FIVE

Who Do You Love?

WHEN I WAS IN HIGH SCHOOL DURING THE 1970S, a popular riddle about a well-liked member of our social group went like this:

Q: Why is Jane bisexual?

A: Because there are only two sexes.

The implication, of course, is that Jane is democratic in her affections and if there were, say, three sexes, then she would be trisexual.¹ I bring up this old joke to highlight that what we have come to call sexual orientation—categorized as gay, straight, or bi—is a crude instrument. If a man had a single early sexual experience with another man, but has since been only romantically and sexually involved with women, do we call him straight or bi? How would he categorize himself, if at all? In 1997, the actress Anne Heche, who had previously only been romantically involved with men, embarked on a well-publicized relationship with the openly lesbian comedian Ellen DeGeneres. That relationship ended after two years, and Heche went on to marry a man. Does that make her presently straight or bi? What about a woman who has been exclusively sexually involved with men but likes to watch feminist lesbian porn? Do these categories even matter?

Is the transgender man who fancies women straight, as befits his gender identity, or gay, as would be indicated by his sex at birth? I’d say straight, but others might argue. To get around this problem, some researchers have employed the terms androphilic (man loving), gynephilic (woman loving), and ambiphilic (woman and man loving) to refer to the object of one’s desires without regard for the biological sex or gender identity of the desirer. This highlights that people are not attracted to “those who share my gender identity” or “those who don’t share my gender identity”—they are attracted to men or women or both. When people change their gender identity, they almost never simultaneously experience a change in whom they are attracted to, which would serve to maintain a constant sexual orientation of gay or straight.

For most people, their romantic and sexual interests are aligned, but not for everyone and not always. For example, some men in prison will resort to sex with other men for release but would deny any romantic attraction. And lots of us have sexual fantasies that we have no intention or even desire to act on, no matter how much delight they may provide in the realm of the imagination. Some people experience neither romantic nor sexual attraction, or one without the other. It is legitimate to ask if a terminology of sexual orientation should reflect fantasy, desire, overt behavior, or some combination thereof. The answers to these questions are not obvious or straightforward and have been the subject of much debate.

IN THE SOUTHERN LOWLANDS of Papua New Guinea and some neighboring islands, there is a traditional belief that there exist two essential bodily fluids: breast milk and semen. The people who live there hold that while all infants require breast milk to grow, boys cannot become men without consuming semen from adult males. Around age ten, boys are removed from their mothers and placed in a separate boys’ house at some distance from the village, where they live for months to years. During this enforced segregation, it is the duty of men, particularly maternal uncles, to deliver their semen to boys in order to help them grow through puberty. In some groups, this is accomplished by the man receiving oral sex from the boy, in others by rubbing the semen over the boy’s body, and yet others by the boy receiving anal intercourse. This sexual initiation is obligatory and it must always involve the boy receiving semen from the man (the other way around would impede the boy’s growth and maturation).

After a boy’s initiation ceremony to manhood, he is expected to have a transition period where he is sexually involved with both sexes, but after a few years, he is encouraged to marry a woman and eventually have sex with her exclusively.² This example shows how deeply held cultural ideas can profoundly influence sexual behavior. People in the villages that maintain this sexual practice have no words for or concept of gay, straight, or bi. Describing the development of young males as progressing from gay to bi to straight is a construct that we can impose from outside their culture, but it would make no sense to them, and hence its utility and validity are minimal.

THESE DAYS, SOME PEOPLE are offended by the lack of subtlety inherent in the traditional sexual orientation categories of straight, gay, and bi. They may adopt newer terms like pansexual (not limited in sexual choice with regard to biological sex, gender, or gender identity), demisexual (sexual attraction only to people with whom they have an emotional bond), or heteroflexible (minimal homosexual activity in an otherwise primarily heterosexual orientation), or even refuse any single-word descriptor of their sexual or romantic feelings. Psychologist Sari van Anders has proposed to replace sexual orientation with a matrix theory, with identity/orientation/status in one dimension and gender/sex/partner number in another.³ This is a useful endeavor and should be applauded for its accuracy and inclusion, but, in my view, it’s a bit unwieldy for everyday conversation. Here, I’m going to use the terms straight, gay, and bi, with the understanding that they are imperfect tools that do not entirely capture the subtlety, range, and dynamism of human sexual expression.

