2. A NEW SOLUTION TO AN OLD PROBLEM

Consider a prospective mother, Wilma. Wilma wants to conceive a child, but has just learned after a visit to her doctor that she is affected by a medical condition which, if she conceives a child now, would cause her child to be born incurably blind. This bad news is, however, accompanied by good news: Wilma's condition can be cured, so if she takes a pill once a day and delays the conception by two months, the child that she will then conceive will be sighted.

Would it be morally wrong if Wilma did not wait until her condition is cured before she conceives? Taking a pill once a day is, after all, a small sacrifice for Wilma, but it would prevent a bad condition affecting her child.

If the answer to this question is ‘yes’, then this conflicts with other deep-seated moral intuitions that most people seem to have. First, due to basic facts about human physiology, the child that Wilma could conceive now would be different from the one that she would conceive if she waited for two months, because (for one thing) it would originate from a different ovum. To distinguish the child that would be born now from the child that would be born in two months, we will call the former Pebbles and the latter Rocks. Since Pebbles could only be conceived now, it seems that if Wilma waited and conceived Rocks instead, this would not be better for Pebbles. In other words, waiting would not benefit Pebbles by making her sighted, but it would simply ensure that a different person (Rocks) would be born instead, albeit a sighted one. Therefore, Pebbles does not appear to be harmed by Wilma's choice of conceiving a child now. But then, in what sense could we say that Wilma's act is morally wrong, if it doesn't harm her child?

There are many instances of the non-identity problem. Wilma's story exemplifies what Boonin (2–7) calls the direct version of the problem, but there is also an indirect version of it. The following example illustrates how the indirect version can arise. Imagine that a wealthy society needs to decide whether or not to build a new energy plant. If the plant is built, the wealthy society will get an immediate benefit from having access to cheaper energy production. However, the plant will produce waste, which can only be disposed of safely for a determinate amount of time. Even if the highest precautions are taken, it is known that the waste will leak in 500 years’ time. When this happens, some 10% of the population that will then be alive will die as a result of being exposed to the toxic leak.

As in Wilma's case, it seems that the wealthy society should not build the plant, because this would harm future people. However, as in Wilma's example, a case can be made that the people who will be born in the future if the wealthy society chooses to build the plant would not be born otherwise. This is because having a different energy policy would affect people's lives, which as a result would change the timing of conception of children in that society. Taking this into account, we can be reasonably sure that, after a few generations, none of the people who will be born would have been born if another policy had been chosen. So, as in Wilma's case, it appears that it is not possible to say that choosing to build the plant would harm anyone. So, building the plant could not be wrong.

Boonin's solution to the non-identity problem is to accept the conclusion that actions such as Wilma conceiving Pebbles or the wealthy society building the energy plant are not morally wrong. This solution is surprising, because virtually all other proposed solutions to the problem in the literature have focused on finding ways to resist that intuitively implausible conclusion.

If Boonin's ‘solution’ is right, this should have implications for the way we evaluate reproductive decisions and policies that have consequences reaching into the future, for example in the context of climate change. If it is not morally wrong to bring about sub-optimal outcomes in non-identity cases, then it follows that our obligations towards future people are much less stringent than we may have thought. Unless other side constraints apply, on which Boonin does not go into much detail, we may be under no obligation to change our CO₂ emission patterns for the sake of leaving a cleaner environment to future generations.

3. THE STRUCTURE OF THE ARGUMENT

To establish his solution, Boonin first identifies the five premises that give rise to the non-identity problem and what he calls the implausible conclusion to which they jointly give rise. The book is then structured around the detailed discussion of the premises and the conclusion, each of which occupies a chapter.

Throughout the book, the discussion focuses on the direct version of the non-identity problem, exemplified by Wilma's case. Boonin seems to take this case to be more fundamental and argues that most of the arguments targeted to the direct version can be re-formulated to target the indirect version.

The five premises that Boonin identifies are the following (27):

1. P1: Wilma's act of conceiving now rather than taking a pill once a day for two months before conceiving does not make Pebbles worse off than she would otherwise have been.

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

3. P3: Wilma's act of conceiving now rather than taking a pill once a day for two months before conceiving does not harm anyone other than Pebbles.

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

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

Which jointly give rise to the problematic conclusion:

• C: Wilma's act of conceiving Pebbles is not morally wrong.

According to Boonin, any successful solution to the problem should satisfy three requirements. First, there is the independence requirement, which states that any argument for a proposed solution should be plausible in its own right, independently from the fact that it solves the non-identity problem. For instance, this requirement would forbid denying P5 simply on the basis that rejecting this premise would block C.

The second requirement is robustness. According to this requirement, the reason for accepting a solution to the non-identity problem should apply to all instances of the problem. For instance, one could deny P3 by pointing out that there may be other people who are harmed by Wilma's choice of conceiving Pebbles, for example Pebbles’ siblings who may receive fewer resources, or members of society at large who would have to make special arrangements to accommodate blind Pebbles’ needs. But even if these considerations turned out to be relevant, it would still not be enough to solve the non-identity problem in general, because we could easily construct cases where it is more straightforward that nobody is harmed.

Finally, Boonin formulates a modesty requirement: a successful solution should not prove ‘too much’. In other words, it should not generate even more implausible implications than the problem that it is supposed to solve. For instance, if we denied P5 on the grounds that Wilma's act is morally wrong, even if it does not wrong anyone, because it is selfish, then this solution would force us to accept that all selfish acts are morally wrong. But this claim would clearly be too strong, so this solution would not pass the modesty test.

