This article addresses three questions about well-being. First, is well-being future-sensitive? I.e., can present well-being depend on future events? Second, is well-being recursively dependent? I.e., can present well-being (non-trivially) depend on itself? Third, can present and future well-being be interdependent? The third question combines the first two, in the sense that a yes to it is equivalent (given some natural assumptions) to yeses to both the first and second. To do justice to the diverse ways we contemplate well-being, I consider our thought and discourse about well-being in three domains: everyday conversation, social science, and philosophy. This article’s main conclusion is that we must answer the third question with no. Present and future well-being cannot be interdependent. The reason, in short, is that a theory of well-being that countenances both future-sensitivity and recursive dependence would have us understand a person’s well-being at a time as so intricately tied to her well-being at other times that it would not make sense to consider her well-being an aspect of her state at particular times. It follows that we must reject either future-sensitivity or recursive dependence. I ultimately suggest, especially in light of arguments based on assumptions of empirical research on well-being, that the balance of reasons favors rejecting future-sensitivity.
Getting perspective on the temporal character of well-being
Some days are better than others. Although this might be an observation about the variability of our day-to-day circumstances, it is also an observation about how well-being varies over time. The underlying thought is that well-being is something one has (or lacks) at a moment or during some period, and it is subject to change over time. When we consider a person’s well-being—how she is faring, how she is doing—we often adopt this temporal perspective. We are interested in how a person was faring at some point in the past, how she is faring now, and how she might be faring in the future. And we are often keen to track the changes and compare how she is faring at these different times.
Sometimes, it seems, the manner in which our levels of well-being change over time actually affects our well-being, too. It is better, one might think, to be on one’s way up than on one’s way down. Perhaps it is better to be leading a life with some variation, than to be coasting along with a constant middling level of well-being. Of course, if the ups and downs are too steep, maybe that would not be so good for us either. It is worth wondering which of these patterns might be good for us and why. However, this article addresses the more foundational issue about the extent to which our concept of well-being, understood as an aspect of a person’s state at a particular time, has room for these sorts of macroscopic or diachronic factors.
I will focus on three questions: First, can present well-being depend on future events? Second, can present well-being (non-trivially) depend on itself? Third, can present and future well-being be interdependent? The third question combines the first two, in sense that a yes to it is equivalent (given some natural assumptions) to yeses to both the first and second. The central argument in this article concludes that we must answer the third question with no. The reason, in essence, is that interdependence of well-being across time would yield what I call a problem of too many moving parts: A person’s well-being at a time would be so intricately bound up with her well-being at other times that it would not make sense to consider her well-being an aspect of her state at particular times. This negative verdict on the third question means, of course, that we must say no to at least one of the first two questions as well. The choice between them is complicated. Ultimately, I will suggest that we should reject the possibility that present well-being can depend on future events, while remaining open to the possibility that present well-being may depend on itself.
One standard approach to issues like these would be to test the relevant theses and their consequences for accord with our intuitions, appealing to what would or would not make sense in our ordinary thought and talk. I will indeed discuss such considerations, but they will ultimately prove inconclusive. Hence, I also take a slightly broader perspective. In addition to considering our everyday assumptions about well-being, I draw on the assumptions about well-being that underlie social scientific study of it, as well as the assumptions that structure our philosophical inquiries about it.
Three domains for the temporal character of well-being
Our interest in well-being is broad, including at least everyday conversation, empirical policy evaluation, and philosophical inquiry.Footnote 1 For the arguments that follow, we will draw on each of these domains. So, we will begin by reminding ourselves how, in each domain, we are interested in tracking and comparing individuals’ well-being over time and across circumstances.
In everyday conversation, well-being commonly wears its temporal character on its sleeve. When I see a colleague upon her return from a leave of absence, I may ask how she is. I may ask if she is doing better than before her leave. My concern might be narrow; I might just want to know if the situation that required the leave has been remedied. More often, I will be interested in anything bearing on how she is faring more generally. Suppose a mutual friend informs me, “Well, it’s taking her longer than expected to regain use of her right leg, but she’s very excited because she got that fellowship she applied for last year.” When I hear this, I will not interpret the second clause as a change of subject. Rather, it provides additional information about my concern. And our friend may naturally continue, “Overall, I’d say she’s doing pretty well, certainly better than this time last year.” Even if I ask how someone is doing just after some significant life event—say, starting a new job, returning from vacation, getting divorced, losing a loved one, etc.—I am likely to be just as interested to hear about any other factors bearing on her state and outlook at that moment. If so, then I am asking about her well-being at the time. I take it this is quite familiar.
Not so universally familiar but perhaps equally significant in our lives is the focus on well-being in empirical research, motivated by governmental and organizational decision-making and policy. The last 40 years have witnessed a proliferation of what we may think of as well-being-related constructs. Prominent examples include subjective well-being (Diener 1984; Diener et al. 1999), happiness interpreted as global life-satisfaction (Veenhoven 1984, 2000), and the state of flow (Csíkszentmihályi 1990). In addition to these broadly encompassing constructs, social psychologists, policy-makers, and management professionals also measure domain-specific well-being-related constructs tied closely to occupations. These include job satisfaction and work engagement (Bakker et al. 2008).
