Sex Differences Are Real, Overlapping, and Morally Non-Prescriptive

Steve Stewart-Williams frames the study of sex differences around a deliberately narrow moral claim: average differences between men and women can be real, partly evolved, and socially important without implying that either sex deserves less respect, opportunity, or freedom. His preferred practical ethic is: “let people be themselves.” If someone fits traditional gender roles, that is fine; if they do not, that is fine too. Evolutionary origin, in his view, carries no automatic moral status.

That distinction matters because Stewart-Williams thinks the subject is often treated as if explanation equals endorsement. If an evolutionary psychologist explains why men might be more inclined toward casual sex, aggression, risk-taking, or sexual jealousy, critics may hear an excuse for infidelity, violence, or coercion. Stewart-Williams rejects that inference. Natural does not mean good, permissible, or inevitable. It only means there may be an evolved contribution to a trait or motivation.

The controversy, as Stewart-Williams sees it, has an understandable historical source. Science has been used to justify sexism, and he cites Gustave Le Bon’s 19th-century claim that intellectually accomplished women were as rare as a “two-headed gorilla” as an example of the sordid background. That history makes people understandably nervous that talk of evolved sex differences will reopen old arguments about women’s inferiority. Stewart-Williams’s counterclaim is that modern science is not a continuation of that sexism but part of the correction: many of the differences under discussion are modest, many concern preferences rather than cognitive ability, and many put men in a worse light than women.

Chris Williamson presses the concern from the other side: the atmosphere around the topic has become so accusatory that even finding sex differences fascinating can be read as evidence of hidden reactionary motives. Stewart-Williams agrees that the subject should be handled carefully. The necessary caveats are not cosmetic. Most psychological sex differences are not huge; within-sex variation is large; overlap between men and women is extensive; and averages do not prescribe what any individual should do.

Williamson suggests that modest differences may become large when stacked across many traits — violence, risk-taking, parental investment, sexual interests. Stewart-Williams accepts that multivariate sex differences can be statistically real, but warns against using them to inflate the practical size of differences. If a population were split randomly by coin toss, small accidental differences across many traits could also be aggregated into apparently large multivariate separation. For most real-world questions, he argues, people usually care about a single trait or domain — parenting, aggression, sexuality, interests — and the size of the difference in that domain should be discussed directly.

The result is a restrained view rather than a maximalist one. Men and women are not interchangeable on every trait, but neither are they two alien species. The differences are often meaningful enough to affect social patterns, institutions, health, relationships, and policy. They are also usually overlapping enough that individual freedom is the only defensible rule.

Steve Stewart-Williams defines sex in the traditional biological way: by gamete size. In sexually reproducing species with two gamete sizes, males are the sex that produces the smaller gametes, sperm, and females are the sex that produces the larger gametes, eggs. He describes this as a definitional truth in biology rather than a statistical tendency: “male” and “female” are defined by reproductive role.

He distinguishes that definition from the empirical fact that, across the vast majority of sexually reproducing species, there are two gamete types rather than a continuum of gamete sizes. The technical term for two differently sized gametes is anisogamyA reproductive system with two unequal gamete sizes: small sperm and larger eggs.. Stewart-Williams says a minority of sexually reproducing species are isogamousA reproductive system in which gametes are the same size rather than divided into sperm and eggs., and that isogamy likely came first in evolutionary history. Once some gametes became slightly larger by carrying more resources for offspring, smaller gametes gained an advantage by seeking out the larger ones and being produced in greater numbers. Intermediate gametes were disadvantaged: not large enough to carry the survival benefits of big gametes and not numerous or mobile enough to compete with small ones.

Chris Williamson summarizes the logic as a kind of barbell: the middle is where gametes “go to die.” Stewart-Williams agrees. Medium-sized gametes lose on both sides. They do not survive as well as resource-rich larger gametes, and they cannot be produced in the same numbers as smaller ones.

This gamete definition does not deny variation. Stewart-Williams says there is “tons” of within-sex variation in masculinity, femininity, body size, and other phenotypic traits. Intersex conditions and variation within the sexes do not, in his account, disprove the binary definition any more than an exceptionally tall woman disproves the claim that men are taller on average. The binary applies to gamete type; it does not imply that every male or female body is identical, nor that every trait has a clean split.

