Introduction
In several papers,Footnote 1 John Harris has argued that if
i) it is morally permissible to engage in reproduction, whether natural or artificial, despite knowledge that a large number of embryos will fail to implant and quickly die, then
ii) it is morally permissible to produce embryos for other purposes that involve killing them, for instance, to harvest stem cells that may be used to save lives.
For the purposes of this paper, we make an assumption that Harris also makes, namely, that, on average, only one out of five embryos results in a live birth.
Harris seems to use (i) to argue also that
iii) it is morally permissible to use the spare embryos produced in the course of artificial reproduction for other purposes (than implantation) that involve killing them, such as the harvesting of stem cells.
Perhaps he thinks that he need not bother to keep (ii) and (iii) apart because (i) supports one as much as it supports the other. We would, however, like to concentrate on (ii). Our argument is to the effect that there is a plausible view of the moral status of human embryos, according to which the truth of (ii) does not follow from the truth (i). Nor does the truth of the weaker claim (iv) follow:
iv) It is morally permissible to produce embryos for purposes other than implantation if the purposes for which they are produced include the purpose of implantation.
However, even if the truth of these propositions does not follow, they may still be true. This is what we believe, and, at the end of the paper, we outline an argument that, with the help of (i), establishes the truth of (ii), (iii), and (iv). This argument turns on the idea that an embryo's potential to develop into a human person might not be actualizable.
Harris wants to target those who accept natural reproduction—which includes those who believe that the embryo has a moral status and a right to life—but who oppose embryo research. He aims to show that they are logically inconsistent in engaging in a practice that foreseeably results in the death of a majority of embryos but refusing to engage in a practice that intentionally kills embryos. We argue that there is no logical inconsistency in engaging in natural reproduction with a high embryo loss but opposing embryo research. There is a possible view of the moral status of embryos that would make it at least logically consistent to support natural reproduction with high foreseeable embryo loss but oppose the intentional killing of embryos. Moreover, this view of moral status is not the controversial view that embryos have a right to life. However, we offer a revision of Harris's argument that would force even those who take this possible view of moral status to concede that natural reproduction is acceptable only if embryo research is acceptable.
Embryos Have No Right to Life
Before we propose this possible view of moral status of embryos, imagine that we adopt the apparently common view that a human embryo is “one of us” with a right to a worthwhile life. If in the process of creating one being who will lead a worthwhile life we have to create four other beings with rights to worthwhile lives who will lead only short, truncated lives that could not reasonably be said to be worthwhile, it seems that we should abstain from creating the one. This is because this result would be achieved at the cost of violating the rights of four embryos. Thus, if one accepts that the embryo has a right to a worthwhile life, it seems more consistent to reject (i) as false than to regard it as true. That is, that it is wrong to engage in natural reproduction with high rates of embryo loss. This, of course, would be something of an embarrassment to the people who take this view of the moral status of embryos.
The following analogy may make this clearer. Suppose that a couple knows that, for some reason, they have to have four babies so severely disabled that they will die early in infancy, without having developed any consciousness and, so, without having experienced either enjoyment or suffering, in order to have a healthy child. Then many would think that it would be wrong were the couple to set about to have a healthy child. They may think this for the reason that it would be wrong to violate the rights of four children in order to have a fifth.
For our own part, we do not share this view. We think that the couple would do wrong to the four babies only if they were made to suffer.Footnote 2 But this is because we deny that human beings who have never been conscious could have any rights. However, we are now considering what is implied by the common view that anything human has a right to worthwhile life. We believe it to imply that it would be wrong to create the four babies and, similarly, that it would be wrong to engage in natural reproduction if it involves the death of four out of five embryos.
Therefore, if natural reproduction is to be permissible, as it normally is, the moral status of embryos must be lower than that of babies, who are assumed to have a right to worthwhile life. For if they have a right to worthwhile life, we will violate this right by bringing them into existence when we know that they will quickly die without getting anything out of life. The greater the number of such beings we bring to life, the more rights we will infringe, and the greater the wrong we will do, other things being equal. But in the embryo case, there seems to be no wrong that mounts up: No matter how many such embryos we need to create in order to get a healthy child, it seems permissible to do so. Thus, embryos have no right to worthwhile life.