For present purposes, more important than the differences among these constructs is what they share: Whether we want to evaluate a governmental policy or, say, an employee benefit, we look for changes in well-being across time, tracking how the constructs respond to interventions. Some constructs, such as global life-satisfaction, we expect to be relatively stable, not fluctuating extremely, especially over short periods (Veenhoven 1984). But we expect all of the constructs to exhibit at least some variation over time. Hence, in a standard experimental design—a randomized field trial—we might have two randomly assigned groups, each of which is surveyed twice, at the beginning and end of the study, regarding some well-being-related construct. One group is subject to the intervention being studied, while the other is administered a suitable placebo. A significant difference between the intervention group and the placebo group in the change of measured levels of the well-being-related construct tells us something about the power of that intervention to affect that construct.Footnote 2
In short, regarding interventions and policies intended to affect subjects’ well-being, it is a background assumption, if not an explicit premise, that some factors within our sphere of influence may increase or diminish individuals’ well-being over time, yielding higher or lower measurements of the constructs that reflect it.Footnote 3
Finally, the concept of well-being is prominent in various areas of moral philosophy. For instance, consider a consequentialist view of moral rightness that evaluates consequences in terms of the well-being of the persons involved. In this vicinity questions naturally arise about how to evaluate aggregates of well-being. Aggregation may be across persons: Are two extremely happy persons better than five who are moderately happy? Or spanning time periods: Is a period that includes high highs and low lows better than a period where the person’s well-being remained stably at a moderate level? Another much-debated question has been about whether the goodness of a life is equal to the sum of the goodness of its parts.Footnote 4 Here, the parts typically in question are temporal parts, the episodes a life comprises. In order for this sort of question to make sense at all, we already must think of well-being as attaching to those episodes.
Instead of focusing on substantive questions about the relationship between well-being over longer periods and the shorter periods they contain, note what such questions presuppose: They presuppose that well-being is something a person has to a particular degree or at a particular level at a particular time. Regardless of how we characterize such a level and the scale we might use to measure it, it is something we attribute to a person at particular times.
To say that well-being is future-sensitive is to say that a person’s present well-being may depend on future events.Footnote 5 Might well-being be future-sensitive? Recent philosophical debate has been inconclusive.Footnote 6 Ultimately, I will argue that we should not countenance a concept of well-being that allows significant future-sensitivity, but the argument will depend on upcoming sections. In this section, in light of an initial examination of the three domains I identified, we will see that the case is not clear-cut.
How could it even be that present well-being might depend on the future? More intuitive is the thought that present well-being is past-sensitive, that it may depend on what has happened in the past. For example, suppose I am a mathematician who has just found an elusive proof for some famous theorem T. Proving T may improve my well-being, but, arguably, the degree of improvement depends on my history. If I have been working for years on T, and the proof is the culmination of my labor, then my well-being when I discover the proof may be greater due to this history. In contrast, suppose I just discover the proof in a flash, as it were, without much prior effort. In that case, although my well-being may increase when I prove T, perhaps it does not improve as much as in the more laborious scenario. If that is plausible, and if we are careful to stipulate that the difference in increased well-being is not due to a difference in feelings of satisfaction, then it is plausible that well-being is past-sensitive.
We can exhibit future-sensitivity with a similar case. Consider two ways my life might go. In one, I work for years to prove T, but I never succeed. In the other, I equally work for years, and I finally do succeed in proving it on the last day of my life. During the corresponding periods of labor, did I have greater well-being in the life where I eventually succeeded than in the life where I failed? We might be tempted to affirm this, perhaps in accord with a general thought that labor that will eventually pay off does more for one’s well-being during the period of labor than does labor that will ultimately be for naught. Such thoughts depend on the possibility that well-being is future-sensitive.
Questions about future-sensitivity naturally arise in the context of desire-satisfaction theories of well-being. Desire-satisfaction theories compel consideration of future-sensitivity because a desire and its satisfaction may not take place at the same time. Specifically, a desire may be satisfied by events that occur later, perhaps after the desire has been extinguished. For instance, I might now desire that my novel be published some day, but, as it happens, it is not published until 5 years later, at which point I no longer care. In such a case does the desire satisfactionFootnote 7 benefit me at all? If so, when? At the time when I had the desire? At the time when it was satisfied? If we say the benefit occurs at the time of the desire, then the view allows future-sensitivity.Footnote 8 According to such a view, my well-being when I have a particular desire depends on future contingencies about whether or not the desire is ever satisfied. With this as a paradigm conception of future-sensitive well-being, we can examine whether this is consistent with talk about well-being across our three domains.