Williamson reads from a ChatGPT screen that puts “total global sperm production” at 200 quadrillion and “eggs released” at 70 million. The exchange is jokey, and the displayed source is only a ChatGPT interface, so it should not carry the argument as evidence. Its function is illustrative: Williamson uses the contrast to dramatize the asymmetry between sperm and eggs, while Stewart-Williams’s substantive claim rests on the gamete-size distinction rather than on the displayed global totals.

Stewart-Williams does not deny socialization. He says culture plainly affects the size and details of sex differences: differences vary across cultures and over time. His argument is narrower: many differences are unlikely to be only socialized. The case, he says, comes from multiple imperfect lines of evidence pointing in the same direction.

The first line is early development. Some differences appear so early that a purely social explanation becomes less plausible. Aggression and risk-taking differences appear as soon as children are physically able to move around, take risks, and aggress. Stewart-Williams gives one concrete example: from early toddlerhood, more boys than girls end up in emergency rooms from risk-taking. Puberty is another clue. Some differences may exist earlier and then expand sharply after puberty, across cultures, which is difficult to explain without a biological contribution.

The second line is that some differences appear despite cultural pressure rather than because of it. Stewart-Williams uses male aggression as the central example. Most cultures do not celebrate uncontrolled young male violence; they try to suppress it. Parents and teachers tell boys off for aggression more often than girls, not because they approve of girls’ aggression, but because boys give them more occasions to intervene. Yet male aggression still rises after puberty.

Williamson adds fetal brain differences detected in the womb as an example that strains purely social accounts. Stewart-Williams agrees that the idea of fetuses being socialized into different brain structures is not plausible.

The third line is persistence across time. Stewart-Williams says sex differences in mate preferences were consistent throughout the 20th century in the United States, and sex differences in career-related interests were already evident when first measured in the early 1900s. Men, on average, were more interested in things and things-related occupations; women, on average, were more interested in people and people-related occupations. That pattern persisted despite sustained social pressure to move women into traditionally male roles.

The fourth line is hormonal correlation. Higher prenatal testosterone exposure is associated with later aggression, risk-taking, lower parental inclinations, and other male-typical traits. Stewart-Williams highlights congenital adrenal hyperplasia, or CAH, in which girls exposed to very high prenatal testosterone show, on average, more male-typical patterns: less interest in dolls, less interest in marriage and children, greater interest in things-related professions, and a higher likelihood of bisexual or lesbian orientation.

The fifth line is cross-cultural universality. Stewart-Williams says the key differences under discussion tend to appear in every culture or the vast majority of cultures: men are more aggressive; men predominate among homicide perpetrators; men take more risks and suffer more accidental fatalities; women are more involved in parental care.

The sixth line is continuity with other species. Many sex differences seen in humans are also found in species subject to similar selection pressures. Stewart-Williams does not claim every species shows the same differences; the point is that when similar pressures recur, similar patterns often recur.

The strength of the argument is cumulative. None of these lines alone “seals the deal,” he says. But when a given sex difference appears early, persists despite counter-pressure, recurs across cultures, tracks prenatal hormones, and resembles patterns in other species, the purely social explanation becomes harder to maintain.

The gender equality paradox is one of Stewart-Williams’s central examples because it reverses a common prediction. Social role theory and patriarchy theory, in their simplest forms, predict that sex differences should be larger in more patriarchal societies with stricter gender roles and smaller in more gender-equal societies with looser roles. Often, Stewart-Williams says, the opposite appears: more gender-equal societies show larger sex differences.

He does not present this as something he predicted from the start. Even as an evolutionary psychologist, he originally expected patriarchy and stricter sex roles to push men and women further apart. The surprising result is that more egalitarian societies often show wider gaps in traits such as personality, career interests, and even some cognitive or physical measures.

One rebuttal he attributes to sociocultural explainers is that people in more patriarchal societies may be more sex-segregated. Men compare themselves mostly with other men, women with other women, and self-report inventories may therefore compress apparent sex differences. Stewart-Williams says that rebuttal makes sense for self-report measures such as personality. But it is weaker where the paradox appears in domains not reducible to self-report, such as spatial ability or height.

