Animal Experimentation

Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.   ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• ANIMAL EXPERIMENTATION ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• alastair norcross Introduction •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• FN:1 I take the central issue concerning the ethics of animal experimentation to be the moral status of animals.1 Since most animal experimentation involves treating experimental subjects in ways that would clearly not be morally acceptable if the subjects were human, and since no animal experimentation involves the informed consent of the experimental subject(s), any attempt to justify such experimentation must include a defense of the claim that the moral status of animals differs significantly from that of humans. The influence of animal welfare advocates, in particular Peter Singer, Tom Regan, and their followers, but certainly dating back to Bentham and Mill, seems to have resulted in at least the grudging acceptance by the research community that animals have some moral status. That is, that the interests of animals should be taken into account when designing and justifying experiments involving them. For example, Baruch Brody argues for what he calls ‘a reasonable pro-research position on animal research’, which is committed to at least the following propositions: 1. Animals have interests (at least the interest in not suffering, and perhaps others as well), which may be adversely affected either by research performed on them or by the conditions under which they live before, during, and after the research. 1 In keeping with common conventions I use the term ‘animal’ in this chapter to refer to non-human animals. Page 648 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  2. The adverse effect on animals’ interests is morally relevant, and must be taken into account when deciding whether or not a particular program of animal research is justified or must be modified or abandoned. 3. The justification for conducting a research program on animals that would adversely affect them is the benefits that human beings would receive from the research in question. 4. In deciding whether or not the research in question is justified, human interests should be given greater significance than animal interests. (Brody 2003: 262–3) In clarifying 4, Brody argues that human interests should be given proportionally greater significance than animal interests, as opposed to lexically greater significance. He does not, therefore, claim that any benefit whatsoever for humans can justify the infliction of any harm, no matter how great on animals. He doesn’t attempt to say precisely how much greater significance should be given to human interests. It seems reasonable to say, though, that if this approach is to justify much (though perhaps not all) of the research that currently involves animals, the difference in significance must be vast. Consider such examples of animal experimentation as the Draize Eye Irritancy Test, in which quantities of cleaning fluids are tested on rabbits’ eyes, or the infamous learned helplessness experiments of Martin Seligman, in which dogs were subjected to repeated painful shocks from which they couldn’t escape. If these experiments, or many others like them, are to be justified by appeal to the claim that human interests should be given greater significance than animal interests, the difference in significance cannot be small. If human interests are merely somewhat more significant than animal interests, it should be acceptable to perform such experiments on humans, so long as the humans suffer somewhat less than the animals (or perhaps so long as somewhat fewer humans are subjected to the experiments). I know of no defenders of animal experimentation who are also prepared to defend painful experiments on humans just so long as these conditions are met. Attempts to justify the widespread practice of giving little or no consideration to the vital interests of animals (the most obvious one being the interest in avoiding suffering) have been made from several different ethical perspectives. This chapter will explore three of the most common perspectives—utilitarianism, natural rights theory, and social contract theory—and explain why none of them is likely to justify the claim that the interests of humans are vastly more significant than the like interests of animals. While many people may be somewhat disturbed at learning the details of many medical and psychological experiments involving animals, relatively few seriously challenge the moral permissibility of such practices. The status quo in this regard appears to be that, minor details aside, our treatment of animals raises no serious moral questions. I will discuss the utilitarian approach in the first section, where I will argue that the utilitarian case against the status quo is overwhelming. In the next section I will consider various attempts to defend Page 649 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross the status quo from within a natural rights framework, and will argue that all such attempts fail. Finally, I will turn to social contract theory, which appears to hold out the most hope for the defender of the status quo with respect to our treatment of animals. In a recent book, Peter Carruthers has vigorously defended the view that social contract theory can justify the claim that all and only humans have basic moral rights. His approach, he claims, provides the only satisfactory way to justify giving greater weight to the interests of severely retarded humans than to those of animals with equal or greater cognitive capacities. That is, it gives an answer to what is commonly called ‘‘the argument from marginal cases’’. I will argue both that social contract theory fails to give such an answer, and that all the well-known versions of the theory actually beg the question against attributing basic moral standing to animals. The ways in which both a natural rights approach and a social contract approach attempt to answer the argument from marginal cases embody a deeply flawed view of morality. Utilitarianism •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• Most forms of utilitarianism consist of both a theory of the good and a theory of the right. The theory of the good tells us what states of affairs are intrinsically valuable or desirable, while the theory of the right tells us what actions are right or wrong, morally obligatory or morally forbidden. The standard utilitarian account of the good is that happiness, or more broadly, well-being, is intrinsically good, and unhappiness is intrinsically bad. The early utilitarians Jeremy Bentham and, to a certain extent, John Stuart Mill equated happiness with pleasure and unhappiness with pain. More recent utilitarians give a broader account of well-being, some including desire-satisfaction as an essential component, but most agree that pain and other forms of suffering are intrinsically bad. All suffering is bad, not just my suffering, or that of my family, or nation, or race, or species. The standard utilitarian account of the right is that the right action is that action, of all possible alternatives, that results in the greatest balance of good over bad. If more than one action results in the same balance of good over bad, and no actions result in a greater balance, all such actions are right, although none is obligatory. Any action that is not right is wrong. This approach to the rightness and wrongness of actions can also be applied to moral evaluations of character, rules, social practices and institutions, and so on. So, for example, a system of government will be judged morally acceptable or unacceptable by a utilitarian depending on whether there are any viable alternative systems that would result in a greater net balance