By normal scientific standards, the hypothesis that a woman is an adult human female is the natural default answer to the question “what is a woman?”. Call that the sex-based account of woman, since it explains being a woman in terms of being female. The sex-based account of woman is simple and informative. It is stated in perspicuous and independently well-understood terms. “Female” in particular, along with its counterpart “male”, is among the central theoretical terms of evolutionary biology, one of the most successful branches of modern science. Indeed, few terms for living things are under as good theoretical control as “male” and “female”.

The sex-based account of woman also fits well with the evidence: everything we independently know about who is a woman and who is not. Uncontroversially, almost all adult human females are women and almost all women are adult human females. That observation generalises to arbitrary counterfactual circumstances. No matter which people there had been, it still would been true that almost all adult human females were women and that almost all women were adult human females.

Of course, there are some tough cases for the sex-based account of woman, and in particular for its consequence that there are no male women. For example, adult human males with complete androgen insensitivity syndrome (CAIS) look to the naked eye just like adult human females. Reliance on the natural and highly reliable method of assessing whether someone is a woman by whether they look like an adult human female encourages classifying such males as women. In a different kind of case, some adult human males with uniformly male-typical primary and secondary sex characteristics say that they are women. Reliance on the natural and highly reliable method of assessing whether someone is a woman by whether they say that they are a woman similarly encourages classifying such males as women.  

However, such tough cases for the sex-based account of woman do not undermine its status as the default view. For normal scientific practice is to tough it out in tough cases. That is because there are anomalies for any scientific theory: cases on which the natural verdict is in tension with what the theory would lead one to expect. Sometimes, those anomalies are indeed clues for how to go deeper. For instance, famously, the anomalies for the geocentric model of the universe eventually led to its replacement by the heliocentric model. But often enough, such anomalies merely reflect noise in the data, such as measurement error or some other bias. Given that there are always anomalies, and that anomalies are often enough mere noise, scientists do not lose confidence in a theory just because it has some unwelcome consequences.

For example, consider the heliocentric model itself. From its perspective, the appearance that the sun revolves around the earth is a kind of anomaly. Superficially, it looks as if the sun orbits the earth (just as the moon actually does), not the other way round. One way in which defenders of the heliocentric model made progress was to show how that appearance is unsurprising if the heliocentric model accurately describes the universe. Today, the heliocentric model is so familiar that it can require some effort to recover a feeling for the force of the appearances against it. Thus anomalies for a theory do not always motivate even minimal revisions of the theory. In the long run, they may be satisfyingly resolved without any change of paradigm.

Moreover, the anomalies for the geocentric model of the universe are only one part of the story of its eventual rejection. Another important factor was what happened when defenders of the geocentric model tried to accommodate the data. To restore fit with the observed positions of celestial entities, they were notoriously forced to add epicycles to the model, extra loops describing the motion of the sun and planets. The geocentric model took on a staggering complexity. The initially simple and attractive picture of the earth as centre of the universe became increasingly ugly and baroque, with no end of the multiplication of complexities in sight.

On a popular stereotype of the Copernican revolution, defenders of the geocentric model were guilty of blithely dimissing the data, dogmatically sticking to a too-simple picture even as violations of its predictions piled up. That popular stereotype in turn suggests a crudely empiricist moral: scientists must be willing to revise their theoretical preconceptions whenever new data come in. But on a more accurate account, the ad hoc and undisciplined way in which defenders of the geocentric model were led to add more and more epicycles—to complicate more and more—was more decisive evidence of underlying error. After all, through the addition of epicycles, the geocentric model could formally be made to fit any new data. Thus in contemporary philosophy of science, a common moral of the Copernican revolution is quite different from the crudely empiricist one. It is: beware epicycles!

**

Remarkably, the sex-based account of woman received its first explicit defence in mainstream philosophy only in 2020. The culprit was Alex Byrne, a well-known and well-respected philosopher of mind at MIT. In his paper “Are women adult human females?”, Byrne arranged some basic considerations in support of the sex-based account of woman and responded to objections from the tough cases identified above. Several of Byrne’s positive arguments for the sex-based account were in effect from its fit with the evidence. Others were based in a robust form of sex realism, on which facts about sex play an important role in human thought and action, as they do in the thought and action of many other species.

In particular, given sex realism, it is to be expected that natural languages have common general terms for human females and males, which enable the expression of information about sex in humans. But presumably, if natural languages have any common general words for sex in humans, then “woman” is one of them, along with its counterparts “man”, “girl” and “boy” in English, synonyms “femme” in French and “mujer” in Spanish, and so on. Thus if the sex-based account of woman fails and “woman” does not mark a sex-based distinction, there risks being no perspicuous general-purpose means in natural languages for expressing information about sex in humans. Hence sex realism directly motivates the sex-based account of woman.

