1. Are adaptive preferences characterized by absence of perfectionist flourishing?

Before I turn to well-being theorists’ accounts of adaptive preferences, I should briefly explain why I do not begin with the well-theorized work of Serene Khader, who offers what I take to be the best and most explicit account given by someone who actively shares the political project. In brief, I reject her account because of its commitment to perfectionism about value. I do not intend what follows to convince anyone sympathetic to perfectionism to reject it, since a much fuller discussion of the issues would be required for that project. I have undertaken a more complete discussion elsewhere (Terlazzo Reference Terlazzo2014), Footnote ⁶ and intend what follows to situate rather than justify my non-perfectionist project.

Among adaptive preference theorists who subscribe to the political project, perfectionist theories of adaptive preferences are very popular.Footnote ⁷ Khader is no exception. She justifies her perfectionism in the following way:

Accordingly, she uses a substantive account of human flourishing to define adaptive preferences on the grounds that such an account can ‘provide us with an objective sense of what human beings need.’ (Khader Reference Khader2011, 18, italics in original)Footnote ⁸

Understanding adaptive preferences as (in part) preferences non-conducive to basic flourishing is in some ways promising for the political project, since it provides a clear and pressing justification for political intervention: if any circumstances or preferences are in need of change, it must surely be those that prevent people from pursuing even a basic level of flourishing for themselves. Nevertheless, I reject perfectionist accounts, on the grounds that our account should be compatible with recognizing what John Rawls calls ‘the burdens of judgment’ (Rawls Reference Rawls2005). These burdens explain the ways in which reasonable and well-informed people may continue to disagree about questions of value without either party being mistaken, and recognizing these burdens causes us to recognize that ‘a plurality of reasonable yet incompatible comprehensive doctrines is the normal result of the exercise of human reason within the framework of free institutions‘ (Rawls Reference Rawls2005, xvi). Since disagreement about comprehensive doctrines includes, in part, disagreement about what constitutes flourishing, I take it that we should avoid adopting an account of adaptive preferences that requires us to take a stand on what constitutes basic flourishing and which preferences violate it. The possibility of fully reasonable disagreement makes it morally unacceptable to necessarily count preferences as adaptive (and, therefore, flourishing non-conducive) on the basis of their content, even when those preferences are endorsed in considered, informed, and sincere ways. This will especially be the case for adaptive preference interventions in development contexts, where such significant power differentials exist between those staging interventions and those for whose benefit the interventions are staged. Persons who have already been disadvantaged and marginalized should not be told by those in positions of power that what they desire in a considered way could not be conducive to their flourishing. Rather, they must ultimately retain the authority to decide for themselves what constitutes flourishing for them.

The above should not be taken to suggest that Khader is unconcerned with these power differentials, or with defending a significant degree of moral agency for those taken to have adaptive preferences. Indeed, much of her work on the political project is focused on ensuring that these differentials are minimized and that agency is not violated or denied (Khader Reference Khader2011, Reference Khader2013). But ultimately, Khader and I have different bottom lines in this area. In response to the criticism that perfectionism involves a disrespectful claim that a person has corrupted values, Khader emphasizes that her perfectionist list only aims to capture flourishing at the most basic level. She writes that persons retain much space to decide on their own good beyond some basic minimum, and are not harmed simply because they have not ‘personally participated in forming the most fundamental normative conceptions according to which their lives are judged’ (Khader Reference Khader2011, 150–151, italics added). This stipulation helps to some degree, since if a perfectionist conception of the good establishes only a very limited set of goods, then persons are free to disagree about goods on which no stand is taken. But this does not get to the heart of things – it can offer nothing to the person whose argument is precisely with the set of goods that has been put beyond the pale. And as long as we can imagine even one person seeing one of the proposed basic goods as no good at all in her own life, even such minimal perfectionism is therefore a failure of respect.Footnote ⁹ Khader disagrees. Elsewhere, she writes that ‘it is [not] disrespectful to treat another as though she has some bad values. Being subject to criticism for one’s values is just part of being a normal moral agent’ (Khader Reference Khader2013, 314, italics in original). Again, this may generally be true – but in cases in which fundamental questions of value are at stake, and in which such great power differentials exist, it becomes even more important to give deprived persons the final say on the value of their own values. And these are precisely the cases with which the political project of adaptive preferences is concerned.

