frameworks / adaptive-preferences · article
frameworks/ · 2,720 w · 12 min · ✎ dialogue

Adaptive Preferences

Adaptive preferences is the philosophical and welfare-economic vocabulary for what happens when a person's wants shift in response to what they can actually get. The fox in La Fontaine's fable decides the unreachable grapes are sour; a defeated worker decides she never wanted the promotion; a caregiver whose career stalled decides family is what she values most. The question the literature addresses is whether such preference shifts are rational updates (you should listen to them), irrational cope (you should discount them), or something in between. The concept matters outside philosophy because it sits underneath real decisions — whose preferences count in health-economics QALY calculations, whether oppressed people's satisfaction with their lot invalidates calls for reform, and — for Jason's deep-ambition-book-thesis — whether a driven person's value shift from shallow to deep ambition is wisdom or defeat dressed up as wisdom. This hub page connects the three primary sources now in the wiki: Bovens (1992) on sour grapes vs. character planning, Baber on why the concept is a red herring, and Mitchell (2018) on the adapted/adaptive distinction in health contexts. Elster's original Sour Grapes (1983) and Nussbaum's capabilities work are engaged through these three papers rather than directly — the primary texts remain on the acquisition list at deep-ambition-sources-to-acquire.

The Core Puzzle (Elster's Setup)

Jon Elster's Sour Grapes (1983) set the canonical frame. Two kinds of preference revision look identical from outside but feel radically different:

  • Sour grapes (SG). Aesop's fox can't reach the grapes, so he decides they were sour all along. His taste for grapes has changed because he failed to get them. We judge this irrational.
  • Character planning (CP). A reforming poker player decides cheating is wrong, commits to fair play, habituates herself to it, and comes to prefer it. Her taste has also changed, and also in response to what she can (or should) pursue. We judge this rational.

Both are adaptations to feasibility. Why does one count as genuine preference change and the other as rationalization? Elster's answer: intentionality. CP is consciously chosen; SG is an unconscious drive reducing the frustration of unmet desire. The rest of the literature is largely an argument about whether Elster's answer works.

Three Positions Now in the Wiki

1. Bovens (1992) — the Davidsonian criterion

Full article. Bovens accepts Elster's puzzle but rejects Elster's intentionality answer as internally inconsistent (Elster's own commitment to rational-choice explanations requires preferences the agent isn't conscious of) and extensionally wrong (people acquire perfectly rational preferences through learning or habituation without any explicit "character plan"). His alternative, developed from Davidson's theory of akrasia, locates the irrationality of SG in a specific structural fact: in typical SG, the agent's preference flips but none of her underlying reasons, criterial judgments, or weights change. She still, all things considered, judges the old alternative better — her new preference contradicts her unchanged all-things-considered verdict. In CP, the reasons and weights restructure alongside the preference, so the preference remains informed by the (now also revised) all-things-considered judgment. The test isn't "did you choose this?" It's "did your reasons restructure too, or only your ranking?"

Bovens's move also predicts the intuitive gray zone: most real cases are partial restructurings somewhere between pure SG and pure CP. And it explains the auxiliary facts — why sudden post-failure flips look suspicious (real reason-restructuring takes time) and why SG invites descriptions in terms of self-deception (the old reasons are still producing the old all-things-considered verdict, so the agent appears to be suppressing that verdict).

2. Baber — the rationalist skeptic

Full article. H. E. Baber argues the concept is a red herring. The women Nussbaum cites as exemplars of adaptive preference — Jayamma the brick-kiln worker, Vasanti in the abusive marriage, Saida who married off her 12-year-old daughter, Srey Mom the Cambodian sex worker — are not in fact people whose preferences for better conditions have been extinguished. They are rational satisficers under constraint, playing lopsided odds, choosing between bundles rather than isolated goods, and declining to feel frustration because futile frustration is prudentially costly. Baber's operational test is sharp: would the fox jump at the grapes if they came within reach? If yes, he still wants them — his preference is thwarted, not revised. The apparent "adaptive preference" dissolves once preference is understood dispositionally (what someone would choose from an adequate option set) rather than as an occurrent felt state.

Baber's political worry matters for Deep Ambition: invoking "preference deformation" can substitute for the expensive work of providing real material options. Rights, dignity, and self-esteem are cheap; money, education, and legal infrastructure are not. Philosophers who diagnose preference deformation risk offering consciousness-raising in place of the options people actually want.

