Triumphs of Experience

George E. Vaillant

• Publisher: Belknap Press
• Year: 2012
• ISBN:
  9780674071810
• Topic: Psychology
• Subtopic: Developmental Psychology
• Index: Psychology

At a time when many people around the world are living into their tenth decade, the longest longitudinal study of human development ever undertaken offers some welcome news for the new old age: our lives continue to evolve in our later years, and often become more fulfilling than before.Begun in 1938, the Grant Study of Adult Development charted the physical and emotional health of over 200 men, starting with their undergraduate days. The now-classic Adaptation to Life reported on the men's lives up to age 55 and helped us understand adult maturation. Now George Vaillant follows the men into their nineties, documenting for the first time what it is like to flourish far beyond conventional retirement.Reporting on all aspects of male life, including relationships, politics and religion, coping strategies, and alcohol use (its abuse being by far the greatest disruptor of health and happiness for the study's subjects), Triumphs of Experience shares a number of surprising findings. For example, the people who do well in old age did not necessarily do so well in midlife, and vice versa. While the study confirms that recovery from a lousy childhood is possible, memories of a happy childhood are a lifelong source of strength. Marriages bring much more contentment after age 70, and physical aging after 80 is determined less by heredity than by habits formed prior to age 50. The credit for growing old with grace and vitality, it seems, goes more to ourselves than to our stellar genetic makeup.

1

MATURATION MAKES LIARS OF US ALL

No man ever steps in the same river twice; for it is not the same river, and he is not the same man.

—HERACLITUS

This book is about how a group of men adapted themselves to life and adapted their lives to themselves. It is also about the Grant Study, now seventy-five years old, out of which this story came. In it I will offer tentative answers to some important questions: about adult development in general, about the people who engaged us in this exploratory venture, about the study itself, and, perhaps above all, about the pleasures and perils of very long scientific projects.

Originally the Grant Study was called the Harvard Longitudinal Study. A year later it became the Harvard Grant Study of Social Adjustments. In 1947 it received its now-official name, the Harvard Study of Adult Development. But to its members and its researchers, and in early books, it has always been the Grant Study.

It began in 1938 as an attempt to transcend medicine’s usual preoccupation with pathology and learn something instead about optimum health and potential and the conditions that promote them. The first subjects were 64 carefully chosen sophomores from the all-male Harvard College classes of 1939, 1940, and 1941, who took part in an intensive battery of tests and interviews. That first group was joined by sophomores from the next three Harvard classes, resulting in a final cohort (as the panel of subjects in a study of this kind is called) of 268 men. The original intention was to follow these healthy and privileged men for fifteen or twenty years, supplementing the intake data from time to time with updates. Thus an abundance of information would accumulate about the men and the lives they constructed for themselves—information that could be analyzed at will over time and across different perspectives. (Interested readers will find much, much more on the history and structure of the Study in Chapter 3.)

That plan was realized, and more. Almost seventy-five years later, the Grant Study still, remarkably, goes on. We’re asking different questions now than the founders asked when the Study began, and our investigative tools are different. Of course the participants are no longer the college sophomores they once were; those who are still with us are very old men indeed. Time has called many of the beliefs of those days into question, and some much more recent ones, too. How long our current conclusions will hold up we cannot know.

But whatever the uncertainties, asking questions and trying to answer them is always a fruitful process. We actually have learned some of what they wanted to know back in 1938: who would make it to ninety, physically capable and mentally alert; who would build lasting and happy marriages; who would achieve conventional (or unconventional) career success. Best of all, we have seventy-five years’ worth of data that we can refer back to (over and over again, if need be) as we try to understand why these things turned out as they did.

We can use those data to try to answer other questions, too. There are old ones still open from the early Study days—about the relative importance of nature and nurture, for instance, or how mental and physical illness can be predicted, or the relationship between personality and health. There are new ones that could never have been dreamed of in 1938, like what the emotions of intimacy look like in the brain. And some old questions have morphed into new ones as we learn to formulate ever more cogently just what we’re seeking to accomplish in even making inquiries like these. This last in particular is an abiding concern of good science.

