A letter reviewing Seligman's 'Flourish'

Dear <redacted>

I read Seligman yesterday. I must admit that I liked it more than I thought I would. I had associated Seligman with the work he did in Authentic Happiness, which I thought was rather poor from a holistic ‘well-being’ point of view, but he’s changed his position quite dramatically. I’m sympathetic to PERMA and a lot of the actions he wants to take to increase well-being at the social level, especially the education programs. The case study from the army was outstanding and I would love to get my hands on that data set. Look forward to the results. I am also very glad to see his emphasis on non-emotional factors in well-being, and his awareness that the meaning of the term happiness is rather confused in the literature. Finally, I loved the sections on drugs and CBT in treating depression because it confirmed all my biases. In particular, that a large number of people suffering from depression are suffering from structural changes in the psychological systems they use to make sense of themselves in the world as opposed to biological changes in their brain chemistry. Drug therapies cannot treat these structural factors and CBT is also unlikely to unless they have a behavioural source. 

I can't believe they call this visionary...bitch plz

However, I was a little disappointed that after alluding to structural factors like life narratives, ethics, values and meaning, Seligman engaged with them very little thereafter, and most of the ‘techniques’ in the book were from the Authentic Happiness days. The personal strengths metric is great, but I think it leaves a huge amount out of the picture. I will elaborate my misgivings below (about 3000 words):

The most fundamental is that a lot of the concepts that make up PERMA are quite shallow and under-theorised (and as someone who left philosophy because I was fed up with theory, I do not use that term lightly). This is despite there being quite good theory on a lot of them, from both philosophy and psychology. My specialty is the M for meaning, so let me focus on that here. I have quite an extensive section on relatedness as well in the paper that I sent you.

Seligman introduces meaning with the following anecdote. He was standing in front of the mirror one day in late middle-age and thought: isn’t a pleasant life not just fidgeting until you die? You must need more—you must need meaning. This notion of fidgeting until you die is what philosophers call existential dread. I suspect that the reason why most top academics don’t experience it until middle age or indeed, ever (as in the case of Kahneman or Bertrand Russell, whose lives I find instructive on a lot of things), is that their academic pursuits are a calling. If you are called you never need to question the meaningfulness of your life because such meaning is granted to you by your deepest inner drives. I am remiss to talk about myself, but I think my turn towards this material is instructive. From age 5–18, if someone had asked me what I wanted to do I would have said: ‘be a lawyer’. When I got a little older I would have added ‘and then a judge’. Then I got to law school and it was absolutely not what I had signed up for. So I spent the next five years (perhaps 8), in the wilderness. When you’re that deep in the wilderness you go looking not so much for your calling as an explanation for the nature of your existential vacuum, so that’s my specialty.

The question that comes immediately after ‘am I just fidgeting until I die’ is ‘is what I do transcendentally meaningful’, which is to say ‘won’t it crumble to dust after I die and is it not therefore actually not meaningful at all?’ This is what philosophers call despair, and such nihilism is inevitable once God is dead (c. 1890). This is the topic of all of Nietzsche’s work, and Kierkegaard’s as well. They are both mentioned but never discussed by Seligman. We have tried to replace the cosmic order that ensures the just will prosper with various ethical axioms like the veil of ignorance and the utilitarian criterion, but without a cosmic justice system these things are only as powerful as they are rationally compelling. The axioms themselves aren’t written into the firmament—people have to choose to abide by them, and they can only do this with their reason. Such axioms are then no longer moral in the sense that they are objectively binding, but merely ethical in that they are concerned with normative statements rather than strictly empirical statements. <redacted>

As the paragraph above alludes, after we have answered the question of values (meanings) we can also ask ‘is what I am doing good or evil?’ There is also ‘who am I?’, which is hard to answer if you don’t have answers to the other questions because your values and ethics are fundamental constituents of your identity. These kinds of questions will need to be answered if PERMA is to become a vertically integrated system (down to metaphysics and up to social philosophy) and not just a slogan.

Now there has been a huge amount of work on these questions, notably from the existentialists in philosophy, from the founders of psychoanalysis, and more recently from researchers in self-discrepancy theory, self-determination theory and life narratives theory (and me). Much of it is empirically verified, but these things are difficult to package neatly into empirical analysis so the going is very slow.

