I

Many of our choices effect who will exist in the future. These include the choice of having a child, of giving priority to young people over old when saving lives, and of offering generous state-funded parental leave. In many of these cases it seems that one outcome is worse than another in virtue of the wellbeing of people who do not exist in both. For example, it seems, creating a very unhappy person makes the world worse, other things being equal. Some would also say creating a very happy person makes the world better, other things being equal. It would be easy to justify these verdicts if it can be better, or worse, for a person to exist than not to exist. But that seems to require that things can be better, or worse, for a person even in a world in which she does not exist, which sounds paradoxical.

This paradoxical-sounding claim is defended in Ingmar Persson's latest book Inclusive Ethics.Footnote ¹ He argues that in a world in which a person does not exist, she is a merely possible being – a being that never exists – and that for such beings it is worse not to exist than to exist with a good life. Furthermore, he argues for this claim from what he calls ‘incontestable’ premises. We argue that the premises are far from incontestable. Persson's argument, as stated, has false premises and is invalid. One can reconstruct the argument to make it valid, but the premises are still highly contestable. Finally, we argue, one can make sense of our obligations to future generations without letting merely possible beings into the moral club.

II

Persson's argument is presented in the following passage:

1. (1) for a being who has never existed nothing is either intrinsically good or bad,

and

1. (2) the fact that nothing is either intrinsically good or bad for a being is worse for it than the fact that things are overall intrinsically good for it.Footnote ²

Claims (1) and (2) are the premises, which are claimed to be ‘incontestable’, and the conclusion is:

1. (3) never having existed is worse for possible beings than an existence in which things are overall intrinsically good for them.

Note that (1) talks about a being who has never existed. This does not rule out that it will exist in the future. But Persson wants to say that non-existence can be worse for merely possible beings, beings that never have existed and never will exist. So, readers should interpret (1) and (3) as talking about merely possible beings.

Persson finds (2) ‘incontestable’, but as stated it is false, since the two facts mentioned in (2) are incompatible. Necessarily, if it is a fact that nothing is intrinsically good or bad for a being, then it is not a fact that things are overall intrinsically good for this being. Hence, it cannot be true that the fact that nothing is intrinsically good or bad for a being is worse for it than the fact that things are overall intrinsically good for it.Footnote ³

Another problem is that the facts compared in (2) are evaluative facts, facts about whether things are good or bad for a being. But it is unclear what it means to say one evaluative fact is worse than another for a being.

Finally, the argument is invalid: (3) does not follow from (1) and (2). According to (2), a certain evaluative fact is worse for a being than a certain other evaluative fact. But ‘never having existed’ in (3) refers to the non-evaluative state of affairs (or fact) of a being never having existed.

III

The following reconstruction of Persson's argument, where evaluative fact-talk is replaced by non-evaluative state-of-affairs-talk throughout, avoids the problems of the originally-stated argument:

1. (4) For a merely possible being, non-existence is neither overall good nor overall bad.

2. (5) Any state of affairs which is neither overall good nor overall bad for a being is overall worse for it than a state of affairs that is overall good for it.

Therefore,

1. (6) For a merely possible being, non-existence is overall worse than an existence that is overall good for it.

This reconstruction is valid, assuming ‘existence’ and ‘non-existence’ pick out states of affairs. Moreover, it avoids comparisons of incompatible evaluative facts. Finally, it makes clear that the relevant comparisons concern the overall values of states of affairs (existence or non-existence), i.e. the total intrinsic value for the being of all the things that the states of affairs would realize, if they obtained.Footnote ⁴ We should add that the reason why (4) is true is that everything lacks positive and negative intrinsic value for the being.

But the argument is still problematic. According to (5), something that is overall good for a being is better for it than something that lacks overall positive and negative value for it. But something is better for a being than something else only if both things have value for the being. The same holds for any comparative notion. One thing is taller than another only if both have height; one thing is heavier than another only if both have weight; one thing has a higher temperature than another only if both have temperature. The question, then, is what value non-existence has for a merely possible being.

Persson states that non-existence has neutral value for a merely possible being. Indeed, he claims that everything is neutral for a merely possible being. However, he characterizes ‘neutral value for a being’ as ‘being neither intrinsically good nor intrinsically bad for a being’.Footnote ⁵ This is an implausible characterization of ‘neutral value for a being’ because, plausibly, some states of affairs are undefined in value for a being. Candidates include contradictory states of affairs (e.g. that 2 + 2 = 5), necessary states of affairs (e.g. that 2 + 2 = 4), evaluative states of affairs (e.g. that happiness is good for people), and states of affairs that concern other people's wellbeing (e.g. that a stranger is unhappy).Footnote ⁶

Furthermore, being neutral for one is different from being neither good nor bad for one. If something is neutral for one, it has a value that is neither positive nor negative, but which can be compared to positive and negative values. This is analogous to:

1. (a) having zero temperature, which is having a temperature that is neither positive nor negative, and thus different from lacking positive and negative temperature just because one lacks any temperature, or

2. (b) having a weight that is neither heavy nor light, which is different from being neither heavy nor light just because one lacks any weight, or

3. (c) having a height that is neither tall nor short, which is different from being neither tall nor short just because one lacks any height.Footnote ⁷

Persson could deny that things must have value to stand in value relations or that neutral value is a value in its own right. But such claims are highly contestable and Persson does not give arguments for why value comparisons differ so radically from other comparisons.

