The Morality of Creating Lives Not Worth Living: On Boonin's Solution to the Non-Identity Problem

Abstract

David Boonin argues that in a choice between creating a person whose life would be well worth living and creating a different person whose life would be significantly worse, but still worth living, each option is morally permissible. I show that Boonin's argument for this view problematically implies that in a choice between creating a person whose life would be well worth living and creating another person whose life would not be worth living, each option is also morally permissible.

1. Introduction

The non-identity problem concerns the question of whether we morally ought to create a person with an excellent life, rather than another person whose life would be significantly worse but still worth living. While most people believe that creating the worse-off person would be wrong, justifying this verdict has been difficult, since creating the worse-off person arguably does not harm anyone. In The Non-Identity Problem and the Ethics of Future People (2014), David Boonin proposes a radical solution to this problem. On the ground that creating the worse-off person is indeed harmless, Boonin argues that – contrary to what it seems – it is also morally permissible.

What Boonin does not note, however, is that his argument for this view has a consequence that is even more counter-intuitive, and, it seems, clearly unacceptable: That it is also permissible to create a person whose life would not be worth living, rather than to create a person whose life would be well worth living. After formulating the non-identity problem and Boonin's solution to it (sect. 2), I explain why this problem arises for his argument (sect. 3) and consider some possible responses to the problem (sect. 4). I then offer a diagnosis of why the problem arises, on which the underlying reason is that Boonin's argument implies a ‘person-affecting’ moral view (sect. 5). Section 6 concludes.

2. The non-identity problem

Consider this case:

Worth Living: Wilma has decided to have a baby. Her doctor tells her that if she conceives now, she will get a daughter, ‘Pebbles’, with an irreversible disability. While Pebbles's disability will have a substantially negative impact on her well-being, her life will nonetheless be worth living (i.e. it will contain more positive well-being than negative well-being). If Wilma instead takes a pill once a day for two months before conceiving, she will get a different child: a son, ‘Rocks’, who will be perfectly healthy and whose life will therefore be significantly better than Pebbles's life would be. The pill is easy to take, has no side effects, and will be paid for by Wilma's health insurance. However, Wilma finds it too inconvenient to take the pill and so chooses to conceive at once. As a result, she gives birth to Pebbles, who has a significant and irreversible disability.¹

Intuitively, Wilma's decision to conceive now rather than later is wrong. However, if Wilma had conceived later, she would have given birth to Rocks instead of Pebbles, thereby bringing a different person into existence. Hence, since Pebbles would not have existed if Wilma had not decided to conceive now, Pebbles would not have been better off if Wilma had not decided to conceive now. This suggests that Wilma's decision does not harm Pebbles. If we also assume, as Boonin does, that Wilma's decision does not harm anyone else, this suggests that her decision is not wrong. And this contradicts our initial assessment of the case – intuitively, again, Wilma's decision to conceive now rather than later is wrong. This is the non-identity problem.

Boonin construes the problem as an argument that consists of ‘five plausible premises and one implausible conclusion’ (2014: 1):

The non-identity argument

P1: Wilma's act of conceiving now rather than later does not make Pebbles worse off than she would otherwise have been.

P2: If A's act harms B, then A's act makes B worse off than B would otherwise have been.

P3: Wilma's act of conceiving now rather than later does not harm anyone other than Pebbles.

P4: If an act does not harm anyone, then the act does not wrong anyone.

P5: If an act does not wrong anyone, then the act is not morally wrong.

The Implausible Conclusion: Wilma's act of conceiving Pebbles is not morally wrong. (Boonin 2014: 27)

Most solutions to this problem have consisted in rejecting one of the premises. For instance, according to harm-based solutions, Wilma's decision harms Pebbles even though it does not make Pebbles worse off than she otherwise would have been (Harman 2004; Gardner 2015). This conflicts with P2. Similarly, according to act-consequentialism, Wilma's decision is wrong because one of her alternatives – creating Rocks – would have led to more total well-being. Given that an action can be suboptimal in this way without wronging anyone, act-consequentialism conflicts with P5. Boonin argues, however, that all solutions that consist in rejecting one of the premises are bound to fail.² Thus, he proposes that we instead accept the argument's conclusion, despite its apparent implausibility: Wilma's decision, Boonin holds, is not wrong. As he puts it, ‘[if] we accept the Implausible Conclusion […] then the fact that the five premises of the non-identity argument entail it will no longer give rise to a problem’ (Boonin 2014: 189).

