Desires and time are connected in several ways. Every desire is had at a particular time or during an interval, and our desires can change as our commitments, concerns and priorities evolve. Furthermore, desires can have times built into their objects. We might have desires for certain things to happen now (present-directed desires), to happen sometime in the future (future-directed desires), or to have happened sometime in the past (past-directed desires).
The ways in which desires and time are connected raise difficult conceptual questions. Parfit illustrates the issue as follows:
When I was young what I most wanted was to be a poet. . . . . . . Now that I am older, I have lost this desire. . . . . . . Does my past desire give me a reason to try to write poems now, though I now have no desire to do so?Footnote 1
This is a crucial question for those who would give an account of how one ought, from the perspective of rationality or prudence, to act at a given time.Footnote 2 More importantly for present purposes, the phenomenon of desire change poses difficulties for Desire Satisfactionism about well-being (or, equivalently, welfare). If the Parfit of today were to start writing poetry, this would satisfy the desire he had as a youngster. Since Desire Satisfactionism takes it (roughly) that what enhances well-being is for one to desire something and have it obtain, the satisfaction of young Parfit's desire would increase the total amount of well-being in his life. But some might find this to be an odd consequence, given that the Parfit of today – when the poems would be written – has no desire whatsoever to write poetry.
In this article, I will be concerned with these temporal problems for Desire Satisfactionism. In particular, I aim to clarify how the Actual Desire Satisfactionist – who thinks that well-being depends on the desires one actually has (as opposed to would have under ideal conditions) – should formulate her view so as to accommodate the connections between desire and time.Footnote 3 I consider a number of attempts to capture the temporal nature of desire, and I will argue that one of these, Weak Concurrentism, represents the most promising route for the Desire Satisfactionist to adopt. I consider the Desire Satisfactionist's other main options, but argue that none succeeds. This leaves Weak Concurrentism looking attractive. However, Weak Concurrentism might also be thought to have some implausible consequences of its own. Nonetheless, I argue that, on closer inspection, these consequences are not implausible at all – at least by the lights of a Desire Satisfactionist. My aim, then, is not to offer a full-blown defence of Weak Concurrentism, but rather only to defend a conditional conclusion: in so far as one is committed to Actual Desire Satisfactionism, Weak Concurrentism represents the best way to tackle the problems raised by the temporal nature of desires.
WHY TAKE TIME INTO ACCOUNT?
The Desire Satisfactionist must take time into account somehow. To show why, I will sketch a plausible version of Actual Desire Satisfactionism and explain the trouble it would face absent some attempt to accommodate the temporal nature of desires.
Actual Desire Satisfactionism takes it that well-being depends on the desires one actually has, as opposed to the ones one would have under ideal conditions.Footnote 4 On this view, what is intrinsically good for a person is to desire some state of affairs and have it obtain. Conversely, what is intrinsically bad for someone on this view is to desire a certain state of affairs and have it fail to obtain. The intrinsic value (disvalue) of such an episode of desire satisfaction (frustration) equals the product of (i) the duration of the episode and (ii) the intensity of the desire in question. The total amount of welfare in a life, then, equals the sum of the intrinsic values of the episodes of desire satisfaction in that life minus the sum of the intrinsic disvalues of the episodes of desire frustration in the life.
This basic view must be refined in a number of ways. First, it is generally agreedFootnote 5 that the view should take welfare to be enhanced only by the satisfaction of one's intrinsic desires – that is, the desires one has for things as ends, not merely as the means to other things. For instance, if one desires happiness for its own sake, and not merely as a means to something else (e.g. one thinks it makes one more charming), then this would be an intrinsic desire.
Second, the view should be restricted to say that one's welfare would not be enhanced by the satisfaction of desires for things that intuitively have no connection to one's life. Consider Parfit's well-known ‘stranger on the train’ case:
Suppose that I meet a stranger who has what is believed to be a fatal disease. My sympathy is aroused, and I strongly want this stranger to be cured. We never meet again. Later, unknown to me, this stranger is cured. On the Unrestricted Desire-Fulfilment Theory, this event is good for me, and makes my life go better. This is not plausible.Footnote 6
It would indeed seem implausible for the Desire Satisfactionist to allow one's welfare to be enhanced by the satisfaction of desires that bear such a tenuous connection to the events that intuitively make up one's life. Thus we would increase the plausibility of the view by not allowing such intuitively ‘irrelevant’ desires to bear on welfare.
There are other ways an Actual Desire Satisfactionist might want to restrict further the preferences that can impact welfare – say, by ruling out desires that seem trivial,Footnote 7 idiosyncratic, masochistic or immoral.Footnote 8 However, such restrictions are controversial,Footnote 9 so I will set them aside for present purposes.
With such refinements as these, Actual Desire Satisfactionism might appear quite attractive. Nonetheless, the view is defective as it stands because it fails to take time into account. The trouble stems from the fact that the view sketched so far employs an overly simple notion of desire satisfactions and frustrations. As sketched, the view takes it that what has intrinsic value for a person is any state consisting of one's actually desiring some state of affairs, p, and p's in fact obtaining. By contrast, what has intrinsic disvalue for a person is any state that consists of one's actually desiring that p obtain and p's failing to obtain.
The source of the trouble is that these simplistic definitions of desire satisfaction and frustration do not make any mention of times. As it stands, the view allows that if one has a desire, at time t, for a given state of affairs, p, then one will be harmed as long as there is some time at which p fails to obtain. Suppose that right now I have a desire (an atemporal one) for my book to be finished sometime, but not at any particular time. The state of affairs I now desire did not obtain when I was 15 years old. So the simple version of Actual Desire Satisfactionism sketched above seems to imply that I am thereby harmed. For here we have an episode of desire frustration, on the simple definition. There is a state consisting of (i) my desiring a certain state of affairs – viz. being done with my book – and (ii) this state of affairs not obtaining while I am 15. However, it is absurd to say that I am harmed by this state. It is entirely natural that one's book not be finished over a decade before one starts working on it. There is no harm in that. An exactly analogous problem arises in connection with desire satisfactions as well.
