The Prudential Rationality of Risking Traumatic Brain Injury in Dangerous Sport: A Parfitian Defense

Abstract

It is an open question whether participation in dangerous sports is prudentially rational given the high risk of traumatic brain injury they involve. This paper explores the merits of one attempt to rationalize participation in dangerous sport, which is based on Derek Parfit’s idea that it is rational for you to care less about the well-being of your future selves the more distant they are, because time diminishes the degree of their psychological connectedness to your present self. I respond to two sets of concerns about this defense of dangerous sport. First, there are concerns that the Parfitian defense is too weak to justify taking such serious risks as dangerous sport involves—specifically, objections based on an appeal to the overall shape of one’s life, as well as objections from self-directed duties. Second, there are concerns that the Parfitian defense would prove too much—specifically, that it would rationalize generally reckless risk-taking for fun; and that it would make one exploitable. I show that none of these objections succeeds, and I conclude that the Parfitian defense remains tenable.

1. Introduction

I take it to be an open question whether it is prudentially rational—for anyone—to engage in dangerous sports such as American football, mixed martial arts, and auto racing. Such sports involve a high risk of traumatic brain injury (TBI), which can lead to neurological and psychiatric disorders that impact athletes’ rational capacities, thereby diminishing their ability to pursue a good life in their later years. It is not clearly rational to engage in an activity that puts one’s future self at such risk.1 So I distinguish sports such as football from sports such as BASEjumping, in which death is as likely as permanent injury. (See e.g., [1,2]). In addition, while I think it is a serious question whether any instance of participation in dangerous sport is prudentially rational, my interest is primarily in dangerous sports at the highest level, where competition pushes elite athletes to their limits, leading them to take risks to their well-being that non-elite athletes may avoid. When I speak of “dangerous sport”, I intend to refer only to such sports.

One argument that it can be prudentially rational appeals to Derek Parfit’s [3,4] views about personal identity and their implications for rational concern about one’s future well-being. Roughly, on Parfit’s account, you have prudential reasons to care about your future selves only insofar as those selves are psychologically connected to you. Because the passage of time diminishes the degree to which future selves are connected to present ones, you have less reason to care about your future selves the more distant they are. This, combined with the fact that there is a relatively small window in which an athlete can effectively pursue the goods of dangerous sports at the elite level, may seem to rationalize taking serious risks in such sports.

This paper explores the merits of this defense of the rationality of risking brain injury in dangerous sport (hereafter, I substitute “DS” for “dangerous sport”).2 Specifically, I discuss two sets of concerns about it. First, there are a number of reasons to think the Parfitian defense of DS cannot deliver the desired justification for engaging in DS. (Call these the “too weak” objections.) For one, the amount by which it is rational to discount one’s future well-being, if Parfit is right about what matters in personal survival, may simply be too little to justify risk-taking in DS given the seriousness of those risks. At best, Parfit’s view rationalizes less concern for one’s distant future self; it does not rationalize apathy. Consequently, it is possible that although some significant degree of discounting of one’s future well-being is rational, it is nevertheless prudentially irrational to take the sorts of risk that wholehearted engagement in DS typically requires. Moreover, the rationality of Parfitian discounting is consistent with the view that we have certain duties to our future selves that would be violated by taking certain risks in the present.

The latter view forms the basis of another (deontic) objection to the Parfitian defense. It might be argued that we have special duties to future individuals to whom we are strongly psychologically connected and that fulfillment of these duties precludes taking such risks in DS. Such duties might be grounded in a principle of respect for autonomy, or they might be grounded in a principle of utility (the thought being that preserving your rationality is a precondition of your being able to act so as to maximize the good over the course of your life). Provided there is an important link between prudence and morality, such views may support the conclusion that pursuit of DS is prudentially irrational. The deontic objection I will discuss is a version of the argument that pursuit of DS involves a failure to respect one’s autonomy. It argues that (self-imposed) threats to rational capacities are different from other kinds of threats (e.g., to life or physical well-being) in a way that makes pursuit of DS prudentially irrational.

The second set of concerns are based on the worry that the Parfitian defense of DS would prove too much. (Call these the “too strong” objections.) It might be argued that if the Parfitian defense succeeds, then it also rationalizes the use of dangerous drugs for recreational purposes, as well as generally reckless risk-taking for fun. In addition, there is a problem of exploitability, which suggests that the Parfitian defense goes too far: as Ahmed [7] has recently shown, someone who rationalizes her concern for future well-being on Parfitian grounds will voluntarily take a course of action that reduces her overall well-being at some time without increasing it at any other.

I will argue that none of these objections succeeds. (I discuss “too weak” objections in §3, and “too strong” objections in §4.) My ultimate conclusion will, however, be somewhat tentative. If I am right, the objections to the Parfitian defense are not convincing. This means that the Parfitian defense is viable; it does not mean that there are no further grounds (in the form of direct arguments) that taking such serious risks to one’s future well-being fails to comport with prudential rationality. At the same time, the discussion shows that the issue of the prudential rationality of risking TBI in DS is complex, insofar as it cannot be adequately addressed independently of consideration of the rational discounting of future well-being.

Let me clarify what is at issue. The defense of DS that I am concerned with does not rest on uncertainty about the future (hence, it does not rest on uncertainty about whether one will obtain certain goods, through participation in DS or otherwise).3 Nor does it depend on athletes’ wills being granted normative authority within the domain of DS. I will assume that it is generally impermissible for others to interfere in athletes’ choices and actions with respect to their participation in DS. In general, the impermissibility of interfering with certain choices and actions does not imply that those choices and actions are rational any more than it implies that they are morally permissible. We may also suppose that athletes generally know that, independently of their participation in DS, they will undergo what I will refer to as “decayed connectedness”: their distant future selves will be psychologically connected to their present self to a much smaller degree than their near future selves (and the more distant the future self is, the less psychologically connected it is to the present self).4 The question is whether, supposing it is rational to discount due to decayed connectedness, this can rationalize participation in DS.

