Parenting in the poor neighborhood involved “hard defensive individualism.” The neighborhood was rife with addiction, homelessness, incarceration, death—and parents’ aim was to shelter their child from the literal and metaphorical street. Their speech was full of metaphors about not losing what was achieved—standing your ground, keeping up your pride, not letting others get under your skin. Parenting was authoritarian, toughening the goal. For example, parents teased kids far more than in the other neighborhoods.
In contrast, working-class parenting involved “hard offensive individualism.” Parents had some socioeconomic momentum, and kids were meant to maintain that precarious trajectory. Parents’ speech about their hopes for their kids contained images of movement, progress, and athletics—getting ahead, testing the waters, going for the gold. With hard work and the impetus of generations of expectations, your child might pioneer landfall in the middle class.
Parenting in both neighborhoods emphasized respect for authority, particularly within the family. Moreover, kids were fungible members of a category, rather than individualized—“You kids get over here.”
Then there was the “soft individualism” of upper-middle-class parenting.* Children’s eventual success, by conventional standards, was a given, as were expectations of physical health. Far more vulnerable was a child’s psychological health; when children could become anything, parents’ responsibility was to facilitate their epic journey toward an individuated “fulfillment.” Moreover, the image of fulfillment was often postconventional—“I hope my child will never work an unsatisfying job just for money.” This, after all, is a tribe giddied by tales of the shark in line to become CEO chucking it to learn carpentry or oboe. Parents’ speech brimmed with metaphors of potential being fulfilled—flowering, blooming, growing, blossoming. Parenting was authoritative or permissive, riddled with ambivalence about parent-child power differentials. Rather than “You kids, clean up this mess,” there’d be the individuated, justifying request—“Caitlin, Zach, Dakota, could you clean things up a bit please? Malala is coming for dinner.”*
We’ve now seen how childhood events—from the first mother-infant interaction to the effects of culture—have persistent influences, and how biology mediates such influences. When combined with the preceding chapters, we have finished our tour of environmental effects on behavior, from the second before a behavior occurs to a second after birth. In effect, we’ve done “environment”; time for next chapter’s “genes.”
But this ignores something crucial: environment doesn’t begin at birth.
NINE LONG MONTHS
The Cat in the Hat in the Womb
The existence of prenatal environmental influences caught the public’s imagination with some charming studies demonstrating that near-term fetuses hear (what’s going on outside the womb), taste (amniotic fluid), and remember and prefer those stimuli after birth.
This was shown experimentally—inject lemon-flavored saline into a pregnant rat’s amniotic fluid, and her pups are born preferring that flavor. Moreover, some spices consumed by pregnant women get into amniotic fluid. Thus we may be born preferring foods our mothers ate during pregnancy—pretty unorthodox cultural transmission.59
Prenatal effects can also be auditory, as shown by inspired research by Anthony DeCasper of the University of North Carolina.60 A pregnant woman’s voice is audible in the womb, and newborns recognize and prefer the sound of their mother’s voice.* DeCasper used ethology’s playbook to show this: A newborn can learn to suck a pacifier in two different patterns of long and short sucks. Generate one pattern, and you hear Mom’s voice; the other, another woman’s voice. Newborns want Mom’s voice. Elements of language are also learned in utero—the contours of a newborn’s cry are similar to the contours of speech in the mother’s language.
The cognitive capacities of near-term fetuses are even more remarkable. For example, fetuses can distinguish between two pairs of nonsense syllables (“biba” versus “babi”). How do you know? Get this—Mom says “Biba, biba, biba” repeatedly while fetal heart rate is monitored. “Boring (or perhaps lulling),” thinks the fetus, and heart rate slows. Then Mom switches to “babi.” If the fetus doesn’t distinguish between the two, heart rate deceleration continues. But if the difference is noted—“Whoa, what happened?”—heart rate increases. Which is what DeCasper reported.61
DeCasper and colleague Melanie Spence then showed (using the pacifier-sucking-pattern detection system) that newborns typically don’t distinguish between the sounds of their mother reading a passage from The Cat in the Hat and from the rhythmically similar The King, the Mice, and the Cheese.62 But newborns whose mothers had read The Cat in the Hat out loud for hours during the last trimester preferred Dr. Seuss. Wow.
Despite the charm of these findings, this book’s concerns aren’t rooted in such prenatal learning—few infants are born with a preference for passages from, say, Mein Kampf. However, other prenatal environmental effects are quite consequential.
BOY AND GIRL BRAINS, WHATEVER THAT MIGHT MEAN
We start with a simple version of what “environment” means for a fetal brain: the nutrients, immune messengers, and, most important, hormones carried to the brain in the fetal circulation.
Once the pertinent glands have developed in a fetus, they are perfectly capable of secreting their characteristic hormones. This is particularly consequential. When hormones first made their entrance in chapter 4, our discussion concerned their “activational” effects that lasted on the order of hours to days. In contrast, hormones in the fetus have “organizational” effects on the brain, causing lifelong changes in structure and function.