There have now been several large-scale anonymous surveys of sexual orientation performed using random sampling in the United States and Europe. They indicate that approximately 3 percent of men and 1 percent of women identify as consistently homosexual, about 0.5 percent of men and 1 percent of women as bisexual, and the remainder heterosexual. While the results from these surveys have been fairly consistent, the caveat remains that people do not always answer surveys honestly and so there may be some systematic error in the results. In addition, these surveys have mostly been conducted in higher-income, Christian-majority countries, so the results might be somewhat different in other populations.⁴

ASK A STRAIGHT MAN, “When did you decide to be straight?” and he’ll likely answer that it didn’t feel like a decision at all, but rather like a deep compulsion that became evident around puberty or before. Gay and bi men will give the same answer: in one survey in the United States, only about 4 percent of all gay and bi men reported that they chose their sexual orientation—the rest felt that they were “born that way.”⁵ In 2019, presidential candidate Pete Buttigieg reinforced this point: “If me being gay was a choice, it was a choice that was made far, far above my pay grade. And that’s the thing I wish the Mike Pences of the world would understand—that if you got a problem with who I am, your problem is not with me. Your quarrel, sir, is with my Creator.”⁶

Not surprisingly, the scientific question “Is sexual orientation immutable?” has become a political issue. Many on the political right see homosexuality as a harmful and immoral choice made of free will and hence, in their view, undeserving of civil rights protection. Conversely, most people on the political left have advocated for the rights of gay and bi people on the basis that sexual orientation is an innate and immutable trait, hence requiring civil rights protection. Indeed, in Obergefell v. Hodges, the landmark US Supreme Court decision that established a constitutional right to same-sex marriage, the majority wrote: “Psychiatrists and others [have] recognized that sexual orientation is both a normal expression of human sexuality and immutable.” The court went on to declare that, because sexual orientation is fixed, gay people are compelled to enter into committed same-sex relationships: “Their immutable nature dictates that same-sex marriage is their only real path to this profound commitment.”⁷

While men’s self-reports would suggest that sexual orientation is nearly always set early in life and remains constant through adulthood, the situation with women is not so clear. Although many lesbians report that they felt solely and consistently attracted to women starting at an early age, a significant fraction have had a more fluid experience of sexual and romantic interest (like Anne Heche did with Ellen DeGeneres). When psychologist Lisa Diamond interviewed seventy-nine lesbian and bisexual women over a ten-year period, she discovered that about two-thirds of the women changed their stated sexual orientation and about one-third changed two or more times. Importantly, and contrary to some widespread cultural ideas, bisexuality was rarely a transition stage on a journey from straight to lesbian. Diamond found that, in many cases, women who had only experienced attraction to men suddenly found themselves falling in love with and being sexually attracted to a woman, or vice versa.⁸ In many cases, their atypical attraction was specific to a particular person. For example, a self-identified lesbian who found herself attracted to one particular man might not then become attracted to men generally. Diamond calls this malleability of attraction “female sexual fluidity.”⁹

The existence of sexual fluidity complicates the case for gay and lesbian civil rights that has been made to date. Certainly it questions the immutability argument that forms a central pillar of the Obergefell decision on gay marriage. However, I would hold, in agreement with Lisa Diamond and legal scholar Clifford Rosky, that immutability should never have been central to the argument for lesbian and gay civil rights in the first place. People who have experienced their sexual orientation as fixed—whether they are gay, straight, or bi—should not be privileged in their civil rights over those who have experienced their sexual orientation as fluid. The fundamental moral argument for gay and lesbian civil rights should be one of individual liberty, not of immutability. By comparison, in the United States we already have laws in place to protect against religious discrimination, even though the practice of a particular religion is clearly not an immutable trait. And we protect the rights of people who change their religion. Sexual orientation is sometimes fluid, more often in women than men, but that is not a valid rationale for denying civil rights to sexual minorities.

THERE’S A DEEPLY HELD idea that the body doesn’t lie. You may seek to project a cool and collected exterior in a stressful situation, but your sweaty armpits and racing heart will give you away. There’s a similar idea about men and penile erections. I, as a straight guy, may protest that I’m not sexually excited by watching women’s Olympic beach volleyball on TV, but when my wife points to the tent in my pants and says, “You’re busted!” I must agree, sheepishly.

In the lab, this embarrassment can be studied with a bit more rigor. Cisgender men’s erections are measured using a device called a circumferential penile plethysmograph, which is a fancy term for a broad rubber band with a strain gauge inside. As the penis becomes erect and increases in girth, the strain gauge is activated and the signal is sent to a recording device. Volunteers are fitted with this gizmo and are then asked to watch a video screen and simultaneously report their degree of sexual arousal using a computer mouse. The erotic stimuli are various porn films. As a control, videos of nature or sports are used. One might imagine that the ideal porn film to measure heterosexuality in men would involve heterosexual couples. The problem is that, by definition, such a video will feature sexual activity involving both a man and a woman. Experiments have shown that both exclusively gay and exclusively straight men will have erections in response to hetero porn. By contrast, most straight men neither self-report arousal nor do they become erect to gay male porn videos. Similarly, most gay men neither report arousal nor do they become erect to videos depicting sex between two women. More recently, it has been shown that at least some bisexual men respond to male-male and female-female sex videos with both erections and self-reported arousal.¹⁰ The common finding here is that men—bi, gay, or straight—have strong concordance between their self-reports and their erections. Generally speaking, if men say that something excites them sexually, they will become erect in response to it; if they say that something does not, they won’t.¹¹