2. Parfit’s Argument and Its Application to Dangerous Sport

I will first briefly explain Parfit’s argument in a bit more detail and show how it can be applied to DS. For Parfit, the rationality of time preference—discounting of one’s future well-being, in particular—is grounded in a correct view of personal identity over time. In Parfit’s view, this means it is grounded in considerations of psychological connectedness and continuity (as in his view there is no enduring self, which is numerically one and the same at different stages of a person’s life). These are the relations that fundamentally matter when it comes to the kind of special concern you have for your own future well-being. (Parfit refers to the relation consisting of psychological connectedness and continuity as “Relation R”.) Continuity is defined in terms of connectedness: A person $A_1$ at time $t_1$ is continuous with a person $A_{\mathit{n}}$ at a later time $t_{\mathit{n}}$ iff there is a chain of persons-at-times, $A_2$-at-$t_2$, $A_3$-at-$t_3$ …$A_{n-1}$-at-$t_{n-1}$, with $t_1$<$t_2$<…$t_{n-1}$<$t_n$, such that each is strongly connected to the next. The passage of time diminishes the degree of psychological connectedness earlier selves bear to later ones: that is, it diminishes the number of direct connections5 between memories, experiences, character traits, and other mental states. Since connectedness does, but continuity does not, come in degrees, it is connectedness that can diminish over time. So, in Parfit’s view, if there are fewer connections between you now and your distant-future self than between you and your near-future self, and connectedness is what grounds your concern for your future well-being, then it is rational for you to discount your distant future well-being. (“[If] what matters holds to a lesser degree, it cannot be irrational to care less” [4], 99).6

The argument can be stated as follows:

• Psychological connectedness and continuity are the grounds for caring about one’s future selves.
• Psychological connectedness is a matter of degree, and connectedness to one’s present self diminishes over time (distant future selves are connected to one’s present self to a much lower degree than near future selves).
• One’s grounds for caring about one’s distant future selves are comparatively weak, and therefore it is rationally permissible to care less about them than one’s near future selves.

The Parfitian defense applies this conclusion to the case of DS as follows:

P1

Because it is rationally permissible for an athlete to discount her distant-future well-being, it is rationally permissible for her to take certain risks that may diminish that well-being—provided the level of risk is (inversely) proportional to the degree of connectedness between present and future selves.7

P2

The risk of TBI in DS is, for many athletes, within the range of acceptable risk.

P3

Risking TBI in DS can be rational.8 (from P1 and P2)

P3, combined with the fact that athletes’ access to the goods of dangerous sports is conditioned by their age (and related considerations, chiefly health), makes it rationally permissible to take serious risks in such sports.

The first set of objections to the Parfitian defense targets P2. The second targets P1.9

3. Is Future Discounting Sufficient to Rationalize DS?

3.1. “Arc of a Life” Considerations

Consider the idea that the amount by which future well-being is rationally discounted (if Parfit is right) is insufficient to justify present risk-taking in DS. This might be the case, if the discounted value is offset by some other factor that calls for greater concern for one’s distant future self. For instance, “arc of a life” considerations may countervail the relevant Parfitian considerations. In particular, it may be that later stages of a life have more bearing than earlier ones on the overall goodness of the life [10,11]. If this is the case, then the amount by which future well-being is rationally discounted due to decayed connectedness may be less than the amount by which future well-being is “marked up” due to its greater bearing on overall goodness. In that case, decayed connectedness will not help justify greater risk-taking in the present.

There are examples that suggest that later stages of a life can sometimes have more impact on its overall value than early stages. However, as Bloodworth [12] notes, it is difficult to see why in general harms in one’s later years should weigh more heavily when considering overall well-being. This is because, in assessing the overall goodness of a person’s life, it is difficult to avoid any appeal to her own preferences, including, crucially, preferences about the overall shape of her life. Granted, we do not (and should not) regard an individual’s well-being as solely a function of preference satisfaction. Moreover, while preference satisfaction does plausibly play a role in contributing to an individual’s well-being, some preferences are given more weight than others, and some may be given no weight at all.10 However, some preferences are given significant weight in contributing to the individual’s particular conception of what constitutes a good life for her.

Some people simply prefer a life involving certain risks—along with certain distinctive accomplishments and experiences of high intensity that are thereby made possible. In many cases, they do so without a lot of concern for the overall shapes of their lives. Such preferences seem to play an important role in determining what constitutes the good life for them. Not anything goes, of course, but we grant individuals a certain authority on the matter of what is best for them, based in part on such personal preferences about the kind of life they want for themselves. Consequently, it is difficult to judge that risk-seeking individuals are prudentially irrational in virtue of not giving greater weight to later stages of life than many risk-avoidant individuals.

3.2. Deference to Individuals’ Risk Attitudes

Lara Buchak’s recent work on risk-taking and rationality is relevant here [14,15]. While Buchak is concerned with instrumental or “means-end” rationality, her notion of risk-weighted expected utility is relevant to prudential rationality insofar as individuals’ risk attitudes help determine what is good or bad for them.11 Provided an individual’s risk-seeking disposition makes it reasonable for her to engage in certain high-risk activities, this fact may rationalize participation in DS even in the face of countervailing reasons.

If Buchak is right, then maximizing expected utility is just one way of making rational decisions. Another way is to maximize risk-weighted expected utility, according to your risk attitudes. Buchak explains the notion of risk-weighted expected utility as follows:

[T]here are actually three psychological components in preference-formation and decision-making: how much an individual values consequences (utilities), how likely an individual thinks various states of the world are to obtain (probabilities), and the extent to which an individual is willing to trade off value in the worst-case scenarios against value in the best-case scenarios (the risk function). There are two different ways to think about the risk function: as a measure of distributive justice among one’s “future possible selves”—how to trade off the interests of the best-off possible self against the interests of the worst-off possible self—and as a measure of how one trades off the virtue of prudence against the virtue of venturesomeness.12 ([14], 91)

Buchak’s view is that there is a fairly wide range of acceptable attitudes towards risky choices and actions and no single risk-benefit ratio that all rational agents require. Some agents generally care more about potential costs and potential benefits, and give greater weight to worst-case scenarios than best-case ones; others may do the opposite; still others may give equal weight to them. None of these attitudes is inherently more rational than any other ([14], 91).

