4.1 A Response to the problem of prudential neutrality

In Faith, the desire to become an astronaut is presumed to be the only actual desire of Faith's that is relevant to whether she prudentially ought to take the pill. I think that, in addition to this desire, Faith is likely to have a variety of desires that are relevant but hidden, in the sense that they are easy to overlook on the basis of how the case is described. These desires include a desire to retain the desires that are central to one's identity, a desire not to change one's psychology by artificial means, and a desire to live a meaningful life. If Faith has any of these hidden desires, then choosing to undergo the last set of tests to become an astronaut would satisfy not just the desire to become an astronaut but also some of the hidden desires, whereas choosing to take the pill would only satisfy the desire to be an Air Force pilot. Thus, if Faith has at least one of the hidden desires, she has stronger prudential reason to undergo the last set of tests to become an astronaut because that is the option that would make her better off. This solution therefore allows desire theorists to avoid claiming that Faith prudentially ought to be neutral between the two options.

When people perform thought experiments, they bring in their own way of thinking about the world and implicitly assume that agents in the thought experiments would think about the world in the same way. When people read about Faith, they think about what they would desire if they were in her situation and implicitly assume that she would have the same desires. As I will argue, people tend to have a desire to retain the desires that make them who they are, a desire not to change their psychology by artificial means, and a desire to live a meaningful life, and it is reasonable to assume that people are at least implicitly aware of these desires in themselves. In addition, I take it that desires can be merely dispositional, as most desire theorists would agree (Schroeder Reference Schroeder2015): if a person is disposed to act in ways that she believes will bring about p or is disposed to feel enthusiasm or attraction for p when thinking about p (or revulsion toward ~p when thinking about ~p), then she has a desire for p. So, as long as people have the dispositions that constitute the relevant hidden desires, they can be said to have these hidden desires. And as long as people are at least implicitly aware of these dispositions in themselves, they are at least implicitly aware of having these desires.

The first desire that people might attribute to Faith is a desire to retain desires that are central to one's identity. It is not true in general that whenever one has a first-order desire, one also has a second-order desire to retain that desire: you might want to eat potato chips without wanting to keep wanting this, for example. But we do have (and implicitly recognize ourselves as having) a special set of desires that we want to keep: the desires that shape our identity and make us who we are. A lifelong fan of the Boston Red Sox, for whom being a Red Sox fan is a core part of her identity, not only wants the Red Sox to win but wants to continue wanting this: she is disposed to feel unhappy at the prospect of losing this desire and to refrain from behaving in ways that she thinks would extinguish it. A person whose love of poetry figures prominently in his self-conception not only wants to read poetry but wants to keep wanting this, since he would, in an important sense, no longer be who he is if he were to stop wanting it.Footnote ⁶ Since Faith is described as having ‘long desired to become an astronaut’, and since people's long-held career aspirations are typically part of their identity, it is natural to suppose that Faith's desire to become an astronaut is central to her identity: if she were to lose this desire, then an important part of who she is would be gone. For this reason, it is natural to suppose that Faith desires to retain this desire.

The second desire that people might attribute to Faith is a desire not to change one's psychology by artificial means. We all go through natural psychological changes as we age and have new experiences, and we are usually not disturbed by these changes. When psychological changes are caused by pharmacological interventions or other artificial means, however, we do have a tendency to be disturbed by them – a tendency that is constitutive of a desire not to undergo psychological changes via such means. For example, you would not be concerned if you've come to appreciate classical music more and more because you've been hanging out with friends who play it for you, but you would be at least somewhat concerned if your appreciation for classical music were caused by a drug that you've been taking. Consider, too, our preferences regarding how mental illnesses are treated. If you were to have a mental illness that could be cured either by talk therapy or by taking pills that modify the chemicals in your brain, you would probably prefer the former cure to the latter even if the pills have no side effects. This is further evidence of our desire not to have our psychologies changed artificially. Whether such a desire can be justified and what it is for an intervention to be artificial in the relevant sense are, of course, interesting open questions. But what matters for our purposes is that this is a desire that can plausibly be attributed to Faith and that would clearly be frustrated if she were to take the pill.

