Valuing the “Afterlife”

Abstract

To what extent do we value future generations? It may seem from our behavior that we don’t value future generations much at all, at least in relation to how much we value present generations. However, in his book Death and the Afterlife, Samuel Scheffler argues that we value the future even more than we value the present, even though this is not immediately apparent to us. If Scheffler’s argument is sound, then it has important ramifications: It would give us a strong motivation to put more energy into abating environmental crises like climate change, and it supports at least a limited form of ethical longtermism. However, in this paper, I show that Scheffler’s argument is fallacious. Scheffler claims that we do not regard the fact that we in the present generation will all die relatively soon as a catastrophe, but we do regard the non-existence of future generations as a catastrophe. But the particular scenario used by Scheffler to illustrate this point—the plot of the book The Children of Men—is one in which both the present generation will perish and there will be no future generations, and it is this conjunction that is catastrophic, thus giving no information about which is worse. I suggest other ways to compare our valuations of present and future generations, and recommend that philosophers who are interested in the moral psychology of how we value future generations ought to engage with social science, as it is an empirical issue.

1 Introduction

In an important and much-celebrated book, Scheffler (2013) argues that we value the lives of the generations of future humans who will succeed us more than we value our own lives and the lives of others today. As he puts it (2013, p. 45): “The coming into existence of people we do not know and love matters more to us than our own survival and the survival of the people we do know and love.” (Scheffler also discusses this argument in his 2018, Ch. 2, but for my purposes here I focus primarily on his 2013). However, as I argue in this paper, Scheffler’s argument is based in fallacious reasoning.

Scheffler’s aim (in his 2013) is not to argue from first principles that future generations of humans are more valuable than we ourselves are. Rather, he undertakes the task of convincing us that we already do value future humans to a significant extent, but for various reasons that fact is obscured from view. Since value in a more objective sense is not Scheffler’s concern in his (2013), in what follows, when I say that a state has a certain value, I intend it only in the subjective sense that the state is valuable to us. Ultimately, answering whether future generations are more or less valuable than the present one requires substantial philosophical theorizing that goes well beyond what I discuss in this paper. Although I do make some brief speculations in that regard, my primary aim is to assess the structure of Scheffler’s argument and to show that the considerations that he brings to bear in support of his conclusion do not demonstrate that it is true. More positively, this paper clarifies ways to compare our valuings of present and future generations, and shows the potential benefits of philosophers engaging with social scientists.

This paper should be viewed in light of other philosophical projects regarding future generations. For instance, in a review essay of Scheffler’s Death and the Afterlife, Davidson (2015) points out how environmental philosophers have long discussed the need to care for future generations. Furthermore, longtermists argue that future people matter just as much as present people, and thus much of our ethical attention should be turned to increasing the chances of a good long-term future (see Ord 2020; Cargill 2021; MacAskill 2022). Although Scheffler’s books preceded the growth of the longtermist movement in its present incarnation, they can be seen as supporting it, in a motivational sense: it might seem that we don’t value future people very much, given some of our behaviors (e.g., humanity’s present levels of emitting greenhouse gases, causing climate change), but if Scheffler is correct, then people really do view a collapse of future generations as catastrophic (either insofar as future generations might not come to exist, or that future generations might be stuck with low levels of wellbeing), and thus should dedicate much more of our energy to benefitting future generations.

Wolf (2013) argues, in her response to Scheffler, that if we were to find out that there will be no more future generations, we could still find value in our present existence, and that would motivate us to live worthwhile lives. (Also see Rulli 2021 for similar arguments.) While my argument here accords with Wolf’s point, this paper gives even more reason to believe that we aren’t dependent on future generations for our own lives to be meaningful or happy.

One final clarificatory point: Scheffler (2013; see pp. 17, 18) uses the first-person plural terms “we” and “our”, and explains (2018, pp. 3, 4) that he is “trying to characterize patterns of belief that I hope readers will recognize as constituting familiar (even if not universal) tendencies within contemporary thought and discourse.” Despite have some qualms about this usage, in this paper I will follow Scheffler by using the first-person plural in this way.