Buchak also notes—crucially, as regards the issue of the prudential rationality of participation in DS—that on the view that we ought to respect others’ risk attitudes, allowing them to choose the tradeoffs they’re willing to make between worse scenarios and better ones, we are not recognizing their right to choose over doing what is in their best interests. That is, it is not that individuals may be regarded as rational for acting against their best interests, provided they have certain risk attitudes. Rather, we are recognizing that it can truly be in an individual’s best interests to take certain high-risk gambles. Buchak’s account is thus relevant to prudential rationality. An individual’s risk function helps to determine what is good or bad for her and, hence, what she has reason to do as concerns her own well-being or quality of life.

I am suggesting that whether it is prudentially rational for an individual to participate in DS will depend (among other things) on the shape of her risk function. Provided she has a concave risk function (i.e., as the best consequences are realized in less likely states, she cares proportionally more about them) it is not unreasonable to think that her participation in DS is, for her, prudentially rational.

Let me be clear on the dialectic. My appeal to Buchak’s account is meant to be merely a defensive move against the “arc of life” objection to the Parfitian defense of DS. The thought is that if Buchak is right that there is no single risk-benefit ratio that all rational agents require, then we need some reason to think that the range of permissible attitudes excludes those involved in DS, if we are going to say that DS is not prudentially rational. I do not see any clear reason to think that such choices and actions lie outside of the relevant range, although I acknowledge that there may be such a reason. Certainly, the mere fact that some individuals have risk functions that permit participation in DS does not establish that such participation is prudentially rational for them. Nor does the fact that some individuals have risk functions that forbid such participation establish that such participation is irrational for them. The rationality of DS will depend in part on the kind of value it creates. Consideration of the risk functions involved in participation in DS does however show that the ’arc of a life’ objection to the Parfitian defense of DS is weak. Even if later stages of life typically have greater weight than earlier stages, it is possible that the actual value of early stages of a life involving participation in DS is great enough that later stages are rationally discounted, such that it is acceptable to impose serious risks on well-being at those stages.

3.3. Duties to Future Selves

We might have special duties to our future selves that make it immoral to take certain kinds of risks in the present. The risks involved in DS threaten to severely diminish not only the well-being of one’s distant future selves but also their connectedness to one’s present self.13 Although it would not immediately follow that participation in DS is prudentially irrational—it is conceivable that although it is morally unacceptable for you to participate in DS, it is not necessarily bad for you to do so—it would be reasonable to infer this, if acting immorally is typically not in your best interest, and there is no clear reason to think the case of DS is atypical in this regard. I think these are plausible claims. However, they are substantive, and they require defense that I cannot provide here. Nevertheless, I will proceed on the assumption that generally, at least as a contingent matter, acting immorally does not contribute to your well-being, and tends to diminish it.14

3.3.1. Schofield’s Theory of Self-Directed Duty and Hypothetical Retrospection

The view that it is immoral to take risks that threaten to severely diminish the well-being of your distant future self might be defended by appeal to the following principle endorsed by Paul Schofield [17]: X owes it to Y that she ϕ iff Y can legitimately demand that X ϕ. In general, it seems that you can sometimes legitimately make certain demands on your selves at times other than the present (e.g., that they not overindulge in food and drink, or act selfishly or impetuously) and that such demands can generate duties. For instance, it may seem reasonable for you presently to commit to keeping a promise, thus issuing a demand and imposing a duty on your future self. Duties of one’s future selves thus seem explicable in terms of future-directed demands. Duties to one’s future selves are less clearly explicable in terms of past-directed demands, however. It can of course be reasonable for your future self to regret certain actions of yourself at an earlier time, and to have certain other negative reactive attitudes, such as blame and resentment, towards earlier selves. However, it is not clear that past-directed demands are intelligible. Given that what’s done is done, what sense could it make to demand that one’s past self act in some way? The answer, for Schofield, is that duties to future selves are grounded in hypothetical demands—demands that would be made, under sufficiently idealized conditions ([17], Ch. 3). That is, we are to imagine that an individual who occupies, at different times, two different temporal perspectives, $P_1$ and $P_2$ might, as occupant of $P_2$, make certain demands of himself as occupant of $P_1$—supposing the occupants of $P_1$ and $P_2$ could interact as normal people do in the interpersonal case. Duties to one’s future selves are generated by such hypothetical demands, in Schofield’s view.15

Schofield’s theory of self-duty draws from Stephen Darwall’s account of the ’second-person standpoint’, which Darwall characterizes as “the perspective you and I take up when we make and acknowledge claims on one another’s conduct and will” ([18], 3). For Darwall, it is the fact that we can take up the second-person standpoint that makes us morally accountable. According to Schofield’s understanding of this idea, the capacity to make and receive demands from the second-person standpoint, either interpersonally or intrapersonally, is what makes it possible to have certain duties to one another. The “normative pressure” that grounds such duties “emanates not merely from norms of rationality or from the moral law, but from a person addressing demands to her” ([17], 22).

Using this framework, we may consider what an athlete’s distant future self would demand of her in the present, regarding her participation in DS: Would she demand that the present self not take such risks? Would she have legitimate reasons for doing so? It is difficult to say. After all, in speaking of “the” future self of the athlete, we seem to be presupposing that there is a unique such individual; and whether she would, or could, reasonably, make such demands seems to depend on how things actually turn out. Since it is not inevitable that, if the athlete continues to pursue DS she will have a brain injury and suffer loss of autonomy, there is a possible future in which this never happens. If the future self is considered from the perspective of this possible future, it is not at all clear that she would have any grounds for complaint, or any reason to demand that the present self act otherwise.16

Despite this difficulty, it might be argued that it is necessary to consider a range of possible future scenarios, each of which could plausibly materialize from the agent’s present actions and decisions, and to consider, in addition, how those present actions and decisions would be evaluated by the future self who occupies the temporal perspective corresponding to the relevant future scenario. The thought then may be that if there are any such scenarios in which the future self in question could reasonably demand that the agent have acted otherwise, then she presently has a duty to do so.