The people-versus-things divide is the most consequential application. Stewart-Williams says it is one of the larger psychological sex differences, around one standard deviation or more. Men are more interested in things and things-related work; women are more interested in people and people-related work. He refers to one study of about 200,000 people that, in his telling, found the career-interest difference in 53 out of 53 nations, and another with half a million people that found it in 80 out of 80 nations. In more gender-equal countries, he says, the interest gap grows, and occupational choices reflect it.

Chris Williamson connects this to what he calls the “soft bigotry of male expectations”: the tendency to treat male-typical domains, such as STEM and things-oriented work, as the status benchmark and to push women toward them while showing less urgency about men entering female-dominated caring professions. Stewart-Williams agrees that this framing can smuggle in a negative valuation of female-typical interests.

At the same time, Stewart-Williams acknowledges a practical reason people care about STEM gaps: such jobs often pay more. But he argues that pay does not erase the interest difference. If interventions assume all occupational gaps are caused by bias and barriers, they may produce resentment, costly interventions that miss the cause, and eventually more coercive attempts to force parity.

His proposed alternative is not complacency about discrimination. It is to remove bias and barriers, present people with accurate information, and then respect choices. If some residual gaps remain because average preferences differ, he argues, those gaps are not automatically evidence of injustice.

The deepest evolutionary explanation in Stewart-Williams’s account begins with reproductive variance: the range in how many offspring males and females can produce. In most species, average reproductive success must be the same for males and females when the sex ratio is roughly equal, because each offspring has one male and one female parent. But the variance can differ dramatically. A few males may have many offspring, while many males have few or none. Females may have a lower ceiling and less variance.

This is the logic Stewart-Williams associates with Bateman’s Principle: reproductive variance is greater in males than females in most species. Robert Trivers’s refinement, he says, was to identify parental investment as the major driver. Females in mammals invest in pregnancy, birth, nursing, and often care. That limits the maximum number of offspring they can produce and makes each mating decision more consequential. Males, in principle, can produce offspring through a single sex act, which creates selection for traits that help them compete for mates, seek multiple partners, display, fight, or acquire status and resources.

Parental investment includes gestation because it forecloses other reproductive opportunities. A pregnant female cannot simply start another pregnancy at the same time. Nursing and childcare extend that investment. The asymmetry lowers female maximum offspring number and raises the importance of choosiness.

Stewart-Williams is careful not to reduce everything to parental investment alone. Ecology matters too. In solitary species, a male may not have access to many other females even if females invest more in offspring; in such cases, male and female maximum offspring numbers may be closer. But in many species, parental investment is the major cause of the sex difference in reproductive variance.

Humans complicate the standard mammalian pattern. Stewart-Williams says humans are unusual among mammals because we pair bond and exhibit high levels of biparental care. Most mammalian males are, in his phrase, “deadbeat dads.” Humans are more like birds in this respect: he says around 90% of birds form pair bonds and have biparental care, whereas perhaps 5% to 10% of mammals do. In primary reproductive behavior, humans are more like the average bird than the average mammal.

That human pair bonding “nerfs” the size of many sex differences. Men still have greater reproductive variance than women, and the usual mammalian patterns remain: greater male interest in sexual variety, greater male aggression, more male risk-taking. But because human males often invest substantially in offspring, the differences are muted relative to many mammals.

Stewart-Williams is skeptical of the commonly repeated claim that, ancestrally, about 40% of men and 80% of women reproduced. He has heard the number but thinks human reproductive skew was probably smaller, because male parental investment and biparental care constrain men’s maximum offspring number relative to females. He declines to give a replacement figure.

The reproductive-variance framework also explains why sex ratios tend toward 50-50 rather than one male for every 20 females. If females greatly outnumbered males, parents producing sons would have a reproductive advantage because sons would face abundant mating opportunities. Selection would then favor more sons until the ratio moved back toward equilibrium. If males greatly outnumbered females, daughters would become the better bet. Selection operates on individual reproductive advantage, not on what would be efficient for the group.

When the discussion turns to sex itself, Stewart-Williams identifies one of the largest sex differences: sociosexuality, or interest in casual sex, sexual variety, and no-strings-attached sex. Men score higher on average, however the trait is measured. But he stresses that the gap is not as large as some assume.

