Reply to Spears’s ‘The Asymmetry of Population Ethics’

Abstract

Is the procreation asymmetry intuitively supported? According to a recent article in this journal, an experimental study suggests the opposite. Dean Spears (2020) claims that nearly three-quarters of participants report that there is a reason to create a person just because that person’s life would be happy. In reply, I argue that various confounding factors render the study internally invalid. More generally, I show how one might come to adopt the procreation asymmetry for the wrong reasons by misinterpreting one’s intuitions.

According to the received view, the following procreation asymmetry is supported by its intuitiveness. On the one hand, (1) the fact that a life would be bad for the person living it does, all else being equal, give us a reason against creating it. On the other hand, (2) the fact that a life would be good for the person living it does not, all else being equal, give us a reason to create it (call this the no reason claim). ‘All else being equal’ indicates that we are only concerned with reasons directly generated by the fact that the life would be good for the child, but not with surrounding factors, such as the parents’ well-being.

Against the received view, ¹ Dean Spears (2020) suggests in an excellent article in this journal that, in fact, most people find a procreation symmetry to be intuitive. In an experimental study, nearly three-quarters of the participants, while agreeing about (1), reject the no reason claim (2) and adopt, instead, the following reason claim: (2*) the fact that a life would be good for the person living it does, all else being equal, give us a reason to create it. Or so Spears interprets the data. In reply, I argue that Spears’s experimental study is internally invalid. ² Various confounding factors might explain the surprising results. Generally, my argument shows how one might come to adopt the procreation symmetry for the wrong reasons.

By way of motivation, let me point out that the question of whether the procreation asymmetry is true has, prima facie, important implications beyond one’s personal procreative choices. It also seems to matter for administrative decisions that influence population size on a large scale. In particular, it seems to have implications for the strength of our reasons to avoid human extinction. ³ The number of happy people that might live in the future is vast (Greaves and MacAskill 2021). This consideration might generate strong moral reasons to avoid human extinction if the procreation symmetry is true and we have reasons to create people just because they would be happy. If, on the contrary, the procreation asymmetry is true, then our reasons for avoiding human extinction might be significantly weaker.

Thus, it seems important to determine whether the procreation asymmetry is true. In this regard, various authors suggest, the alleged intuitiveness of the asymmetry is not just some relevant evidence. Instead, several defences of the negative claim crucially rely on the supposed intuitiveness of the asymmetry. They do not even aim to show that independent premises force us to accept it (e.g. Roberts 2011; Kolodny 2022). Instead, based on the intuitiveness of the asymmetry, they take the procreation asymmetry for granted and try to show how the asymmetry fits into a possible consistent and connected account of population ethics. Accordingly, it is important to scrutinize Spears’s results.

1. The study

According to Spears’s interpretation of the data, in two studies (No. 1a, n = 232; No. 1b, n = 186), few survey respondents support the procreation asymmetry, while roughly three-quarters support the procreation symmetry.

Here is the design of his studies. First, the participants read about a couple who are unsure about whether they should have another child. “They already have two children, and their family is happy. … The couple sees reasons for and against conceiving another child and has not yet decided what the right thing to do is” (Spears 2020: 439). The participants were then presented with various “possible facts” that the couple might (“somehow”) come to know, e.g. “the new child would be especially good at sports” (Spears 2020: 439–440). The participants were asked whether these facts should morally influence the parents’ decision to have a child. Here is the fact that is supposed to settle whether the procreation symmetry or the asymmetry is intuitively supported: “The new child would have an especially good and happy life, well worth living and full of very much joy and well-being” (Spears 2020: 440). To determine whether a fact should have moral influence, participants must classify it as ethically irrelevant, an ethical reason against having the child, or an ethical reason in favour of having the child.

Only a small minority of participants in Spears’s study (12.5% in 1a; 15.1% in 1b) classified the prospect of an especially happy child as ethically irrelevant while classifying the fact that “the child would have an especially bad and unhappy life” as an ethical reason against having the child. Only these participants, Spears concludes, find the procreation asymmetry to be intuitive. Nearly three-quarters of respondents (74.1% in 1a; 74.7% in 1b) classified the fact that the child’s life would be “especially good and happy” as an ethical reason in favour of having the child while agreeing that the prospect of a bad and unhappy life speaks against procreation. These participants, Spears concludes, find the procreation symmetry intuitive.