This account can be enriched and supported, I suggest, by incorporating insights from Hansson’s [20] framework of “hypothetical retrospection” (HR). HR is intended as an aid to moral judgments under conditions of risk and uncertainty. According to this strategy, ethical values of decisions are evaluated under the assumption that one of the branches of possible future developments has materialized. The evaluation is based on the values and information available when the action in question took place, rather than on values and information available at the future time in question. It is carried out from the point of view of the imagined future point of retrospection. Hansson explains:

[I]n order to be decision-guiding, hypothetical retrospection has to refer to the decision that one should have made given the information (actually) available at the time of the decision, not the information (hypothetically) available at the time of the retrospection. The decision-relevant moral argument is not of the form “Given what I now know I should then have….”, but rather “Given what I then knew, I should then have…”. The purpose of hypothetical retrospection is to ensure serious consideration of possible future developments and what can be learnt from them, not to counterfactually reduce the uncertainty under which the decision must be taken.…

Hypothetical retrospection aims at ensuring that whatever happens, the decision one makes will be morally acceptable (permissible) from the perspective of actual retrospection. To accomplish this, the decision has to be acceptable from each viewpoint of hypothetical retrospection. ([20], 148–149)

The aim in employing HR is to choose an alternative that is as defensible as possible in all possible future scenarios that could develop as a result of the choice. If we adopt this approach in conjunction with Schofield’s framework for duties to future selves, we will look for future branches in which future selves could reasonably make certain demands of the present self, based on her values and knowledge at the time of the action. This seems to be a reasonably good guide to determining one’s obligations to one’s future selves, at least in a range of cases.17 On this approach, the crucial question is whether there is a possible future scenario in which one’s future self can reasonably demand that one not participate in DS, in light of the values and knowledge of the present self.

The existence, as well as the nature and scope, of self-directed duties is controversial. (This is despite the fact that the language of self-directed duty is pervasive—it is common to speak of what one owes oneself, etc.18) Accordingly, there are many further worries that might be raised about the appeal to duties to one’s distant future selves in criticism of the rationality of risk-taking in DS—and, specifically, about the appeal to the Schofield–HR framework I have briefly outlined. In the remainder of this section, I will discuss two such worries.

3.3.2. Against Duties to Distant Future Selves: Diminishing Bonds of Concern

First, when thinking about self-directed duty in this context it is important to distinguish two ways in which an athlete’s pursuit of DS might be thought to wrong his distant future self: (i) it might be thought to wrong him by causing him to suffer from the effects of TBI; or (ii) it might be thought to wrong him by imposing a significant risk of such suffering. I claim that only (ii) is relevant. By participating in DS, the athlete is not directly responsible for his distant future self’s suffering from effects of a brain injury, because he is not aware that by participating in DS he is causing such an outcome, and he has no intention to do so. (Indeed, he might do everything he can, compatible with his pursuing DS, to avoid such an outcome.) If he wrongs his distant future self, he does it by imposing a certain risk of such an outcome. The question, then, is whether it is permissible for him to impose such a risk. In terms of the Schofield–HR framework, the question is whether his distant future self can (hypothetically) reasonably demand that he refrain from imposing such a risk. My suggestion is that in light of the fact that decay of connectedness over time leads to diminished “bonds of concern”19 (to use Dorsey’s [22,23] expression) between present and future selves, it is reasonable to regard imposition of such a risk as justified, and as permitted by prudential rationality.

Let me explain. First, consider that, on the Parfitian view, there would appear to be a kind of diminished marginal utility of preserving R over the course of a human life. That is, decay of connectedness over time will, it seems, eventually lead to a condition in which preserving R matters much less than it once did. From this perspective, hypothetical demands by distant future selves have little force by comparison with demands of highly connected, nearer future selves; they are analogous to demands by strangers (in the interpersonal case) whose situation is not significantly causally connected to ours. While such individuals can legitimately demand that we not interfere with their lives in certain ways (thus respecting their negative rights, including the right not to be attacked), it is controversial whether they can legitimately demand any sort of positive assistance.20

Can distant future selves at least legitimately demand that their early selves not impose significant risks to their well-being? It might be argued that they can. By comparison, distant strangers in less developed nations, as well as future generations, can perhaps legitimately demand that we act to curb our carbon emissions and take further steps toward abatement of global warming, lest we impose serious risks to their well-being. Similarly, your distant future self can perhaps demand that you not participate in DS, lest you impose serious risks to her well-being. However, there is an important difference between these cases in terms of the foreseeability of harm. By failing to curb our emissions (etc.), we foreseeably cause harm to people in less developed nations, and to future generations. Arguably, it is the foreseeability, and not merely the risk, of harm that grounds the reasonableness of future individuals’ hypothetical demand that we refrain from activities that will bring about that harm. Although an athlete in DS imposes a risk of harm on his distant future self, he does not foreseeably cause such harm. If hypothetical demands are the basis of duties to future selves, then duties to distant future selves would seem to be fairly weak.

To further illustrate the idea that decay of connectedness leads to diminished bonds of concern, consider one type of psychological connectedness, that involved in episodic memory. It is plausible that, as a person ages and his episodic memories fade, it is all the more important to retain the memories he still has and cherishes—up to a point. As connections continue to decay, there comes a point at which it actually becomes less important for him to retain remaining connections, rather than more so. Connectedness has at that point decayed to such an extent that there are no longer substantive ’bonds of concern’ between present and past selves. On this picture, there is a kind of tipping point where bonds of concern between one’s selves at different times become so weak that they no longer impose significant constraints on one’s choices and actions. After this point, any demands the later self is entitled to make on the early self would be very limited.21

If this view is correct, then the strength of your duty to preserve R may diminish rapidly as R itself decays once R has decayed beyond a certain level (See Figure 1). In that case, an athlete’s duties to her distant future self—i.e., the self that lies beyond the point at which the strength of her duty begins a precipitous decline—may not be strong enough, after all, to countervail her right to pursue her own good in the present through DS.