He estimates the effect size at roughly Cohen’s d of 0.8 to 1 — large for a human psychological sex difference, but still with substantial overlap. If one man and one woman were selected at random, the man would score higher perhaps two-thirds to 70% of the time. That means a significant minority of women score higher than a significant number of men.

The evolutionary explanation is direct. If male animals can increase reproductive success by mating with multiple females, selection favors a stronger motivation for sexual variety. In humans, that motivation is moderated by pair bonding and biparental care, but not erased. Stewart-Williams likens the male motivation, in its extreme form, to a “heat-seeking missile” seeking multiple partners.

Williamson raises infidelity and cheating as examples where explanation can sound like excuse. Stewart-Williams repeats the distinction: evolutionary causation no more excuses cheating than a sociocultural explanation would. All behavior has causes. If identifying a cause removed responsibility, then every cause — cultural, developmental, genetic, situational — would do the same.

Williamson cites one striking claim about Ashley Madison, the cheating website: 20 million active male users and 1,492 active female users, despite women receiving free lifetime membership while men had to pay. Stewart-Williams calls the difference “wild” and treats it as consistent with the broader sex difference in interest in casual sex. In the source, this is presented as Williamson’s cited figure rather than independently examined evidence.

The same difference appears, Stewart-Williams says, in gay and lesbian behavior. He says gay men have more casual sex than straight men, and lesbians have less casual sex than straight women. His interpretation is that same-sex behavior removes the need to compromise with the other sex’s typical sexual inclinations. Gay men are not constrained by female gatekeeping; lesbians are not pulled upward by male interest in casual sex. The pattern, in his view, provides a clearer window into male and female sexual motivations.

Within heterosexual marriage, Williamson cites a related finding: men often say they would like roughly twice as much sex as they are having, while women report being satisfied with the current amount. Stewart-Williams interprets that as evidence that women are more often setting the level, though he adds that some women also desire more sex than their male partners. The average pattern is not a universal rule.

Stewart-Williams emphasizes that men and women want many of the same things in long-term partners. Both sexes tend to value intelligence, kindness, love, attraction, good looks, and emotional stability. The differences are not usually about one sex caring and the other not caring; they are about relative weighting.

Men, on average, place more weight on physical attractiveness in long-term mates. Stewart-Williams links that to fertility cues, especially youthfulness. Because female fertility is more tightly linked to age, and because menopause ends female fertility well before the end of life, men evolved to weight youthfulness and associated physical cues more strongly.

Women, on average, place more weight on status and resources in long-term mates. Stewart-Williams gives two possible explanations. The first is biparental care: a man with resources and standing is better able to invest in offspring and co-raise children. The second is a “peacock’s tail” explanation: resources and status may indicate good genes, because they reveal traits that helped a man climb the hierarchy and may be passed to offspring. He thinks both may contribute.

Context changes the pattern. In short-term mating, women place much more weight on physical attractiveness, and the long-term sex difference in looks largely disappears. In a low-commitment encounter, the man is less likely to provide future investment; the main potential benefit is genetic. Physical attractiveness becomes a fast proxy for mate quality.

Williamson discusses sperm-donor preferences as a way of separating attraction from desired offspring traits: if courtship and seduction are removed, the choice may reveal more directly what traits someone wants for a child. Stewart-Williams says mate preferences in that context appear similar to those that emerge in actual social interactions. He treats that as suggestive, not as a complete account of attraction.

The Clark and Hatfield campus study, as Stewart-Williams summarizes it, provides the most vivid illustration of sex differences in casual-sex willingness. Attractive male and female confederates approached strangers and said they had noticed them around and found them attractive, then asked one of three questions: whether the stranger wanted to go out that night, come up to their room, or go to bed with them. For the date request, Stewart-Williams says the original study found roughly equal acceptance, around 50-50, though later replications found men somewhat more likely to say yes. For the room request, he says about 67% of men said yes, compared with about 6% of women. For the sex request, he says 75% of men said yes and 0% of women did. These percentages are Stewart-Williams’s in-conversation recall of the study.