It should be stressed that if it is agreed that it is wrong to have the four severely disabled babies before having a fifth healthy one, this is not because it involves using the four babies as (mere) means to the fifth. Our illustration did not specify that the four babies are used as means to the fifth. That is, it did not specify that the creation of the four severely disabled babies is a causally necessary condition for the creation of the fifth healthy one. The situation may instead be as in the case of, say, a high jumper who had to make some failed attempts before he succeeded for the reason that, because there are uncontrollable factors, several attempts are needed to make success (practically) certain. This seems to be the situation in natural reproduction: The production of the embryos that fail to implant are not causal means to the production of the ones that successfully implant. As in the case of the high jumper, success might come in the first as well as in any later attempt, for all we know.Footnote 3
There is one obvious difference between stem cell research and natural reproduction. Whereas natural reproduction involves letting embryos die rather than killing them, stem cell research involves actually killing them. This is a morally relevant difference according to most deontologists and some consequentialists. If this is a morally relevant difference even in the case of human embryos, it will open up a gap between (i) and (ii). In reply, it might be argued that when the moral status of embryos is adjusted so that (i) becomes true—that is to say, embryos are denied a right to worthwhile life, so that there is nothing wrong in bringing into existence embryos with stunted lives—it follows that the killing/letting die distinction has no moral import for them (even if it has in the case of beings with a right to life). In any case, we will make the assumption that there is no difference between killing embryos and letting embryos die. The gap between (i) and (ii) that we insist upon is a different one that is based on another view of the moral status of embryos. To bring this out, let us begin with an analogous case involving people instead of embryos.
Selection by Fair Lottery
Suppose five people have contracted a terminal illness due to no fault of their own. They are now unconscious and will die from the disease without waking up unless they are given a substance D, which is present in small, insufficient amounts in each of their bodies. The only way to obtain enough D to save a life is to let four of them die by not giving them any life support (or killing them) and extracting D from their dead bodies. Enough pooled D can then be injected into the body of the fifth patient with a life-saving effect. This means that four patients will, in fact, be used as means to the end of saving the life of the fifth. Imagine further that there are no medical or other morally relevant grounds (such as life expectancy, social usefulness, etc.) to select any one particular patient for survival.
One obvious way to select the survivor would be by means of a procedure that is a fair lottery, FL for short (say, each patient is given a random number, and the winning number is drawn from a ballot). Another way would be by a different procedure that we call direct selection, DS, to walk around in the ward and directly select the survivor on some ground that is morally irrelevant, for example, personal liking, race, sex, height. (Remember that we have stipulated that there is no morally relevant ground for selection.)
Now it is possible to claim that in this case, Five Patients, it is morally permissible (we leave aside whether it is obligatory) to save one patient if we use procedure FL but not procedure DS. This could be so even though we will, in fact, use four people as a means to saving one whichever method we use. The justification for this difference could reasonably be that, if we choose method FL, we do not regard any patient as a mere means to the welfare of any other; we rather simultaneously regard each patient as an end (or, more intelligibly, as having ends). For we regard each patient as an end to an equal extent by giving her the same chance to survive (and pursue her ends) as any other. This is not so if we opt for DS. By picking out one patient in the ward on morally irrelevant grounds, we would by implication be regarding the others as mere means to her survival.Footnote 4 We would be disregarding the fact that they, too, have ends. This is also why all five patients could rationally consent to method FL being used if, before being unconscious, they were asked what should be done. They could justifiably object to DS being used on grounds of unfairness and to the policy of doing nothing on the ground that it involves needlessly letting one of them die. But because FL is in everyone's self-interest, it is justified by both contractualist (based on rational egoistic considerations) and utilitarian theories.Footnote 5
This should not be taken to imply that the usage of DS is never morally justifiable. It could be justifiable when sufficient D is something that you own or are entitled to use as you see fit. Thus, suppose that one of the five patients is a brilliant scientist who has higher than normal amounts of D in her body. Because of her extraordinary scientific knowledge, she can extract a sufficient amount of D from her own body to save one of the other patients. If she chooses to do so—perhaps at the expense of dying herself—she could be allowed to give D to whomever of the four she likes best. This is analogous to the familiar case in which, at the expense of your life, you are allowed to choose to save your own child or spouse rather some stranger.
Notice that our reasoning in Five Patients also excludes a combination of the two procedures, if this decreases the chances of survival of any of the patients. Suppose that only one of the five needs to be sacrificed to save the other four. Then it would be impermissible to pick out two patients and run a fair lottery between just the two of them to settle who should be sacrificed for the sake of the other four, because this would reduce the survival chances of the two from 80% to 50%.
