Innocence and Consequentialism: Inconsistency, Equivocation and Contradiction in the Philosophy of Peter Singer
  • : Function ereg() is deprecated in /home/sites/ on line 647.
  • : Function ereg() is deprecated in /home/sites/ on line 647.
  • : Function ereg() is deprecated in /home/sites/ on line 647.

Innocence and Consequentialism: Inconsistency, Equivocation and Contradiction in the Philosophy of Peter Singer

Jacqueline A. Laing*

In Human Lives: Critical Essays on Consequentialist Bioethics, eds. David S. Oderberg and Jacqueline A. Laing (Basingstoke: Macmillan, 1997), pp. 196-224.

Dr Jacqueline Laing

'Ubi innocens formidat damnat iudicem' - Sententiae No. 7091

The dictum of Pubilius Syrus expresses a central idea of traditional morality. When we say an action is gravely wrong because it betrays the innocent, we assume that morality addresses, at least in part, questions of justice, and further that these questions of justice are inextricably linked with the concept of innocence. The concept of innocence however is as broad as it is multifaceted. An agent may be innocent (of an act, say) or she may be an innocent (a child, perhaps). Condemning an innocent human being is ordinarily considered an evil and abusing or harming the innocent is regarded as a disgraceful thing. Innocence is often seen as finding its opposite in guilt, so that the term necessarily involves a determination of an agent's liability. But in fact the concept extends far beyond questions of liability and culpability into realms that are not ordinarily discussed by contemporary moral philosophers.

Traditional morality has always regarded the concept of innocence as central to right conduct. Indeed many principles of traditional morality are precisely principles which safeguard the innocent. Accordingly, it is traditionally taught that where innocence is concerned, the desires of third parties must be reasonable. Threats of violence or destruction if an innocent is not executed cannot alter the wrongness of killing that innocent. Again, according to traditional morality, it is not merely those who have the actual and actuated characteristics of the mature human being who are properly accorded respect and protection. Traditional morality has always acknowledged the need to protect the vulnerable, for example the sleeping, the comatose and the very young, as well as healthy mature human beings.

Contemporary utilitarians deny these principles of traditional morality. Talk of innocence is all but non-existent in their work. One influential commentator writing on the subject of bioethics in this vein is Professor Peter Singer. Among his recommendations for society are infanticide of the severely disabled (or in his own words 'euthanasia for defective infants') and the destructive use of human embryos for scientific experiment. To arrive at these conclusions Singer takes seriously the proposition that the stuff of ethics is precisely the maximization of good consequences. He also promotes the doctrine of personism,2 the view that a human may have more or less of a claim to life according as he or she displays certain defined characteristics (like rationality and self-consciousness). Furthermore, he insists that the desires of third parties may affect a person's claim to life. But the principles of traditional morality which safeguard the innocent present Singer with cases which threaten the logic of his arguments. Equivocations, inconsistencies and contradictions emerge in his writing. My aim here is to point out in his work certain instances of these kinds of irrationality.

Innocence Threatened Or Singer Misunderstood? Proposals for 'Non-Voluntary' Killing

Peter Singer's espousal of the cause of what he calls 'non-voluntary euthanasia' is well known in philosophical circles. He was the longtime director for the Centre for Human Bioethics, a government-funded organization connected to Monash University in Victoria, Australia, and is the author of several influential books and papers on human bioethics.3 Whereas some consider his notions respectable, disabled people in some European countries have disrupted his lectures. On occasion, his seminars have been cancelled owing to protest. At other times he has been disrupted or shouted down by disabled people in the audience who presumably felt threatened by his proposals. There is considerable public revulsion over his repudiation of long-accepted norms. However, Singer himself believes he has been misunderstood.4 On the one hand he characterizes the protests by the disabled at his lectures as a peculiarly German thing, even comparing them with the growing tide of German Nazism in the latter days of the Weimar Republic.5 On the other, he claims that the disabled have nothing to fear from his proposals. He also suggests that those who boycotted his lectures were in fact members of a small group of feminists and left-wingers. Whether the reaction he encountered was the result of dictatorial thinking, or the consequence of genuine fear of discrimination, together with a collective conscience sensitized by relatively recent experience of totalitarian attacks on the innocent, is itself an intriguing question. These issues are not, however, the subject of this paper. What is of concern here is first, whether Singer has been misinterpreted and secondly, whether his arguments are self-contradictory, equivocal or inconsistent precisely where rationality is crucial. Singer's recommendations are self-confessedly practical. They counsel the killing of innocent human beings. It is therefore important to be clear about the foundations of his scheme and to understand the principles he applies to arrive at his revolutionary conclusions. The common notion that it is wrong to attack the innocent seems to make no appearance in his work. I want to show to what extent various principles of traditional morality, explicitly rejected by him, must explain certain of his positions. If flaws in his argument are apparent, we may want to think twice before embracing his justification of the killing of the very young, the disabled and the very old. Let us see how this is so.

The Language of Utility and Respect for One Another


It is worthwhile making some preliminary observations about the very language Singer uses. The first edition of Practical Ethics is replete with references to 'defective infants' and 'euthanasia for defective infants'.6 In the second edition Singer becomes coy about using the adjective 'defective' and systematically drops the reference to 'defective' infants'. Instead he talks about 'Life and Death Decisions for Disabled Infants'7 and 'disabled infants' more generally. He is right to make this modification. The very language of the first edition (prior to his experience of the boycott of his lectures and publications by disabled people) indicates how we are asked to regard the vulnerable. Disabled infants are indeed doubly vulnerable where Singer's proposals are concerned: first in virtue of their youth, and second in virtue of their disability. We ordinarily speak of goods, products, artefacts and material objects as defective if they are not fit for their purpose. Singer was initially content to apply this language to infants who are handicapped or disabled. The very use of this terminology is repugnant. Ordinary language, it could be pointed out, does not apply the language of products to humanity. There is, however, a serious objection to this kind of response to Singer. It is not enough to demonstrate that ordinary language does not support the usage Singer would advocate. The fact that certain language is not 'ordinary' is no argument against its rightful use. Discourse that is unacceptable at one time might be acceptable in another. Furthermore, were the language of products liability to become the language of bioethics, for example, it would still be open to doubt that such use was appropriate. If we are examining substantive moral issues, then, we will need to do better than to appeal to ordinary language. In any case, in the interests of a fair and faithful interpretation of Singer's approach, it will be necessary to have recourse to the terms he actually uses - language which we ourselves mights balk at using. Moreover, it might well appear repugnant or disgraceful that the powerful in society should turn on the most vulnerable and needy by defining them out of the moral universe or by, quite simply, attacking them. That fact alone is however no argument against such attacks and their justifications. Philosophy is not built on the tastefulness of propositions. What appears repugnant to me might seem virtuous or even saintly to you. Gut reactions need to be articulated in rational terms before they can be understood. We need only note for the time being that singer, in the first edition and prior to his experience in Europe, applies the word 'defective' not simply to products and objects but to infants with disabilities.

