A Philosophy for the Science of Well-Being
A review of Anna Alexandrova’s recent book
I’m interested in the good life. I want to know what it means to live well. And I’m convinced that academic research on such questions is best done in an interdisciplinary fashion. Philosophers and scientists (in a range of fields) should be working together to understand the nature of well-being and how it can be promoted.
This article is a review of Alexandrova’s book, which makes a nice contribution to a burgeoning academic literature at the interface of normative, philosophical theorizing and empirical, social scientific research. As the author herself writes, “this book is about well-being as an object of science: how science should define well-being, how it should measure it, and the role of philosophy in all this” (p. xv). Philosophy, Alexandrova recognizes, is relevant to this scientific work. Yet sometimes the best way to advance the state of knowledge is to ignore nagging philosophical questions — at least for the time being. A central task for the book is thus to distinguish the philosophical questions that we need to address from those that it can safely ignore to go off and do some science.
Part I: Tools for Philosophy
Part I, argues that philosophers should not try to give fully general accounts of what’s good for a person — any person no matter who, when or where. Yet this is what all the standard theories in the philosophical literature do. The standard theories are: hedonism, which claims that only pleasure is good for a person; desire-satisfaction theory, which claims that only the satisfaction of one’s desires is good; and “objective list” theories, which claim that getting certain objectively valuable things is good for a person. (See here for an intro to theories of well-being.) These theories are meant to be all-encompassing in the sense that they say for anyone anywhere what’s good is XYZ. Alexandrova calls these “high-level theories”.
In contrast, Alexandrova thinks that philosophers should aim to produce contextual theories, theories of what it takes for members of specific population groups to do well. (In Chapter 3 she gives a theory of children’s well-being as a case study.) Her case against high-level theories is based on the claim that any “philosophy of the science of well-being worth its salt must come with recommendations for how researchers should choose their constructs” (p. xxxv). “Constructs” are the phenomena that social scientific (particularly psychological) theories appeal to. Alexandrova claims that mid-level theories are useful for science because they can guide construct choice. High-level theories cannot. Since philosophers love high-level theories so much, she claims, well-being scientists have ignored philosophical theories (pp. xxix–xlv).
I have two objections to Alexandrova’s claims about high-level theories: (1) high-level theories can actually provide guidance for construct choice; and (2) this is evidenced by the fact that they already do.
It’s not at all clear to me why Alexandrova thinks high-level theories can’t guide scientists. Hedonism provides remarkably clear guidance: measure pleasure and pain. Objective list theories, similarly, just are lists of things one could measure (e.g., knowledge, loving relationships, and achievement). Moreover, it seems quite clear that the standard constructs within the science of well-being are based directly off of these high-level philosophical theories. Three of the standard constructs are: hedonia, life satisfaction, and eudaimonia. These are drawn directly from the standard philosophical theories. Hedonia — clearly inspired by hedonism — is the ratio of one’s positive to negative affect. The researchers who came up with this construct even cite hedonist philosophers like Jeremy Bentham (see here, for instance). Life satisfaction is a cognitive assessment of how well one’s life is going, which looks a lot like an operationalization of the desire-satisfaction theory. And eudaimonia involves “actualizing one’s human potentials” (Deci & Ryan, 2008, p. 2) — clearly drawn from an Aristotelian version of the objective list theory.
If Alexandrova just wants to say that philosophical theories of well-being tend to have underdeveloped “how to” sections, then I concur. But clearly — if nothing else — they at least offer inspiration for science. So, it seems that, rather than abandon high-level theorizing, we should just do a better job of applying and recommending these high-level theories for scientific research. I for one wholeheartedly endorse this recommendation.
Part II: Tools for Science
Part II of the book defends and then criticizes the current scientific practice. Chapter 4 addresses worries about the “objectivity” of a science of well-being. The concern here is that investigations into well-being will necessarily be value-laden. Thus, the science of well-being cannot live up to the (putative) ideal of impartiality — not making value judgments or giving guidance, only describing the way things are. Alexandrova argues that, while this science is certainly value-laden, this is unproblematic. The kind of objectivity worth having, she argues, requires only that the evaluative judgments and presuppositions are made fully explicit, and subjected to appropriate forms of public scrutiny. Alexandrova recommends that researchers state explicitly their normative assumptions in the methods section of their academic publications.