The refusals differed too, in Stewart-Williams’s account. Men who declined tended to be polite, apologetic, or even asked for a rain check. Women were more likely to react as if the request were alarming or insulting. Stewart-Williams gives two reasons. First, men’s greater interest in casual sex makes the offer more likely to be flattering. Second, a woman going off with an unknown man faces greater physical risk than a man going off with an unknown woman. The size of the gap reflects both desire and danger.

Men are more responsive to visual sexual stimuli, according to Stewart-Williams, though both sexes care about looks. He connects that to men’s stronger weighting of fertility-linked cues in women and to men’s much higher consumption of pornography. Williamson contrasts that with women’s greater interest in romance and “romantasy” novels.

Stewart-Williams finds the romance-novel pattern interesting because it is not simply the inverse of men’s casual-sex interest. Men and women, he says, appear about equally interested in long-term committed relationships, falling in love, and forming a bond. Some recent papers, he adds, even suggest men are in some ways more romantic than women: they fall in love more quickly, say “I love you” earlier, and may suffer more after breakups. Why women are much more interested in romance novels remains unclear to him. He jokes that perhaps men are too busy with porn, but does not offer that as an explanation.

The discussion also turns to sexual overperception bias, associated in the exchange with David Buss and Martie Haselton. Men are more likely than women to falsely infer sexual interest from the other sex. Williamson describes it as a failure of cross-sex mind reading: men’s estimate of a woman’s attraction to them may track their own attraction to her. Stewart-Williams agrees that projection may be involved.

The evolutionary logic is the smoke-detector principle. For ancestral men, missing a genuine sexual opportunity could be reproductively costly; suffering an awkward rejection was less costly. Selection could therefore favor a bias toward over-detection. The point helps explain why misreads occur; it does not turn misread signals into consent or excuse unwanted advances.

This theme recurs in jealousy and friendship, but the speakers keep some uncertainty in the frame. Stewart-Williams says many men would be interested in sleeping with female friends, while fewer women are interested in sleeping with male friends. Because of overperception bias, male friends may assume women are more interested than they are. Williamson extends this into a broader claim about partners interpreting same-sex rivals for one another — men warning women about male friends, women warning men about other women — but Stewart-Williams’s narrower point is about asymmetry in sexual interest and the likelihood of misreading it.

Sex differences in aggression are among the most consistent patterns Stewart-Williams discusses. The gap is modest for verbal aggression, larger for low-level physical aggression, larger still for serious violence, and largest at homicide. Men are not only more often perpetrators; they are also more often victims.

For verbal aggression, Stewart-Williams estimates an effect size around 0.5: if a man and woman are selected at random, the man will be higher perhaps two-thirds of the time. But as the severity rises, the difference expands. In every nation for which he has seen data, more than 90% of homicides are perpetrated by men. Men also constitute the majority of homicide victims in almost all datasets.

Williamson cites numbers comparing humans and chimpanzees: in humans, he says, males commit 95% of homicides and are 70% to 80% of victims; in chimpanzees, males commit 92% of “chimpicides” and are 73% of victims. Stewart-Williams cautions that chimpanzees are more aggressive than humans overall. The point he accepts is not that human violence equals chimpanzee violence, but that the sex ratio in lethal aggression and victimization is similar.

The explanation returns to male reproductive variance. If some males can gain much more reproductive success through status, resources, territories, and mating opportunities, selection favors male-male competition. Aggression can be a route to those goods. For females, the reproductive ceiling is lower, so fighting other females for status and resources tends to produce fewer reproductive gains relative to the costs.

This does not mean women are incapable of aggression, nor that all male aggression is adaptive. It means the selection pressure has been stronger on males, especially for same-sex competition.

The same framework helps explain why protectiveness can be attractive while aggression remains dangerous. Williamson raises a tension: if violence is such a major fear for women, why are many women attracted to powerful men who could be violent? Stewart-Williams distinguishes aggression from protectiveness. Aggression itself does not straightforwardly increase attractiveness; protectiveness does. Muscularity, height, and physical formidability can increase attraction, but they are a mixed blessing. A physically dangerous man may protect the family, but he may also turn that danger inward. For that reason, women also care about character — whether the man can and will regulate his aggression.