Persons and persons*

Singer also applies the term 'person' in a way that is at odds with ordinary language. For him, a person is anything that answers to the description 'rational and self-conscious'.8 What of autonomy (the capacity to choose)? On this he is unclear. He does not reject the idea that autonomy is part of the concept of personhood but denies that it has any bearing on the wrongness of killing. 'The classical utilitarian might have to accept that in some cases it would be right to kill a person who does not choose to die on the grounds that the person will otherwise lead a miserable life'.9 Again 'preference utilitarians ... must allow that a desire to go on living can be outweighed by other desires'.10 He also holds that it might be better to encourage the belief that autonomy matters even though it does not: 'This is true [that autonomy does not matter] only on the critical level of moral reasoning ... utilitarians my encourage people to adopt, in their daily lives, principles that will in almost all cases lead to better consequences ... The principle of respect for autonomy would be a prime example of such a principle.'11 So Singer's favoured morally significant description of a person, the description which is the criterion when it comes to killing, seems to be a 'rational and self-conscious' being. Much could be made of the arbitrariness of his chosen definition of a 'person', his lack of interest in qualities like humour, musicality, artistry and spirituality. But I do not propose to enter into a debate about the meaning of 'person'. I will assume Singer's definition for the purposes of argument.

Singer's use of the term is still at best technical, understood largely only by philosophers familiar with his work. Ordinary language, however, does not place use of the term 'person' under the strictures that Singer employs. In ordinary parlance it is roughly equivalent to 'human being'. And in our ordinary dealing with one another we do not detach moral significance from human beings and attach it solely to 'persons' (in Singer's sense). Nevertheless, that it is a technical term in contemporary philosophy is not yet any reason to reject it outright. In order to mark the difference between ordinary language and the use urged by Singer I shall speak of persons (as we ordinarily understand them) and persons* (as Singer understands the term).

Rights-Based Feminist Arguments Distinguished

There is a distinction between rights-based arguments for the killing of human beings on the one hand and consequentialist and personist arguments on the other. Among rights-based arguments I count those which analyse the rightness or otherwise of killing in terms of the threat that the proposed victim poses to another human being. Judith Jarvis Thomson,12 for instance, uses a memorable example to argue by analogy that it is not always obligatory to refrain from killing others. She asks us to consider someone, A, who wakes up to discover himself in a hospital connected to an unconscious man, B, in the nearby bed. The man B, it transpires, is a famous violinist with a kidney disease. His survival depends upon his circulatory system's being plugged into the system of another of the same blood type. A society of music lovers has kidnapped A because he is the only one with the requisite blood type. If A chooses to be disconnected from the violinist B, B will certainly die. If B remains connected for nine months, he will have recovered and will be able to be unplugged without injury. Thomson argues that A is not obliged to act as a life support system. Such a choice would be supererogatory, or beyond the call of duty. She uses the analogy as an argument in favour of abortion. Her argument is directed to the special position of the woman qua life support system and mother of the gestating infant. The example stresses the threat posed by the unborn to the mother, upon whom the former is altogether dependent. Unlike Singer, Thomson accepts a system of rights and duties which allows for justification and excuses on the basis of the threat that is posed by the putative victim. It is non-consequentialist in that it allows that an agent may act for less than the best outcome. It operates, for all intents and purposes, on non-personist assumptions in that it does not rest on the idea that a human being becomes more or less significant or valuable according to the degree to which she satisfies the description of a person*. I do not propose to discuss rights-based defences of abortion here. Suffice it to say that these defences are entirely different from Singer's in that they do not repudiate the claims of innocence. On the contrary, a theory like Thomson's gives content to the notion of innocence by elucidating its meaning. For her, a human being who is a threat to another (in the case of abortion, her mother) and dependent on that other, might cease to be an innocent properly understood. In such a case, according to Thomson, it would be permissible for the threatened party to retaliate against the aggressor. More debates would, given these assumptions, centre on what it is to be an innocent and what constitutes a threat or form of aggression.13

Theories of Culpability Distinguished

Altogether different again from Singer's defence of killing are theories which analyse the culpability of human behaviour. We often say that some killings are not murderous because the person who causes the death in question is not acting in the fullest sense of that term. When a person is thought to be acting involuntarily, as an automaton, or in her sleep, or even under the influence of drugs or alcohol, there are often thought to be grounds for exculpation or reduction in blame because the accused person in question is not altogether in control. Debates in these contexts may centre around the question of the degree to which a person is thought to lack control when she deliberately drinks or takes drugs, and then kills someone. What is not doubted in these contexts is that there is a class of cases in which it would clearly be wrong to blame or punish someone for causing death. A person who through no fault of his own falls from a great height onto an innocent below and kills him ought not be blamed or punished for the resulting death of the innocent. Again, one who has an epileptic fit whilst holding a knife slicing his bread cannot be held responsible for the death of the innocent who is stabbed in the neck whilst trying to assist him. Loss of control explains why it is wrong to blame these people who cause death. There are different ways in which a person may lose control. But in all these kinds of cases the killing is not justified but excused. here it is not the victim who is the problem (there is no question of the victim 'having a life not worth living') but the kind of action or behaviour that causes the victim's death. Consequentialist and personist arguments are not of this kind. They are not assessments of culpability based on the degree to which the agent of the resulting death may be thought to be 'in control' of his or her actions.

Personism and Consequentialism: Principles Rejected by Singer

Among the concepts and principles that Singer rejects as having any relevance to the morality of action are these:

(i) Potentiality. According to traditional morality, the potentiality of a thing or species is a matter of moral significance. A noteworthy feature of Singer's ideology is his actualism. For him the only thing that matters ethically are occurrent properties. A favourite technique of Singer's is to compare infants with animals like snails, and the intellectually disabled with more ntelligent animals. He points out that if it is rationality and self-consciousness that matter to the rightness or wrongness of killing, then there is nothing except unwarranted bias ('speciesism') to justify our favourable treatment of these disabled humans. Accordingly, a day-old infant should, ceteris paribus (e.g. in the absence of parents who desire her existence) be considered as having the same kind of claim to life as a non-human animal with like characteristics. A human embryo, needless to say, would have very few claims indeed. Its claims would be the same as those of any animal with the same occurrent properties. Its parents presumably would be wrong in thinking they had any special obligations (qua parents) in respect of it.

(ii) Commonness of kind. Singer uses the term 'speciesist' to condemn as biased one who discriminates in favour of human beings. There are two aspects of his argument. First, he rejects moral generalizations on the basis of class or species. Secondly, he sees no moral relevance in origins. He appears to reject the idea that special obligations flow from parenthood or common ancestry. I do no propose an elaborate discussion of the 'speciesism' point within the confines of this paper. It is an independent topic and deserves analysis in its own right, but some important allied points will be made below (5.2).

(iii) The resonableness of third-party desire. When an innocent victim is concerned, the desires of third parties must be reasonable. The fact that parents desire the abuse of their young infants (too young to know what is happening to them and powerless to object) in no way alters the fact that the desires are unreasonable and the abuse wrong. Only reasonable third-party desires have any moral significance where the innocent are concerned. Singer, by contrast, espouses a preference utilitarian foundation for his ethic. It is the main feature of preference utilitarianism that maximization of preference satisfaction is the standard by which the rightness and wrongness of actions are judged. There is no independent standard which discounts the satisfaction of some preferences as distinct from others. All preferences, then, are thrown into a philosophical melting pot in order to get the 'right' answers about the morality of particular actions.

Having seen then some of the principles that govern traditional morality, we must now consider Singer's own position for it is within his own work that contradictions appear. Furthermore, unless Singer is prepared to adopt some principles of traditional morality, he is forced to accept conclusions which he himself acknowledges to be grossly unpalatable.