2 Scheffler on the “Afterlife”

Scheffler uses the term afterlife to denote generations that come after one dies, and not the idea that one somehow, in a literal sense, lives on after the death of their body. Scheffler’s main example in arguing that we have deep concern for the afterlife comes from the book The Children of Men by James (1992), and its subsequent film adaptation, directed by Cuarón (2006).¹ The plot involves a doomsday scenario: no one (just about?) is able to reproduce, and thus humanity will soon die off. How would we react to such a scenario? The people in the universe of Children of Men take this as catastrophic, and lose interest in many of life’s normal activities. Scheffler uses Children of Men to show that we view the absence of future generations of humans as catastrophic.

I should note that in James’s original book (1992, p. 45), part of the origin of the breakdown of society is that people, knowing that there will be no younger generation to look after them when they are old, hoard goods for their own long-term benefit. As might be expected, this hoarding does not in the end work out well for the society. Thus the catastrophe in James’s The Children of Men is at least partly due to straightforwardly egoistic behavior. And (spoiler alert!) Cuarón’s (2006) film ends with the birth of a child, but factional fighting nevertheless continues, suggesting that the lack of future generations is not the core reason for the breakdown of the social fabric in the movie’s universe. So there is reason to doubt Scheffler’s interpretations of the lesson of these works. Nevertheless, I shall set aside these concerns, except to note, in anticipation of what follows, and in accord with a longstanding tradition in philosophy dating back at least to Kant, that it is easy to (mis)interpret prosocial actions as being prosocially motivated when a large number of underlying psychological factors, including highly egoistic ones, may better explain them.

Scheffler claims that we recognize the finitude of our own lives, but that such recognition does not make us despair. He writes (2013, p. 44):

Every single person now alive will be dead in the not too distant future. This fact is universally accepted and is not seen as remarkable, still less as an impending catastrophe. There are no crisis meetings of world leaders to consider what to do about it, no outbreaks of mass hysteria, no outpourings of grief, no demands for action.

Scheffler continues (2013, p. 45): “Not only is that fact not regarded as a catastrophe, it is not even on anybody’s list of the major problems facing the world.”

We must be careful in setting out Scheffler’s exact argument, because, as the discussion below shows, it is not exactly clear how the argument is supposed to work. So my hope is that the formalisms I use will help us both understand both the nature of Scheffler’s argument and its flaws. Let I_F represent the scenario that Future generations of persons will continue Indefinitely. To be clear, I_F is not the scenario where some specific future generation or generations will live indefinitely; rather, it is that each generation will live only for a limited duration of time but successive generations will continue to be brought into existence indefinitely many times, in the familiar way.

I use the term “indefinitely” rather than “infinitely,” since Scheffler holds that humanity itself will likely someday perish. So for Scheffler, we are not aware of any set limit on how long humanity will continue to exist, and we have good reason to believe that it will be for a very long time, and it is this indefinitely long succession of future generations of humanity that is more valuable to us than our own existence would be.²

Now, let I_P represent the possible scenario that our own lives (and the lives of our loved ones and others in the Present generation) will continue Indefinitely long beyond the expected duration of its lifespan (which I shall assume to be 80 years). So while ~ I_P represents the present scenario, I_P is the scenario in which members of the present generation will live as long as we currently expect humanity as a whole to survive into the future.³

On the face of it, Scheffler’s argument seems to be the following (A):

• A1. We don’t regard ~ I_P as catastrophic.

• A2. We do regard ~ I_F as catastrophic.

Thus,

• A3. We regard ~ I_F as much worse than ~ I_P.

Scheffler’s support for A1 is that we don’t regard the actual present situation as catastrophic, and the actual present situation is such that ~ I_P. Scheffler’s evidence for A2 is that we do regard the Children of Men scenario as catastrophic, and that is a scenario of ~ I_F.