Height, in Stewart-Williams’s view, probably belongs somewhere in this protective cluster, along with muscularity and physical formidability. Williamson argues that the preference is not merely about relative size to the woman: a shorter woman may still prefer a very tall man, even though a shorter man would still be larger than she is. Stewart-Williams does not turn that into a settled model of height preference; he treats protectiveness as a plausible part of the explanation.

Stewart-Williams says sexual violence is much more often perpetrated by men than women. He does not give exact numbers, saying he wants to remain “directionally correct, specifically vague,” but suggests the effect size for extreme forms of sexual violence, assault, and harassment may be around two standard deviations.

The explanation he discusses is stark. Men have, on average, greater interest in casual sex and sexual variety. Men are also more willing to use violence in some contexts. Combine those facts and, in some men, there will be circumstances where a man wants sex with a woman who does not want sex with him and uses force to try to get it. Stewart-Williams says some researchers have argued that this tendency in some men has evolutionary roots, and he notes that male sexual coercion occurs in many species, not only humans.

Stewart-Williams adds a second evolutionary point: female choice is deeply important across the animal kingdom because females typically have fewer reproductive opportunities and higher investment per offspring. Women have evolved to be choosier, with mate-choice criteria that matter. Sexual coercion forcibly overrides that choice. This may help explain why rape and sexual coercion are so profoundly distressing: they do not merely cause harm in a generic sense; they violate a core evolved domain of female agency.

Williamson suggests that men often fail to grasp the felt reality of women’s fear of sexual violence, especially lower-level harassment or unwanted sexual attention. Stewart-Williams agrees and cites this as another failure of cross-sex mind reading. Men may underestimate how upsetting harassment is because, if the situation were reversed, they might not experience it with the same fear, vulnerability, or threat.

The political asymmetry, in Williamson’s view, is that people who deny sex differences rarely argue for parity in male-bad domains. Few insist that, if women were fully liberated, they would commit as many homicides as men. Stewart-Williams uses violence to make a broader point about equality: equality is not good in itself. Reducing male violence would narrow a sex gap and make the world better, but the reason is that violence is bad, not that the gap is bad. Raising female violence to male levels would also create equality and would be a disaster.

Stewart-Williams says women do more direct parenting in every culture where good data exist. They invest more in the young and do more parental care. Men, he adds, do far more than the average male chimpanzee or walrus, but less than women in every culture.

The controversial part is not the descriptive claim alone. It is the fear that men will use it to excuse doing less work: if women are naturally more parental, then women should change the nappies and men can opt out. Stewart-Williams rejects that normative move. But he also argues that treating the claim as insulting reveals something telling: it implies that parental care is lower-status or less valuable. If parenting were valued properly, saying women are more inclined toward it on average would not be heard as derogatory.

Williamson’s “soft bigotry of male expectations” returns here. If male-typical activities are treated as the standard of value, female-typical traits are implicitly downgraded. Stewart-Williams calls that sexist if women really are, on average, more parental by nature: it assigns a negative valuation to a more female-typical trait.

The Israeli kibbutzim serve as Stewart-Williams’s strongest example of parenting inclinations emerging despite culture. Some kibbutzim created highly egalitarian communal arrangements in which children lived in communal houses rather than under the same roof as their biological parents. Dedicated adults cared for the children, and the system was designed to undermine traditional family structures and private parental attachment.

The arrangement worked for a while, but parents hated it. In Stewart-Williams’s telling, mothers especially hated it and agitated against it. He presents this as a case where women were not forced by men into primary caregiving; rather, women were central to the pressure for a more direct maternal role. For him, it is another instance of a pattern appearing despite cultural pressure rather than because of it.

Williamson raises the modern work-family conflict, including a video of a mother at work watching her child’s first steps or first words via a daycare screen. Stewart-Williams says both sexes face trade-offs, but the clash appears still to be a bigger deal for women. Many women partially solve it by taking more time off when children are young and returning to work later, especially once children start school. Neither speaker treats this as an easy policy conclusion. The point is that choices always trade one good against another, and social narratives that devalue motherhood can make those trade-offs harder to discuss honestly.