Actuality and potentiality

The traditional moralist often points out that it is insufficient simply to protect those who have the actual and actuated characteristics of the adult (e.g. rationality, autonomy, artistry, musicianship, etc.). Even the most talented musician must sleep. The most honoured among us were once dependent foetuses and infants. In the course of a lifetime many of us will suffer serious illness. Accordingly, traditional morality acknowledges that protection is due not merely to adults but to the sleeping, the young and the sick. Moreover, traditional morality usually points out that it is central to a proper understanding of beings and their treatment that we consider not only an individual's occurrent characteristics nor even its occurrent abilities (to the extent that an ability might be thought to be occurrent at all). It is necessary to consider its potentialities, since these potentialities identify treatment that is proper to it qua member of a kind or species. Singer is fond of pointing out similarities between snails and day-old infants.14 Indeed they might be thought to be alike as far as occurrent properties are concerned. But snails and humans thrive in entirely different conditions and immature snails and immature humans survive in distinct circumstances. Their potential characteristics are unalike, though their actual characteristics might be very similar. One distinction between snails and human infants is that the former have the ability to withstand cold, wet conditions in the garden while the latter do not. It is such distinctions, expressed in generalizations about species and the potentialities of various species, that inform the proper analysis of how they should be treated. Another difference between snails and immature humans is that the latter have human parents, a matter which ipso facto raises the question of parental affection and obligation.

The consequentialist denies these claims, arguing that it is actualized characteristics like rationality and consciousness (these are Singer's preferred criteria) that we value. This proposition I call moral actualism, because it values that which is actual and not that which is potential. It is this proposition combined with a stated belief that it is 'speciesist' to discriminate on the basis of our common humanity that does the moral work for Singer. We must then examine whether in fact Singer abides by his moral actualism.

Before proceeding to this question, however, it is important to make plain some simple errors in Singer's work. These errors take the form of misinterpretations of the opposing position. In Practical Ethics15 Singer considers the question of the potential of the human foetus and announces that the opposing camp holds that '[a] human foetus is a potential human being'. This however is not the position of those who deny moral actualism. There can be no doubt that a human foetus is an actual human being - it is tautologically true. Only an immature human being can become a mature human being. Its potentiality lies precisely in its capacity to mature as a human being. Accordingly, there is no sense whatsoever to Singer's protracted discussion of whether a potential X has the same rights as an X.16 First, no one seriously suggests that a human foetus should have the right to vote or equal opportunity in employment - some of the rights of the human adult. A foetus cannot vote and cannot work. He or she can, however, be killed or allowed to live. It is precisely because a foetus can be killed that the question of his or her value arises. Singer's misleading discussion of potentiality begins with the false claim, placed in his opponent's mouth, that a human foetus is only a potential human being (as distinct from an actual human being who is a potential infant, teenager or adult) and then proceeds by making the obvious but irrelevant point that a potential X need not have the rights of an X. Once we accept that a foetus is an actual human being who is a potential infant, child, teenager or adult, we can hardly assert that he or she has no value whatsoever. By the same token, it is the same potentiality which - qua lack of actuality - explains why the foetus does not have the right to vote, to equal opportunity in employment, etc. Although Singer conducts his discussion of potentiality in the context of the term 'human being', it can just as easily be understood in terms of persons*. If foetuses are only potential persons* it might be argued that they do not have the same rights as persons*. With this we may happily agree, since some of the rights persons* have are rights to vote, equal opportunity in employment, and so on. But these are not at issue. It is the right to live that is in question. Living is precisely what very young unborn humans are doing while they are subject to these moral disputes. They are not potential lives (as, say, human sperm and ova are). They are actual lives, and so to assert that we need not ascribe to foetuses the kinds of rights we do to persons* still does not settle the issue.

(i) The significance of potentiality according to Singer. These misinterpretations of his opposition aside, Singer's argument is susceptible to deeper objections. For whilst denying that there is any moral significance to the potentialities of individual members of particular species, Singer nevertheless later assumes that there is some moral significance to the concept. Singer holds that infants are not persons*. He says:

Self-consciousness, which could provide a basis for holding that it is wrong to kill one being and replace it with another, is not to be found in either the foetus or the newborn infant. Neither the foetus nor the newborn infant is an individual capable of regarding itself as a distinct entity with a life of its own to lead ...17

He also holds that it is a significant factor in the determination of an infant's claim to life that it have a serious defect:

[I]t is, rather, characteristics like rationality, autonomy and self-consciousness that make the difference [to the wrongness of killing a human being]. Defective infants lack these characteristics. Killing them, therefore, cannot be equated with killing normal human beings, or any other self-conscious beings.18

It, however, Singer holds that infants per se are not persons* and so lack the features that make them valuable and worthy of protection, then we have to ask what relevance their disability or 'defect' makes to the question of whether they are justifiably killed. If it is permissible to kill any infant on the grounds that it is not sufficiently developed, there is no good reason to think that killing very young (perhaps unborn) human beings who are also handicapped needs any justification at all. That Singer devotes much energy to proving that 'euthanasia' might be the best thing for 'defective infants' indicates precisely that he considers the potentiality of non-disabled infants to develop in the ordinary way to be of moral significance. That potentiality is recognized as having any moral significance at all demonstrates that Singer does not adhere rigorously to his actualism. Once liberated from the idea that it is only actual characteristics that count towards a creature's worth, we are bound to reject the view that infants are for the purposes of practical moral decision just like snails. Accordingly, there is no need to suppose that unborn humans are for moral purposes just like snails, a favourite personist contention designed to erode respect for the immature.

It may be objected that this argument from potentiality cuts two ways. Disabled infants have only some of the potentialities that normal infants do. Why then should we not kill them? It is, however, not the argument from potentiality of particular members of species that confirms that it is wrong to kill disabled infants. It is the argument from the potentiality of species per se that in part demonstrates this. It is also our common humanity that is the succour of the disabled. I propose no discussion of Singer's condemnation (as biased) of those who insist that the fact that a being is human does count in his or her favour. This is an issue in its own right warranting independent examination. Suffice it to say that, if Singer is wrong to think that generalizations of the moral and practical sort about species are always discriminatory and he is wrong to think that our common humanity counts for nought in life-and-death decisions about disabled infants, then we are bound to respect the immature amongst us, whether or not they are also disabled.

(ii) Potentiality and the sleeping. That Singer both wants and does not want to recognize the significance of potentialities is evidenced by the concessions he makes to the sleeping. In 'On Being Silenced in Germany',19 reprinted as the appendix to the second edition of Practical Ethics, Singer writes that his 'views cannot be a threat to anyone who is capable of wanting to go on living, or even of understanding that his or her life might be threatened'.20 Later on in the article, perhaps realizing that this definition would threaten sleeping adults, since they too while asleep are incapable of wanting to go on living, Singer modifies the claim. The new definition is rather more verbose. This time it is not those who are 'capable of wanting to go on living' who are safe, but 'anyone who is, or ever has been, even minimally aware of the fact that he or she has a possible future life that could be threatened'.21 The only reason we respect the sleeping is then, according to Singer, because prior to falling asleep the sleeper lays a claim to her life when later awake. This claim, as it were, protects her, acts as insurance, while she is asleep.

Now the main problem with Singer's 'awareness of a possible future life' criterion of moral significance is that 'awareness' is not the feature that is doing the moral work for him. It is a central part of his position that comatose people who are judged medically to be 'irreversibly comatose' may be killed or allowed to die. The fact that the comatose person was once aware and concerned for her future life prior to falling into a coma matters not at all. It is the objective potential to survive as a person* that matters for Singer - a potential possessed by the sleeping but not by the irreversibly comatose. This is so whether or not he adds that awareness of a possible future life counts in a human being's favour. Once potentiality is accepted as the conceptual instrument which demonstrates why, according to Singer, it is permissible to kill the comatose as distinct from the sleeping, he cannot backtrack. He cannot then decide that potentiality has no moral relevance in the determination of a creature's worth.