To have any plausibility, therefore, the Actual Desire Satisfactionist must take account of time. But what is her best strategy for doing so? This is the question I will address in the remainder of the article.
THE VIEW TO BEAT: WEAK CONCURRENT DESIRE SATISFACTIONISM
The trouble with a form of Desire Satisfactionism that fails to take time into account is that it allows one to be benefited (harmed) as long as there is any time that the objects of one's desires obtain (do not obtain) – even if this is not the same time that one holds the desires in question. Thus, the most natural way for the Desire Satisfactionist to avoid such problems is to say that what is good for one is to get what one wants want while one wants it, and what is bad for one is to fail to get what one wants while one wants it.Footnote 10 This is the idea behind a family of views we can call Concurrent Desire Satisfactionism.
However, there are several ways in which the basic concurrentist insight might be spelled out. In this section, I focus on the most straightforward version of the view: Weak Concurrentism. This is the view I will end up defending later on. However, as we will see, it has certain consequences that one might find problematic, and this motivates the search for other ways in which the Desire Satisfactionist might take time into account. Let us begin, therefore, by considering the view to beat and its prima facie problems.
Weak Concurrentism
Weak Concurrentism takes the concurrentist motto literally. On this view, one's welfare is enhanced if one desires, at time t, some state of affairs (even one with a time built in), and this state of affairs in fact obtains at t. Similarly, one's welfare is diminished if one desires, at t, some state of affairs (even one with a time built in), and this state of affairs fails to obtain at t. To formulate the view precisely, let me introduce two notions:
S receives an episode of concurrent desire satisfaction during the whole interval <t1, t2> iff for every time, t′, that occurs during <t1, t2>, it is true both that (i) S has, at t′, a desire that some state of affairs, p, obtain, and (ii) p in fact obtains at t′.
S receives an episode of concurrent desire frustration during the whole interval <t1, t2> iff for every time, t′, that occurs during <t1, t2>, it is true both that (i) S has, at t′, a desire that some state of affairs, p, obtain, and (ii) p does not obtain at t′.
Notice that this lets us get around the simple problem that refuted the atemporal version of Desire Satisfactionism sketched earlier. If, for instance, I now have a desire for the non-temporal state of affairs [my book is finished], we cannot say that my desire is concurrently frustrated just because my book was not finished back when I was 15. In order for this desire to be concurrently frustrated, it must be the case that the state of affairs I desire now fails to obtain now.
Given the notions of concurrent desire satisfaction and frustration, we can state the view as follows: Weak Concurrentism
(i) The intrinsic value for S of an episode of concurrent (intrinsicFootnote 11) desire satisfaction equals the duration of the episode times the average intensity of the relevant satisfied desire; the intrinsic disvalue for S of an episode of concurrent (intrinsic) desire frustration equals the duration of the episode times the average intensity of the relevant frustrated desire.
(ii) The total amount of welfare contained in S's life equals the sum of the intrinsic values of all the episodes of concurrent desire satisfaction in the life minus the sum of the intrinsic disvalues of all the episodes of concurrent desire frustration contained in the life.Footnote 12
The prima facie problem for Weak Concurrentism
Though I will ultimately argue that Weak Concurrentism is the best way for the Desire Satisfactionist to accommodate the temporal nature of desire, we must confront the fact that the view has certain consequences about time that some find problematic. These concern a widely discussed type of case, nicely illustrated by the following hypothetical of Brandt's:
[A] convinced sceptic who has rebelled against a religious background wants, most of his life, no priest to be called when he is about to die. But he weakens on his deathbed, and asks for a priest. Do we maximize his welfare by summoning a priest? Some would say not in light of his past desires.Footnote 13
HareFootnote 14, GriffinFootnote 15 and HeathwoodFootnote 16 offer structurally similar cases. The vignette from the introduction, involving Parfit's childhood desire to write poetry,Footnote 17 is another example of the same sort. What these cases have in common is that the person in the story begins with a future-directed desire Footnote 18 for some state of affairs (e.g. the desire not to be visited by a priest on one's deathbed), but when the time in question comes around, this desire is not matched by a corresponding present-directed desire.
What does Weak Concurrentism imply about such cases? Let's suppose that the atheist's initial future-directed desire (not to receive a visit from the priest) has the same average intensity as his subsequent present-directed desire (to receive such a visit). Moreover, suppose that the duration of the former desire is much greater than that of the latter. On these suppositions, Weak Concurrentism entails that the total amount of welfare contained in the atheist's life would be greater if no priest visits him than if he were to receive a visit from the priest. After all, if no priest comes to his deathbed, then he would receive one quite long episode of concurrent desire satisfaction, and another shorter episode of concurrent desire frustration. Given that the intensity of the two desires is assumed to be the same, the value of the concurrently satisfied desire to avoid a visit from the priest would outweigh the disvalue of the concurrently frustrated desire to receive such a visit. Thus Weak Concurrentism implies that in such a case, if we want to maximize the dying man's well-being, we should try to satisfy his original future-directed desire, not his subsequent present-directed desire. Bykvist calls this a ‘present-for-past sacrifice’.Footnote 19
Some philosophers find these implications counter-intuitive. Bykvist, for instance, says he ‘want[s] to avoid the present-for-past sacrifices illustrated by Parfit's poet example’.Footnote 20 Heathwood argues that if a theory has this sort of implication about cases like that of the dying atheist, this would be sufficient grounds for rejecting the theory.Footnote 21 For it might intuitively seem that the way to maximize the atheist's welfare would be to give him that which conforms to his current preferences – namely, a visit from the priest. If one were the man's close friend and wanted what is best for him, wouldn't calling a priest to his bedside be the thing to do? Those who think so may find Weak Concurrentism's implications about such a case hard to accept.Footnote 22
In my view, cases of this sort present a challenge for Weak Concurrentism, and I will tackle it head on in the last section of this article. However, we must also take care to formulate the challenge precisely. Given the brevity of Brandt's description of the case, there are some initial ways one might try to defuse the objection. I don't think they succeed, but we must be careful to formulate the objection as strongly as possible.