3. Mitchell (2018) — adapted ≠ adaptive

Full article. Polly Mitchell is concerned with the disability paradox: people with a condition report higher utility for their state than people without the condition predict for it. QALY calculations depend on which set of preferences counts. The standard move against patient testimony is: "they've just adapted." Mitchell's reply is to separate adapted preferences (ubiquitous, often rational, everyone does this) from adaptive preferences in the strict philosophical sense (irrational, malformed). She walks through Elster's, Bovens's, and Nussbaum's accounts and shows that patient adaptation fails to meet the criterion on any of them — much patient preference change is best described as Elster's own "learning and experience" category, or as comprehensive reason-restructuring in Bovens's sense, or as rationally grounded in ways Nussbaum's substantive test can't touch without begging the question.

Mitchell leans on Elizabeth Barnes's testimonial injustice argument: the default should be to take the reports of people whose lives have changed seriously. Overriding their testimony requires showing a specific distorting mechanism (manipulation, gaslighting, structural coercion), not just the bare fact of adaptation. Her key line: "Adaptation should not be regarded as an exceptional phenomenon: all preferences are contextually anchored to some extent, and as such responsive to environment and circumstantial change."

4. Callard (2018) — aspiration as forward-looking value-acquisition

Full article. Where Bovens/Baber/Mitchell all reason backward from a revised preference (how do we evaluate the shift that has occurred?), Agnes Callard's Aspiration: The Agency of Becoming reasons forward from the aspirational stance (how is rational action possible when the agent doesn't yet hold the values that would ground it?). Her central construct is the proleptic reason: a reason that points in the right direction but that the aspirant acknowledges she does not yet fully understand. The aspiring music student who attends concerts "because she's told they're valuable" has a reason, even though it's incomplete — her rational agency develops as the value is acquired, not after. Callard's decisive move for this literature: "Aspiration is a form of ethical movement, and nothing can move in a moment." Preference revision inside an aspirational trajectory is neither sour grapes nor character planning — it's the natural signature of value-learning. The aspirant is rational even though her grasp of why she's doing what she's doing is "not as good as the grasp she will have once she is capable of the simple, unified, and transparent agency she aspires towards." Callard also supplies the decisive aspiration ↔ ambition distinction: "Ambition is marked precisely by value-stasis. The ambitious person, qua ambitious, is engaged in getting what he wants, as opposed to learning what he wants." For Deep Ambition's central argument, that single sentence flips the conventional reading — the "ambitious" person racing up the default scoreboard is, in Callard's vocabulary, less aspirational than the person in the value-shift. Three supporting sources are also in the wiki: callard-replies-to-critics (Callard's responses to Katsafanas/Kraut/Sauvé Meyer/Paul in a 2021 symposium — sharpens the key defenses), callard-on-deliberation (a Boston Review essay on why personal decisions are harder than political decisions, and why "don't overthink it" is often the right advice), and callard-marriage-of-the-minds (Rachel Aviv's 2023 New Yorker profile in which Callard's own transformative value-shift — leaving her marriage to Ben Callard for Arnold Brooks — became a lived test case for her theory).

What the Four Papers Agree On

  • Mere preference change is not evidence of irrationality. Humans adapt constantly — hedonically, contextually, through learning. The bare fact of a shifted preference licenses no diagnosis.
  • The burden of proof runs toward the skeptic. To call a preference "adaptive" in the pejorative sense, you must show what specifically is wrong with it, not just that it changed after a setback.
  • Elster's intentionality criterion is insufficient. Bovens rejects it explicitly; Baber's counterexamples (information-based updating, commitment in Sen's sense) show preferences can be rationally revised without an explicit character-planning project; Mitchell's "learning and experience" category does the same; Callard's proleptic reasons are rational without being fully conscious or intentional in Elster's sense.
  • Self-deception is a separate problem. All four distinguish genuine SG from cases where the agent is suppressing a persisting preference. Baber treats the fox's declaration as pure self-deception; Bovens's theory explains why SG invites a self-deception reading without requiring it; Mitchell treats Barnes's testimonial-injustice argument as orthogonal to the question of authentic preference revision; Callard explicitly rejects reading aspiration as self-deception — it is instead a "painful confrontation with one's own inadequacy."
  • Time is a structural feature of rational preference revision. All four agree that single-moment flips are suspicious; genuine revision takes time. Callard makes this most explicit ("nothing can move in a moment"), but Bovens also foregrounds it.