Granting that ~ I_F would be catastrophic, it should be noted that the Children of Men scenario is one where both ~ I_P and ~ I_F obtain, and it is this conjunction that the people in Children of Men find catastrophic. So, even if we grant that a situation of (~ I_P & ~I_F) is a catastrophe, more needs to be said to provide support for the idea that ~ I_F, on its own, is a catastrophe, or that ~ I_F is worse than ~ I_P.

The considerations given by Scheffler seem only to support the following argument (B):

• B1. We don’t regard (~ I_P & I_F) as catastrophic.

• B2. We do regard (~ I_P & ~I_F) as catastrophic.

Thus,

• B3. We regard ~ I_F as much worse than ~ I_P.

But these premises are not sufficient to demonstrate the conclusion, for they do not properly compare the values of I_P and I_F. For it is consistent with B’s premises that even though the state of (~ I_P & I_F) is not catastrophic, (I_P & I_F) on its own is so fantastically valuable that the loss of value in going from it to (~ I_P & I_F) would not be catastrophic per se, but would still exceed the loss of value in going from (~ I_P & I_F) to (~ I_P & ~I_F). Thus it is still consistent with these two premises that the value of I_P is greater than the value of I_F.

Perhaps it is transitions between states that may or may not be catastrophic, rather than states described on their own. In what follows, I discuss the relevant arguments in terms of transitions rather individual states because doing so helps with the presentation of my view. (There may be isomorphic ways to make the same central point I am making while restricting the discussion to values of states themselves, but I will not discuss these). Using transitions, a better argument in the spirit of Scheffler’s view is C:

• C1. We don’t regard a transition from (I_P & I_F) to (~ I_P & I_F) to be catastrophic.

• C2. We do regard a transition from (~ I_P & I_F) to (~ I_P & ~I_F) to be catastrophic.

Thus,

• C3. We regard ~ I_F as much worse than ~ I_P.

One way to illustrate this argument is to use a toy model of values for the varying scenarios in accord with the intuitions that Scheffler employs. In what follows, in saying V(x) = y, I intend that the value we place on scenario x is y.

Since we are concerned with cases in which either (or both) I_P and I_F do not obtain, we might be inclined to discuss values such as V(~ I_P). However, I don’t mean by this that somehow the absence of scenario I_P has a value (or disvalue) in itself. Rather, it should be understood as the value that we would give to the world in a scenario in which P does not obtain. In particular, it is the state of the present generation living for the normal duration of its lifespan. I will represent this Background state as B. B & I_P thus represents that the present generation will first live 80 years, and in addition, will continue to live beyond that for an indefinite number of years.⁴ So below, when I discuss cases of value such as V(B & ~I_P & I_F), I do not intend that ~ I_P adds some special value (or disvalue). Rather, it should be understood as simply being equivalent to V(B & I_F); my inclusion of “~I_P” just serves as a reminder that the state under consideration is one in which I_P does not obtain.

With these clarifications, consider argument D:

• D1. V(B & I_P & I_F) = 11,000.

• D2. V(B & ~I_P & I_F) = 10,000.

• D3. V(B & ~I_P & ~I_F) = 100.

Thus,

• D4. We regard ~ I_F as much worse than ~ I_P.

D compares what would happen if, at first (in D1), both the present generation, and the set of future generations, will persist for an indefinitely long time; then (in D2), the present generation loses its indefinitely long persistence and the situation becomes as the world is now; and then (in D3), the Children of Men scenario obtains. Assuming that the loss in transitioning to D3 is catastrophic, and the loss transitioning to D2 is not, it seems to follow that the persistence of future generations is more important than the persistence of the present generation. Using the toy values, the disvalue to us in losing I_P, as determined by the 1000 unit loss of value in going from the scenario in D1 to the scenario in D2, is much less than the disvalue in losing I_F, as determined by the 9,900 unit loss in value in going from the scenario in D2 to the scenario in D3.