Nevertheless, Khader’s broader point remains: if we think that adaptive preferences distort people’s understandings of their own needs, then we need an account of what this distortion amounts to. If, in rejecting perfectionism, we simply take everyone to be an infallible expert on their own values, then we have effectively given up on the concept of adaptive preferences – and this will clearly not serve the political project. What we need is a fully theorized account of adaptive preferences that can serve the political project and make sense of the concept, without being hostile to fundamental disagreement about values.

2. Adaptive preferences as procedural autonomy deficits

If we reject substantive perfectionist efforts to theorize adaptive preferences, then the obvious alternative is procedural. Here, the content of a preference will be neither necessary nor sufficient for its adaptiveness; instead, adaptiveness will be determined by the procedure by which a preference is formed or maintained. Offering a full procedural account of adaptive preferences is, unfortunately, beyond the scope of this paper. Rather than attempting that, I will turn in what follows to some of the careful procedural work that has been done by well-being adaptive preference theorists. I will aim to show both what is valuable in their accounts and should be adopted by adaptive preference theorists focused on the political project, and what lessons the political project might have for well-being theory.

Seeing adaptive preferences as autonomy deficits has a long history in this strain of the literature. Indeed, John Elster, who initiated serious philosophical discussion of adaptive preferences, took them to be autonomy deficits (Elster Reference Elster1983). Seeing adaptive preferences in this way is also intuitively attractive since, as in the case of perfectionist accounts, there is an obvious justification available for treating them as suspicious: preferences are suspicious because they are not genuinely a person’s own, and therefore prevent her from living her life as she herself would freely choose to do.

But autonomy is a complicated concept surrounded by a huge and conflicting literature, and one cannot, accordingly, simply say that adaptive preferences should be understood as autonomy deficits. Rather, one needs to explain what precisely is meant by autonomy, and why autonomy deficits are problematic. As Ben Colburn rightly notes, Elster himself does neither of these things (Colburn Reference Colburn2011). But as we will see, while Colburn successfully improves upon Elster’s account in a way that the political project should take note of, consideration of the political project can also show what accounts like Colburn’s fail to do.

In aiming to explain what is bad about adaptive preferences, Colburn appeals to an account of autonomy characterized by two features:

Since at least some sort of endorsement is generally taken to be involved in adaptive preferences, Colburn focuses on the Independence condition to explain the badness of adaptive preferences. His proposal is that we take adaptive preferences to be preferences for which Independence is compromised by ‘covert influence’: that is, commitments (preferences, desires, etc) where ‘the explanation for those commitments is necessarily hidden from [a person], in the sense that it would not be the explanation for their commitment if it weren’t hidden’ (Colburn Reference Colburn2011, 64). To illustrate covert influence, Colburn uses the example of subliminal advertising, where the ‘real’ reason for a person’s desire for a product can no longer count as reason for endorsing desire for that product once it has been made apparent. At that point, the person must, Colburn claims, either find alternative and independent reasons for desiring the product that has been subliminally advertised, or else repudiate the desire on the grounds that it was induced (Colburn Reference Colburn2011, 65). Since covert influence violates Independence, and therefore compromises autonomy, Colburn holds that covert influence allows us to explain what is bad about adaptive preferences.