Premise (4) is also highly contestable. Indeed, that there are merely possible beings is one of the most contestable claims in modal metaphysics.Footnote ⁸ Persson claims not only that there are non-existent beings, but also that things are neutral for them and, given premise (5), that things can be better or worse for them.

Consider the existence claim first. It seems incoherent, since it seems to assert the existence of non-existent beings. But Persson believes that in a certain sense of ‘exist’, the claim is true:

Persson's claim that

follows from

is puzzling. B states that in some possible world some beings will begin to exist in the future. Why would it follow that there are and thus exist, in any sense, merely possible beings here in the actual world? Compare: it is possible that the pope has two children in the future, and thus possible that two children of the pope will begin to exist in the future. But why would it follow that there exist, in any sense, merely possible beings here in the actual world? Regardless of what sense of ‘exist’ is assumed, we need to know how to derive A from B. We could derive it if we accepted the Barcan formula as a bridge principle:

For then we could make use of the following instance of the Barcan formula:

and from B derive that there are beings such that it is possible that they begin to exist, which entails A, assuming that they in fact will not begin to exist. But the Barcan formula is highly contestable, and Persson claims (in correspondence) that this is not what he intended.Footnote ¹⁰ So, we do not see how he can incontestably derive A from B.

Even if we could establish that A follows from B, we need to know what the relation is between merely possible beings in one world and beings that begin to exist in some alternative possible world. Consider Tommy, a child born with a painful condition who lived an utterly miserable life. Persson seems committed to saying that things would have been better for Tommy if he had not existed, and that this holds in virtue of the fact that if he had not existed, things would have been neutral for a merely possible being. Either this merely possible being and Tommy are identical or they are not. Both options are problematic.

If they are identical, then a merely possible being which is not concrete (i.e. lacks space-time location and causal powers) could have existed and been concrete. But being non-concrete seems essential to its bearers: if something is non-concrete, it is essentially non-concrete. Similarly, if something is concrete, it is essentially concrete.

If they are not identical, then it is unclear how Persson can establish that things would be worse for Tommy, if he did not exist. If we had not created Tommy, he would not have been around in any sense, since being concrete is essential to Tommy. Since Tommy would not have been around, things would not have been worse for him, for the relation ‘worse for’ requires a subject for whom things are worse. Even if there is a distinct merely possible being in the Tommy-less world and things are worse for this being, it does not follow that things are also worse for Tommy in that world, for we have assumed that the merely possible being is not identical to Tommy.Footnote ¹¹

Let us now turn to the crucial value claim: things can be neutral, better or worse for merely possible beings. This claim is far from obvious. After all, a merely possible being is not a concrete being, an animal, a conscious being, a human, and so on. At most, it is a merely possible concrete being, a merely possible animal, and so on. Why think that a being that merely possibly exemplifies these features can stand in value relations? Why not think that standing in a value relation requires being concrete in some way, for example, having a mind, body, space-time location or causal powers? We need an argument for the controversial claim that things can be better or worse for non-concrete beings.

Persson offers an argument from analogy. The analogy involves a comparison between a non-existent being and an anencephalic infant that lacks the capacity for consciousness and so never becomes conscious:

According to Persson, the crucial similarity between the anencephalic and the merely possible being is that both lack consciousness. This is supposed to show that nothing is intrinsically good or bad for them, and hence that lacking conscious existence is neutral for them. The analogy is contestable, however, for there is a clear difference between the two beings. Although both lack consciousness, the anencephalic is a concrete being and an animal. The merely possible being is only merely possibly concrete and merely possibly an animal. Thus, the anencephalic seems a better candidate for a being for which things can be neutral. Remember that being neutral for is not just a lack of the relations being good for and being bad for; it is an evaluative relation in its own right. Moreover, the analogy breaks down for flesh-and-blood individuals, such as the previously discussed Tommy, for he would not exist in any sense if he were not created (assuming such a concrete individual could not have been non-concrete).

Persson could reply that states of affairs can have value for a being if that being has a capacity for being concrete, an animal, and so on, and both the anencephalic being and the merely possible being have these capacities. This reply works at most for merely possible beings, not for the flesh-and-blood Tommy, since he does not exist at all in the Tommy-less world (assuming, again, that he could not be non-concrete), and thus cannot exemplify any capacities in that world. Moreover, the reply succeeds only if possibly having a certain property F is sufficient for having a capacity to have F. But this seems false. For example, the fact that you possibly jump to the moon, i.e. that there is a possible world in which you do this, does not entail that you have the capacity to jump to the moon.Footnote ¹³ Having a capacity to do or have something (in the ordinary sense of ‘capacity’) requires more than having the purely modal feature of being such that one possibly does or has it. Arguably, the capacity must be somehow grounded in features that are not purely modal.

