Constraints and Animals
Excerpted from Anarchy, State, and Utopia, New York: Basic Books, 1974, pp. 35-42
  Acrobat version
We can illuminate the status and implications of moral side con­straints by considering living beings for whom such stringent side constraints (or any at all) usually are not considered appropriate: namely, nonhuman animals. Are there any limits to what we may do to animals? Have animals the moral status of mere objects? Do some purposes fail to entitle us to impose great costs on animals? What entitles us to use them at all?

Animals count for something. Some higher animals, at least, ought to be given some weight in people's deliberations about what to do. It is difficult to prove this. (It is also difficult to prove that people count for something!) We first shall adduce particular examples, and then arguments. If you felt like snapping your fingers, perhaps to the beat of some music, and you knew that by some strange causal connection your snapping your fingers would cause 10,000 contented, unowned cows to die after great pain and suffering, or even painlessly and instantaneously, would it be per­fectly all right to snap your fingers? Is there some reason why it would be morally wrong to do so?

Some say people should not do so because such acts brutalize them and make them more likely to take the lives of persons, solely for pleasure. These acts that are morally unobjectionable in them­selves, they say, have an undesirable moral spillover. (Things then would be different if there were no possibility of such spillover— for example, for the person who knows himself to be the last per­son on earth.) But why should there be such a spillover? If it is, in itself, perfectly all right to do anything at all to animals for any reason whatsoever, then provided a person realizes the clear line between animals and persons and keeps it in mind as he acts, why should killing animals tend to brutalize him and make him more likely to harm or kill persons? Do butchers commit more murders? (Than other persons who have knives around?) If I enjoy hitting a baseball squarely with a bat, does this significantly increase the danger of my doing the same to someone's head? Am I not capable of understanding that people differ from baseballs, and doesn't this understanding stop the spillover? Why should things be different in the case of animals? To be sure,  it is an empirical question whether spillover does take place or not; but there is a puzzle as to why it should, at least among readers of this essay, sophisticated people who are capable of drawing distinctions and differentially acting upon them.

If some animals count for something, which animals count, how much do they count, and how can this be determined? Suppose (as I believe the evidence supports) that eating animals is not necessary for health and is not less expensive than alternate equally healthy diets available to people in the United States. The gain, then, from the eating of animals is pleasures of the palate, gustatory delights, varied tastes. I would not claim that these are not truly pleasant, delightful, and interesting. The question is: do they, or rather does the marginal addition in them gained by eating animals rather than only nonanimals, outweigh the moral weight to be given to animals' lives and pain? Given that animals are to count for something, is the extra gain obtained by eating them rather than nonanimal products greater than the moral cost? How might these questions be decided?

We might try looking at comparable cases, extending whatever judgments we make on those cases to the one before us. For ex­ample, we might look at the case of hunting, where I assume that it's not all right to hunt and kill animals merely for the fun of it. Is hunting a special case, because its object and what provides the fun is the chasing and maiming and death of animals? Suppose then that I enjoy swinging a baseball bat. It happens that in front of the only place to swing it stands a cow. Swinging the bat unfor­tunately would involve smashing the cow's head. But I wouldn't get fun from doing that; the pleasure comes from exercising my muscles, swinging well, and so on. It's unfortunate that as a side effect (not a means) of my doing this, the animal's skull gets smashed. To be sure, I could forego swinging the bat, and instead bend down and touch my toes or do some other exercise. But this wouldn't be as enjoyable as swinging the bat; I won't get as much fun, pleasure, or delight out of it. So the question is: would it be all right for me to swing the bat in order to get the extra pleasure of swinging it as compared to the best available alternative activity that does not involve harming the animal? Suppose that it is not merely a question of foregoing today's special pleasures of bat swinging; suppose that each day the same situation arises with a different animal. Is there some principle that would allow killing and eating animals for the additional pleasure this brings, yet would not allow swinging the bat for the extra pleasure it brings? What could that principle be like? (Is this a better parallel to eat­ing meat? The animal is killed to get a bone out of which to make the best sort of bat to use; bats made out of other material don't give quite the same pleasure. Is it all right to kill the animal to obtain the extra pleasure that using a bat made out of its bone would bring? Would it be morally more permissible if you could hire someone to do the killing for you?)

