal is just as morally serious, in general, as the killing of a human being, then those practices that involve the regular slaughter of animals, such as farming and some forms of animal experimentation, will seem to fall within the same moral category as the Nazi holocaust. And then any form of opposition to such practices, of whatever degree of violence, will seem eminently justified. In fact, those animals rights activists who pursue the methods of terrorism - planting bombs and poisoning baby-foods - are only following utilitarianism through to its logical, but morally abhorrent, conclusion.
Our common-sense, pre-theoretical, view is that it would be very wrong to place the lives of many dogs over the life of a single (albeit old and friendless) human. This belief is probably too firmly held, in the case of most of us, to be shaken by theoretical argument. (Recall from Chapter 1, indeed, that it is a belief shared even by those philosophers who have been most vociferous in defence of animals, namely Regan and Singer.) Moreover, since this common-sense belief will prove to be retained under contractualism (as we shall see in later chapters) but is lost under utilitarianism, and since contractualism is in other respects just as, if not more, theoretically satisfying than utilitarianism, the correct response is to reject utilitarianism altogether. At any rate, the utilitarian approach to animal lives seems just as unacceptable as we found the utilitarian approach to animal suffering to be in the last chapter.
Death is a harm to the one who dies only in so far as it prevents future worthwhile existence. Yet our reason for fearing death is that continued life will be presupposed by almost all of our desires. From a utilitarian perspective there is essentially the same direct moral objection to killing an animal as there is to killing a human - namely, that the death prevents future enjoyments, and that not being killed is a necessary condition for the future desires of the creature to be satisfied. From the same perspective there is no reason to count the life of an animal as less valuable than the life of a rational agent. I have argued that these consequences are too extreme to be believed.
 For further discussion of the issue, see my Introducing Persons, chs. 3 & 7.
 See particularly Thomas Nagel, ‘Death', in his Mortal Questions.
 Similar points are made by Joel Feinberg, ‘Harm and Self-Interest' in P. Hacker & J. Raz eds., Law, Morality and Society (Oxford University Press, 1977).
 See Practical Ethics, ch. 4.
 For discussion of the many puzzles involved, see Derek Parfit, Reasons and Persons (Oxford University Press, 1984), part IV.
 See Practical Ethics, pp. 88-90.
 See Animal Rights, p. 324.
 See Practical Ethics, p. 89.
 See ‘Life's Uncertain Voyage' in P. Pettit, R. Sylvan & J. Norman eds., Metaphysics and Morality (Blackwell, 1987).
 On this, see my Introducing Persons, ch. 8.