But while covert influence certainly compromises autonomy, one might wonder whether Colburn is right to focus on this particular type of compromise to determine adaptiveness. After all, many overt forms of influence can compromise our Independence as completely. Think, for instance, of a case in which a criminal holds a gun to Larry’s head and forces him to marry Mary – a woman whom he does not love. No covert influence is involved in this case, but it is obvious that Larry’s move toward seeing value in marrying without love is anything but independent. Of course, one might object that we can better explain the lack of autonomy in Larry’s preference by pointing out that he does not genuinely value the loveless marriage – instead, he values being alive, and marriage without love seems to be the only route to that outcome. But if this is the concern, then we can simply change the case. Instead, imagine Gary, a gay man who reasonably believes that he will be unable to have a socially accepted family with another man, Footnote ¹⁰ and so works hard to convince himself that he can be happy marrying Mary. Here, he does not simply see a socially acceptable family as valuable – instead, he consciously works to convince himself that his only route to such a family is itself one that he would be better off valuing. In this case, it is again an oppressive social system that violates the Independence of his choice, rather than a covert reason that he is not – and indeed could not be – aware of. But in neither loveless marriage case is there a reason to hold that violation of Independence is any less serious for not being covert. And both cases are clearly ones in which a person’s preferences are problematically adapted to a bad situation. Accordingly, there does not yet seem to be a principled reason for focusing on covert influence rather than overt influence in our discussion of the badness of adaptive preference.

Colburn himself, however, has a reason for focusing on covert influence in his particular context. He aims not only to show what is bad about adaptive preferences – he also aims to show how adaptive preferences can be distinguished from what Elster and others have called ‘character planning, ’ which both Elster and Colburn take to be comparatively unproblematic. Character planning involves ‘being aware of one’s options and molding one’s projects and inclinations so as to settle on preferences which one can fulfill’ (Colburn Reference Colburn2011, 55).Footnote ¹¹ When this option is made available, Gary’s case is clearly one of character planning, and covert influence does indeed pick out the difference between his case and other cases of adaptive preferences. So given Colburn’s desire to distinguish between character planning and preferences formed behind a person’s back, we can see why he appeals to covert influence.

I do not deny that Colburn picks out a genuine distinction between covert and overt influence. It remains open to us, however, to wonder whether this is a useful distinction to focus upon. That is, why take this distinction to be the relevant bad-making feature of adaptive preferences rather than some other? In particular, why take it to be more relevant than the distinction between those preferences adapted to circumstances whose Independence is more significantly violated and those who Independence is only minimally violated?Footnote ¹²

Considering the political project suggests a principled reason for well-being theorists to focus on the second kind of distinction, and to count preferences like Gary’s and Larry’s as adaptive. The political project uses the concept of adaptive preferences to identify a set of preferences that are problematic enough to warrant intervention. This intervention could be aimed at the preference itself, at the circumstances that cause it, or at both. But in any of these cases, the intervention aims to influence those preferences that warrant the most concern. The criterion that determines whether a preference warrants concern will be determined by the theory that one uses, and could range from injustice to inequality to lack of self-worth to lack of basic flourishing. And importantly, any one of these sorts of preferences could be developed, at least initially, through character planning. A person who is frustrated with all of the options available to her, but reasonably takes herself to have no others, might have good reason to convince herself that she should work to become happy with one of the options available to her. But the reasonableness of this decision clearly does not lend legitimacy to her initial circumstances. Whether significant success in her endeavor should lend legitimacy to her new preferences themselves is more controversial, but whether they do so will once again depend upon the criterion that one has used to determine whether concern is warranted. An intentional and successfully cultivated lack of self-worth or preference for circumstances non-conducive to basic flourishing will not become fully legitimate simply for having been formed on purpose. The political project, then, reminds us that it is the reason for which preferences warrant concern that should draw the line between the preferences that warrant inclusion in the theory and those that warrant exclusion.

