Three More Arguments Against Early Abortion

Abstract

The main purpose of this chapter is to examine three arguments – from Hare, Marquis and Harman – that, whether by design or not, place the permissibility of the early abortion in doubt. Each of the three arguments is grounded in positions that seem to assign at least some moral significance to the loss incurred by a merely possible person when that person is left out of existence altogether. Those positions are, in turn, at odds with the account of the moral significance of loss that Variabilism itself suggests.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

4.1 Introduction

The main purpose of this chapter is to examine three arguments – from Hare, Marquis and Harman – that, whether by design or not, place the permissibility of the early abortion in doubt. Each of the three arguments is grounded in positions that seem to assign at least some moral significance to the loss incurred by a merely possible person when that person is left out of existence altogether. Those positions are, in turn, at odds with the account of the moral significance of loss that Variabilism itself suggests.

4.2 Variabilism and the Timing of the Abortion

4.2.1 I begin with a brief sketch of Variabilism’s treatment of loss in the context of abortion. A fuller range of issues will be addressed in Chapter 5. But a brief sketch is in order here for purposes of clarifying the distinction between Variabilism and the three alternative positions I focus on in this Chapter 4.

According to Variabilism, it is critical just when, during the process of gamete production, conception, implantation, fruition, birth and early rearing, a person comes into existence. In particular, whether that process is interrupted before a person comes into existence or after a person comes into existence will often determine the permissibility of the abortion choice itself.

Timing is important, according to Variabilism, since whether the loss imposed by abortion is morally significant – whether that loss bears on the permissibility of the choice of abortion and its alternatives – depends on just where that loss is incurred in relation to the person who incurs it. And that fact is itself a function of timing. Thus, the loss incurred by a person p will have no moral significance at all if the abortion – or non-conception or non-implantation – takes place prior to p’s own coming into existence, since in that case the loss will be incurred at a world where p never exists at all. And the loss incurred by p will have full moral ­significance if the abortion – or infanticide – takes place after p comes into ­existence, since in that case the loss that p incurs is incurred at a world where p exists.

4.2.2 Variabilism itself simply determines what losses are morally significant and what losses are not. On the basis of that moral data, the otherwise plausible permissibility theory – which plausibly includes OPPP1-OPPP4 – then evaluates the abortion choice.

If the loss the person p incurs is not morally significant, then the otherwise plausible permissibility theory will not count that loss against the choice of abortion or, in a roundabout way, in favor of any alternative. But if the loss p incurs is a morally significant loss, then the otherwise plausible permissibility theory will count that loss against the abortion and in favor of its alternatives.

4.2.3 Let’s first take the case where the loss incurred as a result of the abortion is incurred by a person who never comes into existence to begin with. Here we are talking about the early abortion – the abortion that relegates a particular person to the class of the merely possible. According to Variabilism, that loss is devoid of moral significance.

With that loss out of the picture, in any ordinary case, the otherwise plausible permissibility theory can be expected to imply that the early abortion is permissible – as long as it is what the pregnant woman wants – whether it is done for a good reason, a poor reason or no reason at all.

Now, there will be exceptions to this general principle. Conflicts will sometimes arise among persons other than the person who, as a result of the early abortion, never comes into existence to begin with. And the losses such persons face in many cases are losses Variabilism will deem morally significant. Thus, it may be in the interest of the pregnant woman herself to have the abortion and in the interest of the man who impregnated her – or the woman’s grandmother, or still others – that the woman continue the pregnancy. Then and only then does the analysis become particularly complex. For then and only then do we have a situation in which the only way to avoid causing one person to incur a morally significant loss is to impose a morally significant loss on someone else. Just how the tradeoff is to be made will, in some cases, be resolved by OPPP4 – and, more generally, by a more complete version of the otherwise plausible permissibility theory.

Still, we can anticipate that the applicable tradeoff principles will surely resolve many such conflicts in favor of the pregnant woman. After all, she ordinarily has the most at stake. The morally significant loss that she will incur if the early abortion is not performed ordinarily will be considerably greater and deeper than any morally significant loss anyone else will incur if the early abortion is performed.

Moreover, in many cases, there will be a way short of requiring the woman to continue a pregnancy she doesn’t want to compensate for any deep loss the man – or the grandmother or still others – may incur. The man may choose another partner, and the woman’s grandmother may be willing to accept that a great-grandchild will come later.

Finally, at least in some cases what seems like a loss will not be a genuine loss at all, but rather a sense or expectation of loss. While the sense of loss may well make for a loss of sorts, it is also likely that the person who has that sense of loss has, or can be expected to locate, the resources necessary to mitigate it – the inner strength; the help of a friend; professional counseling. It is likely, in other words, that there will be some very good way of dealing with any such sense of loss that is consistent with the pregnant woman’s proceeding with the early abortion.

4.2.4 We now turn to the case where the loss incurred as a result of the abortion is incurred by an already-existing person – the case, that is, where the abortion involves ending the life of an existing person. Here, we are talking about the late abortion. And here the loss incurred is, according to Variabilism, a morally significant loss.

We can anticipate that the otherwise plausible permissibility theory will now analyze things quite differently. It’s true that the loss the early abortion will impose is just as great and deep as the loss that the late abortion will impose. But, as we have already seen, the loss imposed by the early abortion, according to Variabilism, is devoid of moral significance. In contrast, according to Variabilism, the loss imposed by the late abortion on the new person p has full moral significance. –So, of course, does the pregnant woman’s. Still – unless we are in a case where p’s life will unavoidably be less than worth living – it is very likely that the loss p faces will be greater and deeper than any morally significant loss that anyone else will face. And we can accordingly expect the otherwise plausible permissibility theory to resolve the conflict in favor of p.

It is thus critical, according to Variabilism, just when during the process of gamete production, conception, implantation, fruition, birth and early rearing a person p comes into existence. When, during the process of childmaking, is it that we actually have a person rather than something that is a live human organism but not a person?

4.2.5 For purposes of this book, I accept as an assumption what I earlier called the Thinking Thing Account of when existence commences in the case of persons.¹ According to that account, a pregnancy does not involve a person – the kind of thing, that is, in respect of which we have moral obligations; the kind of thing we must generally create more wellbeing for rather than less at least in the case where it exists – until the point in the pregnancy at which thinking itself has emerged.