The third desire that people might attribute to Faith is a desire to live a meaningful life. It is a familiar fact that people typically want their lives to be meaningful: they are disposed to be unhappy at the thought that their lives are meaningless, excited about the prospect of making changes that would make their lives more meaningful, motivated to make such changes, and so on. Kauppinen (Reference Kauppinen2012) gives a theory of meaningfulness that I find compelling. On this account, what we mean when we say that a person's life is meaningful is that it is fitting for the person to be proud of it or to be fulfilled by it, and it is fitting for others to be inspired by it or to admire it. The kind of life that warrants these feelings, according to Kauppinen, is a coherent life in which the person sets goals, makes efforts toward achieving them, and succeeds in achieving them. For example, a life in which a person comes from a disadvantaged background but, through her determination and hard work, comes to have a fulfilling career is a meaningful life. On the contrary, a life in which a person fails to achieve her goals or in which her goals come about due to sheer luck is not as meaningful. If we accept Kauppinen's theory of meaningfulness, then the life in which Faith becomes an astronaut is much more meaningful than the life in which she takes the pill and stays a pilot. In the former life, she aspires to become an astronaut, works hard toward achieving that goal, and finally becomes what she has long wanted to become. It is a life that is fitting for Faith to be proud of and for others to admire. In the latter life, Faith aspires to become an astronaut, works hard toward achieving that goal, but gives up on it at the last minute for no apparent reason. It would be far less fitting to be proud of or admire this life, so this life would be far less meaningful. The life in which she becomes an astronaut would better satisfy her desire to have a meaningful life.

Faith does not need to have all the hidden desires that I have just laid out to have stronger reason not to take the pill. Any one of them would suffice to break the symmetry between the two options. As long as it is plausible that Faith has at least one of those desires, desire theorists can plausibly say that Faith has stronger reason not to take the pill. Since the hidden desires that are likely to be present in Faith are also likely to be present in other cases that display the problem of prudential neutrality, my response is not just a solution to Faith but also a solution to the general problem of prudential neutrality.

One might object to my response in the following way. We can imagine that, for each of the hidden desires that Faith currently has, the pill would give her a ‘counter-desire’ of the same strength that would be satisfied if and only if she were to take the pill. If Faith desires to retain her desire for becoming an astronaut, then the pill would give her an equally strong desire to eliminate her desire to become an astronaut. If Faith desires not to change her psychology by artificial means, then the pill would give her an equally strong desire to change her psychology by artificial means. If Faith desires to live a meaningful life, then the pill would give her an equally strong desire to live a meaningless life. Since, for each hidden desire of Faith's that would be frustrated if she were to take the pill, taking the pill would give her a satisfied desire the value of whose satisfaction cancels out the disvalue of the frustration of the hidden desire, it might seem that my solution implies that, in this version of the case, Faith prudentially ought to be neutral between her two options. But intuitively, even in this version of the case, Faith prudentially ought to become an astronaut instead of taking the pill.Footnote ⁷

My response is that, far from conflicting with this intuition, my solution supports it. We can see this by comparing simple models of the two versions of the case. Suppose that Faith has only one hidden desire (the desire to retain her desire to be an astronaut, let's say), that this desire has a strength of 3, and that her desire to be an astronaut has a strength of 5. In that case, we can represent the original version of the case as follows – ignoring, for simplicity's sake, all of her other desires, which will be the same no matter which option she chooses.

By contrast, the new version of the case that we are now considering can be represented like this:

Although the disvalue that Faith would get from the frustration of her hidden desire is exactly canceled out by the value that she would get from the satisfaction of the ‘counter-desire’ that the pill would give her, and although this means that the difference in welfare between the two options is smaller than it is in the original case, it remains true that Faith would be worse off if she were to take the pill than if she were to become an astronaut. Thus, my solution has the intuitive implication that she prudentially ought to become an astronaut.