Colburn’s project also aims to explain when and why preferences warrant concern, and here consideration of the political project can highlight what is problematic about the way in which Colburn draws the relevant distinctions in his project. Since Colburn is not committed to the political project, he may not be concerned with preferences that warrant intervention, and he is accordingly free to draw the lines for concern in a different place. These lines, however, must be more clearly drawn. If he is interested in determining which preferences contribute to well-being, then it is not clear that those that are covertly influenced detract more than those that are overtly influenced. If he is interested in determining which preferences deserve respect from third parties, then distinguishing overt influence that leads to character planning from covert influence that works without an agent’s knowledge will likely be a valuable project, since it is at least plausible that the intentional formation of a preference goes some way toward counting it as meaningfully an agent’s own. In this case, however, all of the problems associated with political accounts arise, and we are turned back toward concerns about intervention and respect. In either case, discussion of political accounts highlights that well-being accounts must be sure to explicitly address the nature of the concern in question, and to justify why that concern rather than some other is focused upon. Discussion of political accounts also more immediately suggests that by trying to explain both the badness of adaptive preferences and the goodness of character planning by appealing to the same mechanism, Colburn focuses on too narrow a type of violation of Independence – namely, the covert kind. Drawing the distinction in this way at worst suggests that preferences of Gary’s sort (which are, of course, also adapted to circumstances in a way that violates Independence) are unproblematic, and at best suggests that whatever problematic features they have should be explained by an account separate from our main discussion of adaptive preferences. If violation of Independence is really the bad-making feature of adaptive preferences, then Colburn would do well to separate his project of explaining the badness of adaptive preferences from his project of explaining the goodness (or at least neutrality) of character planning. This will require questioning the assumption that character planning is indeed good (or at least neutral), but it will also be the only consistent way of recognizing that what is bad about adaptive preferences is the violation of Independence.

Nevertheless, those adaptive preference theorists concerned with the political project would do well to borrow Colburn’s definition of adaptive preferences as autonomy deficits. As we saw above, such a definition explains why adaptive preferences are worthy of suspicion without requiring us to make any controversial commitments about which preferences are compatible with basic flourishing and which are not. To explain why adaptive preferences are bad, we can point to their lack of autonomy, and to explain the badness of a lack of autonomy, we can point to those features that compromise the extent to which a preference is a person’s own: either the Endorsement of the preference is compromised, or the same is true of the Independence required for that Endorsement to be meaningful. We have already seen, however, that we need a better account of what it means to violate Independence – and, since Colburn’s paper focuses on Independence rather than Endorsement, we will also need to provide an explicit account of the latter. To clarify the Endorsement question, I turn in the next section to Dale Dorsey’s work on value and well-being. To clarify the question of Independence, I turn in the final section to Donald Bruckner’s work on the feasibility and the related liveness of options. In both cases, their theoretical work provides a useful basis for an explicit account of adaptive preferences that those committed to the political project should endorse. But in each case, we will once again see that considering the political project recommends changes in Dorsey and Bruckner’s own accounts.

3. Endorsement and value

The challenge in giving a good account of the requirements of Endorsement is to explain what precisely is wrong with the endorsement involved in adaptive preferences, without inappropriately counting some other preferences as wrongly endorsed. Perfectionist accounts can explain what is wrong with the endorsement involved in adaptive preferences by pointing to some kind of fittingness relation between endorsement and the objects of adaptive preferences: when certain objects or activities are non-conducive to basic flourishing, it is not fitting to endorse them. But by dictating what counts as valuable or flourishing-conducive, perfectionist accounts inappropriately count too many preferences as wrongly endorsed: they count as wrongly endorsed even some of those preferences that persons might take, in a considered, committed, and independent way, to be most central to their own well-being.