3. More implausible conclusions

Boonin notes that his solution apparently implies other, seemingly even more implausible conclusions. The most implausible conclusion that he discusses – and which he refers to as ‘The Most Implausible Conclusion’ (Boonin 2014: 226) – concerns cases such as the following:

Barely Worth Living: Wilma's doctor tells her that if she conceives now, she will give birth to a daughter – ‘Sandy’ – with a serious disease. As a result of her disease, Sandy's life will only be barely worth living (i.e. her life will contain only slightly more positive well-being than negative well-being). If Wilma instead takes a pill for two months before conceiving, she will have a perfectly healthy son – ‘Rocks’ – whose live will be well worth living. Wilma decides that having to take a pill before conceiving is too inconvenient and so chooses to conceive at once. As a result, she gives birth to Sandy, whose life is only barely worth living. (cf. Boonin 2014: 226)

The Most Implausible Conclusion is that Wilma's decision in Barely Worth Living is permissible. The premises of the non-identity argument imply this conclusion, for the same reasons that they imply that Wilma's decision in Worth Living is permissible. Boonin is apparently ready to accept this conclusion as well.³

What Boonin does not note, however, is that the premises of the non-identity argument imply a conclusion that is even more implausible than the Most Implausible Conclusion – one that I am accordingly forced to call ‘The More-Implausible-than-the-Most-Implausible-Conclusion Conclusion’, or ‘The More-than-Most Implausible Conclusion’ for short. Consider another variation on the case of Wilma:

Not Worth Living: Wilma's doctor tells her that if she conceives now, she will give birth to a daughter – ‘Dora’ – who will suffer greatly throughout her entire life due to a serious incurable disease. As a result, Dora's life will not even be close to worth living (i.e. it will contain much more negative well-being than positive well-being). If Wilma instead takes a pill for two months, she will have a perfectly healthy son – ‘Rocks’ – whose life will be well worth living. Wilma decides that having to take a pill before conceiving is too inconvenient and so chooses to conceive at once. As a result, she gives birth to Dora, whose life is not even close to worth living.

The More-than-Most Implausible Conclusion is that Wilma's decision in Not Worth Living is not wrong. Even if it is disputable whether her decision in Worth Living, and perhaps even in Barely Worth Living, is wrong, the conclusion that her decision in Not Worth Living is not wrong is intuitively unacceptable. But the premises of the non-identity argument imply this conclusion. If Dora had not existed, she would not have occupied any well-being level at all – not even a well-being level of zero. This follows from the popular and plausible principle that in this context is usually called ‘actualism’, according to which a subject A (such as Dora) instantiates a property F (such as having a well-being level of zero) in a possible world w only if A exists in w.⁴ Since a low well-being level is not lower than no well-being level at all, actualism implies that Dora would not have been better off if Wilma had not decided to conceive her. P2 therefore implies that Wilma's decision does not harm Dora, which, together with P4, implies that her decision does not wrong Dora. If we also assume that Wilma's decision does not harm anyone else (an assumption corresponding to P3), P4 implies that her decision does not wrong anyone, which together with P5 implies that it is not wrong. And this result is unacceptable – it is wrong to create a person with a life not worth living rather than a person with a life well worth living.

Why accept actualism? In addition to being intuitively plausible, it is supported by the fact that existential generalization (i.e. from ‘A is F’, infer ‘there exists an F’) is standardly taken to be a valid inference rule. After all, if actualism is false, then it could be that A is F (in a world) even though A does not exist (in that world), in which case existentially generalizing from ‘A is F’ would lead from a true premise to a false conclusion. Thus, it is hard to deny actualism without also having to revise first-order logic.