Thus this story of the Grant Study: how it came to be, how it developed, and how those of us who participated in it developed too. It’s the story of what we’ve learned from it and what we haven’t (yet). And it’s a story about Time—studying it and living in it.

LONGITUDINAL STUDIES: THEIR STRUCTURE AND PURPOSE

The Grant Study is a longitudinal prospective study. Let me say a few words about that as we set out. In a longitudinal study, a group of participants (a cohort) is observed over time, and data about points of interest (variables) are collected at repeated intervals. In contrast, observations in a cross-sectional study are made only once. A longitudinal study can be prospective or retrospective. Retrospective studies look back over time, seeking to identify the variables that might have contributed to outcomes that are already known. Prospective studies follow a cohort in real time, tracking target variables as the subjects’ lives proceed, and identifying outcomes only as they occur. Accordingly, the Grant Study collected all kinds of potentially (but not necessarily) relevant information about its cohort members over the many years of their lives, looking to learn what it could about health and success. And it has correlated this information periodically with the levels of health and success that each man actually achieved. You’ll see many examples of this process as we go on.

Less abstractly, longitudinal studies let us contrast eighty-year-olds with themselves at twenty-five or sixty. Biographies and autobiographies, and the debriefings that elders offer their fascinated grandchildren, do the same thing. But these are all retrospective narratives, inevitably influenced by forgettings, embellishments, and biases. Time is a great deceiver.

Prospective studies are like the baby books compiled by doting parents, or like time-lapse photographs; they document change as it happens, allowing us to visualize the passage of time free from the distortions of memory. While butterflies recalling their youths tend to remember themselves as young butterflies, prospective studies capture the reality (hard to believe, and often avoided!) that butterflies and caterpillars are the same people.

But baby books and even year-long time-lapse videos are a dime a dozen. Prospective studies of entire lives are very rare indeed. None are known to exist before 1995; I’ll have more to say on that shortly. In fact, observational data on adult development have been so sparse that when Gail Sheehy and Daniel Levinson produced their groundbreaking books on the subject in the 1970s, the cross-sectional studies they relied on led them to some very erroneous conclusions (such as the inevitability of the so-called midlife crisis).¹

Another important reason for prospective studies is that they establish contexts for the outcomes we’re trying to understand. The most experienced handicapper in the world can’t predict for sure which of the well-bred, handsome horses in the Churchill Downs paddocks each May will win the Derby. Certainty comes only after the race is won, and even then we only know who. The hows and whys remain a mystery. It’s easy to rate a beautiful woman according to how many people think her so, or how well she conforms to some established ideal, or how long she keeps her looks. But ratings like that can’t encompass the difference between being beautiful at one’s high school prom and being beautiful at one’s great-granddaughter’s wedding, nor between the beauty at eighteen that is the luck of the draw, and the beauty at eighty that is the result of a life generously lived. The Grant Study was hoping to learn something about nuances like these in its exploration of success and “optimum” health.

Part of my intent in this book is to show why these nuances matter, and why, therefore, we need longitudinal studies if we want to learn about human lives. I will be focusing primarily on the Grant Study and its cohort, whom I will call the College men. But I will occasionally make reference to two other important cohorts. One is the Inner City cohort of the Glueck Study of Juvenile Delinquency.² The Glueck Study is a second Harvard-based prospective longitudinal lifetime study, begun independently of the Grant Study in 1940; its participants were a group of youths from disadvantaged urban Boston neighborhoods. Since 1970, one cohort of this group has been administered together with the original Grant Study, as part of the Harvard Study of Adult Development. (I will use that term when referring to both studies, but will continue to call the Grant Study per se by its original name.) I will also refer from time to time to the Terman women, a cohort from Stanford University’s (1920–2011) Terman Study of gifted children, to whose data and participants I have had partial access.³ It was the coming of age of the Terman cohort (male and female) in 1995 that made prospective lifetime data available for the first time. The Grant Study’s access to the Inner City men and the Terman women permitted us to contrast our privileged, intelligent, and all-male College sample with another all-male sample of very different socioeconomic and intellectual profile, and with a group of even more intelligent, though not particularly privileged, women. When these contrasts are enlightening, I will note them. The Glueck and Terman studies (and some of the correlations among them) are described in their own contexts in Appendix B, and are also considered in my previous books Aging Well and The Natural History of Alcoholism Revisited.⁴

