Introduction
Derek Parfit’s ideas about population ethics have continued to draw attention for decades, also recently.Footnote 1 Much of the discussion has been guided by an ethical conundrum that Parfit presented in his 1984 Reasons and Persons: the nonidentity problem.Footnote 2 , Footnote 3 An analysis of this conundrum from a new angle reveals that Parfit was, perhaps surprisingly, just one factual and one normative premise short of antinatalism, the view that reproduction is wrong. At the heart of his reflections was an admirably progressive view on how our choices can benefit or harm persons who do not exist yet. Had he gone on to apply this view to future lives that are not worth living, he could have reached the conclusion that no one should have children. At the very least, he would have been forced to admit, I believe, that universal childlessness and voluntary human extinction, while unthinkable topics to many, rightfully belong to ethical discussions on future generations.
My analysis proceeds in three main stages. I will begin, in the first main section, by describing Parfit’s problem and its corollaries by using two distinctions that he draws but does not fully utilize. I will then proceed, in the next main section, to give form to my own argument in the light of two further distinctions. I will conclude, in the final main section, by asking and answering a residual question about my argument and assessing what my findings mean both to Parfit scholarship (not much) and antinatalism (potentially a lot).
Two things need to be made clear at the outset. This is not an exercise in Parfit exegesis. I do not have the expertise necessary for that. This is not a normative defense of antinatalism, either. Those are replete in the extant literature.Footnote 4 , Footnote 5 , Footnote 6 , Footnote 7 , Footnote 8 , Footnote 9 , Footnote 10 , Footnote 11 , Footnote 12 , Footnote 13 , Footnote 14 , Footnote 15 , Footnote 16 , Footnote 17 , Footnote 18 , Footnote 19 , Footnote 20 This is a conceptual exploration showing how close to the surface antinatalist premises have been for at least four decades, even in seminal works such as Parfit’s.
The nonidentity problem and its conceptual context
In his introduction of the nonidentity problem, Parfit acknowledged that lives can be of lower or higher quality and hence less or more worth living. He noted that exact comparisons are often impossible but trusted that in clear cases we can intuitively tell the difference. His own example was the following:
The 14-Year-Old Girl. This girl chooses to have a child. Because she is so young, she gives her child a bad start in life. Though this will have bad effects throughout this child’s life, his life will, predictably, be worth living. If this girl had waited for several years, she would have had a different child, to whom she would have given a better start in life.Footnote 21
Views on the aptness of this example and its vocabulary vary, but it is the case Parfit presented and employed as the starting point of his investigation. I will, for brevity, call the first (actual) child Early and the second (possible) one Late.
Early and Late would both have, by definition, lives worth living. The way Parfit describes Early’s situation, however, implies both that his life is only just worth living and that the possibility of lives not worth living is also real.
The second distinction is implicit in the case: causing or not causing a person to exist. Together with judging the life to be worth or not worth living, this gives rise to the double dichotomy presented schematically in Figure 1.
Figure 1. Dimensions of causing or not causing persons to exist.
Explanations of the main points of Figure 1 clarify Parfit’s thinking and its connections with other philosophers’ views. These comments pave the way to my own argument.
We benefit a person by causing the person to exist with a life worth living
In introducing the nonidentity problem, Parfit concentrated on benefiting persons. His conceptual background assumption was that we can benefit a future person by causing the person to exist and have a life worth living. The person need not be our own child. As long as our choices are a necessary condition of a person’s life worth living, we, among others, benefit the person. This point will be crucial to my own argument once I have proceeded to formulating it.
Some actions seem to be wrong although they are not wrong for anyone
The root of Parfit’s conundrum is that, depending on our choices, we can cause different persons to exist, as in the case of The 14-Year-Old Girl. If both Early and Late would have lives worth living, as he assumes, they would both benefit from coming into existence. The Girl’s choice would not be wrong for Early. Yet Parfit seems to believe that The Girl’s act would be wrong. Hence the conundrum. There seem to be actions that are wrong, although they are not wrong for anyone (do not wrong anyone).