Around eight weeks postconception, human fetal gonads start secreting their steroid hormones (testosterone in males; estrogen and progesterone in females). Crucially, testosterone plus “anti-Müllerian hormone” (also from the testes) masculinize the brain.
Three complications, of increasing messiness:
In many rodents the brain isn’t quite sexually differentiated at birth, and these hormonal effects continue postnatally.
A messier complication: Surprisingly few testosterone effects in the brain result from the hormone binding to androgen receptors. Instead, testosterone enters targets cells and, bizarrely, is converted to estrogen, then binds to intracellular estrogen receptors (while testosterone has its effects outside the brain either as itself or, after intracellular conversion to a related androgen, dihydrotestosterone). Thus testosterone has much of its masculinizing effect in the brain by becoming estrogen. The conversion of testosterone to estrogen also occurs in the fetal brain. Wait. Regardless of fetal sex, fetal circulation is full of maternal estrogen, plus female fetuses secrete estrogen. Thus female fetal brains are bathed in estrogen. Why doesn’t that masculinize the female fetal brain? Most likely it’s because fetuses make something called alpha-fetoprotein, which binds circulating estrogen, taking it out of action. So neither Mom’s estrogen nor fetal-derived estrogen masculinizes the brain in female fetuses. And it turns out that unless there is testosterone and anti-Müllerian hormone around, fetal mammalian brains automatically feminize.63
Now for the übermessy complication. What exactly is a “female” or “male” brain? This is where the arguments begin.
To start, male brains merely consistently drool reproductive hormones out of the hypothalamus, whereas female brains must master the cyclic secretion of ovulatory cycles. Thus fetal life produces a hypothalamus that is more complexly wired in females.
But how about sex differences in the behaviors that interest us? The question is, how much of male aggression is due to prenatal masculinizing of the brain?
Virtually all of it, if we’re talking rodents. Work in the 1950s by Robert Goy of the University of Wisconsin showed that in guinea pigs an organizational effect of perinatal testosterone is to make the brain responsive to testosterone in adulthood.64 Near-term pregnant fema
les would be treated with testosterone. This produced female offspring who, as adults, appeared normal but were behaviorally “masculinized”—they were more sensitive than control females to an injection of testosterone, with a greater increase in aggression and male-typical sexual behavior (i.e., mounting other females). Moreover, estrogen was less effective at eliciting female-typical sexual behavior (i.e., a back-arching reflex called lordosis). Thus prenatal testosterone exposure had masculinizing organizational effects, so that these females as adults responded to the activational effects of testosterone and estrogen as males would.
This challenged dogma that sexual identity is due to social, not biological, influences. This was the view of sociologists who hated high school biology . . . and of the medical establishment as well. According to this view, if an infant was born with sexually ambiguous genitalia (roughly 1 to 2 percent of births), it didn’t matter which gender they were raised, as long as you decided within the first eighteen months—just do whichever reconstructive surgery was more convenient.*65
So here’s Goy reporting that prenatal hormone environment, not social factors, determines adult sex-typical behaviors. “But these are guinea pigs” was the retort. Goy and crew then studied nonhuman primates.
A quick tour of sexually dimorphic (i.e., differing by sex) primate behavior: South American species such as marmosets and tamarins, who form pair-bonds, show few sex differences in behavior. In contrast, most Old World primates are highly dimorphic; males are more aggressive, and females spend more time at affiliative behaviors (e.g., social grooming, interacting with infants). How’s this for a sex difference: in one study, adult male rhesus monkeys were far more interested in playing with “masculine” human toys (e.g., wheeled toys) than “feminine” ones (stuffed animals), while females had a slight preference for feminine.66
What next, female monkeys prefer young-adult fantasy novels with female protagonists? Why should human toys be relevant to sex differences in monkeys? The authors speculate that this reflects the higher activity levels in males, and how masculine toys facilitate more active play.
Male rhesus monkeys show a strong preference for playing with stereotypically “masculine” versus “feminine” human toys.
Visit bit.ly/2o8ogEL for a larger version of this graph.
Goy studied highly sexually dimorphic rhesus monkeys. There were already hints that testosterone has organizational effects on their behavior—within weeks of birth, males are more active than females and spend more time in rough-and-tumble play. This is long before puberty and its burst of testosterone secretion. Furthermore, even if you suppress their testosterone levels at birth (low, but nevertheless still higher than those of females), males still do more roughing and tumbling. This suggested that the sex difference arose from fetal hormone differences.
Goy proved this by treating pregnant monkeys with testosterone and examining their female offspring. Testosterone exposure throughout pregnancy produced daughters who were “pseudohermaphrodites”—looked like males on the outside but had female gonads on the inside. When compared with control females, these androgenized females did more rough-and-tumble play, were more aggressive, and displayed male-typical mounting behavior and vocalizations (as much as males, by some measures). Importantly, most but not all behaviors were masculinized, and these androgenized females were as interested as control females in infants. Thus, testosterone has prenatal organizational effects on some but not all behaviors.