I acknowledge readily that the Grant Study is not the only great prospective longitudinal lifetime study. There are others, three of which are better known than ours. Each has its own strengths and weaknesses. The Berkeley and Oakland Growth Studies (1930–2009) from the University of California at Berkeley include both sexes and began when the participants were younger; they provide more sophisticated childhood psychosocial data but little medical information.⁵ These cohorts have been very intensively studied, but they are smaller and have suffered greater attrition than ours. The Framingham Study (1946 to the present) and the Nurses Study at the Harvard School of Public Health (1976 to the present) boast better physical health coverage, but they lack psychosocial data.⁶ These are wonderful world-class studies, invaluable in their own ways, and more frequently cited than the Grant Study. But even in this august company the Grant Study is unmistakable and unique. It has been funded continuously for more than seventy years; it has had the highest number of contacts with its members and the lowest attrition rate of all; it has interviewed three generations of relatives; and, most crucial for adult development, it has consistently obtained objective information on both psychosocial and biomedical health.⁷ Finally, perhaps alone among the world’s significant longitudinal studies, the Grant Study has published, with the men’s permission of course, lifespan histories as well as statistical data.

Other studies exist now in Great Britain, Germany, and the United States that are larger and more representative than these older ones, and will join them in length of follow-up in another decade or two. The Wisconsin Longitudinal Study, for example, began in 1957 and included about a third of all of Wisconsin’s high school graduates of that year; it has endured for over half a century so far.⁸ Eighty-eight percent of its surviving members are still active in the study at age sixty-five. (By way of comparison, 96 percent of the surviving Grant Study members are still active at age ninety!) The Wisconsin Study is more demographically representative than the other studies, and its economic and sociological data are richer and better analyzed. It has a weakness too, however; it lacks face-to-face medical examinations or interviews. We can anticipate a great wealth of prospective life data as these younger studies come into their own. But they will supplement, not supplant, the riches already offered by the Grant Study and its contemporaries.

When I came to the Grant Study in 1966, I was a very young man of thirty-two, and had not yet achieved the pragmatic relativism of Heraclitus. I had been studying stable remissions in schizophrenia and heroin addiction—recovery vs. no recovery, white vs. black. I spent my first ten years at the Grant Study identifying thirty unambiguously good outcomes and thirty unambiguously bad ones out of a randomly selected sample of one hundred middle-aged men from the classes of 1942–44. In 1977 I published a book, Adaptation to Life, demonstrating this accomplishment.⁹ It made quite a splash. I was forty-three. What did I know?

Today I am seventy-eight. The men of the Grant Study are in their nineties. They’re not the same as they were when they joined the Study, and neither am I. I have learned to appreciate how few blacks and whites there are in human lives, and how we and our rivers change from moment to moment. The world we live in is different; science is different; even the technology of documenting difference is different. As the Grant Study becomes one of the longest studies of adult development in the world, it is not only the men of six Harvard classes who are under the microscope. All of us who have worked on the Study—the Study itself, in fact—are now as much observed as observers.

ADULTS, GROWING

The laws of adult development are nowhere near as well known as the laws of the solar system or even the laws of child development, which were only discovered in the last century. It wasn’t all that long ago that Jean Piaget and Benjamin Spock were still kids, and the phases of childhood so stunningly elucidated by them still regarded as unpredictable. Now, however, we watch children develop the way our ancestors watched the orderly waxings and wanings of the moon. We worry; we pray; we weep; we heave sighs of relief. But we are no longer particularly surprised. We know what to expect. Our libraries are full of studies of human development—up to the age of twenty-one.