Notice also that, as previously, it is not crucial that Five Patients involves the use of anyone as a means. For suppose we could only save one of the five, but the means were provided by something other than the other four. Then it would still be impermissible to pick out directly the one to be saved rather than to do so by means of a fair lottery. It is the fact that the four are treated merely as entities having no ends of their own that is objectionable. If you do not take into consideration that a being with ends has ends when you propose to do something that negatively affects the ends of this being, you obviously risk acting wrongly toward it.
Why Harris's Argument Fails
This provokes the objection that it is precisely at this point that the analogy between Five Patients and the case of using embryos for the cultivation of stem cells breaks down: embryos do not have ends, so it cannot be wrong to treat them merely as entities that do not have ends. Those who, like us, want to argue that (i) may be true (i.e., that it is permissible to engage in reproduction) without it logically following that (ii) is true (i.e., that it is permissible to destroy embryos for research) must avoid a dilemma. On the first horn, if they endow embryos with ends, they run the risk of making (i) false—because their moral status may be hard to distinguish from that of persons (as it will be if we take embryos to have a right to life)—whereas on the second horn, if they deny that embryos have ends, they may undermine their case for disputing (ii).
It may be possible, however, to avoid this dilemma and block Harris's argument by employing the familiar idea that embryos have a potential to become persons with ends. Commonsense morality apparently gives some moral weight to the potential of embryos to become persons. It could then be claimed that we must take this into consideration and not treat them merely as things without such a potential. For present purposes, it is not necessary to defend this view of human embryos. We merely wish to show that, in this view, Harris's claim that his argument from (i) to (ii) has the virtue of sidestepping debates about the moral status of embryos is mistaken.
Harris's argument from (i) to (ii) is like saying that, because the application of method FL is morally permissible, so is the application of method DS. For reproduction, whether natural or artificial, is analogous to FL, whereas producing embryos for life-saving stem cell research is rather analogous to DS. This is so because, in the case of research, the embryos created are given no chance of survival—it is decided beforehand that the ones who will (possibly) benefit by the practice are outside that class. But in reproduction, for all we know, any embryo produced, at the time of production, has some chance of surviving and developing into a person who will lead a worthwhile life, and this chance is greater than the zero chance there would have been in the absence of embryo production. Because, as one of us has elsewhere claimed, “imposing risk on the embryo can be acceptable if it is the best strategy to improve the chances of that embryo having a healthy happy life as a person,”Footnote 6 (i) may be true without (ii) being so. In contrast to (i), (ii) consists of using embryos merely as means: it gives them no chance to develop into persons who successfully pursue ends. Hence, Harris's inference from (i) to (ii) is invalid. Furthermore, as our elaboration of Five Patients shows, his argument from (i) to (iv) (i.e., the claim that it is permissible to produce embryos for purposes other than reproduction if one of the purposes one has is that of reproduction) is also invalid.
More generally, Harris fails to appreciate that natural reproduction is different from deliberately creating and killing embryos for research. Reproduction does not treat embryos as a mere means whereas research does. Natural reproduction gives each embryo the greatest chance possible of survival and shows no greater concern for anyone's welfare. When an embryo is created to be destroyed in research, there is no concern for its welfare, whereas there is concern for some others. The research embryo is used as a mere means for the benefit of the latter. Embryo research (insofar as it creates some embryos with no chance of survival), but not natural reproduction, treats the embryo as a mere means.