Such examples and questions might help someone to see what sore of line he wishes to draw, what sort of position he wishes to take. They face, however, the usual limitations of consistency arguments; they do not say, once a conflict is shown, which view to change. After failing to devise a principle to distinguish swing­ing the bat from killing and eating an animal, you might decide that it's really all right, after all, to swing the bat. Furthermore, such appeal to similar cases does not greatly help us to assign precise moral weight to different sorts of animals. (We further discuss the difficulties in forcing a moral conclusion by appeal to examples in Chapter 9.)

My purpose here in presenting these examples is to pursue the notion of moral side constraints, not the issue of eating animals. Though I should say that in my view the extra benefits Americans today can gain from eating animals do not justify doing it. So we shouldn't. One ubiquitous argument, not unconnected with side constraints,  deserves mention:  because people eat animals,  they raise more than otherwise would exist without this practice. To exist for a while is better than never to exist at all. So (the argument concludes) the animals are better off because we have the prac­tice of eating them. Though this is not our object, fortunately it turns out that we really, all along, benefit them! (If tastes changed and people no longer found it enjoyable to eat animals, should those concerned with the welfare of animals steel themselves to an unpleasant task and continue eating them?) I trust I shall not be misunderstood as saying that animals are to be given the same moral weight as people if I note that the parallel argument about people would not look very convincing. We can imagine that pop­ulation problems lead every couple or group to limit their children to some number fixed in advance. A given couple, having reached the number, proposes to have an additional child and dispose of it at the age of three (or twenty-three) by sacrificing it or using it for some gastronomic purpose.   In  justification,   they note  that the child will not exist at all if this is not allowed; and surely it is bet­ter for it to exist for some number of years. However, once a per­son exists,  not everything compatible with his overall existence being a net plus can be done, even by those who created him. An existing person has claims, even against those whose purpose in creating him was to violate those claims. It would be worthwhile to pursue moral objections to a system that permits parents to do anything whose permissibility is necessary for their choosing to have the child, that also leaves the child better off than if it hadn't been born.8 (Some will think the only objections arise from dif­ficulties in accurately administering the permission.) Once they exist, animals too may have claims to certain treatment. These claims may well carry less weight than those of people. But the fact that some animals were brought into existence only because someone wanted to do something that would violate one of these claims does not show that the claim doesn't exist at all.

Consider the following (too minimal) position about the treat­ment of animals. So that we can easily refer to it, let us label this position "utilitarianism for animals, Kantianism for people." It says: (i) maximize the total happiness of all living beings; (2) place stringent side constraints on what one may do to human beings. Human beings may not be used or sacrificed for the bene­fit of others; animals may be used or sacrificed for the benefit of other people or animals only if those benefits are greater than the loss inflicted. (This inexact statement of the utilitarian position is close enough for our purposes, and it can be handled more easily in discussion.) One may proceed only if the total utilitarian benefit is greater than the utilitarian loss inflicted on the animals. This utilitarian view counts animals as much as normal utilitarianism does persons. Following Orwell, we might summarize this view as: all animals are equal but some are more equal than others, (None may be sacrificed except for a greater total benefit; but persons may not be sacrificed at all, or only under far more stringent con­ditions, and never for the benefit of nonhuman animals. I mean (i) above merely to exclude sacrifices which do not meet the utilitar­ian standard, not to mandate a utilitarian goal. We shall call this position negative utilitarianism.)

We can now direct arguments for animals counting for some­thing to holders of different views. To the "Kantian" moral philos­opher who imposes stringent side constraints on what may be done to a person, we can say:

You hold utilitarianism inadequate because it allows an individual to be sacrificed to and for another, and so forth, thereby neglecting the strin­gent limitations on how one legitimately may behave toward persons. But could there be anything morally intermediate between persons and stones, something without such stringent limitations on its treatment, yet not to be treated merely as an object? One would expect that by sub­tracting or diminishing some features of persons, we would get this intermediate sort of being. (Or perhaps beings of intermediate moral status are gotten by subtracting some of our characteristics and adding others very different from ours.)

Plausibly, animals are the intermediate beings, and utilitarianism is the intermediate position. We may come at the question from a slightly different angle. Utilitarianism assumes both that happiness is all that matters morally and that all beings are interchangeable. This conjunc­tion does not hold true of persons. But isn't (negative) utilitarianism true of whatever beings the conjunction does hold for, and doesn't it hold for animals?