At some points during our everyday conversations, future-sensitivity seems rather implausible. Suppose I bump into a colleague and ask her how she is, and suppose she responds, “Good, I think. But I’ll have to let you know next week.” This strikes me, for one, as odd. I think my interpretation would be that she was informing me that present happenings are much less significant to her than what will be happening next week.Footnote 9 However, suppose she clarified, “I’m not sure about my well-being at the moment because it depends so much on how next week goes.” This strikes me as an evasion, or perhaps a change of subject. However, intuitions may differ here. Suppose I buy shares of a mutual fund today, and you ask me if I made a good investment. I might well say, “We’ll see,” which seems like an apt response. Maybe well-being is like a good investment: it depends on what pans out. I grant that everyday conversation is flexible enough to accommodate talking about well-being this way.Footnote 10 If that is right, then everyday conversation does not rule out future-sensitivity.
Now consider empirical research. There appears to be a problem in how we would have to interpret self-reports of well-being (as in a survey), supposing that well-being were future-sensitive. An immediate consequence is that the person may not be able to know how she is faring. Nonetheless, her estimate of her well-being—how she thinks she is faring—may be a reliable indicator of her actual well-being. So far, this is no more objectionable than accepting that we sometimes can do no better than to provide an imperfect estimate of our present well-being. The only difference is that instead of an estimating how she is at that moment, she is estimating how she will have been at that moment. Due to the dependence of this moment’s well-being on the future, the estimate is sort of like a forecast of objective retrospection of the present, a forecast of how this moment will have been in retrospect.
That subjects must be offering forecasts like this contradicts the assumptions of researchers who intend to measure properties that are fully realized at the moment of measurement. This remark is emblematic: “Since well-being is defined as an evaluation of an individual’s actual life, the assessment of well-being has to be retrospective” (Zou et al. 2013: 1249). Nevertheless, our question is not about researchers’ own assumptions, but about the implicit assumptions on which the coherence of their endeavors depend. Although interpretation of well-being self-reports as forecasts of the past may seem odd,Footnote 11 it is not incoherent, nor, it seems, contrary to the aims and practices of empirical research on well-being. Hence, this does not yet rule out future-sensitivity.
However, serious problems with future-sensitivity emerge when we reflect on the assumptions behind experimental design. Recall the standard setup for a field trial, as described earlier: Begin with two randomly assigned groups and an initial measurement of well-being; perform some intervention on one group but not the other; then measure well-being in the two groups again; and, finally, compare the pre- and post-intervention levels to see whether the intervention had any effect. The aim of such a study is to register the effects of the intervention on the subjects’ post-intervention well-being levels. But suppose well-being is future-sensitive. If so, then the intervention may also affect the subjects’ pre-intervention levels of well-being. Not only is the study not designed to catch this; it is designed not to be sensitive to it: Any pre-intervention measurement cannot, in principle, register differences due to membership in the control group or the experimental group. Remedying this would require, at the very least, post-intervention measurements of pre-intervention well-being. Without this, we ignore potentially significant effects of the intervention on subjects’ well-being. And this would be inconsistent with the aims and practices of empirical research on well-being.
Recognition that well-being was future-sensitive would require us to rethink well-being-based policy-making as well. If well-being is future-sensitive, then policy-makers have been operating with an incomplete perspective on the range of possible interventions by which we might advance our policy goals of increasing well-being. After all, it could turn out that some of the most effective ways to increase well-being are through policies that affect past well-being. This is a striking thought, and, if correct, it would call for radical new directions in policy innovation. Hence, we can conclude that even if future-sensitivity would not preclude well-being-based policy-making or the empirical research which would guide it, it is inconsistent with the goals and practices as they now stand.
How does future-sensitivity sit with the philosophical discourse about well-being at a time? Specifically, is future-sensitivity consistent with the assumptions grounding the debate over the relationship between well-being over long periods and over the shorter periods they contain? It would appear so. Far from worrying that future-sensitivity is inconsistent with these debates, it has sometimes been thought to provide solutions to problems in this vicinity. For instance, Dorsey (2015) argues that countenancing future-sensitivity allows a reconciliation of an additive account of a person’s well-being over time with the plausible claim that the distribution of well-being over a life itself affects a person’s well-being.Footnote 12
Hence, philosophical assumptions about the temporal character of well-being do not straightforwardly rule out future-sensitivity. However, in Sect. 5, we will see that future-sensitivity is ultimately untenable given that well-being exhibits the feature, recursive dependence, to be discussed in the next section.
Can a person’s level of well-being at a moment affect her well-being at that very moment? For instance, can the fact that I am doing better right now than I was earlier affect how I am doing right now? To put the general question more carefully, is it possible for a person’s well-being at a time to depend on factors that may themselves (non-trivially) depend on her well-being at that time?Footnote 13 To answer yes is to hold that well-being has a feature we can call recursive dependence. At first glance, recursive dependence may seem absurd. Or it may seem abhorrent from the standpoint of tidy theory-building. Suppose we thought that a person’s well-being was partially constituted by her health, and that her health depended on her well-being. Then how do we assess the person’s well-being? Well, first, we need to assess her health. But for that, we need to assess her well-being. A vicious cycle ensues.
