In his new Dana Press book, The Temperamental Thread, developmental psychologist Jerome Kagan draws on decades of research to describe the nature of temperament—the in-born traits that underlie our responses to experience. Along the way he answers such questions as, How does the temperament we are born with affect the rest of our lives? Are we set at birth on an irrevocable path of optimism or pessimism? Must a fussy baby always become an anxious adult? In this interview with Dana Press journalist Aalok Mehta, he explains how temperament affects personality, whether it can predict your future, and how it might influence a doctor deciding which medical treatment may work best for you.
AM: What is temperament?JK: It's a general term meaning a set of biological conditions that bias infants to react to the world with special feelings and behaviors. A temperament is analogous to the psychological profile of a dog breed. Everyone knows a beagle is different from a pit bull and a pit bull is different from a collie because the two breeds have different neurobiologies.
Although most temperaments are due to inherited neurobiological profiles, some are not inherited. Research shows that the season of conception, fall or spring, can affect the young embryo because short or long hours of daylight affect the pregnant mother’s secretion of melatonin, which influences the embryo. Severe trauma or infection in the pregnant mother can also cause changes. I think most temperaments are due to inherited differences in neurochemistry, in the concentrations of the more than 150 molecules in the human brain, either the density of the receptors for those molecules, or the locations of those receptors.
Because there are so many ways to produce a temperamental bias, there will be many thousands of temperaments. Some will be common, some rare. When the more complete story is in, decades from now, we’ll find many temperamental biases.
How did you become interested in studying temperament?
Three events were critical. My first major research assignment was at the Fels Research Institute in Ohio, which had a large sample of middle-class infants born in the 1920s that the staff had followed from their birth through adolescence. I was hired in 1957 to evaluate these subjects when they were adults and discover if we could predict what they were like in their twenties as a function of their early behaviors. Howard Moss and I published the results in 1962 in the book Birth to Maturity.
The most important result was that a small group of children in the first three years of life were very timid, shy, and cautious. It seemed obvious to us that these behaviors were due to their temperaments and not caused by how they were handled. When they were adults, these shy, timid children were different from others. They were more cautious and introverted than the rest of the sample. I thought about that finding, and should have pursued it then, but didn’t.
In the late 1970s, some colleagues and I were studying the effects of day care on infants because of the possibility that the government might fund day care centers for infants so that mothers could go to work (it never did pass the legislation). Many psychologists were concerned that placing young infants in day care was not a good idea. We ran a day-care center in Boston for three years. We started with three-month-old infants, and for political reasons half of the infants were Caucasian and half were Chinese-American whose parents lived in Boston's Chinatown. We followed them into their third year.
The main result was that the day care experience had a minimal effect on development, compared with children who didn't attend our day care center and were raised only at home. But the Caucasian and Chinese children were different from three months on, and these differences seemed to be temperamental in origin.
Second, by the 1970s, the work of [Stella] Chess and [Alexander] Thomas was becoming popular. Third, neurobiologists studying strain differences in mice, rats, and monkeys reported that the small genetic differences among the strains were associated with behaviors; that is with temperament. All these facts came together, and in the 1980s I began to study temperament in earnest.
What do we get out of studying temperament?
Two implications are obvious. One is relevant to contemporary psychiatry. The Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) is the document psychiatrists use to classify patients into various disorders based primarily on the report of their symptoms. As you may have read in the press, the next edition, which has been released in preliminary form, is going to do the same thing. Here's the problem: Every one of those categories—whether social anxiety disorder, depression, substance abuse, conduct disorder, or ADHD—is heterogenous in its origin or etiology. Some patients in a category acquired their symptoms by experience alone; others did because of their temperamental biases and life history.
If you don't assess some signs of a temperamental bias, you are pooling together patients with different biologies and life histories and treating all of them in the same way. A social phobic who is shy and doesn't like crowds due to a temperamental bias should probably get a different treatment than a social phobic who is shy because he or she feels homely, obese, or does not have enough education.
Gerard Bruder, a scientist at Columbia University tested a group of patients with the same diagnosis of depression. He measured their EEG (the electrical activity along the scalp produced by the firing of neurons within the brain), and found that those who improved on antidepressant drug therapy were likely to have had left hemisphere activation in the frontal lobe. Those for whom the drugs were less helpful were right hemisphere active. The latter group of depressives is probably temperamentally different from the former. Psychiatrists should consider that fact and not waste the patient's time or the HMO’s money by giving all depressives the same drugs or psychotherapy. Psychiatrists will make substantial progress when they begin to look into the origins of each symptom category instead of redefining the disorders based only on the presenting symptoms. Headaches and stomach cramps have distinct causes and call for distinctly different therapies.