If Boonin has overlooked the problem I have outlined in this section, it might be because he finds it intuitive that when someone's life is not worth living, she ‘would be better off never being conceived’ (2014: 239). If we say this about Dora, we may insist that P2 allows that Wilma's decision harms Dora. But again, we should not say this about Dora unless we are ready to deny actualism. (After all, if Dora does not occupy any well-being level in worlds in which she does not exist, then she would not have occupied a higher well-being level than she actually does if she had not existed.) If this seems problematic, it is worth noting that there are nearby claims that actualists can accept instead. For instance, even if Dora would not have been better off if she had never been conceived, it may be that her existence – that is, the state of affairs that she exists – is bad for her.⁵ However, even if Dora's existence is bad for her, and even if various other nearby claims about her are also true, P2 still excludes that Wilma's decision to conceive Dora harms Dora, since (given actualism) it remains true that Dora's well-being level would not have been higher if Wilma had not made that decision.

4. Possible responses

4.1 Abandoning actualism

One possible idea is that adherents of the Implausible Conclusion should reject actualism (despite its plausibility) in order to avoid the More-than-Most Implausible Conclusion. If we do reject actualism, we may claim that a subject occupies a well-being level – presumably zero – in worlds in which she does not exist. Since this implies that Dora would have been better off if Wilma had not decided to create her, it allows us to claim, consistently with P2, that Wilma's decision harms Dora.

However, while this proposal lets us avoid the More-than-Most Implausible Conclusion, it also lets us avoid the Implausible Conclusion for similar reasons. After all, if a subject has zero well-being in worlds in which she does not exist, then this also applies to Rocks (who does not exist) in Worth Living.⁶ And if Rocks's well-being level is zero, then he would have been better off if Wilma had chosen to create him. This suggests that Wilma's decision to create Pebbles harms Rocks, and thus, that P3 in the non-identity argument is false. In particular, P2 does not exclude that Wilma's decision harms Rocks, since we are now supposing that Rocks would have been better off if Wilma had not decided to create Pebbles. Moreover, adherents of the counterfactual comparative view of harm that underpins P2 usually also take the relevant counterfactual condition to be sufficient for harming (e.g. Feit 2015) – given the anti-actualist proposal, this entails (and does not merely not exclude) that Wilma's decision harms Rocks.⁷ The view that we can harm non-existent subjects is of course counter-intuitive, but that is simply because denying actualism is counter-intuitive. The present point is just that denying actualism provides little help for supporters of the Implausible Conclusion who want to avoid the More-than-Most Implausible Conclusion.⁸

4.2 Modifying P2

Another response is to modify P2. After all, contrary to what P2 allows, Wilma's decision to conceive Dora does seem to harm Dora – even given actualism. The most popular way to account for this intuition is by invoking a causal view of harming (Harman 2004; Gardner 2015). Roughly, and at a minimum, causal views claim that an action harms a subject if it causes a state of affairs that is, in some relevant way, bad for her. While different versions of the view spell out the notion of a bad state of affairs in different ways, Wilma's decision plausibly causes some such state – for example, that of Dora's having a miserable existence, or that of Dora's suffering horribly. Thus, plausibly, Wilma's decision harms Dora according to causal views.

However, causal views also imply that Wilma's decision harms Pebbles in Worth Living. The reason is that this decision causes Pebbles to exist with a significant disability, which, Harman and Gardner argue, also counts as a bad state of affairs in the relevant sense. Thus, just like the anti-actualist strategy, the appeal to causal accounts of harming does not help supporters of the Implausible Conclusion.

As a reviewer of this article for this journal, Boonin suggested that P2 could instead be modified as follows:

P2*: If A's act harms B, then either (i) A's act makes B worse off than B would otherwise have been, or (ii) A's act makes B exist in a condition that is not worth existing in, and B would otherwise not have existed in a condition that is not worth existing in.

Wilma's decision of conceiving Dora makes Dora exist in a condition that is not worth existing in, and Dora would otherwise not have existed in such a condition (since Dora would otherwise not have existed at all). Hence, P2* allows that Wilma's decision harms Dora. However, P2* still excludes that Wilma's decision harms Pebbles in Worth Living, since Pebbles's life is worth living (and hence it is false that Pebbles exists in a condition that is not worth existing in). P2* thus promises to provide a way of avoiding the More-than-Most Implausible Conclusion while accepting the merely Implausible Conclusion.

