THREE QUESTIONS

Do animals have rights? More precisely: is it true that many non-human animals – at least the mammals and the birds, at most the sentient ones – do have moral rights? My main aim is to find out how a rule-consequentialist should answer this question. And my main conclusion is that rule-consequentialism, unlike the ethical theories that shaped the animal rights debate, provides a solid foundation for our intuitions about what we owe to the members of other species.

The first thing to note is that the question of whether animals have rights remains ambiguous because there are at least two importantly distinct ways to understand the term ‘moral rights’. In the broadest or weakest sense of ‘moral rights’, the statement that x has moral rights just means that x has moral status or standing. If the beings of a given kind have moral status, then we have some moral duties or obligations to them – and not just duties regarding them. The view that animals have moral status commits us to this claim: their interests count, and count directly, from an ethical point of view. We can act wrongly or fail to fulfil our moral obligations by disregarding the interests of animals, even if such disregard does not negatively affect any human interests.

But what interests do animals have? Surely an interest in not feeling pain, as well as an interest in having pleasant experiences – we have to acknowledge this unless we doubt or deny the very existence of animal mentality. And in the case of the mentally more evolved animals – vulnerable not only to sensory pain, but also to tedium, intense fear and other varieties of distress – it is proper to speak of an interest in not suffering. Arguably, many of these animals also have a significant interest in continuing to live, in experiencing the future goods of their lives – even if they do not now desire those goods. For no one would think that painlessly killing an animal would always be better for the animal if the killing prevented some future pain in its life.Footnote ¹

Taken in a more stringent sense, the term ‘moral rights’ conveys the idea of deontological constraints – those strong agent-relative reasons for not performing acts of certain kinds. If there is an absolute constraint against ϕ-ing, each one of us has a reason for not ϕ-ing that will be decisive in all possible circumstances. But deontological constraints can be moderate. A moderate constraint against ϕ-ing amounts to a reason for not ϕ-ing that will be outweighed by other reasons (at least, by the reason to promote the overall good) in some possible circumstances – although in most cases the reason for not ϕ-ing will be morally decisive even if ϕ-ing would result in better consequences. If we believe some animals are protected by constraints, either absolute or moderate, we are ascribing a very strong moral status to them. To attend to their interests or well-being, even with the strictest impartiality, will not always ensure the respectful treatment that morality requires.

Given the above senses of ‘moral rights’, our initial question, ‘Do animals have rights?’, splits into two distinct questions:

1. 1. Do animals have moral status?

2. 2. Are animals protected by deontological constraints?

We should also consider a third question:

1. 3. Are human beings protected by deontological constraints?

With reference to these questions, we can now contrast three of the main contenders in the animal rights debate – and point out what makes each of them clash with strongly held moral intuitions.

Typical act-consequentialists, like Peter Singer, answer ‘yes’ to question 1, but ‘no’ to questions 2 and 3. As anyone can learn from any decent textbook on moral philosophy, the act-consequentialist answer to question 3 is deemed too counterintuitive. According to act-consequentialism, it would be permissible, in principle if not in practice, to perform painful and lethal experiments on human beings, without their consent, in order to promote the overall good. It is widely believed that such a view fails to give sufficient moral protection to human beings.

The standard act-consequentialist reply is that common moral intuitions are not to be trusted. Singer, for instance, dismisses them as ‘the biological residue of our evolutionary history’.Footnote ² On this matter I side with Shelly Kagan, who suggests moral intuitions do not deserve much confidence until we find an explanatory foundation for them.Footnote ³

Typical contractualists, either of Hobbesian or Kantian inspiration, answer ‘no’ to questions 1 and 2, but ‘yes’ to question 3.Footnote ⁴ This time the trouble comes from the answer to question 1. Typical contractualists, like Peter Carruthers, argue that any moral significance animals might have is entirely indirect, being grounded in human interests. Carruthers is bluntly clear about the practical implications of this view. ‘[P]ainful methods of factory farming’, he says, ‘and the testing of detergents on animals, would not be ruled out, even granting that the purposes subserved by such activities (cheaper meat, and new varieties of shampoo) are trivial.’Footnote ⁵

This idea that animal suffering is a matter of moral indifference, so that it is often permissible to inflict great suffering on animals in order to get some benefits acknowledged as trivial, is at odds with very common moral intuitions. Thus many – probably most – of us would say that typical contractualism does not give sufficient moral protection to animals. It is worth adding that although contractualists try to accommodate some nearly irresistible intuitions about the proper treatment of animals, they do so in a very counterintuitive way. Consider the torturing of animals in private just for the fun of it – something (almost) everyone agrees is plainly wrong. The contractualist explains the wrongness of such conduct not in terms of the harm done to the animals, but with the claim that torturing animals would degrade our character, making us more dangerous to other human beings.Footnote ⁶ This means that we would have no reason to avoid cruelty to animals if such cruelty did not pose any considerable threat to human beings.

