Evolution depends on characteristics being passed from parents to offspring. Behavior, like other traits, is the result of a combination of influence from the genes and from the environment. But while it is relatively easy to understand how the environment changes behavior, the connection between genes and behavior is a little harder to comprehend.
Of course, one doesn’t have to know anything about genes to think about evolution. Darwin was able to develop his theory of how characteristics—including behavior—evolved without knowing the mechanism of how parents pass traits to their offspring. It took the monk Gregor Mendel, working in his garden in the Augustinian St Thomas’s Abbey in Brno, now part of the Czech Republic, to find the key to inheritance. During the mid-nineteenth century, Mendel used careful experiments to show how pea plants could pass on smooth or wrinkled seeds to their offspring. Working out the proportions of each type in the seedlings of a particular set of parents, he was able to show that each parent plant had an element that it contributed to the seed, and that those elements stayed as distinct entities in the seedlings. When one crossed a smooth-seeded parent and a wrinkly-seeded parent, the seedlings were either smooth or wrinkled, but not somewhere in between. This notion contrasts with the common view at the time, which was that the characteristics of the parents were always blended together, as if one were mixing paint, so that a red parent and a white parent (to switch metaphors) would have pink offspring. We now know that traits that follow Mendelian inheritance are due to one or a few genes together.
Determining the nature of genes was a huge breakthrough, and once Mendel’s work was recognized in the early twentieth century, scientists began to study which traits were associated with single genes. The early history of genetics is filled with discoveries about the genetic basis of everything from eye color in fruit flies to the number of kernels in an ear of corn. Not all the traits were inherited in a Mendelian fashion, as I discuss later, but many were.
A few behavioral traits are inherited almost as simply as seed texture in Mendel’s peas. For many decades, beekeepers have been vigilant about a nasty disease called American foulbrood, which invades beehives and eventually kills the larvae inside, decimating the colony. The queen bee lays eggs that hatch into larvae inside the hexagonal wax cells of the hive. The workers will then rear the young bees. A diseased larva can quickly spread its infection through contact with the bees in the colony tending to it. But some colonies show what is called hygienic behavior: adult bees tending the young will recognize a sick larva, saw through the top of the wax cover of the cell, remove the diseased individual, and take it away from the hive, preventing further transmission.
In the 1960s, scientist Walter Rothenbuhler did experiments on the control of foulbrood similar to those performed by Mendel on the peas.1 He mated bees from hygienic colonies with bees from colonies that were not hygienic, and then examined the behavior of the colonies that were produced. According to his results, the hygienic behavior depended on two genes: one that controlled the bee’s propensity to uncap the wax, and another that caused the bees to remove the larva, limiting the spread of the infection. Thus, in the “hybrid” bees (I put hybrid in quotation marks because the bees were all the same species, just with different behaviors), some were completely hygienic—meaning that they successfully rid their colonies of foulbrood—and some were not. The non-hygienic bee colonies were of two types. In some, the bees could remove a larva if the beekeeper uncapped the cell first, while in others, workers would uncap the cell, but then leave the larva rotting away. Each behavior was associated with one gene, and both genes were needed to complete the entire sequence. Since that time, more genes have been discovered that moderate the behavior, and colonies have been found to vary in the extent to which they enact the removal of the diseased young, but it is still a relatively simple correspondence between genes and action.
Rothenbuhler’s work couldn’t identify the genes that were responsible for the behavior, largely because well into the twentieth century, scientists knew little about genes in most animals. The necessary delicate microscope work and tedious experiments made only a few living things possible subjects. During the early 1900s, most genetic research relied on a single animal, Drosophila, commonly called fruit flies. No one would call them charismatic—they are the tiny brown creatures attracted to rotting bananas in your kitchen—but they were of immense importance in the development of genetics in the early to mid-twentieth century, and they continue to be essential animal models for studies ranging from genetics to physiology. Drosophila are easy to rear in the laboratory, they have chromosomes that can be seen more easily under the microscope than those of other organisms, and they can be bred to show different genetic changes, or mutations, in their physical appearance. In the days before DNA sequencing, or even before DNA was known to be the stuff of genes, much progress was made by breeding fruit flies with obviously different traits and analyzing the resulting offspring.
However, although scientists diligently placed fruit flies with different eye colors or wing shapes together in glass bottles (the preferred way to harbor the insects) and counted the types of juveniles produced from these crosses (as they are called), for a variety of reasons, they rarely if ever bothered to watch the fruit flies mating. As an aficionado of insect sex myself, I find that mind-boggling. How can you not wonder about the drama of courtship and mating evaluation that constitutes sexual behavior? Reluctantly, however, I admit that not everyone, including other scientists, shares my enthusiasm.