Julian Savulescu abandoned Parfit’s comparative caution in his 2001 principle of procreative beneficence and argued that.
couples (or single reproducers) should select the child, of the possible children they could have, who is expected to have the best life, or at least as good a life as the others, based on the relevant, available information.Footnote 22
Savulescu’s focus was on genetic selection, but, by the ethos of his principle, The Girl’s choice to have Early is wrong if having Late is a tangible possibility. This straightforwardly utilitarian solution sidesteps, however, Parfit’s reliance on a more person-centered approach, and has its own potential weaknesses.Footnote 23
We do not harm a person by not causing the person to exist with a life worth living
We do not harm persons by omitting to provide them with lives worth living. Literally speaking, this is self-evident. Before a person comes into existence, there is no one there to be harmed.Footnote 24 Parfit does not use this conceptualization, but the same conclusion follows from the logic of the nonidentity problem. If The Girl’s choice is not wrong for anyone, as stated, it is not wrong for either Early or Late. No one is harmed or wronged by continued non-existence.
We harm a person by causing the person to exist with a life not worth living
We do harm persons by causing them to have lives that are not worth living. Parfit’s concentration on benefits leaves this point unexplicated, but it is implicit in his premises. Causation plus positive or negative outcome equals positive or negative verdict. If Early’s life would have been not worth living, he would have been harmed.
The estimation that a life is not worth living does not allow further comparisons
Life can be more or less worth living, but presumably not more or less not worth living. The latter seems to be a threshold evaluation: if a person’s life does not have value, the discussion ends there. In purely utilitarian assessments—like Savulescu’s—further comparisons make sense; not in Parfit’s framework.
A possible shortcut to antinatalism: no lives are worth living
David Benatar provided in 1997 a shortcut to antinatalism by maintaining that all human lives are bad and hence not worth living.Footnote 25 In his own 2006 words, “even the best lives are very bad.”Footnote 26 In Parfit’s framework, this is neither here nor there because he believed that both kinds of existence, good and bad, are possible and postulated that both Early and Late evade Benatar’s judgment. Overturning the postulation would make Parfit an antinatalist, but this is not my point here.
We do not benefit a person by not causing the person to exist with a life not worth living
If nonexistence persists, there is no change for the better and hence no benefit. This is in line with the earlier verdict that neither Early nor Late would be harmed by not causing them to have lives worth living. Similarly, neither Early nor Late would be benefited by not causing them to have lives not worth living.
We fail to benefit a person by not causing the person to exist with a life worth living but that might be of no consequence
Having said this, while we do not harm persons by letting them remain nonexistent, we might fail to benefit them. Seana Shiffrin in 1999 used this distinction to show that it may be wrong to have children in the first place. While preventing harm, according to her, could justify making life decisions for them without their consent, bestowing a benefit does not. For Shiffrin, this was primarily evidence that wrongful-life cases should succeed even if the claimant has a life worth living.Footnote 27 An antinatalist conclusion is also possible, but she has consistently shunned away from that.
For Parfit, the question does not arise. If having Early is not wrong for anyone, it is not wrong for either him or Late, and withholding the benefit cannot be the decisive wrong-doing factor.
Persons and needs
Talking about the consent of “a person” in the bottom boxes of Figure 1 is, of course, conceptually strained. If the entity does not exist now and will not exist in the future, who or what are we talking about? Nothing, it seems.
The situation could be clarified by describing it in terms of needs. Once in existence, a person with a life well worth living would have a limited number of frustrated major needs, while a person with a life barely worth living would have a great many of them, and a person with a life not worth living an overwhelming amount. In turn, a being not in existence does not currently have any needs. It cannot because it does not exist. There is no one there to have needs.Footnote 28
If we add to this the simple moral rule that contributing to major need frustration is, other things being equal, wrong, the results are clear.Footnote 29 It might be permissible to cause a person like Late to exist with a life well worth living. It would be prima facie wrong to cause a person like Early to exist with a life barely worth living. It would be conclusively wrong to cause a person to exist with a life not worth living. And not causing persons to exist is never wrong.
