The non-identity problem
The non-identity problem originates with the work of Derek Parfit and it concerns the creation of future people. The non-identity problem comes in two forms – either Direct or Indirect.Footnote 1 Imagine a girl becomes pregnant at the age of fourteen (1984, p. 358). If she has a child now then it will suffer from her inexperience and financial instability – alternatively she can wait until she is older and have a child who will not suffer from her young age. Parfit thinks that most would claim that in this situation she should wait and have the later child. But by choosing to have a child either now or later, the girl does not make anyone's life better than it could otherwise have been. The first child would not exist at all if the mother waits – waiting does not lead to the existence of the same child in better circumstances: a completely different child exists. The mother faces a choice between separate distinct children. This is the Direct version of the non-identity problem.
The Indirect version of the non-identity problem occurs when selecting a policy which affects where and when people are born, thus altering the make-up of the resulting population. Parfit's famous example is the social policy choice between Conservation and Depletion. A community chooses either depletion or conservation of resources; if depletion is chosen the living standard ‘over the next two centuries would be slightly higher than it would have been if we had chosen Conservation’ (1984, p. 362). However, the living standard for many generations beyond that would ‘be much lower than it would have been if we had chosen Conservation’ (1984, p. 362). The policy chosen affects the lives of those who exist at the time it is implemented and, plausibly, it will affect those who will come into existence because people will live different lives which would not be available under the alternative policy. The implication of these two scenarios is that if our actions result in different people existing, because the behaviour of existing people changes, resulting in the creation of different offspring, then our actions do not make them better or worse off. Those generations who might exist with a low living standard as a result of the depletion policy could not otherwise exist and so cannot complain that their lives could have been better.
What underpins the non-identity problem? Consider Wilma: she can conceive a child now who will be blind (Pebbles) or take a pill once a day for two months and conceive another child who will not be blind (Rocks). Boonin sets out the following five propositions:
P1: Wilma's act of conceiving now rather than taking a pill once a day for two months before conceiving does not make Pebbles worse off than she would otherwise have been;
P2: if A's act harms B, then A's act makes B worse off than B would otherwise have been;
P3: Wilma's act of conceiving now rather than taking a pill once a day for two months before conceiving does not harm anyone other than Pebbles;
P4: if an act does not harm anyone, then the act does not wrong anyone;
P5: if an act does not wrong anyone, then the act is not morally wrong;
C: Wilma's act of conceiving Pebbles is not morally wrong. (p. 27)
Validity of the non-identity problem
Unsurprisingly, a number of arguments have been brought to bear against the non-identity problem because it undermines an area of major moral concern – the status of future persons. If creating people does not make their lives better or worse, then we cannot claim that by having disabled children we are making their lives worse than they otherwise would have been, nor can we claim that we are benefiting those who come to exist.
David Boonin's book addresses a number of objections to the non-identity problem and finds them inadequate – he concludes that we must accept the non-identity problem as the correct explanation of future people. Boonin covers the arguments targeting each of the premises and provides counter-arguments against them. Boonin's book is, therefore, an excellent introduction and reference guide for anyone interested in the non-identity problem.
By considering each proposition in turn, Boonin is able to consider a wide variety of otherwise disparate issues in a straightforward and clear manner. The most useful feature of Boonin's work is the book's structure: each chapter covers one of the premises above, setting out the arguments objecting to the premise followed by Boonin's challenge to those objections. These objections are always taken in their strongest form, with Boonin arguing that even if we accept the objections they still fail to undermine the relevant proposition. Even more helpfully, arguments against the objections based on their internal coherence are included as a set of appendices, allowing the reader to engage with a wide variety of issues relating to the non-identity problem. For example, Chapter Three covers the second harm-related proposition and considers the objections to it along with different conceptions of harm. The relevant appendix (A) considers arguments concerning the internal coherence of accounts of harm, while the substantive chapter itself takes as read the objections to the comparative account of harm and then argues that this account would still fail to undermine the second proposition because the alternatives are either absurd or entail that harm does not occur (p. 102). In this way Boonin successfully collates a range of issues concerning the non-identity problem in an accessible and informative book which, if nothing else, serves as a springboard for those considering non-identity problems.
For example, one objection Boonin analyses is the de re de dicto distinction. Boonin's analysis concerns two interpretations of the reproductive scenarios: the de re understanding focuses on the child Wilma actually has and the de dicto understanding that ‘whatever person turns out to satisfy the description “her child”’ (p. 31) may be worse off than some other person who could have satisfied that description. On the de dicto interpretation Wilma's act does make ‘her child’ worse off because both possible children can be ‘her child’. However, this is insufficient to counter the non-identity argument because an additional component is needed, namely that harm in the de dicto sense is relevant to determining whether an act is morally wrong.