Once it is allowed that it is potential, and not merely actual, characteristics that count towards a creature's worth, we are free to reject the view that infants and other immature human beings are, for the purposes of practical moral decision, just like snails.

'A morally indefensible preference for members of our own species': 'speciesism' and 'sentientism'

In Practical Ethics, Singer says:

[N]onhuman animals and infants and severely intellectually disabled humans are in the same category ... If we make a distinction between animals and these humans, how can we do it, other than on the basis of a morally indefensible preference for members of our own species?22

It is this sort of statement that Singer himself admits23 has generated fear and, he believes, misunderstanding among some disabled people. Although I do not propose an elaborate discussion of the idea that it is discriminatory to make moral generalizations of a practical nature on the basis of species, a few points need to be made. Singer himself makes more generalizations on the basis of kind. Instead of locating interests in species according to their functions, purposes and potentialities, he locates them in sentience. If we accept terms like 'speciesism' we ought logically to condemn as discriminatory any theory which denies beings' interests on the basis that they lack sentience. Lest we misinterpret him, this argument is best understood in the context of
Singer's own discussion.

In Practical Ethics, Singer claims to have discarded the classical utilitarian interest in pleasure and pain. He expressly abandons the ethic of acting to increase pleasure and reduce pain in favour of one which requires that one do 'what, on balance, furthers the interests of those affected ...'.24 But his alleged rejection of the pleasure-pain calculus is only apparent because he elsewhere resuscitates the notion of sentience, the capacity to feel pleasure and pain, in order to reject another position. The position in question is the ethical basis of deep ecology.25 Freya Mathews in The Ecological Self locates interests in plants and ecosystems, amongst other things. Singer protests against Mathews's account that we should confine ourselves to arguments based on sentience.26 There are two points to be made here. If it is only arguments based on sentience that matter where interests are concerned, he simply has not, in the final analysis, discarded the classical utilitarian pleasure-pain account of morality. Secondly and moreover, the fact that he limits his prescription (to further the interests of all concerned) to sentient creatures suggests that he is himself prepared to make moral generalizations on the basis of kind. For him, in the context of issues surrounding ecology, sentient creatures are the kind whose interests he is prepared to countenance. Now, by applying his own methodology and using his own mode of condemnation, he is 'sentientist'. He discriminates against the interests of plants, for example, whose interests are clearly demonstrated by their potentiality to grow, reproduce and flourish in their ow right (in ways that are not merely useful to sentient creatures) by forcing his analysis of interests into the strait-jacket of the requirement of sentience. His location of the boundaries of moral concern in such a way as to exclude the interests of non-sentients is precisely the discrimination on the basis of kind that he abhors. As a matter of contingent fact, I believe that this failure to comprehend arguments for the interests of, for example, trees and plants derives from a faulty metaphysic, one which deprives itself of the conceptual apparatus which would allow an understanding of those interests. The sorts of concepts necessary to the right account are those already mentioned: potentiality, capacity, function and purpose.

Finally, Singer's sentientism in the context of ecophilosophy sits uneasily with his personism in the context of human bioethics. He was, after all, director of the Centre for Human Bioethics. If we should confine ourselves to arguments on the basis of sentience in the case of non-human interests, why is the same not true of human interests? Why do we not simply confine ourselves to arguments from sentience in the human context? By demonstrating that he is prepared to distinguish between non-human interests and human interests, Singer acknowledges that there is some morally significant distinction between human (qua type) and non-humans. When this acknowledgment is taken in conjunction with the failure of his actualism (supra), we begin to understand precisely what it wrong with the idea that it is sheer bias that motivates our distinguishing the severely intellectually disabled from other non-human animals. We are able to see too why disabled people and indeed many able-bodied and intellectually unimpaired people might find the comparison in the context of his argument so threatening.

Desires of third parties and the inviolability of innocence

One feature of traditional morality is that where innocence is concerned, the actual desires of parties other than the innocent are, for all intents and purposes, secondary. Threats of violence and destruction if an innocent is not executed27 cannot alter the fact that killing the innocent is, according to traditional morality, entirely wrong. Again, the fact that parents desire to abuse their young infants (who for the sake of argument are too young to know what is happening to them and powerless to object) in no way alters the fact that the abuse is wrong. It is a part of the very concept of innocence that it is not susceptible to variation by the desires or consent of others. On the contrary, desires of third parties, to be morally relevant, must conform to the standard of reasonableness. Indeed it is this feature of traditional morality that explains the drama and the tragedy of many real and fictitious predicaments. Innocence cries out for respect. The mob desirous of the punishment of an innocent must cease its bellowing. The parents of the abused child must abandon their desire to dominate and abuse. This position is to be contrasted with one which argues that the desires of third parties are to be taken at face value in determining the morality of actions, even where innocent victims are concerned. According to this argument, the wrongness of punishing or abusing the innocent is not intrinsic nor connected with the innocence of the victim but arises from extrinsic reasons. If it is admitted at all that actions like punishing and abusing the innocent are wrong, it is admitted on extrinsic grounds which balance the desires of the innocent against the desires of the mob or the parents respectively as well as the desires of existing and future communities. So, it is argued, it would be wrong to punish and abuse the innocent because it would threaten people's belief in systems of justice and make them fearful, but not, it must be said, because it does violence to an innocent.

Singer equivocates about the role desires play in his scheme. On the one hand he suggests that what parents want for their 'defective' newborn children and the degree to which they are distressed should be taken at face value when considering whether the newborn should be killed. Their actual desires, whether or not they are reasonable, must be taken into account when assessing moral options. On the other hand, he seems to suggest that the kinds of desire that should be taken into account for ethical purposes are not actual desires but reasonable desires. In one passage, quoted earlier, Singer notes that someone's choice, and presumably too his desire to live, may be outweighed where the person will otherwise lead a miserable life:

The classical utilitarian [whose position Singer accepts in the context] might have to accept that in some cases it would be right to kill a person who does not choose to die on the grounds that the person will otherwise lead a miserable life.28

In this sort of case, what a person actually wants is subject to the demands of reasonableness. Where the desires are unreasonable they may be outweighed by other considerations. In the above example, the unreasonable desire to live is subject to the demands of (utilitarian) rationality. Singer nowhere rejects this view outlined by him as the classical utilitarian and the preference utilitarian view, though (as we have noted) he equivocates inasmuch as the demands of rationality seem to consist, for a preference utilitarian, simply in maximizing the satisfaction of actual preferences. In any case, for present purposes we may assume that the position outlined in the above quotation also reflects his own. It seems, then, that to the extent that he subscribes to the position outlined, Singer wants to maintain the independence of rationality,so that desires may be classified as irrational and consequently irrelevant or outweighed in his moral calculus. At the same time he wants to use the actual desires of third parties (parties other than the innocent) as a factor in his utility equation. But he cannot insist on both a functional definition of desire (that is, desires conceived as they are) and an ideal definition of desire (that is, desires conceived as they ought to be, rationally considered). The inconsistency becomes obvious when a clear and rational test of a justified killing is sought.