A landmark exception was Margaret Bastock, a PhD student working under the illustrious ethologist Niko Tinbergen at the University of Oxford in the 1950s. Tinbergen and most of his group studied vertebrates, mainly fish and birds, but Bastock wanted to use Drosophila to see how the relatively new science of genetics could help explain differences in behavior, and in turn, the way that behavior could shape evolution.2 She decided to use a mutant called yellow, caused by a single gene which, as one might imagine, renders the fruit flies golden colored. These mutant flies also occur in the wild, meaning that one can collect yellow fruit flies at garbage dumps and other places where the flies congregate, but yellow fruit flies are more common in the laboratory. Earlier researchers had seen that the male yellow fruit flies fathered fewer offspring, but didn’t know why.
Bastock first arranged some matings between the normal “wild type” (the Drosophila ordinarily found in nature) and yellow individuals. She reared their young to create groups of fruit flies that differed only in the yellow gene. She then proceeded to perform a series of exquisitely detailed experiments, including observations of courtship and mating behavior in which she sat in front of a pair of fruit flies and spoke into a microphone to record, on a reel-to-reel tape recorder, what the fruit flies were doing every one-and-a-half seconds. It is worth pausing for a moment to think about this less glamorous part of watching animals behave: peering at fruit flies under a microscope for hours on end is not much like watching lions in the Serengeti, though of course that can be tedious too.
Part of an amorous male fruit fly’s courtship includes vibrating his wings and licking the female while he positions himself behind her. The female has to cooperate for mating to occur, and males will often persist for long periods before either giving up or getting a chance to mate. Bastock found that yellow males did all of these courtship behaviors less often and for less time than the non-yellow flies, and that this lowered vigor seemed to be what was responsible for the difference in success at reproduction. Bastock wondered if the reason the males were lackluster was that the females did not react to them the same way they did to the wild-type fruit flies, rather than because of something in the males themselves. She did yet more observations to find out. Nope, the females employed the same behavior for all of the male flies. It really was, as she said in the title of her paper, “A Gene Mutation Which Changes a Behavior Pattern.”3
This was a big deal. No one had demonstrated a clear link between a known gene and what an animal did. What is more, Bastock pointed out that such genetically driven behavioral differences can play an important role in that most important of evolutionary processes, the formation of new species. As she recognized, if some of the males in a population of flies stopped vibrating their wings as much, perhaps some females would turn their attention to other features of courtship, like odor, or the pattern of spots on the wings. That in turn would mean that males with the newer method of attracting a mate were favored by selection, which could lead to the separation of these fruit flies from those that followed the more old-fashioned technique. Eventually, the populations might diverge so much that females from one group would not mate with males from the other, and vice versa. All because of a single gene that affected behavior.
Plastic Expressions
Since Bastock’s time, of course, research into the role of genes influencing behavior has exploded. But the novelty of her finding endures—it is extremely rare to find just one gene that has a big effect on complex behavior. And when it happens, it’s often not the presence or absence of a gene that is important, as it was with the yellow fruit flies, but a change in whether or not a gene is expressed.
To understand the distinction, remember that a gene is just the term we use for the DNA chunk that sits on a chromosome (there are actually philosophical debates about just what a gene is, but we’ll set those aside for our purposes). How does the DNA end up making a muscle fiber, or a nerve cell, or any of the other parts of the body? The answer is that the genes have to be expressed, which means that first the information they contain must be made into a different genetic material and then into a final product, like the aforementioned muscle fiber. Crucially, different genes are expressed at different times, which is why you have had the same genes all your life but only started growing body hair at puberty. It is a complicated process, and one that modern techniques have only recently allowed us to explore.
Back to behavior, and this time to a different set of insects: ants. Like honeybees, ants are highly social. Most ant species live in colonies with a queen that does all the reproducing and workers, all female, that do pretty much everything else—find food, care for the larvae, and defend the nest. The queen and workers are often genetically identical. So what makes a queen, a queen?
To answer this question, Daniel Kronauer from Rockefeller University in New York and his colleagues compared gene expression in seven different species of ants, some with queens and some that live in groups of workers that reproduce without mating, making what are essentially clones of themselves.4 The queenless groups appear to have arisen more recently in evolution, though like other ants, they cooperate to care for the larvae. Across the species, one gene was activated to manufacture more of its product in the reproducing individuals of a colony than in the workers: insulin-like peptide 2, or ILP2. This means that the most recent common ancestor of ants likely had high ILP2 expression in its reproducing individuals and low expression in its workers.