Two responses are particularly pressing.Footnote 23 First, one might argue that the atheist's desires in this case are merely instrumental, in which case their satisfaction or frustration would not bear on welfare. In particular, one might think the atheist originally desired not to have a priest visit him only because this would be the means to, say, a pleasant feeling of inner strength and independence of mind. Similarly, one might think he subsequently adopted the desire for a visit from the priest only because he comes to believe that it is the means to avoiding eternal damnation (or at least to obtaining the comfort of knowing he didn't decisively burn the bridge to heaven). If the atheist's desires in this case are instrumental, then Weak Concurrentism would not have the implication that some find implausible.
A second response appeals to the notion of a desire that is conditional on its own persistence. As several philosophers have pointed out,Footnote 24 sometimes one desires a given thing only on the condition that one still desires it when it is obtained. To take a simple example, suppose that in the afternoon one desires a hamburger for dinner, but only on the condition that this is still what one wants when dinnertime arrives. If one has stopped wanting the hamburger when one sits down to eat, then not receiving it would in no way frustrate one's desire. For what was desired was to [eat a hamburger provided one still wants it]. Since the condition built into the desire is not met, one is not harmed. Now, one might think that the atheist's long-standing desire not to have a priest visit his deathbed is similarly conditional on its own persistence. If so, then when he changes his mind later in life and has a priest called to the bed, this would not frustrate his earlier desire to avoid priestly visits. After all, the condition built into the desire would simply not obtain. Accordingly, calling a priest would not generate an episode of desire frustration. In this way, too, one might be able to block the perhaps troubling implications of Weak Concurrentism about the case in question.Footnote 25
However, I don't think the objection can be disposed of as easily as this. After all, these attempts to defuse the objection can be accommodated by adding more detail to the case. Most importantly, suppose that opposition to organized religion is a core feature of the atheist's identity up to his dying days. Liberating himself from religious influences is one of his central and long-standing goals. Making this supposition about the case has two benefits. First, it lets us construe the atheist's desire for the end of his life to be free of priestly visits as a desire that is not conditional on its own persistence. As Bykvist points out, ‘preferences that express ideals seem not to be conditional on their own persistence’.Footnote 26 Second, it allows us to construe this desire as intrinsic as well. Let us stipulate that the atheist desires not to be visited by a priest on his deathbed not because it is the means to something else he wants (e.g. happiness, pride, etc.); rather he desires it for its own sake because he sees it as part and parcel of remaining stoically opposed to religion. Thus the desire is intrinsic and would bear on his welfare. Moreover, we can make a similar stipulation about the atheist's sudden new-found desire to call a priest to his deathbed. Let us suppose that when faced with imminent death, he experiences a profound transformation and becomes deeply religious. Accordingly, he comes to desire a visit from the priest not because he thinks this is the means to avoiding eternal damnation, but rather because he sees it as an intrinsically worthy conclusion to a life. Thus we can avoid the worry above by telling the story in such a way that both the desires in question here are not merely instrumental.
When the case is described in this more nuanced way, we see that Weak Concurrentism cannot wiggle out of the challenge it faces in cases like that of the dying atheist. The view does indeed have the potentially problematic implication that what would maximize the atheist's welfare is a ‘present-for-past’ sacrifice. The view really would recommend satisfying the man's long-standing past desire, even though this would frustrate his current desires.
Note also that this is not a rare or idiosyncratic case. As we saw, a number of structurally similar cases have been suggested in the literature, and they can be strengthened in the very same ways as the case of the dying atheist. For instance, it would be easy to describe a version of the case involving Parfit's childhood desire to write poetryFootnote 27 in which the desire is not instrumental or conditional upon its own persistence, with the result that Weak Concurrentism would recommend a present-for-past sacrifice in that case too. Moreover, the cases suggested by Hare,Footnote 28 GriffinFootnote 29 and HeathwoodFootnote 30 can be brought into line in similar ways. Thus Weak Concurrentism has many implications of this kind.
Accordingly, one who finds present-for-past sacrifices to be problematic might prefer to seek an alternative desire satisfaction theory that does not recommend them. In the last section, I will argue that present-for-past sacrifices are in fact not as problematic as they might at first seem. Before that, however, we should investigate whether there are other formulations of Desire Satisfactionism that do not support present-for-past sacrifices.
OTHER FORMS OF CONCURRENTISM
Strong Concurrentism
Perhaps the most natural way to avoid these implications about cases like that of the dying atheist is to say that one simply receives no benefit in cases where one used to desire some future event and it later happens, but one no longer desires it when it does. Instead, the thinking goes, what is good for a person is to desire, at t, that something should happen at t and then have it in fact happen at t. This is the idea behind Strong Concurrentism. Its basic claim is that one's welfare is impacted only by the concurrent satisfaction or frustration of one's present-directed desires – i.e. desires that things happen now (at the time the desire in question is held). Hare seems to endorse such a view of welfare.Footnote 31 Bykvist and Heathwood also briefly mention views of this sort (though they find them implausible for various reasons).Footnote 32
Given the notions of concurrent desire satisfaction and frustration introduced above, it is easy to state Strong Concurrentism precisely. The view takes the bearers of welfare value (disvalue) to be episodes of what I'll call strong concurrent desire satisfaction (frustration):
S gets an episode of strong concurrent desire satisfaction during the whole interval <t1, t2> iff S has a present-directed desire, Dpr, throughout <t1, t2>, and Dpr is concurrently satisfied throughout <t1, t2>.