Both sexes get jealous, but Stewart-Williams says the focus differs. Men are more focused on sexual infidelity; women are more focused on emotional infidelity. If asked which would be more upsetting — a loved partner sleeping with someone else or forming a close emotional bond with someone else — both sexes generally find both upsetting. But men are more likely to rank sex as worse, while women are more likely to rank emotional involvement as worse.

The explanation is paternity uncertainty for men and desertion risk for women. No woman gives birth and wonders whether the baby is hers. Men, however, face a non-zero possibility that a child they invest in was fathered by another man. Sexual infidelity is therefore a uniquely salient reproductive threat for men.

Women face a different threat. A man’s emotional attachment to another woman may signal that he will leave and redirect investment. In ancestral conditions, being left “holding the baby” could reduce a woman’s ability to raise current offspring and delay future reproduction. Sexual infidelity is still painful and can destroy a relationship, but emotional defection more directly threatens the continued partnership.

Williamson asks about a modern case: a person discovering through ancestry testing, decades later, that their father had conceived another child in a one-night stand long ago. By then, the father may have stayed, invested, raised the family, and the outside child may no longer need provisioning. Why would the discovery still be devastating? Stewart-Williams suggests the proximate emotional mechanism may not track the current adaptive stakes. Jealousy evolved as a gut reaction to threats of sexual or emotional defection. It may still activate 30 years later even when the original reproductive consequences are gone.

Mate guarding follows from jealousy. Both sexes may surveil partners, but for different reasons: men to reduce paternity uncertainty, women to reduce desertion risk. At low levels, this can mean ordinary vigilance about a partner’s closeness to attractive others. At extremes, it becomes controlling surveillance.

Stewart-Williams describes most personality sex differences as modest but consistent. In the Big Five framework, the clearest differences are neuroticism and agreeableness. Women score higher on both, on average, with effect sizes he places roughly between 0.2 and 0.5 standard deviations. Women may also score slightly higher on conscientiousness, extraversion, and openness in some studies, but those differences are smaller and less consistent.

He interprets higher female neuroticism as part of a broader self-protective tendency. Men take more risks, which he links to status competition and higher reproductive variance. The flip side is greater female caution. Anxiety, in his functional description, protects like a turtle shell: it helps an organism avoid danger. The same protective tendency can become costly at the extreme.

Agreeableness may be the flip side of male aggression and status striving. Women, on average, are more compassionate, friendly, and oriented toward understanding others. Stewart-Williams is less confident about the exact evolutionary explanation, but sees it as connected to different social strategies and lower selection for direct physical competition.

The people-versus-things difference is larger than these Big Five differences. Stewart-Williams treats it as one of the most important psychological sex differences because it maps onto education and work. It is also unusually robust: he says it has been documented since the early 20th century, found across many cultures, associated with prenatal hormones, and resistant to interventions designed to move women into things-oriented fields.

Cognitive differences are more controversial and generally smaller. Stewart-Williams says there is basically no average sex difference in IQ or general cognitive ability. Some IQ tests have been designed to balance male- and female-favoring subtests, but he says even tests not deliberately constructed that way show essentially equal averages in representative samples.

There is, however, a small difference in variance: male cognitive ability varies slightly more. That means more males at both extremes — more male geniuses and more males with very low cognitive ability. Stewart-Williams stresses that this is a small effect and one of the most controversial claims in an already controversial area.

Specific cognitive abilities show small average differences in both directions. Men tend to do somewhat better on spatial abilities. Women tend to do better verbally. Among the minority who are highly gifted at math, women are more likely than men also to be verbally gifted — what Stewart-Williams calls a “double threat.” His conclusion is that cognitive sex differences exist in specific domains, but they are smaller than personality differences and much smaller than interest differences. Cognitively, men and women are much more alike than different.

Physical and mental health differ by sex in ways Stewart-Williams thinks are too important to ignore. Men are more prone to cardiovascular disease and heart attacks, especially in middle and later adulthood before very old age. Women are more prone to immune disorders and pain disorders. Men are more prone to most cancers other than sex-specific reproductive cancers. Men die younger than women in the vast majority of cultures, even many where women face high childbirth mortality.