We should note that this picture is consistent with the view that an athlete who participates in DS has relatively strong duties to nearer future selves who are significantly more R-related to her present self. It might be thought that this is more pertinent since there are a number of known cases in which athletes with chronic traumatic encephalopathy (CTE) have died by suicide at a young age—evidently much earlier than any significant decay of connectedness.22 However, two points can be made. First, there is yet much ignorance of the nature and causes of CTE23 and its connection with suicide24, and such cases appear to form a relatively small percentage of cases of CTE in athletes.25 So, the risks of brain injury that are, at present, most relevant to prudential rationality appear to concern relatively long-term consequences. If it is not clear whether those risks are imprudent due to their being at odds with duties to future selves, then it is likely that this is because such duties are directed toward distant future selves whose connectedness to the present self is greatly diminished. Second, the issue of how risks to the well-being of one’s distant future self bear on the prudential rationality of present participation in DS is important, even if there are other pertinent risks regarding one’s nearer future. Here we are interested in challenges to the prudential rationality of DS based on its potential impact on athletes’ long-term well-being. In addressing such challenges, we may allow that there are both further reasons to be cautious, and further reasons to worry that participation in DS is not rational.

3.3.3. Against Duties to Distant Future Selves: The Nonidentity Problem

On the diachronic conception of duty to protect autonomy, you have a duty to protect your rational capacities for the sake of future selves whose livelihood depends on them. On the Parfitian view of personal identity, this view about future self-directed duty faces a version of the nonidentity problem. If your actions now, including your participation in DS, determine which future individuals come into existence—individuals who are not identical to you but are psychologically connected to you—then it seems none of those individuals will have been harmed by those actions, for they simply would not have existed otherwise, and it seems they cannot have been harmed by the very thing that brought them into being.26 As long as all such future individuals have lives that are worth living, they would seem to have no rational basis for complaint. The duty to respect their humanity is not violated, it seems, as long as you do nothing that results in any future life that is not worth living. Your present choices do not treat the humanity of any such individual as a mere means (to your present flourishing), since the relevant instances of humanity do not exist prior to those choices. To be clear, the problem for the appeal to self-directed duty is not that future selves are not identical to the present self; it is that their very existence depends on the present self’s decision to participate in DS (or not).

In response, it might be objected that everyone has a right not to be brought into a flawed existence [33,34,35]. Thus, if your choice to engage in DS results in a certain future individual with greatly diminished rational capacity, etc., coming into existence, rather than one without such flaws, then such a choice might reasonably be regarded as immoral regardless of whether the relevant harm is unavoidable if the individual is to exist at all.

An investigation of this complex and difficult problem is well beyond the scope of this paper. I will note, though, that there are significant problems for the appeal to the right not to be brought into a flawed existence in this context. One is that even if we suppose that your distant future selves have such a right, it is also plausible that you (now) have certain liberty rights that entail a right to bring such selves into existence (see [36]). Such rights seem to include an athlete’s right to participate in DS, as the latter is arguably grounded in a more basic right to self-determination. I am not suggesting that the right to self-determination entails the right to impose harm on others. However, arguably, it does entail the right to impose certain risks to others, including one’s future selves. As it is not clear how to respect both the right to pursue DS and the right of future selves not to be brought into a flawed existence, or that either right trumps the other, the present objection does not succeed.

The more important point, however, is that the choice to participate in DS cannot be regarded as involving the intention to bring about a future flawed existence, nor is it foreseeable for the athlete that such an event (involving one of her futures selves) will occur. Consequently, it seems she cannot be morally responsible for bringing about a flawed existence. If she is not, then she does not violate a right of her distant future self.27 It is important to note that in assuming that it is not foreseeable for any particular athlete that they will suffer long-term debilitating effects of TBI due to their participation in DS, I do not assume that it would be prudentially rational to risk this outcome. (This would beg the central question.) My point is that in taking the risk, one is not violating a right of the future self. Thus if one is being prudentially irrational in risking TBI, it is not for this reason.28

3.4. The Duty to Protect One’s Autonomy

3.4.1. Synchronic vs. Diachronic Duties to Self

I have considered whether participation in DS may be prudentially irrational because it conflicts with duties to future selves, and I have argued that there is reason to think it does not, if the strength of such duties depends on the strength of ’bonds of concern,’ and the latter depends on the psychological connectedness of the individuals who bear such relations. However, there may be relevant duties besides those to future selves, which conflict with the pursuit of DS.29 Consider the idea of a duty to protect and promote autonomy, in oneself and in others. Is there a general duty to refrain from taking risks that threaten, at some future time, to undermine this capacity, regardless of whether doing so would wrong any particular individual? It is conceivable that there is such a duty, which is independent of any duty to preserve connectedness over time. If there is such a duty, then it might be argued that it is prudentially irrational to engage in DS (again, given a suitable tie between prudence and morality), regardless of any decay of connectedness that may occur over time. The claim of such a duty might be supported by an appeal to broadly Kantian considerations pertaining to the dignity of humanity, or of rational agency. Consider, for instance, Christine Korsgaard’s account in The Sources of Normativity, according to which a person’s human identity requires that she value her humanity as an end-in-itself—in particular, that she protect and promote her rational capacities, not allowing them to be usurped by another. According to Korsgaard, a person may have multiple practical identities at a single time (e.g., parent, citizen, athlete), and each may generate certain practical requirements. However, demands from her perspective as a human being are ones that she is required to make, since her identity as human is not contingent and cannot be shed. For Korsgaard [38], there’s no escaping this normative requirement.

We can imagine, then, the perspective of an athlete’s distant future self, who suffers from the effects of a TBI brought about by her participation in DS, and who now has moralized regret and resentment toward her past self’s actions. These reactive attitudes may be justified from a Korsgaardian perspective. It seems the actions of the athlete’s past self have failed to treat the humanity of her future self as something which imposes substantial constraints on her present actions.