At a second glance, recursive dependence may not be so bad. To ease into the idea, consider that past well-being might influence our present well-being. We might think this: Having fared poorly in the past makes the present better than it would have been. This is like the commonplace thought that having had a tough time might make doing okay be actually pretty good, by comparison. One way to unpack this kind of thought is that past well-being affects present well-being in some standard way, regardless of what the present level of well-being would have been—like being at level -5 in the past gives you a +1 bonus in the present. But another way to unpack this is sensitive to the difference between past and present well-being, yielding something like this: Having fared poorly in the past makes faring well in the present even better. More carefully, improvement signified by the difference between my past and present well-being would make the present even better than it would have been without the improvement. According to such a principle, present well-being would depend, in part, on itself. Thus, well-being would be recursively dependent.
Before confronting the alleged theoretical abhorrence of recursive dependence, consider whether recursive dependence is consistent with the assumptions of everyday conversation and empirical research. Elaborating the principle just described, recursive dependence does not seem to be in tension with our ordinary conversations about how we are doing. Imagine the following exchange:
Anke: How are you?
Dries: I’m pretty well. In fact, a lot better than a few months ago. Considering how poorly I was doing then, I’d actually say I’m quite well now.
Anke: That’s good to hear!
Dries: Yeah, you know, considering where I am and where I’ve been, I’d say I’m really well now.
Anke: Okay, hang on. Are you pretty well, quite well, or really well???
Dries: Haha! That’s what I’m saying. When I step back and look at the improvement, I’d say I’m really well.
Anke: I see. I’m glad you’re on such a positive upswing.
Here we have Dries reassessing his well-being as he takes his upswing into account. Note that we need not think it is the experience of reassessing that is the cause of the upswing. Rather, let us suppose, Dries increasingly recognizes that his present well-being has been improved by the relationship in which it stands to his past well-being. Perhaps these kinds of iterative assessments are not particularly common, but they do not seem to be at odds with the background assumptions of our ordinary conversations about our well-being and how it changes over time.
With regard to empirical research and associated decision-making, recursive dependence is of no greater cause for concern than it is in the context of everyday conversation. Given that individuals’ self-assessments of their own present levels of well-being can approximately reflect those levels, the possibility of recursively dependent well-being does not jeopardize the goals or methods of empirical well-being research.
Of course, if it were the case that, due to the recursive dependence of well-being, individuals’ self-assessments reflect very inaccurately those individuals’ actual well-being, then this could be a problem for empirical research. But it would be a measurement problem and a matter of degree, similar to problems that arise with any conception of well-being according to which a person’s well-being may not always be immediately transparent to her. If, however, recursive dependence entails a vicious circular dependency—one that does not bottom out and leaves well-being levels genuinely indeterminate at certain moments—then recursive dependence would be inconsistent with assumptions and practices of empirical study. The problem would be the systematic error in subjects’ assessments of their own well-being. Subjects’ self-assessments would be deeply erroneous (not just inaccurate) at those moments when there was nothing determinate to assess.
So, now we must see whether or not the circularity is indeed vicious. Campbell (2015) suggests that recursive dependence has absurd consequences. He considers a case where the shape of a person’s life—the shape of the curve representing his well-being levels over time—is supposed to benefit the person by affecting his well-being at a particular moment. In such a case we would have to wonder whether or not the curve already registers the benefit at that moment. Regarding the affirmative horn of this dilemma, Campbell remarks, “If so, this implies that he was mysteriously bootstrapped into having his exact synchronic well-being curve with the aid of the benefit of having that exact synchronic well-being curve! This is a very bizarre result.” (2015: 573). But the negative horn looks even worse: If we suppose that bootstrapping does not work, then present well-being must be adjusted to reflect the shape of the curve, which means the shape of the curve changes to reflect the adjusted present well-being, which means present well-being must be adjusted, ad infinitum. We would lose our grip on any purportedly fixed level of well-being at the moment in question.
Although some bootstrapping may seem mysterious and although recursive dependence is a condition that makes possible the kind of vicious circularity that yields indeterminacy, it turns out that recursive dependence does not, by itself, entail absurd consequences. In fact, iterative reevaluation could itself be a model of how recursively dependent present well-being bootstraps to a stable level. Consider a view of well-being that does exhibit recursive dependence in how it accounts for the value of improvements in well-being: At moments when a person’s well-being has an upward trajectory, her present well-being gets a boost—a kind of improvement bonus—according to how much the present improves on the past.
Recall Dries, whose well-being is boosted because of how much better he is now than before. For illustration, make the example artificially precise. Suppose Dries’s level of well-being at each moment gets a precise real number, and the value of the improvement bonus is 20% of the person’s total well-being increase over the relevant period. Dries’s level of well-being could be figured as follows:
Earlier, his well-being level was 5.000.