Once again, well-being theorists can provide us with a fully theorized alternative to the perfectionist account. Dale Dorsey writes about ‘status-quo bias’ rather than using the terminology of adaptive preference, but his concerns are similar (Dorsey Reference Dorsey2010).Footnote ¹³ Like adaptive preference theorists, he is concerned that persons’ preferences can become inappropriately skewed toward the options that they have available to them. What is wrong with the endorsement involved in adaptive preferences, on his account, is that they reflect the circumstances of the person’s status quo, rather than capturing what is genuinely relevant to a person’s well-being. But he does not endorse a perfectionist explanation of this failure. Instead, he proposes assigning value in a way that privileges a person’s own authority, but is ‘abstracted from the agent’s actual conception of the good … [instead, value] must be identified with what a person would prefer under certain counterfactual conditions’ (Dorsey Reference Dorsey2010, 548, italics in original). So the endorsement involved in adaptive preferences is suspect because it does not take into account a wide enough set of candidates for possible endorsement – thereby missing the candidate that most deserves it. This kind of proposal should be attractive to those concerned with the political project. Abstracting away from a person’s actual conception of the good when it is molded by her constrained circumstances allows the concept of adaptive preferences to get off of the ground, but relying on the person’s own counterfactual preferences rather than some objective account of the good makes it possible to avoid the hostility to fundamental value disagreement found in perfectionist accounts of adaptive preferences.

However, Dorsey’s criteria for a preference to count as welfare relevant – that is, the counterfactual conditions that he takes to be relevant – are extremely demanding: he requires that a person’s actual preference be judged against a complete preference ordering that involves judgments about all possible lives.Footnote ¹⁴ Dorsey holds that such an ordering solves the problem of status-quo bias, because when ‘welfare relevant preferences are drawn from a complete ordering of all possible lives … there is no room for an agent’s good to be more heavily weighted towards the one she actually lives’ (Dorsey Reference Dorsey2010, 550). Such a demanding condition would indeed solve the problem of adaptive preferences. Adaptive preferences are problematic because they are stunted in order to reflect impoverished circumstances; if we count all preferences calibrated to circumstances as wrongly endorsed, then we have necessarily captured adaptive preferences in that set.

But considering the political project can once again highlight for us what is wrong with such a demanding account of well-being relevance. Since the political project is concerned with intervention aimed at ameliorating adaptive preferences, the preferences that our criteria capture as adaptive will be presumptive objects of intervention. Accordingly, political accounts of adaptive preferences ought to reject Dorsey’s strong criterion for adaptiveness, since it would open far too many preferences up to intervention. My preference for the diet that I have, for instance, or the clothing that I wear, is likely highly culturally conditioned and therefore unlikely to hold up against a full ordering of possible lives. And opening so many trivial preferences up to intervention would clearly take us far beyond the political project – a project focused on micro-managing all of persons’ suboptimal preferences would be very different from a project aimed at ending the oppressive or dominating circumstances that those living under them might not object to.

Since Dorsey is not committed to the political project, he himself need not be concerned with whether his theory has any implications for which preferences are open to intervention. Given this, it matters little to his theory that it identifies many trivial preferences as welfare irrelevant. But considering the political project also highlighted that on such a strong criterion, welfare-irrelevant preferences will be ubiquitous. All of my preferences are shaped by my circumstances, and the chance that any of my actual preferences are ones that would fall at the very top of my full rank-ordering is infinitesimal. This will be true even of the preferences that fall extremely high on my ordering. I may achieve and be delighted by a tenured job in the Harvard Department of Philosophy – but chances are good that I would still prefer to have an endowed chair there that raised my salary and lowered my teaching load. This seems to be a more significant problem for Dorsey, since such a view treats virtually every one of our actual preferences as impoverished. Any real instantiation of even my wildest dreams will be lacking in at least some small way. If even my desire for and achievement of the Harvard job counts as consolation prize, then it seems that no actual contribution to my welfare will ever be robust – and this is a profoundly depressing way of viewing the world.