Suppose an infant who has the potential to live happily, say a baby girl living in China, is not wanted by her parents and causes her parents severe distress. Does this severe distress justify her being killed? If the parents' actual desires are taken to be the criteria by which this matter is judged, the baby girl must die. If the parents' reasonable desires are the criteria by which her life is justified, a quite different result might be achieved, depending on the test supplied for the rationality of desires. This same point about Singer is made by Uniacke and McCloskey:

The 'parental distress' argument is very indeterminate. How much distress and unhappiness to parents and siblings warrants killing an infant who will lead a worthwhile life? Will any noticeable amount suffice if a replacement infant with at least equally good prospects would enhance parental and sibling happiness?29

Singer states that '[p]arents may, with good reason, regret that a disabled child was ever born. In that event the effect that the death of the child will have on its parents can be a reason for, rather than against killing it'.30 It is difficult to know what the phrase 'may, with good reason' is meant to signify in this context. Are parents permitted, when they have good reason, to regret that their child was ever born and permitted thereby to count these regrets in the calculus against his or her continued existence? Or are all parents of disabled children correct in regretting that their disabled children were ever born and ipso facto right in supposing there are good reasons against his or her continued existence? Or are the regrets of the parents of the disabled child the very reason why the disabled child in question should be 'helped to die'? If we construe his remarks in the sense that Singer has a kind of reasonableness requirement on the desires and preferences he is prepared to count in his moral calculus, we realize that he is not neutral between the kinds of desire people actually have. On this account he outlaws forms of desire incompatible with his conclusions and he loses the preference utilitarian foundation for his own ethic. The task then for him would be to give an account of the rationality of desires. This enterprise would be an enormous one. Importantly, it would not be preference utilitarian.31 On a second interpretation of his argument, the parents' desires regarding a particular severely disabled child are not legislated out of the moral calculus but must always be outweighed by other considerations. What the parents want for a particularly severely disabled child cannot ever affect the final decision because their distress and the fact that they do not want the child, or even the fact that they want the child desperately is, in the end, simply not relevant. Other factors determine the issue. On the final interpretation, the parents' actual desires are taken at face value and there is no requirement that parental distress and preference be a reasonable response to an infant's predicament.32 This interpretation is incompatible with the view that it is wrong to kill the little girl living in China. It is also incompatible with a conclusion which Singer seems to want to allow, that there are some lives that will be just so miserable that the desires of the family (or the self) are not a reasonable ground in the circumstances for allowing the life to continue. As Singer says: 'The classical utilitarian might have to accept that in some cases it would be right to kill a person who does not choose to die on the grounds that the person will otherwise lead a miserable life.'33 Which the correct interpretation of Singer's text is in the circumstances, remains mere speculation. There are, however, problems inherent in the discussion.

Baby-farming and harvesting unwanted non-persons*

If it does not matter what species a creature belongs to and if we ought to be doing whatever is necessary to bring about the best outcome (furthering the interests of everyone involved), why should we not grow human beings (perhaps without brains, or at least intentionally damaged ones) for use as spare parts in transplant surgery? Singer himself tackles this thorny question and urges caution over the growing of human embryos with deliberately damaged brains for organ transplant and other uses. Growing brain-damaged children for use as spare parts cannot contravene any prohibition on killing persons* (where persons* are, recall, self-conscious beings) because by definition 'embryos [grown] until they resemble normal babies but with brains deliberately damaged so that they are in a permanent coma'34 are non-persons*. These human beings have had their capacities for sentience, self-consciousness and rationality removed. Nor can growing brain-damaged children for use as spare parts contravene any consequentialist principle of maximizing good states of affairs (or furthering the interests of everyone involved) since these transplant organs and limbs would certainly further the interests of those patients who needed them and their friends and families. Whatever it is that is wrong with growing them for the greater good must be a powerful moral principle - since it is doing a great deal of work - and it must be independent of the prohibition on killing persons*. What then could this principle be? Singer replies that to grow human beings for use as spare parts in transplant surgery would do violence to our basic attitude of care and protection for infant human beings. Furthermore, he even seems to suggest that there is, after all, a morally significant distinction between setting up conflicts deliberately and facing situations when they are forced upon one. This is how he puts it:

To face these situations when they are forced upon us is one thing; to set up such conflicts deliberately is another. For the sake of the welfare of all our children, the basic attitude of care and protection of infants is one we must not imperil.35

At this claim we should pause.

Singer uses two principles to explain his intuition that this sort of activity against non-persons* is wrong. Both may be used to demonstrate inconsistencies in his own account. Perhaps the clearer of the two principles is that we must not imperil 'the basic attitude of care and protection of infants'. We must consider 'the welfare of all our children'. The second principle seems to insist that there is a morally significant distinction between facing situations when they are forced upon us and setting up conflicts deliberately.

(i) The attitude of care and protection. Consider first the content of the first of the two points thus stated: baby-farming does violence to our basic attitude of care and protection and this feature demonstrates that such activity is unacceptable. If our basic attitude of care and protection matters morally even where non-persons* are concerned, then they matter where the lives of the embryonic, the unborn, the disabled, the suffering, comatose and elderly are at issue. It is, after all, precisely this kind of argument that is often relied on to demonstrate the wrongness of killing vulnerable people. Killing the terminally ill or comatose, it may be argued, does violence to our attitude of care and protection for the terminally ill and comatose. It is precisely the terminally ill and comatose who need simple medical care as well as ordinary respect. Killing them, it might be argued, is incompatible with the care and respect that is owed to them. Likewise, killing the unborn does violence to our attitudes of care and respect for immature humanity. Unborn humans very often need medical and other care as well as simple respect. Killing them is at odds with the care that is due to them.

The inconsistency in Singer's work may be elucidated by posing a dilemma. On either horn of the dilemma Singer must contradict an important aspect of his theory. If our basic attitude of care and protection matters morally then the principle which asserts that it does may be used to safeguard other non-persons* like the embryonic, the unborn, the disabled, comatose and the elderly. On this horn of the dilemma, Singer must deny his conclusions about these groups. Conversely, if Singer insists that our attitudes of care and respect for one another are not undermined by, for example, creating human embryos for scientific experiment, we may likewise insist that our attitudes of care and respect for infants are not affected by creating non-persons* for use in transplant surgery. On this analysis Singer must deny his conclusions about baby-farming. he is simply wrong to think that it is morally problematic.

If Singer is wrong to urge caution in respect of baby-farming and if he is wrong to think that the prospect of baby-farming is 'repellent', a great many other kinds of 'repellent' cases are permissible according to his own theory. So long as no members of the victim's family have any objection or, better still where the victim is an orphan, comatose patients may be harvested for their organs and tissue, the terminally ill may be used for scientific experiment and foetal tissue may be used for collagen creams. If Singer is forced onto this horn of the dilemma he must acknowledge that his theory lacks the conceptual apparatus which would allow him to urge caution in respect of any of these cases.

If we wanted to help Singer onto this horn of the dilemma so that he was committed to some really revolting kinds of activity, we could protest that loss of the attitude of care and protection for these groups is only contingently related to particular acts of harvesting, experimentation or exploitation. We might use an example. Imagine a society in which the baby-farming activity is conducted by only a few concerned individuals in private. This small team understands practical morality just as it should be understood according to the combined precepts of personism and consequentialism. The group does not practise baby-farming on brain-damaged babies whose parents object, and there are no abuses of theory. Baby-farming is conducted precisely as it should be, that is to say only in cases where there are no additional moral reasons which speak in favour of the disabled baby's being treated with the sort of care due to ordinary babies. Moreover, the team understands rightly that many other individuals are being helped by their activities. In such cases, we might argue, Singer can have no grounds to object to the activity of baby-farming. There are simply no important attitudes of care and protection at stake so he is wrong to suppose that this is a moral reason against it. His intuition that this sort of activity is unacceptable is entirely mistaken. What this example would show is that given that there is no good reason to suppose that basic attitudes of care and concern for persons* would necessarily be threatened by baby-farming, we would have no reason to balk at the prospect of it. Once committed to this kind of case, Singer would have to admit that his theory could not supply any reason against harvesting the comatose and the terminally ill, taking the gold from their teeth for charitable causes and so on.