And when is that? For purposes here, I do not try to answer that question in any complete way. It does seem clear that thinking itself requires more than electrical activity in the brain or a mechanical response to pain stimuli. Yet it also seems plausible that thinking often takes place beyond our own conscious awareness. Moreover, intermittent thinking alone is enough to keep one in existence. Once in existence, a person can continue in existence even if that person is not continuously thinking. The person doesn’t cease to exist, then, until the conclusion of that last thought – even if that person’s body, sustained by natural or artificial means, clings to life far beyond that point.

On this way of looking at things, the embryo and the early fetus are live human organisms. But, until thinking has emerged, there is no person there. Since person itself is simply defined in moral terms, the most contentious point here is that persons are the kinds of things whose coming into existence is signaled by their thinking. The most contentious point here, in other words, is that being a thinking thing and mattering morally – being, that is, the kind of thing in respect of which we have obligations; being the kind of thing we must create more wellbeing for rather than less at least if it exists – come together.

I will not try to say why that is so here. Plausibly, however, the reason that being a thinking thing and mattering morally come together is that it is in the case of thinking things that it matters to them whether we treat them in one way or another.

On this view, those who matter morally and those who don’t matter morally – persons and non-persons – are not divided in accordance with their species membership. Some individuals who are arguably human count as non-persons, in other words, and some non-humans count as persons.

4.2.6 Putting the view just described together with Variabilism, we can say that, if the biological organisms that constituted our own early bodies never had had their first thoughts, then we never would have existed at all. And any losses imposed on us at any world where we never existed at all are devoid of moral significance. They do not count against the choices that impose those losses or, in a roundabout way, in favor of the alternatives that avoid those losses.

But once we begin to think – more precisely, once we come into existence; once the human organism begins to think – things are very different. For any loss we incur at any world where we do or will exist has full moral significance. Depriving us of our entire futures once we exist thus constitutes a morally significant loss – and in many cases it will constitute a greater and deeper loss than the loss the pregnant woman or anyone else would be forced to suffer if the abortion never takes place at all.

4.2.7 Many of the points I have just made – that the embryo and early fetus are not persons; that the late fetus and human neonate ordinarily are persons; that thinking and mattering morally come together – are adopted, for purposes here, as assumptions. But they don’t beg any question. That is so, since the issues they raise are not in dispute here. Even if all of those issues happen to be resolved entirely in my favor, the issues that are in dispute here will remain quite alive. For the arguments against early abortion that I examine here are not driven by the claim that the embryo, or the early fetus, is a person. The authors of those arguments – Hare, Marquis and Harman, and also McMahan, – can easily accept the assumptions that I have made here even if they regard those assumptions in need of tweaking at the margins. What they won’t accept is just how I have claimed the thing that is unequivocally a person – that is, the possible person that the embryo or early fetus may eventually develop into, depending on how the choice of early abortion is itself made – matters morally. What they won’t accept is my claim that the merely possible person matters variably.

Thus for purposes here the open questions are these. Can the loss that a merely possible person incurs, as a result of the early abortion, make that otherwise permissible choice wrong? What are the implications, for our evaluation of the choice of early abortion, that any one of a vast number of merely possible persons could have had just as wonderful a life as I have had ? What are the implications, for the evaluation of that choice, that the future any such merely possible person faces is not really particularly distinguishable from my own? What does it mean, for the evaluation of that choice, that one or more possible persons will have existences worth having if the choice of early abortion is left unperformed?

4.3 The Golden Rule: Hare

4.3.1 It may not be surprising that some early discussions of abortion flirt with significant problems in population ethics – including, among others, the Repugnant Conclusion. Surely by 1975 it was beginning to become clear just how interesting those problems in fact were. Thus, in one short paper published in just that year, R.M. Hare worried not just about abortion but also about the “general duty to produce children” – which he somewhat hesitantly endorses on a rebuttable basis – and whether, by accepting that that duty exists, he thereby commits himself to “unlimited procreation.”² (He admits, wryly I hope, that “it is morally all right for the parents to decide, after they have had, say, 15 children, not to have any more.…”³)

But he unfortunately – in a move Parfit would question only a few short years later in Reasons and Persons, and that many today would reject on the basis of the Repugnant Conclusion⁴ – relies on the “classical utility principle” – what I have here called Totalism – to set all appropriate limits. The more the merrier, in other words, until we get to that point at which the additional person “impose[s] burdens on the other [persons] great enough in sum to outweigh the advantage gained by the additional [person]. In utilitarian terms, the … total utility principle sets a limit to population which, although higher than the average utility principle, is nevertheless a limit.”⁵

4.3.2 While Hare’s conclusions have the support of Totalism, his argument against abortion is in fact formulated in terms of the Golden Rule: “we should do to others as we wish them to do to us.”⁶ Extended a bit, the rule becomes: “we should do to others what we are glad was done to us.”⁷ If we are glad that no one “terminated the pregnancy that resulted in our birth, then we are enjoined not, ceteris paribus, to terminate any pregnancy which will result in the birth of a person having a life like ours.”⁸

A like argument – Hare is the first to admit – applies to the choice not to conceive.⁹ The fact that the one object – the pair of gametes that might combine to produce a child – is spread out in space and the other – the early fetus – is not does not, in his view, represent a material distinction between the two cases.

4.3.3 I agree with this last point. If the argument he presents against – say – early abortion is sound, so is the argument against non-conception. As he puts it, “[i]f I have a duty to open a certain door, and two keys are required to unlock it, it does not seem to me to make any difference to my duty that one key is already in the lock and the other in my trousers.”¹⁰

I do, however, object to his particular extension of the Golden Rule. From the fact that I am glad that no one interfered with my bank heist, it does not follow, not even ceteris paribus, that I am obligated not to interfere with the bank heists of others. It may be hypocritical for me to do so, but it isn’t necessarily morally wrong. Similarly, I am thrilled to pieces that my parents decided to have a child soon before they conceived me and that my mother did not then change her mind and opt for abortion. But how we infer from that fact that I am obligated not to interfere with the coming into existence of still others is a mystery.