Finally, Reganian deontologists – i.e. those who accept Tom Regan's view on animal rights at least in its essentials – answer ‘yes’ to all three questions. It is their answer to question 2 that is most controversial. According to Regan, all animals that qualify as subjects-of-a-life – a concept defined in terms of a cluster of psychological capacities – are protected by some very strong deontological constraints. These constraints consist in individual rights to life, bodily integrity, and freedom. Since mammals and birds, at least, qualify as subjects-of-a-life, all the animals in this very large class have these rights. ‘Morally’, says Regan, ‘we are never to take the life, invade or injure the body, or limit the freedom of any animal just because we personally or society in general will benefit.’Footnote ⁷ In practice this means, among many other things, that we should abolish immediately all biomedical research involving the use of mammals and birds, no matter how large are the expected benefits of this practice. Killing mice to discover new treatments for cancer, for instance, or using pig organs for xenotransplantation, is morally out of the question.

Faced with such conclusions, most of us conclude that Reganian deontology grants excessive moral protection to animals. Reganians will claim that any moral theory that leaves animals unprotected by deontological constraints will either rest on an arbitrary view of the significance of species, or imply that too many human beings are similarly unprotected. I will show they are wrong.

What about rule-consequentialism? In the answers it yields to questions 1–3, does this view agree with any of the ethical theories just mentioned? I think not. A rule-consequentialist should say, in opposition to the typical contractualist, that animals have moral status. However, unlike the Reganian deontologist, he should deny that we have sufficient grounds to ascribe rights, understood as deontological constraints, to any non-human animals – although he should insist that human beings possess such rights, thus disagreeing also with the typical act-consequentialist. I will now argue that these are indeed the general implications of rule-consequentialism regarding what we owe to animals. If I am right, then, any truly consistent rule-consequentialist will answer questions 1–3 in the most intuitively plausible way.

RULE-CONSEQUENTIALISM

Everyone ought to follow the ideal moral code. This is the code that comprises the optimific rules: those rules which, if generally accepted or internalized, would have the best consequences.

These two statements summarize the most influential version of rule-consequentialism as we find it in the writings of John C. Harsanyi, Richard B. Brandt, Brad Hooker and others. Obviously, they leave open many important questions that generate much more specific versions of the theory. For example, should we have in mind actual or expected consequences? Strictly universal acceptance or something more realistic, like a 90 per cent level of acceptance? Acceptance everywhere or just in the agent's society? And what code should we follow if we find two or more incompatible sets of optimific rules?

Fortunately we do not need to address such questions in order to find out what a rule-consequentialist should think, in a general way, about our moral obligations to animals in the context of the advanced technological societies of our time. But two further points must be made to clarify the theory.

First, the optimific rules should be understood as those which would have the impartially best consequences – and not, say, the best consequences for the agent or for some restricted social group. To use Sidgwick's famous metaphor, alternative moral codes are to be assessed from ‘the point of view of the universe’.

Second, the evaluation of rules must be sensitive to their internalization costs, more precisely to the costs of getting them widely accepted by the human beings of each new generation.Footnote ⁸ These costs are of two main kinds: cognitive and motivational. This means that, ceteris paribus, simpler and less demanding codes are always to be preferred to those which are more complex or involve a stronger opposition to natural human inclinations, so to speak. And it also means, of course, that a moral code might turn out to be much better, on rule-consequentialist grounds, than some more complicated or more motivationally demanding alternative set of rules, even if the latter would result in a greater good after attaining the required level of acceptance.

The version of rule-consequentialism just outlined is the one I will have in mind in the discussion that follows. And I will make two important substantive assumptions concerning the ideal code. Both, I believe, have already been adequately justified.

