frameworks / mitchell-adapted-vs-adaptive-preferences · article
frameworks/ · 2,022 w · 9 min · ✎ dialogue

Mitchell — Adapted vs. Adaptive Preferences

Polly Mitchell's 2018 paper "Adaptive Preferences, Adapted Preferences" (Mind, Vol. 0.0, doi:10.1093/mind/fzy020) argues that the vocabulary philosophers have used to discount the preferences of the oppressed, the defeated, and the diminished has been smuggled into health economics to discount the testimony of patients — and that this move doesn't survive scrutiny. Mitchell, then at University College London, introduces a crucial distinction most of the literature elides: adapted preferences are not the same as adaptive preferences. Adaptation is ubiquitous, often rational, and frequently positive. Adaptive preferences are the narrower subset that are irrational or malformed. Conflating the two lets policy-makers dismiss patient self-reports on the grounds that "they've just adapted," when in fact most adaptation is the ordinary, defensible human capacity to respond to environmental change. For Jason's deep-ambition-book-thesis — and particularly the "Treading Water" chapter on the cope/defeat ambiguity — Mitchell provides the cleanest philosophical argument that the appearance of preference revision (sour grapes) does not by itself constitute irrationality. The burden of proof is on the skeptic, not on the person whose life has changed.

The Empirical Puzzle: The Disability Paradox

Mitchell opens with what health economists call the disability paradox: people who have not experienced a given disease or disability consistently judge those states to be worse than people who actually live with them report them to be. In Ashby et al.'s breast-cancer study, women without experience of breast cancer rated the utility of living with a mastectomy significantly lower than women who had actually undergone mastectomy. In Boyd et al.'s colostomy study, patients with colostomies assigned higher utilities to their state than patients treated for the same rectal cancer without colostomy. The pattern is documented across many diseases and disabilities (Ubel et al. 2005; Dolan & Kahneman 2008).

This creates a policy problem. QALY calculations (quality-adjusted life years) depend on utility values, and the choice of whose utilities to use — public (the unaffected majority) or patient (those with the condition) — radically changes which treatments are cost-effective and which populations get resources. If patient utilities are "correct," public values systematically over-fund conditions patients find manageable. If public utilities are "correct," patient reports can be discounted as distorted, and many treatments look less valuable than patients experience them to be.

One way to resolve the puzzle in favor of public preferences is to argue that patients have adapted to their condition in ways that compromise their testimony. This is the move Mitchell is setting up to refute.

The Three Accounts of Adaptive Preferences

Mitchell surveys the three dominant philosophical accounts of adaptive preferences and argues that none of them actually show that patient adaptation is irrational:

1. Elster's procedural account (Sour Grapes, 1983). Elster distinguishes adaptive preference formation from deliberate character planning. The fox in La Fontaine who decides the unreachable grapes are sour has adapted non-autonomously — a causal, non-conscious process driven by the need to reduce cognitive-dissonance frustration. A "character-planning fox" who decides to cultivate a taste for the sour cherries within reach has adapted autonomously, with a genuine second-order preference. For Elster, the adaptive fox is irrational; the character-planning fox is not.

2. Bovens's formal account (1992). Bovens criticizes Elster's conscious/non-conscious line and argues instead that irrationality consists in conflict within a preference set. When the sour-grapes fox says "I don't want these grapes" while holding a general preference to eat grapes, his token preference contradicts his type preference — that's what makes it irrational. The character-planning fox's preferences cohere; the sour-grapes fox's don't. (Bruckner 2009 and Christman 2014 offer formal refinements: irrational preferences are those not reflectively endorsed.)

3. Nussbaum's substantive account (2001). Nussbaum identifies adaptive preferences by their content: preferences are adaptive when they favor outcomes that ought not to be preferred against a normative list of ten central human capabilities. Vasanti accepting marital abuse, Jayamma accepting differential pay, malnourished Andhra Pradesh women not considering themselves malnourished — these are adaptive because they conflict with bodily health and dignity, regardless of how they were formed. Khader's perfectionist account (2011) is structurally similar: adaptive preferences are those inconsistent with basic flourishing, formed under conditions non-conducive to flourishing, and that would not have formed under conditions conducive to flourishing.

Mitchell's Core Move: "Adapted" Is Not "Adaptive"

Mitchell's central claim is captured in a quiet but devastating line:

"Adapted preferences should not be treated as synonymous with adaptive preferences."

Adaptation, in the hedonic sense, is ubiquitous: lottery winners return to baseline within months (Brickman et al. 1978); so do accident survivors; so do newlyweds, the bereaved, and people experiencing income growth. Most adaptation happens to everyone. It is not a pathology of the oppressed. Adaptive preferences, in Elster's or Bovens's or Nussbaum's strict sense, are a narrower subset — the ones that are irrational, malformed, or malign.