In our toy model, seemingly:

• D5. V(I_P) = 1,000 (from D1 and D2).

• D6. V(I_F) = 9,900 (from D2 and D3).

And thus,

• D7. V(I_P) < V(I_F).

And if D7 is the case, then Scheffler would be correct that future generations matter more to us than the present generation, ourselves included, does.

3 Apples and Oranges

However, a further assumption is needed for this construal of Scheffler’s argument to be sound. It is that the value of a conjunction of two (or more) scenarios is simply the sum of the value of each scenario in the conjunction. It is only this assumption that allows us to infer D5 from D1 and D2 and D6 from D2 and D3. I will call this the Additive Assumption. However, the additive assumption is false.

Consider a case where someone sets out on a trek in the desert with a basket containing both a dozen apples and a dozen oranges, which they will use for sustenance as well as for combating thirst. (They are not bringing water). It is most desirable to consume all the fruit in the basket. However, no serious problems will ensue if, say, all the apples are somehow lost or go bad, as long as the oranges are not also lost or go bad. And, importantly, vice versa. But if there are no fruit whatsoever, the individual will die in the desert.

Let Apples represent the scenario in which they eat all the apples. (Let’s also stipulate that either they eat all the apples or no apples whatsoever, and thus ~ Apples means not just that they do not eat all the apples, but that they do not eat any). Oranges is defined in a parallel manner. The following argument uses toy values for the various scenarios:

• E1. V(Apples & Oranges) = 110.

• E2. V(~ Apples & Oranges) = 100.

• E3. V(~ Apples & ~Oranges) = 0.

Thus,

• E4. Oranges is more valuable than Apples.

That’s because, by the additive assumption,

• E5. V(Oranges) = 100 (from E2 and E3).

and thus,

• E6. V(Apples) = 10 (from E1 and E5).

This reasoning is quite clearly fallacious, because we could simply switch “Apples” and “Oranges” to arrive at the opposite conclusion. It thus shows that the Additive Assumption is false. For there is nothing contradictory with accepting E1, E2, and E3, while also holding:

• E7. V(Apples & ~Oranges) = 100.

Just because the transition from (Apples & Oranges) to (~ Apples & Oranges) is not catastrophic, and the transition from (~ Apples & Oranges) to (~ Apples & ~Oranges) is catastrophic, it does not show that Oranges is more valuable than Apples. So more is needed to show that argument D succeeds in demonstrating Scheffler’s conclusion.

One may point out that the values of present and future generations are seemingly intrinsic values, whereas the values of apples and oranges in this scenario are instrumental. Since instrumental goods are typically replaceable in a way that intrinsic goods are not, there is reason to believe that even though the Additive Assumption fails when attempting to compare Apples and Oranges in my example, it may still hold when comparing other values, like those of present and future generations.

However, there is no way to show that this distinction makes any difference to my argument, and perhaps even some reason to think that it would work against Scheffler’s (2013) argument (see also Scheffler 2018, Ch. 5). That’s because, given that future people by definition don’t yet exist, they seemingly are indeed replaceable. We wouldn’t regard it as a tragedy if we had been on a path for a certain set of future people to exist, but given some action of ours, a different set of future people whose lives would be just as many, as just as good, as those in the first set, will now exist. But with present people, we shouldn’t be receptive to such replaceability. This gives reason to think that we do indeed value members of present generations more, because we know and love at least some of them as particular existing individuals. And because present people, rather than future people, are irreplaceable, it would seem to give reason to think that V(I_P) is at the very least equal to, if not greater than, V(I_F), and that the additive assumption will be false in this application.

4 How Can We Determine How Much We Value Future Generations Relative to Present Ones?

4.1 Armchair Arguments

Figure 1 describes the four relevant scenarios and twelve possible transitions between them. (I leave off the Background of the scenarios since it should be included in all four).

