A few years ago, a friend heard a speaker at an academic conference make a disparaging reference to a new book that, supposedly, denied that there were any sex differences between men’s and women’s brains and behaviour. The author of the dreadful tome – ‘this woman from Melbourne’ – turned out to be me. Afterwards, my friend asked the speaker if he’d actually read my book. He hadn’t.
Sadly, this is the type of anecdote for which my colleagues and I can provide many examples. It draws on a powerful narrative in which anyone who criticises accounts of evolved sex differences is seen as refusing to face up to the role of biology. Yet the question I tackled in the offending book, Testosterone Rex (2017), is an important and active one in evolutionary science and neuroscience. How is it that the genetic and hormonal components of sex can create two distinctly different reproductive systems – and yet human male and female behaviour shows itself to be flexible, diverse and often surprisingly similar?
In trying to understand that puzzle, I’ve used the work of the many scientists whose research has revealed the distorting effects of entrenched assumptions about sex differences. Some of this research is informed by a feminist awareness of how science has embedded cultural biases and stereotypes into its theories, hypotheses, methods and interpretations. But this kind of work also comes with an occupational hazard: accusations of being a politically motivated ‘sex-difference denier’. According to psychologists and neuroscientists whose positions I’ve criticised, I put politics before science, confuse equality with similarity, and wish to abolish femininity entirely. This ungrounded ad feminam commentary has even extended to suggestions that my use of the acronym ‘T’ for ‘testosterone’, when discussing its effects on the brain, indicates that I am ‘conflicted about how to incorporate this experimental evidence into [my] world view’, and that my work is a misguided reaction to experiences at dinner parties listening to chauvinists invoke the ‘caveman’ defence of bad male behaviour.
The claim here is that progressive politics lead to the rejection of particular findings or theories – not on intellectual grounds, but because they’re politically unpalatable or inconvenient. The evolutionary psychologists David Buss and William von Hippel recently laid this charge against the field of social psychology, my own work presented as an exhibit for the prosecution. They argued that Left-leaning political ideology has led to the adoption of a ‘blank slate’ view of human nature and group differences.
So, we seem to be at an impasse over what poses the greater risk to the study of the human mind: gender bias or feminist bias. These two perspectives might seem poles apart, but they have something fundamental in common. They’re both concerned about threats to scientific objectivity – the principle that, as much as possible, science should remain untarnished by politics, cultural bias, interests or preferences. This commonality is good news. It means that those who take opposing sides in these debates don’t necessarily need to be adversaries by default. As scientists, we should all be able to get behind correctives that support scientific objectivity. Unfortunately, those in the thick of the controversy often overlook this important overlap of interests.
Something that’s particularly unhelpful in debates about objectivity is a model of science where researchers simply dig facts out of nature, like a farmer harvesting turnips. This is not how science works. Facts about the world don’t passively lie in wait for scientists to uncover them. It’s impossible to do science without background theories and assumptions that influence the many decisions scientists must make: which hypotheses to test; what methods are appropriate; which populations to sample (and what size of sample); how to characterise and analyse data; how to interpret results; which findings to emphasise.
The feminist psychologist Sandra Bem coined the phrase ‘the lenses of gender’ to capture the particularly pervasive and pernicious assumptions people sometimes harbour about men and women. One such hidden assumption, ‘gender polarisation’, relies on the idea that there are ‘mutually exclusive scripts for being male and female’. Since scientists, too, are people, this lens can sharpen the focus on sex differences even in scientific research – and blur sex similarities. Take the popular view of sexual selection based on what are known as ‘Bateman’s principles’, named in honour of the 20th-century biologist Angus Bateman. His important and inventive research was inspired by Charles Darwin’s theory of sexual selection, which held that males are under greater selection pressure to be chosen as a mate. Hence the evolution of male characteristics such as big antlers (for beating off the competition) or elaborate plumage (to draw in crowds of admiring females). If so, Bateman reasoned, there should be greater variation in reproductive success in males than in females – that is, a wider spread between the least and most successful individuals. (This is because, like natural selection, sexual selection needs variations in reproductive success to work.)