The problem with this strategy is that there are versions of Not Worth Living where P2* also implies that Wilma's decision is harmless. We can simply add to the case the further assumption that if Wilma had not done what she actually did, then she would have done something else which would have caused Dora to exist in a condition that is not worth existing in. There are different ways to achieve this result; for instance, we can suppose that if Wilma had not decided to conceive immediately, she would have decided to conceive in 60 minutes, which would have caused Dora to exist at the same well-being level as she actually does. In this version of the case, Wilma's actual decision still seems wrong. In particular, rather than deciding to conceive immediately, she ought to have decided to take a pill for two months before conceiving, thereby creating Rocks instead of Pebbles. But P2* now entails that Wilma's decision is harmless, since it is false that if Wilma had not made the decision that she actually made, then Dora would not have existed in a condition that is not worth existing in.⁹ Given the other premises of the non-identity argument, this leads to the unacceptable conclusion that Wilma's decision is not wrong. Hence, invoking P2* does not fully help to avoid the More-than-Most Implausible Conclusion.¹⁰

One could respond to this argument by removing the counterfactual conjunct from condition (ii) in P2*, as follows:

P2**: If A's act harms B, then either (i) A's act makes B worse off than B would otherwise have been, or (ii) A's act makes B exist in a condition that is not worth existing in.

Unlike, P2*, however, P2** clearly abandons the spirit of the counterfactual comparative view of harm that motivates P2, since condition (ii) is now non-counterfactual and non-comparative. It is hard to see how such a disjunctive view of harm could be motivated. At least, due to its disunity, it is at a significant disadvantage compared to competing ‘pure’ views of harm.

4.3 Modifying P4

As reviewer of this article, Boonin also suggested a second response, which is to modify P4 as follows:

P4*: If A's act (i) doesn't harm B and (ii) doesn't violate B's rights, then A's act doesn't wrong B.

If Wilma's decision in Not Worth Living violates Dora's rights, P4* allows that this decision wrongs Dora (even if it does not harm her). And if Wilma's decision in Worth Living neither harms Pebbles nor violates Pebbles's rights, P4* excludes that this decision wrongs Pebbles. So P4* promises to offer another way of avoiding the More-than-Most Implausible Conclusion while accepting the merely Implausible Conclusion.

The main problem with this strategy is that there are less complex variants of P4 that do not lead to either the Implausible Conclusion or the More-than-Most Implausible Conclusion (even given the other premises of the non-identity argument). This raises the question of why we should accept P4* rather than one of those variants. Here is an example:

P4**: If A's act doesn't even pro tanto harm B (i.e. if A's act doesn't even to some extent, or in some way, harm B), then A's act doesn't wrong B.¹¹

Unlike P4 and P4*, P4** does not lead to either the More-than-Most Implausible Conclusion or to the Implausible Conclusion. The reason is that even if we accept P2's implication that Wilma's decision does not harm Pebbles overall, the claim that her action does not even pro tanto harm Pebbles is much stronger, and, it seems, not particularly plausible.¹² (After all, Wilma's decision causes Pebbles to exist with a disability that substantially negatively impacts her well-being.) At the same time, P4** promises to capture a version of the ‘no harm, no foul’ intuition that initially motivated P4 – that there is a close connection between harming and wronging such that if an action is harmless then it does not wrong anyone (Boonin 2014: 4).

Since options such as P4** are available, then, the question arises why we should accept P4*. Its disjunctiveness speaks against it, and the ‘no harm, no foul’ idea does not favor it. Moreover, the fact that P4* leads to the Implausible Conclusion (given the other premises of the non-identity argument) should itself be seen as a reason to doubt it, since the Implausible Conclusion – as even Boonin agrees – is counter-intuitive.