Then the researchers looked at the clonal raider ant Ooceraea biroi, which is one of the queenless species. They removed the larvae from colonies that were in their caregiving stage, and found that the expression of ILP2 went way up within twelve hours. Conversely, if ants were offered larvae, the expression of that gene went down. What is more, when they injected the peptide into workers in colonies with larvae (a procedure that requires a steady hand, a lot of practice, and a very tiny needle), their ovaries were switched on, as if they were about to reproduce. The higher the dose, the more eggs developed inside the workers’ bodies.
This insulin-like peptide is similar to the insulin found in the human body, and indeed in many kinds of animals. In the ants it appears to increase the likelihood that they will go foraging and then give the food to their larvae. An article in the New York Times about the research drew the analogy to people getting hungry when their insulin levels fall, although in the ants presumably that doesn’t induce preparing macaroni and cheese for the larvae.
In addition to illuminating a tiny part of how a complex social system like that found in ants could have evolved, the ILP2 story shows once again that genes by themselves don’t induce behavior. Having high or low levels of ILP2 activation means nothing without the presence or absence of larvae in the environment. At the same time, you can’t just make larvae without the necessary genetic machinery that switches on the ovaries.
Knowing more about gene expression also sheds light on the question of how new behaviors get incorporated in the genome. Work by renowned bee genetics expert Gene Robinson might have some answers. The African honeybee is very similar to our more familiar variety. It is famous, however, for its aggressive defense of the hive, as people have discovered when the African bees have been inadvertently introduced into places where people and bees come into contact. Robinson and his colleagues discovered5 that, as compared to other honeybees, the brains of the African bees show greater expression in genes that make them respond to the alarm pheromone, the chemical that signals danger to the hive and triggers the colony members to seek out and sting intruders. This means that the degree of ferocity, so to speak, is flexible, depending on gene expression, but not on the presence or absence of new genes.
Why might that flexibility have evolved? Imagine bees in an environment with many threats to the hive. In such a scenario, a colony with denizens that had greater hair-trigger responses to alarm pheromones would be more likely to survive than those with a more lackadaisical response. Thus, colonies with higher levels of gene expression would do better, which would then translate into an overall heightened level of aggression in that type of bee. Same genes, but different activation levels.
On Height and Lovebirds
Most traits, whether behavioral or not, are difficult or impossible to attribute to one or a few genes, and that is true in animals as well as people. Instead, they arise from a complex interaction (that word again) between many genes and the environment of an individual. The most common example of such a characteristic is how tall you are. Your height obviously reflects the heights of both your parents as well as the kind of environment you had as a child: if you were malnourished, you did not grow to be as tall as if you had an adequate diet, even if your parents were both above average in height. Many different genes contribute to height, which means that there are many different heights in any one population—if your mother was five feet four inches and your father five feet nine inches, your height might be the same as one of them, somewhere in between, or even shorter or taller than either of them.
This kind of distribution is common in animals as well, much more so than the single or few-gene associations illustrated by the bees, who either uncapped cells or didn’t (no bees uncapped a cell part of the way and then left the rest of the cap alone, or sniffed at a diseased larva and then did not attempt to remove it). That either-or dichotomy is rare. Instead, behaviors are more likely to occur along a spectrum, like height in humans.
Consider lovebirds, those African parrots often touted as paragons of monogamy. Different species of lovebirds can be bred with each other in captivity, but each species also has its own set of distinctive behaviors. William Dilger studied hybrids between two kinds of lovebirds in the 1950s and 1960s.6 He clearly harbored real affection for his study animals, reflecting in an article: “The partners exhibit their mutual interest with great constancy and in a variety of beguiling activities.” Both the Masked Lovebird and the Peach-faced Lovebird nest in tree holes, and both bring bark and grass back to the trees to use as nesting material, but they do so in completely different ways. The Masked Lovebird carries those materials in its beak, while the Peach-faced Lovebird carries grass by tucking it into its rump feathers. The hybrids that were the result of a male from one species mating with a female from the other acted, as Dilger put it, “completely confused.” They showed a range of behaviors—they might begin to carry one strip of material in the beak and then stop before they get to the nest, or they might carry several in the rump feathers and then lose them all partway. Unlike the bees, the lovebirds did not exhibit the behavior of one or another of the parental species, but did things that had not been observed in either. Interestingly, over time the nest-building behavior improved, so that after a few years, the hybrids successfully managed to build a structure that worked as a nest. Whether that was because they observed other birds or simply learned from their own experiences is unclear.