Boonin points out that the de dicto interpretation can itself be interpreted in two different ways: we can focus on either the ‘magnitude of the harm’ or the individual's total ‘level of health’ (p. 35) that would be suffered by whoever satisfies the descriptive statement. Consider a Minor and a Severe accident in which the driver in Minor suffers a broken arm while the driver in Severe suffers a traumatic brain injury – this is a case concerning the magnitude of an injury. Now consider two cases in which a driver suffers a broken arm. In the first vehicle the driver is Ambidextrous while the second the driver is Handicapped and has the use of only one arm. Here, an injury of the same magnitude – a broken arm – has very different implications for each driver. For Ambidextrous a broken arm is a minor inconvenience because he simply uses his other arm. However, for Handicapped a broken arm means that he is wholly dependent on care and assistance until his one usable arm has healed sufficiently for him to use it again. The total level of health of each driver differs, resulting in an injury of the same magnitude having different effects.
In the case of accident policies we can either choose a policy which reduces the severity of accidents or we can try to ensure that healthier people have the more severe accidents. Boonin argues that the first sense must be used: in improving road safety the ‘job is to reduce the severity of the accidents that occur, not to increase the severity of the accidents by redirecting them onto healthier people’ (p. 36). Similarly, in the case of reproduction, a prospective parent's ‘duty is simply to minimise the harm that will be incurred by whatever person turns out to occupy the role of her child’ (p. 36). Additionally, this brings the notion of harm into reproduction, which Boonin covers in Chapters Three and Four.
Here, I merely highlight the range of Boonin's book; although it primarily deals with reproduction, the topics that are discussed provide an excellent introduction and a useful analysis to many other topics. We can see from the de re de dicto discussion that it extends beyond reproduction and can encourage the reader to consider different and broader areas. Boonin demonstrates this wide applicability with his comprehensive study of the range of issues confronting the non-identity problem.
Chapter Two covers the problems of creating people, classification of our actions towards them and the effects of our actions; here we find the de re de dicto objection, the metaphysics of generating persons, the problem of a life worth living and the apparent asymmetry between pleasure and pain. Chapters Three and Four cover notions of harm, both in conceptions of harm (for example comparative and non-comparative accounts of harms) and who can be harmed (that is, whether those we create can be harmed by our actions). Chapter Five covers rights-, fairness- and respect-based arguments concerning how we treat people, and specifically considers whether we are constrained and obligated to treat potential or future persons in a certain manner. Chapter Six covers notions of aggregate effect on person-groups and optimal outcomes, thus dealing with Utilitarian and other standards of well-being-based accounts of benefit and harm.
After arguing that these issues fail to show that the non-identity problem is incorrect or inconsistent, Boonin proposes that the only option is to accept the non-identity conclusion. Thus, in cases that produce the non-identity problem it is not morally wrong to have any particular child even if they are disabled or otherwise suffering. As a consequence of discussing this broad range of issues, Boonin makes it easy for someone new to the non-identity problem to engage with the topic and for those familiar with the problem to think of the issues in a more schematic way.
The final chapter is primarily concerned with showing that accepting the non-identity conclusion meets Boonin's own requirement for a satisfactory solution, namely, the independence, robustness and modesty requirements. The independence requirement prevents us from denying the validity of a premise purely on the basis that it allows us to reject the non-identity conclusion. For example, Buchanan, Brock, Daniels and Wikler claim that ‘an adequate moral theory should include as well non-person-affecting principles’ (Reference Buchanan, Brock, Daniels and Wikler2006, p. 250). The reason given for this, as far as can be inferred from their book, is that we should ‘abandon the specific feature of typical moral principles about obligations to prevent or not cause harm which generates difficulty … in genesis cases’ (p. 248). Here, Buchanan et al. fail to provide a reason which is independent from being able to deny the non-identity outcome, unlike Parfit himself, who argues for a utilitarian theory of conduct for reasons independent of the non-identity problem (Reference Parfit1984, p. 446).
The robustness requirement means that our reason for rejecting a premise ‘must be strong enough to warrant rejecting any weakened version of the premise’ (p. 22). The third and final requirement is the modesty requirement, which entails that the reason for rejecting a premise does not lead to a more absurd outcome than the non-identity conclusion itself. Consider, if the property of Wilma's act that is determined to be wrong is found in all human action then every human action is morally wrong and so prohibited, but this outcome is more absurd than the non-identity conclusion itself. This final requirement relies on a fair amount of subjective judgment, but it seems to be clear enough that the reason for rejecting the non-identity problem should not lead to an outcome which classifies all reproductive action or all human action as morally prohibited.