A possible reason to favor P4* over P4** may concern cases of harmless right violation. If A reads B's diary or peeps in on B while B is undressing, without B's permission or knowledge, for example, then one might think that A wrongs B even if B never finds out or is otherwise affected by A's action.¹³ However, whether A's actions in these cases really are harmless is unclear. It depends on controversies about the nature of well-being. Typical versions of hedonism support thinking that A's actions are indeed harmless, since her actions do not make a difference to B's balance of pleasure over pain. But typical desire theories of well-being support thinking that A's actions are harmful – at least if we assume, realistically, that A's actions frustrate some of B's desires (such as a desire that no one reads their diary or peeps in on them while they are undressing).

That said, if it turns out that there really are cases of harmless rights violations (and that the actions performed in such cases wrong the person whose rights are being violated), the following variant of P4 is still an option:

P4***: If A's act (i) doesn't pro tanto harm B and (ii) doesn't violate B's rights, then A's act doesn't wrong B.

P4*** is strictly weaker than (and hence more probable than) P4*. It also allows that harmless rights violations are wrongings, and it does not lead to the Implausible Conclusion. Thus, the question again arises why we should accept P4*, which does lead to the Implausible Conclusion.

5. A diagnosis

The reason why the non-identity argument leads to the More-than-Most Implausible Conclusion is, I believe, that its premises imply a person-affecting moral view (e.g. Heyd 1992). The basic idea behind such views is that a world w is worse than a world w* only if w is worse than w* for someone. Most straightforwardly understood, this means that w is worse than w* only if, for some subject A, A's well-being is lower in w than in w*. A well-known problem for this view is that it implies that a world w₁, in which one person has a life that is not worth living, is not worse than another world w₂, in which another person has an excellent life. This conclusion is similar to the More-than-Most Implausible Conclusion – the only difference is that it concerns worseness rather than wrongness. And once we see that the premises of the non-identity argument also imply a person-affecting view, though one that concerns wrongness rather than worseness, it is unsurprising that they generate similar problems. More exactly, as we have seen, P2, P4 and P5 together imply that realizing a world w rather than a world w* is wrong only if, for some subject A, A's well-being level is lower in w than in w*. If there is no such subject, P2 implies that realizing w rather than w* does not harm anyone, which, together with P4 and P5, implies that it is not wrong. And the problem is precisely that there is no such subject in Not Worth Living, since if Wilma creates Rocks, then Dora will never exist, and vice versa.

To avoid these problems, supporters of the person-affecting view may wish to modify their view so that it allows that w₁ is worse than w₂ even if no subject has a lower well-being level in w₁ than in w₂. Arguably, though, such modifications abandon the core idea behind person-affecting views and instead introduce a fundamentally impersonal element (Heyd 1992: 112–15). For example, when discussing a similar case, Nils Holtug suggests that w₂ is better than w₁ because ‘the people that will exist if w₂ comes about will benefit more from coming into existence than the people that will exist if w₁ comes about’ (Holtug 2001: 363, fn. 3, notation adjusted). While Holtug views this as a person-affecting view, this label is misleading – since the view allows that w can be worse than w* even though w is not worse than w* for anyone, it is an impersonal view in personal clothing.¹⁴ In particular, unlike standard person-affecting views, Holtug's suggestion allows that the world in which Wilma creates Rocks is better than the one in which she creates Pebbles. Thus, it is not committed to the Implausible Conclusion.

6. Conclusion

I have argued that the premises of the non-identity argument lead to the More-than-Most Implausible Conclusion, and that the argument is therefore unsound.¹⁵ For this reason, I believe, we have not yet been given sufficient reason to accept the merely Implausible Conclusion.

Can the non-identity argument be rescued? Perhaps there is a way to modify its premises so that they, thus modified, (i) do not entail the More-than-Most Implausible Conclusion, (ii) still entail the Implausible Conclusion, and (iii) are otherwise sufficiently plausible and independently motivated. If such a strategy were to be presented, it would help support the Implausible Conclusion. But I myself, at least, am unable to think of a strategy that satisfies these three conditions. In particular, for each such suggested modification, we should ask ourselves: why not accept only some weaker variants of the premises, which do not entail even the merely Implausible Conclusion – a conclusion that most people (including Boonin) find intuitively implausible? I leave it for supporters of the Implausible Conclusion to address these questions.¹⁶