of happiness. So what does utilitarianism say about the moral status of animals? Consider an animal abuser who tortures dogs and cats out of malevolent curiosity. Our common moral sensibilities are appalled by such behavior. Utilitarianism provides a clear Page 650 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  explanation of what is wrong with the abuser’s behavior. The dogs and cats are made to suffer for no sufficient reason. In this respect, the utilitarian answer accords with ordinary intuitions. But the utilitarian approach also calls into question much commonly accepted animal experimentation (and animal agriculture). Many experimental subjects, such as rats, mice, rabbits, and monkeys, are made to suffer, sometimes severely, in the process of medical, pharmaceutical, and psychological research. Perhaps we could deny the moral significance of this treatment of animals by denying that they feel pain. It is often claimed that this was Descartes’s position, though the truth, as I will explain shortly, is more complicated. Whatever Descartes and his contemporaries may have thought, however, it is hard to find anyone today who seriously claims that animals don’t feel pain. The evidence that they do, both physiological and behavioristic, is simply overwhelming. It seems, then, that in order to justify the widespread infliction of animal suffering, a utilitarian will have to argue for a pretty hefty outweighing benefit. What are the prospects for such an argument to succeed? Perhaps a utilitarian defender of the status quo will deny that she needs to argue for a large benefit to outweigh animal suffering. Perhaps she will say that I was mistaken to claim that all suffering is intrinsically bad. It is only human suffering that is intrinsically bad, she might say. Or perhaps she will admit that animal suffering is, indeed, bad, but nothing like as bad as human suffering. What reason could she supply for such differential concern for animal suffering? Perhaps she will claim that animal suffering is of lesser (or no) moral significance, because animals themselves are of lesser (or no) moral significance. They have less intrinsic value than humans, or maybe none at all. While this line of reasoning is fairly common in discussions of the moral status of animals, it is not one to which a utilitarian can appeal. Utilitarians hold that certain types of states have intrinsic value and disvalue, not types of creatures. Talk of an individual creature’s intrinsic value is best understood in terms of the intrinsic value of the life of the individual, which in turn amounts to the intrinsic value of the states (usually the mental states) that comprise the life. Given the theoretical primacy of judgements about the intrinsic value of mental states of individuals, claims about the intrinsic value of the individuals themselves cannot be used to justify claims about the intrinsic value of the individuals’ mental states. It may well be that the typical human life is of greater intrinsic value than the typical canine life, but this will be because the human life is comprised of a greater and richer variety of experiences, emotions, hopes, aspirations, and the like. The sufferings, however, of a dog, considered in and of themselves, are of no lesser (or greater) moral significance than the like sufferings of a human being. There is one other line of reasoning open to a utilitarian to deny moral significance to animal suffering. Consider the following partial characterization of what Derek Parfit calls Preference-Hedonism: ‘On the use of ‘‘pain’’ which has rational and moral significance, all pains are when experienced unwanted, and a pain is worse or Page 651 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross greater the more it is unwanted’ (Parfit 1984: 493). Some might even claim that it is part of the very concept of pain that it is unwanted. Even if we deny this, it seems plausible to say that a pain is only bad to the extent that it is unwanted. If someone really doesn’t care about a pain, in and of itself, it is hard to see how the pain could be intrinsically bad. So what does this have to do with animals? Recall Descartes. Although he didn’t deny that animals have sensations, such as pain, he did deny that they have what he called ‘‘thoughts’’, which included both beliefs and desires. (His argument for this, which I won’t explore here, has to do with animals’ lack of linguistic ability.) If animals are incapable of desire, they are a fortiori incapable of desiring that painful sensations cease. This would also provide a desire-satisfaction utilitarian with a reason to deny moral status to animals. So, what should we say about the denial that animals have desires? At first sight, it seems almost as unbelievable as the denial that they feel pain. Only a philosopher could make such an obviously false claim with a straight face. Recall some of the other outrageous claims made by philosophers over the ages: motion is impossible; all is flux; all is water; there is no such thing as weakness of will; the physical world is just a collection of ideas; the unregulated free market will work to the benefit of all. Of course animals want things. Any pet owner can tell you that. However, as someone who has been known to make some seemingly outrageous claims myself, I cannot dismiss this one without at least examining an argument for it. A philosopher who argues that animals don’t have desires is R. G. Frey. Here, briefly, in his own words is his argument: I may as well say at once that I do not think that animals can have desires. My reasons for thinking this turn largely upon my doubts that animals can have beliefs, and my doubts in this regard turn partially, though in large part, upon the view that having beliefs is not compatible with the absence of language and linguistic ability. (Frey 1989: 40) So, why does Frey claim that desires require beliefs? Here is the example he uses to argue for this claim: Suppose I am a collector of rare books and desire to own a Gutenberg Bible: my desire to own this volume is to be traced to my belief that I do not now own such a work and that my rare book collection is deficient in this regard . . . . without this belief, I would not have this desire. (Frey 1989: 40) I don’t wish to dwell on this part of Frey’s argument, since the more interesting claim is that beliefs depend on linguistic ability. However, it is worth pointing out that, even if we accept his example of the desire for a Gutenberg Bible depending on a belief, it may well be that other, perhaps more basic desires, such as the desire for food, don’t depend on beliefs. So, what of his claim that beliefs require linguistic ability? Here he is again, still on the example of the Gutenberg Bible: Now what is it that I believe? I believe that my collection lacks a Gutenberg Bible; that is, I believe that the sentence ‘My collection lacks a Gutenberg Bible’ is true. In constructions of the form ‘I believe that,’ what follows upon the ‘that’ is a declarative sentence; and what I Page 652 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  believe is that that sentence is true. The difficulty in the case of animals should be apparent: if someone were to say, e.g., ‘The cat believes that the door is locked,’ then that person is holding, as I see it, that the cat holds the declarative sentence ‘The door is locked’ to be true; and I can see no reason whatever for crediting the cat or any other creature which lacks language, including human infants, with entertaining declarative sentences and holding certain sentences to be true. (Frey 1989: 40–1) The most obvious flaw with this reasoning is that it generates an infinite regress. According to Frey’s approach, my belief that my collection lacks a Gutenberg Bible just is my belief that the sentence ‘My collection lacks a Gutenberg