Of course, one could imagine another variation on the case, in which each ‘counter-desire’ caused by the pill would be twice as strong as the hidden desire to which it corresponds. In this version of the case, Faith would be equally well off regardless of which option she takes, so my solution really would imply that she prudentially ought to be neutral between the two options. Perhaps some would have the intuition that, even in this version of the case, Faith prudentially ought to become an astronaut. I admit that my solution cannot accommodate this, but this strikes me as an acceptably small bullet to bite. After all, even if we think that Faith prudentially ought to be neutral, we can hold that she ought, all things considered, to become an astronaut (e.g., because she has a strong reason of self-respect to take her existing commitments seriously). Perhaps what's intuitively plausible about this case is merely that Faith ought not to be neutral, all things considered – a judgment that my response leaves open.

My solution differs from Dorsey's and is immune to the objection that I made against it. Dorsey assumes that the welfare scores are the same in the two options because (except for what the crucial desire is about) all else is equal between the two options. His solution relies on the rejection of the view that the strength of a prudential reason to realize an option is proportional to the welfare score you would get if you were to realize that option. It is his rejection of this view that made his solution vulnerable to the objection that I raised. My solution is not vulnerable to the same objection because I don't reject this view. My solution relies on the claim that the welfare scores are not the same in the two options because not all else is equal between the two options, given that some hidden desires are satisfied in the first option but frustrated in the second.

My solution does not deliver the result that Faith necessarily has less reason to take the pill, regardless of whether she has hidden desires that would be frustrated if she were to take the pill but satisfied otherwise. It says that she has less reason to take the pill only if she has such desires. If Faith has no hidden desires favoring either option, then my solution says that she prudentially should be neutral between the two options because in the absence of any hidden desires, these two options would give her the same welfare score. This seems to me to be the right thing to say. Alternatively, if, instead of having any hidden desires that count in favor of her not taking the pill, Faith has desires that count in favor of her taking the pill (e.g., a desire not to retain desires that are central to her identity), then my solution says that Faith prudentially ought to take the pill because taking the pill gives her a higher welfare score than not taking the pill would. This also seems to be the right thing to say.

Those who are unconvinced should consider a case in which a small child can get one of two flavors of ice cream. She wants chocolate and could get it, or she could replace that desire with an equally strong desire for vanilla and get vanilla. Supposing (as is plausible) that she has no hidden desires favoring either option, it seems clear that she prudentially ought to be neutral between them, since both would make her equally well off. But stipulating that Faith has no hidden desires favoring either of her options makes her case exactly analogous to this one. We should therefore conclude that, if Faith has no such hidden desires, she prudentially ought to be neutral between her options. I suspect that any inclination to resist this conclusion is due to the fact that the relevant stipulation is psychologically unrealistic.

4.2 A Response to the problem of remote desires

The desire-satisfaction theory entails that Parfit benefits from the satisfaction of his remote desire that the stranger be cured, but this strikes many people as implausible. I think that, if we examine Parfit's psychology more closely, we are likely to find that there is a hidden desire that accompanies his remote desire: a desire to know whether his remote desire is satisfied.Footnote ⁸ It is clear from Parfit's example that this desire to know is frustrated. The frustration of this hidden desire decreases Parfit's well-being to a certain extent, and this might mislead us to think that Parfit does not benefit from the satisfaction of his remote desire. That is, Parfit does benefit from the satisfaction of the remote desire, but the benefit is counterbalanced to some extent by the frustration of the desire to know. If we are not thinking carefully enough, we might conclude that Parfit does not benefit from the satisfaction of the remote desire at all. This response generalizes to other cases of remote desires since these cases all involve a frustrated desire to know whether the remote desire is satisfied. A person who desires that the number of atoms in the universe be prime cannot know whether her desire is satisfied given our limited understanding of the universe. A person who desires that Napoleon's favorite color was blue cannot know whether her desire is satisfied given that nothing was recorded on the matter. And a person who desires that people living centuries from now flourish cannot know whether that desire is satisfied given that we have no way of knowing that.