One of them is about what counts as good or bad consequences of the acceptance of rules, and can be stated in this way: in the assessment of the intrinsic value of any state of affairs, total well-being is at least as important as any other factor or set of factors, like environmental integrity or fairness in the distribution of well-being. I will assume, in other words, that the rule-consequentialist evaluates rules from a standpoint that is either utilitarian or not very distant from the utilitarian standard of goodness. Otherwise he will be endorsing a very implausible theory of the good.

My second assumption – which will receive some support in the course of this discussion – is that the ideal code is fully deontological. Instead of being practically equivalent to the principle of act-consequentialism or to some form of ethical egoism, for instance, it contains all the main ingredients of a deontological ethics. According to the ideal code, then, we have a duty of beneficence (i.e. a reason to promote the good). But this duty can be outweighed by deontological constraints (i.e. agent-centred reasons against harming and deceiving, at least), presumably of the moderate kind, as well as by special obligations (e.g. the duty of parents to educate their children). Even when promoting the good would not involve any violation of deontological constraints or special obligations, however, we are not always morally required to bring about the best consequences. The duty of beneficence is not extremely demanding – although it might be more demanding than common sense morality leads us to assume.

Finally, I must clarify an important point about the proper way to try to figure out what the ideal code says. It might seem that the rule-consequentialist (or anyone who wants to discover which set of principles is supported by the theory) has to proceed more or less as follows:

1. 1. Identify the most plausible candidates for the ideal code, C₁, C₂,. . ., C_n.

2. 2. Estimate the value of the consequences of the general acceptance of C₁, C₂,. . ., C_n.

3. 3. Compare the value of the consequences of C₁, C₂,. . ., C_n – and choose the code with the highest value.

The last stage of this procedure does not seem epistemically daunting, but no one would be able to reach it. For how could we select, from the infinite set of all possible moral codes, the most plausible candidates for the ideal code? How could we get a complete formulation of each one of them? And how could we foresee the morally relevant consequences of a whole series of entire codes? Fortunately, in order to gain knowledge of the optimific principles we do not have to formulate the ideal code in its entirety, and then compare its relevant consequences with the relevant consequences of other fully formulated moral codes. We can follow – and we should do so, since there is no real alternative – a piecemeal approach in the attempt to identify the true optimific principles. A rough characterization of such an approach can be given in terms of two methodological guidelines.

The first is a prima facie rule of increasing specificity: begin by determining the bare bones of the ideal code, and then advance gradually to its more specific features. For example, begin by determining whether the ideal code is equivalent to act-consequentialism, and then – having found compelling reasons to deny such equivalence – try to discover with increasing detail which limits to maximizing the good would have the best consequences if generally accepted, without forgetting to consider their internalization costs. Is a deontological constraint against harming among such limits? If so, does it consist in a prohibition against intending harm? If so, does it allow us to intend harm in self-defence? In what circumstances?

According to the second guideline, we should follow this procedure or some variation of it:

1. 1. Consider two codes, C₁ and C₂, which (a) have all the already identified features of the ideal code, and (b) differ just with respect to feature F, such that only C₁ has F.

2. 2. Focusing on that difference, estimate the value of the consequences of the general acceptance of C₁ and C_2.

The assessment of consequences requires careful reflection supported by empirical evidence. If we conclude that the best code is C₁, that will be a reason to believe that the ideal code has F; if we arrive at the opposite conclusion, we should take it as a reason to think that the ideal code does not have F.Footnote ⁹ It will not always be easy – and sometimes it might be too hard – to discover which of the contrasted codes would have the best consequences, but an extreme pessimism concerning the prospects of gaining knowledge of the optimific principles through this piecemeal approach would be unwarranted.

THE CASE FOR MORAL OBLIGATIONS TO ANIMALS

The rule-consequentialist case for the existence of moral obligations to animals is quite straightforward. We are assuming that total well-being is at least a very important factor in the evaluation of states of affairs. But whose well-being counts? Since we are supposed to adopt the point of view of the universe, we have absolutely no grounds either to restrict relevant well-being to human well-being or to give it more weight just because it is human well-being. Speciesism is ruled out at this level.