Fig. 1

Diagram of scenarios for valuing present and future generations

Scheffler claims that (I_P, I_F) → (~ I_P, I_F) is not a catastrophe, but (~ I_P, I_F) → (~ I_P, ~I_F) is a catastrophe. If my arguments in the previous section are correct, this on its own is insufficient to show that V(I_F) > V(I_P). It also does not show that if we had a choice to go from our present (~ I_P & I_F) scenario to (I_P & ~I_F), we would/should choose not to. In this section, I will first give reason to think that we do value (I_P & ~I_F) more than (~ I_P & I_F). However, part of the ingenuity of Scheffler’s argument is that he compares V(I_F) and V(I_P) indirectly. I will thus discuss some other transitions to show that even if we accept Scheffler’s view that (~ I_P, I_F) → (~ I_P, ~I_F) is a catastrophe and (I_P, I_F) → (~ I_P, I_F) is not, it still does not show that V(I_F) > V(I_P).

One important question for our purposes seems to be:

• Q1. Given that the present scenario is (~ I_P & I_F), would we opt to undergo transition (~ I_P & I_F) → (I_P & ~I_F) if given the choice?

There are reasons to think that we value present generations more than future ones, and that we would choose to undergo (~ I_P & I_F) → (I_P & ~I_F) if given the choice. Arguably, we value the current situation (~ I_P & I_F) much less than Scheffler thinks. By restricting his focus to those who reject the existence of a supernatural afterlife from the “we” in his writings, Scheffler thereby excludes at the outset a large portion of people who likely do think that (~ I_P & I_F) would be catastrophic. A religious skeptic might even say that the best explanation of why people believe in a supernatural afterlife, when there is arguably no empirical evidence of one, is the very fact that they would indeed view the present situation of (~ I_P & I_F) as catastrophic—if only they were to bring themselves to believe that it is going to occur (as Freudians such as Becker 1973 hold).

Additionally, we should revisit some of what Scheffler says to support the idea that the present scenario is not catastrophic. As noted above, he writes that there are (2013, p. 44) “no outbreaks of mass hysteria, no outpourings of grief.” This is an odd thing to say: the lack of hysteria is due to the fact that we as a society have known for a very long time, and individual people know from a very young age, that we are all going to die, and there is typically no single immediate occasion for hysteria. Furthermore, many people, at least in their formative years, are indeed struck by the difficulties that death presents. And virtually every death leaves at least some of those remaining in grief. And ever further, some in the life-extension movement (e.g., de Grey 2005) do in fact explicitly do regard human death as a catastrophe. So it is not clear from the armchair considerations I’ve discussed here that we don’t massively value present generations.

There are questions aside from Q1 that might help us determine the relative values of present and future generations. One complication in teasing out our intuitions is the fact that our present expectations are already of (~ I_P & I_F), and so if we were to learn, say, that (~ I_P & ~I_F), it would mark a change in expectations, and that change might cause special difficulties. If one has lived one’s life working for a goal, any goal, and then the goal becomes unachievable, then it is easy to see how one could be sent into despair. If one’s goal is to cure cancer, and then one discovers that everyone will perish very soon regardless of whether cancer is cured, then it is no doubt that it would cause one to devalue those projects that one had valued. So (~ I_P & I_F) → (~ I_P & ~I_F) will devalue lives merely due to the difficulties arising from mere changes in expectation. Scheffler is aware of this issue (2013, pp. 46, 47), but remarks that it is implausible that all the intuitive badness of the Children of Men scenario is due to our change in expectations. So it is worth considering other choices under a stipulation that we haven’t already structured our lives given certain expectations.

First, let’s imagine that a malevolent deity makes it such that the Children of Men scenario (~ I_P & ~I_F) is likely to obtain, and that humans have been aware of this for quite some time. But then a benevolent deity gives us a choice between (~ I_P & I_F) and (I_P & ~I_F). In other words,

• Q2. Given a choice between (~ I_P & ~I_F) → (~ I_P & I_F) and (~ I_P & ~I_F) → (I_P & ~I_F), what would we choose?⁵

Second, let’s imagine that a benevolent deity makes it such that (I_P & I_F), and we are aware of this. Then, a malevolent deity gives us a forced choice between (~ I_P & I_F) or (I_P & ~I_F). In other words,

• Q3. Given a choice between (I_P & I_F) → (~ I_P & I_F) and (I_P & I_F) → (I_P & ~I_F), what would we choose?

