Supernumerary Pregnancy,Collective Harm,and Two Forms of the Nonidentity Problem

“The greatest harm to the children of ART is posed by the probability that they will be born a multiple…Approximately thirty-six percent of all births resulting from [in vitro fertilization] and [intracytoplasmic sperm injection] are multiples…and the numbers continue to grow…Multiples generally are considered a necessary side effect of existing ART practices…It is well-established that multiple births cause serious health problems. Multiples are more likely to be born prematurely…and at [low birth weight]…The mortality rates for multiples are significantly higher than for singletons: for triplets, it is thirteen times that of singletons, and for twins, it is five times…higher…. Even if they survive, multiples (including twins) are more likely to suffer serious physical and mental handicaps, such as cerebral palsy…. In light of this evidence, one noted commentator has wondered why the widespread creation of multiples through ART has not been considered a ‘public health problem’.” Rosato J. L. , “The Children of ART (Assisted Reproductive Technology): Should the Law Protect Them From Harm?” Utah Law Review 57 (2004): 77–78. S. Khanijou notes that “[i]n the last decade, there has been a dramatic increase in multifetal pregnancies as a result of ART…Data published by the Centers for Disease Control…indicate that more than 30% of assisted technology births are multiples compared to the 2% incidence in the general population…The media and much of the public hail multiple births as a ‘miracle,’ as in the case of the septuplets born to the McCaughey family, but the medical community is not so enamored. They have identified such pregnancies as failures, rather than successes, of the IVF enterprise. The press heralded the live birth of sextuplets in California, but the attention soon disappeared when three of the six died. In 1998, one year after the septuplets, the birth of octuplets in Texas made history. But this holiday ‘miracle’ soon turned tragic when one of the infants died and the remaining seven returned to neonatal intensive care soon after Christmas. Perhaps some records were simply not meant to be broken.” See Khanijou S. , “Multifetal Pregnancy Reduction in Assisted Reproductive Technologies: A License to Kill?” DePaul Journal of Health Care Law 8 (2005): at 403, 404. See also Green N. S. , “Risks of Birth Defects and Other Adverse Outcomes Associated With Assisted Reproductive Technology,” Pediatrics 114 (2004): 256–59.

The “no harm done” result turns on a showing that an impaired existence is the only way that a particular child may exist at all. As one court put the point, wrongful life constitutes “a Hobson's choice [between] life in an impaired state and nonexistence.” Becker v. Schwartz, 386 N.E.2d 807, 812 (N.Y. 1978). Notably, the standard “no harm done” wrongful life model may not even be applicable to all instances of wrongful life. Where the child who is negligently brought into existence foreseeably experiences so much pain and misery and so little of whatever it is that we think makes life worth living, it is plausible to say that the child would have been better off never having been born at all and thus is harmed when he or she is brought into existence. See note 11 below.

For discussion of the risks of congenital and chromosomal abnormalities and developmental delays associated with ICSI, see Rosato , supra note 1, at 69, notes 86–88. The criterion of harm set forth in Part III below and elaborated in Part V below implies that, in the case where the agents had no option for bringing the impaired child into existence other than by way of ICSI, that child is not harmed by ICSI despite the fact that ICSI may be a part of the causal sequence that results in impairment to the child. Thus, according to that criterion, where the impairment that is imposed is a necessary condition for the child's very existence, and where that existence is worth having (i.e., at or above the zero level implied by nonexistence, and hence not genuinely “wrongful”), the use of ICSI does not harm the child. Of course, consistent with this result, if agents had available a safer ICSI technique, or the option of substituting the ICSI technique in fact used with some safer infertility treatment, such that the agents could have brought the damaged child into existence without the damage, we can take the position that the form of ICSI in fact used harms the child. The position recently articulated by John A. Robertson seems roughly consistent with this point. See Robertson J. , “Procreative Liberty and Harm to Offspring in Assisted Reproduction,” American Journal of Law and Medicine 30, no. 7 (2004): 14–15. (“In all of the situations surveyed, the child appears to be harmed by the very method of conception, gestation, or social setting of birth. If, as many commentators argue and legislative schemes provide, ‘the prime consideration must be the welfare of the child,’ then use of ARTs in those situations described must be questioned. To prevent the feared ‘injury’ to the child, the person and couple should give up the use of the ARTs that pose those risks…But this leads to a paradox. If offspring are ‘harmed’ by being born in those conditions, then the only way to prevent the harm is to not use those techniques. But this means that the children sought to be protected will never be born. Because their lives will not be so miserable as to be ‘wrongful,’ it would seem that once born they have benefited from rather than been harmed by being born. If that is so, then using ARTs to enable their birth does not harm them and does not justify restriction on those grounds…The logic of this position troubles many people…A key point about the paradox of non-wrongful life is that the person could not have been born without the condition of concern. If so, refusing the act or omission that causes the child to be born with that condition cannot harm the child. Of course, if changes in technique or treatment protocols could reduce the frequency of the condition, there would be an obligation to adopt those changes. However, in situations in which no improvement can be made one cannot show that the child has been harmed as a result.”)

