10.3 Where Does Intelligence Come From?

No one is born knowing calculus and no one has to be taught how to blink. Some things are learned, others are not. But almost all of the really interesting things about people are a joint product of the innate characteristics with which their genes have endowed them and of the experiences they have in the world. Intelligence is one of those really interesting things that is influenced both by nature and by nurture. Let’s start with nature.

Genetic Influences on Intelligence

The notion that intelligence is “in the blood” has been with us for a long time. For example, in The Republic, the philosopher Plato suggested that some people are born to rule, others to be soldiers, and others to be tradesmen. But it wasn’t until late in the 19th century that this suggestion became the subject of scientific inquiry. Sir Francis Galton was a half cousin of Charles Darwin, and his contributions to science ranged from meteorology to fingerprinting. Late in life, Galton (1869) became interested in the origins of intelligence. He did careful genealogical studies of eminent families and collected measurements from over 12,000 people that ranged from head size to the ability to discriminate tones. As the title of his book Hereditary Genius suggests, he concluded that intelligence was inherited. Was he right?

Sir Francis Galton (1822–1911) studied the physical and psychological traits that appeared to run in families. In his book Hereditary Genius, he concluded that intelligence was largely inherited.

MARY EVANS PICTURE LIBRARY/THE IMAGE WORKS
© BETT MANN/CORBIS

Studying Relatives

Small genetic differences can make a big difference. A single gene on chromosome 15 determines whether a dog will be too small for your pocket or too large for your garage.

DEANNE FITZMAURICE

The fact that intelligence appears to “run in families” isn’t very good evidence of this genetic influence. After all, brothers and sisters share genes, but they share many other things as well. They typically grow up in the same house, go to the same schools, read many of the same books, and have many of the same friends. Members of a family may have similar levels of intelligence because they share genes, environments, or both. To separate the influence of genes and environments, we need to examine the intelligence test scores of people who share genes but not environments (e.g., biological siblings who are separated at birth and raised by different families), people who share environments but not genes (e.g., adopted siblings who are raised together), and people who share both (e.g., biological siblings who are raised together).

There are several kinds of siblings with different degrees of genetic relatedness. When siblings have the same biological parents but different birthdays, they share on average 50% of their genes. Fraternal Twins (or dizygotic twins) develop from two different eggs that were fertilized by two different sperm, and although they happen to have the same birthday, they are merely siblings who shared a womb and they, too, share on average 50% of their genes. Identical Twins (or monozygotic twins) develop from the splitting of a single egg that was fertilized by a single sperm, and unlike other siblings, they are genetic copies of each other, sharing 100% of their genes.

These different degrees of genetic relatedness allow psychologists to assess the influence that genes have on intelligence. Studies show that the IQs of identical twins are strongly correlated when the twins are raised in the same household (r = .86), but they are also strongly correlated when the twins are separated at birth and raised in different households (r = .78). In fact, as you’ll notice from TABLE 10.3, identical twins who are raised apart have more similar IQs than do fraternal twins who are raised together.

These photos, taken from the “Genetic Portraits” series by artist Ulric Collette, were made by blending the face of a 32-year old woman with the face of her mother (left) and the face of her father (right). Because physical appearance is highly heritable, the resemblance between family members can often be quite striking.

Ulric Collette

What this means is that people who share all their genes have similar IQs regardless of whether they share their environments. Indeed, the correlation between the intelligence test scores of identical twins who have never met is about the same as the correlation between the intelligence test scores of a single person who has taken the test twice! By comparison, the intelligence test scores of unrelated people raised in the same household (e.g., two siblings, one or both of whom were adopted) are correlated only modestly, about r = .26 (Bouchard & McGue, 2003). These patterns of correlation suggest that genes play an important role in determining intelligence. This shouldn’t surprise us. Intelligence is, in part, a function of how the brain works, and given that brains are designed by genes, it would be quite remarkable if genes didn’t play a role in determining a person’s intelligence. Indeed, a mere 20 genes on a single chromosome are all that separates you from a person with Williams syndrome. Clearly, genes influence intelligence.

Heritability

But exactly how powerful is that influence? The heritability coefficient (commonly denoted as h²) is a statistic that describes the proportion of the difference between people’s scores that can be explained by differences in their genes. When the data from numerous studies of children and adults are analyzed together, the heritability of intelligence is roughly .5, which is to say that about 50% of the difference between people’s intelligence test scores is due to genetic differences between them (Plomin & Spinath, 2004; Plomin et al., 2013; cf. Chabris et al., 2012).