Note that a duty to preserve one’s rational capacity need not be understood as a duty to one’s future selves; it may instead be understood as a duty to oneself at a particular time, rather than a duty that holds across time ([17], Ch. 4). In particular, the duty might be conceived as holding between an individual at a time t and her own rational capacities at t. Indeed, if the duty to protect one’s rationality is to be understood as independent of duties to future selves, then it seems it should be understood as synchronic, rather than diachronic.30

However, once such a duty is understood as synchronic, it becomes less plausible that paradigmatic instances of the pursuit of DS violate it. An athlete who fully understands the risks that are involved in her pursuit of DS, and who prepares conscientiously to face them, is exercising her rational capacities well, perhaps as well as she knows how. She understands further that by pursuing DS she is also putting this very capacity at significant risk of great harm. However, is it plausible that this constitutes a failure of self-duty, a failure of respect for her own autonomy? Or that in putting her rational capacities at such risk, she is failing to treat her humanity as an end-in-itself? I suggest this claim is far less plausible considered in relation to a single temporal perspective than in relation to a cross-temporal demand. It may be conceivable that in making effective use of my faculties now I thereby fail to respect certain future individuals whose existence I have a role in creating. However, the thought that I am failing to show proper respect for my (present) rational nature is difficult to accept. After all, I may be intent on exercising my capacities as well as I can.

3.4.2. Consensual Domination and Risk to Future Autonomy

The claim that it is even possible to exercise one’s rational capacities fully and conscientiously in pursuing DS might be challenged. In particular, it might be argued that DS is analogous to consensual domination in the way it involves putting one’s rational capacities at risk of harm [39,40]. Specifically, it might be argued that the decision to participate in DS involves voluntarily foregoing future autonomy in much the same way that consenting to being a slave does, and as such it is contrary to the duty to respect one’s own autonomy [41]. Sailors, for instance, writes:

Given what we know, and are learning, about CTE, choosing to play football is analogous to choosing to be sold into slavery, since choosing football means choosing the likely brain damage that makes later autonomous choice equally impossible. ([39], 271)

Now, it might be argued that people are rationally permitted to forego their future autonomy, provided their decision to do so is the result of rational deliberation and careful reflection on their values [42]. However, setting this aside, there are a couple of respects in which the decision to engage in DS is not analogous to standard cases of consensual domination. First, in voluntarily participating in DS, the athlete does not necessarily consent to undergo brain damage and consequent loss of autonomy; at most, he consents to a certain risk of this bad outcome. This is important because, in voluntarily taking the risk of losing your autonomy, you do not thereby accept that you will suffer such a loss. It is reasonable to withhold such acceptance, after all, since such a loss is not, for you, foreseeable.31 Compare the case of a healthcare worker who, in providing treatment to desperate patients who have an infectious and potentially fatal disease, risks contracting the disease herself. Such an individual does not consent to becoming ill, or to dying, due to her interaction with the patients, although there is certainly a significant risk that these events will occur. Similarly, a woman who consents to unprotected sex with a man does not necessarily consent to becoming pregnant, even if she does so without requiring that he use contraception and understands that there is a significant risk of this outcome.

The second point is that in the case of voluntary participation in DS, there are significant goods at stake, for the sake of which the athlete is willing to take serious risks. This is an important respect in which DS is unlike consensual domination, which does not yield anything of value. Granted, it might be argued that as long as the decision to sell oneself into slavery reflects one’s considered preferences, there is something of significant value here—the decision has, we might say, personal value. However, I think it is not clear that this would be a genuine instance of personal value, since it is not clear that it is possible for you to take a valuing attitude toward the state of affairs in which you are enslaved—even if you can desire that it obtain (cf. Dorsey [23], Ch. 5). I acknowledge that this is a substantive claim a full defense of which requires an account of the nature of the valuing attitude. While I do not offer one here, I suggest that such an account should distinguish the attitude of valuing $\varphi$ from that of desiring $\varphi$; and it should treat the valuing attitude as having normative significance.32 In light of this, it is not clear that you can truly value your own enslavement. If the only thing of value in this case is that you get what you want, then I think the case is significantly disanalogous to DS which, ideally, involves achievement, as well as other important values, such as the development of virtue and self-affirmation (cf. [47,48]).

3.4.3. Autonomy as Demand

Finally, it might be objected that the choice to pursue DS cannot be regarded as respectful of the duty to protect one’s autonomy, since it clearly involves putting one’s rational capacities at great risk of being impaired. However, on at least one understanding of autonomy—what Stephen Darwall [18] refers to as ’autonomy as claim or demand’, in contrast with “autonomy as benefit”—what is fundamental is the duty to respect a person’s right to autonomy, rather than the duty to protect autonomy (where respect is understood in the sense of recognition or acknowledgement33). Darwall’s view is that the right to autonomy is best understood in terms of an authority to make a kind of claim or demand on others that requires taking up a second-person standpoint. This irreducibly “second-personal” authority is part of what Kantians think of as the dignity of persons.34 (’[A]n important part of autonomy’s value…involves the right, claim, or authority that persons have to demand that they be allowed to make their own choices and lead their own lives.’ ([18], 268)). On this understanding, any duty you may have to protect your autonomy is plausibly secondary to a more fundamental duty to respect autonomy-as-claim-or-demand, as an end-in-itself, whether in others or in yourself.

The duty to protect autonomy has substance, it seems, only if respect for an individual’s autonomy at t requires its protection at t. The case of DS suggests that this does not hold in general.

4. Does the Parfitian Defense Prove Too Much?

4.1. On Rationalizing Generally Reckless Behavior

Suppose that Parfitian discounting, taken together with the fact that there are significant prudential goods at stake, is sufficient to rationalize pursuit of DS. There is a further worry. Namely, why does not the Parfitian defense rationalize all manner of reckless behavior, in addition to the pursuit of DS? Does the rationality of discounting future well-being make it rational for you to to take similar risks in the pursuit of other activities, such as taking dangerous drugs like meth and heroin, just for fun? What is so special about DS? The answer, I think, has to do with the fact that whether something has prudential value is an objective matter—objective, in the sense that it depends on certain attitude-independent value facts.35 Provided DS differs from dangerous drug-taking in terms of the attitude-independent goods it involves, DS can contribute to the goodness of a life in a way that reckless activities such as dangerous recreational drug-taking cannot.

Which attitude-independent goods differentiate DS from such reckless pursuits? I suggest that they include the fact that DS has significant potential to cultivate virtues such as resilience and humility. Add to that the fact that DS affords opportunities for self-affirmation, i.e., for confronting and extending the apparent limits of our capacities [47]. In addition, there is the fact that an athlete’s pursuit of DS can result in great achievement, which, as well as being significant in its own right, can inspire awe in others.