I think in our first meeting I mentioned to you that I am writing a philosophy book that summarises this material and extends it. It is nearly done. I attach the summary, which also happens to be the introduction. In it I lay out a fairly comprehensive theory of meaning, ethics and identity, one that is firmly grounded in political theory, metaphysics and the subjectivity of value judgements. It would plug easily into PERMA, especially because it explains how individual identity affirmation is necessarily a social project, which provides a philosophically rich basis for the relationships part of the PERMA framework.

I can’t summarise it here. Please read the attached document. The coalescence of being, which is the idea at the heart of the book (and its title), provides something that Seligman mentions but then never develops, which is a self-reinforcing system for achieving well-being. It allows the individual to consciously and systematically hone their psychological systems until they are capable of generating meaningful experiences that in turn produce emotional happiness. This psychological system also allows for the rapid analysis of exogenous shocks and the formulation of appropriate responses, which builds resilience in a way that parlour tricks like gratitude do not.

Critically, I think, the coalescence of being accesses and shapes the fundamental psychological structures of values and ethics that underpin behaviour. In so doing, they are able to fundamentally change a person at the roots rather than at the surface, which is where CBT and many positive psych techniques operate, like savouring. These techniques are useful (though their Californian hue gives me the heeby jeebies), but they are superficial. They do not get at the person’s identity, at their existential dread, at their human condition.

Allow me one brief sidebar. Seligman’s personal strengths packet is really great, but what underpins it? What generates strengths and how can we improve our existing ones in a manner that we have ownership of? He says most people have 3-4 strengths from among the 16 or so where they score 9 or 10 out of 10, and 3–5 weaknesses where they score 4-6. I was 9 or 10 on every single criterion except humility (the irony is not lost on me). This isn’t because I’m special (well, the intelligence and athleticism stuff is maybe just genetics); it’s because the coalescence of being allows you to systematically improve your citizenship, integrity, wisdom, willpower etc. In many ways it is founded on these concepts, so it is self-reinforcing. Moreover, the coalescence of being allows you to do this honing in a manner that is entirely your own and owned by you, rather than imposed from the outside. What do I mean by this? In religion we must conform to an external standard of ‘right’ conduct. We must break our own will to follow the will of God. In modern psychological research, the individual simply needs to follow the recommendations of the average effect across large sample sizes. But all the philosophy since the death of God and a wealth of psychological research suggests that the constituents of well-being are profoundly subjective and individual in nature, so using the average effect is misguided. There is no one true path to happiness—you need to find your own. The two Hellenic maxims are on point: know thyself and become what you are. Well-being must come from within. Certainly you must interact with ideas, reasons and evidence from outside, but you must take ownership of the final construct that emerges. Pure comportment won’t work.  

So that was my first gripe. Seligman, like most of his squad (Diener, Lyubomirsky etc), has mostly done research on the emotion of happiness rather than on existential satisfaction. Flourish really just staples some stuff on. There isn’t much deepening. Emotional happiness is largely exogenously determined (shoot, I missed the bus, that sort of thing), so behavioural tricks like savouring work. Existential satisfaction, on the other hand, can’t be tackled with these superficial practices.

My second criticism, which will surprise nobody, is that economics is misrepresented. Rational choice theory does not use people’s happiness as the unit of account; it uses utility, which is deliberately vague and left up to the individual in question. This is to avoid making value judgements, which is something I will return to in a moment. Moreover, rational choice theory is a decision theory—it is not interested in what actually makes people happy; it is interested in what they think will give them utility as it is this prediction (which obviously comes with uncertainty) that determines how they behave. Economics is a behavioural science, which is why it dominates public policy. We can infer the preferences that inform individuals’ behaviour from their behaviour. This is why economists have only limited interest in expressed as opposed to revealed preferences. The revealed preferences do not need to be rational. They do not necessarily make the agent ‘happy’ (as you always say: people have dumb preferences). They don’t even need to be fitness-enhancing. A human can only act in a manner that it thinks is in its interest, unless it is under mind control. An ‘altruist’ is acting in such a way because it thinks such behaviour is in its interest, through group selection, collective action dynamics or reciprocity, among other explanations. To suggest that economics thinks that humans are self-interested in a ‘dumb’ manner is to have stopped reading at either the end of the 19^th century or the end of an introductory course in microeconomics (or to be speaking to a mathematician working within mathematical models with heaps of assumptions built in).