To the utilitarian we may say:

If only the experiences of pleasure, pain, happiness, and so on (and the capacity for these experiences) are morally relevant, then animals must be counted in moral calculations to the extent they do have these capaci­ties and experiences. Form a matrix where the rows represent alternative policies or actions, the columns represent different individual organisms, and each entry represents the utility (net pleasure, happiness) the policy will lead to for the organism. The utilitarian theory evaluates each policy by the sum of the entries in its row and directs us to perform an action or adopt a policy whose sum is maximal. Each column is weighted equally and counted once, be it that of a person or a nonhuman animal. Though the structure of the view treats them equally, animals might be less important in the decisions because of facts about them. If animals have less capacity for pleasure, pain, happiness than humans do, the ma­trix entries in animals' columns will be lower generally than those in people's columns. In this case, they will be less important factors in the ultimate decisions to be made.

A utilitarian would find it difficult to deny animals this kind of equal consideration. On what grounds could he consistently dis­tinguish persons' happiness from that of animals, to count only the former? Even if experiences don't get entered in the utility matrix unless they are above a certain threshold, surely some animal ex­periences are greater than some people's experiences that the utili­tarian wishes to count. (Compare an animal's being burned alive unanesthetized with a person's mild annoyance.) Bentham, we may note, does count animals' happiness equally in just the way we have explained.

Under "utilitarianism for animals, Kantianism for people," animals will be used for the gain of other animals and persons, but persons will never be used (harmed, sacrificed) against their will, for the gain of animals. Nothing may be inflicted upon persons for the sake of animals. (Including penalties for violating laws against cruelty to animals?) Is this an acceptable consequence? Can't one save 10,000 animals from excruciating suffering by inflicting some slight discomfort on a person who did not cause the animals' suf­fering? One may feel the side constraint is not absolute when it is people who can be saved from excruciating suffering. So perhaps the side contraint also relaxes, though not as much, when animals' suffering is at stake. The thoroughgoing utilitarian (for animals and for people, combined in one group) goes further and holds that, ceteris paribus, we may inflict some suffering on a person to avoid a (slightly) greater suffering of an animal. This permissive principle seems to me to be unacceptably strong, even when the purpose is to avoid greater suffering to a person!

Utilitarian theory is embarrassed by the possibility of utility monsters who get enormously greater gains in utility from any sac­rifice of others than these others lose. For, unacceptably, the theory seems to require that we all be sacrificed in the monster's maw, in order to increase total utility. Similarly if people are util­ity devourers with respect to animals, always getting greatly coun­terbalancing utility from each sacrifice of an animal, we may feel that "utilitarianism for animals, Kantianism for people," in requiring (or allowing) that almost always animals be sacrificed, makes animals too subordinate to persons.

Since it counts only the happiness and suffering of animals, would the utilitarian view hold it all right to kill animals pain­lessly? Would it be all right, on the utilitarian view, to kill people painlessly, in the night, provided one didn't first announce it? Utilitarianism is notoriously inept with decisions where the number of persons is at issue. (In this area, it must be conceded, eptness is hard to come by.) Maximizing the total happiness requires con­tinuing to add persons so long as their net utility is positive and is sufficient to counterbalance the loss in utility their presence in the world causes others. Maximizing the average utility allows a per­son to kill everyone else if that would make him ecstatic, and so happier than average. (Don't say he shouldn't because after his death the average would drop lower than if he didn't kill all the others.) Is it all right to kill someone provided you immediately substitute another (by having a child or, in science-fiction fashion, by creating a full-grown person) who will be as happy as the rest of the life of the person you killed? After all, there would be no net diminution in total utility, or even any change in its profile of distribution. Do we forbid murder only to prevent feelings of worry on the part of potential victims? (And how does a utilitarian explain what it is they're worried about, and would he really base a policy on what he must hold to be an irrational fear?) Clearly, a utilitarian needs to supplement his view to handle such issues; perhaps he will find that the supplementary theory becomes the main one, relegating utilitarian considerations to a corner.

But isn't utilitarianism at least adequate for animals? I think not. But if not only the animals' felt experiences are relevant, what else is? Here a tangle of questions arises. How much does an animal's life have to be respected once it's alive, and how can we decide this? Must one also introduce some notion of a nondegraded existence? Would it be all right to use genetic-engineering tech­niques to breed natural slaves who would be contented with their lots? Natural animal slaves? Was that the domestication of ani­mals? Even for animals, utilitarianism won't do as the whole story, but the thicket of questions daunts us.