I track Jan Narveson's language here, who writes that “[w]e are in favour of making people happy,” and “neutral about making happy people.” See Narveson J. , “Moral Problems of Population,” in Ethics and Population, Bayles Michael D. , ed. (Cambridge, MA: Schenkman, 1976): at 73.

Parfit D. , Reasons and Persons (Oxford: Clarendon Press, 1987): at 363 (emphasis eliminated in part).

Government regulation in this area can be through the existing tort law framework or through direct regulation or both. “[T]wo visions of federal regulation have been proposed: one by The President's Council on Bioethics (‘Council’); and another by Erik Parens and Lori Knowles of The Hastings Center. Although both proposals are extremely thoughtful, they go further than necessary and ultimately will lead to overregulation of ART.” Rosato , supra note 1, at 90. For the argument that the FDA should restrict or even withdraw certain drugs used to stimulate ovulation from the market, see Noah L. , “Assisted Reproductive Technologies and the Pitfalls of Unregulated Biomedical Innovation,” Florida Law Review 55 (2003): at 603. For further discussion of the constitutional limitations on the government's ability to regulate conduct in the context of supernumerary pregnancy, see Rosato , supra note 1.

We might, for example, eschew the highly person-based “harm principle” associated with John Stuart Mill's On Liberty in favor of an affiliational or a communitarian model of the U.S. Constitution and the state constitutions, and, indeed, of the role of law in society. More narrowly, we might revise constitutional privacy law and the law of negligence to include “Principle N” developed by Buchanan , or Parfit's own same-numbers principle “Q.” According to Principle N, agents are morally obligated to take steps to insure that any new person they produce will have an existence that avoids “serious suffering or limited opportunity or serious loss of happiness or good” in any case in which the agents could have produced a better-off, though “nonidentical,” child in place of the impaired child without imposing “substantial burdens” on anyone, including the agents themselves. Buchanan A. Brock D. W. Daniels N. and Wikler D. , From Chance to Choice: Genetics and Justice (Cambridge University Press, 2000): at 249. Parfit's Q suggests that it is morally better to bring a better-off person into existence rather than a worse-off person. See Parfit , Reasons and Persons, supra note 5, at 369. For a critical discussion of these approaches, see, among other sources, Robertson , supra note 3, at 15–16.

This is so, whether the impersonal approach under scrutiny is a traditional form of utilitarianism (sometimes called “totalism”) or a form of consequentialism that considers the overall good to be constituted not by aggregate wellbeing alone but by other values as well, such as equality (“pluralism”). The repugnant conclusion, the infinite population problem and an array of extreme inequality problems clearly challenge totalism. While pluralism avoids some variations on each of those problems, other variations are more difficult for the pluralist to provide a plausible account of. The very case at hand illustrates potential problems with a move to totalism. Totalism, applied to ITISP, seems to generate, depending on the specific facts of the case, what is in effect a mini-repugnant conclusion: the result that it is perfectly fine, indeed, obligatory, to continue, or even to commence, a supernumerary pregnancy in order to bring six, seven, or eight babies into existence in one fell swoop rather than just a single healthy baby, on the grounds that the former option is the one that increases total wellbeing. For additional discussion of these points, see Roberts M.A. , “Supernumerary Pregnancy and the Limits of the Constitutional Privacy Guaranty,” Journal of Philosophical Research (special supplement), Adams F. , ed. (Philosophy Documentation Center, 2005): 105–117; and “A New Way of Doing the Best That We Can: Person-Based Consequentialism and the Equality Problem,” Ethics 112, no. 2 (2002) 315–50.