Of course, we can say that human deaths are typically much worse than the deaths of animals because they involve a greater deprivation of goods. And we can add that only in human lives is there an opportunity for the realization of some especially valuable states of well-being, e.g. Millian higher pleasures. But in a truly impartial assessment of total well-being the interests of the members of our species will count as much as the similar interests (some of which are quite strong) of the members of any other species. Therefore, in evaluating alternative moral principles, the rule-consequentialist has to take into consideration the impact of their general acceptance on the well-being of animals without discriminating among them on the basis of species.

Two general empirical facts must now be pointed out. First, our interaction with animals is pervasive, and we have the power to affect their well-being in many significant ways. Second, we are often in a position to gain considerable benefits from actions and practices that impose great harms on animals, sometimes to an unimaginable scale.

With these facts in mind, let us now follow the procedure above outlined and consider two contrasting codes, C₁ and C₂. While C₁ includes a prima facie duty not to harm animals, in C₂ we do not find any such duty. Regarding the proper treatment of animals, C₂ just forbids the kind of strict cruelty – torturing animals for the fun of it – that Carruthers has in mind. To which of these codes should the rule-consequentialist give his preference? Even if C₁ has higher internalization costs, it seems quite clear that it is much better than C₂. Given our natural empathy, it is not very difficult for us to develop a strong aversion both to harming animals with our own actions and, in a more general way, to actions and practices that harm them. Internalizing a principle against harming animals is not unreasonably demanding, something at odds with human nature – and a code with such principle, like C₁, would result in a much greater total well-being after being widely internalized. For under C₂ everyone would feel free to impose disproportionally great harms on animals in the pursuit of self-interest, without taking minimum care to lessen those harms, and this attitude would be highly detrimental to general well-being.

How do we know this? Well, to a large extent we are living under a code like C₂, and so we can see its results in practices like factory farming, the testing of cosmetics on animals, or the trapping of animals for their fur to make luxury items. From the point of view of the universe, the world would be definitely much better without these practices.

We may conclude, then, that the ideal code accords moral status to animals.Footnote ¹⁰ At the very least, it includes a prima facie duty not to harm animals. By a similar reasoning, we can support the claim that a prima facie duty to help animals is among the optimific principles. But since helping is generally harder than simply not harming, a due consideration of the internalization costs will presumably lead to the conclusion that the duty to help animals is much weaker (i.e. more easily overridable by other factors) than the duty not to harm them.

I must add that the duty to help does not generate, for instance, an obligation to save antelopes from being killed by cheetahs. Quite the opposite. For saving the antelope would amount to harming the cheetah. And even if the harm of losing a vitally precious meal were less serious than the harm of dying, it would be wrong, in virtue of the greater strength of the duty not to harm, to harm the cheetah in order to prevent the antelope from being harmed. But when an animal (even a wild one) is in distress because of a natural disaster, a crippling disease or a human-made hazard, the duty to help might retain its force.

To what amounts the principle against harming animals that the rule-consequentialist should endorse? Without doubt, to a reason for not inflicting immediate harm to easily identifiable animals. But also, I believe, to a reason for not supporting those practices where animals are systematically harmed without real necessity, even if by supporting them we would not cause any immediate harm to some definite animal. For the general acknowledgement of a reason for not supporting such practices would have much better consequences than the internalization of a duty of non-maleficence narrowed to immediate harms to easily identifiable animals. Hence, a rule-consequentialist should avoid the consumption of factory-farmed products or of cosmetics tested on animals, among many other things. As a matter of fact, in the current circumstances the seemingly modest prima facie duty not to harm animals might be quite demanding. But these costs of living up to the optimific principles – costs that result from inhabiting a world where most agents do not really accept those principles – are not relevant, according to rule-consequentialism, to determine which of the possible moral codes we should follow.

THE CASE AGAINST ANIMAL RIGHTS

For convenience, from now on I will use the term ‘rights’ only for deontological constraints. According to the optimific principles, I am assuming, human beings have rights that protect them from being harmed in certain ways, e.g. a right to life and a right to bodily and psychological integrity. To determine whether or not the ideal code extends such rights to members of other species, we have to identify the reasons a rule-consequentialist might have for ascribing rights to human beings, and then to consider if those reasons are also reasons for believing in animal rights.