In most cases, when we desire something, we also desire to know whether that desire is satisfied. For example, if you desire that you are healthy, then you probably also desire to know whether you are healthy: you are probably disposed to act in ways that you believe will give you this knowledge, disposed to feel anxiety when lacking this knowledge, and so on. Likewise, if you desire that your mother is happy, then you probably also desire to know whether she is happy. Besides being confirmed by ordinary cases like these, this principle is also supported by the observation that there is a tight connection between caring about something and wanting to know facts about that thing. As Jennifer Hawkins (Reference Hawkins2014: 233) puts it:

If it is possible to desire something without caring about it, then these considerations cannot show that first-order desires are inevitably accompanied by second-order desires to know whether they are satisfied. But since desiring something typically does involve caring at least somewhat about it, these considerations support the weaker claim that I am making: that first-order desires are usually accompanied by such second-order desires. This claim is not only plausible but widely accepted, at least implicitly: when people are asked to imagine that someone desires that p, they implicitly assume that this person also desires to know whether p obtains – unless, of course, they are told something that suggests otherwise. So, when people read about Parfit's example, they are likely to assume that Parfit wants to know whether his desire is satisfied.

It is possible, though unusual, to have a desire without also wanting to know whether that desire is satisfied.Footnote ⁹ If people assume that Parfit is indifferent about knowing whether his desire that the stranger be cured is satisfied, then my response says that they should think that Parfit benefits from the satisfaction of the remote desire, and that he benefits more than he would if he had a desire to know. This seems to me to be the right thing to say.

It would be even more unusual to have a desire while also wanting not to know whether that desire is satisfied. But if someone assumes that Parfit for some reason desires not to know whether the stranger is cured while desiring that the stranger be cured, then my response says that they should think that he benefits from the satisfaction of the remote desire, and that he benefits more than he would if he were to desire to know (and also more than he would if he were indifferent about knowing). This also seems to me to be the right thing to say.

According to this response, remote desires are relevant to well-being, but they seem irrelevant to well-being because they are typically accompanied by other desires that are frustrated. For those who are inclined to say that remote desires really are irrelevant to well-being, it might be helpful to compare two versions of Parfit's example. In the first version, Parfit meets a stranger, learns that the stranger is ill, and forms a desire that the stranger be cured. Parfit never sees the stranger again, and unbeknownst to him, the stranger is cured. Crucially, Parfit neither forms a desire to know nor a desire not to know whether the stranger is cured. The second version is the same as the first version in every way except that Parfit does not form the desire that the stranger be cured. I find it plausible that Parfit's life goes better, even if it is just a little bit, if he were to have the desire that the stranger be cured than if he were not to have that desire, given that the stranger is eventually cured. If you share this intuition, then you should think that remote desires are relevant to well-being.Footnote ¹⁰

My response bites the bullet, just as Lukas's does, but it is an improvement on his in one crucial respect: it can make sense of why, in a case where Parfit has many strong desires about strangers whom he never meets again, he still does not seem to benefit much when these desires are all satisfied. Lukas has to say that Parfit benefits a great deal in this case because his response relies solely on the strengths and quantity of remote desires relative to the strengths and quantity of non-remote desires: he does not identify any desires that are typically frustrated when remote desires are satisfied and whose frustration cancels out at least some of the value of the satisfaction of remote desires. My response identifies desires of precisely that sort: it says that whenever Parfit has a remote desire, it is likely that he also has a desire to know whether that remote desire is satisfied, and the disvalue of the frustration of this desire cancels out at least part of the value of the satisfaction of the remote desire. Thus, my response implies that Parfit benefits less than Lukas's response implies that he does in the case in which he has many satisfied remote desires about strangers. It therefore does a better job than Lukas's response does on its own of explaining why we might mistakenly think that Parfit does not benefit at all in that case. More generally, it does a better job of making it credible that when we judge that people do not benefit at all from the satisfaction of their remote desires, we are mistaking a situation in which they benefit only modestly for one in which they do not benefit at all. When combined with Lukas's observations about the rarity and relative weakness of remote desires, my response makes it more credible than those observations do on their own that remote desires are, despite appearances, relevant to well-being.Footnote ¹¹

There is one way in which my response leaves something to be desired, however.Footnote ¹² Even if remote desires are typically accompanied by desires to know whether they are satisfied, it is possible for a remote desire to be far stronger than its accompanying desire to know. Your desire that people living centuries from now flourish might, for example, be much stronger than your desire to know whether they will flourish – as evidenced by your willingness to expend far more time, effort, or money to satisfy the former desire than to satisfy the latter. My response implies that, in a case like this, one benefits a great deal on balance from the satisfaction of the remote desire, since only a small fraction of the value of its satisfaction is canceled out by the disvalue of the frustration of the accompanying desire to know. Perhaps many of those who have the intuition that people do not benefit at all from the satisfaction of remote desires could be persuaded that people benefit somewhat from their satisfaction. But presumably, far fewer of them could be persuaded that people sometimes benefit a great deal from their satisfaction. For this reason, I am not prepared to declare victory over the problem of remote desires on behalf of the desire-satisfaction theory. It nonetheless seems to me that my response to the problem is the best one developed so far that does not require a departure from the standard version of the desire-satisfaction theory.