Assuming this axiology and rule would make the nonidentity problem go away. Having Early is wrong—and wrong for him—because it causes him major frustration. This was not Parfit’s interpretation, albeit that his conundrum seems to imply either something like this or a leaning towards a Savulescu-type principle of procreative beneficence.
Axiological asymmetry
Considerations of not harming and not benefiting persons by not causing them to exist push to the fore the idea of axiological asymmetry. This was formulated early on by Jan NarvesonFootnote 30 and made widely known by David Benatar.Footnote 31 We seem to have a duty not to bring into existence individuals with bad lives, but no corresponding duty to bring into existence individuals with good lives.
Fumitake Yoshizawa has analyzed the relationship between Parfit’s nonidentity problem and Benatar’s arguments for antinatalism,Footnote 32 and I will not return to it in detail here. Suffice it to say that Parfit’s reflections on The 14-Year-Old Girl are not entirely free of asymmetry thinking. He concentrates on Early at the expense of Late, suggesting that the “person” whom The Girl could harm by a life not worth living (almost the case with Early) is somehow more real than the “person” whom The Girl would only fail to benefit (both Late and the postulated Early).
Taking lives not worth living seriously
Many topics of my comments in the preceding section are peripheral to Parfit’s ruminations because he focuses on the top left box of Figure 1—benefiting persons by causing them to exist. They are all needed for my own narrative, though, as building blocks of my argument. To formulate it, I concentrate on the top right box of Figure 1 with the help of two more distinctions.
We can cause a person to exist more or less immediately, as in the case of The 14-Year-Old Girl; or we can cause a person to exist sometime in the future, maybe long after we ourselves are gone. And the lives these persons eventually have can be not worth living unforeseeably or foreseeably. The alternative outcomes of the four emerging scenarios are depicted in Figure 2.
Figure 2. Dimensions of causing persons to exist with lives not worth living.
The value-and-norm verdict remains the same in the top and bottom boxes on the left and on the right. This is due to Parfit’s view on causation—and crucial to my argument. The verdicts of the boxes on the left and on the right differ in a way that is not explicit in Parfit but probably corresponds with popular intuitions.
Causing immediate unforeseeable harm may be excusable
We may harm persons in the near future unintentionally, believing that we benefit them. Consider this case:
The 14-Year-Old Heiress. This girl chooses to have a child. She is very mature, intellectually and emotionally, and has good reason to believe that she gives the child an exceptionally good start in life. If this girl had waited several years, her financial prospects would have deteriorated. She could then have had a different child, to whom she would have, for all she knows, given a much worse start in life. The child she has now meets, however, an unforeseeable adversity that renders his life not worth living.
The Girl harms her child by causing him to exist with a life not worth living, but as long as causing a person to exist with a life worth living benefits the person and the adversity is genuinely unforeseeable, blaming the girl would seem excessive. She did her best.
Causing immediate foreseeable harm is wrong
We may also harm persons knowingly or intentionally, thinking that we have good external reasons to do so. Consider this:
The 24-Year-Old Woman. This woman chooses to have a child. Because she has a latent inheritable disease, she gives her child a life not worth living. She knows this but believes that the Seven-Headed Cannelloni Monster has ordered her to have the child.
Unlike The 14-Year-Old Heiress, this woman does not have probability-based child-related reasons for her choice. By causing her child to exist, she harms him without any universally recognizable mitigating factors. If the Seven-Headed Cannelloni Monster is widely worshipped in the woman’s community, her action may be understandable, but it is still inexcusably wrong.
Causing distant unforeseeable harm may be excusable
Taking Parfit’s definition of causation strictly and literally, we harm innumerable persons in the distant future just by going about our daily lives. Any of our actions can eventually become a necessary condition of a person’s life not worth living or of other harms persons may encounter. The distinction between foreseeable and unforeseeable consequences moderates this rule without challenging its essence.