Persson's analogy also assumes that lacking consciousness is worse for the anencephalic than having a conscious good life, that the anencephalic would benefit from having a conscious good life. But whether this is true seems to depend on which theory of our identity is correct. Suppose our identity is essentially tied to our capacity for consciousness.Footnote ¹⁴ Then none of us could exist as the anencephalic in Persson's example, since it lacks this capacity. Suppose the anencephalic somehow acquires a fully functional brain that is necessary and sufficient to realize the capacity for consciousness. Then a new being, one of us, comes into existence. Call this being Beth. If Beth has a good life, this does not imply that the anencephalic benefits. For it seems Beth is the only welfare subject in this case, and she is not identical to the anencephalic.

Persson might claim that the anencephalic benefits in virtue of standing in some intimate relation to Beth. For example, suppose the anencephalic is essentially a human body. Then when Beth comes into existence, this human body might constitute Beth, or it might acquire Beth as a spatial or temporal part. If Beth has a good life, then the human body might also have a good life in virtue of standing in some such relation to Beth.

But this reply is problematic. Beth is a welfare subject. If Beth's body is also a welfare subject, there are too many welfare subjects. To emphasize the problem, consider the case of body-switching. Suppose that Beth's brain, which realizes her capacity for consciousness, is transplanted into a different human body that somehow developed without a brain. Since Beth's capacity for consciousness is essential to her, Beth switches bodies. Now suppose the body-switch benefits Beth because her new body is superior in ways conducive to her well-being. Moreover, suppose Beth's life after the body-switch is overall good. If human bodies are welfare subjects, then Beth's new body, call it Body, should also benefit. There should be two beneficiaries – Beth and Body. In that case, Body's benefit would be distinct from Beth's. Their benefits might even be of different sizes. (For example, Beth might benefit only marginally from the body-switch while Body benefits greatly.) But it seems implausible that there are two beneficiaries here. Surely, Beth is the only beneficiary of the body-switch.

If our identity is essentially tied to our capacity for consciousness, we should deny that Body benefits from acquiring Beth's brain. But then it seems we should also deny that the anencephalic in Persson's example (a human body) would benefit from acquiring a brain.

IV

We conclude that Persson's argument that non-existence is worse for a merely possible being than an existence in which things are overall intrinsically good for that being fails.

Is this bad news? Not if our obligations of beneficence to future generations in non-identity cases can be explained by appealing to their existence being good or bad for them. We can have a reason to create a person if her life would be good for her. We needn't assume that her existence would be better for her than her non-existence.

Persson is not convinced that appealing to what would be good or bad for a person (as opposed to what would be better or worse for her) can justify creating or refraining from creating a person. He provides both a specific argument and a general argument. His specific argument is this. Suppose you can either create A with a good life or benefit a different person, B, who would exist independently of your actions.

This is unconvincing. Persson assumes that defenders of a reason to create the non-comparative benefit of a good life must be moral actualists – i.e. that they must say that you have a reason to confer a non-comparative benefit only if the beneficiary actually exists (or will exist).Footnote ¹⁶ This view is vulnerable to Persson's objection. But consider the following non-actualist view: if an action would bring about an outcome in which someone exists with a good life, then this fact provides a reason to perform the action. This view avoids Persson's objection, since on this view the existence of the reason does not depend on who does or will exist. It is enough that the relevant counterfactual is true.

Persson's general argument for there being no reason to benefit a being in a non-comparative sense unless one thereby also benefits her in a comparative sense is as follows:

We agree that there is no contrastive reason to give a person a non-comparative benefit rather than do what results in her having an even greater benefit. However, this does not show that there is no contrastive reason to give her a certain non-comparative benefit – a pleasure, say – rather than do what results in her not having this benefit. Thus, we think, Persson has not established that the only reason to create a being with a good life is that existence with a good life is better for that being than non-existence. We can have a contrastive reason to create a person who will have a good life rather than do what results in her not existing at all, and thus not having any non-comparative benefit.

V

We have scrutinized Persson's argument for the claim that non-existence is worse for a merely possible being than an existence in which things are overall good for that being. Far from being ‘incontestable’ the argument's premises, as initially stated, are false, and the argument is invalid. We presented what we take to be the most promising reconstruction of Persson's argument. Although the reconstructed argument is valid, the premises are highly contestable. They assume that there are merely possible beings and that things can be better or worse for them. Persson's argument for the existence claim is unconvincing and his argument for the value claim is based on a questionable analogy involving anencephalic infants. Finally, we showed that Persson is not justified in his dismissal of the idea that we can explain our obligations to future generations by appealing to facts about their existence being good or bad for them.

In sum, Persson has not made a convincing case for giving merely possible persons membership in the moral club. They will have to look for a different guarantor.Footnote ¹⁸