The problem becomes even clearer when we consider not the achievements or possessions we would prefer to have, but the people we would prefer to be with. Recall Mary, the potential bride of both Gary and Larry – but now consider her third suitor, Harry. He thinks that Mary is genuinely wonderful, adores her company, and is delighted at the idea of spending a lifetime with her. Like many other adults, however, he rejects the idea of there being one true love for each of us, and believes that there are many different individuals in the world with whom he could have a wonderful marriage and valuable shared life. As it happens, while he takes Mary to be one of them, he also believes (or at least would find on a complete ordering) that Beyoncé is the one person in the world with whom he could be most happy. Since she is uninterested, he enthusiastically desires to marry Mary. In the case of achievements, while it might be frustrating, there might also be something gained in terms of ambition by recognizing that each achievement can be superseded by a greater one. But it would be significantly damaging to see our relationships in the same way, so that each actual partner, no matter how dearly loved, could be traded up for one who would actually suit us better. We should certainly want the status-quo bias to be able to explain what is wrong with Gary’s preference to marry Mary – but we should not say that the difference between his marriage and Harry’s marriage is at most a matter of degree.

Dorsey’s demanding account does, however, have one strong point in its favor: it chooses the only non-arbitrary line that can be drawn when it comes to the status-quo bias. It would be tempting to draw some sufficiency line for individual persons, so that preferences like Harry’s are above it and preferences like Gary’s are below – but any line drawn would necessarily be arbitrary. The only genuine difference in kind on such a scale is between that preference which dominates all others and those that are dominated, and Dorsey captures that difference with his strong requirement. We need to be careful, however, that aiming to draw a non-arbitrary line does not lead us to draw the line in the wrong place – and the problems raised above suggest that the line might indeed be drawn in the wrong place in this case.

To keep the strong requirement without the undesirable consequences discussed above, we could perhaps supplement the status-quo bias with another value, so that we introduced another difference in kind into the equation.Footnote ¹⁵ For instance, while all preferences but one would suffer from status quo bias, some preferences would additionally be ones that were the result of injustice, domination, or oppression.Footnote ¹⁶ By distinguishing between the two groups, we could reintroduce some kind of sufficiency criteria. Such a solution, however, brings us back to the problem with substantive perfectionist accounts. Injustice, domination, and oppression are themselves controversial concepts, and there is little consensus on what their substantive content might be. If we define them substantively, we are overly hostile to fundamental value disagreement, but if we cannot give a substantive account of them, then they do not help us to draw a useful line. The problem is even greater for those who use such a criterion in the service of the political project. If they define them substantively, then their hostility to fundamental value disagreement risks intervention into sincerely held, informed, and considered preferences, but if they rely only on persons’ own perceptions of which of their preferences result from domination, injustice, and oppression, we have already given up on the idea of adaptive preferences.

To escape this dilemma, both Dorsey and those committed to the political project might be able to use Dorsey’s hypothetical endorsement account of value, but give up on drawing a sufficiency line that is both non-arbitrary and bright. While I find this kind of solution promising, I can only offer a brief sketch of it here. In place of Dorsey’s account, we could appeal to some kind of ‘natural lumping’ in the ordering, so that preferences that naturally clustered nearby to the optimal preference could count as having passed a sufficiency threshold, while those that fell into lower clusters did not. This would explain why it was good for me to be a tenured professor at Harvard (or indeed many other universities and colleges), but not good for me to be a precariously employed adjunct. It could also help to explain why Harry’s marriage genuinely contributes to his welfare, while Gary’s does not.Footnote ¹⁷ This will not entirely solve the problem, since natural clusters may be very different. In some cases, a great many options may be closely and evenly spaced, although the first and last options in the cluster are very far from one another. In such cases, there will be no great difference between each option and the next, but a great difference between the first and last options. In such cases, a fuzzy sufficiency line may simply have to be drawn somewhere. The line will admittedly be fuzzy and somewhat arbitrary, but it is better to have a fuzzy and arbitrary line in roughly the right place than a distinct and non-arbitrary line drawn in the wrong place. Doing so would continue to capture what is wrong with many cases of status-quo bias, while still allowing us to say that many of the actual lives that people lead are valuable and welfare conferring. In cases where first options are unavailable, second and third options will no longer need to be seen as consolation prizes. Instead, these options could rightly count as genuinely endorsed ones – which, in turn, would count toward their status as autonomous preferences on an account like Colburn’s. This would allow us to pinpoint what is bad about adaptive preferences without counting far too many preferences as adaptive.