Such criteria fall into one of two classes: “adversity-based” approaches to harm and “substitutional” approaches to harm. Adversity-based criteria discern harm in any cases involving substantial pain or hardship, even when that pain or hardship is itself physically necessary to secure some clear good for the very individual who suffers the pain or hardship. Such criteria thus by design overlook the fact that acts that impose pain or hardship may in some cases also happen to be the very acts that maximize wellbeing. The adversity-based approach basically equates harm with something like pain: an unpleasant or miserable experience that is not in itself a plus but that we immediately understand to be something we may very badly want simply because, all things considered, it's good for us. Here we might point to Rosato's apparent identification of harm with any serious physical or emotional adversity, or Elizabeth Harman's explicit identification of harm with “significant bodily damage” – including, e.g., the damage that is done when the surgeon makes an incision in one's abdomen for the purpose of saving one's life through the removal of a ruptured appendix. Thus Harman writes: “I claim that causing pain, early death, bodily damage or deformation is harming.” Harman E. , “Can We Harm and Benefit in Creating?” Philosophical Perspectives (Ethics) 18 (2004): at 92, 93. “[T]he fact that adopting the Policy would harm her [in this sense] is a reason against adopting the Policy,” at 92. “[A]dopting the policy is…wrong because we have an alternative that would involve parallel benefits, without parallel harms,” at 93. Substitutional criteria provide that (with certain qualifications) harm has been done to one person just in case agents could have brought a distinct (“nonidentical”) better off person into existence in place of that one. Peters Philip G. Jr. has developed a substitutional approach to harm. Peters P. G. Jr. , How Safe is Safe Enough? (Oxford University Press, 2005). As Robertson observes, however, Peters' approach “only discerns harm in a case where the agents have the option of substituting a better off person in for a less well-off person. In the context of fertility treatment, however, that precondition may very well not be satisfied…. The [same number] principles that Brock and Parfit enunciate are attractive, and can be adopted without accepting a full-scale utilitarianism in all areas of life. But the need to keep the numbers the same and not unreasonably burden parents in making substitutions constrains its impact.” Robertson , supra note 3, at 16.

For brief sketch of Peters' substitutional theory of harm, see note 9 above.

Consistent with the view that the provider's negligent act does not impose a procreative harm on the child who owes his or her very existence to that negligent act, we may think that that same act imposes distributive harms on persons other than the child and, perhaps, on the child as well. Thus, while the procreative harm may itself be unavoidable if the child is to exist at all, that same proposition does not hold for the various distributive harms that may be separately imposed on, for example, the child's siblings as well as the child himself or herself, depending on how the parents go about balancing their own scarce resources among family members. Distributive harm to the parents is recognized by almost all courts as creating the basis for anaction for “wrongful birth.” The suggestion here, however, is that such distributive harm to both the child's existing and future siblings as well as to the child himself or herself may also create the basis for a cause of action. See, e.g., Roberts M. A. , “From Chance To Choice To Harm To Persons: Why We Should Favor a Person-Affecting Account of Wrongful Disability” (unpublished manuscript, on file with author). A second point is important in this context. Courts may be correct in holding that most wrongful life cases – in which the child is born, say, deaf, or perhaps with Down syndrome – as a matter of empirical fact do not involve procreative harm to the child. However, it seems a mistake for any court to hold that, as a matter of law, the case in which the existence is so full of pain and misery as to itself constitute a harm cannot arise. Surely, after all, we understand that, at least in rare cases, never having existed at all will constitute the better alternative from the perspective of the individual who exists and suffers. For further discussion of the cogency of the wrongful life claim, see “Can It Ever Have Been Better Never to Have Existed At All? Person-Based Consequentialism and a New Repugnant Conclusion,” Journal of Applied Philosophy 20, no. 2 (2003) 159–85.