Presently, his well-being level would be 7.000.
This improvement of 2.000 would yield a bonus of 0.400.
Presently, his well-being level would be 7.400.
This improvement of 2.400 would yield a bonus of 0.480.
Presently, his well-being level would be 7.480.
This improvement of 2.480 would yield a bonus of 0.496.
Presently, his well-being level would be 7.496.
. . .
Presently, his well-being level is 7.500.
Thus, even at a moment when the level of well-being exhibits recursive dependence, that level can be finite, determinate, and precisely figured—at least under some conditions. Therefore, recursive dependence itself does not entail absurdities. Of course, it is easy to adjust premises to upset the story. If we make the improvement bonus too large, then Dries’s well-being level would not stabilize after all and indeterminacy threatens. But that problem should be blamed on the new premises, and the instability or indeterminacy would amount to a reason to reject those premises. Without some good reason to posit anything like a very large improvement bonus (or other details that yield indeterminacy), recursive dependence itself is fine.
From the standpoint of the philosophy of well-being, this is a welcome conclusion, since some prominent theories of well-being naturally open the door to recursive dependence. This is most evident with desire-satisfactionism. Consider: I can have desires about my own present well-being—e.g., a desire to be doing better than I had been. Or, I may have desires about someone else’s well-being while that person has desires about mine. If our theory does not constrain the sorts of desires whose satisfaction affects well-being,Footnote 14 then some quite ordinary desires will yield recursive dependence. But, as we have just seen, this is not necessarily a problem.Footnote 15
We may conclude, then, that although initially worrisome, recursive dependence per se is consistent with our assumptions about well-being at a time in our three domains. Since recursive dependence causes problems only under certain conditions, we should not consider exclusion of recursive dependence a constraint on our concept of well-being. In the next section, we will see that one such problematic condition is future-sensitivity.
Mutual dependence of present and future well-being
We have just considered two possible requirements on the concept of well-being—first the exclusion of future-sensitivity, and, second the exclusion of recursive dependence—each in isolation. In each of these discussions, we did not consider how these two possible requirements (or lack thereof) interact. What, then, would be the consequences of a theory that simultaneously ran afoul of both of these possible requirements? Could it be that a person’s present well-being depends on both her present level of well-being and what happens in the future?
At this point, we need to consider everyday conversation and empirical research only briefly. We saw that manifestations of future-sensitivity and recursive dependence, though perhaps not very common, appear consistent with our everyday conversations about well-being. Any conversation intricate enough to exhibit both future-sensitivity and recursive dependence would plausibly qualify as a philosophical conversation regardless of the context in which it occurred. So, there is little to be gained now by lingering in the domain of everyday conversation. Considering the empirical study of well-being and the decision-making it informs, a very brief remark suffices. We already saw that future-sensitivity is inconsistent with the assumptions behind empirical well-being research and policy. It follows that this domain must also exclude combinations of future-sensitivity and recursive dependence. So, without further ado, we can turn to the philosophical domain. To preview: We will find that the sort of problem we initially spied but managed to dissolve in the case of recursive dependence becomes intractable when combined with future-sensitivity.
Among the central aims of the philosophy of well-being is to explain what it is to be faring well at a moment, what is to be faring well over a long period, and what it is to have a good life. If all goes well, progress in each of these areas sheds light on the others, and we end up with a comprehensive view. I will contend that countenancing mutual dependence of present and future well-being threatens this project by yielding an unusable concept of a person’s well-being at a time. It is unusable because it makes a person’s well-being during every period of any length depend, at least potentially, on her well-being during every other period. I will call this a problem of too many moving parts. If a person’s well-being at a moment is already interdependent with her well-being at past and future times, then it is not a discrete element which, along with the other discrete elements, can compose a narrative or other evaluatively significant diachronic features of a life. That is, sufficiently robust interdependence of well-being across time effaces the very possibility of viewing well-being as discrete levels arranged in a series, thus eliminating the possibility of evaluating the significance of such a series.
We can now examine the robustness of the interdependence entailed by a combination of future-sensitivity and recursive dependence. Consider, for example, a view that includes the claim that the shape (the direction and steepness of its slope, its smoothness or bumpiness, etc.) of a sequence of ups and downs in a life can be intrinsically valuable or disvaluable for a person.Footnote 16 Further, suppose, consistent with recursive dependence, that a person’s level of well-being at one time can depend on itself, as well as on the well-being levels at other times in the life. Accordingly, the value and disvalue of the life’s shape is reflected in the person’s well-being at various moments. And suppose also, consistent with future-sensitivity, that not only can well-being at one time depend on well-being at an earlier moment, but also on well-being at a future moment.Footnote 17
With this as the theoretical backdrop, take the perspective of the end of a person’s life and consider the person’s well-being at a particular moment during that life. Suppose, for now, that the person’s well-being at every other time is fixed. If the well-being at the particular moment in question were to change, then this would ipso facto change the sequence of well-being levels in the life, and, so, let us suppose, how good the sequence was. Thus, we can think of the goodness of the sequence as a function of the well-being level at that particular moment. But now note that the goodness of this sequence can itself affect the level of well-being at the moment in question. Let us suppose that everything else that affects the well-being at that moment is fixed, and so the level of well-being at that moment is a function of the goodness of the sequence. We have the goodness of the sequence as a function of the well-being at the moment, and the well-being at the moment as a function of the goodness of the sequence.