Bible’ is true. But by the same reasoning, my belief that the sentence ‘My collection lacks a Gutenberg Bible’ is true just is my belief that the sentence ‘the sentence ‘‘My collection lacks a Gutenberg Bible’’ is true’ is true. And so on. How plausible is it, for example, that my belief that my collection lacks a Gutenberg Bible just is my belief that the sentence ‘the sentence ‘‘the sentence ‘the sentence ‘‘the sentence ‘My collection lacks a Gutenberg Bible’ is true’’ is true’ is true’’ is true’ is true? Perhaps a less problematic way of tying beliefs and desires to language could be found, but it seems doubtful that it could do the moral work necessary for justifying the infliction of suffering on animals. There may well be a whole range of beliefs and desires that does require linguistic ability. However, the ethically significant ones, such as the desire that a pain cease, do not seem to do so. Even if we define desires in such a way that no nonlinguistic creature has them, there is clearly some mental state of the suffering dog that is importantly similar to a human’s desire that the pain cease. So much for any utilitarian attempt to dismiss the intrinsic moral significance of animal suffering. Isn’t it nonetheless possible that the suffering involved in animal experimentation is outweighed by the benefits thereby produced? Notice that a utilitarian demands of an action or institution not that it result in a greater amount of happiness than unhappiness, but that it result in a greater balance of happiness than available alternatives (ignoring the possibility of ties). This detail is important, though sometimes ignored in discussions of the justifiability of animal experimentation. Let me illustrate the difference, with reference to a common criticism of utilitarianism. Some critics charge that utilitarianism is defective on the grounds that it could be used to justify the institution of slavery. Imagine, they say, a society with a small number of slaves and a large number of free citizens. Perhaps the slaves are exceedingly unhappy. Perhaps, indeed, the unhappiness of each slave is many times greater than the happiness of each free citizen. However, if there are enough free citizens, their happiness will outweigh the unhappiness of the slaves. But this is still not enough for the system to be justified on utilitarian grounds. Perhaps the free citizens could have been just as happy, or even happier, in a society without slaves. In which case, assuming that the slaves would have been happier not being slaves, there would have been a bigger balance of happiness over unhappiness in the free society. (The point of this example is not to argue that utilitarianism Page 653 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross couldn’t justify some system of slavery, but to point out that the possibility of such a system being justified on utilitarian grounds is even more remote than it might initially appear.) The relevance of this point to the moral status of animal experimentation should be clear. To justify a particular practice that inflicts significant suffering on animals it is not enough to argue that the benefits of the practice (probably to humans) are greater than the suffering of the animals. What needs to be argued is that nothing like as much benefit could be achieved without significant animal suffering. What of the benefits of animal experimentation? Aren’t there enormous benefits to humans (and maybe other animals) that can only be achieved through the use of animals in research? I won’t explore this empirical question in detail here. It doesn’t require more than a cursory glance at the literature, though, to conclude that huge numbers of animal experiments provide little or no benefit, and could never have been reasonably expected to do so. Many drugs are tested on animals in order to compete on a market already glutted with drugs that do the same job. Much psychological research merely confirms what common sense tells us, and serves only to advance the career of the researcher. Even many of those experiments that do, arguably, give results that have beneficial applications may not be justified on utilitarian grounds. Perhaps only a lesser benefit could have been achieved without animal suffering. Nonetheless, the difference in benefit may well be smaller than the suffering in question. It is sometimes objected that we cannot apply a utilitarian approach to the justification of individual experiments, because we simply never know when we might make a significant breakthrough. If we had to justify each experiment in advance, we wouldn’t justify any, and would thereby miss out on those that do lead to great benefits. If the utilitarian approach had been used in the past, it is claimed, we would have missed out on many of the beneficial advances in medicine. This line of reasoning, though, either fails in its own terms or begs the question against the utilitarian approach. Either the benefits from the use of animals in research really do outweigh the animal suffering or they don’t. If they do, an expected utility calculation will give the result that at least some experiments are justified. If they don’t, the fact that we would miss out on the benefits if we abandoned animal research is not sufficient, morally, to justify such research. But perhaps supporters of research will claim that we simply never know which experiments will result in benefit, even though, on balance, the benefits outweigh the harms. So we can never justify an experiment in advance, on utilitarian grounds, even though we have good reasons to believe that the practice of animal experimentation as a whole can be so justified. This response assumes far too pessimistic a view of our powers of prediction. Researchers don’t select lines of enquiry at random, simply hoping to get lucky. There is plenty of evidence on which to base decisions. It is surely reasonable that, in order to justify the certain infliction of suffering on animals, there has to be some reason to expect a significant benefit. In the absence of such Page 654 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  a reason, we cannot simply resort to the claim that the unexpected sometimes happens. Despite these considerations, there may well be some animal experiments that are justified on utilitarian grounds, but it is likely to be a small fraction of the number actually performed. To summarize the conclusions of the present section, it seems likely that a utilitarian approach to morality will condemn much, and perhaps most, animal experimentation. Whatever benefit, if any, that comes from most experiments is simply not enough to justify the amount of suffering involved. Natural Rights Theory •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• In this section I will discuss an approach to the moral status of animals that, for the sake of convenience, I refer to as ‘‘natural rights theory’’. This approach focuses on identifying certain natural features or properties of individuals or species as the basic grounds for the attribution of differing moral status. So, for example, rationality has often been claimed as the grounds for the superior moral status of human beings over animals. For the purposes of this discussion, to claim that humans have a superior moral status to animals is to claim that it is morally right to give the interests of humans greater weight than those of animals in deciding how to behave. Such claims will often be couched in terms of rights, such as the rights to life, liberty, or respect, but nothing turns on this terminological matter. One may claim that it is generally wrong to kill humans, but not animals, because humans are rational, and animals are not. Or one may claim that the suffering of animals counts less than the suffering of humans (if at all), because humans are rational, and