You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

The Science of the Science of Happiness

Fox Photos/Getty Images

Consider Charles Boatwright. His name is one of the charming pseudonyms from the vignettes in George Vaillant’s new book, Triumphs of Experience: The Men of the Harvard Grant Study. Boatwright grew up with a manic-depressive father and was married to a miserable woman for 30-odd years, but he almost always called himself happy. He bounced from job to job—from boat work to cattle-inseminating (the artificial kind)—but considered himself a success. Vaillant, who studied Boatwright and 267 of his college classmates over a 38-year period, from 1966 to 2004, at first saw him as a failure in chronic denial.

On Vaillant’s retelling, however, it is Vaillant himself who had been too selfish to respect Boatwright’s selflessness, too defensive to see Boatwright’s problems as consistent with his optimism, too ambitious to see his dilettantish do-goodery as truly fulfilling. Boatwright’s life was poor in the terms of Vaillant’s 1970s standards. Now he finds that judgment absurd, not just because the narrowness of his perspective became clear to him but because a lifetime of studying people led him to believe that people grow and substantially change after age 20, 30, and even 55.

This shift in perspective sets the stage for Valliant’s book, in which he aims to trace the contours of his subjects’ growth in late adulthood and its connection to their earlier lives. The study offers broadly applicable evidence about how everything from early maturity to grandparents’ longevity is likely to affect flourishing throughout life. Like a good doctor, Vaillant has written a book whose conclusions generalize most clearly when they concern physical and mental health. His arguments about class-influenced factors such as IQ are limited by his sample, which is confined to the upper class, and his arguments about abstract human qualities such as wisdom and flourishing are limited by his impatience for abstraction. But there is much to like.

It is hard to overstate the wealth of the data provided in Triumphs of Experience or the ambition of the project, composed of survey responses, health records, and interviews. This archive of human life is poised to answer questions shorter studies can barely hint at. Take character—does it remain static after 30, as William James and the young George Vaillant thought it did? You would need a lifetime to answer, and a lifetime not just of possibly self-deluded self-reports, but a lifetime of interviews with the people surrounding the subject.

In turn, Vaillant offers striking conclusions about a range of factors affecting human flourishing. Your relatives’ mental health affects yours, regardless of whether you live with them. Smoking takes years off your life, if not as many as alcoholism. The warmth of the environment of your childhood and the maturity of your early-life emotions more closely predict flourishing in old age than IQ, visceral masculinity, parental class, or income. A high IQ predicts no more success than a semi-high one. Your ancestors’ life-expectancy does not affect your flourishing—except for your maternal grandfather’s, which matters a lot. Warm childhoods do not predict good first marriages; over half of divorces follow alcoholism; liberals flourish no more than conservatives but have sex till much later in life. About politics, both sides remain rigid for life.

On the other hand, these subjects are white men from Harvard when it was even more Harvard than it is now, so some of Vaillant’s results might not apply across different races, classes, genders, and historical periods. A parents’ class and education might not have affected the subject once he joined Harvard as a freshman, but it probably helped to get him there. Your IQ of 150 might actually earn you significantly more than your IQ of 110 if you never had the connections and prestige of the Crimson beanie. Even Vaillant’s metrics for the stages in human development—“Identity” is measured by financial independence and having left home; “Career Consolidation” is measured by “commitment, compensation, contentment, and competence”—are so much easier for the wealthy to achieve.

Vaillant probably knows all of this. He compares a few of his results with those of a long-term study of men from the inner city, and explains that the Harvard study’s founder, Arlen Bock, wanted to measure optimal health rather than the average, and so he chose a population likely to flourish. Vaillant defends this with a medical analogy:

If we want to learn what people eat, we have to study many different populations. If we want to learn about gastrointestinal physiology, however, we try to keep variables like cultural habits and preferences uniform.

I take the analogy to mean that you want to keep variables fixed if the variables cause the situation you want to observe but do not help to explain the questions your observations would answer. But in Vaillant’s case, class-related factors are a major part of any explanation of well-being. The analogy seems suited more to medical science than to social science.