The situation for cisgender women is different. There are several ways to measure sexual responses of the female genitals. The one that has been used most widely is called vaginal photoplethysmography. It uses a tampon-size probe containing a light source and a photocell that measures the color of the light reflected from the vaginal wall. The idea here is that female genital arousal includes the production and secretion of natural vaginal lubricant, which is derived from blood plasma. When women are aroused, increased blood flow to the vaginal wall will change its color as a prelude to the production of vaginal lubricant, and this can be detected by the probe. Female sexual arousal is also accompanied by increased blood flow to the external portion of the genitals, and this can be measured using a technique called vulvar laser speckle imaging or another called vulvar thermal imaging.

In some ways, women’s responses are like men’s: stimuli that women report to be arousing also caused a vaginal response. The difference is that most women also respond vaginally to at least some sex videos that they report as not arousing. Most straight women show vaginal responses to male-male, female-female, and male-female sex videos, even if they self-report arousal to only some of those stimuli. In one study, most women even had vaginal responses to videos of two bonobos (pygmy chimpanzees) having sex, even though nearly all of them reported not being aroused by such activity.

Overall, women’s genital responses and their reported arousal are less concordant than those of men. However, there are two interesting details. First, gay and bi women are more sexually concordant than straight women. In other words, those gay and bi women who report no arousal to male-male or hetero sex videos are less likely to have a vaginal response to them.¹² Second, for all women, the vulvar blood flow measures are a better match to self-reported arousal than vaginal blood flow, with less vulvar blood flow in response to stimuli reported as non-arousing.¹³

Why would most women have vaginal wall blood flow responses to stimuli that they report as non-arousing? One possibility is that women’s self-reports are unreliable and that they are actually mentally aroused by sexual situations that they deny.¹⁴ I think that this explanation is unlikely. Keep in mind that the subjects for these experiments are volunteers who know that they will be watching sex videos with a probe inserted in the vagina, and so are self-selected to be sex positive. These women are unlikely to be of the inhibited type who would deny stimuli that they found arousing. A more likely possibility is that, because vaginal wall blood flow produces lubrication, it is an adaptive response to situations where vaginal penetration is rapid or nonconsensual (which may have been somewhat more prevalent in human evolutionary history than it is now). Sex researcher Meredith Chivers has suggested that reflexive vaginal lubrication triggered by a wide array of sexual stimuli would reduce the chance of pain, injury, or infection.¹⁵ This explanation is consistent with the observation that vaginal responses are less concordant with reported arousal than vulvar responses, as blood flow to the vulva is unlikely to be as protective.

It’s tempting to draw a connection between women’s sexual fluidity and women’s discordant vaginal responses. Both phenomena suggest that, on average, women are more open to sexual experiences that run counter to their stated sexual orientation than men are. That said, this may be a case of “true, true, unrelated.” The reduced concordance between cognitive and vaginal responses in women, particularly women who identify as straight, and the increased fluidity of women’s sexual orientation as compared to men may spring from a common neural origin or evolutionary adaptation, but at present there’s no evidence either for or against that hypothesis.

MIGHT SEXUAL ORIENTATION BE determined by early social experience, even before overt sexual feelings emerge?¹⁶ Children raised by single mothers are no more or less likely to be straight than those raised by straight couples. Likewise, being raised by a lesbian couple does not increase or decrease one’s chance of being straight.¹⁷ Remarkably, a major meta-analysis of the scientific literature by the American Psychological Association failed to find clear evidence that any practices of child-rearing—from religion to discipline to education—influence adult sexual orientation.¹⁸ In this way, sexual orientation is like most personality traits: there is some influence of child-rearing (reflected in the shared environment statistic) on behavior early in life, but little influence remains by early adulthood. If you are raised in a household or community that stigmatizes homosexuality, then you may be less likely to come out as gay or bi, but it doesn’t mean that you won’t feel same-sex attractions.