This fact may tempt you to conclude that half your intelligence is due to your genes and half is due to your experiences, but that’s not right. To understand why, consider the rectangles in FIGURE 10.5. These rectangles clearly differ in size, and if you were asked to say what percentage of that difference was due to differences in the rectangles’ heights and what percentage was due to differences in the rectangles’ widths, you would correctly say that 100% of the difference in their sizes was due to differences in their widths and 0% was due to differences in their heights (which are, after all, identical). Good answer. Now, if you were asked to say how much of rectangle A’s size was due to its height and how much was due to its width, you would correctly say, “That’s a silly question.” It is a silly question because a rectangle’s size is a product of both its height and its width and it can’t be “due” to one more or less than the other.

Figure 10.5: How to Ask a Silly Question These four rectangles differ in size. How much of the difference in their sizes is due to differences in their widths and how much is due to differences in their heights? Answer: 100% and 0%, respectively. Now, how much of rectangle A’s size is due to width and how much is due to height? Answer: That’s a silly question.

Tamara Rabi and Adriana Scott were 20 years old when they met in a McDonald’s parking lot in New York. “I’m just standing there looking at her,” Adriana recalled. “It was a shock. I saw me” (Gootman, 2003). The two soon discovered that they were twins who had been separated at birth and adopted by different families.

©Angel Franco/The New York Times/Redux

Research shows that intelligence is more heritable in wealthy neighborhoods than in poor ones.

©Noah Addis/Corbis

Similarly, if you measured the intelligence of all the people at a basketball game and were then asked to say what percentage of the difference in their intelligences was due to differences in their genes and what percentage was due to differences in their experiences, you could reasonably guess that about 50% of the difference was due to each of these factors. That’s what the heritability coefficient of .5 suggests. But if you were next asked to say how much of the intelligence of the annoying guy with the bad haircut in Row 17, Seat 4 was due to his genes and how much was due to his experiences, you could only reply, “That’s a silly question.” It is a silly question because the intelligence of a particular person is a joint product of both genes and experience–just like the size of a particular rectangle is a joint product of both its height and width–and cannot be “due” to one of these things more than another.

The heritability coefficient tells us why people in a particular group differ from one another; thus its value can change depending on the particular group of people we measure. For example, the heritability of intelligence among wealthy children is about .72 and among poor children about .10 (Turkheimer et al., 2003). How can that be? Well, if we assume that wealthy children have fairly similar environments–that is, if they all have nice homes with books, plenty of free time, ample nutrition, and so on–then all the differences in their intelligence must be due to the one and only factor that distinguishes them from each other, namely, their genes. Conversely, if we assume that poor children have fairly different environments–that is, some have books and free time and ample nutrition while others have little or none of these–then the difference in their intelligences may be due to either of the factors that distinguish them, namely, their genes and their environments (Tucker-Drob et al., 2010). The value of the heritability coefficient also depends on the age of the people being measured and is typically larger among adults than among children (see FIGURE 10.6), which suggests that the environments of any pair of 65-year-olds tend to be more similar than the environments of any pair of 3-year-olds. In short, when people have identical experiences, then the difference in their intelligences must be due to the difference in their genes, and when people have identical genes, then the difference in their intelligences must be due to the difference in their experiences. It may seem paradoxical, but in a science-fictional world of perfect clones, the heritability of intelligence (and of everything else) would be zero.

Figure 10.6: Age and Heritability of Intelligence The heritability of intelligence generally increases with the age of the sample measured.

Does this imply that in a science-fictional world of individuals who lived in exactly the same kinds of houses and received exactly the same kinds of meals, educations, parental care, and so on, the heritability coefficient would be 1.00? Not likely. Two unrelated people who live in the same household will have some but not all of their experiences in common. The shared environment refers to those environmental factors that are experienced by all relevant members of a household. For example, siblings raised in the same household have about the same level of affluence, the same number and type of books, the same diet, and so on. The nonshared environment refers to those environmental factors that are not experienced by all relevant members of a household. Siblings raised in the same household may have different friends and teachers and may contract different illnesses. This may be why the correlation between the IQ scores of siblings is greater when they are close in age (Sundet, Eriksen, & Tambs, 2008). Being raised in the same household is only a rough measure of the similarity of two people’s experiences (Turkheimer & Waldron, 2000); thus, studies of twins may overestimate the influence of genes and underestimate the influence of experiences (Nisbett, 2009). As psychologist Eric Turkheimer (2000, p. 162) noted:,