Vaillant is at his most ambitious on the topic of the human psyche. On human development, one of his foci, he was inspired by developmental psychologist Erik Erikson’s psychoanalytic models and by Shakespeare’s “seven ages of man” from As You Like It, but thought that they “didn’t have much by way of real data to go on,” unlike him. Where genius was, there science shall be. But the risk of this kind of research, conducted in the terms whereby people make sense of themselves, is that its broad lens will not do justice to the nuance of those personal terms. It’s like photographing a stomachache.

In a paper Vaillant wrote a few years ago, he argued that spirituality leads to well-being without aid of religion. (He wrote a book on the same theme.) Being religious was measured by how often you went to “religious or spiritual services” or engaged in “church/temple activities (e.g., dinners, volunteer work, and church related organizations). Spirituality was measured by how often people thought that they felt “a sense of deep appreciation” and “a profound sense of caring for others.” Well, come on. A Secular Sam could volunteer at his local church for its philanthropy but not for its dogma, and even apostates could go to services to see their communities. A surgeon could work too much to go to services but have her whole ideology shaped by the church she was raised with—ideology that shapes whether she attributes her “sense of deep appreciation” to YHWH or another spiritual being or to a positive dopamine imbalance.

The issues raised by Vaillant’s treatment of “religion” apply to a number of other terms in Triumphs of Experience. Consider wisdom. The wisest of all, according to Vaillant: Charles Boatwright. Is wisdom really the relentless optimism of a Boatwright, or the kind of discernment that helps you to feel good only about things that will actually benefit you and the people around you? Does the wisest man marry an unhappy woman for 30 years? Well-being is more of a puzzle. Vaillant prefers to call it “flourishing,” after psychologist Martin Seligman, to emphasize a list of achievements rather than just reports of how happy you think you are (the kind of report favored by psychologists and behavioral economists under Daniel Kahneman’s influence). A problem with this sort of list is that for nearly everything it includes you can imagine someone living well without it or not living well with it. I doubt that Vaillant even meant his list as a literal measure of well-being—“judgments about the ‘good life,’” he writes, “can be very annoying”—and his “flourishing” is most useful as a kind of conceptual six-pack ring for things people generally want in their lives. We learn that a Charles Boatwright type will likely have “a good marriage between ages 60 and 85” and be “close to [his] kids between ages 60 and 75,” even if we don’t really learn that he flourishes.

I have less confidence in what he writes about character. “I now argue with friends not whether personalities change in adulthood,” writes Vaillant, “but about how best to measure the changes.” His main reason is that his data failed to square with received ideas of character development, which state that people mature as their capacities deepen for love and for work. “Periods of success at work and at love came and went unsystematically and according to circumstance, and there were no universally discernible processes of ‘deepening’ with age and experience.” That likely tells us a great deal about outcomes but not much about the characters that led to them. It is hard to see how character could exist in anything but subjunctives, how someone would be if such and such were to happen, so the same character could seem very different “according to circumstance.” Boatwright had the same optimistic selflessness through his whole adult life, Vaillant the same respect for new data even when the data changed his mind. One can say the same sort of thing for every man Vaillant writes about. If their changes seem “unsystematic,” perhaps this says less about the vagaries of character than about the vagueness with which we often understand it, how it is easy to feel but ineffable in metrics.

William James could well be wrong to say that “in most of us, by the age of thirty, the character has set like plaster, and will never soften again.” He had little in the way of data (taken as evidence you can observe and replicate in the relevant ways). But he did not live under a rock, and neither did Tolstoy or Shakespeare. Sensitivity, nuance, and genius are reasons to believe things about humanity inferred from experience, even if the experiences are hard to observe and to replicate. Still, the Grant Study is an archive, and an ingenious researcher could dig through it and tell against James for new reasons—a possibility that Vaillant, with his respect for research no matter whether it confirms his analyses, would wisely welcome.