As Harsanyi has cogently argued, moral rights find a prominent place in the ideal code primarily in virtue of their ‘indirect effects on people's expectations and incentives’.Footnote ¹¹ By the expectation effect of a principle, Harsanyi means the effect of its acceptance on the ‘agent's expectations, on their very ability to form definite expectations, and on their feelings of confidence and security’; the incentive effect of a principle, in its turn, amounts to the effect of its adoption on the ‘agent's incentives to engage in various types of social beneficial behaviour’.Footnote ¹²

It is quite clear that these effects support, on rule-consequentialist grounds, a preference for a moral code with very strong (even if not absolute) deontological constraints. Let us try to imagine, for instance, a society where everyone (or nearly everyone) accepted a requirement of impartial beneficence unconstrained by any right to property. In such a society each person would always be subjected to being deprived of his or her property whenever that was deemed necessary to increase total well-being. Too many would be unable to form definite expectations about their future wealth, and thus would lack a proper incentive to work. Widespread poverty would be the foreseeable result. Or suppose that no right to life was accepted. Then each person would have to live under the dreadful expectation of being killed by anyone who believed, sometimes even without an adequate justification, that his or her death was necessary to promote the overall good. Widespread fear and lack of confidence would be the foreseeable result. As Harsanyi remarks, ‘it is our moral duty not to take the lives of innocent people, even in order to achieve some very noble objective – because in the long run society will be better off if innocent people need not fear being murdered for the purpose of achieving some very noble objectives’.Footnote ¹³

At this point it could be objected that since a requirement of impartial beneficence had been generally internalized by the members of that society, in fact nobody (or almost nobody) would be deeply disturbed by the prospect of losing his property, or even his life, for the sake of the greater overall good. However, we must not forget that, in any moral code worthy of attention by rule-consequentialist standards, the duty of beneficence will not be extremely demanding. The ideal code has agent-centred options or prerogatives.Footnote ¹⁴ According to rule-consequentialism, then, each agent is morally allowed not to promote the overall good if bringing about the best consequences is too costly for him or her. In other words, the ideal code allows a considerable partiality towards ourselves and our loved ones. Now, as it is well known, a moral code with prerogatives but no deontological constraints is highly counterintuitive in its practical consequences.Footnote ¹⁵ But the ideal code does not belong to this kind. As can now be seen more clearly, moral rights are indeed among the optimific principles. For we must assume that in our imagined society the accepted duty of beneficence, being limited by prerogatives, is not extremely demanding. But then, because each person is allowed to be partial – and nearly all will be partial – to a considerable extent, nearly everybody would be really disturbed by the prospect of being deprived of his life, or even just of his property, for the sake of the greater overall good. Furthermore, given the absence of constraints, in virtue of that same partiality nearly everyone would feel too free to harm or deceive others not for the sake of the greater overall good, but just for their own good or for the good of their loved ones. In a society ruled by a moral code without constraints (but with prerogatives), therefore, each person would have to live under the permanent threat of being seriously harmed or deceived by agents moved not by an impartial concern with the overall good, but by the not so noble motive of self-interest.

In a due consideration of the expectations and incentives resulting from the general acceptance of principles, then, the rule-consequentialist finds a decisive reason to endorse a moral code where several rights are accorded to each moral agent. But has he found any reason to extend those rights to animals? Not at all. For suppose such extension is not generally accepted. As a matter of fact, even the most intelligent animals lack the required capacities to understand which ethical principles have been widely internalized by the moral agents of a given society. It is thus quite obvious that they will not form negative expectations in virtue of not being protected by deontological constraints, i.e. their ‘feelings of confidence and security’ will not be significantly affected by the fact that moral agents do not accord them rights (although they care about their well-being). No animal can be truly aware of this fact. No animal can think anything like ‘My God, I am living in a world where human agents do not accept animal rights! They just accept a prima facie duty not to harm us!’, and so no animal can be distressed by any such thought.

For the same reason, and even more obviously, a general non-acceptance of animal rights would not be detrimental to animals’ ‘incentives to engage in various types of social beneficial behaviour’. Hence, from a rule-consequentialist perspective, we are left with no reason to extend the scope of deontological constraints to members of other species.

This line of reasoning, however, exposes an important problem for rule-consequentialism. After all, many human beings also lack the required capacities to develop negative expectations, or to lose incentives to act in a socially beneficial way, in virtue of not being accorded moral rights. Does the ideal code, then, restrict rights to those human beings who are moral agents?