Some researchers have claimed that both male and female homosexuality can result from childhood emotional or sexual abuse. My reading of the contentious literature on the topic has led me to believe that evidence for this relationship is flimsy. First, there are arguments about the facts. Some surveys have found a positive association between emotional or sexual abuse and later same-sex attraction. But others have failed to find any statistical relationship, or have found an association in women but not in men, or for sexual abuse but not emotional abuse.¹⁹ When there is an association, it tends to be statistically weak. For example, one study found that a history of sexual abuse increased the prevalence of adult same-sex partners, but only by 1.4 percent. Second, even if we were to assume that there are small statistical associations between childhood abuse and adult homosexuality, that doesn’t mean that the former caused the latter. One possibility is that reported differences in childhood maltreatment result not from a higher incidence of abuse but rather from a higher incidence of recall among gay men and lesbians. To my thinking, the most likely explanation for this small association is that nascent homosexuality, as expressed by less gender-typical behavior in childhood, slightly increases the probability of child abuse.

Sigmund Freud famously posited that male homosexuality was caused by distant fathers and close-binding mothers (he didn’t have so much to say about female homosexuality). One can critique Freud’s conclusion by pointing out that his patients were not a representative sample of gay men; rather, they were a subset sufficiently troubled to have sought psychotherapy. That’s true, but there is evidence from broader surveys that, on average, gay men report stronger childhood bonds with their mothers and weaker bonds with their fathers than straight men.²⁰ But just like the purported association with childhood sexual abuse, the devil is in the details. It’s probably not that child-rearing strongly influences sexual orientation. Rather, variation in brain circuits that influence gender-typical behaviors, first evident in young children, affect how parents and other adults respond to them. Let’s examine how that might come to pass.

If you’re a man with a gay brother, your chance of being gay is about 22 percent (versus 3 percent in the general population). If you’re a woman with a lesbian sister, your chance of being a lesbian is about 16 percent (compared to 1 percent in the general population).²¹ However, there are no such statistical links for opposite-sex siblings. If a woman has a gay brother, that situation is not associated with an increased probability of her being attracted to women. Similarly, if a man has a lesbian sister, that does not increase his likelihood of being attracted to men. There is no familial co-occurrence for gayness or straightness that binds brothers and sisters. The relevant traits here are not “attracted to the same sex as me” or “attracted to the opposite sex as me,” but rather “attracted to women” or “attracted to men.”

These statistics tell us that sexual orientation clusters in families but do not tell us why. For example, two brothers share, on average, 50 percent of their gene variants with each other, but they also tend to share a similar upbringing. So, for example, if Freud was right and a close-binding mother causes a boy to become gay, then this effect could be found in brothers raised together. As we’ve discussed, one way to disentangle genetics from upbringing is to analyze same-sex twins. If sexual orientation had no heritable component, we’d expect that the percentage of twin pairs where both are gay would be roughly the same for fraternal and identical twins. Conversely, if sexual orientation were entirely determined by genes, then every gay identical twin would have a gay twin sibling. The best estimates to date, from a population of 3,826 randomly selected twin pairs in Sweden, indicate that about 20 percent of the variance in sexual orientation in women is determined by genes; in men, it’s about 40 percent.²² Some previous twin studies had produced higher estimates of heritability for both male and female sexual orientation—around 50 percent for both—but those studies used self-selected volunteers (in part through advertisements in the lesbian and gay press and at gay pride events) rather than randomly chosen twins, and it is likely that the design biased the results.

The conclusion is that genetic variation is one factor in determining sexual orientation, but it is far from the whole story and it is a somewhat stronger effect in men than in women. As always, it’s important to remind ourselves that these estimates of heritability are measures for populations, not individuals. It may be that there are some individual women and men who carry gene variants such that their sexual orientation is entirely genetically determined, and others for whom their sexual orientation has no genetic contribution whatsoever. As with all behavioral traits, there is no single gene that determines human sexual orientation. Rather, many different genes each appear to make a small contribution, and at present we do not have a useful list of these genes.²³ The lack of a single gene for sexual orientation in humans is not an argument against a heritable contribution to variation in this trait. Recall that height is a very heritable trait, but its heritability is determined by small variations in hundreds of genes.

IF CHILD-REARING HAS LITTLE or no effect on sexual orientation, and gene variants have only a partial effect, then why are some people straight while others are bi or gay? What accounts for the missing variation? For cisgender men, fraternal birth order appears to be one contributing factor. What this means is that, if you are male, having older brothers increases your chances of being attracted to males as an adult. This effect, though small, has been found in many different cultures and geographic regions. Having older sisters, younger sisters, or younger brothers has no effect. This does not appear to be a result of child-rearing, as having a biological older brother increases the chance of same-sex attraction in men even if that brother was raised in a different household. Likewise, having an older adopted brother has no influence.²⁴

Interestingly, it has been found that men with older brothers have lower-than-expected birth weights, suggesting that the fraternal birth order effect may operate prenatally. One potential biological explanation for the fraternal birth order effects on both sexual orientation and birth weight is that they reflect a progressive maternal immune response directed against foreign male proteins (presumably those encoded by the Y chromosome). In this view, male proteins (or perhaps entire cells) cross into the maternal circulation, are recognized as foreign, and direct the production of maternal antibodies. During the next pregnancy with a male fetus, these antibodies cross through the placenta and influence development of the fetal body and brain. One fascinating recent study has found that mothers of gay sons, particularly those with older brothers, had significantly higher levels of antibodies directed against a protein called neuroligin-4 Y-linked than did the control samples of women, including mothers of heterosexual sons.²⁵ This potentially exciting finding awaits replication.