Williamson cites eunuchs in the Korean Chosun Dynasty living longer than intact males of similar social standing as evidence that testosterone carries a biological cost. Stewart-Williams treats the example as supporting a provocative claim in his book: there is, in principle, an intervention that could “obliterate” the average male-female lifespan gap and extend male lifespan — castration before puberty. He is not presenting it as a practical recommendation. His point is that testosterone appears to contribute to men being bigger, stronger, more aggressive, and shorter-lived.

Williamson also cites Randy Nesse’s calculation that lowering male mortality to female rates would save more years of life than curing cancer. Stewart-Williams says people are so accustomed to men dying younger that they do not register the scale of the gap. In the source, both the Chosun eunuch example and the Nesse comparison are introduced as Williamson’s cited claims and then discussed by Stewart-Williams, not independently audited.

Risk-taking is a major behavioral part of that story. Williamson says 88% of Darwin Award winners — people who die through spectacular self-inflicted stupidity — are men, while 90% of Carnegie Hero Awards, given for risking one’s life to save a stranger, also go to men. Stewart-Williams treats the cited contrast as an illustration of the same trait cutting both ways. Risk can produce catastrophic failure or extraordinary prosocial sacrifice.

The same logic applies to entrepreneurship and achievement. More men may start businesses that become extremely successful, but more men also go bankrupt. Risk produces both winners and losers. Stewart-Williams resists the simple prescription that women should be encouraged to take more risks, because increasing risk exposure increases both upside and downside. His preferred approach is to tell everyone the trade-off clearly and let them choose.

Mental health follows similar distributional logic. Women’s higher neuroticism corresponds to higher rates of depression and anxiety, including clinical depression and clinical anxiety disorders. Stewart-Williams does not think clinical depression or severe anxiety is adaptive. Rather, he sees them as maladaptive extremes of traits that may be protective nearer the average. Men are somewhat more prone to schizophrenia, much more prone to antisocial personality disorder, sociopathy, and psychopathy, and boys are more often diagnosed with autism and ADHD. He does not offer a settled theory for the schizophrenia difference. ADHD, he suggests, may be an extreme of boys’ higher average activity and distractibility.

Bipolar disorder, by contrast, shows little sex difference. Stewart-Williams calls it “the least sexist disorder on record” because it appears at similar frequencies in men and women across cultures.

Stewart-Williams ends with a symmetrical warning. The danger of exaggerating sex differences is familiar: people may moralize averages, force individuals into traditional roles, or treat exceptions as defects. Because most psychological differences are modest and overlap is large, many people will not fit sex-typical patterns. His “let people be themselves” principle rejects coercing them into roles.

But denying or minimizing sex differences has costs too. It can replace traditional gender straitjackets with a unisex or reversed straitjacket. It can misread occupational gaps as pure discrimination, creating resentment and interventions that do not address the main causes. It can produce reverse discrimination in the name of eliminating gaps, which Stewart-Williams describes not as undoing injustice but adding new injustice. And it can discourage women from entering fields by portraying those fields as uniformly hostile when preferences may also contribute to the gap.

Health is another cost. If sex differences are ignored, symptoms may be missed. Cardiovascular disease can present differently in women, with symptoms such as shortness of breath rather than the stereotypical signs associated with men. Autism may be underdiagnosed in girls because it presents differently, including less repetitive behavior. If clinicians use a male template and ignore sex differences, girls may not get help.

Social problems also require accurate sex-specific understanding. If intimate partner abuse is assumed to run only male-to-female, female-to-male abuse can be missed. Stewart-Williams says low-level forms such as verbal abuse and pushing or shoving are often bidirectional in the West, with some data suggesting verbal abuse may be more common from women to men. Severe aggression, however, remains more male-predominant. Accuracy requires holding both facts at once.

Exaggeration creates the mirror-image health and social errors. If cardiovascular disease is treated as only a male problem, it may be missed in women. If depression is treated as only a female problem, it may be missed in men, even though roughly a third of depression cases are male. If sexual harassment or intimate partner violence is treated as exclusively male-to-female, reverse cases will be ignored, even if men commit the majority of severe cases.

The irony Stewart-Williams notes is that some gender progressives minimize sex differences in many domains but exaggerate them in domains such as sexual harassment or partner abuse, treating them as if they run only one way. His position is not that differences should be minimized or maximized according to political convenience. It is that the size, direction, overlap, and context of each difference should be described as accurately as possible.