In further studies, many carried out by Goy’s student Kim Wallen of Emory University, pregnant females received lower doses of testosterone, and only in the last trimester.67 This produced daughters with normal genitalia but masculinized behavior. The authors noted the relevance of this to transgender individuals—the external appearance of one sex but the brain, if you will, of the other.*
And Us
Initially it seemed clear that prenatal testosterone exposure is also responsible for male aggression in humans. This was based on studies of a rare disorder, congenital adrenal hyperplasia (CAH). An enzyme in the adrenal glands has a mutation, and instead of making glucocorticoids, they make testosterone and other androgens, starting during fetal life.
The lack of glucocorticoids causes serious metabolic problems requiring replacement hormones. And what about the excessive androgens in CAH girls (who are typically born with ambiguous genitals and are infertile as adults)?
In the 1950s psychologist John Money of Johns Hopkins University reported that CAH girls had pathologically high levels of male-typical behaviors, a paucity of female-typical ones, and elevated IQ.
That sure stopped everyone in their tracks. But the research had some problems. First, the IQ finding was spurious—parents willing to enroll their CAH child in these studies averaged higher levels of education than did controls. And the gender-typical behaviors? “Normal” was judged by 1950s Ozzie and Harriet standards—CAH girls were pathologically interested in having careers and disinterested in having babies.
Oops, back to the drawing board. Careful contemporary CAH research has been conducted by Melissa Hines of the University of Cambridge.68 When compared with non-CAH girls, CAH girls do more rough-and-tumble play, fighting, and physical aggression. Moreover, they prefer “masculine” toys over dolls. As adults they score lower on measures of tenderness and higher in aggressiveness and self-report more aggression and less interest in infants. In addition, CAH women are more likely to be gay or bisexual or have a transgender sexual identity.*
Importantly, drug treatments begun soon after birth normalize androgen levels in these girls, so that the excessive androgen exposure is solely prenatal. Thus prenatal testosterone exposure appears to cause organizational changes that increase the incidence of male-typical behaviors.
A similar conclusion is reached by an inverse of CAH, namely androgen insensitivity syndrome (AIS, historically called “testicular feminization syndrome”).69 A fetus is male—XY chromosomes, testes that secrete testosterone. But a mutation in the androgen receptor makes it insensitive to testosterone. Thus the testes can secrete testosterone till the cows come home but there won’t be any masculinization. And often the individual is born with a female external phenotype and is raised as a girl. Along comes puberty, she’s not getting periods, and a trip to the doctor reveals that the “girl” is actually a “boy” (with testes typically near the stomach, plus a shortened vagina that dead-ends). The individual usually continues with a female identity but is infertile as an adult. In other words, when human males don’t experience the organizational prenatal effects of testosterone, you get female-typical behaviors and identification.
Between CAH and AIS, the issue seems settled—prenatal testosterone plays a major role in explaining sex differences in aggression and various affiliative prosocial behaviors in humans.
Careful readers may have spotted two whopping big problems with this conclusion:70
Remember that CAH girls are born with a “something’s very different” Post-it—the ambiguous genitalia, typically requiring multiple reconstructive surgeries. CAH females are not merely prenatally androgenized. They’re also raised by parents who know something is different, have slews of doctors mighty interested in their privates, and are treated with all sorts of hormones. It’s impossible to attribute the behavioral profile solely to the prenatal androgens.
Testosterone has no effects in AIS individuals because of the androgen receptor mutation. But doesn’t testosterone have most of its fetal brain effects as estrogen, interacting with the estrogen receptor? That aspect of brain masculinization should have occurred despite the mutation. Complicating things, some of the masculinizing effects of prenatal testosterone in monkeys don’t require conversion to estrogen. So we have genetically and gonadally male individuals with at least some brain masculinization raised successfully as females.
The picture is complicated further—AIS individuals raised female have higher-than-exp
ected rates of being gay, and of having an other-than-female or neither-female-nor-male-sex/gender self-identification.
Argh. All we can say is that there is (imperfect) evidence that testosterone has masculinizing prenatal effects in humans, as in other primates. The question becomes how big these effects are.
Answering that question would be easy if you knew how much testosterone people were exposed to as fetuses. Which brings up a truly quirky finding, one likely to cause readers to start futzing awkwardly with a ruler.
Weirdly, prenatal testosterone exposure influences digit length.71 Specifically, while the second finger is usually shorter than the fourth finger, the difference (the “2D:4D ratio”) is greater in men than in women, something first noted in the 1880s. The difference is demonstrable in third-trimester fetuses, and the more fetal testosterone exposure (as assessed by amniocentesis), the more pronounced the ratio. Moreover, CAH females have a more masculine ratio, as do females who shared their fetal environment (and thus some testosterone) with a male twin, while AIS males have a more feminine ratio. The sex difference in the ratio occurs in other primates and rodents. And no one knows why this difference exists. Moreover, this oddity is not alone. A barely discernible background noise generated by the inner ear (“otoacoustic emissions”) shows a sex difference that reflects prenatal testosterone exposure. Go explain that.
The 2D:4D ratio is so variable, and the sex difference so small, that you can’t determine someone’s sex by knowing it. But it does tell you something about the extent of fetal testosterone exposure.
Behave: The Biology of Humans at Our Best and Worst Page 22