Our intuitive responses to Q3 might not mirror our intuitive responses to Q2.

Third, since Scheffler claims that (~ I_P & I_F) → (~ I_P & ~I_F) is a catastrophe, one might also wonder:

• Q4a. Would (I_P & ~I_F) → (~ I_P & ~I_F) also be a catastrophe?

And if so,

• Q4b. Would (I_P & ~I_F) → (~ I_P & ~I_F) be a worse catastrophe than (~ I_P & I_F) → (~ I_P & ~I_F)?

My main aim in posing these questions is not to answer them, but to note that they are much more relevant in comparing the value of present vs. future generations than Scheffler’s comparison between (I_P & I_F) → (~ I_P & I_F) and (~ I_P & I_F) → (~ I_P & ~I_F). I do invite the reader to consider Q2–Q4 carefully; my suspicion is that many if not most people would choose (~ I_P & ~I_F) → (I_P & ~I_F) in Q2, (I_P & I_F) → (I_P & ~I_F) in Q3, and would also say regarding Q4b that (I_P & ~I_F) → (~ I_P & ~I_F) would be a worse catastrophe than (~ I_P & I_F) → (~ I_P & ~I_F).

One issue complicating the intuitive responses to these questions is that some people might regard an indefinitely extended life of a single person or generation as becoming tedious or perhaps incoherent (in the spirit of Williams 1973). For this reason, some may respond to the questions above in ways that do show a preference for I_F over I_P, contrary to what I surmise. But this would not show that people value present generations over future generations—it would only show that people believe that attempts to instantiate the goods within a single generation marginally decrease over time, and so to maximize overall value, humanity would have to be refreshed (so to speak) on occasion.

More empirical work should be done to examine these issues more carefully than the armchair conjectures I have just made. And the questions I state here can be a basis for qualitative empirical studies of people’s values. For there indeed is a lot of empirical work on these empirical questions, as I will now discuss.

4.2 On the Empirical Study of Our Concern for Future Generations

Given the questions I have posed, it is still unclear (to me at least) how much we value people hundreds or thousands of years in the future. As I have already noted, the relative degree to which we value present and future generations is an empirical question,⁶ and though the main focus of this paper is on Scheffler’s armchair arguments, I would like to discuss briefly some of the empirical work on related issues.

A reason to doubt that we give much value to future generations is humanity’s current treatment of the environment, and in particular our propagation of global climate change, which seems to indicate our significant disregard for the long term (see Broome 2012, Stern 2006, and Broome 1992 for analyses of the difference between the actual temporal discounting of the value of future generations and a normative view of how we ought to discount it). Would we be as reckless as we are in changing climatological conditions if we believed that we ourselves would be the main ones harmed by it? The consensus amongst environmental economists and psychologists in an enormous literature⁷ is that people do temporally discount, quite significantly, the goods of possible future generations (see e.g., Heal 1998 and Arrow 2013 on the economics of discounting). Even if, in surveys, we might say that we value future generations, many of our consumer choices do not indicate that we really believe it.

It is difficult to compare the relative importance of a vast swath of empirical evidence on discounting with Scheffler’s armchair use of the Children of Men scenario. However, I should note something in support of Scheffler’s methodology. It could serve as a reason to be skeptical of the empirical work, and as evidence that the empirical work has overlooked something important. In particular, Scheffler tries to elicit deep values that may not be demonstrated by our consumer choices, and one of Scheffler’s main points is that our true values may be obscured from even our own view. Relatedly, McShane (2017) argues that economists’ methods of valuation don’t correspond well to philosophers’ notion of intrinsic value, partly because economists typically just study surface-level, unreflective behaviors and values. Furthermore, it may be worth noting on Scheffler’s behalf that the empirical evidence is not unanimous. In fact, some believe that the way we discount future generations is similar to how we temporally discount within our own generation (see e.g., Chapman 2001).