S gets an episode of strong concurrent desire frustration during the whole interval <t1, t2> iff S has a present-directed desire, Dpr, throughout <t1, t2>, and Dpr is concurrently frustrated throughout <t1, t2>.
Thus if I desire, between 1pm and 2pm this afternoon, to be skiing during this whole interval, and in fact get to ski for that whole time, then I receive an episode of strong concurrent desire satisfaction during the entire interval. By contrast, if I have this same desire, but do not get to do any skiing between 1pm and 2pm, then I receive an episode of strong concurrent desire frustration during this whole interval. What happens if I don't get to go skiing for this whole time, but, say, for just 45 minutes between 1pm and 1:45pm? In that case, I would still receive an episode of strong concurrent desire satisfaction – it would just be a shorter episode. In particular, I would receive an episode of strong concurrent desire satisfaction during the 45 minutes I am skiing, followed by an episode of strong concurrent desire frustration during the 15 minutes in which I want to be skiing but am not.
Using these two notions, Strong Concurrentism can be stated as follows: Strong Concurrentism
(i) The intrinsic value for S of an episode of strong concurrent (intrinsic) desire satisfaction equals the duration of the episode times the average intensity of the relevant satisfied desire; the intrinsic disvalue for S of an episode of strong concurrent (intrinsic) desire frustration equals the duration of the episode times the average intensity of the relevant frustrated desire.
(ii) The total amount of welfare contained in S's life equals the sum of the intrinsic values of all the episodes of strong concurrent desire satisfaction in the life minus the sum of the intrinsic disvalues of all the episodes of strong concurrent desire frustration contained in the life.
In cases like that of the dying atheist, Strong Concurrentism does not recommend present-for-past sacrifices. If one thinks this is the intuitively correct result, then one might think that Strong Concurrentism has an advantage over Weak Concurrentism. According to Strong Concurrentism, the future-directed desire that the dying atheist had during most of his life not to have a priest visit him on his deathbed would not bear on his welfare. After all, this is not a present-directed desire. What would matter to his welfare, however, is the present-directed desire he has while on his deathbed to have a priest visit him now. Thus if a priest does not visit him when he has a present-directed desire for one, he would receive an episode of strong concurrent desire frustration, and his welfare would be decreased. So Strong Concurrentism implies that the present-directed desire that the atheist holds while on his deathbed cannot be overridden by the earlier desires he held during the prime of his life.
Even if one thus finds the consequences of Strong Concurrentism to be more plausible than those of Weak Concurrentism, I think we should not endorse Strong Concurrentism. For it has other unacceptable consequences, which give sufficient reason to reject the view.
First, the view fails to account for the fact that our desires often are not directed at any time – past, present or future. For instance, I might desire that my life would make for a compelling narrative. Or I might desire that my life display a certain structure: say, adversity followed by success (as opposed to the other way around).Footnote 33 Or I might have an atemporal desire to lead a highly autonomous life in which others do not manipulate me. These global desires are not such that I want their objects to obtain at any particular time. Thus they are not present-directed desires, and their satisfaction or frustration cannot affect my welfare according to Strong Concurrentism. However, surely a Desire Satisfactionist would want to say that I would be made better off if I got such desires as these satisfied during my life.
Even more worryingly, Strong Concurrentism mistakenly implies that past-directed desires are entirely irrelevant to one's welfare. Consider the following case:
The college student who stuck it out
Jeremy hated every minute of college. At no time during his college tenure did he desire to be there. The whole time he desired to drop out and do something else. But he didn't. His parents threatened to punish him in various ways, so he grudgingly stuck it out and completed his degree. Suppose that in Jeremy's thirties, he becomes extremely thankful for his college education. He realizes it allowed him to attain professional success and the respect of his peers. He starts to see his education as an important source of personal pride and comes to attach a great deal of value to it (not merely as the means to other desirable things, but as an end in itself). Jeremy thus forms a past-directed desire to have received a college education in his youth, and he holds this desire for the rest of his life.
Since the satisfaction of past-directed desires cannot impact one's welfare according to Strong Concurrentism, the view implies that Jeremy receives no benefit whatsoever from his coming to value the fact that he went to college as a youngster. However, this is implausible. Anyone who is sympathetic to a Desire Satisfactionist approach to welfare should think that Jeremy's life goes better for him because of the fact that he comes to attach value to his college education. After all, the degree to which Jeremy's desires actually are satisfied in life is clearly greater than the degree to which they would have been satisfied if he had never come to appreciate his college education. If Jeremy, for the rest of his life, had continued to wish that he had never gone to college, then the total amount of desire satisfaction contained in his life would be much lower than the total amount of desire satisfaction that his life actually contains, given that he actually comes to value his college education. But this is something that Strong Concurrentism cannot capture.Footnote 34 Because Strong Concurrentism is incapable of accounting for the ways in which past-directed desires intuitively can affect one's welfare, it is an untenable formulation of the Desire Satisfactionist position.
A Strong Concurrentist Response?
The Strong Concurrentist might attempt to respond to this objection as follows.Footnote 35 In the case of the college student who stuck it out, one might think there is indeed a present-directed desire that is satisfied. Jeremy might be thought to have a present-directed desire now to be such that he has completed college. Thus the response involves a translation strategy: for every past-directed desire (or future-directed desire, for that matter) whose satisfaction or frustration intuitively seems to affect one's welfare, an analogous present-directed desire can be found that actually would affect that person's welfare according to Strong Concurrentism.