First-born children tend to be more intelligent than their later-born siblings. But when a first-born child dies in infancy and the second-born child becomes the oldest child in the family, that second-born child ends up being just as intelligent as the average first-born child (Kristensen & Bjerkedal, 2007). This suggests that first-borns are smarter than their siblings because they experience a different family environment. So if Joe and Nick murdered Paul…um, never mind.

Dan Hallman/Invision/AP

Environmental Influences on Intelligence

At 7′, Pieter Gijselaar is taller than most of his friends–but not that much taller. The Dutch government recently adjusted building codes so that doors must now be 7′6.5″ high.

AP PHOTO/PETER DEJONG

Americans believe that every individual should have an equal chance to succeed in life, and one of the reasons why we bristle when we hear about genetic influences on intelligence is that we mistakenly believe that our genes are our destinies–that “genetic” is a synonym for “unchangeable” (Pinker, 2003). In fact, traits that are strongly influenced by genes may also be strongly influenced by the environment. Height is a heritable trait, which is why tall parents tend to have tall children; and yet, the average height of Korean boys has increased by more than 7 inches in the last 50 years simply because of changes in nutrition (Nisbett, 2009). In 1848, 25% of all Dutch men were rejected by the military because they were less than 5′ 2″ tall, but today the average Dutch man is over 6′ tall (Max, 2006). Genes may explain why two people who have the same diet differ in height–that is, why Chang-sun is taller than Kwan-ho and why Thijs is taller than Daan–but they do not dictate how tall any of these boys will actually grow up to be.

Is intelligence like height in this regard? Alfred Binet (1909) thought so:

It turns out that Binet was right. As FIGURE 10.7 shows, intelligence changes over time (Owens, 1966; Schaie, 1996, 2005; Schwartzman, Gold, & Andres, 1987). For most people, intelligence increases between adolescence and middle age and then declines thereafter. The sharpest decline occurs in old age (Kaufman, 2001; Salthouse, 1996a, 2000; Schaie, 2005), and may be due to a general slowing of the brain’s processing speed (Salthouse, 1996b; Zimprich & Martin, 2002). Age-related declines are more evident in some domains than in others. For example, on tests that measure vocabulary, general information, and verbal reasoning, people show only small changes from the ages of 18 to 70, but on tests that are timed, have abstract material, involve making new memories, or require reasoning about spatial relationships, most people show marked declines in performance after middle age (Avolio & Waldman, 1994; Lindenberger & Baltes, 1997; Rabbitt et al., 2004; Salthouse, 2001).

Figure 10.7: Absolute Intelligence Changes Over Time

Data from Kaufman, 2001.

105-year-old Khatijah (front row, second from right) sits with five generations of her family. The Flynn Effect suggests that intelligence is increasing across generations.

AP PHOTO/BINSAR BAKKARA

Not only does intelligence change over the life span, but it also changes over generations. The Flynn effect refers to the accidental discovery by James Flynn that the average IQ score is 30 points higher than it was about a century ago (Dickens & Flynn, 2001; Flynn, 2012; cf. Lynn, 2013). The average person today is smarter than 95% of the people who were living in 1900! Why is each generation scoring higher than the one before it? Some researchers give the credit to improved nutrition, schooling, and parenting (Lynn, 2009; Neisser, 1998), and some suggest that the least intelligent people are being left out of the mating game (Mingroni, 2007). But most (and that includes Flynn himself) believe that the industrial and technological revolutions have changed the nature of daily living such that people now spend more and more time solving precisely the kinds of abstract problems that intelligence tests include–and as we all know, practice makes perfect (Flynn, 2012). In other words, you are likely to score higher on an IQ test than your grandparents did because your daily life is more like an IQ test than theirs was!