Bateman’s research with fruit flies, published in 1948, reported that males did indeed show more variation in reproductive success, and that there was a stronger link between promiscuity and reproductive success in males than in females. Bateman’s principles are the foundation of the familiar idea that, because dispensing sperm is cheap, but harbouring and nurturing an egg is costly, females tend to be sexually choosy and inclined to reserve their favours for the best male on offer. Meanwhile, males – who, unlike females, have much to gain from winning multiple mates – are ardent and competitive by near-universal evolutionary design.
But consider an unexpected observation made in the 1970s by the primatologist Sarah Blaffer Hrdy: female langur monkeys in South Asia tended to have multiple sexual partners. This flew in the face of received wisdom; theoretically, as Hrdy drolly observed, female promiscuity ‘should not have existed’. But by paying attention to females, she produced unexpected data that challenged existing scientific models.
To my critics, my account is distorted by feminist bias. Pointing out bias becomes, in itself, evidence of bias
Hrdy’s observations with langurs also resonate with more recent critical scrutiny of Bateman’s work with fruit flies. As the evolutionary biologists Zuleyma Tang-Martinez and Brandt Ryder note, Bateman’s conclusion that only males benefit from promiscuity applied to just two of his six series of experiments. The other four series showed the same beneficial pattern for females, albeit to a weaker extent. Yet Bateman focused on the results that fit the polarised notion of competitive males and choosy females. This selective emphasis was then perpetuated by others, meaning that the unexpected benefits to females of multiple mates gained no traction in the literature. Analysing the last two series separately was a decision, part of the construction of a scientific finding, made for reasons that remain unclear. A subsequent re-analysis of all of Bateman’s data, pooled together, led to a finding of sex similarity in the effects of multiple mates on reproductive success. The researchers Patricia Gowaty and Brian Snyder, concluded that:
[T]here is no serious statistical basis in Bateman’s data for his conclusion that the reproductive success of females does not increase with the numbers of mates females have.
There have been many more challenges to bedrock assumptions, as I describe in Testosterone Rex. These include a proper accounting of reproductive costs for males, which go well beyond the single sperm required for conception. There are also species in which sex roles change in response to social and ecological conditions, such as the hedge sparrows that can wind up in a variety of different sexual arrangements. And, of course, we mustn’t forget humans’ own spectacularly inefficient means of reproduction. More than any other species, we have sex in ways and at times that won’t lead to conception; we should therefore be especially careful not to overestimate the reproductive advantages of promiscuous sexual behaviour for men. Look at mean sex differences in interest in casual sex, and you do indeed see a decisive male/female gap. An alternative characterisation of the data, however, yields a no less important conclusion – that for men and women alike, the vast majority prefer to be in a sexually exclusive relationship.
Yet in the eyes of my critics, my account is distorted by feminist bias. In line with the narrative of sex difference denialism, the act of simply pointing out bias becomes, in itself, evidence of bias. So according to Buss, von Hippel and co-author George Richardson, my work poses ‘a fundamental threat to the integrity of the scientific enterprise’. As a supposed demonstration of just how badly I misrepresent the evidence, they referred their readers to an article led by the evolutionary biologist Tim Janicke titled ‘Darwinian Sex Roles Confirmed Across the Animal Kingdom’ (2016). Yet their findings and conclusions, based on a meta-analysis of 66 animal species, are remarkably resonant with the case made in Testosterone Rex. I note that ‘contemporary research has identified many species to which Bateman’s principles do appear to apply’; that greater reproductive variance in men is seen in some human populations, but not all; and that ‘Bateman’s principles aren’t obsolete, but nor are they omnipotent and omnipresent’, due to the many additional factors that ‘enter the mix’.
Meanwhile, the meta-analysis reveals diversity in sex roles (as well as a significant proportion of species in which no sex differences in sexual selection pressure were found), and the researchers acknowledge that the ‘exceptions to the rule highlight the importance of incorporating environmental conditions when interpreting animal mating systems’ and that there ‘is increasing evidence that sexual selection varies not only between but also within species as a function of demographic and ecological factors’. Indeed, the studies of sexual selection in humans included in their analysis illustrate this very point. It seems that, even if we are trading in scientific turnips, they take on a quite different appearance depending on who is dishing them up.