The federal district court in Morgan v. Christman , a case discussed aptly by Lars Noah in a recent paper, thus seemed to get this point just exactly right. See Noah , supra note 6, at 603.

At present, doctrines of intra-family immunity suggest that that patient will not be held liable for the harm she has done her offspring. At the same time, laws against parental abuse and neglect may be applicable. What account of harm we ultimately adopt will be of critical importance to whether doctrines of intra-family immunity require revision and just how laws against parental abuse and neglect apply in the ITISP context.

In part to capture the causal element in harming, we may require for harm both an increased risk as well as a flawed outcome. If so, then MCH1 could be revised to incorporate not just the concept of actual wellbeing but also the concept of expected wellbeing.

This is the approximate description of the plight of the Chukwu octuplets. For a description of that case, see Chester R. , “Double Trouble: Legal Solutions to the Medical Problems of Unconsented Sperm Harvesting and Drug-Induced Multiple Pregnancies,” St. Louis Law Journal 44 (2000): at 451, 462–63. See also Roberts , “Supernumerary Pregnancy and the Limits of the Constitutional Privacy Guaranty,” supra note 8.

I am grateful to David Wasserman and Arthur Caplan, respectively, for these points.

My suggestion here, then, is that the person-based consequentialist might appeal to the principle of “leximin” – the idea, that is, that in a trade-off situation, the agent's obligation will be first to increase the wellbeing of the least well-off, then to increase the wellbeing of the next least well-off, etc. For further discussion of this point, see Roberts M. A. , “A New Way of Doing the Best That We Can,” supra note 8.

Coleman C. H. , “Conceiving Harm: Disability Discrimination in Assisted Reproductive Technologies,” UCLA Law Review 50 (2002): at 44–45.

Coleman , supra note 18, at 46–47, citing Steinbock B. , Harm and Future Persons, 2000 (unpublished manuscript); see also Steinbock B. and McClamrock R. , “When Is Birth Unfair to the Child?” Hastings Center Report 24 (1994): at 15, 16.

Coleman , supra note 18, at 44–47.

Emphasis added. Consistent with this point, Robertson elsewhere acknowledges that, even outside the case where the life is less than worth living, harm to offspring may be imposed by ARTs in cases where agents have some way of “improving welfare” for offspring – some way, that is, of avoiding imposing “harm” on them while still bringing those same persons into existence. In this context, he cites the case of “multiple gestation.” Robertson , supra note 3. Robertson then goes on to say that, “[u]nfortunately, few ART situations will be covered by the principle of avoidability.”

Supra note 1.

See Child Versus Childmaker: Future Persons and Present Duties in Ethics and the Law (Rowman & Littlefield, 1998): at 179–211, Chapter 5; “Cloning and Harming: Children, Future Persons and the ‘Best Interest’ Test,” Notre Dame Journal of Law, Ethics & Public Policy 13, no. 1 (1999): at 37–61; and “Human Cloning: A Case of No Harm Done?” Journal of Medicine and Philosophy 21, no. 5 (1996): at 537–54. If the man-made proliferation of my genome counts as a diminution in wellbeing for me, that is, a harm to me – and I think that it is must, else we would not insist on the right of consent to the use of our genomes in connection with the production of others via cloning – then we can explain why cloning harms the very people it helps to bring into an existence worth having. Suppose, for example, that four of us have been produced by way of cloning – to keep things simple, let's say by way of embryonic cloning. By logic analogous to that used in Part III of this present paper, agents had the option of creating additional wellbeing for me – that is, the option of bringing me into a still better existence. All they had to do was bring me into existence and refrain from bringing the remainder of my cohort into existence. Rather than transferring all four embryos to one or more uteruses for development, they could have, in other words, transferred just the single embryo, the one who becomes me, and let the others perish. But that means they had the option of creating more wellbeing for me when they instead created less. And that means harm.