The rationality of future discounting is not sufficient for the prudential rationality of present risk-taking. In addition, there must be some prudential good at stake. The pursuit of alpine mountaineering, for instance, must have value for the person who does it, in order for such pursuit to be prudentially rational for her. An account of prudential value may be able then to distinguish between these cases. It may explain why DS has prudential value, and why various other dangerous activities lack such value. For instance, such an account might take prudential goods to be things that are, necessarily, valued36; and it might hold that DS differs crucially from these other activities in that the former, but not the latter, is valued by its participants.

Here I appeal to Dale Dorsey’s [23] theory of prudential value. Dorsey argues for a subjectivism about prudential value, according to which a necessary feature of prudential goods (i.e., goods that contribute to one’s personal good or well-being) is that they are valued, under the right conditions, by the agent in question. He thus defends the following principle:

Good-Value Link: for any object, event, state, etc., $\varphi$ and agent x, $\varphi$ is good for x only if and (at least in part) because $\varphi$ is valued, under conditions c, by x. ([23], 80)

The Good-Value Link imposes a substantive condition on prudential value. In order for $\varphi$ to be prudentially valuable for x, it is not enough that x has some pro-attitude or other to $\varphi$. (It is not enough, for instance, that x desire $\varphi$.) x must take a valuing attitude toward $\varphi$. Although it is not entirely clear what having such an attitude consists in (and there is plenty of disagreement in the literature on this point), it is clear at least that valuing $\varphi$ cannot simply be identified with a desire for $\varphi$. Moreover, it is plausible that there is a cognitive component of the valuing attitude. On Dorsey’s account, for instance, for x to value a particular state, object, or event (in a prudential way) is for it to be the case that x would believe that it is intrinsically good for her assuming her attitudes are coherent and fully considered ([23], Ch. 6).

If this theory is at least roughly correct, then generally reckless risk-taking does not contribute to an individual’s personal good just because she wants these things. Moreover, it may not even be possible to take a valuing attitude toward these activities. It is plausible that provided your attitudes are coherent and fully considered, you will not judge that these activities are intrinsically good for you.

I am suggesting that the athlete who pursues DS values the sport, whereas the reckless risk-taker does not count as valuing either the risky activities he engages in or their intended outcomes. Consequently, DS can be good for the athlete who pursues it, whereas reckless risk-taking for fun cannot be good for the individual who pursues it. Note that this is not because the athlete is alienated from his desire to take such a risk (cf. [53]). Whether or not he disavows his first-order desire, he may lack the belief that it is good for him to fulfill it. Hence, he need not value the activity. Further, unless he values it, the activity will not have prudential value for him. In addition, I should note that I think we can reasonably allow that, although they are rare, there are cases where dangerous recreational drug-taking is truly valued by the agent (because he believes, justifiably, that it is good for him, for its own sake); and that in such cases, discounting future well-being might rationalize this sort of risky behavior. Note that this is consistent with the claim that in fact most cases of drug abuse are not ones in which the agent truly values the activity. Moreover, it is consistent with the claim that in most cases where he does value it, he would not value the drug under favored or idealized conditions.

In sum, the Parfitian defense does not imply that any sort of dangerous risk-taking can be prudentially rational, because it does not imply that a risky activity has prudential value provided it has some appeal for the agent. The Parfitian defense is consistent with the idea that some sorts of risk-taking involve prudential goods, whereas others do not. Moreover, it is consistent with the idea that whether a pursuit has prudential value depends on whether the agent would have pro-attitudes toward it under idealized counterfactual conditions.

4.2. Exploitability and Discounting

Finally, I turn to the worry that discounting future well-being on Parfitian grounds makes one exploitable—specifically, Arif Ahmed’s [7] recent argument. In order for Parfitian discounting to rationalize participation in DS despite its serious risks, the discounting must itself be rational. Thus, if it is irrational due to the fact that it entails exploitability, then the Parfitian defense of DS fails.

Ahmed’s argument is quite rigorous and sophisticated, and I cannot go into it in detail here. Fortunately, it will not be necessary to do so. For my purposes, the following brief outline of Ahmed’s case will suffice. There are two main steps. First, Ahmed argues that Parfitian time preference is delay-inconsistent. That is, one who discounts on the basis of future decayed connectedness will fail to treat delays of the same duration as having the same discount value.37 For example, suppose an individual, Al, discounts on such a basis. The claim is that there will then be delays, ($T_1$, $T_2$) and ($T_3$, $T_4$), such that $T_2$ − $T_1$ = $T_4$ − $T_3$, and such that V($T_1$, $T_2$) ≠ V($T_3$, $T_4$), relative to Al. Second, Ahmed argues that delay-inconsistency is irrational, because it leads to exploitability. Thus, Al, who is delay-inconsistent, will voluntarily make trades that make him worse off on some days and better off on none.

The argument for the claim that Parfitian time-preference is delay-inconsistent rests on the assumption that different psychological connections decay at different rates. For instance, vivid memories last longer than memories of everyday perceptual experiences; intellectual capacities change less frequently than certain other traits; and so on. Because of this, the proportion of direct connections that an individual maintains with her current state will decrease at a declining rate: connections that break quickly will initially make a big difference in terms of diminishing connectedness, and will make less and less of a difference as time goes on.38

Figure 2. Rate of increase in connectedness increases, then begins to decrease; then, connectedness begins to decay—at first quickly, then more slowly.

Figure 2. Rate of increase in connectedness increases, then begins to decrease; then, connectedness begins to decay—at first quickly, then more slowly.

This means that if Al discounts simply on the basis of connectedness, the rate at which he discounts will depend on when the relevant delay occurs. In particular, the discount value Al assigns to a given delay will increase with increasing futurity of the delay. (Al is at any time more patient about delays that occur in the relatively distant future than he is about delays that occur in the near future, i.e., he is willing to exchange a greater number of units of well-being at $T_1$ for extra units of well-being $T_2$, the farther in the future $T_1$ and $T_2$ are.) This makes Al liable to exploitation, provided he discounts at the same rate no matter which time happens to be present.