To answer this question, let us examine again two contrasting codes: according to C₁, only moral agents have rights; C₂ extends certain basic rights at least to a much larger class of human beings.Footnote ¹⁶ To determine which of these codes would be better if generally accepted, it is very instructive to consider what Carruthers has to say in his attempt to justify the claim that rational contractors would favour a moral code that accorded a strong moral status to all (or nearly all) human beings, and not just to themselves. His argument from social stability runs as follows:

Foreseeing this, a rule-consequentialist will prefer C₂ over C₁. Not only would a code that restricted the possession of rights to moral agents pose a serious threat to social stability, but it would also have higher costs of internalization than a code that extended rights to many other human beings. Carruthers would agree. As he remarks, ‘if anything counts as part of “human nature”. . ., then people's deep attachment to their infants and aged relatives surely belongs within it’.Footnote ¹⁸ In virtue of our natural tendency to develop this kind of attachment, getting a code like C₁ widely internalized would require extensive brainwashing. And let us not forget that by suppressing the development of that tendency we would thereby make things go worse in some ways, since a very important source of well-being lies in deep relations with human beings that lack moral agency.Footnote ¹⁹

As Carruthers also makes clear, social stability provides no argument for extending rights to animals – and the rule-consequentialist should agree with him on this point.Footnote ²⁰ As a matter of fact, we are all living in societies where the vast majority of agents do not accept animal rights, and this does not endanger social stability. Although many people care significantly about their pets, this attachment (which anyway is generally much weaker than the attachment to relatives and other human beings) surely does not count as part of human nature.Footnote ²¹ Hence, internalization costs are no objection to a moral code that does not accord rights to animals – quite the contrary.

Besides, since it is reasonable to assume that the ideal code gives property rights to moral agents, its general acceptance would ensure that pets enjoyed a high level of protection in virtue of being owned by those who cared about them. Thus it would be pointless to accord rights to pets in order to prevent their owners from endangering social stability in an attempt to defend them.

Those who believe that animals are under the scope of deontological constraints, let us concede, might be willing to take extreme measures, even to resort to violence, in order to abolish all those practices that in their view involve a violation of animal rights. In a society with many animal rights advocates, then, social stability could be seriously compromised if those rights were not also acknowledged by many others. But this kind of possible threat to social stability is utterly irrelevant. Since the ideal code is the set of principles that would maximize the good if generally accepted, to identify those principles we should only consider general acceptance scenarios. What would happen in a mixed social scenario where a large proportion of agents accepted animal rights, and a large proportion of agents did not acknowledge them, simply does not matter in a rule-consequentialist assessment of moral codes.

I have to conclude that there is no good rule-consequentialist case for animal rights. Although rule-consequentialists should evaluate principles in the light of a non-speciesist theory of value, it turns out that the ideal code embodies a speciesist view on what concerns the possession of moral rights. No animals have rights, but even human beings who do not qualify as Reganian subjects-of-a-life (e.g. infants and the severely demented) fall under the scope of deontological constraints.

What we have here, of course, is not bare speciesism, i.e. the claim that some human beings have a higher moral status than the members of other species simply because they are human. Discrimination on the basis of species is justified by the higher-order principle that we ought to live according to the optimific principles. On rule-consequentialism, the fundamental wrong-making property is the property of being disallowed by these principles – and it is this property, not the property of being human, that grounds the moral significance of species. This principled speciesism, as we may call it, effectively blocks Singer's famous analogy between speciesism and the bluntly unjust forms of discrimination of racism and sexism, since it is quite clear that the rule-consequentialist's reasons to accept deontological constraints do not provide any basis to limit their scope on the grounds of race or sex.Footnote ²²

It is worth adding that rule-consequentialists can consistently support the introduction of legal rights for some animals as a means to ensure that, in the world as we now know it, their well-being is better safeguarded. This seems to make sense especially in the case of those animals that, in virtue of their higher mental capacities, tend to have stronger interests. Arguably, self-conscious animals, such as great apes, have a much stronger interest in continuing to live, and killing them means imposing a serious harm. Protecting their lives by introducing a legal right to life for them might then be a good idea in a world where too many people do not wish to comply with their duty not to harm animals.