Implications of the non-identity problem
The final chapter helps secure Boonin's conclusion by showing how accepting the non-identity outcome satisfies these three requirements, but he only hints at some of the broader implications for the actions individuals may take. If there is any criticism of Boonin's book, it is that he spends very little time on the implications of accepting the non-identity conclusion (a mere two-and-a-half pages). If arguments against the non-identity conclusion fail, then we should accept that non-identity is the best explanation for the effects of our actions that create people. He says, for example, that his ‘solution tends in the direction of more’ (p. 216) reproductive freedom. Beyond this and a few other equally brief statements, Boonin does not explore the implications of accepting the non-identity conclusion further. This seems odd given the radical implications of accepting the non-identity conclusion.
The non-identity conclusion has wide-ranging and startling implications for a number of legal issues – most obviously for law relating to reproduction and for population policy choices. Accepting the non-identity conclusion makes civil liability for pre-natal harm more uncertain, because no person exists at the time the harm occurs, thus a new person may be created by that action (see Walker, Reference Walker2014a); errors in genetic testing prior to conception would also be affected by the non-identity problem; assessing the welfare of a child created through artificial reproduction would become incoherent; and there are implications for selection and genetic modification of offspring (see Walker, Reference Walker2014b). The non-identity conclusion prevents us from claiming that future people can be made better or worse off, that any future people are entitled to exist, and that we should conserve resources now for future people because only once a person exists can effects happen to them rather than constituting them.
The thalidomide case brings out some of the difficulties in accepting this conclusion. Ordinarily, we would say that a foetus affected by thalidomide would harm the person born because without the introduction of thalidomide the foetus would have developed according to its genomic structure. But this depends on one's conception of a person – our ordinary discourse implies that a person is reducible to their genomic structure. The non-identity conclusion arises because our actions do not affect a person but change which person exists. Where the timing of pregnancy changes so too does the person created. This proposition is less clear when an action occurs during pregnancy – does changing an embryo or a foetus lead to a different person existing?
This is difficult terrain because, unlike the cases where a woman has a child at a different time, there is a lack of metaphysical precision when it comes to a person existing. If one takes the existence of a unique genome to be a person, then everything past fertilisation is a person; if one takes sentience to mark a new person, then this may cover a broad period of gestation; or if one takes self-awareness as a distinct entity as indicating a new person, then only an infant of at least fifteen to eighteen months would count. Thus, there is ambiguity where damage occurs during pregnancy as it depends on when one considers a person to have come into existence. This would limit the application of the non-identity conclusion, depending on how early one places the existence of a new person. The ambiguity problem does not arise when the decision affects the timing of pregnancy – it only arises once pregnancy begins and, therefore, does not apply in indirect population policy cases.
In generational population policies, those born in worse circumstances, due to the adoption of policy A, cannot complain of this fact because the alternative policy B would have led to different circumstances and to the existence of different people. These implications have profound consequences for how we think of reproduction and generational responsibility and liability, yet these discussions are absent from Boonin's account. Perhaps Boonin is reluctant to venture too far into these areas as he is concerned with showing the validity of the non-identity conclusion. Even so, a greater exploration of the implications would have provided a useful and suitable end to his book.
Boonin even seems to underestimate the impact of the non-identity conclusion. He suggests that there is ‘no reason to think that ordinary cases of pollution and of risking pollution are sufficient to generate the non-identity problem’ (p. 216). Yet this seems to underestimate the impact of (to use Boonin's example) the opening of a nuclear power plant which would cause people to move and work in different places than they otherwise would have. Consequently, there is no way for us to know who would have existed in both scenarios and who would exist in only one. This means that we have no epistemological justification for treating the existence of some people as more certain than others.
There is no way to know who will exist as a result of some actions rather than other possible actions, and the sheer number of variables that we would need to comprehend means we cannot know who will exist until they actually do exist. We cannot, therefore, treat any one future person's existence as unaffected by our actions because even though we can comprehend the idea that someone's existence will be unaffected we have no way of determining who those individuals are. Moreover, every future person's existence is contingent on someone's choice and, in this sense, no one future person's existence is certain. Thus, ‘if there do turn out to be particular cases in which we act in ways that are worse for no one’ (p. 216), or if there is no way to know who those worse off people are, then the non-identity conclusion would ‘make it harder to argue against those practices’ (p. 216).
Consequently, Boonin should have explored the implications of accepting the non-identity conclusion further because of the widespread impact such a conclusion would have on ethics, population theories, law, education, culture and religion. However, this should not detract from the great detail and clarity with which Boonin treats the premises of the non-identity problem and the thoroughness with which he works through the issues relating to non-identity. This is an excellent book for anyone wanting to explore the non-identity problem.
