The Aporetics of Longtermism: Are You Morally Obligated to Future Persons?
Longtermism is a relatively new philosophical position. It is a thesis in moral philosophy which holds that persons alive in the present have a moral duty to future persons (i.e., those who are not alive now but will be born and live in the long-term future) to act in ways that contribute to their welfare. Let us call this the Longtermist Thesis (LT).
In What We Owe the Future, moral philosopher William MacAskill claims that the case for longtermism is not particularly controversial (p. 9), and that it rests on common sense (p. 10). Pace MacAskill, however, his argument is questionable, as Kieran Setiya (2022) notes. For example, it seems that longtermism rests on dubitable utilitarian assumptions. Utilitarianism is a controversial moral theory, and as Émile Torres (2021) writes, a utilitarian-charged longtermism is objectionable.
I share the concerns of Setiya and Torres. My moral views are non-consequentialist and thus I am not a utilitarian. But MacAskill and other longtermists raise intuitively appealing points. MacAskill correctly notes that concern for future generations is common (p. 11); we invest in various long-term projects which we hope will last for generations. In fairness to longtermists, we should explore the plausibility of a non-consequentialist longtermism. It seems that longtermism, as I have articulated it in the LT, does not presuppose utilitarianism. A Kantian moral philosopher, for instance, could commit to a version of longtermism if a key metaphysical position about the nature of time is assumed. I will present this case below.
An Aporetic Triad
According to Nicholas Rescher (p. 106), philosophical deliberation often starts with an aporia. Consider the following set of propositions, each of which might seem plausible, depending on one’s philosophical commitments, and yet the set is inconsistent. If any two of the propositions are true, the third must be false.
- A moral duty is an obligation that an existing entity with moral standing (e.g., a person) has to an existing entity with moral standing (i.e., either to himself or to another entity with moral status).
- Present persons have a moral duty to future persons.
- Future persons do not exist.
Proposition (1) of the triad seems plausible. It is common among moral philosophers to hold that a moral duty is something that an agent (e.g., a person, an organization of persons) morally ought to do, i.e., an action that an agent is morally required to perform. Moral duties are thus possessed by agents who are bound by them. Furthermore, one might argue that moral duties are held by agents only in relation to beings of moral standing, whether they be human, persons of some other sort, or as some have argued, non-human animals.1placeholder In other words, if Smith has a moral duty, it is a duty to someone or something. Consider persons. Arguably, if a person has a moral duty to another person, the latter possesses the property of being owed regarding that duty. For example, if Smith is obligated to help Jones move into a new apartment because Smith has promised to help, then Jones is owed that assistance. Jones hence possesses the property of being owed Smith’s assistance. But it is quite plausible to claim that existence is a necessary condition for having properties. Non-existent things cannot possess properties. For instance, minotaurs don’t have properties, since they don’t exist.2placeholder It follows that if a person has the property of being owed a duty, then that person must exist. In other words, any right-holder who has a right to receive that which is owed to him or her must exist.
The previous paragraph contains an argument in favor of (1) in the triad. Yet the reader might suspect that I have stacked the deck in favor of (1), or perhaps that I have committed the definist fallacy, unfairly defining (1) in way that makes my position easier to defend.3placeholder To address this concern, let us examine briefly the denial of (1). Note first that by “existing” I do not mean “alive,” which will become clear in the next section. Given the debate between presentists, eternalists, and growing-block theorists, it would be question-begging at this point to define human existence in terms of being alive. Moreover, there are abiotic and other non-living things that nevertheless exist, such as rocks, beliefs, and numbers. Hence, ‘existence’ and ‘life’ are different concepts.
Now, one might press against (1) that it is possible for a person to have a moral obligation to a non-existent thing. I am not aware of any philosopher taking such a position. But let us imagine that it were so. It is hard to understand what it would mean for a person to have a moral obligation to a non-existent thing. Yet let us try to understand. Suppose that we start by recognizing the distinction between necessary non-existence and contingent non-existence; the former is such that it is metaphysically impossible; the latter is possible but not actual. Next, we can safely eliminate non-existent things like square circles and married bachelors, for these are not merely contingently non-existent, but necessarily non-existent because they are logical absurdities. It is hence metaphysically impossible for such things to be owed something on the basis of some duty that an existing person has. We will therefore not saddle our objector with the unnecessary baggage of an absurd objection.