By the same reasoning, sex realism directly motivates sex-based accounts of the categories of man, girl and boy, on which a man is an adult human male, a girl is a human female child and a boy is a human male child, respectively. On a popular minimal characterisation of what unifies the categories of woman, man, girl and boy, they are all gender categories, whether or not they are also sex categories. Thus sex realism motivates a sex-based account of gender, on which some central phenomena associated with the term “gender” are to be explained in terms of sex. More abstractly, given a robustly sex realist framework, a natural conjecture is that gender is the same as sex, so that talking about gender amounts to another means of talking about sex. The difference between the terms “sex” and “gender” is not in what they mean, but in their cognitive associations: “sex” is more overtly biological.

The response to Byrne’s defence of the sex-based account of woman by orthodox philosophers of sex and gender was unsympathetic. Undeterred, Byrne has now written a monograph Trouble with Gender, in effect an exposition and defence of sex realism and the sex-based account of gender for a general audience. It will not convince his critics; little could. Nevertheless, Trouble with Gender is an excellent work of public philosophy. It is clear and rigorous, despite largely avoiding formal notation. It is relevant to questions that organically arise in everyday life, while remaining focused on their general aspects. And it makes use of ideas from the natural and social sciences, especially biology and psychology, without denigrating the philosophical questions it addresses as derivative or second-rate.

More generally, Byrne keeps things simple. He avoids the dangerous temptation to plunge straight into the nuances in the naïve hope that the basics will take care of themselves. Such an approach is crucially enabled by the nature of the subject matter of Trouble with Gender itself. For as Byrne discusses, the distinction between female and male has a striking track record of independent confirmation. It has survived the test of systematic biological enquiry. More specifically, whereas many superficially analogous ways of classifying organisms have been substantially displaced in biologists’ theoretical vocabulary by more explanatorily basic distinctions, the classification of organisms into female and male has proved remarkably robust.

For instance, many distinctions between species are superficially striking, but do not correspond to anything biologically fundamental: species differences are not at the heart of modern evolutionary theory. By contrast, at least at the level of phenotype, the difference between female and male is among the most central in evolutionary developmental biology. Indeed, the distinction between female and male features in a common working definition of “species”, according to which (roughly speaking) some organisms are members of the same species just in case if one is female and the other is male, they will be able to have fertile offspring.

Given the explanatory significance of the difference between female and male across much of the biological kingdom, it is only natural to attempt to understand as much as possible about human life in terms of it. Of course, there are many interesting intersex similarities, as well as many interesting intrasex differences, in humans just as much as in other species. It would be foolish to expect a full explanation of all interesting patterns in human life in terms of female and male. But that minimal anti-reductionist point does nothing to undermine the naturalness of a sex-based approach to understanding some basic aspects of human life.

For example, consider a quite mundane generalisation like the following: almost always, men are sexually attracted to women and women to men. Such patterns of sexual attraction are so familiar that they might seem in no need of explanation. But what is the non-sex-based explanation of why almost always, men are attracted to women and women to men? Sceptics of the sex-based account of gender may be surprised to find how much of human life becomes puzzling if they are disciplined in their repudiation of sex-based explanations, and do not oscillate between explicitly disavowing sex-based reasoning and implicitly relying on it.

**

The first three chapters of Trouble with Gender concern the equation of gender with sex, the nature of female and male, and the sex-based account of woman, respectively. Much of the work in these chapters consists in carefully distinguishing between different levels of the phenomena under discussion. For instance, to understand why it is uncontroversial in biology that sex is binary, sex must be distinguished from mechanisms for determining sex, such as possessing certain chromosomes, as well as from sex characteristics, such as having certain genitalia. In brief, the crucial point is that such mechanisms and characteristics vary across and even within species, while sex remains the same.  

The importance of not over-prioritising a few salient cases is an implicit theme of Byrne’s book. The specific bias at issue is clearest in the case of theorising about the general categories of female and male, where the obvious danger is of over-prioritising the case of human females and males: most straightforwardly, by treating features characteristic of humans of a given sex as necessary for being a member of that sex. Byrne shows how the temptation to build specific features of females and males of the human species into the general definitions of “female” and “male” depends on a failure to appreciate how robust the sex binary is across the biological kingdom and how variously it is realised in different species. Another, explicit theme of his book is that, at least when it comes to sex, humans are unexceptional.

Analogous points apply to theorising about the category of woman, where the more heterogenous danger is of over-prioritising cases of high social-political salience. In a reflection of that danger, the most historically prominent alternatives to the sex-based account of woman have tended to build reference to specific social statuses into the definition of “woman”. The issue is that the specific social statuses had by women vary across and within actual and possible societies, and have often coincided with the social statuses had by some men. To avoid gross reductionism, defenders of social alternatives to the sex-based account of woman have tended to define “woman” in terms of some logically complex combination of specific social statuses. But recall the earlier methodological moral: beware epicycles!

A related challenge for social accounts of woman is this. As Byrne emphasises, the answer to the question “what is a woman?” should generalise naturally to the question “what is a girl?”, since the difference between women and girls is just that women are adults and girls are children. The sex-based account of woman of course generalises to the category of girl with ease. But what specific social status do all girls have, including infant girls (who may yet die in infancy)? Even if the difficulty can be met in principle, the fact that social accounts of woman struggle to handle the simple test case of infant girls is evidence that they are starting in the wrong place.