Alexandrova suggests that researchers (and especially government bodies with policy interests in well-being) conduct “deliberative polls”.
Groups of deliberators could be presented with various options for conceptualising well-being… and with the relative advantages of each option normatively and practically. The deliberators will attempt to reach agreement according to whatever consensus-building and voting rules they decide to put in place. Even if not everyone favours the values that survive such an exercise, the resulting consensus has some legitimacy and deserves trust at least from those whose views are admissible in a democracy and have been heard. (pp. 103–104)
These polls, Alexandrova thinks, should include scientists, philosophers and other scholars as well as lay members of the relevant populations. These recommendations are a good place to start. I’d like to see more attention from philosophers on these questions — especially given that governments are already monitoring and attempting to influence their citizens’ well-being.
In the final chapter, Alexandrova turns a critical eye to the scientific practice of construct selection and measure development, which she argues constitutes a problematic kind of “theory avoidance”.
Systematic philosophical accounts of well-being and related constructs offer the only opportunity to theorise normatively, that is, to ask what happiness, or life satisfaction, or flourishing are such that they can play the role that these thick concepts play in people’s lives. And yet this normative test is not part of measure validation, except perhaps as an initial inspiration. This is insufficient if the outcome is supposed to be a measure of a property denoted by a thick concept. Part of measure validation should be whether the measure captures a construct that is worth caring about. (p. 142)
The standard measure validation procedure goes (roughly) as follows. The researchers will offer a pre-theoretical gloss of the concept in question (e.g., happiness), sometimes citing dictionary definitions. They use this gloss to produce a lengthy list of questionnaire items which are then given to participants. The responses from this initial sample are then factor analyzed. (This is a statistical technique for cutting that list down while making sure that the shorter list gets basically the same results. See here.) The resulting factors are then tested for “convergent validity”. To display convergent validity is to be correlated to other constructs in intuitive ways, or as predicted by a background theory. For instance, we would expect an accurate measure of generosity to be (amongst other things) positively correlated with altruism and compassion, and negatively associated with egotism and psychopathy. So, in developing a measure of generosity, researchers would administer their new questionnaire along with previously validated measures of these other constructs and look for these relationships. If a measure passes this battery of psychometric tests, it is said to be validated.
One problem here is that participants’ intuitive, pre-theoretical understanding of the central concepts may be normatively lacking. As Alexandrova reminds us, there are better and worse theories of well-being, happiness, flourishing, etc. By not engaging in philosophical reflection on these topics, the science of well-being misses an important opportunity to ensure that it is researching something that actually matters.
Another problem appears in the process of testing convergent validity. As indicated, the process involves looking at correlations between the measure under development and existing constructs. Consider, for instance, the Satisfaction with Life Scale (SWLS), and the Positive Affect Negative Affect Schedule (PANAS). The SWLS correlates strongly with wealth and social status, while the PANAS correlates strongly with a sense of personal growth and closeness with others. Either of these could plausibly be seen as demonstrating good convergent validity. Wealth and social status do seem like they should be positively associated with well-being. But, then, so do personal growth and social connection.
The important question is: Which of these is (more) important for one’s well-being? Answering this question requires making a value judgment. (Speaking for myself, the latter seem much more important.) It requires saying that money isn’t good for people if it doesn’t improve the quality of one’s experience, or that social status is a good thing in itself, etc. However, the standard approach amongst scientists at the moment is to abstain from explicitly making such judgments. Rather, researchers tacitly make these judgments in their choice of measures or and research questions. Researchers who favor the SWLS, for instance, are tacitly assuming a conception of the good life according to which wealth and status are more closely related to well-being than growth and close relationships. This keeps the evaluative presuppositions of the research concealed, preventing them from being subjected to public scrutiny. This is a problem for both the quality and legitimacy of the research.
I want to conclude this review by reiterating how important it is for philosophers to do the kind of work that Alexandrova undertakes in this book. Despite my criticisms, I highly recommend the book — particularly Part II — to everyone interested in well-being. The book’s central questions and argument should be agenda-setters for future philosophical work of the science of well-being.