Another promising explanation involves exposure to hormones in utero and early postnatal life. In this view, when female fetuses and babies are exposed to higher levels of testosterone, their brains are partially masculinized and this increases the chance that, later in life, they will become sexually attracted to women. Similarly, when male fetuses and babies are exposed to lower levels of testosterone, their brains become partially feminized, thereby increasing the chance that they will eventually become sexually attracted to men. There is evidence in support of this idea from human conditions that alter steroid hormone signaling. For example, as we have discussed, females with a condition called congenital adrenal hyperplasia have increased testosterone levels in fetal life. Even when these girls are treated with testosterone-blocking drugs starting at birth, their brains appear to have been partially masculinized. About 21 percent of women with congenital adrenal hyperplasia report consistent sexual attraction to women (compared to 1.5 percent in the general female population).²⁶

This finding is congruent with some experiments in laboratory animals: when guinea pigs, rats, or sheep receive treatments that boost fetal testosterone signaling, the females grow up to display male-typical sexual behaviors. That is, they mount females and fail to exhibit a posture called lordosis, which encourages males to mount them. Similarly, treatments that attenuate testosterone signaling in developing males reduce male-typical sexual behavior in adult rats and sheep. The caveat to these observations is that we don’t always know the meaning of the behaviors. When a female sheep mounts another female, is that expressing sexual interest in females, social aggression, or both? Similarly, when a male rat displays lordosis, is that a sign of sexual interest in males, social submission, or both? Or perhaps these behaviors have a meaning that doesn’t even occur to us humans.

Can we use average differences in the brain structure of straight men and straight women to test the hypothesis that the brains of gay men are more likely to be partly feminized and the brains of lesbians are more likely to be partly masculinized? The relevant parts of the brain include a portion of the hypothalamus called INAH3, which is larger in straight men, and a bundle of fibers that connects one side of the brain to the other, the anterior commissure, which is larger in straight women. While there have been some well-publicized reports suggesting that the size of the anterior commissure and INAH3 are more female-like in gay men,²⁷ there have yet to be clear, independent replications of these findings.²⁸ This doesn’t mean that there aren’t important differences in the brains of straight versus gay women and men, but, just as we discussed for differences between the sexes, most of the variation is unlikely to manifest as changes in the size of brain regions—the sort of measurement that can be performed with a brain-scanning machine or from autopsy tissue. As a hypothetical example, when comparing lesbians with straight women, lesbians might have lower expression of a gene coding for a voltage-sensitive potassium channel in certain neurons that help to direct sexual behavior toward females. This would render these neurons more electrically excitable but would not change their shape or the volume of the brain region that contains them.²⁹

From the earliest portion of childhood, there are, on average, some behavioral differences between girls and boys. As we’ve discussed, boys are more likely to engage in rough-and-tumble play and to interact with inanimate object toys, while, on average, girls are more likely to play less aggressively and to choose dolls and animals for their toys.³⁰ When populations of girls and boys were evaluated for gender-typical behavior and then followed to adulthood, an amazing result emerged. Boys who showed highly female-typical behaviors early on were much more likely to become sexually attracted to men as adults (75 percent, compared with 3.5 percent of the general population) and girls who showed male-typical behaviors were far more likely to become sexually attracted to women as adults (24 percent, compared with 1.5 percent of the general population).³¹ However, this isn’t a universal result: not all tomboys develop sexual attraction to women and not all effeminate boys become sexually attracted to men. And, of course, as adults not all lesbians are mannish and not all gay men are effeminate. Yet, these findings are striking and suggest a general explanation: sexual orientation is just one aspect of variation in brain function that produces a set of behaviors that can be more or less gender-typical. For example, gender nonconforming girls are more likely to engage in rough-and-tumble play, are less likely to engage in cooperative social play, and are more likely to become sexually attracted to women as they grow up. The most likely explanation is that some combination of social experience, genes, fetal signals such as circulating hormones and immune molecules, and perhaps other biological factors we do not yet understand influence the relevant brain circuits for gender-typical behavior, of which sexual orientation is just one part of the package.