Suppose instead that we consider contingently non-existent things such as Sherlock Holmes and unicorns. Such things do not exist, but since they are possible, they could exist. Perhaps our objector has contingent non-existents in mind, since it is possible for one to have duties to them (i.e., there is a possible world in which someone has a duty to Holmes). Now, an appeal to the plausible ought-implies-can principle seems to rule out the feasibility of there being moral duties to contingently non-existent things like our fictional master-detective. According to the ought-implies-can principle, for any action x, if a person morally ought to x, then that person can x (i.e., x is a possible action for that person). Thus, if the person cannot x, then it is not the case that the person ought to x. Yet notice that no existing person can do anything for Detective Holmes, since Holmes does not exist but instead is a fictional character. No existing person stands in any relation whatsoever to Holmes the man, since Holmes the man is not real. It follows that no existing person can do anything for Holmes, and therefore that no existing person ought to do anything for Holmes. In sum, no existing person has a moral duty to Holmes or to any other non-existent thing.
Proposition (2) from the triad above is the LT, which is accepted by longtermists and is prima facie plausible, particularly if motivated by the kinds of long-term projects MacAskill astutely discusses in Chapter 1 of What We Owe The Future. The longtermist cannot deny (2) without denying longtermism.
(3), that future persons don’t exist, is also plausible. MacAskill himself admits it (p. 10).4placeholder (3) is particularly reasonable if one is a presentist or a growing-block theorist. According to Emery, et alia, presentism holds roughly that “no objects exist in time without being present (abstract objects might exist outside of time).”5placeholder On presentism, only the present moment and its objects and events exist temporally. Thus the Battle of Salamis does not exist, nor does Themistocles, nor does the event of Lionel Messi’s retirement from soccer since these three things are not now. Schopenhauer articulates the presentist intuition: “The present alone is true and actual; it is the only time which possesses full reality, and our existence lies in it exclusively.”6placeholder
The growing-block theory holds that the past and present exist but the future does not. On this view, the ontology of time is like a growing block which possesses a vanguard edge. This edge is where the present moment resides; a constantly expanding larger section behind the vanguard is filled with past moments which are also real. The manifold continually grows, and as it does, the formerly vanguard present moments recede to the past section of the block, while the new forefront of the block becomes the new present moment.
Assuming either presentism or the growing-block theory, (3) is hard to deny. (1) and (2) are also reasonable. Yet these propositions are collectively inconsistent. Which one ought to be rejected? Arguably, (1) is the strongest of the three. This leaves a conflict between (2) and (3). The longtermist cannot reject (2), since this proposition is the longtermist’s cynosure. Hence, the longtermist seems pressed to deny (3), namely that future persons do not exist. In the next section, I will suggest a coherent way to do this.
Longtermism and Eternalism: An Argument
How can a longtermist provide a principled rejection of (3)? Consider eternalism, which is roughly the view that past, present, and the future things exist. On eternalism, past, present, and future events and entities have being. Past and future things do not exist here in time (i.e., now), but exist elsewhere in time, either at past points or at future points. An eternalist might compare time with space: Greece exists, but is not spatially here in Florida, where this paper is being written. Similarly, the Battle of Salamis and the summer of 2025 exist, but not temporally here in October 2022.7placeholder If the longtermist assumes eternalism, he has a reason to deny (3) and instead affirm that future persons exist. This dialectical move raises the following argument.
- If eternalism is true, then future persons exist.
- Eternalism is true.
- Therefore, future persons exist.
- If future persons exist, then we have moral obligations to them with respect to their welfare.
- Therefore, we have moral obligations to future persons with respect to their welfare (i.e., longtermism is true.)
In short, if the longtermist is prepared to provide an adequate defense of eternalism, then some version of longtermism seems acceptable.
The eternalist might hold that eternalism is a defensible position because, say, it provides a better explanation for why we can refer to or make true claims about past and future objects or events, such as Socrates, Socrates’ trial in Athens, or the President of the United States in 2028. According to the eternalist, presentism has trouble grounding such references and claims, but eternalism provides a reasonable account for the grounding of true propositions about the past and future. Similarly, the growing-block theorist has trouble grounding propositions about the future. Since eternalism can offer such metaphysical grounds, eternalism is preferable to presentism and to the growing-block theory.
