The Ethics of WellBeing Matthew Adler Duke University
The Ethics of Well-Being Matthew Adler, Duke University and LSE Warwick Summer School
Overview of Lecture • Three families of philosophical theories of well-being: preference-based, experientialist, objective-good • Ethical welfarism • Social welfare functions and the measurement of well-being; preference utility (decision utility). • How to choose among different theories of well-being: wellbeing as an ethical concept, reflective equilibrium • The case for a preference-based view of well-being • Preferences for happiness • If preference utility is the fundamental measure of well-being, what is the role of SWB surveys?
Warning! • What philosophers mean by “well-being” is not what psychologists and some economists mean by “well-being” • SWB (happiness) = some measure of the quality of an individual’s mental states (affects, feelings of satisfaction, feelings of happiness, etc. ) • Philosophers: “well-being” is goodness-for. A type of value concept: how good a life is for the subject. Flourishing, quality of life. Identifying the “right” account is an evaluative issue. Open question whether well-being is SWB. • Psychologists and some economists. “Well-being” is just a synonym for SWB • This lecture: uses “well-being” in philosophers’ sense
Three Families of Philosophical Theories of Well-Being • Some symbolism: x, y …; i …. ; (x; i); ≽ (∼, ≻); ≽W; weak preference/strict preference. • Preference based. (x; i) ≽W (y; i) iff i weakly prefers outcome x to outcome y. Preferences may be “laundered” (rational, informed, self-interested) • Experientialist. Whether (x; i) ≽W (y; i) depends on i’s mental states in x and y. (Affects, memories, perceptions, feelings …) Mental state supervenience: if i’s experiences in x and y are identical, then (x; i) ∼W (y; i) • Objective Good. Whether (x; i) ≽W (y; i) depends on i’s realization of various objective goods. Finnis: Life, knowledge, play, aesthetic experience, sociability, practical reasonableness, religion. Griffin: Accomplishment, autonomy, physical integrity, understanding, enjoyment, deep personal relations. Nussbaum: life, bodily health, bodily integrity, senses-imagination-thought, practical reason, affiliation, other species, play, control over one’s environment. Sher: moral goodness, rational activity, development of one’s abilities, having children and being a good parent, knowledge, awareness of true beauty. Different from both preference-based and experientialist accounts
The Difference between Preference. Based and Experientialist Theories • A preference-based view of well-being is not an experientialist view. Doesn’t satisfy mental state supervenience. It’s possible that i has the same mental states in x and y but i prefers x to y in virtue of nonexperiential features of his life. If so, according to the preference view, (x; i) ≻W (y; i). – In x, June’s spouse is faithful to her. In y, June’s spouse betrays her but she falsely believes he is faithful. If asked about these possible outcomes June prefers x to y. – In x, Matt is an accomplished academic with a strong reputation among his colleagues. In y, Matt’s scholarship is worthless and teaching is weak--his colleagues and students mock him behind his back--but Matt has deluded himself into believing he is accomplished. If asked about these possible outcomes, Matt prefers x to y. – In x, Sara has achieved a genuine understanding of advanced mathematics. In y, Sara falsely believes she understands advanced mathematics. If asked about these possible outcomes, Sara prefers x to y.
Nozick’s Experience Machine • Nozick (1974): “Suppose there were an experience machine that would give you any experience you desired. Superduper neuropsychologists could stimulate your brain so that you would think and feel you were writing a great novel, or making a friend, or reading an interesting book. All the time you would be floating in a tank, with electrodes attached to your brain. Should you plug into this machine for life, preprogramming your life’s experiences? If you are worried about missing out on desirable experiences, we can suppose that business enterprises have researched thoroughly the lives of many others. You can pick and choose from their large library or smorgasboard of experiences. … Would you plug in? …. . We learn that something matters to us in addition to experience by imagining an experience machine and then realizing we would not use it. • Nozick uses this thought experiment to conclude that people do or should prefer more than experiences. We can use it (less ambitiously) to illustrate that people may prefer more.