BEYOND GENDER IDENTITY AND sexual orientation lies the even more difficult question about why we are attracted to one person but not another. One thing that’s become clear from the huge data set derived from online dating is that—male or female, trans or nonbinary, gay or straight or bi—what we say are our requirements for a mate rarely turn out to be quite so important in the flesh. A straight woman may say that she needs a tall guy who likes the opera, but she may also be happy with a medium-size guy who hates the opera and would rather go to death metal shows. A gay man looking for an extrovert ginger might well find love with a shy blond instead. While certain traits can truly be deal breakers—these days, political affiliation is a line that many will not cross—we’re generally not as good at predicting who we’ll fancy as we imagine.

The perfume industry would like you to believe that attraction is all about human pheromones, and they are happy to sell you expensive concoctions designed to make your chosen romantic partners consumed with lust. The term pheromone was coined in 1959 to mean signal molecules that are characteristic of a particular species (or, for example, just the females of a particular species) and that trigger a stereotyped behavior in a conspecific target group (like the reproductive-age males).³² The first pheromone to be discovered was a sexual attractant produced by female silkworm moths (Bombyx mori) called bombykol. A tiny bit of this chemical attracts the amorous attention of male silkworm moths from miles around. Importantly, a dab of bombykol smeared on a twig (at concentrations mimicking those secreted naturally by randy female silkworm moths) will still attract males, even if there is no female moth around, and it will have no effect on any other species, insect or otherwise.³³ This is what defines a pheromone: it is produced by every group member of a species and acts on every member of a target group of that same species. While many pheromones are important for sexual behavior, they can also be used to signal social rank, individual or group territory, or the presence of food or danger.³⁴ Crucially, pheromones are not individual odors; every member of a species group can deploy them and every member of the target group will respond.³⁵

Pheromones are used by many animals, from insects to fish to mammals. Some pheromones diffuse freely in the air or water for long distances, while others are so sticky that they must actually be wiped on the recipient.³⁶ While the original pheromone, bombykol, was a single chemical, it is now understood that some pheromones are mixtures of a few different chemicals. Certain pheromones, like one in the urine of the male house mouse (Mus musculus domesticus) can elicit both rapid stereotyped behaviors, like aggression in other males and sexual attraction in females, and slower developmental effects, like accelerating the onset of puberty in young female house mice.³⁷

Our mammalian cousins like goats, mice, and rabbits use pheromones, so there’s no obvious reason why humans couldn’t get in on the party too. We have perfectly fine and sensitive odor detectors in our noses, which are the main way of receiving pheromones. And we secrete all kinds of odor molecules of many chemical classes, often with the metabolic help of bacteria living on our skin. The production of these odor molecules is different in males and females and, as anyone who has attended middle school can attest, they change at the onset of puberty.

One promising indicator of human pheromones, which received wide attention when it was first published in 1971, involved menstrual synchrony in women living together in a dorm at Wellesley College.³⁸ The idea was that women living together would signal to each other with pheromones to synchronize their cycles (it’s not entirely clear why this would be a good idea). A follow-up experiment in 1998 reported that sweat from the armpits of women in their fertile (late follicular) phase, while producing no conscious perception of odor when smeared on the upper lip of recipient women, nonetheless accelerated the recipients’ cycle (by advancing the preovulatory surge of luteinizing hormone), thereby promoting menstrual synchrony.³⁹ Unfortunately, the candidate chemicals from the armpit secretions of fertile women have yet to be identified, and, more importantly, the entire phenomenon of menstrual synchrony has failed to replicate in several subsequent studies.⁴⁰ While the idea of pheromone-based menstrual synchrony is still often mentioned in the press, it is now widely regarded as unproven by biologists.

Another line of human pheromone research was inspired by pig breeding. The hormone androstenone is secreted in the saliva of male pigs, and when detected by sows in heat will cause them to reflexively adopt a sexually recipient posture. It’s sold to pig farmers by DuPont as the product Boarmate and is used to determine when the sows are sexually receptive and hence ready to breed. When it was discovered that androstenone and some related compounds—androstenedione in men and estratetraenol in women—were also present in human armpits, this was enough to lead several groups of researchers to embark on a set of underpowered and poorly controlled studies to evaluate them as human pheromones. One involved spraying certain chairs in a waiting room with these compounds to see which were later chosen by women and men. The main problem here is that there was never any solid rationale to examine these particular compounds, rather than the hundreds of others present in human armpits. Despite many studies on these chemicals, there is no convincing evidence for their action as human pheromones.