This over-simplification of economics extends through the book but is especially prominent in the concluding chapter. GDP does not go up when a window breaks. This is the glass window fallacy. Yes, a trader comes and fixes the window, increasing economic activity. But if the window hadn’t broken then its built capital would have remained in the system and could have been used for some other production. The trader wouldn’t have been demanded, so his labour inputs could have been applied somewhere else. We can never see these effects because we don’t have the counterfactuals, but a broken window makes GDP go down.

That said, there are plenty of good arguments to be made that alongside GDP as a measure of activity we need some measure of wealth, like the four capitals, as this can allow us to better analyse sustainability. Two points are worth making here. The first is that some of the biggest names in economists are on this (journal article) (short summary on pg. 34). The second is that such things are crazy hard to measure in a way that GDP is not. 

This brings me to the issue of measurement, where I think Seligman made some great and some not so great points. I agree that well-being should be the yardstick by which we measure social progress; it’s why I came across to economics and public policy. But I don’t think it should replace GDP. GDP not only correlates extremely well with just about any other objective indicator of well-being (e.g. environmental quality, health, education etc.), but is also critical for managing the unemployment rate, and no single item has more consistently demonstrated powerful effects on all dimensions of well-being as unemployment. Moreover, if your society isn’t ticking over at least a little then you don’t have any Schumpertarian dynamism and you are heading for stagnation a la France, which is inter-generationally iniquitous. This doesn’t hold with an aging population, but as a global society we are decades from an aging population, so in the meantime, we need growth and migration (though of course growth without destroying natural capital). I think suggesting that economists are fixated on wealth is unkind to economists, who are often very human people with very humanitarian agendas, they just recognise that wealth is one of the few amoral things that we can get decent data on and that correlates closely with human improvement.

Next point on measurement: the sophisticated packets psychologists use to measure well-being are excellent. My view is that if we are going to do national well-being accounts anytime soon then random samples of ~1000 citizens filling these out on an annual basis are our best bet. Seligman is at the cutting edge of developing these packets.

In contrast, the 1-10 life satisfaction scales I think are horribly mired in measurement error. As Seligman notes, about 70% of the responses come from mood on average, but we don’t want states involved in people’s moods. For one, that is horrifyingly paternalistic. We might as well be in Huxley’s Brave New World. We will be putting anti-depressants and ecstasy in the water supply in no time. Second, bad mood is important for good mood; suffering is important for achievement. You don’t win marathons without working. You don’t finish PhDs without anguish. Third, psychological health requires you to be down sometimes. We would be a weird society if we saw grief after the death of loved ones as a sign of regression and pathology. There are numerous other problems with these scales—endogeneity with the independent variables, measurement error in the independent variables, arbitrary cut-offs across scale points and a bunch of other stuff. The big one though is that happiness is arguably not on a scale, so by using a scale, you impose a structure on the data that impedes your ability to correctly measure the phenomenon you are interested in. I elaborate on this quite a lot in the paper I sent you. The so called adaptation effect might just be re-scaling i.e. the qualitative meaning of scales change over time, but we don’t see that. You could still describe this as preference drift, but there is no way to be sure that is what is going on rather than people simply being on a structurally higher level. Indeed, there is evidence that structural change is what is actually going on rather than preference drift (c.f. the study of Tongan migrants). I have elaborated on this in the document about the empirical project I sent you last time. <redacted>, but the current status quo in research seems fine with just running regressions on life satisfaction scales, so you can safely ignore it.

Despite the abundance of evidence that life satisfaction scales are riddled with measurement error and the profound theoretical problems with interpreting them, people are still hanging their hats on the counterintuitive results they produce. Seligman does this at length. He suggests that people in underdeveloped countries are just as happy as those in developed ones, even though this is not supported by the most up to date literature (he even cites Stevenson and Wolfers and then just brushes it off) or by migration flows. He makes the standard points about flat life satisfaction in America and Japan despite GDP growth; results that would be totally refuted were my suspicions about rescaling to be confirmed. There is also a similar result for China over the past 20 years of growth, which suggests that people were happier in the immediate aftermath of the Tiananmen incident (which was nationwide and followed by huge crackdowns) than in the next decades of liberalisation and growth. This seems bizarre. I suspect that Seligman is here overly reliant on discussions with Layard. I like a lot of what Layard writes about, especially with regards to positional competition, but he is on the left of happiness economists who are on the left of economists, so he is not exactly a balanced reference point on these matters. Seligman perhaps needs to diversify a bit.