That kind of variety of behaviors in the offspring of a hybrid means that many genes are involved in influencing the behavior. If it were only one or two genes, we would see just one or two forms of behavior, as with the hygienic bees, and nothing in between. And the effect of learning shows that, once again, behaviors, no matter how rigid they may seem, are not produced in a vacuum.
What now? If many genes influence a behavior, and the environment, or culture, or learning all affect the behavior as well, is that it? Do we simply say, “It’s a lot of genes, and it’s complicated” and leave it at that? We could—and sometimes I think the truism that “things are complicated” gets less credit for its profundity than it deserves—but we don’t have to, because farmers got there first, and their discoveries have helped us understand how behavior can evolve even when many genes are involved.
Well, it is not entirely true that farmers made the discovery. But the interest of agriculturalists who wanted to get faster horses, bigger ears of corn, or cows that produced more milk helped motivate scientists and mathematicians to develop a way to numerically measure the degree to which ancestry mattered. As Darwin’s ideas about evolution were developing into their modern version, biologists argued about how heredity might work. They had read Mendel, but also acknowledged that everything was not inherited the same way. Farmers didn’t care about Darwin, but were interested in which plants or animals to use as seed or breeding stock, because it was important for them to know how much change they could expect after a certain number of generations. Breeding a cow that produced a pint more milk per day would be wonderful, but it would be of limited use if it took a thousand generations to get there. Eventually, the scientists developed a method to describe the way that differences in traits like height, which show a wide range from short to tall, could be attributed to genes and hence could likely be successfully selected to improve.
We call height a “quantitative trait,” and it and others like it can be studied using an approach called quantitative genetics. Rather than looking at individuals, quantitative genetics examines populations and the variation in characteristics that they contain. Why, for example, are Icelandic people taller and Argentinians shorter relative to other groups of humans? The answer is that they are genetically different, at least in part. But more than that, quantitative genetics asks about the variation within a group of individuals. Every Icelandic person isn’t the same height, and neither is every Argentinian. What is more, we know that if a given Argentinian has a poor diet as a child, he or she might end up shorter than if the same person were well fed. Using techniques originally developed for animal and plant breeders, it is possible to measure how much genes or the environment can explain variation in a trait, whether behavioral or not. The amount of variation that can be explained by the genes is referred to as “heritability,” a term that is so often misunderstood that virtually every discussion of it bemoans the confusion. Heritability is measured as a percentage or proportion, so that a given characteristic, whether height or anything else, can be assigned a number as a percent, between 0 and 100, or, if you prefer, a proportion, between 0 and 1. But that score is not something you carry around with you, or can use to describe yourself. In fact, it isn’t a characteristic of an individual at all.
Part of the problem is that people tend to gloss over the words variation and population and go straight for the “how much is genes” part. But that is a mistake. To illustrate, let’s think about plants, as the Harvard University geneticist Richard Lewontin did in his classic example of the concept.7 Imagine that you take a group of basil seedlings and plant them in two trays of soil. You put one tray in a place with sun, and you water and fertilize the seedlings as they grow. The other tray is left in a gloomy corner of your yard, and you hardly ever remember to water it. Both groups of plants eventually mature, and the plants in the first tray are, on average, taller than those in the second tray. What is the heritability of plant height? You can’t answer that question right away, and no answer would pertain to all the basil plants. First, you have to know the variation in height among the plants within each tray. In the first tray, say that some plants are ten inches high, some are twelve, some are fourteen, and one each is eighteen and six inches. Why are the differences there? The heritability within that tray is close to 100 percent; the variation among the individual plants only occurs because the seeds each had different genes, since they were all in the same environment. Note that I said “close to 100 percent.” That is because it is virtually impossible that each seedling truly experienced the same environment. Perhaps water was more likely to pool at one end of the tray, or perhaps the seedlings at the corners had somewhat more room. Still, it is reasonable to conclude that the variation among the plants—not the height of any single individual—can be attributed to their genes.
Now let us turn our attention to the second, neglected, tray. In contrast to the first, say that its inhabitants are between four and ten inches tall. Once again, we can measure the heritability by examining the variation among the individuals within the tray. And we can also determine that the difference between the trays is because of the difference in their environments. However, we can never talk about heritability as an individual characteristic, and we can never talk about it separate from the environment in which we measured it. People sometimes say that “heritability is a local measure,” emphasizing that the degree to which genes account for the variation in a trait depends on the circumstances when it is measured. This also means that heritability can increase or decrease: if you reduce the variation in the environment that an organism is in, the remaining variance you see in the trait you are measuring must be due to the genes, and since heritability is a proportion, the number describing that gene variance has to go up.