However, I do not think this response succeeds. In particular, it allows problems for Strong Concurrentism to re-emerge in cases like that of the dying atheist. Suppose the idea behind the translation strategy is correct – namely, that any past-directed desire that seems to affect welfare can be translated into a present-directed desire that actually can affect welfare on Strong Concurrentism. If this idea is accepted, then during all the years in which the atheist had a future-directed desire for no priest to visit him on his deathbed we should also say that the atheist held a present-directed desire to be such now that he will not receive any such visit at the end of his life. That is, the future-directed desire for no priestly visit that he held during the prime of life can in fact be mapped onto a present-directed desire that Strong Concurrentism would allow to affect his welfare. The result is that if a priest actually visits the atheist on his deathbed, the present-directed desire he held for all those years (i.e. to be such now that no priest will visit him) would count as frustrated. Accordingly, what would maximize the atheist's welfare is to have no priest visit – precisely the implication that the Strong Concurrentist hoped to avoid.
Thus if the Strong Concurrentist seeks to accommodate the case of the college student who stuck it out by adopting the translation strategy, then the view will end up recommending precisely the sort of present-for-past sacrifices in cases like that of the dying atheist that the view was designed to avoid in the first place. Accordingly, Strong Concurrentism faces a dilemma: either adopt the translation strategy and run into trouble with cases like that of the dying atheist, or else reject the translation strategy and face the problems raised by cases like that of the college student who stuck it out. In other words, the view is either unmotivated or false.
Discount Concurrentism
To avoid the problems with Strong Concurrentism, we might try to formulate a version of Concurrentism that allows for past-directed desires, future-directed desires and atemporal desires to count towards one's welfare, though at a discounted rate. Thus we might have a theory on which, all other things being equal, the satisfaction or frustration of a present-directed desire would affect one's welfare n times as much as the satisfaction or frustration of a past-directed, future-directed or atemporal desire would. I will not fully formulate the view, since it should be obvious how it would go.
Even though Discount Concurrentism does not suffer from the same defects as Strong Concurrentism, Discount Concurrentism still faces two problems that make it a less than ideal way to accommodate the temporal nature of desire. First, the theory does not solve the problems it was intended to solve. The motivation for abandoning Weak Concurrentism and seeking some other theory was that one might want to avoid the implications that Weak Concurrentism has in cases like the dying atheist. And at least at first sight, one might think that Discount Concurrentism can avoid such consequences. After all, on Discount Concurrentism, the atheist's desire later in life to have a priest visit him counts for more than his earlier future-directed desire not to get a priestly visit. So one might think the view implies that it would be best for the dying atheist to be paid a visit by the priest.
But on closer inspection, the Discount Concurrentist solution will not work. For we can just construct a version of cases like this in which Discount Concurrentism gives the supposedly counter-intuitive result.Footnote 36 Suppose that present-directed desires count for, say, twice as much as future-directed desires and past-directed desires do.Footnote 37 Moreover, suppose that the atheist holds, for a total of twenty years, his future-directed desire for his dying days to be priest-free. And then, on his deathbed, he has a change of heart and loses this desire, instead acquiring a present-directed desire for a visit from the priest. Although the welfare value of the satisfaction of the future-directed desire for no priest is discounted, the atheist still held it for two decades. So its satisfaction would provide a great deal more welfare value for the atheist than the satisfaction of the much shorter present-directed desire to get a visit from the priest – even though this present-directed desire is counted fully. Thus Discount Concurrentism, too, implies that the total amount of welfare contained in the atheist's life would be enhanced much more by having no priest visit than by having a priest come by. If one doesn't like the fact that Weak Concurrentism yields such results, then one should not look favourably on Discount Concurrentism either.
There is another, perhaps even more important, reason to reject Discount Concurrentism. In particular, it seems arbitrary to discount the contributions that past-directed and future-directed desires, as well as atemporal desires, make to a person's welfare. Why make Desire Satisfactionsm biased in favour of desires that are directed at the same moment the desire is held, as opposed to some time that comes later or earlier than the time at which it is held, or that is not directed at any time at all? There seems to be nothing particularly special about present-directed desires, which could provide reason to think that their satisfaction counts for more than the satisfaction of other desires. I can see no independent rationale for this kind of temporal bias.Footnote 38
Thus since Discount Concurrentism not only is unmotivated, but also contains an arbitrary temporal bias, it does not seem to be a promising route for the Desire Satisfactionist.
BYKVIST'S THEORY
In the previous sections, we have seen several attempts to make good on the idea that the temporal problems for Desire Satisfactionism can be resolved by taking it that what is good for you is to get what you want while you want it. Krister Bykvist has proposed a very different answer to the question of how one's well-being is impacted by the temporalized desires one might hold during life.Footnote 39 In particular, the idea behind Bykvist's solution is that since a person's temporal parts are properly regarded as autonomous in an important sense,Footnote 40 future-directed and past-directed desires will impact one's welfare only if a certain kind of ‘consensus’ exists between one's various temporal parts. Thus if you currently desire that something happen at some future time, but your future self at that time does not desire that thing, then there is no ‘consensus’ between your current and future selves. Accordingly, you would not be benefited if this future-directed desire of your current self were to be satisfied.
Stating Bykvist's view
The required kind of ‘consensus’ is not a matter of reaching an agreement or bargain. Rather, it is a matter of one's future-directed desires being matched by a corresponding present-directed desire. So consider, for example, the atheist's future-directed desire, which he held for all those years earlier in life, that no priest visit him on his deathbed. This desire would be matched in the relevant sense iff at the time specified in the object of the desire – the time he is on his deathbed – he also has a present-directed desire not to get a visit from the priest. Since the atheist's future-directed desire is in fact not matched in this way – i.e. since there is no ‘consensus’ of the relevant sort between his temporal parts – the satisfaction or frustration of his future-directed desire not to get a visit from the priest will not impact his welfare, on Bykvist's view. When it comes to past-directed desires, although Bykvist doesn't explicitly mention them, presumably something similar holds for them: viz. that a past-directed desire counts towards one's welfare iff it is matched by a present-directed desire.