I treat here an issue of harm. But issues of causation arise as well. Green Michael D. discusses the intersection of these issues. Plausibly, what the provider has done at least should be counted as a “cause” of the outcome in which the baby both exists and suffers. At least, the provider's acts would seem to satisfy the NESS (the “necessary element of a sufficient set”) criterion for causation described by Green as having been developed by Wright and based on the work of Hart and Honore. Since the patient might always have gone to another provider, however, those same acts would seem not to satisfy the traditional “but for” test for causation. Green M. D. , “The Intersection of Factual Causation and Damages,” DePaul Law Review 55 (2006): at 671, 683 (citing Wright R. W. , “Causation in Tort Law,” California Law Review 73 (1985): 1735, 1795; and Hart H. L. A. and Honore T. , Causation in the Law (2nd ed.) (Oxford University Press, 1985): at 239).

“Cause the tortfeasor's act is, but liability should not follow.” Green , supra note 24, at 700.

Parfit raised this question, though in a different context, in a recent talk entitled “What We Together Do: Questions About Collective Responsibility,” University Center for Human Values, DeCamp Lecture (Princeton University, October 10, 2005).

The Restatements of Torts describe classic cases involving reciprocally limited options as well as the broader problem of causal overdetermination (or multiple sufficient causes). In one such case, two fires have been negligently set. Each fire is sufficient to destroy the owner's property. The fires then converge on the owner's property, destroying it. In that case, the agents have, by hypothesis, each engaged in a negligent act; they have, that is, acted (or failed to act) in a way that is deemed to breach a duty of care they owed to the property owner. A variation on the two fires case involves two agents negligently firing their guns and hitting the same eye of the same person at the same moment. Case III, in contrast, does not involve a double shooting. Rather, b2 simply stands to the side, prepared to shoot if b1 doesn't, as b1 in fact does the shooting. But it is not b2's simply standing there that makes Case III an interesting one, nor his churlish failure to rescue (he can't). It is rather the fact that – like the second fire in the two fires case – he is limiting the options that b1 would otherwise have had to just one: the boy's being shot. Michael D. Green describes the view expressed in the Restatement (Third) of Torts: Liability for Physical Harm 27 cmt. i (Proposed Final Draft No. 1, 2005)(William C. Powers, Jr. and Michael D. Green, reporters): “If two hunters fire negligently at their quarry and their bullets arrive in the plaintiff's eye at the same moment, is either one a cause of the plaintiff's harm? This is just a variation on the familiar ‘two fires’ hypothetical that requires some consideration of the role of the sine qua non test for factual cause. Each hunter is a cause of the harm and liable for the full extent of damages suffered by the victim. Sine qua non will not do here, but there is no doubt that instinctively, rationally, and normatively, we are entirely comfortable describing each hunter's action as a cause of the victim's harm and holding each liable for damages affiliated with the loss of an eye.” Green , supra note 24, at 681. Green notes that the draft Restatement takes a similar position in the case where the overdetermination is due not to two tortious acts but rather to one tortious and one innocent act: “In conclusion, causation has taken the day and determined that the tortfeasor who concurs with an innocent cause is liable for damages reflecting the overdetermined harm. Note that the audience is no worse off because of Landsman's tortious (and heinous) act, as they would have suffered equally in a tort-free world.” Green, at 687. Thus, the relevant section “applies regardless of whether the other cause involves tortious conduct or consists only of innocent conduct. As long as the competing cause was sufficient to produce the same harm as the defendant's tortious conduct, this Section is applicable.” Green, at 687 (citing draft Restatement at 27 cmt. d). MCH2, below, similarly addresses the case of the one tortious and one innocent act in the same way it addresses the case of the two tortious acts. Another challenging scenario arises when both an agent's act and some natural event are sufficient to impose a diminution in wellbeing for p. Do we still say that the agent has harmed p? MCH2, as drafted, will not have that implication. We are left then, under MCH2, with the result that the agent may be considered to harm p when another agent's innocent act is sufficient to cause that same harm but not when nature is sufficient to cause that same harm.