Spears does not discuss the fact that these results are, to a certain degree, in tension with the data of his other studies (No. 2, n = 239 ⁴ /No. 3, n = 264). In these studies, participants did not evaluate the ethical relevance of a happy life or a miserable life simpliciter. Instead, they looked at a certain quality of life, ranging from 0 (worst) to 10 (best). Spears observes a tendency: the higher the quality of life, the higher the percentage of people who report that the quality of life counts ethically in favour of having the child. But he leaves undiscussed that, overall, the percentage reporting that high well-being counts in favour of procreation remains comparably low: even for the highest quality of life, the percentage doesn’t rise above roughly 47%. ⁵ Leaving these peculiarities aside, I shall argue in the following that the experimental setup of Spears’s study is internally invalid. Various confounding factors might explain why many participants classified the fact in question as an ethical reason in favour of having the child.

2. Ceteris paribus

First, the questionnaire does not clarify that people should abstract from ways in which the child’s well-being might benefit others. Presumably, the happy child would be a source of joy and inspiration for her siblings and parents.

In Study No. 3, Spears tried to exclude this confounding factor by adapting the experimental setup of Study No. 2, which involved considering various qualities of life, by including the following ceteris paribus statement:

For the purposes of this survey question, you should assume that, although life will be different for the parents if they have the additional baby, the good and bad consequences are balanced: the parents, their other children, and every other person (except the new baby itself) will be just as well off whether the parents have the baby or not. (Spears 2020: Statistical Appendix)

Spears finds that this addition has no effect on participant judgements (note that this was only tested with respect to Study No. 2, where support for the reason claim seemed considerably lower in the first place). The statement, however, faces two problems. First, it does not say that parents also believe that well-being levels will remain unaffected by having the child. The participants, who are only asked what facts should matter “for the couple’s decision”, might think that there are subjective norms that are relative to the decision maker’s beliefs. ⁶ For example, it might be intuitively appealing that if a doctor is deeply convinced that medicine will cure her patient, then she has reason to administer it even if, unbeknown to the doctor, the medicine will be without effect. Similarly, participants might have the intuition that parents have reason to have an additional child, given that the parents think that their child would be a source of inspiration and joy for them and their other children. To be sure to exclude such confusions and to achieve greater precision, it would help to clarify that the parents also believe that well-being levels remain unchanged. ⁷ Another point to note about Spears’s ceteris paribus addition is that it presents, as it were, an act utilitarian understanding of “all else equal”. A pluralist, for instance, might object that even if the person’s well-being levels are unaffected, there might still be extraneous deontological considerations that are affected by having an additional child, deriving, for example, from the relation of the parents to their already existing children. I will return to this in section 4.

3. Removing worries vs. being a standalone reason in favour

Crucially, the limited option set of ethical irrelevance, reason in favour of and reason against having a child does not do justice to two important distinctions. This section considers the first.

Recall that the parents were undecided concerning whether to have another child. Probable explanations for this include worries that their child might have a severe illness or would not be happy. Regarding themselves and their existing children, the parents might worry that an unhappy child would be very demanding to care for and put their other children’s well-being at stake. In such a situation of indecision, coming to know that the child would be very happy would certainly be ethically relevant, indicating that these risks are inexistent or lower than expected.

Thus, even if the procreation asymmetry were true, the prospect that the additional child would be very happy would be ethically relevant because it removes certain worries. Presumably, intuitions rooted in such considerations will be classified as reasons in favour of procreation. After all, the participants in Spears’s study could only choose between ethical irrelevance, reason in favour of and reason against procreation.

However, such intuitions do not amount to intuitive support for the claim that there are reasons to create a life just because it would be good for the person living it. We must distinguish between considerations that remove worries about a choice option and considerations that present standalone reasons in favour of that option.

4. Mere dampeners vs. standalone reasons

Spears neglects another distinction that is crucial to various philosophers who endorse the no reason claim (e.g. Frick 2014: Ch. 3). They believe that ethically relevant factors cannot be divided into reasons in favour of and against an action, or metaphorically speaking, into weights on one or the other side of the balance of reasons. This is because some factors are mere dampeners of reasons. They can justify an action by dampening the reasons against the action without, therefore, being reasons favouring the action (see Kagan 1988; Gert 2007).