Ethical Welfarism • Strong welfarism: ethical requirements are reducible to the ethical goodness of outcomes, in turn reducible to the pattern of well-being. • Weak welfarism: The goodness of outcomes (in light of the pattern of wellbeing) is a significant component of ethics, along with non-welfarist factors (deontological constraints and options) • The social welfare function (SWF): a formal framework for ranking outcomes in light of well-being. (A formalization of welfarism. ) Adler, Well. Being and Fair Distribution (2012); Adler and Fleurbaey, Oxford Handbook of Well. Being and Public Policy (2016). Developed by theoretical welfare economists, used in various applied literatures (optimal tax theory, public finance, environmental economics, …) Can be used both to assess governmental policies and to measure social condition. Corresponds to cost-benefit analysis with distributional weights.
The SWF Framework • The framework has two key components: a well-being measure w(. ) for translating outcomes into well-being vectors; and a rule (the SWF) for ranking such vectors. w(. ) numbers are interpersonally comparable. • x ≽E y iff (w 1(x), w 2(x), …, w. N(x)) ranked by the SWF at least as good as (w 1(y), w 2(y), …, w. N(y)) • Plausible candidates for the SWF. Utilitarian. Well-being vectors assigned the score Σi wi and ranked according to those scores. Prioritarian. Wellbeing vectors assigned the score Σi g(wi), with g(. ) strictly increasing and concave, and ranked according to those scores. The utilitarian SWF: formalizes ethical tradition of utilitarianism going back to Bentham. The prioritarian SWF: Parfit. • The SWF framework is generic. It can be coupled with any account of wellbeing (preference-based, experientialist, objective). We arrive at the specific content of w(. ) by adopting a specific account.
Possible Axioms for an SWF • Pareto superiority: (3, 4, 10, 13) ≻ (3, 4, 10, 12) • Anonymity/impartiality: (7, 12, 4, 60) ∼ (12, 60, 4, 7) • Separability: (7, 100, 7) ≽ (4, 100, 12) iff (7, 7, 7, 7) ≽ (4, 7, 7, 12) • Continuity: If (1, 3, 50000) ≻ (1, 3, 6, 8), then (1, 3±ε, 50000, 50000) ≻ (1, 3, 6, 8) for ε sufficiently small • Pigou-Dalton: (3, 6, 8, 12) ≻ (3, 4, 10, 12)
A Prioritarian SWF g(w. H) g(w. H – Δw) Transformed wellbeing, g(w) g(w. L + Δw) g(w. L) w. L w. H − Δw w. H w. L + Δw Wellbeing, w
Why SWFs Need Interpersonally Comparable Well-Being Numbers one array of well-being numbers a second array with same intrapersonal information x y z Amy 49 64 36 4. 9 6. 4 3. 6 Barry 16 9 25 Utilitarian 65 73 61 20. 9 15. 4 28. 6 Prioritarian ∑√wi 11 11 11 6. 21 5. 53 6. 90
v. NM Utility Functions • A v. NM utility function, ui(. ), is a mathematical representation of her preferences over outcomes and outcome lotteries. i weakly prefers x to y iff ui(x) ≥ ui(y), and similarly for lotteries. v. NM utility functions are unique up to positive affine transformations. An individual’s preference utility in an outcome, ui(x), is just the number assigned to that outcome by a v. NM utility function representing her preferences. (“Preference utility” = “decision utility”). • On a preference-based theory of well-being, we can construct w(. ) based upon individuals’ v. NM utility functions. Harsanyi; Adler (2016). Recall that the SWF framework and w(. ) are generic. Having adopted a specific theory of well-being (the preference-based theory), we can now construct w(. ) in line with that theory, by setting well-being equal to preference utility.