This gives a rough picture of Bykvist's view. To state it precisely, consider Bykvist's own notion of a desire with full inside support. On Bykvist's theory, it is only the desires with full inside support whose satisfaction or frustration can impact one's welfare. Bykvist explains the notion of full inside support as follows:
my preference for something to happen at a certain time or during a certain time period t has full inside support just in case it is perfectly matched by those preferences that occur at or within t. More formally, my wanting, at t1, with intensity i, that p at t2, has full inside support iff for every time t within t2, either
(a) I do not have any preferences at t1, or
(b) I want, at t, with intensity I, that p at t2.Footnote 41
This definition implies that present-directed desires will always count as having full inside support. Since such a desire is about the very time at which the desire itself is held by the agent, these desires will count as ‘supporting themselves’ and thus will impact welfare. Next, future-directed desires typically need to be matched by an appropriate present-directed desire in order to have full inside support. Suppose I have a desire at some time t1 – call the desire ‘D1’ – that p obtain at some later time t2. If I still desire at t2 that p obtains at t2, then it follows that my future-directed desire D1 has full inside support. There is another way in which D1 could have full inside support, too, namely if I am completely out of commission (e.g. unconscious, dead) at t2.Footnote 42 Bykvist includes this way for a desire to get full inside support in order to ensure that the satisfaction of your desires for things that happen after your death can enhance your welfare.Footnote 43 Things work similarly when it comes to past-directed desires. Suppose I have a desire at some time t1 – call the desire ‘D2’ – that p obtain at some earlier time t0. D2 would have full inside support if it was the case, at the earlier time t0, that I also desired that p occur at t0. The other way in which D2 would count as having full inside support is if I were completely out of commission (e.g. unconscious, not born yet) at t0.
Bykvist endorses the thesis that when it comes to welfare, ‘a . . . preference counts iff it has full inside support’.Footnote 44 We can state a version of Desire Satisfactionism that conforms to this thesis. To state it simply, I will coin two terms:
S gets an episode of Bykvist-ian desire satisfaction during the whole interval <t1, t2> iff S has a desire, D, with full inside support during all of <t1, t2>, and D is concurrently satisfied during all of <t1, t2>.
S gets an episode of Bykvist-ian desire frustration during the whole interval <t1, t2> iff S has a desire, D, with full inside support during all of <t1, t2>, and D is concurrently frustrated during all of <t1, t2>.
Now we can state a version of Desire Satisfactionism that conforms to Bykvist's views: Bykvist-ian Desire Satisfactionism
(i) The intrinsic value for S of an episode of Bykvist-ian desire satisfaction equals the duration of the episode times the average intensity of the relevant satisfied desire; the intrinsic disvalue for S of an episode of Bykvist-ian desire frustration equals the duration of the episode times the average intensity of the relevant frustrated desire.
(ii) The total amount of welfare contained in S's life equals the sum of the intrinsic values of all the episodes of Bykvist-ian desire satisfaction in the life minus the sum of the intrinsic disvalues of all the episodes of Bykvist-ian desire frustration contained in the life.
Bykvist's view is designed to avoid the implications that a view like Weak Concurrentism has for cases like that of the dying atheist, Parfit the poet, and so on. In other words, the theory is supposed to avoid
the present-for-past sacrifice illustrated by the cases about Hare and Parfit. Other things being equal, it is never right to sacrifice a person's present preferences about her present life for the sake of her past preferences about her present life.Footnote 45
The reason that this is a desirable result, Bykvist thinks, has to do with considerations of autonomy:
I argued that the problem of past preferences arises when these preferences of past selves ‘poke their nose’ into present selves’ lives and private concerns. This way of putting things suggests that what is at issue is some kind of respect for the autonomy of person-stages . . . [E]ach person-stage has a veto over what they should do with their lifestage so any conflicting preferences of other persons or other temporal stages of the same person are to be completely disregarded.Footnote 46
Bykvist's idea here is that since each person-stage is autonomous, it would be misguided to allow the preferences of a particular person-stage to be overridden by the conflicting preferences of earlier or later person-stages.Footnote 47 Rather, ‘consensus’ between one's various person-stages about what is desirable is required in order for one's past preferences to impact well-being. Bykvist-ian Desire Satisfactionism is a theory that seems to give these results.
Against Bykvist's view
Nonetheless, I do not think that Bykvist-ian Desire Satisfactionism represents a good way for those who are sympathetic to a Desire Satisfactionist approach to welfare to formulate their view. There are three reasons for this. First, I think the intuitions that Bykvist has designed his theory to capture are misguided. I will argue in the last section of this article that the Desire Satisfactionist should allow that your total welfare would be enhanced by the satisfaction of desires that you held earlier in your life, even if you no longer hold those desires at the current time. Thus I will argue, for example, that there is nothing wrong with thinking that the total amount of welfare contained in the atheist's life would be enhanced more by declining to call a priest to visit him on his deathbed – even though such a visit is precisely what he wants at the end of his life. Accordingly, since Bykvist-ian Desire Satisfactionism is specifically designed so as not to give these results about cases like that of the dying atheist, I think Bykvist-ian Desire Satisfactionism is unmotivated. However, more on this in the last section.