animals are not. These claims may proceed through the intermediate claim that the rights of humans are more extensive and stronger than those (if any) of animals. Alternatively, one may directly ground the judgement about the moral status of certain types of behavior in claims about the alleged natural properties of the individuals involved. What can a proponent of this approach say about the moral status of animals? The traditional view, dating back at least to Aristotle, is that rationality is what separates humans, both morally and metaphysically, from other animals. With a greater understanding of the cognitive powers of some animals, recent philosophers have often refined the claim to stress the kind and level of rationality required for moral reasoning. Let’s start with a representative sample of three. Consider first these claims of Bonnie Steinbock: While we are not compelled to discriminate among people because of different capacities, if we can find a significant difference in capacities between human and non-human animals, this could serve to justify regarding human interests as primary. It is not arbitrary or smug, I think, to maintain that human beings have a different moral status from members of other Page 655 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross species because of certain capacities which are characteristic of being human. We may not all be equal in these capacities, but all human beings possess them to some measure, and non-human animals do not. For example, human beings are normally held to be responsible for what they do . . . . Secondly, human beings can be expected to reciprocate in a way that non-human animals cannot . . . Thirdly . . . there is the ‘desire for self-respect’. (Steinbock 1997: 467–8) Similarly, Mary Anne Warren argues that ‘the rights of persons are generally stronger than those of sentient beings which are not persons’. Her main premise to support this conclusion is the following: there is one difference [between human and non-human nature] which has a clear moral relevance: people are at least sometimes capable of being moved to action or inaction by the force of reasoned argument. (Warren 1997: 482) Carl Cohen, one of the most vehement modern defenders of what Peter Singer calls ‘‘speciesism’’, states his position as follows: Between species of animate life, however—between (for example) humans on the one hand and cats or rats on the other—the morally relevant differences are enormous, and almost universally appreciated. Humans engage in moral reflection; humans are morally autonomous; humans are members of moral communities, recognizing just claims against their own interest. Human beings do have rights, theirs is a moral status very different from that of cats or rats. (Cohen 1992: 462) So, the claim is that human interests and/or rights are stronger or more important than those of animals, because humans possess a kind and level of rationality not possessed by animals. How much of our current behavior towards animals this justifies depends on just how much consideration should be given to animal interests, and on what rights, if any, they possess. Both Steinbock and Warren stress that animal interests need to be taken seriously into account. Warren claims that animals have important rights, but not as important as human rights. Cohen, on the other hand, argues that we should actually increase our use of animals. One of the most serious challenges to this defense of the status quo involves a consideration of what philosophers refer to as ‘‘marginal cases’’. Whatever kind and level of rationality is selected as justifying the attribution of superior moral status to humans will either be lacking in some humans or present in some animals. To take one of the most commonly suggested features, many humans are incapable of engaging in moral reflection. For some, this incapacity is temporary, as is the case with infants, or the temporarily cognitively disabled. Others who once had the capacity may have permanently lost it, as is the case with the severely senile or the irreversibly comatose. Still others never had and never will have the capacity, as is the case with the severely mentally disabled. If we base our claims for the moral superiority of humans over animals on the attribution of such capacities, won’t we have to exclude many humans? Won’t we then be forced to the claim that there is at least as much moral reason to use cognitively deficient humans in experiments Page 656 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  (and for food) as to use animals? Perhaps we could exclude the only temporarily disabled, on the grounds of potentiality, though that move has its own problems. Nonetheless, the other two categories would be vulnerable to this objection. I will consider two lines of response to the argument from marginal cases. The first denies that we have to attribute different moral status to marginal humans, but maintains that we are, nonetheless, justified in attributing different moral status to animals who are just as cognitively sophisticated as marginal humans, if not more so. The second admits that, strictly speaking, marginal humans are morally inferior to other humans, but proceeds to claim pragmatic reasons for treating them, at least usually, as if they had equal status. As representatives of the first line of defense, I will consider arguments from three philosophers, Carl Cohen, Alan White, and David Schmidtz. First, Cohen: [the argument from marginal cases] fails; it mistakenly treats an essential feature of humanity as though it were a screen for sorting humans. The capacity for moral judgement that distinguishes humans from animals is not a test to be administered to human beings one by one. Persons who are unable, because of some disability, to perform the full moral functions natural to human beings are certainly not for that reason ejected from the moral community. The issue is one of kind . . . What humans retain when disabled, animals have never had. (Cohen 1992: 460–1) Alan White argues that animals don’t have rights, on the grounds that they cannot intelligibly be spoken of in the full language of a right. By this he means that they cannot, for example, claim, demand, assert, insist on, secure, waive, or surrender a right. This is what he has to say in response to the argument from marginal cases: Nor does this, as some contend, exclude infants, children, the feeble-minded, the comatose, the dead, or generations yet unborn. Any of these may be for various reasons empirically unable to fulfill the full role of right-holder. But . . . they are logically possible subjects of rights to whom the full language of rights can significantly, however falsely, be used. It is a misfortune, not a tautology, that these persons cannot exercise or enjoy, claim, or waive, their rights or do their duty or fulfil their obligations. (White 1989: 120) David Schmidtz defends the appeal to typical characteristics of species, such as mice, chimpanzees, and humans, in making decisions on the use of different species in experiments. He also considers the argument from marginal cases: Of course, some chimpanzees lack the characteristic features in virtue of which chimpanzees command respect as a species, just as some humans lack the characteristic features in virtue of which humans command respect as a species. It is equally obvious that some chimpanzees have cognitive capacities (for example) that are superior to the cognitive capacities of some humans. But whether every human being is superior to every chimpanzee is beside the point. The point is that we can, we do, and we should make decisions on the basis of our recognition that mice, chimpanzees, and humans are relevantly different types. We can have it both ways after all. Or so a speciesist could argue. (Schmidtz 1998: 61) There is