Now, here is a fact that is almost sure to confuse you: Although intelligence changes over the lifetime, there is a strong correlation between an individual’s performance on intelligence tests that are taken at two different times (Deary, 2000; Deary et al., 2004; Deary, Batty, & Gale, 2008; Deary, Batty, Pattie, & Gale, 2008). TABLE 10.4 shows the results of several studies that demonstrate this fact. How can that be? If intelligence changes over time, then why is there such a strong correlation between tests taken in childhood and tests taken in old age? The answer is that intelligence does change over time, but it changes in much the same way for everyone. The large correlations between tests administered at different times merely tell us that the people who got the best (or worst) scores when the test was administered the first time also tended to get the best (or worst) scores when it was administered the second time. You already know that the same thing happens with height. People get taller as they go from childhood to adulthood, and yet, the tallest child is likely to be among the tallest adults. Like height, a person’s absolute level of intelligence changes over time, but his or her level of intelligence relative to others stays about the same.

The fact that intelligence changes over the life span and across generations shows that it is not “a fixed quantity that cannot be increased.” Our genes may determine the range in which our IQ is likely to fall, but our experiences determine the exact point in that range at which it does fall (Hunt, 2011; see FIGURE 10.8). Two of the most powerful experiential factors are economics and education.

Economics

Figure 10.8: Genes and Environment Genes may establish the range in which a person’s intelligence may fall, but environment determines the point in that range at which the person’s intelligence will fall. Although Jason’s genes give him a better chance to be smart than Josh’s genes do, differences in their diets could easily cause Josh to have a higher IQ than Jason.

PHOTODISC

Maybe money can’t buy love, but it sure appears to buy intelligence. One of the best predictors of a person’s intelligence is the material wealth of the family in which he or she was raised–what scientists call socioeconomic status (SES). Studies suggest that being raised in a high-SES family rather than a low-SES family is worth between 12 and 18 IQ points (Nisbett, 2009; van Ijzendoorn, Juffer, & Klein Poelhuis, 2005). For example, one study compared pairs of siblings who were born to low-SES parents. In each case, one of the siblings was raised by his or her low-SES parents and the other was adopted and raised by a high-SES family. On average, the child who had been raised by high-SES parents had an IQ that was 14 points higher than his or her sibling (Schiff et al., 1978). Although these siblings had similar genes, they ended up with dramatically different IQs simply because one was raised by wealthier parents.

Exactly how does SES influence intelligence? One way is by influencing the brain itself. Low-SES children have poorer nutrition and medical care, they experience greater daily stress, and they are more likely to be exposed to environmental toxins such as air pollution and lead–all of which can impair brain development (Chen, Cohen, & Miller, 2010; Evans, 2004; Hackman & Farah, 2008). The fact that low SES can impair a child’s brain development may explain why children who experience poverty in early childhood are less intelligent than those who experience poverty in middle or late childhood (Duncan et al., 1998).

SES affects the brain, and it also affects the environment in which that brain lives and learns. Intellectual stimulation increases intelligence (Nelson et al., 2007), and research shows that high-SES parents are more likely to provide it (Nisbett, 2009). For instance, high-SES parents are more likely to read to their children and to connect what they are reading to the outside world (“Billy has a rubber ducky. Who do you know who has a rubber ducky?”; Heath, 1983; Lareau, 2003). When high-SES parents talk to their children, they tend to ask stimulating questions (“Do you think a ducky likes to eat grass?”), whereas low-SES parents tend to give instructions (“Please put your ducky away”; Hart & Risley, 1995). By the age of 3, the average high-SES child has heard 30 million different words, while the average low-SES child has heard only 10 million different words, and as a result, the high-SES child knows 50% more words than his or her low-SES counterpart. These differences in the intellectual richness of the home environment may explain why children from low-SES families show a decrease in intelligence during the summer when school is not in session, whereas children from high-SES families do not (Burkham et al., 2004; Cooper et al., 1996). Clearly, poverty is the enemy of intelligence (Evans & Kim, 2012).

HOT SCIENCE: Dumb and Dumber?

For most of human history, the smartest people had the most children, and our species reaped the benefits. But in the middle of the 19th century, this effect began to reverse and the smartest people began having fewer children, a trend that scientists call dysgenic fertility. That trend continues today.

But wait. If the smartest people are having the fewest children, and if IQ is largely heritable, then why–as James Flynn showed–is IQ rising over generations?

Some researchers suspect that two things are happening at once: Our inherited intelligence is going down over generations, but our acquired intelligence is going up! In other words, we were all born with a slightly less capable brain than our parents had, but this small effect is masked because we were born into a world that greatly boosted our intelligence with everything from nutrition to video games! How can we tell whether this hypothesis is right?