Consider this case:
The Woman with the Bottle of Olive Oil. This woman chooses to buy a bottle of olive oil. She takes every precaution in storing it and recycling it. A hundred years later, the bottle is broken in a forest and a person steps into the splinters. Had the woman not bought the bottle when she did, it would not have ended up in that forest for that person to step into.
Like in the case of The 14-Year-Old Heiress, it would be excessive to blame the woman for the future mishap. We may accept Parfit’s idea of distant causation in principle, but some constraints in assigning moral responsibility and blame are needed in practice. If the harm is reasonably unforeseeable—nothing we choose to do can insure us against causing it or something similar—we can probably be exonerated without offending any major ethical sensitivities.
Causing distant foreseeable harm is wrong
The excuses evaporate when the harm in the future is foreseeable, like in the case of The 24-Year-Old Woman. The grounds on which we make our decisions need not be relevant when we know, or should know, the harm that will befall an as-of-yet unidentifiable future person. Consider the following:
Responsible Reproducers. These persons choose to have children. They take every precaution to ensure that their children’s lives are good, and they are in good positions to do this. Partly due to these persons’ efforts, their children’s lives are well worth living. In the distant future, however, each created family line produces at least one individual with a life not worth living. By not having children, these persons would have prevented the lives not worth living from coming to existence.Footnote 33
I mentioned in the first main section that Parfit seems to be one factual and one normative premise short of an antinatalist outlook. The case of Responsible ReproducersFootnote 34 introduces the first, factual premise. By having children, parents make themselves necessary conditions of lives not worth living sometime in the future. Their own immediate progeny may have good-quality lives, but somewhere along the line someone will draw the short straw. The Responsible Reproducers harm those persons—and they should know it.Footnote 35
Is it inevitable that every family line will end up producing lives not worth living? Given enough time, I think that it is. Be that as it may, assuming this factual premise would take Parfit a step closer to the antinatalist view that reproduction is always wrong.
Harm could be justified by condoning sacrifices in the name of greater benefits
When I state in the bottom right corner of Figure 2 that there are no mitigating factors to inflicting the harm, I mean future-progeny-related mitigating factors. There can be others, both rule-based and outcome-based.
We can exclude rule-based reasons like the orders of the Seven-Headed Cannelloni Monster because Parfit does not recognize divinities or community customs in his account. This leaves outcome-based reasons.
The Responsible Reproducers can estimate that while their choice would allow the eventual emergence of lives not worth living, it would also bring benefits to many individuals in their family line. They could argue that the overwhelming amount of future benefits outweighs the harm inflicted on a few individuals with lives not worth living.Footnote 36
The Responsible Reproducers do not need to be through-and-through utilitarians to make this claim. They can appeal to moderate interpretations of risk-taking.Footnote 37 They would, however, in doing so have to accept that the one or the few can be sacrificed on the altar of the greater good of the many. Especially critics of utilitarianism have found this principle difficult to accept.Footnote 38 , Footnote 39 , Footnote 40
This brings us to the normative premise separating Parfit from antinatalism. It is that the deliberate sacrifice of the one or the few to benefit the many should not be routinely condoned. Admitting that every family line will eventually cause a life not worth living to exist and holding that sacrifices like this should not be made would make Parfit an antinatalist. This, in sum, is my conceptual argument.
A residual concern and the nature, limits, and meaning of my result
One detail in Parfit’s introduction of the nonidentity problem might trip my analysis so far. And although—as I will show—it does not, the argument for his potential antinatalism has its limits. I will conclude by considering these points and the meaning of my observations to Parfit scholarship and to investigations into antinatalism.
Parfit is not adamant about his distant-causation thesis
While Parfit favors the idea that we can benefit persons by distant causation in the remote future, he also allows its denial if the concept of a benefit is understood “in its ordinary sense.”Footnote 41 This would mean excluding far-away consequences as too abstract or too impracticable to be included in moral considerations.