A second practical implication is relevant for mothers, especially American and European mothers who often experience guilt if they have a very anxious or difficult child. Many assume that they did something wrong; perhaps they went to work too early or took a long holiday at the wrong time. But the mothers of some of these children are not to blame for the child’s behavior. They gave birth to a child with a temperament that made it easy for them to become anxious or difficult, and the mother should not feel guilty about being unable to socialize such a child. Many mothers who read about our work send me e-mails saying that this fact has been helpful to them because they now realize that they were blaming themselves unnecessarily.
Why did you decide that now is the right time to write a book about temperament?
Because many parents had called me to say they had learned a lot from our work, I thought it would be useful for the general public to have a book on temperament that was accessible, didn’t have a lot of jargon and statistics, and explained what scientists have learned.
There's a continuing debate about nature and nurture, and which has more influence, genetics or upbringing. Does temperament tell us anything about this?
All experts agree that we must stop asking that question. The two conditions come together like the warp and woof in fabric. It's snowing today in Boston. To ask which is more important, temperature or humidity, is like asking whether nature or nurture is more important. One can't answer that question. A temperature less than 32 degrees and high humidity cause snow. The two come together. Every recent textbook declares that we should stop asking whether biology or experience is more important. The two conditions come together to produce a personality profile.
Some temperament research delves into controversial subjects, such as the contribution of gender or ethnic background to temperament. What does temperament research tell us about these subjects, and how do we study these subjects in a way that minimizes controversy and is fair to the groups involved?
Scienists want to be responsible. Temperamental differences between the sexes or among ethnic groups are touchy issues. But we know from the history of the sciences that suppressing knowledge in the service of political correctness turns out to be an error. The book says that although males, females, and reproductively isolated ethnic groups differ in temperamental biases, these differences have no implications for political rights.
A group of 1,000 Caucasian-American males would differ biologically and psychologically from 1,000 white females, and 1,000 Chinese-Americans would differ from 1,000 African-Americans. The book suggests that each group has temperaments that have advantages and disadvantages in any given society. Consider the sex differences. Most males have a higher threshold for becoming fearful to dangers and that seems to be an advantage in our society. On the other hand, males also have a temperamental bias toward taking risks and an absence of high levels of empathy, and, therefore, are more likely to commit murder. Those traits are disadvantageous. When we add up the pluses and minuses, the final account turns out to be balanced. Each gender or ethnic group has some temperamental biases that are advantageous in their society and some that are disadvantageous, and that's the way it is.
How would you characterize our understanding of temperament so far? Are we still in the early stages?
Serious scientific research on human temperaments is only 53 years old, for it began with Chess and Thomas’s writings in 1957. The research currently funded by the National Institutes of Health and the National Science Foundation supports a small proportion of proposals on human temperaments.
When our understanding is more complete, future historians will realize how primitive our present understanding is. For example, consider the two temperaments that have been studied extensively: the very cautious, quiet child and the outgoing, ebullient child. These are the two temperaments my laboratory is studying. We have learned a great deal but we still do not know the complete neurobiology of those two types. That fact reveals how immature the field is today.
What are the limits of temperament research? What is this line of research not going to tell us?
Knowledge of a child’s temperaments does not predict their adult personality profiles very well. Temperaments act primarily by preventing a person from developing a certain personality, rather than determining a specific type. For example, we call one of our infant temperaments high-reactivity. These infants are easily aroused by any event that is new or unexpected. When we assessed a large group of high reactives at 15 years of age and asked what proportion were extremely shy, anxious, timid, or cautious we learned that about 25 percent had such a profile. One might conclude that their infant temperament was not very predictive. That conclusion is reasonable because their environments have shaped their temperaments in distinct ways. But now let us ask: what proportion of high reactives has been, over the past 12 or 13 years, consistently ebullient, risk-prone, and socially outgoing? The answer is 5 percent. We do very well when we predict what they will not be for we predict correctly for 95 percent of these adolescents.
Almost everyone can learn to play tennis. But how many can play like Roger Federer? Not many, because their biology prevents them from attaining this level of talent. If you knew my genome the day I was born, you might be able to predict with great confidence that I would not play as well as Roger Federer. But you would be wrong if you said that this boy will never learn to play tennis. Temperaments constrain, rather than predict, the future. Their primary role is to limit what might develop.