The latter principle is known as stationarity. It states that when evaluated at different times a delay of fixed length and fixed distance from the index of evaluation creates the same discount value. In other words, stationarity requires that the discount value is a function only of a temporal interval measured from the present and is not sensitive to which time is now. Thus, one who discounts on Parfitian grounds will, at any time, take the same attitude towards her well-being at the same distance in the future. For instance, if Al’s discount function obeys stationarity, then, if on January 1 he is indifferent between 1 additional unit of well-being on January 1 and x additional units of well-being on February 1, then on February 1 he is indifferent between 1 additional unit of well-being on February 1 and x additional units on March 1.

Given stationarity it evidently follows that one who discounts on Partifian grounds will be exploitable.39 How worrying is this? In response, I think it is important to note, first, that future bias40—apart from any discounting—appears to lead to the same kind of exploitability, at least when combined with risk aversion ([54]). While we might conclude that future bias is, like near term bias, irrational, the fact that this would undermine a very wide range of preferences that people commonly do not regard as involving any sort of error should give us pause.41 An error theory should not be ruled out. However, neither should we be too quick to dispense with the Parfitian defense because of a problem that arises not only for Parfitian discounting but for alternative views of time preference as well.

Now, Ahmed notes that it would be possible to avoid exploitation, while remaining delay-inconsistent, by using a non-stationary discount function. However, he argues that such a function is unmotivated:

[Non-stationarity] makes [Al’s] present evaluation of a future delay depend not only on the futurity and length of that delay but also on what date it is now. More specifically, her rate of time-preference at any time is a function of the date t—typically a declining function, so that the longer she lives the more patient she gets. But why should her past longevity have normative bearing on her present concern for her future self? Why should it be a demand of rationality that (for instance) Alice on 1 January 2016 cares less about her consumption on 10 January than Alice on 1 June 2017 cares about her consumption on 10 June 2017? ([7], 253)

Here I think Ahmed raises a very interesting question about time preference (despite its rhetorical intent), as I do not think it is obvious that one’s past cannot have normative bearing on one’s concern for one’s future self. First, suppose Ahmed is right that a non-stationary discount function cannot be a demand of (prudential) rationality. This is consistent with its being rationally permitted (cf. [56], 195). As long as it is not prudentially irrational to adopt a non-stationary discount function, such that one becomes increasingly patient (hence, increasingly willing, as time passes, to forego increases in present well-being for the sake of increases in later well-being), one can avoid exploitation consistently even while being delay-inconsistent.

Moreover, although it might be objected that there is no apparent rational basis for adopting such an attitude, there are arguably grounds from my development of the Parfitian point of view for rejecting stationarity—that is, for regarding non-stationarity as a demand of rationality, given contingent facts about decay of connectedness over time. If the relationship between the strength of one’s duty to preserve R is related to the decay of R in the way I have suggested (§3.3.2), then we have an explanation for why it is rational to discount in a non-stationary way. If it becomes more important to preserve direct connections, the more they decay—up to a certain point—then—even if such decay occurs at a declining rate, as Ahmed argues—one should discount one’s future well-being at a declining rate (up to that point). That is, one should care increasingly more about one’s nearer future well-being, as time passes (where such well-being is understood to depend in part on connectedness to prior selves), and as decay increases. The rational discount rate thus seems to depend on which time is now.

In sum, the worry about exploitability faces two difficulties. The first is that exploitability seems to arise independently of near-term bias, and thus independently of considerations of rational discounting of future well-being. As such, it does not seem to be a problem specifically for the Parfitian defense. The second problem is that the case for exploitability in connection with the Parfitian defense rests on stationarity, which the proponent of the Parfitian defense may have grounds for rejecting. Finally, even if the latter point is not granted, it looks as though it may be sufficient for the Parfitian defense that non-stationarity is rationally permissible. As it is not clear why stationarity should be rationally required, this puts the burden on the critic of the Parfitian defense.

5. Conclusions

The Parfitian defense of DS is based on the idea that it is rational to discount distant future well-being, due to the fact that psychological connectedness to one’s present self decreases over time. Because it is rational to discount one’s distant-future well-being, one can be rational in putting one’s distant-future self at risk for the sake of certain goods in the present, including those involved in DS. I have considered a number of challenges to the Parfitian defense, and I have tried to show how the proponent of this defense of DS can respond to each of them. First, I have argued that even if there is some reason to think that later stages of a life have more bearing on its overall goodness than early stages, this does not mean it is prudentially irrational to impose significant risks to one’s future well-being. Buchakian considerations regarding the shape of one’s risk function are relevant to the question of whether it is in one’s best interest to impose such risks, and therefore, they are relevant to the question whether it is prudent for one to do so. It is not clear why the range of acceptable attitudes towards risky choices and actions should not include those involved in DS.

I have also argued that there are difficulties facing challenges to the Parfitian defense based on self-directed duties. In effect, I have posed a dilemma for such challenges: either self-directed duties are diachronic, or they are synchronic. If such duties are diachronic, then arguably duties to distant future selves are too weak to undermine the Parfitian defense, because as psychological connectedness diminishes, so do bonds of concern, and as bonds of concern diminish, so does the strength of one’s duty to preserve connections between selves. On the other hand, if self-directed duties are synchronic, then there is reason to think DS does not necessarily violate them. Pursuit of DS may involve careful and deliberate use of one’s rational capacities (including careful risk assessment and mitigation). Challenges from self-directed duties thus do not show that DS is prudentially irrational.

Finally, I have discussed challenges to the sufficiency of the Parfitian defense. I have argued that the objection that it would rationalize much more than DS, including activities that are clearly imprudent, is unfounded. The Parfitian defense does not rationalize reckless risk-taking for fun, because, unlike DS, such activities lack significant prudential value. Moreover, the exploitability worry about Parfitian discounting can be addressed, provided non-stationarity is rationally permissible.

I conclude that the Parfitian defense of DS remains tenable. While there may be further challenges to the prudential rationality of DS, I suggest that they should engage with this line of defense.