something deeply troublesome about the line of argument that runs through all three of these responses to the argument from marginal cases. A Page 657 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross particular feature, or set of features, is claimed to have so much moral significance that its presence or lack can make the difference to whether a piece of behavior is morally justified or morally outrageous. But then it is claimed that the presence or lack of the feature in any particular case is not important. The relevant question is whether the presence or lack of the feature is normal. Such an argument would seem perfectly preposterous in most other cases. Suppose, for example, that ten famous people are on trial in the afterlife for crimes against humanity. On the basis of conclusive evidence, five are found guilty and five are found not guilty. Four of the guilty are sentenced to an eternity of torment, and one is granted an eternity of bliss. Four of the innocent are granted an eternity of bliss, and one is sentenced to an eternity of torment. The one innocent who is sentenced to torment asks why he, and not the fifth guilty person, must go to hell. St Peter replies, ‘Isn’t it obvious, Mr Ghandi? You are male. The other four men—Adolph Hitler, Joseph Stalin, Richard Nixon, and George W. Bush—are all guilty. Therefore the normal condition for a male defendant in this trial is guilt. The fact that you happen to be innocent is irrelevant. Likewise, of the five female defendants in this trial, only one was guilty. Therefore the normal condition for female defendants in this trial is innocence. That is why Margaret Thatcher gets to go to heaven instead of you.’ As I said, such an argument is preposterous. Is the reply to the argument from marginal cases any better? Perhaps it will be claimed that a biological category such as a species is more ‘‘natural’’, whatever that means, than a category like ‘‘all the male (or female) defendants in this trial’’. Even setting aside the not inconsiderable worries about the conventionality of biological categories, it is not at all clear why this distinction should be morally relevant. What if it turned out that there were statistically relevant differences in the mental abilities of men and women? Suppose that men were, on average, more skilled at manipulating numbers than women, and that women were, on average, more empathetic than men. Would such differences in what was ‘‘normal’’ for men and women justify us in preferring an innumerate man to a female math genius for a job as an accountant, or an insensitive woman to an ultra-sympathetic man for a job as a counselor? I take it that the biological distinction between male and female is just as real as that between human and chimpanzee. A second response to the argument from marginal cases is to concede that cognitively deficient humans really do have an inferior moral status to normal humans. Can we, then, use such humans as we do animals, and experiment on them (and raise them for food)? How can we advocate this second response while blocking such uses of marginal humans? Warren suggests that ‘there are powerful practical and emotional reasons for protecting non-rational human beings, reasons which are absent in the case of most non-human animals’ (Warren 1997: 483). Here is Steinbock in a similar vein: I doubt that anyone will be able to come up with a concrete and morally relevant difference that would justify, say, using a chimpanzee in an experiment rather than a human being Page 658 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  with less capacity for reasoning, moral responsibility, etc. Should we then experiment on the severely retarded? Utilitarian considerations aside, we feel a special obligation to care for the handicapped members of our own species, who cannot survive in this world without such care . . . . In addition, when we consider the severely retarded, we think, ‘That could be me’. It makes sense to think that one might have been born retarded, but not to think that one might have been born a monkey . . . . Here we are getting away from such things as ‘morally relevant differences’ and are talking about something much more difficult to articulate, namely, the role of feeling and sentiment in moral thinking. (Steinbock 1997: 469–70) This line of response clearly won’t satisfy those who think that marginal humans really do deserve equal moral consideration with other humans. It is also a very shaky basis on which to justify our current practices. What outrages human sensibilities is a very fragile thing. Human history is littered with examples of widespread acceptance of the systematic mistreatment of some groups who didn’t generate any sympathetic response from others. That we do feel a kind of sympathy for retarded humans that we don’t feel for dogs is, if true, a contingent matter. Perhaps we could claim that the practice of giving greater weight to the interests of all humans than of animals is justified on evolutionary grounds. Perhaps such differential concern has survival value for the species. Something like this may well be true, but it is hard to see the moral relevance. We can hardly justify the privileging of human interests over animal interests on the grounds that such privileging serves human interests! Although the argument from marginal cases certainly poses a formidable challenge to any proposed criterion of full moral standing that excludes animals, it doesn’t, in my view, constitute the most serious flaw in such attempts to justify the status quo. The proposed criteria are all variations on the Aristotelian criterion of rationality. But what is the moral relevance of rationality? Why should we think that the possession of a certain level or kind of rationality renders the possessor’s interests of greater moral significance than those of a merely sentient being? In Bentham’s famous words ‘The question is not, Can they reason? nor Can they talk? But, Can they suffer?’ What do defenders of the alleged superiority of human interests say in response to Bentham’s challenge? Some, such as Carl Cohen, simply reiterate the differences between humans and animals that they claim to carry moral significance. Animals are not members of moral communities, they don’t engage in moral reflection, they can’t be moved by moral reasons, therefore (?) their interests don’t count as much as ours. Others, such as Steinbock and Warren, attempt to go further. Here is Warren on the subject: ‘Why is rationality morally relevant? It does not make us ‘‘better’’ than other animals or more ‘‘perfect’’ . . . . But it is morally relevant insofar as it provides greater possibilities for cooperation and for the nonviolent resolution of problems’ (Warren 1997: 482). Warren is certainly correct in claiming that a certain level and kind of rationality is morally relevant. Where she, and others who Page 659 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross give similar arguments, go wrong is in specifying what the moral relevance amounts to. If a being is incapable of moral reasoning, at even the most basic level, if it is incapable of being moved by moral reasons, claims, or arguments, then it cannot be a moral agent. It cannot be subject to moral obligations, to moral praise or blame. Punishing a dog for doing something ‘wrong’ is no more than an attempt to alter its future behavior. So long as we are undeceived about the dog’s cognitive capacities, we are not, except metaphorically, expressing any moral judgement about the dog’s behavior. (We may, of course, be expressing a moral judgement about the behavior of the dog’s owner, who didn’t train it very well.) All this is well and good, but what is the significance for the question of what weight to give to animal interests? That animals can’t be moral agents doesn’t seem to be relevant to their status