A more plausible principle would include a moral evaluation of some sort in its antecedent. An example of such a principle would be this: if our parents had an obligation not to interfere with our coming into existence in virtue of the fact that we are glad that we exist, then we are obligated not to interfere with the coming into existence of others in virtue of the fact that they will be glad that they exist. If we take the view that they were obligated to conceive and not to abort us in virtue of the fact that that is what is good for us, then we may be compelled to say that we have an obligation to conceive and not to abort still others in virtue of the fact that that is what is good for them. The original principle, in other words, should itself be amended, so that it tries to tell us something about the views we are bound to take regarding what we are obligated to do, rather than something about what we are obligated to do. And such an amended principle would be powerful: if the antecedent held, the principle would certainly have important implications for the sorts of conclusions regarding non-conception and abortion we would be bound to reach here.

4.3.4 Now, it may seem that this amended principle can be challenged as well, on the ground that there is a morally relevant distinction between the class of persons who do exist (1975) and the class of persons who never will exist (1975). Hare himself considers that objection: the “logician,” he writes, might counter that “these potential people do not exist … and therefore cannot be the objects of duties”; “only actual people have [rights or interests].”¹¹

George Sher subsequently put the point in a more general way. Not all entities, he argued, qualify for treatment under the Golden Rule.¹² For the Golden Rule to imply that it is wrong to interfere with a new person’s coming into existence, we must first establish that that new person is the sort of entity to which the Golden Rule applies. If we just assume that point, then our argument is circular.

4.3.5 Now, if the abortion is not in fact performed, and the new person does exist, then it would be hard to argue that that the Golden Rule would not apply to that person. But suppose the abortion is performed, and the person who might have existed is left as a merely possible person. In that case, Sher’s objection seems to hit home (1977): it may seem at least plausible to think that there is an important moral distinction between persons who do or will exist (think Hare’s actual persons) and merely possible persons, persons who will never in fact exist at all.

I want to suggest, however, that this way of objecting to Hare’s argument will not in the end be satisfactory, for two reasons. First, as Hare himself notes, “it would be strange if there were an act whose very performance made it impossible for it to be wrong.”¹³ We don’t make acts right by performing them or wrong by failing to perform them.¹⁴ But there is second difficulty as well. Surely our obligations in respect of the merely possible truly are governed by something like the amended form of the Golden Rule. Thus, just as I don’t want to be enslaved and take the view others are obligated not to enslave me, I am obligated myself not to enslave others. And those others I am obligated not to enslave will include the merely possible: it would be wrong for me to bring them into existence and then, or in order to, enslave them.

This is the point that Addition Plus (Case 2.3) makes: whatever else we say about the case, we think that the choice to bring Fen into a much lesser existence, when agents had the option of bringing Fen into a much better existence, was wrong. And that’s so, even if Fen never actually exists.

We should say instead that there is some rough sense in which we are all – existing, future and merely possible – subject to the Golden Rule. We should think that others should be treated as we think we should be treated. It does not seem we are going to make much headway understanding moral law if we think one set of rules applies to us and another set of rules, a set of rules that the relieves us of any obligation we happen not to want to discharge, for example, applies to everyone else.

But it is not going to follow that the abortion itself is wrong. Even if we think the amended version of the Golden Rule itself seems plausible, and even if we then say that that amended principle applies to the merely possible just as well as it does to you and me, there remains the question of whether the antecedent of that principle is itself satisfied.

In its amended form, and applied to abortion, what we have is just this: if we think that our parents had an obligation not to interfere with our coming into existence in virtue of the fact that we are glad that we exist, then we should also think that we are obligated not to interfere with the coming into existence of others in virtue of the fact that they will be glad that they exist. The problem is that it is not at all clear that we do think our parents had an obligation not to interfere with our coming into existence for our sake, or because our coming into existence is good for us or because we are glad that we exist. I think that – intuitively – we don’t think that. We – intuitively – think instead that just as we are not obligated to produce additional persons for their own sake, our parents were under no obligation to produce us for our own sake.

4.3.6 Variabilism helps explain why it would be a mistake to think that our parents had the obligation not to interfere with our coming into existence – or, more accurately, why, if they do have that obligation, it is grounded in something other than what is good for us. Leaving you (1975) out of existence altogether would have, let’s suppose, imposed a loss on you. According to Variabilism, however, since that loss is incurred by you at a world where you never exist at all, that loss is devoid of moral significance. That loss thus cannot count against the abortion choice. If that choice is otherwise permissible – if it appropriately balances the interests of others whose wellbeing might be at stake – then the loss that you yourself would have incurred as a result of that choice cannot make that choice wrong.

Of course, we here distinguish between never bringing a person into existence and removing a person from existence. Variabilism is perfectly capable of making that distinction. Removing someone from existence is to impose a loss on that person at a world where that person exists; the loss that is imposed is, thus, a morally significant loss. While the otherwise plausible permissibility theory will not condemn all such acts, it will surely condemn many.

4.3.7 One final note. Hare “abjures” as unhelpful the discussion of whether the fetus itself is a person.¹⁵ It may be thought that a significant advantage of Hare’s approach is that we need not address that question: the abortion is forbidden, as long as we can agree that it is a person that the pregnancy in the end will generate. That is a plus, he thinks, since person itself is “not a fully determinate concept” and we can argue endlessly whether the early fetus, or the later fetus, or the embryo, the human embryo or even the human neonate, counts as one.¹⁶ I have acknowledged that point above.

On the other hand, isn’t there at least some sense of the term person that we do clearly understand, and that is useful for purposes here? I argued earlier that we think of ourselves as persons, and of ourselves as things that begin to exist when we begin to think and things that cease to exist when we cease to think. With this sense of the term firmly in mind, we can plausibly take the position that persons are also exactly the kinds of things that matter morally – at least if they exist – perhaps simply in virtue of the fact that it matters to them how we treat them. So, if we think that our task is that of giving the one, true definition of the term “person,” then we should give up. But there is no reason that we need to consider that our task.