The Power of Two, and Evolutionary Bookkeeping
The plant example is clear in part because the experiment is so controlled. But we can calculate heritability in animals, as well as in people, and it is perfectly possible to determine heritability in behavior as well as in physical attributes. In humans a common way of doing so is to take advantage of the genetic similarities in twins. As you probably know, human twins come in two basic flavors: identical, which means they started out as a single fertilized egg that then split, with each of the resulting halves genetically the same as the other; and fraternal, which are like ordinary siblings in that they arose from two different eggs fertilized by different sperm.
Scientists recognized the power of examining twins many years ago, and my own university, the University of Minnesota, has a famous place for doing just that, the Minnesota Center for Twin and Family Research.8 The basic idea is simple: if we compare identical twins that were raised in different environments, perhaps because each member was adopted into a different family outside of the birth family, then we can see if they grow up to be more like each other or if they become more like the siblings in their respective adoptive family. If the twins are different, that difference may be said to be due to their different environments. Similarly, we can compare the traits of fraternal twins raised in the same environment: if they are then different, that difference is due to their genes, not the environment.
Everyone, including the Minnesota Center, recognizes that these broad generalizations are just that. The research is quite complicated, using repeated questionnaires, family history data, sophisticated statistical modeling, and more. Since its start in 1989, the Center has gathered information from over 9,800 people to date, and they have examined behaviors ranging from eating disorders to personality traits to happiness. The results have been illuminating in many ways, perhaps most by acknowledging the importance of both genes and environment in a wide range of characteristics. For example, even fingerprints are not identical in identical twins, though they are more similar in the pair than they are between fraternal twins. And attributes such as general levels of happiness or a tendency to some personality disorders, while more similar in identical twins, are also influenced by family life. Sometimes genetic influences change over time.9 The twins’ concerns about body shape and weight were not as affected by genetic similarity in preadolescent twins as they were in twins from early adolescence onward. Twins understandably fascinate us, but their similarities don’t always mean what people think, as I detail later on.
Animals don’t lend themselves to twin studies in quite the same way, but in many species the offspring are born or hatched in groups, and while the siblings are not genetically identical, they still share more of their genes than a random pair of individuals. To measure heritability of a characteristic in animals, one can approximate the scenario of a twin raised by a family other than the one he or she was born in by doing what is called a cross-fostering study. These are most frequently accomplished using songbirds, with their handy nests full of chicks as well as their general inability to recognize their own offspring. (Side note, and public service announcement: the idea that if you touch a baby bird on the ground and put it back in its nest, its mother will reject it because it “smells of human” is a myth, perhaps perpetuated by overworked mothers who did not want their children messing with chicks in the first place. Please, by all means, if you see the nest, put the baby back—the mother, like most birds, has a rather poor sense of smell and will feed it again without skipping a beat, or a worm. Or you can call a wildlife rehabilitation center.) For cross-fostering, a researcher selects two nests with eggs of similar age and waits for the chicks to hatch. Once they do, half of each brood is swapped between them, so that the parents of nest number one raise half their own chicks and half from nest number two, and vice versa. The researchers can then measure the characteristics of the chicks in each nest and compare them to the parents that reared them and to their genetic parents.
One can also do an animal heritability study experimentally, depending on the species in question. My one experience with such a study used Red Junglefowl, the ancestor of domestic chickens, when I was a postdoctoral researcher at the University of New Mexico in Albuquerque. Chickens have been domesticated for so long that their husbandry is well understood, so we could use many of the same techniques employed for studies of flies or beetles. We know how to feed and house them, and we know how big an area they like to have to thrive. In our research, we were interested in how the elaborate ornaments of the roosters, including their fleshy combs and wattles, had evolved, and whether the hens preferred characteristics that might indicate a male’s health.
I loved working with the birds, and it gave me a fondness for chickens and their kin that I harbor to this day, but that experiment was a nightmare. We had to ensure which hens mated with which roosters, take their eggs, individually mark them with pens, and rear them in incubators so their rearing environment was as standardized as possible. We also had to keep track of how many eggs were produced, how many chicks hatched from each family group, and then measure various attributes of the chicks as they grew up, including the size of the roosters’ combs, the length of various feathers, and the kinds of mates the hens preferred. A large sample size is essential because of the complex calculations one needs to perform, so we were rearing hundreds of chicks, each of which had to be tracked individually. It was a gigantic exercise in bookkeeping, and I developed enormous respect for accountants who work with large data sets of any kind. Not to mention for poultry farmers.