The three requirements for a successful solution put forward by Boonin are key to his discussion of the arguments that aim to reject each of the five premises. For the purposes of this review, I will focus on Boonin's discussion of P5 and C in Chapters 6 and 7.

P5 is particularly interesting because its discussion highlights some important aspects of Boonin's position. P5 asserts that if an action is morally wrong then it necessarily wrongs someone. Not all ethical theories agree on P5; consequentialism is a notable exception (Parfit Reference Parfit1986: ch. 16). On a consequentialist view, an action is wrong if it produces a suboptimal outcome relative to some appropriate measurement criterion, for example aggregate total utility. It is not hard for a consequentialist to solve the non-identity problem simply by negating P5: in Wilma's case, for example, a total utilitarian would claim that conceiving Pebbles now would be wrong because it brings about a consequence associated with lower total utility than conceiving Rocks later (provided that the disutility to Wilma of waiting is small).

The ease with which consequentialism appears to solve the non-identity problem could perhaps be taken as an argument in its favour. The solution also satisfies the robustness requirement, as it is equally applicable to the different versions of the problem. Boonin argues, however, that we shouldn't accept consequentialism only for the sake of solving the non-identity problem, as this would violate the independence and modesty requirements.

According to Boonin, a consequentialist could invoke the following (152):

In order to reject P5, the consequentialist needs to produce an impersonal criterion to judge the goodness of the overall consequences of an act, according to which what matters is the absolute value of aggregate welfare. Since impersonal goodness can be compared across states where different sets of people exist, this provides a reason to claim that an act may be wrong (impersonally), even if it does not wrong anybody. However, Boonin denies that we have reasons to accept an impersonal criterion other than the fact that it helps to solve the non-identity problem, so the argument falls short of the Independence requirement (section 6.1.2). Moreover, Boonin thinks (section 6.1.3) that the Moderate Principle violates the modesty requirement because it seems (to him) to require the creation of additional people when the welfare balance of so doing is positive, or the replacement of existing people with possible persons whose welfare would be greater.Footnote ¹

Boonin then argues in Chapter 6 that the conclusion C might not be so implausible after all. Not only does accepting C allow us to vindicate our intuitions with respect to P1–P5, Boonin argues; it also delivers intuitive results in a variety of cases where other solutions seem to violate the modesty requirement (Chapter 7). Boonin's arguments depend on what we may call the person-affecting intuition, i.e. the idea that a state of affairs x is better than y only if there is someone for whom x is better than y. Therefore, anybody who challenges this intuition will be less moved by Boonin's arguments against the Moderate Principle.

4. POSSIBLE OBJECTIONS AND PRACTICAL APPLICATIONS

A direct challenge to the person-affecting intuition indeed comes from consequentialism, on grounds that are independent of the non-identity problem, which may include an appeal to rational intuitions (Singer Reference Singer2005). A consequentialist may argue that a state of affairs x is better than y if the same number of people exist in each state and the aggregate well-being of the people alive in x is greater than the aggregate well-being of the people alive in y, even if the two populations are different and, therefore, there is no-one for whom x is better than y.

A further observation, which is not discussed by Boonin, is that it may be possible to reconcile the person-affecting intuition with a broadly consequentialist solution to the non-identity problem by focusing on the epistemic circumstances of the agent at the moment they have to make a choice. Most of the time, decision-makers face some level of risk; in non-identity cases, the risk may be relative to both the welfare and the identity of future people. Even assuming that the uncertainty about welfare is resolved, the idea is that, from the decision-maker's perspective, the identity of the possible people who would be brought into existence could still be unknown. This is evident in the indirect version of the non-identity problem: at the moment of choice, the wealthy society may have no way to determine who, among all the theoretically possible individuals, would be brought into existence as a result of each policy. In particular, for any theoretically possible individual, the present wealthy society might be unable to determine whether this individual would be born under either policy. Given this uncertainty regarding the identity of future people, the wealthy society may well come to the conclusion that, from the present point of view, choosing to build the plant would damage the prospects of all possible people, at the time when the choice is made, including those who will actually come into existence. Since imposing a risk of a loss on someone is bad, not building the plant would be the morally better option.

This line of argument may be understood as denying P4, the premise stating that a necessary condition of wronging is harming a person. When the wealthy society chooses to build the plant, knowing that it is lowering the expectations of possible future people, one could argue that it is wronging those future people, whoever they subsequently turn out to be. However, after the uncertainty about the identity of the future people is resolved (i.e. when the future generation comes into existence), it would still be plausible to say that these actual people are not harmed, because it would not have been possible (reasoning ex post) to bring them into existence under better circumstances. In other words, the people who would turn out to be actual in the future would be have been wronged, because someone took an unfavourable bet on their behalf at the time that the plant was built, but not harmed, since it would be revealed ex post that they in fact had no chance of coming into existence under any other scenario.

An advantage of this line of argument is that it would allow us to keep the person-affecting intuition, along with the intuitions which ground our acceptance of all the premises with the exception of P4, while consistently employing the Moderate Principle to resist the implausible conclusion. The person-affecting intuition would still be respected, because the interests of possible people would be taken into account only in so far as among them are the people who will be actual (presently, we just don't know who they are), and it would be only these future actual people who would be in a position to raise a complaint for having been wronged. While this argument against P4 shares some of the features of Veil of Ignorance arguments discussed by Boonin (Chapter 5), it is not explicitly discussed there.

Boonin could object that this argument draws too stark a distinction between the cases in which we know who will come into existence and the cases in which we don't. If that is true, the proposed solution would not meet the robustness requirement. I don't have a definitive answer to this objection. However, I would argue that as a matter of fact it might be empirically impossible for us, in most non-identity cases, to know the identity of future people in advance of their conception.