4.4 Futures of Value: Marquis

4.4.1 In both his highly regarded 1989 paper arguing against abortion and in a more recent paper, Don Marquis aims to (1975) reinvigorate the intuition motivating Hare’s argument, (1975) avoid, as Hare does, any need to say just when during the process of conception and fruition a person comes into being and (1975) avoid, as Hare does not, a commitment to what seems an implausibly oppressive procreation obligation.¹⁷

4.4.2 Marquis begins with a general account of why death is often a misfortune and why killing is “presumptively” wrong – wrong, that is, in the absence of adequate justification. On Marquis’s view, most of us have before us a future of value, a future that, however it precisely unfolds, will include many good things. If we are killed now, we lose that future of value – we lose something that is of value to us. For that reason, to impose that loss on us – to kill us – is, in the absence of justification, a serious wrong.¹⁸

Marquis’s view of death is consistent with the idea that whether a particular death is bad at all, and, if it is, how bad, can be determined by comparing how that person fares at a world where he or she continues to live against how that person fares at a world where he or she does not continue to live.¹⁹ Of course, there are many different possible futures that may be open to us at any given time, depending, among other things, on the choices that we and others happen to make. Some of those possible futures will be wonderful for us, and some will be at least nice and a few may well be less than worth having. The comparison between each of these possible futures in which we do continue to exist against a possible future in which we do not continue to exist will yield an array of particular values – measures of just how much better (1994) for us it is that we continue to exist than that we not. As long as the value that is correlated with a particular future that includes us falls in the positive range (as long as the life is not itself wrongful), the death is, relative to that future, a bad thing for us. Since it is very likely that, whatever future in fact unfolds for me, that future will be wonderful or at least nice for me, it is very likely, as well, that death now would be a bad thing for me. And it would be bad for me in a sense that is relevant to whether the act that causes it is permissible. Death isn’t just a loss. It’s a morally significant loss.

4.4.3 This account of why death is often bad seems reasonably uncontroversial – and is consistent with Marquis’s own. And it seems roughly to explain, as he puts it, why killing is presumptively wrong.

What happens next is less plausible. For Marquis then extends his account of the badness of death for the purpose of grounding an argument against early abortion.

Suppose that we do find this account of the badness of death reasonably uncontroversial. I think we are then surely making a certain critical background assumption regarding just when it is to be applied. We are assuming that the account applies – generates the result, that is, that the death is bad – just in the case where there exists a person whose future of value is eliminated by that death. We are assuming, that is, that the account generates the result that death is bad only when (1) the thing whose future of value we strip away is a person and (2) that thing exists.

In order to extend the account of the badness of death to the case of early abortion, Marquis must revise (1). But he can at the same time – and this is why his account diverges from Hare’s – leave (2) as it stands.

The conditions he thus adopts on when the stripping away of a thing’s future of value – when the killing of that thing – is bad and presumptively wrong are just these: (1′) that thing is an individual – it need not be, or in any event need not yet be, a person – and (2) that thing exists.

Now, not all individuals have a future of value that can be stripped away from them. What sorts of individuals can expect – provided that they continue in existence at all – to have futures like ours, that is, futures of value? Daffodils and ticks we can perhaps rule out; cats and dogs we can leave unjudged. The individuals that most clearly will have futures of value are, of course, human embryos and early human fetuses. They look forward, not to the future that a daffodil or a tick or even a cat will have, but to the future that individuals who are uncontroversially persons look forward to. They look forward to futures like ours, futures of value.

Marquis thus concludes that it is bad and presumptively wrong to end the life of a human embryo or early human fetus. The early abortion is, like any ordinary killing of any ordinary person, bad and presumptively wrong.

4.4.4 Marquis thus nowhere claims that the human embryo or early human fetus is a person. He can easily live with the assumptions that I make starting out – in particular, that the embryo and early fetus are not persons in virtue of the fact that they have not had that first thought. Marquis’s argument instead proceeds from the more modest idea that the embryo and human fetus are existing individuals that have futures of value. And that’s all – he thinks – that he needs.

4.4.5 It is Marquis’s adoption of condition (2) – that the individual whose future value would be eliminated by abortion in fact exists; what I will call the existence condition – that allows Marquis to circumvent the stringent procreative obligation that Hare himself seems committed to. It is that condition that blocks the extension of Marquis’s own argument to the case of non-conception. The idea – and we will look more closely at this idea below – is that until the individual who has the future of value itself exists, there is nothing there to take anything away from.

While Hare seems to insist that we conceive a child and not abort that child, what Marquis says is perfectly consistent with the idea that non-conception is permissible. It’s just that, on Marquis’s view, when we do conceive – when we do bring the embryo or early fetus into existence – we must not then abort it.

4.4.6 The beauty of Marquis’s argument against early abortion is its facial neatness. But beneath the neatness things are a little messy. Conditions (1′) and (2) both require a very close look.

4.4.7 Condition (1′) asserts that a thing must be an individual in order for the elimination of its future of value to count as bad and presumptively wrong. Marquis uses the concept of the individual to argue that his account has no application to the human embryo in its very earliest stages. I take it that that concept would similarly rule out the application of his account to the scattered object consisting of just egg and sperm. That thing, after all, has every bit as much of a future of value as does the early embryo. But let’s start here with the early embryo.

Until approximately the 14th day after conception, when the primitive streak emerges, the embryo itself is capable of dividing into two genetically identical (1′) successor embryos.²⁰ It is capable of twinning. Marquis considers – but seems to reject – the view that the original embryo ceases to exist at the moment the twinning takes place.²¹ The following argument might lead one to think otherwise. If one of the successor embryos is identical to the original, then, since the other embryo occupies the same position relative to the original, the other embryo must be identical to the original as well. But by the transitivity and symmetry of identity, we cannot consistently say that any two things are identical to any one. The argument concludes that the original embryo must not have survived as either of the two successor embryos; instead, it has ceased to exist at all, and been replaced by the two embryos.

The glitch in this approach that leads Marquis to reject it is the case where that identical original embryo does not in fact have a fission in its future.²² In that case, it becomes implausible to say that that embryo ceases to exist. But it also seems that what we say about the embryo in the one case we should say about the embryo in the other: if it survives in the case where twinning doesn’t take place, then it survives as well in the case where twinning does take place.