4. Independence and live options

Let us return, then, to Independence. Using Dorsey’s demanding account of value to render Endorsement legitimate forced us to count too many preferences as adaptive (or problematic in a closely related way), but if we use the wrong account of Independence, we might err in the opposite direction. In discussion of Colburn we saw that a lack of options can violate Independence, but it is not yet clear what this violation amounts to. Donald Bruckner’s work on feasible options can help the political project to explain this violation, but there is once again room for caution. I will argue that if we wrongly focus on which of a person’s options are accessible, rather than which of their options are live, we will end up counting too few preferences as adaptive.

Bruckner defines an adaptive preference as ‘a preference that is regimented in response to an agent’s set of feasible options’ (Bruckner Reference Bruckner2009, 308), and his aim, like Dorsey’s, is unrelated to the political project. It is, in fact, directly opposed: he aims to defend the claim that adaptive preferences can be fully rational and worthy of pursuit, and that we should judge them to be so when the person who holds them herself judges that they are. Whether or not this claim is plausible, however, depends on what we mean by ‘feasible.’ Feasibility could be related to either the accessibility or the liveness of options. By ‘accessible’ options, I mean those that a person could successfully take up, given the set of external resources that she in fact has. By ‘live’ options, I mean those options that persons could actually see themselves pursuing, given the values and ambitions that they in fact have.Footnote ¹⁸ Often, the liveness of an option and the accessibility of an option will come together. The fact that I know that I can successfully achieve a certain outcome will make it more likely to be live for me than an outcome that I have grave doubts about my chances of achieving. But the two categories can also come apart. For the Jew who keeps kosher in New York City, the option of eating bacon is accessible but not live; for the child who grows up in a family that cannot afford higher education, the possibility of attending college may be live despite the fact that it is inaccessible. When Bruckner defines a preference as adaptive only when it is ‘conditioned by what is and is not in an agent’s feasible set’ (Bruckner Reference Bruckner2009, 311), he seems to intend feasibility to be much more closely related to accessibility of options than to liveness. He holds, for instance, that an upper-class young man who would receive significant ridicule and lack of support from friends and family for taking up a blue-collar trucking job nevertheless has that option in his feasible set. Accordingly, when he abandons the preference for trucking that he would otherwise have, Bruckner insists that he does not do so adaptively (Bruckner Reference Bruckner2009, 312–313).

When properly conceived, the concept of feasibility can give those committed to the political project an account of Independence: a choice is Independent when it is the one that a person would consciously and intentionally select from a broad feasible set. This explains why neither Gary nor Larry developed their preference to marry Mary in an Independent way. The gun to Larry’s head clearly limited his feasible set of options, as did Gary’s hostile social circumstances. It also explains the lack of Independence of a covertly influenced preference for some object when that object would not have been consciously selected.Footnote ¹⁹ And more generally, it explains in an intuitive way what is wrong with adaptive preferences: they are preferences that people are forced by a constrictive option set to form, which they would not have chosen otherwise.