It is also worth pointing out that, in the current circumstances, the disagreement over the moral significance of animal interests has a much greater practical import than the further question over whether they have rights. But this question is far from being practically innocuous. Reganian deontologists advocate the abolition not only of animal experimentation in scientific and medical research, but also of the use of animal products in medicine. Rule-consequentialists have to disagree with them on this important matter. Of course, unlike the typical contractualists, they will support a ‘3Rs’ policy of reducing, refining and replacing animal testing or other measures designed to safeguard animal well-being. But at the same time they will approve the use of animals in science and medicine when there are strong enough reasons to believe that using them is necessary to promote the overall good.

Consider the case of xenotransplantation. To a Reganian deontologist, breeding genetically modified pigs to harvest their organs for transplantation in order to save human lives is morally out of the question. The rule-consequentialist will reject this deeply counterintuitive judgement. His theory also clearly agrees with common moral intuitions – to change the subject just a little – as regards the view that it is not wrong to kill mammals or birds for food when the alternative to not killing and eating them is to die from starvation or become malnourished. The Reganian deontologist, however, will be unable to explain how such a thing can be permissible. For if it is wrong to take an organ from a pig in order to avoid death by disease, how can it be right to do pretty much the same in order to avoid death by starvation?

SPECIAL OBLIGATIONS TO ANIMALS

Although there is no place for animal rights in the ideal code, this does not mean that a rule-consequentialist has to think that our moral duties to animals are exhausted by general prima facie obligations to promote their well-being, either by not harming or by helping them. As it has been widely acknowledged, rule-consequentialism provides a firm foundation for special obligations of several kinds.Footnote ²³ There are good reasons to believe that among them we will find special obligations to animals.

Consider the principle according to which those who own animals or have animals under their care ought to provide for their animals’ needs. The general acceptance of this principle would have the obviously good result of nearly eliminating gross negligence to animals – and also of reducing negligence to human beings, given the plausible assumption that there is a significant correlation between the two forms of negligence. Besides, we should bear in mind that the ideal code, in what concerns our relations to animals, is a code with prerogatives but no constraints. In the absence of special obligations to animals, then, we would be too free to harm animals for personal benefit.

McMAHAN'S BENIGN CARVIVORISM – AND BULLFIGHTS

It is now pretty clear that rule-consequentialism is intuitively very appealing in its general implications regarding what we owe to animals. If – as I believe to be the case – an ethical theory becomes more plausible in virtue of cohering well with our ‘considered judgements’, this constitutes a significant reason (albeit an overridable one) for accepting rule-consequentialism at least over the other theories referred to above. This will only be of little significance if someone manages to show that yet another ethical theory, genuinely different from rule-consequentialism, accounts better for the intuitions at stake.

Quite unsurprisingly, however, in some important cases it will not be easy to discover what the ideal code says about what may be done to animals. I will now illustrate this point with a brief discussion of the practice of benign carnivorism, as Jeff McMahan calls it.Footnote ²⁴

Benign carnivorism is quite unlike factory farming. Although in both practices mammals and birds are raised for food, not for meeting any vital human need that could not be met without using animals, but just for satisfying our dietary preferences, in benign carnivorism they live happy and reasonably long lives before being painlessly killed for consumption. They will be replaced by other animals. And each one of them owes their very existence to the practice. Benign carnivorism ‘thus yields a continuous bounty of contented animals and contented diners’.Footnote ²⁵ So what can be wrong with it?

A Reganian deontologist, of course, would say that the whole practice is wrong because it involves a systematic violation of animals’ right to life. For her, benign carnivorism is not more acceptable than a practice of benign slavery, since keeping slaves in ‘humane conditions’, so to speak, would never remove the inherent injustice of slavery. McMahan is not willing to accept animal rights, but nevertheless he argues that benign carnivorism is wrong, only conceding that it could be permissible in some far-fetched circumstances. His argument rests on an assessment of the interests at stake. Although animals are killed in a painless way, killing them is still against their interest in continuing to live. And this interest, even if it is normally much weaker than our interest in continuing to live, is still much stronger than our interest in eating meat. What each animal loses by dying is not even closely compensated by the human gains, i.e. by the additional pleasure several human beings get from eating meat a few times instead of having meals prepared only with vegetables (or with lower animals, like insects, clams or shrimp). Given this assessment, what can be said in favour of benign carnivorism? McMahan considers and dismisses an appeal to ‘impersonal reasons’:

From a rule-consequentialist perspective, however, the acknowledgement of such a reason is far from being surprising. For the ideal code consists of a set of principles tailored, at least to a great extent, for the promotion of total well-being – and causing ‘animals to exist with lives that would be good for them’ might increase total well-being to a very considerable extent.