An alternate picture of things may accordingly seem more tenable: that, whether the embryo twins or not, the two, or four, or eight, genetically identical cells that constitute that original embryo are themselves individuals, each of whom has its own future of value. On this view, twinning does not involve the elimination of a single individual, or the survival of one individual as two or more individuals. Rather, twinning involves the survival of two or more individuals, from one moment to the next, as two or more individuals.²³

What keeps Marquis from embracing this way of looking at things is his sense that, in the case where the twinning does not in fact take place, we are compelled to say that only one individual human being can emerge from the process.²⁴

Facing these questions, Marquis seems to conclude that we should delay saying that we are dealing with an individual – and hence delay application of the future of value account – until the possibility of twinning is itself passed – until, that is, the 14th day after conception.²⁵ “Whatever assumption we make concerning the first phase of a human being, the future of value analysis of the wrongness of killing does not support an objection to human embryonic stem cell research.”²⁶

4.4.8 The difficulty with Marquis’s position I want to focus on here is not that he wants to say that the embryo is not yet a single individual until the multiplicity of potential individuals have knitted themselves together into one unravelable thing (2008). The difficulty isn’t that he wants to say some things are individuals and others aren’t. The difficulty isn’t that he doesn’t want to count the 5 day embryo as an individual, or alternatively as many individuals, but does want to count the 14 day embryo as an individual. At least, that’s not an obvious difficulty (though defining individual is never going to be a simple task).

The difficulty I want to focus on here is the moral weight that Marquis sotto voce attaches to the term individual. He wants to say that the 14 day embryo is a single individual having a future of value that we are enjoined from eliminating. He wants to say that the 5 day embryo is not an individual and not a multiplicity of individuals, and hence that we are not enjoined from eliminating whatever futures of value that would otherwise be theirs.

In effect, he is saying that some things, that is, individuals, have futures of values we are enjoined from eliminating and other things, that is, non-individuals, do not. He is committing himself – without explicitly saying so – to the view, in other words, that some things, that is, individuals, matter morally and others, that is, non-individuals, do not.

4.4.9 The same tactic, it seems, would be deployed in the case of the barely scattered object consisting of a single human egg cell and single sperm cell residing in a single glass dish. Fertilization, let’s suppose, is imminent. To avoid the view that it would be wrong to interfere with conception in this case, Marquis could take the position that that barely scattered object is not itself an individual. His account then would be consistent with the view that there is nothing wrong with our interfering with conception – and thereby depriving that scattered object of its future of value – by way of, saying, grabbing the sperm by its tail and insisting on its having its own dish. We’ve stripped a thing, but not an individual, of its future of value. So what we have done is not necessarily wrong.

Or recall Hare’s example of the two keys and the lock he has a “duty” to open. Marquis would presumably deny that any arbitrary pair of egg and sperm, from any arbitrary pair of potential gamete providers, constitutes an individual. Hence, though any such pair will have a future of value in any scenario in which the relevant gamete providers are poised to choose the relevant manner of conception over non-conception at the relevant time, depriving that pair of its future of value, by not proceeding to conceive a child, is not wrong.

4.4.10 To address the scattered object cases, Marquis might insist that a thing is not an individual unless its parts are spatiotemporally continuous. A future of value may well be in the picture, relative to the scattered object consisting of the single egg and single sperm, whether residing in the same glass dish or in human bodies on opposite sides of the planet. But if the scattered object is not an individual, then it would not be wrong, on Marquis’s account, to eliminate that future of value.

Perhaps that way of looking at things is correct. But even if it is, my objection is left unaddressed. Marquis is using the distinction between individuals and non-individuals to distinguish between cases where it is wrong to eliminate a future of value and cases where it is permissible.

More generally, let’s suppose that Marquis can clearly articulate a line between individuals and non-individuals. We shall still need an explanation for why things that belong to the former class matter morally in a way that things that belong to the latter class do not – why, that is, it is wrong to deprive individuals of their futures of value but not wrong to deprive still other sorts of things – non-individuals – of their own futures of value.

4.4.11 We now turn to Marquis’s second condition – condition (2). According to that condition, the thing whose future of value is eliminated – that is, the individual whose future of value is eliminated – must exist, in order for the elimination of that future of value to be considered bad or presumptively wrong. Accordingly, depriving the merely possible individual of its future of value does not, according to Marquis, even begin to trigger the result that that deprivation – that “killing” by way of relegating an individual to the class of the merely possible – is wrong.

Marquis thus adopts what I am calling the existence condition. But he doesn’t fully explain it. It is way to avoid the embarrassment of Hare’s overly stringent procreative obligation. But it is a way that seems at odds with certain aspects of Marquis’s own account.

Thus, just as we can legitimately ask why Marquis thinks individuals matter morally but non-individuals do not, we can ask why Marquis thinks existing things matter morally but merely possible things do not. After all, the merely possible person has just as much of a future of value, immediately prior to conception, as the 14 day human embryo does. Yet, on Marquis’s view, the fact that embryo has a future of value means that the embryo is the kind of thing in respect of which we have moral obligations: when we take its future of value away from it – when we impose that loss on it – we do something that is presumptively wrong. Why, then, don’t we do something just as wrong when we interfere with conception and take the future of value away from that merely possible person? Why doesn’t that loss, as well, matter morally? Why does the speck that consists of the human embryo in some way inherit the value of its possible future, while the merely possible person, according to Marquis, does not?

4.4.12 Marquis might answer that the importance of existence is simply automatic. The difficulty with that answer is that it seems at odds with the explanation he would give of why the loss of a future of value incurred by the 14 day embryo has moral significance, but the loss of a future of sorts incurred by the daffodil or the tick does not. It seems that it is the fact that the loss is the loss of a future of value, and not just a loss of a future of sorts, that is doing the work of conferring the moral significance. It is the nature of what stands to be lost that seems to be doing the work.

But if that is so, then the loss the merely possible person incurs when that person is left out of existence altogether should itself have moral significance. After all: that merely possible person has just as much of a future of value as the 14 day embryo.

Of course, to block the implications of that way of looking at things for the case of the merely possible person, Marquis can – simply insist on the existence condition. But in doing so Marquis raises the question of just what we are to say about the future of value per se. If it is existence that has the power to imbue some losses but not others with moral significance, then it seems that it is not, after all, the future of value that has that power.

4.4.13 Another option – and in fact the option that Variabilism suggests – would be to say that losses of futures of value are just like any other losses. What makes them have moral significance or not is where those losses are incurred in relation to the individual – or I would say the person – who incurs them.

