Fifteen Eighty Four

The headline shouts, “Genetics is a big reason divorce runs in families.” It is common nowadays to hear that some surprising aspect of human behavior is “heritable.” Often it is said about intelligence, but one also hears it about personality, mental illness, sexual orientation, or even marital status. What does this mean? The word sounds as though it means that behavioral characteristics are passed on from our parents and programmed into our genes somehow, but that isn’t right.  Understanding the scientific basis of the concept of heritability will allow you to see that while it is a scientifically meaningful idea, it doesn’t portend either the most miraculous (designer babies!) or horrific (Gattaca!) implications that are often ascribed to it.

History

Evolution and genetics were discovered long before anyone knew anything about DNA.  Darwin published The Origin of Species in 1859; the double helix structure of DNA wasn’t worked out until 1953.  So over nearly a century, scientists observed the effects of evolution and genetics on the visible characteristics of organisms, with very little understanding of the biological processes that underlie them.

In 1918, the mathematical biologist RA Fisher showed that for a genetically influenced trait like height, there is a relationship between how similar genetically related groups of organisms (I’ll talk about humans here) are to each other, compared to the general population.  Identical twins (who are clones), for example, are almost exactly the same height, compared to groups of unrelated people who vary a great deal. Siblings vary more than identical twins, but still less than the general population. Even second cousins are more likely to be similar in height than you and I.  Fisher showed how to quantify this relationship into a statistic that varied from zero (for traits where genetic relatedness made no difference) to one (for traits where variability was a perfect function of genetic relatedness). Fisher worked out the math but didn’t name the statistic. Twenty years later, an animal geneticist named Jay Lush named this quantity “heritability,” and showed that it was an important parameter for understanding selective breeding of plants and animals on the farm. All of this is the basis of what is called “quantitative genetics” today.

In humans as opposed to farm animals, heritability coefficients have been estimated in two different ways. The first approach, as suggested above, was to use twins. Identical twins are clones; fraternal twins are siblings who share 50% of the DNA they inherit from their parents. To the extent some characteristic of identical twins, like height or personality, is more similar in identical as compared to fraternal twins, it is an indication of heritability.

Twin studies were developed before the discovery of DNA. Nowadays, scientists can collect DNA from hundreds of thousands of people and use the same logic as twin studies to estimate heritability using unrelated people. We are all related, after all, some of us more closely than others. If pairs of people who happen to share more DNA are also more similar in their behavioral characteristics, the trait is heritable.

What Happened

When behavioral scientists first started conducting twin studies, they targeted the obvious human behavioral differences: intelligence, personality, and mental illness. Identical twins were more similar for all of them, leading to the conclusion that they were all heritable. In the late twentieth century, the world of human behavior seemed to be turning out to be far more genetic than anyone had imagined, promising either a great future of genetic understanding or a dystopian nightmare of genetic control.

But then a funny thing happened. Having run out of obvious behavioral characteristics to study, scientists turned to not so obvious ones, like divorce and how much television people watch, and they turned out to be just as heritable as intelligence and mental illness.  In 2000, I named this outcome “The First Law of Behavior Genetics:” Everything is heritable, and that means literally everything. If a behavior can be reliably measured in human beings, differences in the trait will be heritable.

The same has turned out to be true for heritability estimated from DNA in unrelated people. This turned out to be harder than was originally imagined; hundreds of thousands of people are required to do it well. The heritabilities estimated this way are generally lower than those estimated with twins, sometimes much lower, but they are still greater than zero.

What Heritability Means, and What it Doesn’t

Since RA Fisher, scientists have wanted heritability coefficients to be a measure of “how genetic” a trait is. They don’t.  The classic example is having two arms. Heritability is not about the question, “Why do people have two arms?” the answer to which certainly involves genes. Heritability answers the question, “Why do some people have fewer arms than others?” The answer to that question involves mostly congenital disease and accidents. As a result, the heritability of having two arms is zero. Below, I list some do’s and don’ts on the meaning of heritability. One list will be longer than the other.

Does mean: Genes are a reason people differ.  This is the most general meaning. If we were all more alike genetically, ultimately if we were all clones, we would be less varied in our characteristics, including pretty much everything.

Doesn’t mean: There are “genes for” behavioral differences.  The fact that how much coffee you drink is heritable doesn’t imply that there are “genes for” coffee drinking. The same technology that allows us to use DNA to estimate heritability also allows us to search for specific genetic causes of human behavioral differences. Once you take into account the genes with large negative effects that have been known for a long time, like Huntington’s Disease, there is nothing.

Does mean: you have to be careful when interpreting correlations among genetically related family members. An old-fashioned kind of social science tries to understand complex behavior by looking at associations between the traits of family members, especially parents and children.  This can be problematic. Suppose depressed parents have children who show signs of depression themselves. Can you conclude that depression in parents causes depression in children? Would you observe the same thing if you studied adoptive families, with no genetic relationship between parents and their kids? Genetic knowledge allows us to do social science a little bit better.

Doesn’t mean: Traits can’t be changed because they are “genetic.”  Consider eyesight. Visual acuity is strongly genetic, but visual difficulties can be mostly remediated with an environmental intervention: eyeglasses. 

Doesn’t mean: Differences among ethnic and cultural groups are probably genetic in origin.  Groups of people around the world differ in their behavior; to one extent or another (though to a lesser extent than you might think) they also differ genetically. But in general, there is no evidence that genetic differences among groups cause behavioral differences in any kind of systematic way. There is not a single good example of a psychological difference among ethnic groups that is caused by genetic differences.

Doesn’t mean: It might be a good idea to control the breeding of humans, as we have done so successfully with domesticated animals. This idea, called eugenics, unfortunately tried during the first half of the last century, and it stands with the Holocaust as one of the great moral disasters of modern civilization. There is no way for eugenics to work without (a) someone being in charge of what traits should be selected, and (b) exercising a fascistic level of control over people’s ability to reproduce.

Doesn’t mean: Soon we will be designing babies with desirable behavioral characteristics.  This is a modern version of the previous one.  It is routine to test embryos, either in the womb or in vitro, for major genetic conditions like Down’s Syndrome. Some people imagine using modern DNA methods to test for IQ or athletic ability. This idea shares most of the moral flaws of classical eugenics, with the saving grace that it wouldn’t work. Genetic prediction for behavioral traits isn’t close to strong enough.

The Bottom Line

People will always differ in their behavior. The fact that people differ genetically is one of the reasons for the beauty of human diversity. But there are not systematic causal relations between specific genes and specific behavioral outcomes. That is a disappointment for enthusiastic genetic scientists, but all in all it is probably a good thing.  Few would want a world where genes alone dictate behavior..

Title: Understanding the Nature‒Nurture Debate

Author: Eric Turkheimer

ISBN: 9781108958165