The same logic could be applicable to harming. If it is, my Responsible Reproducers would not have to worry about the remote future lives not worth living as long as they do what is in their power to secure their children and perhaps their children’s children the best living conditions that they sensibly can.Footnote 42
This is not, however, Parfit’s interpretation. He believed that “we ought for moral purposes to extend our use of ‘benefit” to cover events in the future.Footnote 43 Parfit’s belief does not, of course, make the reading universally binding. But it does make it his reading, and since I am only showing how close to antinatalism he came, my point remains intact.
How to reach the unthinkable conclusion
Let me sum up the premises and inferences needed to reach the antinatalist conclusion that many see unthinkable. The syllogism in its complete form looks like this (P = premise; C = conclusion):
The last conclusion, C4, is the arguably unthinkable one.
The limits of the argument
Some of the premises of the argument are well-nigh unassailable, while others can be challenged.
The first two premises, P1 and P2, can, I expect, be universally accepted. We should not do what is wrong, and it is, unless shown otherwise, wrong to harm persons. The third premise, P3, is implicit in Parfit’s introduction of the nonidentity problem, if I have understood him correctly. These bring Parfit as far as the second conclusion, C2, that we should not, unless shown otherwise, cause persons to exist with a life not worth living.
By having children, we can produce lives that are worth living or not worth living. P4 states Parfit’s missing factual premise, that sooner or later every reproductive line will create one or more lives not worth living, regardless of the precautions taken by the initial reproducers. Parents could, of course, convince their children or grandchildren to discontinue the procreative tradition, but it is doubtful that they have the inclination or the means to do this.Footnote 44 The only other alternative I can see is that the world becomes so much better that lives not worth living do not exist anymore. It is not a very realistic alternative.Footnote 45
These considerations, if accepted, take us tentatively to the antinatalist position. Unless shown otherwise, childbearing should be stopped because it leads in the future to the emergence of lives not worth living. P5 states that the only way to “show otherwise” involves the sacrifice of the one or the few for the benefit of the many. P6 rejects such sacrifices. Both 5 and 6 can be challenged, and Parfit’s views on them are not entirely clear, but my thesis here holds its force whatever they were. Parfit was, in Reasons and Persons, one factual and one normative premise away from the antinatalist conclusion that we should not have children.
The meaning of the result
For Parfit scholarship, my result is no breakthrough. I have shown that one fragment of his work in Reasons and Persons would, if combined with two premises that he did not make, allow an “unthinkable” antinatalist interpretation. Insofar as his underlying ethos is outcome-based, this is to be expected. Factual premises have a tendency to change consequentialist conclusions. And adding normative assumptions alters any ethical argument.
Perhaps the only surprise here is the relatively non-controversial nature of the additional premises. It is not particularly outrageous to claim that sufficient iteration in childbearing will, unless miracles happen, eventually produce bad lives. And rejecting the sacrifice of minorities to benefit the rest is hardly a sensational stand.
The real breakthrough, however, concerns the potential streamlining of antinatalist argumentation. Very few people buy the view that all lives are bad; and very few people understand what axiological asymmetry and lack of prenatal consent mean.Footnote 46 , Footnote 47 , Footnote 48 , Footnote 49 The ideas are unintuitive or opaque. Not so with the modified Parfitian argument. Lives not worth living are bad; we should not cause them to exist; we cause them to exist in the future by having children; therefore, having children is bad. Every step on the way can be challenged, but at least the way is clearly marked and the destination visible.
Why did Parfit’s contemporaries not see and embrace the route, then? The answer may lie in the destination. If no one has children, humankind will go extinct. This corollary made even consequentialist moral philosophers wary of views that stressed the reduction of suffering at the expense of the promotion of happiness.Footnote 50 These negative utilitarian, and related, views are only now gradually re-emerging in mainstream moral thinking. Whatever their fate, they do, I believe, have their niche in the marketplace of ideas concerning population ethics. They deserve to be discussed.
Open access