This is not to say that I don’t agree with Seligman’s political proposals; I just think that if you want people to take ideas seriously then you need to speak their language instead of using dodgy data to promote an agenda. This leads into my last gripe with Seligman’s work, which is its political naivety, not just in terms of praxis but also in terms of political philosophy.

The beauty of GDP is that it is amoral. Economics goes out of its way to be amoral, and people who claim that economists comes imbedded with values mistake economics for economists. Economists come imbedded with values. I have also argued in the past that if you teach a blank slate undergraduate economics without any context then they are liable to adopt the objects of economic inquiry—like efficiency—as values. But the economic models themselves are valueless. They tell you about efficiency, for example; they do not say that efficiency trumps all other considerations. Indeed, the efficiency equity trade-off is taught as part of most basic micro courses. The beginning of public economics is social welfare functions, which is an attempt to translate collective values into principles that can then be applied to policy problems by amoral economists.

The discussion of alternate measures of progress in Flourish is punctuated by a statement about the need to decide various value judgements in order to proceed. This is not amoral. My attitude is that if you need to decide such questions before you can go forward then you will never proceed. These include things like ‘how much inequality should we care about?’ That is a normative question; not a scientific one. It cannot be answered objectively, only discursively (or violently, but we are thankfully moving away from that). The best you can do is use democracy (i.e. representative surveys and government) to provide you with a moving answer. ‘Progress’ could then be measured in terms of moving towards average social values. But here you run into problems about protecting minorities, especially intellectual minorities whose arguments might be compelling in a different time and place. Moving towards a Sharia State would be anathema to most modern liberals, for example. How do you handle this plurality? 

This goes to issues of social engineering. Seligman is very minimalist in terms of what he advocates for in terms of an ‘end goal’. He says 51% of people flourishing by 2050. If we measure this subjectively then I can get behind it. If we are talking about a flourishing society though, then it makes me crazy nervous. The last time we tried to use social engineering to get to a flourishing society we ended up with fascism and communism. Those opposed to the normative principles of these paradigms were forcefully converted using violence.

One of the few good ideas of conservatism is the notion that our current system has evolved to its present level of complexity and changing it will undoubtedly have unforeseen consequences, so we need to tread carefully. The other side of this coin is that things that we imagine to be great might turn out to be terrible in practice—we won’t know until we’re in it. We might, for example, think that a libertarian society would be paradise, but then when we’re in it we realise that we hadn’t countenanced the abuse of market power that will be available to individual doctors in small communities in such societies. I worry deeply about calls for social engineering to abstract end states because such engineering inevitably involves making normative assumptions, which is only okay if you have a consensus (as opposed to a mere majority). Historically we have been able to get close to consensuses on problems but not on end states or even solutions. For example, libertarians and progressives both agreed prior to Obama-care that the cost of health care was too high, but they couldn’t agree on a solution. Moreover, their solutions were always informed by their worldview (i.e. libertarianism vs. socialism), which impeded their ability to see a politically practicable hybrid solution like the affordable care act that could be extended or rolled back as experience dictated (unlike whole social systems).

It is for this reason that I remain deeply sceptical about calls for social engineering to ‘better’ (according to whom?) societies. We have plenty of problems that we agree need fixing. Let’s just proceed on them. Happiness and well-being science can play a huge roll in public policy without needing to be embroiled in arguments about what society is better. Indeed, the association of happiness research with left-wing agendas has already resulted in scepticism towards the field. Discursive practices can channel individual values into collective values and then democratic mechanisms can channel those into state actions that abide by the harm principle. We should be wary of trying to go above democratic mechanisms by using scientific techniques to investigate normative questions. Normative answers are inherently subjective. The best you can do scientifically with them is make them more articulate and find an average or median position. But voting does that for you anyway.  

So in closing, I think Seligman’s work is a useful building block as far as theories of individual flourishing go, but dangerously naïve when it comes to theories of flourishing society. I also think it is a bit shallow when it comes to the things that he himself thinks are most critical, namely evaluative satisfactions rather than emotional states. I will endeavour to extend these areas.

If you’re interested in any of my thoughts please let me know. Sorry for the long email. <redacted> seem really busy so I sometimes think a long essay is the easiest way to express long ideas rather than meetings. I’m off to India to do a development economics fellowship on Friday. I will be back in mid-June.