When we consider these two functions together, we have something resembling our earlier example of recursive dependence. To see our present scenario in action, start with the person’s apparent (i.e., without any recursive effects) level of well-being at the moment. Plugging this value into the first function yields a value for the goodness of the sequence of well-being levels. Then plug this value into our function for the well-being at the moment. This yields an adjusted level of well-being at the moment. From there we calculate a new value for the sequence. And so on. As in the earlier discussion of recursive dependence, we should note that this might well stabilize. The picture may appear tidy enough, or even elegant, when we notice that our two functions resemble a simple system of equations and that a process of iterative evaluation may not be the only way to reach a solution. So far, this is not terribly complicated or troubling.
However, so far our illustration has relied on a crucial assumption that keeps complexity at bay, an assumption that is inadmissible. The assumption was that, other than at the point in question, the levels of well-being at all the other times are fixed. But, according to the view we are examining, these levels would not be fixed; they, too, are potentially sensitive to well-being at all the other moments of the life. Hence, when we reach a new evaluation for the sequence of levels of well-being, this affects (at least potentially) the level of well-being at every other moment in the life. And, of course, the adjustment of these well-being levels yields a sequence with a different shape. So, not only do we have interdependence between well-being at the moment and the value of the sequence; we have interdependence among the levels of well-being at all the moments of the entire life.
To rein in this additional complexity, we might hope to avail ourselves of the shortcut just mentioned, of bypassing all the iterative calculations and settling all the evaluations in one fell swoop, like solving a system of equations. But note that it would no longer be a system of equations in just two variables; there would be as many variables as the number of discrete moments of well-being in the person’s life. It is hard even to speculate about the conditions under which such a system of equations would have a unique solution. But even setting aside worries about determinacy, this is simply too much interdependence. According to this picture, the person’s well-being at a particular moment is not just her status at that moment, but also a status inextricable from all the other moments of her life.Footnote 18 Of course, the person’s well-being at a moment is still an aspect of her state at that moment, in the simple sense that it is indexed to that moment. But, besides that index, the person’s well-being at any particular moment would be best understood as a mere parameter in the complex diachronic evaluative system of the person’s life, which now appears as the only object of evaluation we can neatly circumscribe.
Thus, the problem of too many moving parts problematizes basic assumptions that have structured questions in the philosophy of well-being. In short, when a person’s well-being at every moment may depend on her well-being at every other moment, this extreme interdependence makes it difficult if not impossible to productively engage questions about how we should evaluate the significance of how a person’s well-being unfolds moment by moment, episode by episode, over a life or across some long period of time.
Of course, this is not yet a reductio ad absurdum of every possible theory of well-being that allows mutual dependence of present and future well-being. The details we just unpacked were based on substantive assumptions about the value of the sequence of well-being levels and how this value is reflected in the well-being at the moments of the life. These assumptions yielded a very high degree of interdependence among well-being levels. Perhaps different assumptions would yield a more tractable account of well-being at a time. But it is difficult to see how a theory could allow many moments of future dependency, still allow present well-being levels to be sensitive to the past, allow recursive dependence, and meanwhile maintain a modest level of complexity. Hence, I conclude that our concept of well-being must not admit mutual dependence of present and future well-being.
Leaving the good of tomorrow for tomorrow
In Sects. 3 and 4, we found that neither excluding future-sensitivity nor excluding recursive dependence was, on its own, required by the assumptions driving philosophical theorizing about well-being. However, in the preceding section, we found that excluding at least one of these is required, since the philosophical debate about well-being requires that there be no mutual dependence of present and future well-being. So, how do we decide which restriction to accept?
One rationale would embrace a larger focus beyond the philosophical discourse. It has appeared that appeals to everyday conversation were no more conclusive than arguments grounded in philosophical discourse. However, considering empirical well-being research turned out to be instructive. We saw that recursive dependence does not by itself yield any special complications for empirical research, while, in contrast, future-sensitivity is, just by itself, incompatible with fundamental assumptions of empirical well-being research and the policies we might intend it to guide. This favors theories that do not allow future-sensitivity.