Variabilism thus adopts something like the existence condition, but in a carefully tailored form. According to Variabilism, whether the loss has moral significance or not depends on just where that loss is incurred in relation to the person who incurs it. Variabilism is itself controversial. But it avoids a bright line between existing, or actual, persons, and the merely possible. It instead adopts the more plausible view that they all matter morally, but on a variable basis. We can then agree that eliminating a future of value for a person who never exists at all – by never bringing that person into existence to begin with at a given world – cannot itself constitute a morally significant loss. But we also want to recognize very clearly that that same person can incur morally significant losses in abundance at still other worlds – worlds, that is, where that person does or will exist.

4.4.14 One of the reasons Marquis’s account seems so neat on first inspection is that it seems not to require us to say just when during the pregnancy a person comes into being. The concepts he wants to work from seem far more neutral. All and only those things that are both individuals and existing are such that depriving them of their futures of values is bad and presumptively wrong.

The problem is we immediately want to know why it is bad and presumptively wrong to eliminate futures of value from all and only those things that are individuals and exist. In asking that question, we immediately open the whole can of worms. To call a 14 day embryo, but not a 5 day embryo and not an egg cell a nanosecond away from fertilization, an individual is not going to justify the result that it is wrong to deprive the 14 day embryo, but not the 5 day embryo or the imminent conceptus, of its future of value. We need to know more. We need to know why the loss incurred by the one but not the other has moral significance – why it is bad and presumptively wrong to impose the one loss, and to eliminate the one future of value, but not the other.

4.4.15 Marquis’s approach is refreshing in part because it tries to find a way to avoid the arduous work of determining when it is during the pregnancy a person comes into being. In contrast, Variabilism accepts the need for this work. For it determines whether a particular loss has moral significance by reference to the class of persons: existing persons, future persons and the merely possible. For purposes here, I have assumed that persons are those things moral law requires us to create more wellbeing for rather than less at least in the case where they exist at all. We ourselves belong to that class. Moreover, the fact that we can give a plausible account of just when we come into existence and when we will go out of existence suggests that we can give a plausible account of when persons in general come into existence and when they go out of existence. I have suggested that these beginning and end points are marked by the individual’s first and last thoughts.

These points are, of course, controversial. But the idea that losses bear on the moral permissibility of the acts that impose those losses and their alternatives only if those losses are incurred by persons at least potentially explains – in a way that calling one an individual and the other a non-individual never can – why it is wrong to deprive the 3-year-old child of its future of value but not wrong to deprive the egg cell whose fertilization is imminent of its future of value.

4.5 The Actual Future Principle: Harman

4.5.1 Hare’s approach coincides with the “classical utility principle” – that is, Totalism. According to Totalism, agents are obligated to make the choice that maximizes wellbeing on an aggregate basis. If leaving a merely possible person out of existence means that less aggregate wellbeing is created rather than more, then it is wrong, according to Totalism, to leave that person out of existence. In this way, Totalism elevates what happens to merely possible persons when we leave them out of existences worth having to what happens to existing and future persons when we create less wellbeing for them when we could have created more: losses of both sorts are considered to have full moral significance.

Marquis adopts an existence condition. Thus on his account the loss incurred by the possible person when we leave that person out of existence altogether has full moral significance only if the individual human organism – the early human fetus, or even the embryo once it no longer has a potential fission in its future – that will develop into that person in point of fact exists. Thus, Marquis’s account, unlike Hare’s, avoids the implication that the failure to conceive a child to begin with, or the in vitro conception of a child followed by the failure to arrange for the 8 or 16-cell embryo to develop in utero, is inherently wrong.

Elizabeth Harman also adopts an existence condition for purposes of addressing certain questions regarding the moral status of the human embryo and early human fetus. The existence condition she adopts is more stringent than Marquis’s in one respect and less stringent in another. According to Harman, a person or other being that has “intrinsic properties that themselves confer moral status” must come into existence if the abortion is to be declared wrong.²⁷ However, that person need not already be in existence in order for the early fetus to have moral status. It is enough that that person will come into existence at some future time. Thus, it is not just the properties a thing now has that are important in determining whether that thing “has moral status.”²⁸ Rather, a “thing’s present nature” – and ultimately its moral status – is “solely determined by the intrinsic properties it ever has.”²⁹ This is what Harman calls the Actual Future Principle.

In terms of worlds, we could say that the embryo or early fetus has moral status at a world w, according to Harman, if the person the early fetus develops into at some world or another w′ in fact eventually comes into existence at w. Thus: “each fetus has its status in virtue of facts about that fetus’s actual life: these facts might have been different.”³⁰

4.5.2 The Actual Future Principle seems to imply, among other things, that “early fetuses have their moral statuses contingently.”³¹ The early fetus that “will become a person has some moral status” while the “early fetus that will die while it is still an early fetus has no moral status.”³² The fetus, we can say, at one world is identical to the fetus at another. But that will not mean that the fetus has the same intrinsic properties at both worlds. According to Harman, then, “there are two significantly different kinds of early fetuses.”³³

On this view, the present nature of a particular early fetus in part depends on what happens next. At a world where there is no early abortion, we will be dealing with an early fetus that has one sort of nature and some moral status. But at a world where the early abortion does take place, we will be dealing with an early fetus that has another sort of nature and no moral status at all.

4.5.2 The Actual Future Principle enables Harman to address two issues – early abortion and how an early fetus can be an appropriate object of parental love.