I should also note that there is substantial literature in academic psychology about what is known as generativity. In his classic work, Erikson (1950) argues that humans, especially those later in life, in proper conditions, develop a deeper concern for future generations, in part due to a desire to leave an enduring legacy. (While legacy concerns may seem on the surface to be egoistic, they are not, for reasons described by Scheffler: concern for one’s future legacy is valuable at least in part because one does intrinsically value those in the future who maintain one’s legacy). Since Erikson’s time, other psychologists have studied generativity in greater detail (see e.g., McAdams 2016, Ch. 9; Kotre 1984).⁸ This paper is not the place for a comparison of Scheffler’s arguments and those of psychologists, but McAdams does provide empirical evidence in favor of generativity. So I encourage other philosophers working on the moral psychology of future concern to engage with the social science literature that is already present.

5 Objections

5.1 Objection 1: Our Contingent Valuations Don’t Matter

I have argued that in principle, we don’t value the future more than the present. However, one could respond by saying that, given that ~ I_p will most likely be the case no matter what we do, comparing the value of present and future generations in the ways that I have suggested is irrelevant to any decision framework, and thus doesn’t really matter. There is an important sense in which, given that we know that our current lives are limited, and there is virtually no possibility that that can change, we do in fact have deep moral concern for future generations. So it might be objected that whether or not we would value future generations if the present generation could live an indefinitely long time, what matters is what we do value given that it won’t, and thus my argument doesn’t undermine the spirit of Scheffler’s view.

This objection misses the mark. Scheffler’s overall argument is that even though we don’t seem to be very concerned with future generations, we do in fact give a significant underlying value to future generations that even exceeds that of our value of present generations. (And Scheffler 2018 gives reasons why we ought to have significant concern for future generations). However, the argument I have given shows that it might be the case that, at an even deeper level, we don’t care nearly as much about future generations as we do about present ones. And if that’s the case, then we won’t be very motivated to act in ways that are helpful towards future generations.

5.2 Objection 2: Scheffler Only Really Intends to Show that Egoism is False

In some places, Scheffler indicates that his argument arises largely from a rejection of psychological egoism. For instance, he later summarizes an important feature of his (2013) argument by saying that (2018, p. 54): “the fact that the prospect of humanity’s extinction would be so devastating for us reveals some limits to our egoism”. All I have shown is that we don’t have evidence that we care more about future generations than the present ones. However, perhaps the spirit of Scheffler’s argument is that we value others at least to some extent because we want to participate (in a positive way) in something bigger than ourselves (which is a theme that looms large in Scheffler 2018). So one holding a view like Scheffler’s can grant to me the point that we don’t value the future more than the present, but nevertheless claim that what matters is that Scheffler’s arguments still show that we have highly other-regarding values, even if these other-regarding values are temporally neutral.

The main problem with this response is that an argument similar to the one I give in § 3 can be given even against this more limited and temporally-neutral anti-egoistic view. Insofar as we do value others, and we want to be part of something bigger than ourselves, perhaps it is due to our knowing that our own individual lives are highly limited. Perhaps we would value our own individual lives much more if we knew that we ourselves, as individuals, could live an indefinitely long time. If this is right, Scheffler’s argument that we do have deep underlying concern for other people (now or later) is not yet successful. So it is not clear to me at least that anything in the vicinity of Scheffler’s 2013 argument can be used against psychological egoism.