Francis Galton was the first to suggest that reaction time (the speed with which a person can respond to a stimulus) is a basic indicator of mental ability, and his suggestion has been confirmed by modern research (Deary, Der, & Ford, 2001). Recently, a group of researchers (Woodley, te Nijenhuis, & Murphy, 2013) went back and analyzed all available data on human reaction time collected between 1884 and 2004 (including data collected by Galton himself) and what they found was striking: The average reaction time has gotten slower since the Victorian era! The figure above shows the average reaction time of people in different studies conducted in different years.

Does this mean that we are innately less clever and that this fact is being obscured by the big IQ boost we get from our environments? Maybe, but maybe not. There are problems with using old data that were collected under unknown circumstances, and the participants in older studies were probably not representative of the entire population. Nonetheless, the finding is provocative because it suggests that modern life may be an even more powerful cognitive enhancer than we realize.

Education

Alfred Binet believed that if poverty was intelligence’s enemy, then education was its friend. And he was right about that too. The correlation between the amount of formal education a person receives and his or her intelligence is quite large, somewhere in the range of r = .55 to .90 (Ceci, 1991; Neisser et al., 1996). One reason why this correlation is so large is that smart people tend to stay in school, but the other reason is that school makes people smarter (Ceci & Williams, 1997). When schooling is delayed because of war, political strife, or the simple lack of qualified teachers, children show a measurable decline in intelligence (Nisbett, 2009). Indeed, children born in the first 9 months of a calendar year typically start school an entire year earlier than those born in the last 3 months of the same year, and sure enough, people with late birthdays tend to have lower intelligence test scores than people with early birthdays (Baltes & Reinert, 1969).

Education increases intelligence. But not everyone is in favor of that. In Afghanistan, for example, the Taliban attack young girls with acid, guns, and poison to keep them from attending school.

© Horizons WWP/Alamy

Does this mean that anyone can become a genius just by showing up for class? Unfortunately not. Although education reliably increases intelligence, its impact is small, and some studies suggest that it tends to enhance test-taking ability more than general cognitive ability and that its effects vanish within a few years (Perkins & Grotzer, 1997). In other words, education seems to produce increases in intelligence that are smaller, narrower, and shorter-lived than we would like. That might mean that education just can’t change intelligence all that much, or it might mean that education is potentially very powerful, but that modern schools aren’t very good at providing it. Researchers lean toward the latter conclusion. Although most experiments in education– from magnet schools and charter schools to voucher systems and Head Start programs–have failed to produce substantial intellectual gains for students, a few have been quite successful (Nisbett, 2009), which shows that education can substantially increase intelligence even if it doesn’t usually do so. No one knows just how big the impact of an optimal education could be, but it seems clear that our current educational system is less than optimal.

Genes and Environments

Both genes and environment influence intelligence. But that shouldn’t lead you to think of genes and environments as two separate ingredients that are somehow blended together like flour and sugar in a recipe for IQ. The fact is that genes and environments interact in complex ways that make the distinction between them a bit murky.

Roundworms that have a gene called NPR-1 dislike low-oxygen environments, which just so happen to be where the bacteria hang out. As a result, these roundworms get fewer infections. Is NPR-1 a health gene?

© MPG/epa/Corbis

For example, imagine a gene that made people enjoy the smell of library dust or that made them unusually sensitive to the glare produced by television sets. People who had such a gene might well read more books and thus end up being smarter. Would their increased intelligence be due to their genes or to their environments? Well, if they hadn’t had those genes, then they wouldn’t have gone to the library, but if they hadn’t gone to the library, then they wouldn’t have gotten smarter. The fact is that genes can exert some of their most powerful influences not by changing the structure of a person’s brain, but by changing the person’s environment (Dickens & Flynn, 2001; Nisbett, 2009; Plomin, DeFries, et al., 2001). A gene that made someone sociable might lead her to have good relationships with peers, which might lead her to stay in school longer, which might lead her to become smarter. Would we call such a gene a “sociability gene” or an “intelligence gene” (Posthuma & de Geus, 2006)? Would we attribute that person’s intelligence to her genes or to the environment that her genes enabled her to create? As these questions suggest, genes and environments are not independent influences on intelligence, and the clear difference between nature and nurture is not as clear as it might first appear.