According to Harman, the early abortion is deemed permissible – or at least a choice that requires “no moral justification” – if the person or conscious subject the early fetus might develop into never in fact does come into existence – if, that is, that early fetus never in fact “become[s] a person.”³⁴

The upshot, according to Harman, is a “very liberal” position on early abortion: it requires “no moral justification whatsoever.”³⁵ The “common liberal view,” in contrast, often finds early abortion permissible but also holds that the early abortion “requires at least some justification, however minimal.”³⁶

The view that Harman describes exactly tracks the results on early abortion that I have proposed here. The early abortion (2000) any ordinary case can be determined to be perfectly permissible and under a very summary analysis.³⁷

4.5.3 Harman also wants to explain how the very fetus that has no moral status at a world where the early abortion takes place can also be an “appropriate object of love” at still another world, a world where the pregnant woman and her partner have no thought of or desire for an abortion.³⁸ Apparently accepting the idea that being an appropriate object of love and having some moral status come together, Harman uses the Actual Future Principle to unravel the puzzle. If there is no early abortion, and the fetus survives, it will develop into a person. And the “present nature” of that early fetus is determined by all the intrinsic properties it “ever has.” Thus, the fetus that is not aborted – the fetus that later develops into a person – is a perfectly appropriate object of parental love now.³⁹

4.5.4 Still, there are issues. Suppose that the woman in fact continues the pregnancy. One issue is how we are to evaluate her choice to do just that in the case where having the early abortion would not have made things worse for her or for anyone else. Harman wants to say that the early abortion itself is perfectly permissible. It “needs no justification.” But if there is no abortion, then the early fetus itself has at least some “moral status.” Moreover, Harman elsewhere argues, it is plausible to think that bringing a person into an existence worth having confers a “benefit” on that person, and that we have a “moral reason” to confer that benefit by way of bringing that person into existence.⁴⁰

The case, then, looks like this:

Harman’s views about the moral status of the early fetus, the benefit of coming into existence and the role that moral reasons would seem naturally to play in our analysis seem clearly to support the result that a1 – continuing the pregnancy – is obligatory. But if a1 is obligatory, then we cannot say as well that a2 – the early abortion – is itself permissible. At least, according to standard deontic axioms, we can’t have things both ways: we can’t say a2 is permissible but that we are obligated to choose a1 instead.

4.5.5 On the face of things, Harman’s view seems very like Exclusion Beta. In fact, Harman herself suggests such a connection, when she elsewhere describes the “Bad Method” as identifying – with Exclusion Alpha – the persons who actually exist and then concerning ourselves just with how our performed (2009) acts and unperformed (2009) acts affect those persons and those persons alone.⁴¹ The Bad Method does seem bad. But what she calls the “Good Method” is not so great, either.⁴² Applying the Good Method, we ask whether an act would harm any things. If it would, we then ask whether those harmed things have moral status.⁴³ And elsewhere: “one must be careful to evaluate actual events with respect to the actual moral statuses of the early fetuses involved; and one must evaluate counterfactual events with respect to the counterfactual moral statuses of the fetuses involved.”⁴⁴

On the face of things, the Good Method seems deficient in the same way that Exclusion Beta is deficient. The truth is that the evaluation of an act performed at one world is going to be strongly connected with evaluations of alternative acts performed at alternative worlds. It is that connection that deontic axioms often try to describe. Allowing the things that matter morally to shift from world to world, as Exclusion Beta explicitly does and as Harman’s Actual Future Principle also seems to do, is going to generate an array of evaluations that just do not obey those axioms.⁴⁵

4.5.6 In fact, however, it seems that Harman’s view is more carefully drawn – and more complex – than this. On her view, that an agent has a moral reason to perform one act, no moral reason not to perform that act and no moral reason to perform any alternative act does not necessarily imply that the agent has an obligation to perform that one act. The fact that an act will create more wellbeing for some individuals who have moral status (e.g., the child at w1) and create less wellbeing for no one at all (e.g., the mother at w1) will not, on Harman’s view, trigger the obligation to perform a1.

Harman’s goal in introducing the notion of the contingent moral status of the early fetus is to show that a certain necessary condition for the early abortion’s being wrong is failed. That is, for the early abortion to be wrong, the early fetus would need to have some moral status, and it doesn’t. But none of this implies a sufficient condition on wrongdoing. Consistent with what Harman says, a sufficient condition on wrongdoing will be highly complex. It would go beyond the idea that, if we have a moral reason to do one thing, no moral reason not to do that thing and no moral reason to do anything else, then we are obligated to do that one thing. And it would go beyond the idea that if an act will create more wellbeing for some individuals who have moral status and create less wellbeing for no one at all, then that act is obligatory.

4.5.7 There are aspects of Harman’s view, to be sure, that suggest that a1 is obligatory in the Actual Future Case. Harman thus writes that the choice not to abort means that the woman “owes [the early fetus] her love.”⁴⁶ However, the relation between our having a moral reason to do something and our being obligated to do that thing is a complex one. The other side of the early abortion issue makes this very point.⁴⁷ Thus, the pregnant woman may have a good reason – a reason having to do with her own future wellbeing, a reason we consider a moral reason – to have an early abortion. Moreover, the early abortion may be perfectly permissible for her. But it is odd to think that she is obligated to have the early abortion. Surely, if she wants to have the baby, it is permissible for her to do so.⁴⁸

Case 4.1

The Actual Future Case

4.5.8 There is good reason to think, then, that we should not attribute Exclusion Beta to Harman. At the same time, the more complex view that Harman seems to adopt is going make her position on early abortion itself more complex that it seemed starting out.

It seems that, in Harman’s view, moral obligation in any strict sense is going to be hard to come by. If a1 is not obligatory in the Actual Future Case, notwithstanding all there is to be said in favor of a1 – the moral reasons that favor a1, the moral status of the child who is benefited by a1 – then the test that Harman sets for obligation is a stringent one. Little is obligatory. But if that is so, then much is permissible, including a2.

But that means that a2 is permissible even though the agent has a moral reason not to perform a2. At least, it seems that the bare fact that the agent has a moral reason to perform a1 means that the agent has a moral reason not to perform a2. And we have already said that the agent does have a moral reason to perform a1.

But that means that when we counsel the young woman that the early abortion is perfectly permissible – that it requires “no moral justification” – the young woman must now turn to the task of reconciling the permissibility result that allows the early abortion and the moral reason against the early abortion. The complexity of the assessment is itself going to make her, and us, wonder about the very meanings of the terms that are being used. Can something really need “no moral justification” when we have a moral reason not to do that thing?

4.5.9 The opposing theses in Harman’s work aren’t inconsistencies. But they do suggest an argument of sorts against early abortion.

Consider, again, the Actual Future Case. Harman wants her view to suggest that a1 and a2 are both permissible. So far, so good. But her view also suggests that the agent has a moral reason to perform a1 and hence a moral reason not to perform a2, and no moral reason not to perform a1. If this is moral law – and Harman doesn’t claim that it is, but if it is – then it is not clear what it is telling us. Something like: both alternatives are permissible, but the one is really better than the other. Or: neither choice is really terrible, but a2 is a little bad. Or: the agent who aims to act as moral reason instructs will choose a1 and leave a2 alone.