To clarify, consider how proponents of the no reason claim deal with the objection that they are committed to antinatalism. I roughly sketch one way to put the objection (see Frick 2014: 130–131). The no reason claim states that there is no reason to create a person just because the person would have overall positive well-being. This suggests that the factors in a life that positively contribute to overall lifetime well-being are not reasons in favour of creating that life. By contrast, there is a reason not to create a life of negative well-being. This suggests that the factors that negatively contribute to a life’s overall well-being are reasons against creating the life. Therefore, if a child’s life would foreseeably contain some factors that will negatively contribute to the child’s well-being, then this will turn the balance of reasons against having the child. Any life would foreseeably contain some pain. Accordingly, there is an overall reason against having children, at least as far as the well-being of the children is concerned.

Proponents of the no reason claim have argued that this reasoning fails. Even if the good things in a life do not count in favour of creating it, they might still justify creating it despite the bad things. They might be mere dampeners, dampening the negative moral reasons generated by the foreseeable bad things in life without being positive weights on the balance of reasons (Frick 2014: 131–133).

The distinction between favouring reasons and mere dampeners is vital in Spears’s case. Recall that the parents are in a situation of indecision. Plausibly, they fear that the third child might suffer from various bad things, such as a severe illness. These considerations generate reasons against procreation. Accordingly, the fact that the child would be joyful might be ethically relevant because her joy could dampen the force of the reasons against procreation (stemming, for example, from worries about illness or concerns that the other children will not get enough attention anymore). If a factor can be a mere dampener, then evidence that a factor neutralizes reasons against procreation does not establish that the factor is a reason in favour of procreation. Again, because dampening of negative reasons is neither ethically irrelevant nor a reason against having the child, corresponding intuitions would presumably get classified as reasons in favour of procreation. We should expect false positives.

These false positives differ from those discussed in section 3. Here, we are considering ways in which foreseeable good things dampen the impact of negative considerations. This might be compared to dampening a shock impulse by introducing a shock absorber. By contrast, consider removing the cause of a shock in the first place, which might serve as an image for section 3. Section 3 discussed how the impact of bad things is reduced by inferring that some of these bad things are not going to happen; it was concerned with removing negative considerations.

Neither of these confounding factors could be easily excluded by suggesting that the parents’ well-being level and those of their other children remain unaffected. First, this is because both confounding factors also have to do with the child itself: ruling out the risk that the child would have a life that is overall bad or, respectively, dampening the bads that the child might suffer from, even in a good life (e.g. not getting enough attention, or suffering from a painful illness).

Second, even if the good and bad consequences that creating the child might have for other people are in balance so that everybody’s well-being remains unaffected, parents might still worry about inflicting some bad on their existing children within this balance and have deontological worries about neglecting certain duties. Some think that one has a duty not to break a promise even if breaking the promise would leave well-being unaffected (Ross 2002: 34–35). Similarly, parents might think that taking care of their children personally is no better for either party than giving them to childcare, but still believe that they, as parents, should take care of their children personally, given the special relation in which they stand towards their children. Thus, even if the well-being of their other children remains unaffected, parents might worry about neglecting a duty of personal parental care because they might have less time for their current children if they conceive another child. The prospect of having another very happy and joyful child might dampen the considerations about such neglect without being a standalone reason in favour of procreation in the sense implied by the reason claim.

5. Religious reasons

Finally, the study does not single out religious reasons for having children. We learn little about the participants in Spears’s study and whether any of them has religious beliefs driving their responses. The procreation asymmetry is usually understood as abstracting from religious considerations in favour of procreation (see e.g. Parfit 1984: 453–454).

6. Conclusion

My argument has two significant consequences. First, Spears’s experimental study is invalid. Second, by drawing attention to various confounding factors, my criticism might help prevent people from misinterpreting their intuitions about similar scenarios as clear support for the procreation symmetry. Thus, it would help prevent them from endorsing the procreation symmetry for the wrong reasons.

Is there a quick way to fix Spears’s experimental setup? One might think that one way to deal with the central criticisms in sections 3 and 4 would be to increase the set of classifications from ethical reason against, ethical reason in favour and ethical irrelevance to represent the plurality of considerations I outlined above. However, participants might have the general intuition that there is “more overall reason” for having the child without clearly seeing to which underlying consideration this judgement is to be attributed. To identify the most plausible attribution, further thought experiments might be necessary in order to isolate the fact in question from competing considerations.