I note that a rule for when agents may defect from a beneficent group project that parallels the strategy behind MCH2(2) (c) in certain ways has been suggested by Joseph Mendola. See “Consequentialism, Group Acts, and Trolleys,” Pacific Philosophical Quarterly 86, no. 1 (2005): 64–87. Cases analogous to Case IV abound. For example, do slaves who lack any ability to escape their plight harm their own child when they bring that child into existence as a slave? MCH2, plausibly, suggests that they do not. Since the slave-parents, in producing the child as a slave, have thereby done as much for their child as they can, they have not, acting as individuals, harmed that child under clause (1) of MCH2. The issue of group-harm is somewhat more complex. It seems that a group-harm has been imposed on the child, given that the group consisting of the parents together with the various agents who maintain the institution of slavery had the ability to make things better for the slave-child than they are. But in this particular case the parents' non-participation in the causal sequence that ends in the child's existence as a slave would not have created as much wellbeing for that child as he or she himself in fact enjoys. The parents' non-participation, in other words, would have made things still worse for the child. (We here make the assumption that the particular existence, despite its flaws, is better, from the child's own point of view, than never having existed at all would have been.) Since what the parents have done thus does not satisfy clause (2) (c) of MCH2, we infer that the parents have not, acting now as participants in the group-harm, harmed that child. Clause (3) of MCH2 then provides that the parents have not harmed the child at all.

In real life, providers do have more sway over what the patient decides, and more options, than b1 has with respect to b2. Suppose that the patient informs the provider that she will not undergo any treatment that includes any safeguard against supernumerary pregnancy. And suppose her reasons are religious, or “ethical,” in nature, and have nothing to do with costs: she just thinks that once those egg cells pop out and begin to drift downwards in the direction of the fallopian tube, ripe and ready for fertilization, the “be fertile and populate the earth” principle or perhaps the still more important and morally weighty “more the merrier” principle is triggered. The provider informs the patient, then, that that's not how he (suppose) or his clinic does business: no fertility treatment can be made available in the absence of precautions to avoid supernumerary pregnancy. And he explains to the patient, in the most graphic terms, why that is the case. Is this option – clearly available to the provider – one that really so obviously would have failed to create additional wellbeing for the disabled child? Isn't the woman highly committed to having a child just as likely to accede to the provider's terms? Perhaps, by taking a firm stand, the provider does reduce in some small way the child's chances of existence – but surely many supernumerary pregnancies themselves are spontaneously aborted and surely a more limited pregnancy has a greater likelihood of success, so there may be something of a wash here. Especially in the case where after the fact, in the context of a civil action, the woman then asserts that, had the provider made precautions against supernumerary pregnancy a condition of treatment, she would have acceded to that condition, despite her vague religious or moral protestations to the contrary, I am not sure that we would not or should not find her testimony credible.

Notably, the option is not really so far-out. Imagine the reaction if a pediatrician charged $100 for stitches administered under a sterile procedure but only $50 for stitches administered otherwise. Do we really think that if the wound then becomes seriously infected the pediatrician has not harmed the child? Don't we rather think that the pediatrician is welcome to charge $50 for the service rather than turn the child away but, having accepted the child as patient, must now clean the wound before stitching it?.

I discuss this particular type of nonidentity problem – the “can't-expect-better” problem – elsewhere. See “The Nonidentity Fallacy: Harm, Probability and Another Look at Parfit's Depletion Example,” Utilitas (forthcoming, 2006).

Kavka G. , “The Paradox of Future Individuals,” Philosophy & Public Affairs 11, no. 2 (1981): 93–112.

Just think: had your parents not attended just the schools that they did, would you have ever existed at all? Had they not married or fallen in love with each other just when they did, would you have ever existed at all? Had they had dinner at one restaurant rather than another the night you were conceived, would you have ever existed at all? The problem is not so much that new eggs are produced about every month or so. The problem is rather all those sperm: millions competing to inseminate a given egg during a given cycle, and many, many factors at play to determine just which one will succeed.

Supra note 31.

That we think probabilities have a role to play in determining whether harm has been imposed does not mean that actual wellbeing does not have a role to play in determining harm as well. Thus, we are free to say that an act that creates a less-than-maximal quantity of expected wellbeing for a given person p, but turns out, contrary to expectations, to be the objectively better-for-p act in light of how the future in fact unfolded, is an act that does not “harm” p. In the case at hand, however, this condition on harm would be satisfied given that the baby is in fact born damaged. Where this condition would come into play to avoid a.