What They Disagree On

  • Whether the concept is worth keeping at all. Baber says no — preference utilitarianism already explains the injustice of restricted option sets without invoking deformation. Bovens and Mitchell say yes — there really are irrational preference shifts, and we need a criterion that separates them from rational ones.
  • Where the criterion lives. Bovens puts it in the internal structure of practical reasoning (all-things-considered judgment). Mitchell puts it in the mechanism of formation (manipulation/coercion vs. learning). Baber puts it in the dispositional availability of the alternative (would you take it if offered?).
  • Whose cases count as paradigm instances. Nussbaum's oppressed-women cases, which motivated much of the literature, are diagnosed differently by each: Baber reclassifies them as rational satisficing, Mitchell uses them as the narrow subset where specific distorting mechanisms can be identified, Bovens's framework handles them as cases where the all-things-considered judgment hasn't actually restructured.

Why This Matters for Deep Ambition

The "Treading Water" chapter circles the question: how do you tell, from inside, whether your value shift from shallow to deep ambition is wisdom or cope? Adaptive preferences is the technical vocabulary for that exact problem. The three papers converge on a practical answer with four moves the book can use:

  1. Of course your preferences changed. Mitchell: preferences are contextually anchored; adaptation is universal. The mere fact that your wants have shifted after a setback is not evidence of anything.
  2. The diagnostic question is whether your reasons restructured too. Bovens: if the ranking of what kinds of considerations matter (relational depth vs. status, craft vs. prestige, presence vs. acceleration) has actually shifted, you're in CP territory. If only the ranking of alternatives flipped while the underlying reason-weights stayed the same, you're in SG territory. Self-interrogation has a specific form: not "do I really want this?" but "what reason-pairs do I take to be relevant here, and has their relative weight shifted, or am I holding them constant and just flipping the ranking?"
  3. The operational test is behavioral. Baber: if the old opportunity re-appeared tomorrow, would you jump at it? If yes, your preferences are still thwarted — you've just stopped letting yourself feel the frustration. If no, and you can watch a peer win the old game with curiosity rather than the flash, the shift was real.
  4. Timing is diagnostic. Bovens: genuine reason-restructuring takes time. A sudden post-failure flip is structurally suspicious. A slow drift over years, with many small reorderings, is structurally trustworthy. This aligns with Pals's (see narrative-identity) two-step redemption requirement — you have to sit with the loss long enough for the new reasons to actually restructure, not paper over the old ones.

Together, the four papers give Deep Ambition the deepest available philosophical answer to the cope accusation. Callard supplies the forward-looking phenomenology (proleptic reasons; acting-as-if is how you acquire values you don't yet hold); Bovens supplies the backward-looking rationality criterion (did your reasons actually restructure?); Baber supplies the operational test (would you jump?); Mitchell supplies the default stance (the burden is on the skeptic). Callard also supplies the book's sharpest one-line reframe: "Ambition is marked precisely by value-stasis. The ambitious person, qua ambitious, is engaged in getting what he wants, as opposed to learning what he wants." In her vocabulary, the reader racing up the default scoreboard is the less aspirational one; the reader in the Treading Water gap is actively learning values, which is the higher form of rational agency.

Operational Synthesis (for the book)

If the book wants a single diagnostic protocol for the "Treading Water" reader, the four papers together yield:

  1. Accept that preferences changed. This is universal and not diagnostic on its own (Mitchell).
  2. Check the reason-weights (Bovens). Has your ranking of what kinds of considerations matter actually restructured? Not just "do I say I value family over status now?" — but: when you face a specific decision, which considerations pull hardest, and is that different from two years ago?
  3. Check behavioral dispositions (Baber). If the old opportunity arrived tomorrow, would you take it? If yes, treat the new preferences with suspicion. If you'd decline without ambivalence, treat them as real.
  4. Check timing and structure (Bovens, Callard). Has this been a long reorientation with many small reorderings, or a sudden flip after a specific loss? Slow drift is more trustworthy than sudden flip. "Nothing can move in a moment."
  5. Default to believing the person whose life changed (Mitchell/Barnes). The burden of proof is on the skeptic — including the self-skeptic. Absent a specific showing of distorting mechanism, the revised preferences stand.
  6. Reframe the aspirational stance as rational, not self-deceptive (Callard). The reader doesn't have to resolve the "is this cope?" question before acting. Acting under proleptic reasons — reasons you acknowledge you don't yet fully hold — is how values are acquired. The appearance of acting-without-full-understanding is structurally identical to genuine aspiration from inside, and Callard's contribution is to show that this is not a bug but the defining feature of value-learning.
Thread · 0 replies+ add reply
no replies yet — be the first to write back to this article.