A second ‘lens of gender’ is biological essentialism, the default attribution of sex differences to immutable biological causes. This hidden assumption has plenty to answer for in the scientific understanding of how creatures inherit evolved sex differences in brain and behaviour.
More than 30 years ago, the psychobiologist Celia Moore set herself the empirical mission of opening up the developmental black box connecting early hormonal differences between the sexes on the one hand, and later sex differences in brain and sexual behaviour on the other. The longstanding assumption held that there was only a direct link – testosterone turned on genes that masculinised the brains of males. But Moore was trained in a tradition that took an organism’s particular life trajectory and its role in evolution seriously. She had documented that rat mothers (dams) treat male and female pups differently, spending more time licking the anogenital region of the males (an activity that, for rats, is a form of care). She discovered that this was because the dams are attracted to the higher concentration of testosterone in the urine of males (a sex difference). Remarkably, it transpired that this difference in maternal care contributes to sexual differentiation of specific brain regions, and the efficiency and effectiveness of males’ later sexual behaviour.
Moore rightly considered these findings ‘exciting’. They showed that this feature of maternal care, an experience a baby rat could take for granted in the normal course of events, can contribute to the development of the neural basis of evolved masculine sexual behaviour. As Moore later put it, following many other illustrative (non-reproductive) examples, such experiences are ‘a completely reliable part of [a developing organism’s] environment’ that are ‘inherited as surely as its genome’.
Moore refers to her scientific standpoint as that of ‘development from’, in which the aim is ‘to unravel the developmental process, following threads to find out how the behaving organism was put together’. She contrasts this with the ‘development toward’ approach, which starts with the endpoint, then tries to ‘search backward for the “difference that made a difference”’. Moore’s work, by taking the unusual step of asking a development-from question about sex differences in brain and behaviour, revealed an additional indirect route by which sex can get its developmental work done. The tangible effects of sex on a developing individual – from the effect on urine odour in rats, to the relentless gender socialisation that follows identification of a human baby’s sex – shape the experiences that further shape brain and behaviour. They help co-construct the organism’s developmental pathway.
Yet development-toward thinking remains common in sex differences research. Take, for example, a much-discussed study from Simon Baron-Cohen’s lab at the University of Cambridge, published in 2000. In this study, babies just a few hours to days old were offered, in turn, a real face to look at (that of the first author, a masters student), and an odd-looking mobile made up of a composite of her facial features. The researchers measured how long each baby looked at the real face and, in a separate trial, how long at the mobile. This allowed them to make statistical comparisons between the behaviour of the male and female babies. The researchers’ primary analysis involved calculating the difference in proportion of time spent looking at the face and the mobile. They then used this ‘difference score’ to categorise babies as showing either a ‘face preference’, ‘mobile preference’ or ‘no preference’. Statistical testing found that more boys than girls fell into the ‘mobile preference’ category. Other analyses found that girls spent, on average, longer looking at the face than at the mobile, while boys spent a similar time looking at both. No analysis found a sex difference in time spent looking at the face. Nonetheless, the authors conclude that their observations ‘demonstrate beyond reasonable doubt’ that female superiority in sociability is ‘in part, biological in origin’.
The experimenter’s behaviour might have introduced bias, unintentionally, by influencing the babies’ responses
Their claim to have found a ‘biological origin’ for male/female differences in sociability falls firmly in the backward-looking development-to approach. As the Australian philosopher Neil Levy observes of this claim:
Socialisation is not a pattern of interferences, which obscures the underlying design of the infant: it is an essential part of the process whereby the phenotype is completed.