What happens after that remains in many ways a mystery. Even the notion that adults do develop, that they don’t reach some sort of permanent steady state at voting age, has been slow to gain traction. Some of the reasons for this are obvious. It is much easier to achieve a long perspective on childhood than on an entire life. And as in physics, once Time gets into the picture you can say good-bye to the old Newtonian verities. There are certainly patterns and rhythms to adult life, and when we circumvent the distorting effects of time upon our own vision, we can sometimes discern them.

But even the most carefully designed prospective study in the world can never free us completely from time’s confounding influence. Lifetime studies have to last many, many years, and over those years everything will be changing—our questions, our techniques, our subjects, ourselves. I’ve been studying adult development since I was thirty, and I know now that many of my past conjectures, apparently accurate at the time, were contingent or just plain wrong. Still, the Grant Study is one of the first vantage points the world has ever had on which to stand and look prospectively at a man’s life from eighteen to ninety. Having it doesn’t save us from surprises, frustrations, and conundrums; quite the contrary. Nevertheless, a continuous view of a lifetime is now possible for human beings. Like that time-lapse film of a flower blooming that Disney made famous in the sixties, it is an awesome gift.

When the Study was undertaken in 1938, what we knew about human development across the whole of life was mostly based on inspiration or intuition. William Shakespeare delineated seven ages of man in As You Like It in 1599; Erik Erikson defined eight stages in Childhood and Society three hundred and fifty years later.¹⁰ But Shakespeare and Erikson didn’t have much by way of real data to go on. Neither did Sheehy and Levinson. Neither did I, in Adaptation to Life. Nobody had access to prospectively studied whole lifetimes. I hope that this book, written in 2012, will begin to correct that lack.

THE ART OF THE POSSIBLE

A few caveats as we proceed. It is well known that the Grant Study includes only white Harvard men; Arlen V. Bock, the physician who founded it, has frequently been criticized for arrogance and chauvinism on that account. It’s less well known that the Grant Study was not an attempt to document average health over time, like the more famous Framingham Study, but to define the best health possible. And we must therefore keep in mind two realities. First, lifetime studies, like politics, are the art of the possible. As Samuel Johnson famously quipped about dogs walking on their hind legs, “It is not done well, but you are surprised to find it done at all.” Second, in such huge undertakings, one must optimize one’s chance of success. Columbia neuroscientist Eric Kandel did not choose a random sample of the world’s population of homo sapiens when he did his Nobel-winning work on the biology of memory; he chose the obscure sea snail called Aplysia. Why? Because Aplysia has unusually large neurons. And it was precisely the gender and privilege of the Grant Study men that made them so useful for a study of human adaptation and development. Men don’t change their names in midlife and disappear to follow-up as women do. Well-to-do men don’t die early of malnutrition, infection, accident, or bad medical care, as happens much too often to poor ones. These men had a high likelihood of long life, a necessity for this sort of study. (A full 30 percent of the Grant Study men have made it to ninety, as opposed to the 3 to 5 percent expected of all white male Americans born around 1920.) Glass ceilings and racial prejudice were unlikely to hold them back from achieving to their fullest potential, or from the careers and lives that they desired. A Harvard diploma wouldn’t hurt either. When things went wrong in the lives of the Grant Study men, they would have a good chance of being able to set them right. Last but not least, they were unusually articulate historians. Bock needed all those advantages. You can’t study the development of delphiniums in Labrador or the Sahara. The Grant Study’s College cohort and Aplysia may not be perfectly representative, but they both afford us windows onto landscapes we have never been able to see before.

One further note on Bock’s choice of a homogeneous population. 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. Societies are forever changing, but biology mostly stays the s...