In contrast, the presentist and the growing-block theorist claim to have a reasonable case against eternalism. For instance, they might hold that eternalism presupposes the B-Theory of time, and they might deny that theory. Or, they might hold that eternalism cannot explain the phenomenology of time with respect to its apparent forward-flowing nature, but that presentism (or growing-block theory) can explain this experience.8placeholder
Which Version of Eternalist Longtermism?
Assume arguendo that eternalism is a defensible position. A longtermist can hence support his case by appealing to the metaphysics of eternalism. Yet, as I noted in the introduction, utilitarian longtermism is problematic. There are many arguments against utilitarianism. Consider two. I’ll call the first the predictability problem and the second the harm problem.
According to a common articulation of utilitarianism, for any morally significant act, we should act in such a way as to bring about the greatest cumulative benefit for the greatest possible number of people.9placeholder This is the utilitarian thesis (UT). The predictability problem (PP) indicates that no human being is cognitively equipped to know or even to have a sufficient number of reasonable beliefs about the long-term consequences of his actions. This fact about the limits of our epistemic capacity counts against UT. The PP becomes vastly more difficult with respect to longtermism. Suppose that, in the long-term future of humanity, 750 trillion human beings will be born and live for an average of 100 years each. There is no reliable method for us to know what precise consequences our decisions now will have regarding these 750 trillion future humans. It would be practically impossible for us to act on the basis of a dependable utility calculus in such circumstances. But a moral theory that requires us to do the impossible is, for that reason, a flawed theory. Thus, utilitarian longtermism is defective.
The harm problem (HP) has been articulated in various ways.10placeholder The basic idea is that, according to UT, one is obligated to do whatever is necessary to bring about the greatest cumulative benefit for the greatest possible number of people. Yet suppose that intentionally harming one person, or a minority of persons, is required to benefit a majority of persons. Utilitarianism would therefore seem to hold that harming the few to benefit the many is not only morally justified, but also obligatory. This conclusion is likely to violate one’s moral intuitions; instead, one might argue, it is morally wrong to injure persons intentionally and/or to violate their moral rights, regardless of the anticipated beneficial consequences.
Suppose that you see your physician for a routine checkup. You are completely healthy. Your physician’s office is in a building adjacent to the local hospital, where there are five sick patients. Each of them would recover their health if they were to receive one organ. The first person needs a new heart; the second, a kidney; the third, a pancreas; the fourth, a liver; the fifth, a lung. Your organs are functioning properly. The hospital staff members accept utilitarianism and hence conclude that they should, without your consent, kill you and distribute your organs to the five sick people because the health of five persons outweighs the health of one person. (Suppose that this is the only way to save the five.) As consequentialists, they hold that the end of saving five justifies the means of harming one. Most people, I suspect, would agree that the hospital staff would be morally prohibited from acting in such a manner, which poses a problem for utilitarianism.
Here is another example of the HP. In the Marvel films Avengers: Infinity War and Avengers: Endgame the villain, Thanos, desires to produce a world in which everyone alive has sufficient resources to thrive. Considered by itself, this is a good end. Thanos, however, reasons that if the world’s population were reduced to half its current number, then his end would be achieved. Hence, Thanos decides to kill roughly half (perhaps slightly less) the population of rational beings in the Marvel Universe, believing that the achievement of his end would justify these gruesome means, since the survivors of this Thanosian culling would inherit the universe’s resources. Note that Thanos is the villain in this film, not the hero. This point indicates that most people’s moral intuitions–including the story’s writers–oppose Thanos’ version of utilitarianism.