Another way to look at the issue is by turning to Joel Feinberg’s classic (1999/1958) discussion of psychological egoism. Feinberg discusses (1999, p. 497) a story in which Abraham Lincoln once saw some pigs struggling, and told his coach driver to pull over so that he could rescue them. Was this an instance of genuine psychological altruism? In the story, Lincoln says no: if he hadn’t rescued the pigs, it would have bothered him all day, and so the core motivation is egoistic. Feinberg says yes: it would not have bothered Lincoln if he did not already have intrinsic concern for the pigs.

However, the case as given by Feinberg is under-described. What else might Lincoln say or do afterwards? For example, Lincoln might later approve farm policy that showed no concern for pigs’ welfare. Perhaps after this event, Lincoln asked his coach driver to always, in the future, steer clear of struggling animals because he just doesn’t want to see them. Perhaps Lincoln complains to his loved ones about how his feelings for pigs is not rational, since he knows that pigs have no value. Perhaps Lincoln would happily agree to take a pill to rid himself of his pro-pig sentiment. In short, Lincoln’s seemingly intrinsic first-order desire for pigs not to suffer might be contradicted by a set of second-order desires that include values that only pertain to himself. And if a dispositional theory of value (such as in Lewis 1989, Ch. 7) is correct, whereby what we value is not simply what we care about but what we would want ourselves to care about, then it cannot be said that Lincoln does value the pigs intrinsically despite his first-order concern and despite Feinberg’s argument.

Some similar things can be said for our valuations of other people, and of the future—that they are not deep-seated values. Perhaps to avoid hearing of others’ suffering, many people indeed take a metaphorical pill—we immerse ourselves in our cellphones, our entertainment, our jobs, our families. Perhaps we do this as an attempt to hide from ourselves our larger prosocial desires because we both don’t really endorse them—we’d rather (a) not have surface distress over others’ struggles and (b) not act on the basis of this distress because we don’t really value others upon reflection (like my hypothesized Lincoln’s response to the pigs). Our very limited concern for others in our choices—especially others in future generations—does in fact give evidence that we don’t truly and deeply value them.

Again, these are mere armchair conjectures. My main claim here is that arguments that purport to show that despite some appearances, we do, in a somewhat hidden manner, value future people, or at least other people in general, on the grounds that we regard certain scenarios as catastrophic, do not provide strong evidence that we do, at the deepest level of our values, value them. Even though, as noted in 5.1, my own argument involves counterfactual scenarios that don’t obtain, that we have certain judgments in these scenarios may show that at a deep level, we don’t in fact value others.

6 Upshots

The argument in this paper is certainly pessimistic for those of us who do explicitly hope that people will come to value the future, and act accordingly. The road to having people act more significantly out of concern for future people is windier and more difficult to traverse than what those who agree with Scheffler’s argument might think. Perhaps it is true that at least some people wish to contribute to something bigger than themselves. But lots of people don’t, or at least don’t extend much moral concern to those beyond a small circle of partiality.

However, there is reason for optimism. Scheffler, especially in his (2018), gives reasons for valuing future generations, such as how it enhances ourselves as persons (2018, pp. 48, 49) and how it shows a love for humanity as a whole (2018, p. 62). I have no objection to what he says there, and perhaps it does give good reason to think that we ought to value future generations more than we do (though see Greaves 2019, who argues that Scheffler’s arguments, at their best, don’t depart from classical utilitarianism as much as Scheffler supposes). Scheffler’s arguments accord with the literature on generativity that I mention above, giving more hope that concern for future generations is indeed a psychologically stable and widespread phenomenon.

So the case for our moral concern for future generations is mixed. On the one hand, we should not ignore empirical research specifically by economists that gives evidence about the degree to which we discount future value. However, such studies may be limited because they have difficulty in determining deeper values, not demonstrated in choices or behavior, that philosophers like Scheffler may elicit using thought experiments. At the same time, I have argued that even Scheffler’s thought experiments showing that we value others fail to get at the core of our valuations, because we may have even deeper-seated values that undermine the lessons of these thought experiments. So in the end, I encourage philosophers to work together with social scientists to help us understand our deep values, and I hope that this paper is at least a small step in that direction.