Harris considers a version of this argument against his position and responds as paraphrased here:
If this argument is persuasive against the production of research embryos it is easily answered by ensuring that the production of research embryos to some appropriate extent mirrors that of the production of embryos in normal sexual reproduction. One would simply have to produce more embryos than were required for research, randomise allocation to research and ensure that the remainder were implanted with a chance to become persons. To ensure that it was in every embryo’s interest to be “a research embryo” all research protocols permitting the production of research embryos would have to produce extra embryos for implantation. So that for every, say, 100 embryos needed for research another 10 would be produced for implantation. The 100 embryos would be randomised 90 for research, 10 for implantation, and all would have a chance of survival and an interest in the maintenance of a process which gave them this chance.Footnote 7
This response fails. Imagine Embryo Nightmare. We create five embryos and sacrifice four for research and implant one. Harris's argument implies that the odds of each of the embryos is the same as in natural reproduction. This is incorrect. Every embryo created in natural reproduction has a 1/5 chance of surviving. Each embryo created in the Harris Embryo Nightmare has a 1/5 chance of surviving naturally and a 1/5 chance of escaping sacrifice for research. Each embryo has a 1/25 chance of becoming a baby. Thus the Harris Embryo Nightmare does not treat the interests of the embryos with equal concern, as it reduces their chances of survival fivefold. This procedure does not treat research embryos as mere means, but it fails to show them equal concern by reducing their chances of survival for the benefit of others.Footnote 8
In the case of Five Patients, there is equal concern and respect because all could consent to a fair lottery that gives each the greatest chance of survival. Harris is famous for his Survival LotteryFootnote 9—which involves killing one healthy person to provide organs for a number of others. Applied to the case of Five Patients, this means killing one healthy person to save or prolong the lives of four others. Harris rejects a number of objections to the Survival Lottery. But he does not entirely endorse it. Now, there is one crucial difference between the conditions of Harris's lottery and our case Five Patients. In the Survival Lottery, the people participating in the lottery may be responsible for being afflicted by the disease(s) to different degrees. We have stipulated that this is not so in Five Patients, to make this case as closely analogous to that of the embryos as possible. But suppose that some people choose to smoke tobacco, although they are aware that they thereby increase the risk of suffering from illnesses like lung cancer. Then it is arguably not fair or just that, in a lottery designed to fight illnesses like lung cancer, these people are subject to no greater risk of being sacrificed and stand an equally great chance of being benefited as people who have refrained from activities known to cause these illnesses. Whether or not a lottery with these conditions would maximize utility, it would not be fair or just. So, we should not be taken to imply that equal concern for individuals’ ends of leading good lives is always morally justified. What matters in addition to utility, or maximizing expected consequences, is fairness, and fairness may not always consist of equality, because there might be morally relevant grounds for giving priority to some.
An Alternative Argument: Actualizable Potential
Our conclusion that Harris fails to establish (ii) and (iv) should not be taken to imply that we think that these propositions are false. We do think that these propositions, as well as (iii), are true, and that there are other arguments that show this. Harris does not mention that when it is proposed that embryos be created for research purposes, the alternative will normally be that they will not be created at all. Creation for reproduction is out of the question because no woman wants to have the resulting embryos implanted. Therefore, Harris's argument might mislead us into thinking that the alternatives are the following: either create certain embryos for other purposes, such as research, or for the purpose of reproduction. This makes appropriate analogues like the choice between methods FL and DS. But this would be to overlook that when the creation of embryos for research purposes becomes an issue, creation for reproduction is almost invariably not a realistic alternative. So, the relevant question is whether, when implantation of embryos is excluded, we can act wrongly toward an embryo by creating it and using it as mere biological material rather than not create it all. This would presumably be wrong if it were wrong to create embryos with very short, truncated lives. But then it would also be wrong to engage in reproduction because of all the embryos wasted, as we have argued.
We therefore suggest that the moral status of embryos be downgraded, so that it would not be wrong to create any number of embryos in order to have a healthy child. This could be the case in the potentiality view we have sketched, because an embryo that has a claim to be something more than an object to be used merely for the benefit of other things, in virtue of its potentiality to become a person, will lose this claim when implantation is ruled out. Its moral standing will then be on par with that of a pair of gametes.
This is because embryos that will not be implanted can be said to have no actualizable potential. When implantation is excluded, we cannot act wrongly toward an embryo if, instead of preventing this, we let a pair of gametes unite to form it for it to be used as a mere means to some end. This also explains why it is permissible to use spare embryos, which cannot be implanted, in stem cell research, as (iii) above asserts. They have no moral standing because they have no actualizable potential.
But what is a potential that is (not) “actualizable”? An embryo has the potential to become a person if it has the appropriate genes, proteins, and other internal structures that, given a certain environment, enable it to develop into a person. Whether this potential is actualizable depends on whether the relevant external conditions, be they biological or social, obtain that are sufficient to enable the embryo to realize its potential. As we conceive of the concept of actualizable potential, the concept of potentiality refers to states internal to the being and the concept of actualizability refers to states external to the being that together cause it to actualize its potential, for example, a normal human embryo that has a potential to become a person to become such a person. Thus, such an embryo in a petri dish lacks actualizable potential to become a person because it is not in an environment that allows it to become one, whereas such an embryo in a normal uterus has this actualizable potential. Likewise, a baby born in a civil war and extreme famine may have no actualizable potential to become an adult person, for, due to social conditions, that individual will never become a person.