Clear as that may be, it is not conclusive. Proponents of future-sensitivity may insist that practices and assumptions of empirical well-being research should be revised. Or, perhaps rather than revision of these practices, all that is required is that they be reinterpreted. Perhaps, one might contend, conceptions of “well-being” predominant in the social sciences are quite different from well-being proper, and better understood as specialized notions suitable to the goals of these disciplines. However, if so, readers and consumers of empirical studies on “well-being” should be apprised that the constructs the studies track and measure do not have the normative significance we might have thought they had.
Putting aside now the significance of empirical well-being research, this article does advance our theorizing about well-being in one other way. The untenability of theories that entail both future-sensitivity and recursive dependence dialectically situates these two potential features of well-being so that an argument for countenancing one is an argument for rejecting the other. This is especially relevant for debates about desire-satisfaction theories of well-being, for which some have argued that we should accept future-sensitivity.Footnote 19 Though these debates are too deep for us to wade into now, it is nevertheless worth emphasizing that they are incomplete without a discussion of recursive dependence. This is notable because, as I pointed out at the end of Sect. 4, it is difficult to craft a desire-satisfaction theory to preclude recursive dependence, since desires about our own well-being or about the well-being of others easily give rise to it. Insofar as well-being may exhibit recursive dependence, future-sensitivity is untenable.
In general, we have seen that scrutiny of the possibility of future-sensitive well-being should not ignore issues about recursive dependence. In short: We may hold that how well a person is doing depends on future features of her life. Or we may hold that how well she is doing depends on how that very status relates to other past and present features of her life. But we cannot hold one without denying the other. As we have seen, this is because a person’s well-being in the present cannot depend on her future well-being. Her good today cannot depend on the good of her tomorrow.
It is difficult to contest the relevance of well-being within these areas, but see Scanlon (1998: ch. 3) who argues that to emphasize well-being, except perhaps in the second context (in relation to policy-making), is a mistake.
One might wonder whether psychologists and policy-makers intend to measure well-being itself or some more narrow category. A distinction is sometimes drawn between well-being as a normative category and happiness as a more narrow, purely psychological one (Haybron 2008). If happiness, conceived in this narrow sense, rather than well-being, is the object of empirical investigation, then arguments from the standpoint of the assumptions of empirical research apply to this concept. Of course these arguments apply to the concept of well-being only insofar as it is indeed the object of empirical investigation.
To reject future-sensitivity is not yet to endorse what Bradley has called internalism, the thesis that, “The intrinsic value of a time for a person is determined entirely by the value atoms obtaining at that time” (2009, 18). Rejecting future-sensitivity is consistent with accepting that present well-being depends on events that were realized entirely in the past.
Clearly, we are not talking about a feeling of desire satisfaction, since the desire no longer exists at the moment of purported satisfaction. Desire satisfaction in the sense relevant here occurs if a person desires that P and it becomes true that P, whether or not the original desire exists when P comes true. Lin (2017) calls this eternalism about satisfaction.
Dorsey (2013) argues that the best form of desire satisfaction has this feature.
Rejecting future-sensitivity, I make a similar suggestion about apparent future-sensitivity in our everyday conversation about well-being (King 2018).
Along these lines, Campbell suggests, “In some cases, people may feel unable to say how well they are doing in the present without having certain information about the past or future” (2015: 570).
In earlier work (King 2018), I suggest that this consequence of future-sensitivity is already too implausible to accept.
We find a similar suggestion in Brännmark (2001).
Of course, everything trivially depends on itself, in the sense that it is what it is and if it were not it would not be. In the cases of recursive dependence of interest here, a property depends on itself in the non-trivial sense that it is a factor among others in its own determination.
Of course, there are good reasons to restrict the range of desires relevant to well-being. Parfit objected to “unrestricted” desire theories that allow that my well-being may be affected by the fate of a stranger whose health I care about (Parfit 1984: 494). See Darwall (2002: ch. 2) for a helpful overview of the issues. However, most of the worries in this vicinity have been about desires for states that are intuitively unrelated to one’s well-being, like desires about remote situations or the good of others unrelated to us. But even if we exclude some such desires, it would not be plausible to say that a desire for, say, one’s spouse to be faring well is irrelevant to one’s well-being. But desires like that are sufficient for realizing recursive dependence.
Heathwood (2006: 543-544) considers a manifestation of recursive dependence when examining what desire satisfactionists should say about a person who desires to be faring badly and is indeed faring badly. Heathwood dissolves the apparent paradox—that if the person is badly off, he is well-off, and if he is well-off, he is badly off—by suggesting “the satisfaction of this desire to be badly off must, of necessity, count for less, in terms of welfare, than all the daily frustrations he racks up” (2006: 544). Although Heathwood leaves unresolved the apparent difficulty of figuring any determinate level of well-being for such a person (2006: 560, n. 5), our example of Dries shows that recursive dependence would not in itself preclude this, especially if Heathwood is right that this particular desire necessarily counts for relatively little.