Torres (2021) underscores a similar yet much more difficult problem with utilitarian longtermism. Thanos is concerned that everyone alive has sufficient resources. Longtermists are concerned about the welfare of everyone who will be alive in the distant future. But the number of future persons might be astronomical, immensely outweighing the number of humans currently living. Torres presents the problem this way:
“Yet the implications of longtermism are far more worrisome. If our top four priorities are to avoid an existential catastrophe – i.e., to fulfil ‘our potential’ – then what’s not on the table for making this happen? Consider Thomas Nagel’s comment about how the notion of what we might call the ‘greater good’ has been used to ‘justify’ certain atrocities (e.g., during war). If the ends ‘justify’ the means, he argues, and the ends are thought to be sufficiently large (e.g., national security), then this ‘can be brought to bear to ease the consciences of those responsible for a certain number of charred babies’. Now imagine what might be ‘justified’ if the ‘greater good’ isn’t national security but the cosmic potential of Earth-originating intelligent life over the coming trillions of years? During the Second World War, 40 million civilians perished, but compare this number to the 1054 or more people (in Bostrom’s estimate) who could come to exist if we can avoid an existential catastrophe. What shouldn’t we do to ‘protect’ and ‘preserve’ this potential? To ensure that these unborn people come to exist? What means can’t be ‘justified’ by this cosmically significant moral end?”
These and other objections to utilitarianism suggest that utilitarian longtermism is objectionable. Longtermism sans consequentialism is another matter. A deontologist or another kind of non-consequentialist might incline to longtermism. For example, a Kantian who is also an eternalist might hold that, since future persons exist, we ought to treat them according to Kant’s categorical imperative (CI), given that the CI is applicable to all existing persons. According to the so-called humanity formulation (i.e., personhood formulation) of the CI, one ought to act in such a way as to treat humanity, whether in one’s own person or in the person of any other, never merely as a means, but always at the same time as an end. (Kant, p. 96) A Kantian eternalist might thus argue that future persons ought to be treated as ends, just as present persons should be so treated.
Alternatively, a deontologist who follows W. D. Ross might hold that we have a prima facie duty of beneficence to future persons. These are not obligations of the utilitarian sort which would require us to do whatever is required to maximize the happiness or welfare of as many persons as possible, which, assuming that the number of future persons is vastly greater than that of present persons, might necessitate that we act in ways that harm present persons for the sake of benefitting future persons. Rather, Ross’ duties are non-consequentialist obligations which can be balanced against other relevant duties. For instance, a Rossian duty of beneficence to future persons should be weighed against our obligations of beneficence, non-maleficence, fidelity, etc. to present persons, as well as a duty of self-improvement that we owe to ourselves, and thus the welfare of present persons ought not be sacrificed to maximize the flourishing of future persons even if the population of the latter is greater than that of the former. Similarly, a Kantian might argue that present persons, including oneself, ought not be treated as mere means for the sake of benefiting future persons. We can call these positions deontological longtermism.
Objection and Reply
One might object against deontological longtermism that (1) should be modified as follows: a moral duty is an obligation that a living entity with moral standing has to a living entity with moral standing. But since future persons are not alive, we have no duties to them. In response, the deontological longtermist can argue that living persons have obligations to nonliving persons and therefore that the modification fails. For example, Smith makes a promise to his father to complete a task for him after he dies. Following the father’s death, Smith retains that duty, despite the fact that his father is not alive. Jones receives an inheritance from his recently deceased wealthy grandmother. Jones promised her that he would invest the money wisely in the years after her death. Jones hence continues to be obligated to her to do so. Green is trying to become pregnant. She prepares her life and finances for the pregnancy, believing plausibly that she has a duty to her future child to provide for the child’s welfare. Now, the deontological longtermist might argue, since we have duties to deceased persons and to persons not yet born, (1) should not be modified and therefore deontological longtermism holds.
Why does this topic matter? If longtermism is correct, every human person alive now is morally obligated to contribute to the welfare of future persons with respect to their physical environment, their intellectual resources, and the moral values that their societies will accept. For example, environmental sustainability is the idea that those alive now are morally responsible for conserving the earth’s environment in the present and for the future. Longtermism makes this idea more precise by holding that we are obligated to future persons to protect the environment for them and their flourishing. Furthermore, as MacAskill argues, it is plausible to hold that we have a duty to future persons to “lock-in” moral values that will benefit them.11placeholder These are formidable tasks and it is unclear how best to accomplish them. Yet if longtermism fails or if future persons do not exist, then arguably we have no duties whatsoever to them, even if we are still morally responsible for the environment and the moral health of our civilization. Moreover, the concerns raised by Torres regarding utilitarian longtermism are matters of existential significance. The idea of a group of politicians who are willing to break millions of current eggs for the sake of a larger future omelette is a dreadful idea indeed. Much of practical significance thus rides on the truth value of longtermism. For this reason, we ought seriously to examine longtermism as a theory, posing critical questions about it, raising objections to it, and considering its consequences for human life.
