Below is a draft of a rather short section of “Experimental Moral Philosophy” (co-authored with Don Loeb for the Stanford Encyclopedia of Philosophy) on xphi and wellbeing. As always, comments, questions, criticisms, and suggestions are most welcome.
The study of wellbeing and happiness has recently come back into vogue in both psychology (Kahneman, Diener, & Schwartz 2003; Seligman 2011) and philosophy (Haybron 2008), including experimental philosophy (Braddock 2010; Phillips, Misenheimer, & Knobe 2011). It is helpful in this context to draw a pair of distinctions, even if those distinctions end up getting blurred by further investigation. First, we need to distinguish between the notion of a life that goes well for the one who lives it and a morally good life. It could turn out that these are extensionally identical, or that one is a necessary condition for the other, but at first blush they appear to be different concepts. Second, we need to distinguish between a hedonically good life and an overall good life. As with the first distinction, it might turn out that a hedonically good life just is an overall good life, but that would be a discovery, not something we can simply take for granted.
With these distinctions in hand, there are a number of interesting experimental results to consider. First, in the realm of hedonic evaluation, there are marked divergences between the aggregate sums of in-the-moment pleasures and pains and ex post memories of pleasures and pains. For example, the remembered level of pain of a colonoscopy is well-predicted by the average of the worst momentary level of pain and the final level of pain; furthermore, the duration of the procedure has no measurable effect on ex post pain ratings (Redelmeier & Kahneman 1996). What this means is that people’s after-the-fact summaries of their hedonic experiences are not simple integrals with respect to time of momentary hedonic tone. For example, if the colonoscopy were functionally completed after minute 5, but arbitrarily prolonged for another 5 minutes so that the final level of pain was less than it was at the end of minute 10 than at the end of minute 5, the patient would retrospectively evaluate the experience as less painful – even though the first five minutes of the ten-minute procedure were phenomenologically indistinguishable from the whole of the five-minute procedure. If you’re inclined, like Bentham (1789/1961) and Singer (Singer & de Lazari-Radek forthcoming), to think that pleasure is the only intrinsic good and pain the only intrinsic ill, then such experimental work will make you distrust retrospective evaluations of goodness and badness.
Another interesting set of results centers on the idea of virtue as a pre-requisite for wellbeing. If the results of Braddock (2010) and Phillips, Misenheimer, & Knobe (2011) withstand scrutiny and replication, it would seem that ordinary people are willing to judge a life as good for the one living it only if it is full of positive moods and affects, and is virtuous, or at least not vicious. This result resonates with the empirically-supported views of Seligman (2011) that happiness contingently turns out to be achievable only in a life that includes both a good deal of positive emotion and the exercise of both moral and intellectual virtues. If these results are solid, not only do ordinary people think that both virtue and pleasure are necessary for wellbeing, but in fact it’s true that both virtue and pleasure are necessary for wellbeing.
 See Haybron (2008).