Interpersonally Comparable Well-Being Numbers based upon v. NM functions Attribute bundle (a) a+ R w(a, R) = s(R) u (a) + t(R) a+++ Attribute bundle (a) a+ a+++ w(a, R*) s(R**)=1 t(R**)=0 w(a, R**) Attribute bundle (a) 50 150 250 1 5 4 s(R*)=50 t(R*)=0 u. R*(a) u. R**(a) 1 3 5 1 5 4 w(a, R*) s(R**)=1 t(R**)=10 w(a, R**) a+ 1 11 a++ 3 15 a+++ 5 14 s(R*)=1 t(R*)=0 Attribute bundle (a) a+ a+++ w(a, R*) s(R**)=100 t(R**)= − 99 w(a, R**) 4 6 8 1 401 301 s(R*)=1 t(R*)=3
How to Choose a Well-Being Theory • Well-being, for purposes of ethical welfarism (the SWF framework) is an ethical value. It is a value (goodness-for) that plays a central role in determining the ethical goodness of outcomes. We choose among theories of well-being by evaluative reasoning, specifically ethical reasoning (reflective equilibrium). • Well-being in this sense is not a scientific or descriptive concept. • Reflective equilibrium: “The key idea underlying the method of reflective equilibrium is that we ‘test’ various part of our system of moral beliefs against other beliefs we hold, seeking coherence …. For example, we might test the appropriateness of a purported principle of justice by seeing whether we can accept its implications in a broad range of cases and whether it accounts for those cases better than alternatives…. Our moral beliefs about particular cases count in the process. …. [However] even firmly held beliefs about particular cases may be revised. For example, if a principle incompatible with such a firmly held belief about a particular case accounts better than alternatives for an appropriate range of cases we seem equally confident about, …. we may revise our belief [about the case]. ” • In a democracy, the fundamental ethical decisions governing public policy are made by elected officials
The Case for a Preference View • Note: the preference view not only has strong philosophical support, but of course is the traditional view in economics. • The sovereignty argument: If Juan is a competent adult, then Juan is authoritative about what’s best for Juan. We should defer to Juan’s judgments about the source of value in his life. • Counterarguments (in favor of objective-good approach): (1) Actual preferences don’t warrant deference. (2) Looking to “laundered” preferences is no different from an objective good approach. (3) There’s no case for deferring to idealized preferences. • Counterarguments (in favor of experientialist approach): (1) An objective good approach, but with all the goods being experiential (affects, feelings of happiness, …). Bentham’s approach? (2) The preferences that matter for wellbeing are self-interested. Parfit’s “stranger on the train” case. But self-interested preferences are nothing other than preferences for experiences.
Preferences for Happiness • The preference view does not imply that the quality of individuals’ mental states is irrelevant to their well-being. That would be absurd. Rather, mental states are, potentially, one argument in an individual’s preference-utility function. ui(x) = ui(mi(x), ni(x)). The extent to which an individual prefers m-type attributes, as opposed to n-type attributes, is up to her. • Lecture 2: Preferences for Happiness. Reports from a novel survey format (Adler/Dolan/Kavetsos [2017]) used to test the strength of individual preferences for m-type attributes versus n-type attributes.
The Role of SWB Surveys? • On a preference-based view, the ethical measure of well-being, w(. ), is based upon v. NM utility functions, ui(. ). What is the role of SWB surveys in this picture? Two quite different possibilities. • (1) Life-satisfaction surveys as overall evidence of preference utility (along with stated-preference surveys and behavioral evidence). LSi(x) as evidence of ui(x). • (2) An SWB survey measures how well an individual is doing with respect to an m-type attribute that is one of the arguments in the individual’s preference-utility function. For example, imagine that an individual prefers to feel pleasant affects (one of the m attributes) , to be in good physical health (one of the n attributes). ui() = (… pleasant affecti, …, healthi, …. ). Then we measure pleasant affecti using a hedonic SWB survey, healthi using a health metric, and ui(. ) as a function of these scores plus information about i is doing with respect to the other arguments in her utility function.
- Slides: 17