We can take a hint for human pheromone discovery from the recent description of a pheromone in male mouse urine by Jane Hurst of the University of Liverpool. Male mouse urine is a sexual attractant for female mice, and a particular protein isolated from the urine—cheekily named darcin after Mr. Darcy, the smoldering heartthrob of Jane Austen’s Pride and Prejudice—can entirely mimic its effects on female mice, suggesting that darcin is a mouse pheromone. One possible objection to this interpretation is that perhaps the female mice in the study, rather than responding instinctively to darcin, have simply learned over time to associate it with attractive male mice. To address that concern, Hurst and her group raised females in a sort of strict girls’ boarding school for mice, no contact with males allowed. They showed that even these innocent, male-ignorant mice, when sexually mature, responded with sexual interest to either pure darcin or male mouse urine.⁴¹

As much as some people might like it, we’re unlikely to replicate this kind of experiment in humans. To dissociate learning from stereotyped action in our own species, perhaps the best bet is to look at instinctive behaviors in newborns, who haven’t had a chance to learn much yet. When mothers are breastfeeding, their areolae swell and secrete tiny droplets of a non-milk liquid from the bumps that surround the nipple called Montgomery’s glands. When this secretion is wiped on the upper lip of three-day-old babies, they purse their lips, thrust their tongue, and root around searching for the nipple (figure 12). Crucially, the areolar goo of one nursing mother can elicit this feeding response in a totally unrelated baby.⁴² Like a true pheromone effect, it is not dependent on the baby learning to associate feeding with a mother’s individual odor.

To prove that there really is a feeding-response pheromone in human lactating mothers’ areolar secretion, a chemical or small group of chemicals must be isolated and shown to fully replicate its effect on babies’ feeding behavior at naturally occurring concentrations. Then, ideally, depletion of those chemicals would be shown to remove the newborn’s feeding response evoked by the areolar secretion.

What all this means is that, contrary to the best efforts of the perfume industry and many popular magazines, at present, there are no proven examples of human pheromones. The best candidate for a human pheromone is in the aforementioned secretion from lactating nipples that causes newborn babies to feed. Not very sexy, that. Does this mean that, unlike mice or moths or goats, we humans simply don’t have pheromones for sexual behavior? Some scientists have said yes, noting that a part of the nose that’s responsible for pheromone detection in some other mammals, the vomeronasal organ, is merely a vestigial structure in humans (and the other great apes) that’s not connected to the brain. But this seems like a hollow argument to me. We know that rabbits and mice can detect some of their own pheromones using the main olfactory system, which we share with other mammals, not the vomeronasal organ.⁴³ As a result, I wouldn’t rule out human pheromones on that basis. More on this topic is to come in chapter 6.

FIGURE 12. Secretion from the areolar Montgomery’s gland elicits stereotyped feeding response in a newborn baby. A: Areola of a lactating woman three days after birth with a droplet of Montgomery’s gland secretion indicated by the arrow. B and C: Lip pursing and tongue protrusion responses produced by application of the secretion using a clean glass rod. From: Doucet, S., Soussignan, R., Sagot, P., & Schall, B. (2009). The secretion of areolar (Montgomery’s) glands from lactating women elicits selective unconditional responses in neonates. PLoS One, 4, e7579. Reused here under the terms of a Creative Commons Attribution License.

Tristram Wyatt, a pheromone researcher from Oxford University, has pointed out that if we’re going to search for human sexual pheromones, a good strategy would be to compare the molecules emitted from men and women before and after the onset of puberty, concentrating on those present solely in the latter.⁴⁴ Furthermore, he suggests that we may be looking at the wrong places on the body. Wyatt reminds us of a fact we touched on in chapter 1—that people with the dry-earwax variant of the ABCC11 gene, who mostly live in northeast Asia, have reduced apocrine sweat gland secretion and hence armpits with minimal odor, yet they manage to attract mates perfectly well. So, sex pheromones may not derive from apocrine gland secretions, but rather from secretions of sebaceous glands, which are present all over the body but particularly on the scalp, face, chest, and crotch. It may be that the armpits are not the best place to hunt for human sex pheromones, and that more embarrassing locations on the body must be swabbed by intrepid researchers.

JUST BECAUSE THERE IS no evidence to date to support the existence of human pheromones does not mean that individual odor is unimportant to sexual attraction. Recently, it has been claimed that attraction can be influenced by a class of molecule called the major histocompatibility complex (MHC), also known in humans as human leukocyte antigens (HLA). The MHC region of the genome, located in a section of human chromosome 6, directs the expression of a group of proteins that play an important role in immune recognition: they bind foreign protein fragments and present them on the cell surface to a key class of immune cell called T lymphocytes. This is a part of the genome that is particularly variable from person to person. It contains thousands of genes variants and hence even more possible combinations. Because you have two copies of chromosome 6, for any protein coded within the MHC region, you may have two identical copies or two different copies, depending on what you inherited from your parents.

The hypothesis, which is well supported in a range of other animals (fish, birds, mice), is that we tend to choose partners with MHC molecules that differ from our own, and that we can detect emitted MHC molecules through the olfactory system. The proposed reason for this choice is that children born of parents with different variants of MHC molecules will have a sort of hybrid immune system and hence be more resistant to infection, as increased MHC diversity improves the immune response to a wide variety of pathogens.