But considering the political project also suggests that Bruckner is wrong to focus on accessibility rather than liveness when fleshing out feasibility. Focusing on accessibility rather than liveness fails to recognize the huge impact that social influence and self-doubt can have on persons. While the option of trucking might still be accessible to the would-be trucker, despite his family’s ridicule and lack of support, such ridicule and lack of support could easily prevent the option from being live for him. To see how this may be so, consider a young woman who grows up in a household where she is taught that women must remain in the home, take care of their husbands, and perform all domestic tasks.Footnote ²⁰ Even if this woman has the accessible option of taking on a career – that is, even if she has the concrete external resources – the pressure from her family constitutes a real obstacle to her ability to do so. If this were not the case, then most feminist work on adaptive preferences would be entirely groundless. The political project of feminist adaptive preference theorists is rightly focused on preferences that are insidious precisely because they appear to be freely chosen and easy to avoid: preferences to diet constantly, to have cosmetic surgery, to wear constricting clothing, etc. While refusing to take all of these routes is certainly accessible to women in places like the USA, anyone who has been a teenaged girl in such cultures knows that decisions in these areas are anything but Independent. Any theory that wants to speak about adaptive preferences as a class must include preferences like these. Doing so calls into question Bruckner’s conclusion that adaptive preferences should be seen as desirable when they are endorsed by their holders – but his conclusion is rendered legitimate only by a stipulation that artificially excludes many of the preferences that adaptive preference theorists committed to the political project are most concerned by. In other words, by focusing on accessibility rather than liveness, Bruckner counts too few preferences as adaptive.

Considering the political project also suggests that Bruckner might be too quick in his conclusions even in cases where accessibility is genuinely what is at issue. Bruckner may be right that it is rational and good to act on adaptive preferences when there are genuinely no better options that exist. There would be little value in encouraging persons to pine after the things that they truly could not achieve, for this would almost certainly lead to painful and constant frustration. Even those committed to the political project could agree here. But those committed to the political project will also note that while social change is often hard and protracted, it is rarely genuinely impossible. In other words, it is rarely the case that a person’s best immediately accessible option is genuinely the only one that she might bring about with some commitment and work. This is doubly so when her own commitment and work can be joined with that of others in a social movement that can collectively bring about great change. While constant striving might indeed be frustrating, and while there is certainly something to be said for settling for a simply good life, striving is all too often the only means of bringing about a possible life that is better. By arguing in favor of preferences adapted to a person’s accessible set rather than her live set, Bruckner is too quick to discount the value of this kind of individual and collective aspiration. As we saw, persons may have options that are live without being accessible, and in such cases, Bruckner’s argument in favor of adaptive preferences mutes a tool for encouraging persons to bring about a better world for themselves.

5. Conclusion

In this paper, I have left aside difficult questions involving the best ways to identify and respond to concrete adaptive preferences in practice. These are questions that those who endorse the political project cannot afford to ignore. I have also not answered the question of whether a single unified account of adaptive preferences is either possible or desirable. It may be that it is both – but it may equally be that no individual account of adaptive preferences can do all of the work required of the concept by well-being theorists, feminist theorists, and development theorists. Much more work will be required to fully address this question. But in this paper, I have made the case that these three types of theorists must at least do more to engage one another. If they do not, all of three of their projects will be weaker for it. And if one unified account is both possible and desirable, then I hope in this paper to have at least uncovered the bones of an account that could serve all three purposes.Footnote ²¹ By appealing to autonomy in the form of Endorsement and Independence, we can explain what is wrong with adaptive preferences without being hostile to fundamental value disagreement. By looking to an agent’s set of options, we count as adaptive those preferences that lack Independence because they were not chosen from a sufficiently broad feasible set. By appealing to an agent’s own hypothetical judgments about value, we can determine roughly how broad that feasible set must be in order to count toward that agent’s genuine Endorsement of her preference: it must be broad enough to include a number of options that the person herself would take to be highly valuable on a complete ordering. By focusing on liveness rather than accessibility of options, we can recognize that failing to see accessible options as live can be as debilitating as lacking accessible options. And by encouraging persons to see options as live even when they can only become accessible through individual or shared striving, we further the political project itself: we call for collective transformation of those social conditions that seem so unavoidable that individuals would otherwise cripple their own aspirations and values to live comfortably within them.