Singer might, however, want to hold on to his intuition that baby-farming is unacceptable on another basis. He might insist that there is an intrinsic and logical connection between baby-farming and bad attitudes towards infants generally. But this admission would concede too much. It would then be open to those who disagree with his conclusions likewise to insist that there is an intrinsic and logical connection between particular killings of the terminally ill or comatose generally. In short, if Singer is entitled to help himself to the idea that we should not do violence to attitudes of care and respect for the vulnerable, then so too are those who disagree with his conclusions. If our basic attitudes of care and protection matter then they matter in respect of other vulnerable groups.

Singer is faced with a dilemma. He must either drop his prohibition on baby-farming and admit that he lacks the conceptual apparatus that would allow him to show that the 'repellent' cases mentioned above are morally problematic. Alternatively, he must admit that a great many other activities, like killing the terminally ill, the comatose and the very young are wrong because they imperil important moral attitudes.

(ii) 'Setting up conflicts deliberately'. Consider Singer's alternative explanation of his prohibition on growing brain-damaged embryos for use as spare parts. There is, he admits, a distinction between setting up conflicts deliberately and facing them when they are forced upon one. It should be remembered that this explanation is morally significant in the context of his theory because, like the principle considered above (that we should not imperil attitudes of care and protection), it renders impermissible the kind of exploitation and killing which, he argues, is impermissible, but not prohibited by reason of consequentialism or personism. This explanation, unlike the one just considered, however, is somewhat unclear. What does Singer mean by 'forced'? Does he mean that baby-farming would be acceptable if it were 'forced' on a person by reason of duress? Is he claiming that if someone were to use force to get the baby-farming business going, it would be justified and acceptable? Or is the point about one-off cases generally? Is he claiming that if force were applied to achieve the use of only one embryo grown to the appropriate size for his or her organs, this would be acceptable because it was a one-off case? Or is the point about deliberation and intention? Is Singer saying that growing even one baby for use of its organs is morally objectionable because it exploits an infant that was been deliberately and intentionally disabled for the use of its organs? There are many possible interpretations of Singer's words of caution on baby-farming. Since he does not elaborate on the issue we should not put words in his mouth.

We may make one point about some arguments against, for example, arranging health care in such a way as to make killings or aiding and abetting suicide a legal option for medical staff caring for the terminally ill. In the killing of the terminally ill case, for example, a conflict is deliberately and systematically set up. Arranging health care so that it supplies not merely pain relief but execution services for the elderly and terminally ill itself sets up a conflict about the function of health care. Health professionals are no longer concerned with health care but with executions. Nurses and carers become bearers of death warrants and participants in the activity of killing. This kind of 'health care' arrangement, moreover, forces a number of conflicts on the aged and terminally ill.36 These patients can no longer rely on the idea that the health professionals surrounding them have their care at heart. They are faced with the conflict of killing themselves when they are most vulnerable.37 Using Singer's own prohibition on setting up conflicts deliberately, doctor-assisted death legislation would be rightly resisted.

I have suggested that Singer is faced with a dilemma. He must either abandon his prohibition on baby-farming and admit that his account can supply no reason for regarding the 'repellent' cases mentioned above as morally problematic. We should not, then, balk at any of these cases. Alternatively, he must admit that a great many other activities, like killing the terminally ill, the comatose and the very young are wrong because they imperil important moral attitudes or set up conflicts deliberately. On one interpretation of his prohibition on setting up conflicts deliberately, systems of medical killing or medically assisted suicide, to use one example appropriate to our times, are properly treated with caution. We can only speculate which path Singer would care to take in this regard. Both routes, however, are equally destructive of certain of his theses.

Problems of Interpretation

Before concluding this section I need to make two important points about the difficulty of interpreting Singer. He often says that 'standard moral principles, for example, telling the truth, keeping promises ... and so on'38 may be departed from when 'it is absolutely plain that departing from the principles will produce a much better result than we will obtain by sticking to them, and then we may be justified in making the departure'.39 The problem with this sort of declaration is that it must make us wonder whether Singer sometimes deliberately misstates his own position in order to bring about what he sees as (and what might or might not in fact turn out to be) a much better result. Does he say things for their shock value or in order to bring about change? How are we to construe the motivation for his comments on baby-farming and killing those who do not want to die? What might the 'much better result' be when he makes this or that shocking statement? The problem with these sorts of questions is that they are well beyond the ken of philosophy. The discipline of philosophy cannot supply answers to questions about a theorist's motivation. It is not merely improper but also fallacious to inquire into the psychology and motivation of a theory's proponent in this or that context. We must assume for the purposes of argument that a proponent of a given thesis is bona fide and is not using lies or deception to bring about a much better result.

There is a second problem of interpretation that deserves fuller discussion elsewhere. When he gives reasons against killing one person (or person*) for use as spare parts for two, or against baby-farming and like activities,40 is Singer reasoning critically or intuitively? Following R. M. Hare, Singer maintains a distinction between critical and intuitive moral reasoning. This ethical distinction between intuitive reasoning reasoning which, he says, does not admit of 'lengthy'41 utilitarian calculation - he believes this is somethinglike traditional morality for everyday use) and critical reasoning (which is the sort of protracted reasoning which would allow us secretly to kill one innocent victim for a greater good, say, body parts for two others) suffers from a number of problems. One is that the criteria supplied for the application of the distinction (speed and ease of reasoning42 for the purposes of decision making) do not pick out the kinds of cases that need to be explained. When we theorize about certain cases, the length of time we have to consider the problem is not ordinarily an issue. Further, if the critical level of moral reasoning is not to be rendered entirely void of practical significance for everyday life, Singer will need to dissociate critical reasoning, at least in part, from the notion that it is utterly distinct from the reasoning of real life. Again he argues: 'It may be that in the long run, we will achieve better results - greater overall happiness - if we urge people not to judge each individual action by the standard of utility, but instead to think along the lines of some broad principles'43 like those of traditional morality. But it is not clear that this does not simply create two utilitarian principles each of which contradict each other. The first could be stated thus: in order to get better results in the long run in this individual case (of killing one for spare parts for two), we should observer the principles of traditional morality. The second might say this: in order to get better results in the long run, it is better in this individual case (of killing one as spare parts for two) to observe the principles of utility. To point out that both principles are stated in terms of maximization does nothing to remove the contradiction. In order to be a truly practical and rational ethic there must be some solutions to this problem.

Some Red Herrings in Singer

(i) Sanctity of life doctrine. One of Singer's methods of encouraging us to accept his conclusions is the setting up of a false dichotomy between religious ('sanctity-of-life-based') views and 'rational' views. It is argued as follows:

  1. Either sanctity-of-life-based views are true or Singer's 0ersonism-cum-consequentialism is true
  2. Sanctity-of-life-based views are irrational and false, therefore
  3. Singer's view is true.