Mitchell then walks through each of the three accounts and argues that patient adaptation doesn't actually fit them:

  • Against Elster: Much patient adaptation looks more like Elster's own "learning and experience" category than like sour grapes. A patient who comes to a more holistic view of health — one that values "the ability to live a full, purposive, and valuable life" rather than high physical functioning — has not performed cognitive dissonance reduction. She has acquired new information about the option set and updated her preferences informedly. "Preference change through learning is not, for Elster, irrational, because it generates informed preferences."
  • Against Bovens: Patient preferences don't uniformly conflict with the patient's other preferences. Someone who finds new meaning in creative writing after losing the ability to cycle has adopted a comprehensive change in goals, not a contradictory token preference. Much patient adaptation sits in what Bovens himself calls the "large grey area" between sour grapes and character planning — and in the formal accounts that require reflective endorsement (Bruckner, Christman), most such preferences turn out rational.
  • Against Nussbaum: The argument is structurally circular. To classify patient preferences as adaptive because they tolerate "sub-optimal" health states presupposes that health is measurable independently of patient testimony. But whether patient testimony should count is exactly the question at issue. "To reject patient testimony on the grounds that it contradicts objective measures of health begs the question, if the existence of the testimony serves to question the claim that health should be measured objectively."

The Barnes Argument: Testimonial Injustice

Mitchell leans heavily on Elizabeth Barnes's work (The Minority Body, 2016) to strengthen the case. Barnes argues that unsubstantiated diagnoses of adaptive preferences are a form of testimonial injustice — we are refusing to take the reports of disabled people seriously on the grounds that their disability has distorted their reporting, which is precisely what a theory of disability-as-suboptimal is supposed to be proving. As Mitchell puts it, channeling Barnes:

"To say that someone's preferences are incorrect or misguided is to say that other people can determine their preferences better than they can themselves. Unjustified testimonial denial is a serious injustice, so great care should be taken to ensure that the preferences in question really are defective."

The bar for attributing adaptive preferences should be set very high. Barnes's own list of clear cases — domestic abuse, forced incarceration, deeply sexist and racist contexts — all share features of psychological manipulation, gas-lighting, or structural inequality that produce inferiority. Ordinary illness and disability usually don't involve that machinery.

Key Distinction: Six Elements of Patient Adaptation

Mitchell draws on Menzel et al. (2002) for a more granular taxonomy. Patient adaptation is a combination of:

  1. Cognitive denial of functional health state
  2. Suppressed recognition of what full health would be like
  3. Skill enhancement
  4. Activity adjustment (new interests, new friendships)
  5. Substantive goal adjustment
  6. Altered conception of health; lowered expectations; heightened stoicism

(1), (2), and (6) look like Elster's sour grapes. They involve denial, suppression, and lowered expectations — the "diminished life" pattern. But (3), (4), and (5) are different in kind. Skill enhancement, new activities, and goal reprioritization are, as Mitchell puts it, "less obvious candidates for adaptive preferences" — they are comprehensive preference changes, often made consciously, often resulting in what looks much more like character planning than sour grapes. Cycling becomes creative writing. Paraplegic athletes find disability sports. A patient discovers that "full" health is not identical with high physical functioning but with the ability to live a purposive life. These are not malformations; they are lives.

Why This Matters for Deep Ambition

The "Treading Water" chapter in deep-ambition-book-thesis circles a specific question: how do you tell the difference between letting go (wisdom) and giving up (cope)? The synthesis at outputs/2026-04-18--letting-go-vs-giving-up-psychology-synthesis.md already flagged Elster as a gap; Mitchell is the paper that most cleanly operationalizes the answer.

Three transferable moves from Mitchell:

1. Adaptation is ubiquitous, not pathological. Everyone adapts. Hedonic adaptation is a universal response to environmental change, not a special failure mode of the defeated. When the shallow-to-deep ambition reader worries their value-shift is cope because their preferences have changed, Mitchell's response is: of course they have — preferences are contextually anchored to begin with. The mere fact of preference change is not evidence of irrationality.

2. The burden of proof is on the skeptic. Barnes's testimonial-injustice argument, which Mitchell adopts, flips the default. The person whose life has changed is the primary authority on that life. To override their testimony requires a specific showing of distorting mechanism — manipulation, gaslighting, structural coercion. Absent that showing, their account stands. Translated for Deep Ambition: the reader doesn't need to prove their new values are real; the observer needs to show what specific distorting mechanism would have produced them falsely.

3. Elster's "learning and experience" is the rational middle path. Mitchell highlights a category in Elster's own work that is usually skipped: preference change that arises not from dissonance reduction and not from character planning, but from acquiring information about the option set through lived experience. The reader who leaves finance and discovers they actually prefer the life they've built has not committed sour grapes and has not pre-planned a character change. They have learned. This is the philosophically clean vocabulary for what Jason's coaching frames call "earning the reframe through lived evidence."

Paired with Agnes Callard's aspiration work (already in the thesis), Mitchell closes the most dangerous objection: that Deep Ambition is sophisticated cope. Mitchell's answer is that the cope accusation only bites if you already know the new preferences are malformed — and the only way you would know that is by appeal to the very preferences under examination.

A Quotable 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."

This is the single sentence to carry into the book. The framing move it enables is: of course your preferences changed. Everyone's do. The question isn't whether they changed but whether the change was the result of learning.

Thread · 0 replies+ add reply
no replies yet — be the first to write back to this article.