Dowden, Bradley. Fallacies. In Internet Encyclopedia of Philosophy. https://iep.utm.edu/fallacy/#Definist (Accessed October 13, 2022).
Emery, Nina, Markosian, Ned, and Sullivan, Meghan. “Time.” In The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/time/ (Accessed August 12, 2022).
Kant, Immanuel. 1964. Groundwork of the Metaphysic of Morals. New York: Harper Torchbooks.
MacAskill, William. 2022. What We Owe the Future. New York: Basic Books.
Miller, Kristie. 2013. “Presentism, Eternalism, and the Growing Block,” in A Companion to the Philosophy of Time, eds. Adrian Bardon and Heather Dyke. Oxford: John Wiley & Sons, Inc., pp. 345-64.
Nussbaum, Martha. 2019. The Cosmopolitan Tradition: A Noble But Flawed Ideal. Cambridge, Massachusetts: Harvard University Press.
Rescher, Nicholas. 2009. Aporetics: Rational Deliberation in the Face of Inconsistency. Pittsburgh: University of Pittsburgh Press.
Schopenhauer, Arthur. 1851. Counsels and Maxims. https://en.wikisource.org/wiki/Counsels_and_Maxims (Accessed October 1, 2022.)
Setiya, Kieran. 2022. The New Moral Mathematics. In Boston Review. Accessed on September 23, 2022 at https://bostonreview.net/articles/the-new-moral-mathematics/
Torres, Émile. 2021. “Against longtermism.” In Aeon Magazine. Accessed on September 23, 2022 at https://aeon.co/essays/why-longtermism-is-the-worlds-most-dangerous-secular-credo
For instance, see the final chapter of Nussbaum (2019) in which she argues that at least some non-human animals have capacities sufficient for moral status, though she does not argue that non-human animals have moral rights, precisely speaking.
The concept ‘minotaur’ is the idea of a man-bull creature with horns, and so it is intelligible to say that the concept ‘minotaur’ is essentially linked to the concept ‘being horned,’ but there is no actual minotaur that possesses the property of being horned, nor does the concept ‘minotaur’ possess the property of being horned, since concepts do not have physical parts such as horns.
See Dowden for a discussion of the Definist Fallacy.
This point suggests that MacAskill is committed to (2) and (3) and thus ought to deny (1). MacAskill appears to be a presentist longtermist, though as I will note, other claims he makes seem to commit him to eternalism, thereby suggesting a metaphysical inconsistency for MacAskill’s version of longtermism.
See Section Six of Time, 2020. To this definition, we might also include the claim that the present moment moves or changes; that is, time t1 is the present and then t2 is the present, and so forth. Kristie Miller calls this the “Dynamical Thesis” (see Miller, p. 346). Eternalists deny presentism and the Dynamical Thesis.
Schopenhauer makes this point in Counsels and Maxims, Chapter 2, Section 5. I do not intend to argue here that he was a presentist, though his language in this place and in others of his oeuvre is that of presentism.
MacAskill (p. 10) seems to assume this eternalist principle. Speaking like an eternalist, he writes: “Distance in time is like distance in space. People matter even if they live thousands of miles away. Likewise, they matter even if they live thousands of years hence. In both cases, it’s easy to mistake distance for unreality …” It is therefore puzzling why he says on the same page that future persons don’t exist.
There is more to say about the debate between eternalists on one hand, and presentists and growing-block theorists on the other, but I cannot address the debate further here.
Utilitarianism is a version of consequentialism, which is a category of normative theories which roughly hold that the consequences of an act make the act that produced them (whatever that act is) morally right. In other words, the end (i.e., result, outcome, consequence) justifies whatever means produced it.
For instance, see the trolley problem, the healthy patient problem, and what I sometimes refer to as the Thanos problem. When examining utilitarianism with my moral philosophy students, I often refer to the Marvel character Thanos, since many of my students have seen the relevant Marvel films in which Thanos pursues his villainous goal on the basis of utilitarian reasoning. I’ll address briefly the healthy patient problem and the Thanos problem below.
See Chapter 4 of What We Owe the Future.