The first premiss is false. One need not enter into any debate about the existence of God to understand the contradiction inherent in personist and consequentialist theories like Singer's. It is a caricature of the debate to envisage 'Holy Rollers' in one corner and the 'Hard-minded Rationalists' in the other. Rationality is by no means incompatible with non-discriminatory respect for human life. It is a part of the logic of some forms of humanism that every human life be given a chance without the threat of attack or destruction.44

When it is understood that Singer himself recognizes the moral significance of potentiality whilst steadfastly denying that it has any significance in the context of embryonic and other immature human life, we begin to wonder just how rational and hard-minded his position is. Again, Singer's intuition that it is wrong to produce brainless children for spare parts in transplant surgery is an intuition that is entirely arbitrary in the context of his theory. His personism does not dictate it. Nor does his maximization principle (that we should act in order to further the interests of persons* affected) prohibit it. When we understand that certain examples are at odds with his theory and rationalizable only on grounds which contradict other of his proposals, we have good reason to want to part company with Singer's brand of thought.

(ii) The violation of a right to freedom of speech. Singer insists that his freedom to speak has been violated by those who either disrupted his seminars or prevented him from being given the opportunity to speak.45 I do not propose to expatiate upon the subject of free speech.46 What is of interest in this context is that Singer himself is no champion of the right to free speech. Modern consequentialists believe that it is often better to propagate traditional morality, or at least principles that are non-consequentialist. It may, for example, be 'better, all things considered' for doctrines which cause offence to be kept quiet even if they are true. This applies as much to consequentialist doctrine as it does to any other. It is, moreover, straight consequentialist teaching. It is peculiar, therefore, to demand rights that one does not believe one has. Given the content of his theory, Singer must do something other than cry out for his right to speak.

(iii) 'Human beings treated with less consideration'. In response to a protester, Singer once claimed that his views do not reduce the consideration owed to human being but raise that due to animals.47 This thought would be an excellent one if it represented Singer's real position. But it does not. There are indeed defences of the rights of animals which are not founded on personist assumptions.48 The proposal that separates Singer from these other accounts is precisely that Singer is not merely extending our ordinary concern for humans to animals. He is asking us to disregard our common humanity in any decision making about the vulnerable, the very young and the disabled. It is just this proposal that makes Singer's work the 'highly provocative' writing it is claimed to be for the purposes of publication and sale. And it is this proposal presumably that persuaded the disabled to secure their wheelchairs to the doors of newspaper editorial offices for so much as reporting Singer's views on euthanasia.


The claims of innocence demand respect. A moral theory which cannot account for them is to that extent inadequate. It is because of his insistence that 'morality' is synonymous with 'the maximization of good states of affairs' that Singer is forced to equivocate or contradict himself over the harvesting of non-persons* for the general good. What is wrong with these kinds of activity and other abuses of the vulnerable is neither that the activity is not secret enough, nor that there may in fact be some people somewhere who want the victim to survive. The wrong is intrinsic: it is an offence against an innocent. The claims of innocence cut across utility calculations.

There is always something paradoxical about discussing in detail a work which one considers ethically dangerous. By citing a philosopher's work we bolster his presence in various citation indexes. Such is the nature of the modern academy that frequency of citation sometimes even takes the place of reasons for citation in assessing the worth of a particular brand of thought. Singer notes with some satisfaction that his opponents foster a climate of debate49 about the topic of killing by criticising his views. But this fact should not cause amusement. It is to be expected that his much-publicized opinions on who may be killed with impunity will attract opposition. While there are good reasons simply to ignore a dangerously false work that is neither public nor gaining respectability, this is not so with one that is government-funded and in the service of a burgeoning industry in human fertility and genetic engineering. It may be necessary in some circumstances to fan the flames of a debate that we should not ordinarily even countenance. Bad theory assisted by public funds and the interests of a growing industry requires more than a refusal to countenance the legitimacy of the very debate.

Singer's theory, however, is not altogether spurious. It does appear to matter to him that his theory is not false. He has no truck with relativism. He points out rightly that if we are going to respect immature human life like human embryos we ought in truth to treat with care our reproductive capacity, since there is the chance that a human being will be created with all the parental obligations that flow from the fact. This is particularly obvious now. Where human ovum and sperm have been abandoned to scientific research, there is the chance that a couple will become parents without even knowing it. Possible parental obligations do flow from our reproductive capacities and Singer is right to point out the continuum. (He is wrong however to think that conception, the point at which parenthood is determined, is a morally irrelevant event. He is wrong to think that a reproductive technologist's promise that he or she will kill the developing human life after experimenting with it frees parents and societies of their obligations to it.) He is even correct to point out that, in some cases, there is no moral distinction between the foetus and the infant. (I limit the generalization, though Singer himself does not, because in some cases the very fact of giving birth can threaten the mother. In such a case there is an obvious moral distinction between the born and the unborn.)

Singer, moreover, writes simply and eloquently about some issues. He points out how barbarously battery farm animals, and animals used for research, are treated. At the same time, however, he articulates the underlying rationalization of the human fertility industry. The danger of Singer's writing is that his compelling defence of the interests and rights of animals is mixed up with the human fertility business and a eugenic agenda. I have suggested that in order to arrive at his false conclusions Singer must equivocate or contradict his own theory. I have also suggested that those who boycott his lectures are not mistaken about his programme. His plans for the eradication of suffering through disability involve the eradication, the systematic killing, of those with disabilities while they are immature, and even in certain cases when they are adult. To support this conclusion and others which counsel the killing of the innocent, Singer uses arguments which he himself does not respect.


*Current address: Human Rights & Social Justice Research Institute, London Metropolitan University, UK, email: This article first appeared as a chapter in Human Lives: Critical Essays on Consequentialist Bioethics, eds. David S. Oderberg and Jacqueline A. Laing (Basingstoke: Macmillan, 1997), pp. 196-224.

1'When innocence is frightened, the judge is condemned.'

2Jenny Teichman coined the term in 1992. See J. Teichman, 'Humanism and Personism: The False Philosophy of Peter Singer', Quadrant, December 1992, pp. 26-9. See also her 'Freedom of Speech and the Public Platform', Journal of Applied Philosophy 11 (1994) 99-105.

3Within the confines of this paper I discuss in particular two editions of the same book: P. Singer, Practical Ethics (Cambridge: CUP, 1979), hereafter referred to as PE(1), and Practical Ethics (Cambridge: CUP, 1993), hereafter referred to as PE(2).

4'On Being Silenced in Germany', New York Review of Books, 5 August 1991, pp. 36-42; 'A German Attack on Applied Ethics: a Statement by Peter singer', Journal of Applied Philosophy 9 (1992)85-91.

5PE(2), p. 357. Of the boycott by disabled people of his lecture at the University of Zürich in May 1991, he says, 'I had an overwhelming feeling that this was what it must have been like to attempt to reason against the rising tide of Nazism in the declining days of the Weimar Republic.'

6PE(1), pp. 131, 134, 136, 156.

7PE(2), p. 181.

8PE(2), p. 131.

9PE(2), p. 100.

10PE(2), p. 99.

11PE(2), p. 100.