So what should a rule-consequentialist think about benign carnivorism? Since he accepts a prima facie duty not to harm animals, he surely will take the fact that in this practice animals are killed against their interests as a reason to be against it. And, let us assume, he will agree that the human interests at stake do not provide a strong enough reason for overriding the duty not to harm animals. But on his view, the fact that benign carnivorism makes possible the existence of new animals with happy and reasonably long lives has to constitute a reason for supporting the practice (even if it does not generate a duty to support it) – and this reason seems strong enough for overriding the duty not to harm animals. The balance of reasons thus seems to favour benign carnivorism.

To make this point clearer, let us consider two contrasting codes one last time. According to C₁, the benefits that future animals would enjoy cannot override the duty not to harm animals, but in C₂ this prima facie duty can be overridden by such benefits. Which of these codes, if generally accepted, would result in a greater total well-being? Finding no reason to doubt that C₂ is the obvious answer, I conclude that, according to the ideal code, the fact that a practice will result in new animals with good lives counts as a reason that can override the duty not to kill animals against their interests. Benign carnivorism, therefore, seems not to be in conflict with the optimific principles.

But the matter is not so simple. Benign carnivorism bears a striking similarity to a practice that, to my knowledge, meets universal opposition even from moderate animal welfare groups: bullfighting. Before their journey into the arena, fighting bulls live several years in quite pleasant conditions. They are replaced by other fighting bulls. And each one of them owes its very existence to bullfighting, since they are bred for the purpose of this traditional entertainment. This practice thus seems to yield a continuous bounty of contented bulls and contented bullfighters and spectators. Obviously, there is an important difference between benign carnivorism and bullfighting: fighting bulls die painful deaths, or at least are subjected to terrible suffering before finally being killed. Nevertheless, this period of suffering amounts to a very small part of the bull's whole existence, and so it can be said that a fighting bull generally has an overall good life.

In the face of these facts, should the rule-consequentialist adopt a permissive stance towards bullfighting? Of course not. For those who participate in this practice, even if solely as joyous spectators, will be unable to take animal interests seriously enough. They certainly can remain very gentle to their own pets or feel outraged by the mistreatment of certain kinds of animals, but enjoying bullfights requires an attitude of indifference (at best) towards blatant animal suffering that is psychologically incompatible with maintaining any impartial concern with animal well-being. Those who doubt this empirical claim are invited to find a single bullfighting enthusiast who, for instance, avoids the consumption of factory-farmed products for the sake of animals. Even if in most cases going to bullfights is not what renders someone too indifferent to animal suffering, enjoying them only reinforces that attitude. And even if the experience of attending a bullfight makes some people feel outraged by what they see, and ultimately leads them to give more importance to animal well-being, this does not change the fact that the whole practice of bullfighting can only exist with the ample support of those who actually enjoy it.

Since the continued existence of bullfighting rests on an attitude towards animal suffering that makes agents unable to fulfil their duties to animals, this practice has to be condemned by the ideal code. Does benign carnivorism deserve the same fate? Would the consumption of meat in the context of this practice hinder our capacity to take animal interests seriously enough? I just do not know the answer to this empirical question. Maybe no one really knows. Thus I cannot reach a definite conclusion on how a rule-consequentialist should evaluate benign carnivorism.

There are, however, additional grounds for opposing this practice, which must be taken into account. In most circumstances, raising animals in really decent conditions, and allowing them to live for a great part of their natural lifespan in those conditions, would require such an amount of natural resources that widespread benign carnivorism would be no less environmentally devastating than factory farming. Furthermore, benign carnivorism would impose great costs on meat producers. The aim of profit-making, especially in a competitive economic system, would thus translate into constant pressure for relaxing the very demanding animal welfare standards of this practice – a pressure for killing animals earlier, let us say, or for not providing them with so much space. Given this economic pressure, a stable practice of benign carnivorism would be very hard to achieve even in a world where most people had a strong motivation for caring about animal well-being. In light of these facts, rule-consequentialists should view benign carnivorism with utter suspicion.Footnote ²⁷