What's more important, Bykvist's view doesn't fully capture the intuitions that it is designed to capture in the first place. Suppose for the moment that we do want our theory of welfare not to recommend what Bykvist calls present-for-past sacrifices (i.e. to not imply that your welfare can be maximized by giving you something you wanted in the past but which you don't want at present). Bykvist's theory fails to rule out such sacrifices because of a problem with his definition of ‘full inside support’. Consider the following modified version of the dying atheist case:
The case of the comatose dying atheist
Between the ages of 20 and 70, the atheist has a strong future-directed desire, D1, that no priest come to visit him when he is on his deathbed. When he is 70, the atheist's resolve weakens and he now forms a new desire, D2, that a priest visit him when he's on his deathbed. He holds D2 for a total of one week before he gets into a car crash and falls into a coma. The atheist is put on life support. He will not wake up. Thus the hospital bed is his deathbed. However, if he were to wake up, he would desire what he did right before the accident, viz. to have a priest visit him on his deathbed.
In fact, Bykvisti-ian Desire Satisfactionism implies that the total amount of welfare contained in the atheist's life would be maximized by having no priest come to give him his last rites. After all, both of the atheist's future-directed desires, D1 and D2, count as having full inside support. This is because the atheist has no desires whatsoever during the time towards which D1 and D2 are directed – i.e. the time during which he is on his deathbed. Thus the satisfaction or frustration of D1 and D2 would indeed have an impact on the atheist's welfare, according to Bykvist-ian Desire Satisfactionism. Because the duration of D1 is so much longer than the duration of D2, the theory implies that getting D1 satisfied would enhance the atheist's welfare much more than getting D2 satisfied.
However, I do not think Bykvist would want to accept this result. What Bykvist-ian Desire Satisfactionism recommends in this case is essentially the same as the sort of present-for-past sacrifice that Bykvist is so concerned to avoid. The theory implies that we could enhance the atheist's welfare significantly more by ignoring his most recent desires – which are also the desires he would have if he were to wake up from his coma – and instead satisfy his prior long-standing desire for no priest. If Bykvist is troubled by theories that recommend present-for-past sacrifices in the original dying atheist case (involving no coma), then he should be troubled by the implications that his own theory has about the present case, too. Thus it seems Bykvist's theory doesn't do the job it was designed to do.
Even more damagingly, Bykvist-ian Desire Satisfactionism has a range of unacceptable consequences. Recall that Strong Concurrentism was rejected because there were certain cases involving past-directed desires that it could not accommodate. Bykvist's view cannot accommodate such cases either. In particular, recall the case of the college student who stuck it out. The whole time Jeremy was in college, he desired not to be. Ten years later he comes to value his college education. Thus Jeremy forms a past-directed desire to have gone to college, but it is unmatched by any present-directed desire in the relevant time period. I claimed that desire satisfactionists should think that Jeremy is benefited by forming a desire after the fact to have gone to college. After all, this makes the overall degree of fit between his life as a whole and his desires higher. His life seems to contain a greater amount of desire satisfaction than it would if he hadn't formed any desire after the fact to have gone to college. But Bykvist-ian Desire Satisfactionism cannot yield this result. After all, Jeremy's past-directed desire to have gone to college does not have full inside support, since it is not matched by any present-directed desire. Thus its satisfaction does not count towards his welfare. The implications that Bykvist-ian Desire Satisfactionism has about this case are implausible and provide sufficient reason to reject the theory.Footnote 48
IN DEFENCE OF WEAK CONCURRENTISM
We have seen a number of unsuccessful ways in which the Desire Satisfactionist might try to deal with the temporality of desire. Other attempts to deal with these problems have been proposed, but I will not discuss them further here because they face well-known problems.Footnote 49 Rather, I will argue that Weak Concurrentism is in fact the best route open to the Desire Satisfactionist when it comes to accommodating the temporality of desire. This view does not have the unacceptable consequences about unmatched past-directed desires that led us to reject Strong Concurrentism and Bykvist-ian Desire Satisfactionism. Nor is it arbitrary or unmotivated in the ways that cast doubt on Discount Concurrentism.
Of course, Weak Concurrentism has challenges of its own to overcome. As we saw, the view might seem to be in trouble because it implies that we can benefit a person by giving him what he no longer wants. That is, the view recommends present-for-past sacrifices.
Nonetheless, my view is that these implications of Weak Concurrentism, which philosophers like Bykvist and Heathwood find troubling, are in fact not a sufficient reason to reject the view. I will argue that a Desire Satisfactionist, at least, should not think there is anything wrong with the idea that we can benefit people by giving them what they no longer want. My strategy thus will be to explain away the offending judgements about cases like the dying atheist.
Weak Concurrentism was supposed to be in trouble because it implies that the total amount of welfare contained in the atheist's life would be greater if no priest gets called to his deathbed than if a priest were called. But, the objection goes, it would intuitively be better for the atheist if he gets what he currently wants when on his deathbed, namely a visit from a priest. However, I do not think Desire Satisfactionists should be troubled by such results. The reason why one might find the implications of Weak Concurrentism about these cases implausible is that certain facts about happiness and unhappiness (or pleasure and pain, if you like) are contaminating one's intuitions.
When reading the story of the dying atheist, we imagine what it would be like to live through the experience of having a religious conversion at the end of one's life, forming a desire to receive religious comfort in one's time of dying, but then not getting it. This evokes images of the fear and anxiety that such an experience would probably involve. When we contrast this with the feelings of contentment and serenity that receiving a visit from the priest would probably entail, the prospect of the priest's visit seems more appealing. Granted this would frustrate the man's earlier atheistic desire not to receive such a visit, but since this desire is in the past, it would lead to no actual feelings of discontent. If this is the way to explain why the visit from the priest seems better for the dying man, then what is driving our intuitions about the case are assumptions about the happiness and unhappiness that he would feel in the two alternatives at issue. Thus the supposedly intuitive judgement that it would be best for the man's welfare to receive a priestly visit seems to be best explained by a tacit appeal to considerations about happiness and unhappiness. The same goes, I suggest, for the other structurally similar cases mentioned earlier.