But note, also, the confidence with which Baron-Cohen’s team draw their conclusion. By implying the discovery of a fact as indisputable as a turnip pulled from the ground, it seems to follow that disagreement can only be politically motivated. As the evolutionary biologist Richard Dawkins Tweeted in 2019:
How is it still possible for social scientists to deny scientific evidence for genetic sex differences in interests? M[ales] statistically more interested in things & how they work, F[emales] statistically more in people. Baron-Cohen on newborns especially convincing.
But, as with any scientific conclusion, this claim is constructed out of layers of assumptions. How did the researchers get from an absence of any sex difference in time spent looking at the face, to a definitive claim about sex difference in social perception? Bem’s lens of gender polarisation, which implicitly frames the sexes as the ‘opposites’ of each other, offers insight. Through this lens, more interest in the mobile must imply less interest in faces – even if this is not, in fact, the case. This same implicit assumption underlies Baron-Cohen’s hypothesis that brains can be aligned along a continuum of ‘types’ that pass through ‘empathising’ (or ‘the female brain’) to ‘systemising’ (‘the male brain’), via the ‘balanced’ brain type. But in adults, at least, higher self-reported interest and skill in understanding systems doesn’t imply lower self-reported interest and skill in empathising, or vice versa. Translated to the current context, a baby with a clear mobile preference might have spent five seconds looking at the face, or 50.
Background assumptions are also involved in interpreting the meaning of babies’ interest in the two kinds of stimuli. The development of social cognition is a lengthy, complex process and newborns have some interesting quirks when it comes to face processing. These include an apparent indifference to whether the features are in the right place or not, so long as the arrangement is top-heavy, as well as to whether the face belongs to a human or a monkey. As for the assumption that neonatal interest portends later developmental superiority, it’s perhaps humbling to consider that two-day-old chimpanzees’ orientation to social stimuli – including human speech – is superior to that shown by human babies of the same age. The inference that the amount of time spent looking at a face foretells future sociability is ‘essentially unargued for’ and ‘questionable at best’, as Levy said of this study. The same caution also applies to relations between visual interest in the mobile, and future mechanical perception.
Each of these background assumptions could be put to an empirical test. In the meantime, all we can say is that any that are incorrect might introduce error into the final claim. So, too, could all the more familiar contributors to error – confounding factors and measurement error, among others – of which the newborn study had its share. The most noteworthy issue was that the first author – whose face served as one stimulus, and who held up the mobile – was not always blind to the babies’ sex. This raises the concern that the experimenter’s behaviour might have systematically introduced bias, however unintentionally, by influencing the babies’ responses. (Although those rating the babies’ behaviour from a video recording were blind to sex, this shuts the stable door after the horse has bolted.) It’s for this reason that a later study of sex differences in newborn looking preferences went to great efforts to make sure that the people interacting with the babies didn’t know their sex.
It’s also worth pointing out that another study – by Baron-Cohen’s research team, and apparently with the same group of newborns – tested the hypothesis that the babies would prefer looking at a face displaying eye gaze, compared with a face with closed eyes. Baron-Cohen has hypothesised that this orientation to eye gaze is an important foundation for the development of social cognition. According to the theory that females tend to be predisposed to be social-cognition specialists, one might expect a stronger preference for eye gaze in baby girls. However, no such difference was found.
Baron-Cohen has said that criticism of his sex differences work is really about prioritising politics over science. But clearly, to raise such questions isn’t to put politics before evidence. These questions are all about the evidence and its interpretation.
There’s an irony to all this rather tedious talk about how ‘my’ camp treats the relationship between science, evidence and politics. Those with whom I disagree would be rightly outraged if I suggested that the way they interpreted evidence could be chalked up to a political preference for patriarchy, or because they have spent too much time at dinner parties listening to woke progressives talking about social constructions. I suppose I should find it flattering that my mental states are the subject of such fascination, but I don’t reciprocate the interest.
