This issue of Journal of Medical Ethics features an “Author meets critics” discussion of Michelle Taylor-Sands' recent book Saviour Siblings.1 In Saviour Siblings, Taylor-Sands departs from standard approaches to the ethics of selective reproduction that (she argues) usually focus on ‘the individual interests of the child to be born’.2 Instead she proposes ‘a new relational model for selective reproduction based on a broad conception of the welfare of the child that includes both individual and family interests’.2 Rather than thinking that the individual interests of the new child are ultimately what (or all that) matter in ethical decision making about selective reproduction, and rather than pitting the individual interests of a potential new child against the interests of other family members, on her approach, ‘the welfare of the child to be born is inextricably linked with the welfare of his/her family’.2 By benefiting a sick sibling and parents—and thus the family as a whole—a savior sibling will often thus herself benefit according to this broad conception of welfare. Though a new child's individual interests still matter and may be at odds with family interests, according to Taylor-Sands, ‘familial duty’ may provide justification for compromising some individual interests … for the welfare of the family as a whole'. This does not mean that family interests always outweigh the interests of a new child; and, according to Taylor-Sands, there should be (policy) limits that protect children ‘from exploitation, abuse, or neglect’.2

As discussed in Stephen Wilkinson's critical commentary,3 part of the motivation behind Taylor-Sands' relational approach to saviour siblings is dissatisfaction with the idea that, according to analyses appealing to the Non-Identity Problem, selective reproduction is justified whenever the new child's life is worth living. Taylor-Sands argues that the case of selecting for deafness is ethically different from the case of savior siblings—because the latter involves avoidable harms, associated with risks of embryo biopsy, for the new child in question. If an embryo is selected because it is deaf, then the child that results is not thereby harmed—because the alternative (for that child) is non-existence. In the case of saviour siblings, on the other hand, according to Taylor-Sands, risks of harm associated with biopsy could be avoided ‘if the child is born naturally … as a result of natural conception’.1 ,3 Wilkinson objects that the embryos produced using IVF and preimplantation genetic diagnosis (PGD) would inevitably be different from embryos produced naturally—and so the alternative for a child that results from an embryo produced when parents decide to employ IVF and PGD in order to produce a saviour sibling would (like in the case of selection for deafness) itself involve non-existence.

Taylor-Sands responds to this objection by arguing that the risks of harm associated with embryo biopsy are incurred after the individual (embryo) in question has already been produced—and so they could have been avoided if the embryo in question was implanted without biopsy. This reply, however, still assumes that we are talking about an embryo produced via IVF as opposed to “natural conception” (as in the quote above). In any case, as Taylor-Sands admits, (assuming we are talking about IVF as opposed to natural conception) there is no guarantee that the same embryo would have been implanted in the absence of testing/selection if, for example, four embryos had been produced (as is usual) and one was randomly chosen for implantation. Her point, however, is that the harms (or risks thereof) in question are not inevitable for the child in question—because there is at least a decent chance that the same embryo would have been implanted in the absence of testing.

Wilkinson's objection, however, is that the same embryos would not have existed in the first place in cases where parents have only employed IVF in order to employ testing to select a saviour sibling. The existence of the embryo in question (in such cases) is thus a product of the decision to test it (along with the others produced) and so it remains the case that the alternative for the embryo in question (and the child it will develop into) is non-existence—and so no one is ultimately harmed (even if biopsy is risky). Taylor-Sands might reply that it is ethically problematic (or at least ethically different from the case of selection for deafness) to decide to produce embryos (i.e., individuals) upon which you plan to impose risk of harm associated with biopsy. Though (being charitable) she perhaps (in the quote above) misspeaks about ‘natural conception’, her reply to Wilkinson appeals to concern about harming individuals after they have been produced. If this is her point, however, then it is no longer obvious (to me anyway) how selection for deafness would be ethically different from saviour siblings (when the former likewise involves a decision to create embryos in order to test—and thus impose risks of harm to—them).

In the second critical commentary, Emily Jackson questions whether we really need a revisionist relational account of welfare to account for the benefits of being a saviour sibling—because (on standard accounts of individual welfare) there are obvious ways in which we might expect such a child to benefit from the role he or she plays in a family.4 In addition to psychological benefits of saving a loved family member, there would be numerous indirect/instrumental benefits for the new child when the family as a whole better flourishes as a result. In reply, Taylor-Sands argues that individual benefits to the new child cannot be taken for granted—e.g., because he or she might suffer psychological harm if the treatment of the sick sibling doesn't work. ‘Given the uncertainty of outcome’, according to Taylor-Sands, ‘it is difficult to defend saviour sibling selection based solely on the interests of the child to be born’.5 Even in cases where the saviour sibling does in fact benefit from family flourishing, in any case, Taylor-Sands would presumably want to argue that this is partly constitutive of the child's welfare rather than a mere indirect or instrumental benefit to the child (via promotion of her individual welfare as standardly conceived). The implications of this for selective reproductive decision making, insofar as interests of the new child are relevant, would thus partly depend on what the welfare metric on the relational approach should look like: How much weight should we attribute to family interests in comparison with other constitutive components of a child's welfare, especially in cases of conflict?

Because Taylor-Sands holds that a new child's welfare may sometimes be compromised in order to promote that of the family, the tension between the interests of the child and that of the family is not altogether avoided by her relational approach to saviour siblings. As noted above, she thus suggests that policy limits will need to be implemented in order to protect children from abuse. Given that it is already common to limit parental discretion regarding decision-making pertaining to children via the ‘significant harm’ test (i.e., state interference with parental decision making is justified when necessary to prevent ‘significant harm’ to children), Jackson questions how different the relational approach would be in practice. Though Taylor-Sands responds by admitting there might not actually be much difference in outcome, she argues that the relational approach to saviour siblings nonetheless provides ‘a clear ethical foundation on which to base the decision-making process’ and that ‘it is particularly important to openly acknowledge relevant individual and collective family interests to facilitate frank and honest discussion’.5 Inter alia, this means admitting that decision-making need not be solely based solely on the new child's interests.

Because potential compromise of a new child's interests is partly justified by Taylor-Sands via appeal to the notion of familial duties, Colin Gavaghan questions the basis upon which such supposed duties are meant to arise. Quoting Taylor-Sands' claim that ‘[w]e owe a higher level of duty to family members than to strangers because we generally favour intimates over strangers’, he implicitly accuses her of the fallacy of ‘appeal to the people’.1 ,6 The fact that most of us think, feel, or do something does not mean that we actually should do so. Taylor-Sands responds that (in an earlier passage of her book) she argues ‘that membership alone forms the basis of family duty … These obligations, which emanate from the relationship itself, are grounded in debt arising out of the receipt of benefits and reciprocity created by mutual reliance’.5 Given that the obligations in question are meant to apply to embryos, an arguably problematic aspect of this reply (given Taylor-Sands' broader line of argument) is that it seems to imply/presuppose that the embryo that is a saviour-sibling-to-be has ultimately benefited (enough) by being brought into existence. Don't individual interests of saviour siblings thus still play a central/crucial role regarding saviour sibling selection after all?