12Judith Jarvis Thomson, 'A Defense of Abortion', Philosophy and Public Affairs 1 (1971) 47-66. For a feminist rejection of reproductive technologies, see Lynda Birke, Susan Himmelweit and Gail Vines, Tomorrow's Child: Reproductive Technologies in the '90s (London: Virago, 1990): '[W]hat we are witnessing is a takeover by scientists of women's role in reproduction ... moving towards a dehumanised (and defeminised) technological future. The position is one of total resistance to scientific and male control of reproductive processes, by a complete rejection of the new technologies' (p. 19). See also Mara Mies, 'Do We Need All This? A Call Against Genetic Engineering and Reproductive Technology', in Patricia Spallone and Deborah Steinberg (eds.), Made to Order: The Myth of Reproductive and Genetic Progress (New York: OUP, 1987). These approaches are quite unlike some of those of the early 1970s which saw the female biological nature, connected as it is to pregnancy and childbearing, as essentially substandard and, as such, perfected and liberated by the new reproductive technologies.

13For a non-consequentialist rejection of Thomson's argument in favour of abortion see J. M. Finnis, 'The Rights and Wrongs of Abortion: A Reply to Judith Jarvis Thomson', Philosophy And Public Affairs 2 (1972) 117-45.

14PE(2), p. 90: 'Killing a snail or a day-old infant does not thwart any desires of this kind [for the future], because snails and newborn infants are incapable of having such desires'.

15PE(2), p. 152.

16PE(2), p. 153.

17PE(2), p. 188.

18PE(1), p. 131. See Also PE(2), p. 182 (with 'defective' removed).

19New York Review of Books, 15 August 1991, pp. 36-42; PE(2), pp. 337-59.

20PE(2), p. 345.

21PE(2), pp. 357-8.

22PE(2), p. 60.

23'One protester quoted from a passage in which I compare the capacities of intellectually disabled humans and nonhuman animals': PE(2), p. 347.

24PE(2), p. 14; emphasis added.

25See Freya Mathews, The Ecological Self (London: Routledge, 1991).

26PE(2), p. 284.

27This example was devised by H. J. McCloskey. See 'A Non-Utilitarian Approach to Punishment', Inquire 8 (1965) 249-63, and Meta-ethics and Normative Ethics (The Hague: Martinus Nijhoff, 1969).

28PE(2), p. 100.

29Suzanne Uniacke and H. J. McCloskey, 'Peter Singer and Non-Voluntary "Euthanasia": Tripping down the Slippery Slope', Journal of Applied Philosophy 9 (1992) 203-19, at p. 213. Uniacke and McCloskey investigate the question of how serious a defect needs to be for Singer before a baby is justifiably killed. They point out that the attitudes of parents are only contingently related to the degree of defect in an infant or adult. Some defects like disfigurements can cause more distress than apparently more serious disabilities. An important feature of the paper is that it demonstrates that killing of the sort favoured by Singer is not euthanasia i.e. gentle death) since it is motivated by concern for the welfare of others or an increase in the total happiness.

30PE(2), p. 183. For 'disabled' read 'defective' in PE(1), p. 132; emphasis added.

31Bernard Williams addresses this general point in J. J. C. Smart and Bernard Williams, Utilitarianism: For and Against (Cambridge: CUP, 1973), p. 131: '[M]odern utilitarianism is supposed to be a system neutral between the preferences that people actually have. To legislate them out is not to pursue people's happiness, but to remodel the world towards forms of "happiness" more amenable to utilitarian ways of thought. But if they are not to be legislated out, then utilitarianism has got to co-exist with them, and it is not clear how it does that.'

32This is the interpretation Uniacke and McCloskey favour: 'Given Singer's position that "euthanasia" need not be "in the infant's interests" and that infants are replaceable, his justification of pseudo-euthanasia infanticide seems not to require that parental and sibling distress be a reasonable response to an infant's defect': 'Peter Singer and Non-Voluntary "Euthanasia"', at p. 213.

33PE(2), p. 100.

34P. Singer and D. Wells, The Reproduction Revolution (Oxford: OUP, 1984), p. 148.

35The Reproduction Revolution, p. 149.

36Rita Marker's book Deadly Compassion: The Death of Ann Humphry and the Case Against Euthanasia (London: HarperCollins, 1994) is a readable antidote to the writings of Derek Humphry. The latter is cited several times with approval by Singer: PE(2), pp. 176, 369, 370. The details of Ann Humphry's last months after being diagnosed with cancer and her callous treatment by her husband and fellow euthanasia campaigner Derek Humphry, and by the Hemlock Society, are chillingly described by Marker. In a law suit against Derek, Ann claimed among other things that he intended to 'impede and oppress [her] recovery from cancer itself' and to 'induce [her] despair and [her] suicide': Humphry v Humphry, National Hemlock Society, Mero, Hemlock of Washington State, Circuit Court, Lane County, Oregon, Oct. 19, 1990.

37I have suggested elsewhere that the very business of arranging death services for the vulnerable may well ensure that one of the consequentialist criteria of a justified killing is present: namely, that the victim in question is made to feel a burden and so wants to die. My point is that the very holding of these fatal views may affect in advance the outcome. It is important to the logic of a theory that it be independent of the consequences it seeks to describe and/or justify: 'Assisting Suicide', Journal of Criminal Law 54 (1990) 106-16.

38PE(2), p. 93.

39PE(2), p. 94.

40Let us assume, of course, that the complete secrecy condition holds in these cases.

41PE(2), pp. 92-4.

42'In real life we usually cannot foresee all the complexities of our choices. It is simply not practical to try to calculate the consequences, in advance, of every choice we make ... in many cases we would be calculating in less than ideal circumstances. We could be hurried, or flustered. We might be feeling angry, or hurt, or competitive ... Or we might just not be very good at thinking about such complicated issues ... . PE(2), pp. 92-3.

43PE(2), p. 92.

44See J. Teichman, 'Humanism and Personism' (n. 2). For an eloquent popular defence of the unborn disabled from an atheist, see Dominic Lawson, 'A Special Kind of Baby', Sunday Telegraph, 18 June 1995. Speaking of the free NHS tests for abnormalities which are available, with the offer of abortion on discovering any, he writes: 'In the People's Republic of China, the authorities wait until such children are born naturally, before starving them to death. In Hitler's Germany, even before the final solution to the Jewish problem, the Nazis were exterminating wholesale the mentally retarded. In this country the weeding-out process is done before birth, and only with the parents' consent. I do not think, however, that this constitutes a triumph for democracy.'

45He chastises 'Germans and Austrians, both in academic life and in the press' for showing themselves 'sadly lacking in the commitment' to the right to freedom of speech. He cites with approval Voltaire's dictum: 'I disapprove of what you say, but I will defend to the death your right to say it'(emphasis added). He notes that no one has as yet been asked to risk death in order to defend his right to discuss euthanasia in Germany (PE(2), p. 359). He is nonetheless 'not convinced that the notion of a moral right is a helpful or meaningful one' (p. 96), except as shorthand for considerations the utilitarian thinks fundamental, such as preferences, interests and feelings of pleasure and pain. But the notion of a right had better be a helpful and a meaningful one in itself independently of these considerations) if his right to discuss euthanasia in Germany is not to fall entirely prey to the whims of public opinion or public preference. Even if public opinion were entirely against him, his right, assuming he had one, would remain. These utilitarian considerations have very little to do with his right to speak, though they may, of course, affect his actually speaking.

46See further J. Teichman, 'Humanism and Personism' and 'Freedom of Speech and the Public Platform' (n. 2).

47PE(2), p. 147.

48See Stephen Clark, The Moral Status of Animals (Oxford: Clarendon Press, 1977); Andrew Linzey, Christianity and the Rights of Animals (London: SPCK, 1987); Mary Midgley, Animals and Why They Matter (Harmondsworth: Penguin, 1983); Tom Regan, The Case for Animal Rights (London: Routledge, 1984).

49PE(2), p. 350.