Those who go on ad hominem attacks might genuinely believe that they’re helping to defend scientific objectivity. In fact, they undermine it. Many philosophers of science consider openness to critical debate to be the foundation of scientific objectivity, perhaps most notably Helen Longino. She considers objectivity to be something that emerges from the collective endeavours of an intellectual community. This idea hails from the 19th-century philosopher John Stuart Mill’s arguments for freedom of speech. Our opponents – regardless of whether they hold all the truth, some of the truth or are indeed mistaken – challenge us to better understand our own case, said Mill. As the philosopher Heather Douglas writes in relation to science:
The hope is that by keeping scientific discourse open to scrutiny, the most idiosyncratic biases and blinders can be eliminated. In this way, other people help to make sure you are not seeing something just because you want to.
‘Biases and blinders’ aren’t psychological phenomena that exclusively afflict female academics taking aim at research claiming to show the biological origins of occupational gender gaps. Being the progenitor or disciple of a particular theory, holding intellectual capital built off the back of a particular set of background assumptions, being steeped within a school of thought in which certain assumptions are so axiomatic that they’re all but invisible – these can all incline a scientist to interpret findings in a certain way. Whether for reasons self-serving or benign, everyone comes laden with prior knowledge, background assumptions and frameworks. That’s why it takes a diverse village, so to speak, to nurture scientific objectivity.
It doesn’t rest solely on the shoulders of individual scientists to maintain objectivity
This nurturing isn’t necessarily gentle. Longino calls it ‘transformative criticism’: a term for the back-and-forth of critical argument, the unearthing of where disagreements lie, and the subsequent adjustments or counterarguments. In her account, scientific objectivity depends not simply on scientists being coolly detached with respect to their data, but ‘upon the depth and scope of the transformative interrogation that occurs in any given scientific community’.
The important philosophical idea here is that it doesn’t rest solely on the shoulders of individual scientists to maintain objectivity. It’s also the job of the collective to hold itself, and its members, to account. This makes clear why dismissing experts who disagree with our scientific conclusions – she puts politics before science; they think women and men can’t be equal if they aren’t the same – is best avoided. As the historian Naomi Oreskes recently put it in her book Why Trust Science? (2019):
Diversity does not heal all epistemic ills, but ceteris paribus a diverse community that embraces criticism is more likely to detect and correct error than a homogeneous and self-satisfied one.
Embracing criticism means engaging, seriously and in good faith, with arguments and evidence that challenge us. Documenting sexist, racist, cis-sexist or ‘blank-slatist’ biases in science is fine – as is disagreeing with the merit of such arguments, if you interrogate them and the evidence behind them in turn. But if you’re not a fan of seeing scientists’ views attributed to personal racism, transphobia or misogyny, then nor should you tolerate commentary about scientists supposedly being biased by their feminist motives.
It’s quite possible to distinguish transformative criticism from the kinds of commentary and condemnation that doesn’t further scientific objectivity. This latter category is marked by persistent unresponsiveness to counterarguments or new evidence, stubborn misrepresentation of opposing views, and a failure to genuinely engage with them. It also includes behaviour that effectively prevents or dissuades others from engaging with opposing viewpoints – such as various forms of deplatforming, or branding work untrustworthy by accusing it of being politically motivated. As Oreskes argued, the more diverse and open we are in our support of free debate, the more objectivity we should be able to achieve – ‘as individual biases and background assumptions are “outed”, as it were, by the community’.
When it comes to politically loaded scientific debates, the Australian philosopher Russell Blackford got it right: ‘We need to focus on evidence and arguments, and on ordinary fairness and compassion to others, even when we disagree.’ Admittedly it’s not always easy to meet these norms. They demand of us fairness and charity toward those whose ideas we might consider idiotic and harmful, or who might have shown themselves unwilling to extend the same generosity to ourselves and our academic tribe.
But if there’s one thing we can be sure of when it comes to human sex differences in the brain and behaviour, it’s that nobody has got it completely right. Whether we are most concerned about bias in scientific investigations of marginalised groups, or worried about the undermining of science by political values, the remedy is the same. As academic authors, editors, reviewers and colleagues, we should support norms and practices that help to create a scientific community that’s diverse, open and free – and insist that those same norms and standards be applied not just to those whose views we oppose, but also to those with whom we agree.
A simple first step? Don’t shoot down books you haven’t read.