It turned out that heritability was highest in rooster traits such as tail feather length or the color of the neck feathers, neither of which were all that important to the hens in choosing a mate. Intriguingly, males with larger combs sired larger chicks than males with scrawnier ornamentation, which is consistent with the idea that the comb shows a rooster’s general vigor. Much remains to be done to pin down exactly how the attractive traits, and the preference for them, are inherited. It just won’t be done by me.
Be that as it may, we, and many other scientists, have calculated heritabilities of behaviors ranging from courtship frequency in fruit flies and junglefowl to preening in Japanese Quail to learning in pigs. One paper10 reviewed fifty-seven of such studies and found that the average heritability was 38 percent, though the range was substantial. This figure is well in keeping with the heritabilities of human behaviors and behavioral disorders such as anxiety disorders or major depression, though it is somewhat lower than the estimates of 50 percent to 60 percent for alcoholism and somewhat higher than the figures of 15 percent to 20 percent for extraversion and assertiveness. In a 2006 paper from the American Journal of Psychiatry, the authors conclude, “With respect to the broad patterns of genetic influences on behavior, Homo sapiens appears to be typical of other animal species.”11
Two important conclusions follow from this kind of research. The first is that these numbers do not mean what many people think they mean. Heritability is not a measure of “how genetically based” a particular characteristic is, whether we are talking about a behavior like cheerfulness or a more objective trait like height. A heritability of 38 percent does not mean that a fly gets 38 percent of its courtship behavior from its genes and the remaining 62 percent from its environment any more than one gets 60 percent likelihood of becoming an alcoholic from one’s parents. Remember, heritability only means something in reference to a population in the environment in which the characteristic was measured.
Like the zombie idea of nature versus nurture, however, heritability has become something of a zombie measurement, so that no matter how many times its limitations are explained, the misunderstanding revives itself. A case in point is the interpretation of a 2015 paper12 that examined hundreds of human twin studies for traits ranging from developmental diseases to personality components to (quite intriguing to my mind) “Mental and Behavioural Disorders Due to Use of Tobacco.” The authors noted that “across all traits the reported heritability is 49%.” Media coverage and commentary on the work used that 49 percent statistic to conclude that nature and nurture were “tied,” as if a longstanding battle had finally been resolved.
Not so fast. For one thing, the authors of the paper obviously could only review studies of traits for which there are data—if someone didn’t look for the heritability of a trait, that trait didn’t appear in the analysis. For another, as noted earlier, all the heritability estimates apply only to a particular place and time for a given population. IQ scores, for example, have famously increased in the United States from an estimated seventy to eighty in the 1940s to closer to one hundred in the 1990s—and even ignoring the many, many problems associated with using IQ tests as a measurement, no one would argue that the human genes associated with whatever it is that IQ measures have changed that quickly. Instead, the environment changed, and the scores are only relevant (again, to whatever extent they ever are relevant) in that environment.
The second conclusion is that although we are fascinated by the idea that aspects of our own behavior are influenced by our genes in the context of our environment, whether a tendency toward optimism or our economic achievement, there is nothing special about the way the environment influences human behavior. All the caveats about heritability and its dependence on the environment, its nature as a population measure and not an individual one, and its lack of correspondence with “how genetic” something is, also apply to animals. Animals are not more or less controlled by their genes than humans, because neither of us is “controlled” by our genes at all. What that means is that while understanding the relationship between genes and behavior is valuable, humans do not have a premium on complicated interactions between genes and the environment. Eric Turkheimer,13 whose observations about genetics I mentioned in the introduction of this book, proposed the “first law of behavioral genetics”: all human behavioral traits are heritable. Turkheimer is, of course, a psychologist trained in behavioral genetics, so perhaps the stipulation of humans is understandable. But, in fact, behavioral traits are heritable whether they are in people or worms, and we shouldn’t be more surprised about that finding in one or the other. Both humans and animals also influence their environment in many ways depending on their genes, as evidenced by the concept of niche construction discussed earlier in this chapter.
A slightly silly but frequently used example may also illustrate the way that heritability cannot be equated with “genetic.” I am completely confident that every person reading this paragraph has a head. Presumably we can all agree that one is born with one’s head—it is not a manifestation of the environment in which one lives, but a complex result of biological processes that occur during the development of an embryo. Although no “head-producing gene” exists per se, the existence of our heads depends on our genes. So what is the heritability of having a head? Zero. That is because any variation in head possession can be attributed to environmental factors, such as the preponderance of guillotines.