as moral patients. Many, perhaps most, humans are both moral agents and patients. Most, perhaps all, animals are only moral patients. Why would the lack of moral agency give them diminished status as moral patients? Full status as a moral patient is not some kind of reward for moral agency. I have heard students complain in this regard that it is unfair that humans bear the burdens of moral responsibility, and don’t get enhanced consideration of their interests in return. This is a very strange claim. Humans are subject to moral obligations, because they are the kind of creatures who can be. What grounds moral agency is simply different from what grounds moral standing as a patient. It is no more unfair that humans and not animals are moral agents, than it is unfair that real animals and not stuffed toys are moral patients. One other attempt to justify the selection of rationality as the criterion of full moral standing is worth considering. Recall the suggestion that rationality is important in so far as it facilitates cooperation. If we view the essence of morality as reciprocity, the significance of rationality is obvious. A certain twisted, but all too common, interpretation of the Golden Rule is that we should ‘‘do unto others in order to get them to do unto us’’. There’s no point, according to this approach, in giving much, if any, consideration to the interests of animals, because they are simply incapable of giving like consideration to our interests. Inasmuch as there is a consistent view being expressed here at all, it concerns self-interest, as opposed to morality. Whether it serves my interests to give the same weight to the interests of animals as to those of humans is an interesting question, but it is not the same question as whether it is right to give animals’ interests equal weight. The same point, of course, applies to the question of whether to give equal weight to my interests, or those of my family, race, sex, religion, etc., as to those of other people. Perhaps it will be objected that I am being unfair to the suggestion that the essence of morality is reciprocity. Reciprocity is important, not because it serves my interests, but because it serves the interests of all. Reciprocity facilitates cooperation, which in turn produces benefits for all. What we should say about this depends on the scope of ‘‘all’’. If it includes all sentient beings, then the significance of animals’ inability to reciprocate is in what it tells us about how to give their interests equal Page 660 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  consideration. It certainly can’t tell us that we should give less, or no, consideration to their interests. If, on the other hand, we claim that rationality is important for reciprocity, which is important for cooperation, which is important for benefiting humans, which is the ultimate goal of morality, we have clearly begged the question against giving equal consideration to the interests of animals. It seems that any attempt to justify the status quo with respect to our treatment of animals by appealing to a morally relevant difference between humans and animals will fail on at least two counts. It will fail to give an adequate answer to the argument from marginal cases, and, more importantly, it will fail to make the case that such a difference is morally relevant to the status of animals as moral patients as opposed to their status as moral agents. Social Contract Theory •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• For the would-be defender of the status quo, the most promising moral approach is social contract theory, or contractualism. Given its classical expression in Hobbes’s Leviathan, Rousseau’s The Social Contract, and Locke’s Second Treatise on Government, contractualism views morality as in some sense a human construct. If human beings were to live without rules, in what Hobbes and Rousseau refer to as a ‘state of nature’, life would be, in Hobbes’s memorable phrase ‘solitary, poor, nasty, brutish, and short’. It would then be in the interests of everyone to agree to abide by certain rules, such as a rule against killing others, on condition that others also agree. The content of the agreement, or contract, provides the rules of morality. It is no part of the theory that there ever was such an agreement. The contract itself is an enlightening fiction, useful to discover the requirements of morality. In the same way, a utilitarian can appeal to the fiction of an ideally informed, impartial, and benevolent observer to explain the content of that theory’s requirements. James Rachels expresses the basic idea of contractualism as follows: ‘Morality consists in the set of rules, governing how people are to treat one another, that rational people will agree to accept, for their mutual benefit, on the condition that others follow those rules as well’ (Rachels 1999: 137). In a recent book, Peter Carruthers has argued that a contractualist approach to ethics supports the status quo with respect to animals. He claims that the most plausible versions of contractualism accord full direct moral status to all humans, including the severely cognitively impaired, and deny direct moral status to all animals. He further claims that such an approach can explain the wrongness of many instances of cruelty to animals, without accepting that animal experimentation (or factory farming) is wrong, or that the animals who are the victims of wrongful cruelty have direct moral significance. Carruthers bases his discussion on two influential contemporary versions of contractualism; Page 661 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross the theories of John Rawls and Thomas Scanlon. Here are Carruthers’s summaries of the main points of the two theories: The basic idea, then, is that we are to think of morality as the rules that would be selected by rational agents choosing from behind what Rawls calls a veil of ignorance. While these agents may be supposed to have knowledge of all general truths of psychology, sociology, economics, and so on, they are to be ignorant of their own particular qualities (their intelligence, physical strength, qualities of character, projects and desires), as well as the position they will occupy in the society that results from their choice of rules . . . . The point of the restrictions is to eliminate bias and special pleading in the selection of moral principles . . . . Hence his proposal is, in fact, that moral rules are those that we should rationally agree to if we were choosing from a position of complete fairness . . . . Most importantly, the agents behind the veil of ignorance must not be supposed to have, as yet, any moral beliefs. For part of the point of the theory is to explain how moral beliefs can arise. [Scanlon’s] account of morality is roughly this: moral rules are those that no one could reasonably reject as a basis for free, unforced, general agreement amongst people who share the aim of reaching such an agreement . . . . here the agents concerned are supposed to be real ones, with knowledge of their own idiosyncratic desires and interests, and of their position within the current structure of society. The only idealisations are that choices and objections are always rational . . . and that all concerned will share the aim of reaching free and unforced agreement . . . the contractors will know that there is no point in rejecting a proposed rule on grounds special to themselves, since others would then have equal reason to reject any proposed rule. (Carruthers 1992: 37–9) So, how do animals fare on these approaches? It is fairly clear that they won’t be assigned more than indirect moral significance. Since the contractors, on both models, are rational agents motivated by self-interest, ‘only rational agents will be assigned direct rights’. The reasoning that leads