This setup approximates Dorsey’s (2015) recipe for reconciling an additive conception of well-being over time with the possibility that the shape of a person’s life may be valuable for her. Specifically, Dorsey allows that the relations among temporally distant features of a life may be valuable for her (without claiming that the sequence of levels of well-being itself has any intrinsic value) and allows that this value may be reflected in a person’s well-being at a moment preceding at least one of the relata. The same complications I am about to describe could be described in Dorsey’s framework, but the explanation would be more complicated.
Brännmark (2001) offers a conception of well-being like this, where a person’s well-being at a moment depends on the role of that moment in the person’s life. I think he may underestimate the degree of the complexity entailed by such a view.
See, e.g., Dorsey (2013) for an argument that desire-satisfactionism requires future-sensitivity. Very briefly, his argument depends on two premises: First, in order for our own well-being to be sufficiently non-alienating to us, desire satisfaction must benefit us at the moments when we have the relevant desires. Second, our present desires regarding the future are relevant to our well-being at some times during our lives. See also Sarch (2013).
Bakker, A., Schaufeli, W., Leiter, M., & Taris, T. (2008). Work engagement: An emerging concept in occupational health psychology. Work & Stress,22(3), 187–200.
Boruch, R., Yang, R., Hyatt, J., & Turner, H., III. (2017). Randomized controlled trials. In B. Geve (Ed.), Handbook of social policy evaluation (pp. 15–37). Cheltenham: Edward Elgar.
Bradley, B. (2009). Well-being and death. Oxford: Oxford University Press.
Brännmark, J. (2001). Good lives: Parts and wholes. American Philosophical Quarterly,38(2), 221–231.
Bruckner, D. W. (2013). Present desire satisfaction and past well-being. Australasian Journal of Philosophy,91(1), 15–29.
Burtless, G. (1995). The case for randomized field trials in economic and policy research. Journal of Economic Perspectives,9(2), 63–84.
Campbell, S. M. (2015). When the shape of a life matters. Ethical Theory and Moral Practice,18(3), 565–575.
Csíkszentmihályi, M. (1990). Flow. Manhattan: Harper and Row.
Darwall, S. (2002). Welfare and rational care. Princeton: Princeton University Press.
Diener, E. (1984). Subjective well-being. Psychological Bulletin,95(3), 542.
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin,125(2), 276.
Dorsey, D. (2013). Desire-satisfaction and welfare as temporal. Ethical Theory and Moral Practice,16(1), 151–171.
Dorsey, D. (2015). The significance of a life’s shape. Ethics,125(2), 303–330.
Glasgow, J. (2013). The shape of a life and the value of loss and gain. Philosophical Studies,162(3), 665–682.
Haybron, D. (2008). The pursuit of unhappiness: The elusive psychology of well-being. Oxford: Oxford University Press.
Heathwood, C. (2006). Desire satisfactionism and hedonism. Philosophical Studies,128(3), 539–563.
King, O. C. (2018). Pulling apart well-being at a time and the goodness of a life. Ergo,5(13), 349–370.
Lin, E. (2017). Asymmetrism about desire satisfactionism and time. In M. Timmons (Ed.), Oxford studies in normative ethics (Vol. 7, pp. 161–183). Oxford: Oxford University Press.
Parfit, D. (1984). Reasons and persons. Oxford: Oxford University Press.
Sarch, A. (2013). Desire satisfactionism and time. Utilitas,25(02), 221–245.
Scanlon, T. (1998). What we owe to each other. Cambridge: Harvard University Press.
Slote, M. (Ed.) (1983). Goods and lives. In Goods and virtues (pp. 9–37). Oxford University Press.
Veenhoven, R. (1984). Conditions of happiness. Dordrecht: D. Reidel.
Veenhoven, R. (2000). The four qualities of life. Journal of Happiness Studies,1(1), 1–39.
Velleman, J. D. (Ed.) (2000). Well-being and time. In The possibility of practical reason (pp. 56–84). Oxford University Press.
Zou, C., Schimmack, U., & Gere, J. (2013). The validity of well-being measures: A multiple-indicator–multiple-rater model. Psychological Assessment,25(4), 1247–1254.
I am grateful for input from Philip Brey, Martijn Burger, Brandt van der Gaast, Robert Kraut, Donald Hubin, Ana Lucic, Sigrún Svavarsdóttir, Bart Voorn, Nils Wagner, and the attendees of the 2018 OSU/Maribor/Rijeka Philosophy Conference in Dubrovnik. Special thanks to Justin D’Arms for guidance and feedback on earlier versions, to Indy Wijngaards for helpful discussions about the psychology of well-being, and to Mayli Mertens for advice on the article’s structure and a close reading of the final version. This paper was written with support from the Netherlands Organisation for Scientific Research (NWO), under Project Number 652.001.003.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
King, O.C. The good of today depends not on the good of tomorrow: a constraint on theories of well-being. Philos Stud 177, 2365–2380 (2020). https://doi.org/10.1007/s11098-019-01314-z
- Social science
- Recursive dependence
- Desire satisfaction