Despite all the limitations to its interpretation, understanding heritability is crucial if you want to examine how behavior can evolve. Recall that evolution means that genes are changing in a population over time in response to selection, so that, as in my earlier example, the birds became greener because the more camouflaged individuals had more babies. How much greener they get, and how fast, depends on two things: the strength of selection and heritability. Strength of selection means how much of an advantage a given difference in a trait confers; if only the birds that were slightly greener survived, while all their less-green compatriots were snatched up by sharp-eyed hawks, selection is stronger than if green feathers meant just a slight edge in camouflage. The greater the strength of selection, and the higher the heritability, meaning the higher the proportion of the variation in the trait is ascribed to genes, the more the population can evolve. So it’s helpful to understand how genes influence behavior. But we won’t get anywhere by pitting genes against the environment and expecting a winner to emerge.
Genes, Income, and Flies
None of the studies of heritability, whether on human twins or cattle, claim to identify particular genes or even groups of genes that are linked to a particular behavior—all they can do is evaluate the proportion of the variation we see that can be attributed to genetic variation, with “genetic variation” as a one-size-fits-all description. Over the last decade or so, however, advances in molecular biology, and the ability to sequence the genome, have allowed scientists to look much more closely at how people or animals with different characteristics differ at particular parts of their genome. It is this latter development that has led to headlines about genes linked to everything from sexual orientation to liking dogs to a predilection for getting divorced. In addition, the rise in genetic ancestry tests such as 23andMe has contributed to the idea that we can survey our genomes and pick out the parts that make us, say, athletic or slothful.
As many people have pointed out, that conclusion—and the idea that we can identify something like a “gene for sexual orientation”—is false. To understand why, it is useful to know just what is being compared in all those studies that garner “Genes Explain Your Income Level” types of headlines. The research uses Genome Wide Association Studies, usually abbreviated as GWAS, and pronounced “gee-wahss.” To perform a GWAS, you first identify a characteristic of interest. In the earlier days of the technique, these were usually diseases, and medical applications are probably still the most common reasons for this research. Then you find a group of people with the disease and a group without it, and you examine some portion of their genomes—no one is actually sequencing all thirty thousand genes in the human genome—and look for differences in those small chunks. If the sample is large enough, the idea is that any consistent differences you see between the two groups are due to the difference in the presence of the disease. One can also use the results to develop a polygenic score or summary of the gene variants associated with the trait of interest.
These surveys are very useful, but they do not tell us “how genetic” any behavior, disease, or other characteristic might be. As I have already pointed out, genes do not code for behaviors directly, and they do not exist in a vacuum.
One of the most sensitive areas for exploration of this issue has been with intelligence, or at least with scores on IQ tests. Intelligence was one of the first behavioral characteristics that early geneticists wanted to understand, and the idea that we cannot change people’s intelligence because they inherit it as a fixed, immutable property has dogged social programs for over a century. A set of articles in the Wall Street Journal in 2020 revisited this issue after a study was published that examined the association between genetic variation and some cognitive traits like how much education people had and their IQ. After seeing how much their work had been distorted in the media, authors Michelle Meyer, Patrick Turley, and Daniel Benjamin said:14
IQ is not a fixed attribute of individuals and can be affected—for better and worse—by the environment in myriad ways. For example, in a society where people of color are denied access to childhood enrichment programs or adequate nutrition, a polygenic score for IQ might reflect genetic variants associated with skin pigmentation. Relatedly, in a sexist society, variants on the X and Y chromosomes, which determine biological sex, might be related to a variety of socio-economic phenotypes. Such polygenic scores would indeed moderately predict the IQ of people on average, but—and this is key—much of that predictive power would simply reflect social choices, not innate or immutable biology.
Everything that Meyer and her colleagues say here is true, but what many people do not appreciate is that it is true not only for people, with our rich social environments and complex development, but for animals. The problem isn’t that GWAS or heritability are unsuitable for explaining how human behaviors are fixed by the genes, it’s that this isn’t what they explain at all, in humans or anything else. A recent study15 of that tiny genetic powerhouse the fruit fly found that in a wide variety of behaviors, especially mating behavior, hundreds and often thousands of genetic variants were important. The research pointed to some interesting prospects for future work—for example, some of the genetic variation associated with body size was also important in how much male fruit flies court females. This raises questions like: What are the shared neurological pathways that cause such a relationship? In other words, we aren’t going to uncover a gene, or even a handful of genes, that by themselves determine come-hither signals in fruit flies. This reality I hope underscores the futility of thinking we can ever do something comparable for income levels or divorce in human beings.