One last objection deserves to be mentioned. I have been assuming that Parfit's desire to know whether the stranger is cured begins to be frustrated at the same time that the desire that the stranger be cured begins to be satisfied: when the stranger is cured. It is for this reason that Parfit appears to benefit less from the satisfaction of his first-order desire than he does: the value of the satisfaction of that desire (say, 5) arrives simultaneously with the disvalue of the frustration of the desire to know (say, –3), with the result that his welfare score increases by 2 instead of 5. But it might seem that the desire to know begins to be frustrated at the much earlier time at which Parfit acquires it, since he does not know, at that time, whether the stranger will be cured. If this is correct, then the frustration of the desire decreases Parfit's welfare long before the satisfaction of the desire for a cure increases it, and the value of the latter cannot be obscured by the disvalue of the former.Footnote ¹³

My response draws on McDaniel and Bradley's (Reference McDaniel and Bradley2008) astute observation that our desires are typically conditional desires (i.e., desires for p on the condition that q) and that a conditional desire can neither be satisfied nor frustrated, regardless of whether its object obtains, unless its condition is met. For example, my desire to watch a movie tonight is not a desire to watch a movie tonight no matter what: it is conditional on my still wanting to do this when the night comes. If I still want to watch a movie tonight but I don't, then since the condition of my desire is met and its object does not obtain, my desire is frustrated. But if I no longer want to watch a movie when the night comes, then since the condition of my desire is not met, my desire is not frustrated but canceled.

The claim that Parfit's desire to know begins to be frustrated immediately when he acquires it presupposes that this desire doesn't have a condition that isn't met at that time. For if the desire does have such a condition, then even though the desire's object does not obtain then, the desire is canceled (or not yet active) rather than frustrated at that time. I think that, on the most psychologically realistic way of imagining the example, the desire does have such a condition: that there be a fact of the matter about whether the stranger will be cured. Surely Parfit doesn't want to know whether the stranger will be cured regardless of whether there is a fact of the matter about this: for if he did, he would want to know the fact of the matter about something regardless of whether there even is a fact of the matter about it. Even if we grant that people can have desires of this sort, they are surely rare and limited to cases in which the agent cares far more fervently about the matter at hand than Parfit does about whether the stranger will be cured (e.g., a case in which a diehard football fan wants to know who his team's starting quarterback will be).Footnote ¹⁴ But if Parfit's desire to know is conditional on there being a fact of the matter about whether the stranger will be cured, then since there is no fact of the matter about this when he acquires that desire, the desire does not begin to be frustrated at that time. Instead, it begins to be frustrated when there is a fact of the matter about whether the stranger will be cured – viz., the time at which the stranger is cured, or some slightly earlier time. Thus, the desire frustration from the desire to know does coincide temporally (at least approximately) with the desire satisfaction from the first-order desire and can therefore obscure the gain from that desire satisfaction.

Suppose, however, that Parfit's desire to know is not conditional in the way that I have just described. My response to the problem of remote desires would still have some force. This is because the satisfaction of his desire that the stranger be cured would still differ in one important way from typical cases of desire satisfaction: whereas in typical cases, the desire to know is satisfied when the first-order desire is satisfied and we simultaneously accrue some welfare from both desires, Parfit gets no welfare from his desire to know while he accrues welfare from his first-order desire. His welfare score increases by 5 (from the first-order desire), whereas in a typical case, his score would increase by 8 (5 from the first-order desire, 3 from the desire to know). The fact that Parfit does not benefit as much as he would in a typical case of desire satisfaction might mislead us into thinking that he does not benefit at all.