The argument with which we would like to replace Harris's would then run something like this. If
i) it is morally permissible to engage in reproduction, whether natural or artificial, despite knowledge of the large number of embryos that fail to implant and quickly die, then
ii*) it is morally permissible to produce embryos for other than reproductive purposes that involve killing them, for example, to harvest stem cells that may be used to save lives, when to produce them for reproductive purposes is not a viable alternative.
The argument from (i) to (ii*) holds good, whereas Harris's argument from (i) to (ii) does not, because the addition that makes up the difference between (ii) and (ii*) rules out the actualization of the potential upon which any moral status of embryos hinges. Furthermore, it is legitimate to move from (ii*) to (iii) (the permissibility of destroying spare embryos for research), if “spare embryos” are understood as embryos that will not be implanted. It is also legitimate to move to (iv), that is, the creation of embryos for several simultaneous purposes, one of which is reproduction, because these purposes are the ones permitted by (i) and (ii*). The pivotal claim is that it is legitimate to create and destroy embryos when reproductive needs have been fully met because those embryos will have no actualizable potential.
Conclusion
In this paper, we have examined Harris's argument that if we believe it is permissible to engage in natural reproduction when four in five embryos perish, it is permissible to create and destroy embryos for research purposes. Harris's own version of this argument fails. But a related argument succeeds. When implantation of an embryo is not an option, that is, when reproductive needs have been met, it is permissible to create and destroy embryos for research. And it is, of course, permissible to use embryos that have already been created and that will otherwise be destroyed. This might be the case when an embryo has some significant abnormality that makes it unfit for implantation. Few would object to experimentation on abnormal embryos that will be destroyed. Indeed, there is broad-ranging majority support for the donation of embryos for research by infertile couples who have completed their families and would otherwise destroy their embryos. Virtually all countries, including conservative administrations in the United States and Germany, accept the destruction of so-called spare or excess embryos. We have sought to elucidate principles that could ground this view and to show that they extend to supporting the creation of new embryos for research when further reproduction is not an option.
Our argument is consistent with common attitudes to reproduction. When a population has reached its optimum size or a family is complete, we and many other people believe it is acceptable to use long-term contraception or sterilization to prevent further children being born. So too can our reproductive services be used for other purposes when they have met our reproductive needs. The only objection to this principle would be that the drive to reproduce is overriding and that we should have as many children as possible. Some versions of some religions hold this view, but it is unacceptable to most people who seek to limit their family size.Footnote 10 If we believe that reproduction is not an overriding value, then there is a place for destructive embryo research. This is not to necessarily endorse population control but merely to show that common intuitions about population control are consistent with the view that what matters or is of value in embryos is actualizable potential.
There remains the objection that this involves creating beings with a right to worthwhile life who will be killed, in circumstances of stable population or family size, for the benefit of others. But as we and Harris have argued, there is reason to downgrade the moral status of the very early embryo. It would be permissible for thousands or millions of early embryos to die if this was necessary to produce one baby with a worthwhile life by natural reproduction. But if reproductive needs have been met, it is also permissible for some embryos to die to promote the values of health, well-being, and longevity, because they will not in any way actualize their potential to become persons.
In a world in which there is underpopulation or a threat to population survival, reproduction takes on a very high value. Contraception might be prevented and people encouraged to have as many children as possible. Masturbation, in such a world, would indeed be wrongful spilling of seed, as the Bible purports. But when the population or family size is stable or excessive, the value of continued procreation falls. In China, at times of massive overpopulation, couples were restricted to having one child. In the Western world, we give considerable freedom to couples to decide on the size of their family and most opt for one to three children. Once the reproductive needs of a family or society have been satisfied, we have argued that gametes and reproductive services can be used to create embryos for research. These embryos would not otherwise exist. No one is harmed by such practices. No one is denied a worthwhile life who would otherwise have such a life. The value of our gametes changes according to whether we need to have a child. When a family is complete, there is no moral problem with sterilization by vasectomy or tubal ligation. So too is there no problem with embryo destruction when no further children will be born.