The Essence, or Instinct, of It All
Despite our efforts to eradicate the zombie of nature versus nurture and related misconceptions about what heritability really means, both are remarkably persistent ideas. In a survey published in 2010 of 1,200 American adults, 76 percent of respondents believed that “single genes directly control specific human behaviors.”16 I realize that scholars have long been lamenting the ignorance of the public on issues ranging from the geographic location of countries to the efficacy of antibiotics for treating viral infections, and that such surveys can be questioned for their propensity to generalize. With that caveat, this particular misconception is still troubling for at least two reasons. First, it means that the influence of the environment on all characteristics, whether behavior or not, is ignored, when in truth genes don’t single-handedly determine anything. Second, and perhaps of even more concern, this misconception can lead to acceptance of bad, or even criminal, behavior. Psychologist Steven Heine from the University of British Columbia notes17 that “men show increased moral acceptance of undesirable behaviors such as date rape when genes are even remotely implicated as opposed to societal forces.”
So why can’t we accept the limitations of these measures? Perhaps, as Heine and other psychologists, particularly Ilan Dar-Nimrod at the University of Sydney, have proposed, we are genetic essentialists. According to this view, many people harbor, as one of Heine’s papers18 says, an “innate set of psychological intuitions that lead us to think about genetic concepts in a highly inaccurate and biased way.” Leaving aside the presumably unintentional irony of referring to misconceptions about genetics as “innate,” essentialism can be a way to view all things as having an internal, immutable “essence” that makes them the way they are. This idea stems from Aristotle, who, as Heine and colleagues say, “famously proposed that every entity possesses an essence that ultimately makes it what it is and that, without such an essence, the entity would no longer be itself.” Hence, as I mentioned in an earlier chapter, an ant has an “ant-like” essence that makes it an ant, apart from the more tangible things like its antennae, its small size, or its love of sugar and picnics.
Applied to genetics, this means that we imagine our genes as surrogates for that essence, a way for us to view our identities. Genes, like essences, are therefore seen as fixed, and so determine who we are. That attitude in turn might make us less inclined to change ourselves, or to attribute poor outcomes to irresistible forces, as with the date rape example. Research also shows that people ate more cookies after reading about potential genetic contributions to obesity than if they had read about the way the environment affects obesity. Both readings were accurate, but apparently being reminded of the way that genes play a role in our lives changed the subjects’ behavior in the short term. A similar view has also been applied to human racial categories, with racial essentialists arguing that people of different races have physical and psychological differences that are both fixed by their genes and similar among all the members of a race. Neither of these generalizations are true, but the errors are, again, remarkably persistent, to the point where educator Brian Donovan suggests,19 “There appears to be a hidden racial curriculum in biology textbooks that is learned by students but never purposefully taught by teachers.”
Psychologists have explored the significance of essentialist views, and the difficulty they pose for social change, but I want to make a slightly different point here. It’s true that genes don’t give humans their “essence,” and indeed the idea of ineffable essences is an odd holdover from ancient times. But here’s the thing: animals don’t have essences either, and their genes don’t do any better at determining their identities. Many of the authors who provide caveats to the notion that genes determine behavior take pains to confine themselves to humans, pointing out that humans are incredibly complex. One such paper states,20 “A typical human behavioral trait is associated with very many genetic variants, each of which accounts for a very small percentage of the behavioral variability.” That is absolutely true. But it is also true for typical animal behavioral traits. Part of why the work on the yellow fruit fly mutants, or the ILP2 gene, is so fascinating is that such relatively simple relationships are so rare.
A stand-in for that amorphous essence that governs what we do is sometimes called instinct. The concept of instincts is also quite old, and it is sometimes used to distinguish animals from humans, with the former supposedly acting instinctively, meaning without any cognitive process, and the latter able to reason and choose. People are sometimes surprised when I tell them that biologists don’t use the word instinct much in animal behavior research anymore. To me it is a non-explanation; saying that a bird building a nest is behaving instinctively, as the Star Tribune newspaper article I mentioned earlier did, just fobs off the question. All the word instinct tells us is that whoever called the behavior an instinct didn’t know—and indeed, we rarely do—how the environment and genes interacted to produce that particular behavior. Instincts don’t exist, the same way that essences don’t exist—animals aren’t born carrying a mating instinct, a feeding instinct, or any other kind of instinct.
Recognizing, once again, that neither humans nor behaviors are special cases frees us from having to explain their exceptionalism. More important, it means that we can see the extraordinary flexibility and complexity of nonhuman animals and most if not all of their characteristics, whether physical or behavioral or an inextricable mix of both.