4.5.10 Is there a way to make things easier? Or must we instead just live with the idea that moral law is complex – and that the morally cautious pregnant woman would do well to avoid the early abortion in any ordinary case?

Variabilism suggests one strategy. The intuition behind Harman’s claim that early abortion “requires no justification” seems to be that at any world where the early abortion is in fact performed that choice imposes no harm – no loss – on any existing or future individual that has any “moral status” at all. This intuition is – in part – shared by Variabilism. The zero wellbeing level implied by never having been brought into existence to begin with – that loss – just isn’t morally on the same plane as the loss that is imposed on the person who exists and is treated so badly that his or her overall wellbeing level is reduced to zero. The Basic Case makes that point. If a person incurs any loss at all in virtue of never having been brought into existence, then that loss, according to Variabilism, has no moral significance whatsoever.

So far, the views coincide. But now let’s turn to the benefit side of things. For Harman, the agent – at, for example, w1 in the Actual Future Case – has a moral reason to confer the benefit of an existence worth having on the child. This is to say that that particular benefit – that gain for the child – itself has moral significance.

Now, it might seem difficult to say that that gain does not have moral significance, since the child in fact exists at w2. Variabilism, however, in fact has the resources to say just that. For according to Variabilism the child matters morally, but matters variably. That means that just as some of that child’s losses have moral significance and some do not, some of that child’s gains have moral significance and some do not. After all, losses and gains – intuitively, and as a metaphysical matter – mirror each other. If one act creates a gain relative to an alternative, the alternative creates a loss relative to the one.

Harman’s view seems to be that a gain for an individual that itself has moral status can create a moral reason to perform an act, but that the corresponding loss does not create a moral reason not to perform the alternative act. In contrast, according to Variabilism, we can take the position that, just as the loss the child incurs in w2 is devoid of moral significance, so is the gain the child enjoys in w1 devoid of moral significance.

Variabilism thus denies that the gain – the benefit – that is created for the child at w1 has any moral significance at all. We can, accordingly, deny that there is any moral reason for the woman to choose a1 – or not to choose a2. We thus avoid the need to explain why a2 is permissible even though the pergnant woman has a moral reason not to choose a2. We simply deny that she has any moral reason not to choose a2.

4.6 Distinction Between Variabilism and Its Competitors

4.6.1 Each of the views described above sets aside the issue of just when the new person comes into existence – just when, that is, we are dealing with a person rather than simply a live human organism.

Hare relies on the Golden Rule – but notes that the results he reaches regarding abortion conform to the “classical utility principle,” that is, Totalism. Under both those approaches, the merely possible person matters morally. And that person matters morally, not just variably, but rather necessarily: the loss that person incurs, when we fail to bring that person into existence, has every bit as much moral significance as the loss that same person incurs at a world where he or she exists.

Marquis, too, considers the merely possible person to matter morally. On Marquis’s view, however, the merely possible person matters morally, but matters contingently: the loss of the future of value – the future that the merely possible person would otherwise have enjoyed – has moral significance only if the individual whose future of value is at stake in fact exists.

Harman, too, relies on the contingency of existence – but she focuses on the contingency of the future person’s existence, not the contingency of the individual’s existence. Accordingly, Harman, in contrast to Marquis, makes the merely possible person – the person who in fact does not and will never exist – matter morally not at all. For her, it is the fact that a person will exist that makes that person matter morally – that the person is a future person, and not a merely possible person – that imbues any losses that person incurs with moral significance.

On none of these views, then, is it critical when the person comes into existence. Each view focuses, instead, on whether that person could have existed or will exist. Intuitively, though, timing seems central to the question of abortion (1993): just when is it that what we are choosing to destroy is not simply a live human organism but rather a thing that moral law itself requires us to create more wellbeing for rather than less.

Thus, it is at least arguable that we can create more wellbeing rather than less for the early human fetus. After all, in such a case we do have a live human body, something that may continue to develop and flourish depending in large part on how we now conduct ourselves.

But we should resist the idea that we are obligated to create additional wellbeing for anything we can create additional wellbeing for. The fact that we can, in a way, create additional wellbeing for the early fetus does not imply that we ought.

Hare, Marquis and Harman do not deny, of course, that at some moment in time – or perhaps over some period of time – a person emerges during the pregnancy or at some later stage, during infancy or childhood. What they deny is the moral significance of the timing of that event.

4.6.2 Variabilism, in contrast, spotlights the timing of the person’s coming into existence in relation to the abortion’s being performed. It isn’t the existence of the person per se that is important, of course: such a view would seem tantamount to Exclusion in either its Alpha or Beta form. According to Variabilism, what is important, instead, is whether the loss that the person incurs is incurred at a world where that person does or will exist. If the abortion takes place at a world prior to the moment at which the person comes into existence – if it keeps that person from ever existing at all – then any loss that person incurs at that world is devoid of moral significance since that person will never exist at that world. In contrast, if the abortion takes place at a world after the moment at which the person has come into existence, then any loss incurred by that person will have full moral significance since that person does or will exist at that world. Any such loss will then count against the act that imposes it.

A respect in which this approach seems intuitively plausible is that it focuses our attention on what is going on with the fetus in utero. Is it yet capable of feeling pain? Or is its neurology simply developed enough to show a mechanical reaction to a pain-stimulus? Is it thinking, rudimentarily or even sub-consciously? Is there a psychological, or cognitive, being there, something that can be pulled one way or another by the activities and responses of its own mind, that can feel aversion to certain events and pleasure in others?

We think – I think we think – that it’s wrong to torture a cat for no reason whatsoever and that it’s wrong to torture a human baby for no reason whatsoever. That means that those beings have moral status. In deciding whether to torture or not, we must take their interests into account alongside our own. The question that we intuitively want to focus on in thinking about abortion, and the question Variabilism brings to the fore, is just when it is that the developing human ­organism acquires that same moral status – just when it is that a person comes into existence. Variabilism thus considers morally relevant what we intuitively think is morally relevant.