However, if this is right, then the current objection to Weak Concurrentism would be based on general Hedonistic intuitions, not intuitions that are specifically concerned with the relation between time and the satisfaction of desire. Now, perhaps it is a good argument against Desire Satisfaction theories in general that they do not perfectly capture the intrinsic value that happiness seems to have for a person or the intrinsic disvalue that unhappiness seems to have for one.Footnote 50 But this would be an altogether different problem from the one we are concerned with in this article. When it comes specifically to the temporal nature of desires – and the resulting phenomena of desire change, future or past directed desires, etc. – Weak Concurrentism gives acceptable results.
This is apparent when the cases are described in such a way as to neutralize any tacit appeal to facts about happiness and unhappiness. The story of the dying atheist must be told in such a way that he receives the same amount of net happiness no matter whether a priest is called to his deathbed or not. The alternatives must be made hedonically equivalent, for it is only in this way that we can isolate our intuitions specifically about the relation between time and the satisfaction of desire.
Once the stories are construed in this way, however, the implications of Weak Concurrentism do not seem troubling any more. If the dying atheist will feel the same amount of net happiness whether or not he receives a visit from the priest on his deathbed, then it does not seem implausible to say that his total welfare would indeed be greater if no priest were to visit him, thus satisfying his prior ideal of remaining free from religious influence in his waning years.
In fact, it seems that this is what the clear-headed Desire Satisfactionist must say. After all, in light of the long-standing future-directed desire not to have a priest visit him on his deathbed, the total amount of desire satisfaction in the man's life would seem to be much greater in the case where he gets no priestly visit than the one where he does. Viewed as a whole, the man's life would seem to fit his various desires to a much higher degree in the possible world where he stays true to his principles and has no priest visit than in the possible world where he weakens and calls for a priest.
Perhaps one will object to what I have said here. If one feels strongly that these implications of Weak Concurrentism are problematic, one might object that even if the cases are described in the anhedonic way that I want, the results of Weak Concurrentism are still wrong. For instance, suppose we grant that a visit from the priest would produce the same amount of net happiness for the dying man as not receiving such a visit. Even so, one might argue, it is counter-intuitive to say that his welfare would be enhanced most by not getting a visit from the priest. After all, the objection goes, if you were the dying man's friend and you were entrusted with caring for him on his deathbed, and then you found out that he had completely lost his long-standing desire to remain free of religious influences, would you disregard that fact and refuse to call a priest for your friend? No. Even if you could effectively neutralize the happiness and unhappiness effects of both choices, any good friend would still try to comply with the man's current wishes and call a priest over. That is what someone who cared about him would do.
Two points are warranted in response to this objection. For one thing, it should be emphasized that even Weak Concurrentism implies that what would be best for the dying man as he is now, on his deathbed, is to receive a visit from the priest. The person-stage (if I may use this phrase without committing myself to any views about personal identity) present when the man is on his deathbed would clearly be benefited more by getting a visit from the priest than not getting such a visit. After all, that is what the man as he is now wants. And Weak Concurrentism does indeed yield this implication. (Of course, it also implies that what would maximize the total amount of welfare contained in his life as whole is to have no priest visit.)
Second, if one still has the intuition that what a good friend would do for the dying atheist (even in the anhedonic version of the case) is to arrange for a visit from the priest, then I would respond by claiming that this does not conclusively show what would in fact maximize the total amount of welfare in his life. For it does not seem that the only thing involved in being a good friend to someone, or in caring about that person, is seeking to maximize the total amount of welfare contained in this person's life. I think that, for contingent, purely psychological reasons, being a good friend to some person, P, or caring about P, is likely to involve a certain amount of temporal bias in favour of P as P is at the current moment.
Suppose I care deeply about my new wife, whom I met just two years ago. I know her as she is now, but not as she was many years ago. I see her as she is now. I interact with her as she is now. I have a very vivid image of who she is now, while I have but a dim inkling of who she was as a child. Thus it is understandable, psychologically speaking, that I would be much more concerned to do what is good for her as she is now (with whom I am intimately acquainted), as opposed to what would be good for her as she was in earlier times (whom I don't know well at all). For that reason, I might – precisely because I care about her – feel a strong (though temporally biased and perhaps misguided) inclination to do what maximizes her welfare as she is now, rather than what maximizes the total amount of welfare contained in her life as a whole.Footnote 51
I think this explains why some might think that what a good friend would do for the dying atheist (even in the anhedonic version of the case) would be to call for a priest. It is not because this is what would maximize the total amount of welfare contained in his life, but rather because it would maximize the man's welfare as he currently is, and his current friends are, for natural psychological reasons, quite likely to be biased in favour of the man's current self.
So my view is that the implications of Weak Concurrentism about cases like that of the dying atheist are in fact quite plausible, as far as the Desire Satisfactionist is concerned. If one is committed to being a genuine Desire Satisfactionist about welfare, then one must resist the Hedonistic intuitions that seem to have been behind the original judgements about the cases of the dying atheist and the rest. A genuine Desire Satisfactionist cannot reject Weak Concurrentism because it conflicts with these Hedonistic intuitions (i.e. because a genuine Desire Satisfactionist cannot credit these Hedonistic intuitions in the first place). And once this source of confusion is eliminated, the implications of Weak Concurrentism should seem perfectly acceptable by the lights of the Desire Satisfactionist.
Thus, given the inability of the other versions of Desire Satisfactionism to account plausibly for the temporality of desire, I think Weak Concurrentism represents the best way for Desire Satisfactionists to accommodate the ways in which desire and time are connected.