For this plan to work, humans must be able to detect and distinguish MHC molecules through smelling. To test this idea, olfaction researcher Avery Gilbert and his colleagues performed one of the more amusing experiments in human psychophysics. He obtained lines of genetically engineered mice that differed only in their MHC genes and asked if humans could tell them apart by odor.⁴⁵ In Gilbert’s own words:

I had blindfolded people sniff live mice in Tupperware containers with holes cut in the sides. Occasionally a mouse tail would get up someone’s nose; this seemed to bother some people more than others. The judges also sniffed tiny test tubes filled with mouse urine or dried fecal pellets.… For every odor source the results were clear: untrained humans could distinguish between the mouse strains based on smell alone.⁴⁶

Subsequent studies have shown that mice can return the favor and distinguish among human MHC variants. In 1995, these findings and others led Claus Wedekind and his coworkers to perform an investigation that has since become famous as “the dirty T-shirt experiment.” Male students were asked to wear the same T-shirt for two consecutive nights, and the next day female students were asked to rate the odors of six different T-shirts. They scored male body odors as more pleasant when the men associated with those odors had MHC types different than their own.⁴⁷

The dirty T-shirt experiment has been replicated and extended several times. Some researchers have found that men also prefer the body odor of women with MHC variants different from their own,⁴⁸ while other studies have failed to see such a difference in odor preference.⁴⁹ There’s a lively ongoing argument in this corner of science, with people ascribing failures to replicate to shaven versus unshaven armpits and fresh versus frozen T-shirts. To my knowledge, all of these experiments have been done on straight, cisgender men and women, so it’s unclear how this plays out for gay and bi and trans folks.

Let’s assume for a moment that the dirty T-shirt experiments are correct: that both straight men and straight women prefer the body odor of opposite-sex partners with MHC variants different than their own. That still doesn’t mean that they will then go on to have children with them. To determine if that is true, one can look at the DNA sequences of mate pairs and, by analyzing both the maternal and paternal chromosomes, ask if MHC variants reflect random mating with regard to MHC, or if people tend to choose mates with MHC variants different from their own. To date, the largest and best study of this type was performed using 239 mate pairs of Dutch ancestry from the Genome of the Netherlands project. Their finding was clear: the distribution of MHC variants reflected mate choice that was independent of MHCs—neither similar nor different MHCs were favored in the chosen mates. The results were statistically indistinguishable from assigning mate pairs at random.⁵⁰ It will be useful to see if this result is replicated and if it holds for other populations.

One interesting possibility is that humans do tend to choose mates with MHC variants different than their own, but only under those conditions where the burden of pathogens is high. In cross-cultural surveys, it has been claimed that people in geographical areas carrying a high prevalence of pathogens value a mate’s physical attractiveness more than those living in places with less of a pathogen burden.⁵¹ Perhaps the same is true of MHC variants. It may be that people in the Netherlands, with their temperate climate, good public health measures, and well-known penchant for cleanliness are simply the wrong population to see mating based on MHC difference detected through odor.

SEX COLUMNIST DAN SAVAGE is completely right when he writes, “When it comes to human sexuality, variance is the norm.” Beyond sexual orientation and even beyond one’s choice of particular sexual partner(s), there are whole other levels of individual variation in sexual behavior. We have little understanding of why people like their sex a particular way—fast or slow, hard or soft, involving this orifice or that one. And the development of various frills and kinks is almost certainly learned and somewhat serendipitous. There is no evidence for a genetic contribution to foot fetishism or a delight in leather underwear, BDSM practice or sexual peeing. There is, however, a genetic contribution to personality traits like novelty seeking, risk-taking, and obsession, all of which can manifest in the area of sexual behavior, even if they do not direct preferences for particular sexual behaviors.

We may be able to explain some individual preferences for certain sexual acts based on genetic variation in touch sensations. The fine pattern of innervation of the genitals and other erogenous zones does vary from person to person.⁵² To wit, some cisgender women may have more nerve endings (or more of a particular type of nerve ending) then average in the vagina and fewer in the labia minora and clitoris. Some cisgender men may have more in the penile shaft and fewer in the glans and anus. But we don’t yet know if these anatomical differences actually cause variation in sexual sensation or preferred sexual behavior. There are also likely to be inborn differences in the nerves and brain regions that underlie sexual sensation that are not structural and hence are not detectable with a medical scanning machine or even with a microscope. Such differences would only be revealed by measuring the electrical or chemical signals in the relevant individual neurons that transduce and process sexual sensations. In this way, the biological contribution to our individual preferences for particular sex acts resides not just in our brains, but also in the nerves that course through our skin and viscera.