to this conclusion is slightly different on the two approaches, so I will consider Carruthers’s treatment of each in turn. First, Rawls’s theory: Since it is rational agents who are to choose the system of rules, and choose self-interestedly, it is only rational agents who will have their position protected under the rules. There seems no reason why rights should be assigned to non-rational agents. Animals will, therefore, have no moral standing under Rawlsian contractualism, in so far as they do not count as rational agents. (Carruthers 1992: 98–9) The story on Scanlon’s approach is slightly different, since the contractors are there conceived as real people with differing preferences. In particular, some of them may care deeply about animals, and thus may be inclined to reject a proposed rule that gives little or no weight to the interests of animals. Carruthers objects to this suggestion on the grounds that such a rejection would not have a reasonable basis: It cannot be reasonable, therefore, to reject a rule merely because it conflicts with some interest or concern of mine. For every rule (except the entirely trivial) will conflict with someone’s concerns . . . If I can reasonably reject rules that accord no weight to the interests Page 662 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  of animals, then others can equally reasonably reject rules that allow us to dress and make love as we wish, and to worship or not worship as we please. (Carruthers 1992: 104) What rules, then, can reasonably be rejected? Carruthers’s answer is ‘rules that accord no weight to my interests in general, or rules that allow my privacy to be invaded, or my projects to be interfered with, at the whim of other people . . . the basic principle that we should agree upon is one of respect for the autonomy of rational agents’ (Carruthers 1992: 104–5). Of course, if one of my projects is to safeguard the interests of animals, a rule that allows others to disregard those interests does allow my project to be interfered with. It seems that respect for autonomy will have to incorporate a very strong moral asymmetry between what is done and what is allowed to happen. Lets assume, for the sake of argument, that such an asymmetry is justified. There are two serious objections that arise from within Carruthers’s approach. First, there is the problem of marginal cases again. For the same reasons that animals don’t get assigned moral standing in the contractualist framework, nonrational humans don’t seem to count either. Carruthers’s response is to suggest two arguments that the contractors would use to justify rules that accord full moral standing to marginal humans. First, there is the following slippery slope argument: There are no sharp boundaries between a baby and an adult, between a not-very-intelligent adult and a severe mental defective, or between a normal old person and someone who is severely senile. The argument is then that the attempt to accord direct moral rights only to rational agents would be inherently dangerous and open to abuse. (Carruthers 1992: 114) It is because starting out with a rule that distinguishes morally between rational and non-rational humans might lead to the mistreatment of rational humans that the rule has to include all humans. Excluding animals, on the other hand, wouldn’t have the same dangerous consequences. Anyone who argued from the accepted denial of moral standing to chimpanzees to the conclusion that some humans shouldn’t have moral standing either would not be taken seriously. Carruthers’s second argument has a similar reliance on psychological claims. It is simply a fact about human beings, he says, that they care deeply for their offspring, ‘irrespective of age and intelligence’. Given this fact, ‘a rule withholding moral standing from those who are very young, very old, or mentally defective is thus likely to produce social instability, in that many people would find themselves psychologically incapable of living in compliance with it’ (Carruthers 1992: 117). There are two pertinent questions with respect to these psychological claims. First, are they true? Second, if they are true, do they provide the appropriate grounds for the claim that the interests of marginal humans have the same moral weight as those of other humans? The answer to both questions is no. We already distinguish between marginal humans and others in the allocation of some rights. The severely mentally defective don’t get to vote, neither do they go to college. This selective treatment has led neither to the withholding of such benefits from Page 663 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross ordinarily rational humans, nor to widespread social instability. It might be objected that these are examples of different treatment of marginal humans, not different consideration of their interests. Severely cognitively deficient humans don’t vote or go to college, because it is not in their interests to do so. This distinction is morally significant, but it is only relevant to Carruthers’s psychological claims to the extent that it figures in the ordinary thinking of most people, which is hardly at all. Suppose, though, that Carruthers’s psychological claims were true. They would provide a very shaky basis on which to attribute moral standing to marginal humans. To see this, imagine that a new kind of birth defect (perhaps associated with beef from cows treated with bovine growth hormone) produces severe mental retardation, green skin, and a complete lack of emotional bond between parents and child. Furthermore, suppose that the mental retardation is of the same kind and severity as that caused by other birth defects that don’t have the other two effects. It seems likely that denying moral status to such defective humans would not run the same risks of abuse and destruction of social stability as would the denial of moral status to other, less easily distinguished and more loved defective humans. Would these contingent empirical differences between our reactions to different sources of mental retardation justify us in ascribing different direct moral status to their subjects? The only difference between them is skin color and whether they are loved by others. Any theory that could ascribe moral relevance to differences such as these doesn’t deserve to be taken seriously. Carruthers might reply that my own treatment of my example undermines its force. My argument demonstrates, he might say, why the denial of moral status to the green-skinned humans really would be subject to the slippery slope and social stability arguments. It is because philosophers such as I can show the moral irrelevance of the differences between the green-skinned humans and other marginal humans that we couldn’t justify rules that distinguished between them. But this response is unavailable to Carruthers, of all people. For my demonstration of the moral irrelevance of the differences between green-skinned humans and other humans is no different from other demonstrations of the moral irrelevance of the differences between many animals and humans. If we can appeal to the supposed persuasive force of one argument we can appeal to a similar persuasive force for the other. Unfortunately, neither argument has the requisite psychological force. Contractarianism fails, then, to give a convincing answer to the argument from marginal cases. It also fails to account for what Carruthers calls our common-sense attitudes towards animals. It seems to deny direct moral status to animals at all. The prevailing view may be that animals’ interests are not as significant as those of humans, but it is not that they count for nothing. According to this view, the cat torturer may not be doing something as bad as the child torturer, but his behavior is nonetheless morally abominable. Furthermore, it is what is done to the cat itself that is morally objectionable. A contractarian approach might suggest rules against Page 664 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  cruelty to animals, on the grounds of protecting the interests of animal owners and lovers. But this doesn’t capture the central wrong of torturing a cat. It would still be wrong, even if it were a stray and no one else found out about it. Carruthers’s response to this problem is similar to Kant’s, who objected to cruelty to animals on the grounds that ‘he who is cruel to animals becomes hard also in his dealings with men. We can judge the heart of a man by his treatment of animals.’ Similarly, Carruthers claims that cruelty to animals (in venues other than factory farms and laboratories) is a sign of a defective character. Anyone who treats animals with wanton cruelty will also probably treat rational agents with disregard for their legitimate interests. Rational contractors, therefore, would have a good reason to agree to rules that discouraged the development of such characters. This argument is subject to the same two objections as Carruthers’s response to the argument from marginal cases. Even though there is fairly strong evidence of a correlation between cruelty to animals and antisocial behavior towards people, it is by no means obvious that everyone who is wantonly cruel to animals is a danger to people. But even such evidence as exists doesn’t apply to factory farms or most laboratory experiments. Are we supposed to say that the interests of such animals don’t count at all, because they are tortured in ways that don’t warp their torturers’ characters? Besides, the ordinary view that the cat torturer’s behavior is morally abominable is in no way contingent on the belief that the torturer is also likely to mistreat people. If you were to discover that Mother Theresa routinely tortured cats for fun, you wouldn’t think ‘Well, what do you know! I guess torturing cats for fun isn’t always wrong.’ Neither would you think, ‘Well, what do you know! I guess Mother Theresa was actually a danger to people. What luck that she died before she got around to torturing any.’ You would probably be dismayed to learn that someone who had so much compassion for people could be so callous towards animals. The reason for your dismay, though, would be your belief that such callousness towards animals is wrong in itself. The problem with the contractarian approach, at least as presented by Carruthers, is that the specification of the rules as those chosen by rational self-interested individuals begs the question against ascribing moral status to the non-rational. So long as the contractors are motivated by self-interest and are aware of their own rationality, the result is bound to favor rational beings over the merely sentient. Of course, we could modify the approach, at least Rawls’s version, to eliminate this feature. If we simply specified that the veil of ignorance prevented the contractors from knowing whether they would, in the society whose rules they are choosing, be rational, the result wouldn’t give a privileged status to rational beings. Carruthers considers this move, as suggested by Tom Regan. His reasons for rejecting it expose the fundamental defect in the whole contractarian approach: The real line of reply to Regan is that his suggestion would destroy the theoretical coherence of Rawlsian contractualism. As Rawls has it, morality is, in fact, a human construction. Morality is viewed as constructed by human beings, in order to facilitate interactions between Page 665 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M.  alastair norcross human beings, and to make possible a life of co-operative community. This is, indeed, an essential part of the governing conception of contractualism . . . [In my own contractualist account of the source of moral motivation] the basic contractualist concept . . . is held to be innate, selected for in evolution because of its value in promoting the survival of our species. (Carruthers 1992: 102–3) For all I know, Carruthers’s claim that the basic contractualist concept is selected for is true. If true, it might tell us something, though it’s not clear how much, about the conditions of human flourishing. The most that such a claim could generate, though, would be a hypothetical imperative of the form ‘‘in order to promote human flourishing, treat animals and humans in the following ways’’. Even if the content of such an imperative included injunctions against making animals suffer, such injunctions would not have the status of basic moral rules. When we ponder the cat torturer’s behavior, we may well be moved, and rightly so, by the inconsistency of such behavior with realizing the goal of human flourishing. We are right to regard such considerations as morally relevant. However, if we believe that such considerations exhaust the realm of moral relevance, if, in particular, we believe that the cat’s suffering is of no direct moral relevance, we have a sadly impoverished view of morality. That the contractualist approach, and some versions of the natural rights approach, relegate the significance of animal suffering to the merely instrumental renders them unacceptable as moral theories, as opposed to theories of human flourishing. In conclusion, to the extent that we view morality as not simply a human creation, a device whose sole purpose is to ensure cooperation among humans, and thereby promote human flourishing, we have powerful reasons to reject the view that the interests of animals are less significant than the like interests of humans. Such a rejection will render much animal experimentation morally unacceptable. This is not a conclusion that will be eagerly embraced by the scientific community. It is, however, the conclusion best supported by a careful examination of the relevant moral reasons. References Brody, B. (2003), ‘Defending Animal Research: An International Perspective’, in E. F. Paul and J. Paul (eds.), Why Animal Experimentation Matters: The Use of Animals in Medical Research (New Brunswick, NJ: Transaction Publishers and the Social Philosophy and Policy Foundation, 2001), 131–47; repr. in S. J. Armstrong and R. Botzler (eds.), The Animal Ethics Reader (London: Routledge), 262–71. Carruthers, P. (1992), The Animals Issue: Moral Theory in Practice (Cambridge: Cambridge University Press). Cohen, C. (1992), ‘The Case for the Use of Animals in Biomedical Research’, New England Journal of Medicine, 315 (1987), 865–70; repr. in T. A. Mappes and J. S. Zembaty (eds.), Social Ethics, 4th edn. (New York: McGraw-Hill), 458–67. Page 666 Steinbock chap27.tex V1 - September 7, 2006 8:03 P.M. animal experimentation  Frey, R. G. (1989), ‘Rights, Interests, Desires, and Beliefs’, American Philosophical Quarterly, 16 (1979), 233–9; repr. as ‘Why Animals Lack Beliefs and Desires’, in T. Regan and P. Singer (eds.), Animal Rights and Human Obligations, 2nd edn. (Englewood Cliffs, NJ: Prentice Hall), 39–42. Parfit, D. (1984), Reasons and Persons (Oxford: Oxford University Press). Rachels, J. (1999), The Elements of Moral Philosophy, 3rd edn. (New York: McGraw-Hill). Schmidtz, D. (1998), ‘Are All Species Equal?’, Journal of Applied Philosophy, 15/1: 57–67. Steinbock, B. (1997), ‘Speciesism and the Idea of Equality’, Philosophy, 53/204 (1979), 247–56; repr. in J. E. White (ed.), Contemporary Moral Problems, 5th edn. (St Paul, Minn.: West), 464–70. Warren, M. A. (1997), ‘Difficulties with the Strong Animal Rights Position’, Between the Species, 2/4 (1987), 433–41; repr. in J. E. White (ed.), Contemporary Moral Problems, 5th edn. (St Paul, Minn.: West), 479–85. White, A. (1989), Rights (New York: Oxford University Press, 1986); repr. in T. Regan and P. Singer (eds.), Animal Rights and Human Obligations, 2nd edn. (Englewood Cliffs, NJ: Prentice Hall), 119–21. Page 667