Could it similarly be objected that the sex-based account of woman results from illicitly prioritising paradigm cases of women? Given the sex-based account, there are as many ways of being a woman as there are of being an adult human female, but perhaps the ways of being a woman are more various still.

If the case for the sex-based account of woman reduced to its fit with the evidence, such an objection would be compelling. However, the case for the sex-based account of woman does not reduce to its fit with the evidence. The sex-based account is also distinctively simple: it avoids ad hoc epicycles. Its minimum of complexity is evidence that it has not been invidiously gerrymandered to fit biased data. That is why it has methodological priority, and would retain such priority even if some alternative could be constructed which fit the data equally well. As a corollary, even if the tough cases for the sex-based account turned out to be somewhat more prevalent than had initially been realised, to lose confidence that a woman is an adult human female in the absence of any comparably simple alternative would be a failure of nerve.

**

Trouble with Gender is proof that Byrne’s nerve has not failed him. He is one of few philosophers who can exhibit such proof (other monographs that take a broadly sex-realist approach include Material Girls by Kathleen Stock and Gender-Critical Feminism by Holly Lawford-Smith). According to orthodoxy in the philosophy of sex and gender, the sex-based account of woman is not a live option. It is not even allowed as a good approximation to the truth. Rather, it is widely treated as unacceptable to rely on even as a simplifying assumption. From an outsider’s perspective, it is as if physicists had collectively adopted a ban on ever explicitly invoking Newton’s first law of motion. As would probably happen in the physical case, much orthodox theorising about sex and gender thrives on implicitly reasoning in a way which subverts the ban.

Outside of mainstream philosophy of sex and gender, the situation is better, but not by much. Many philosophers privately recognise that the sex-based account of woman is the natural default hypothesis and are yet to be persuaded of the alleged counterexamples to it. However, almost none is willing publicly to challenge the orthodox rejection of the sex-based account. Tenured professors have proved at least as shy about speaking out as more junior members of the profession; among senior faculty in particular, Byrne is almost unique in his vocal defence of sex realism. When issues in the vicinity arise professionally, some philosophers are prepared to challenge publicly the de facto taboo on defending the sex-based account of woman, on the basis of generic principles of freedom of enquiry and related professional norms. A standard justification of liberal norms on enquiry is that restrictions on what can be expressed risk inadvertently ruling out the expression of knowledge. It is often felt safest not to mention that the case of the sex-based account of woman is plausibly exactly of the kind for which, given that justification, such liberal norms are most needed. 

An important related consideration is that even well-informed and well-intentioned people get things wrong. In practice, for any given philosophical question, no matter how consequential, a serious and intelligent thinker can get the answer wrong, while getting the answer to other difficult philosophical questions right. It may even be that the very same dispositions which enable them to get those other questions right are the source of their error: like how someone on the look-out for poisonous berries is at once likeliest to spot the genuinely poisonous ones but misidentify the only apparently poisonous ones. Thus an intellectual community which imposed agreement on any specific question as a requirement for participation would be impoverished in the long run, even if it happened to be right about the very case at issue.

In particular, even if the sex-based account of woman is ultimately wrong, there is a live danger for the philosophical community in raising the costs on philosophers sympathetic to it. For one salient route to accepting the sex-based account is through an aversion to complexity, either instinctive or acquired after exposure to other instances of theorising distorted by a tolerance for epicycles. Arguably, philosophy needs more philosophers with a robust reluctance to over-complicate, even if such a disposition sometimes leads them to over-simplify. Unfortunately, in the present climate, it risks getting fewer, though those philosophers best able to keep their interests under control—and away from social philosophy—should be safe.

Although in the robustness of his defence of sex realism, Byrne is an exception among philosophers, philosophers are not themselves an exception among members of their social-political class, broadly construed. Whatever they might say to each other in private or anonymously online, most people interested in succeeding in politically progressive environments have taken care not blatantly to violate the taboo on using “woman” as a sex-based term.

The reason that such care has so widely been felt necessary is the recent salience of adult males who say that they are women, and who there is intense social-political pressure to take at their word. Crucially, some such adult males are fully sex-typical, in roughly the sense that by almost all interpersonally accessible criteria, they are indiscernible from typical adult human males. For instance, as well as being adult males, they look and sound like adult males, have the usual male genitalia and male levels of testosterone, are exclusively sexually attracted to females, and grew up being perceived and treated by everyone around them as male.

Of course, not every adult male who claims to be a woman is sex-typical in all those ways: for instance, males with CAIS often grow up perceiving themselves as female. But such cases are misleading, since the relevant social-political pressure extends to the avowals of some adult males who are fully sex-typical, so militates even against fallback versions of the sex-based account of woman. Indeed, it also militates against classical social accounts of woman, which invoke sex-based forms of oppression. The result has been the widespread adoption of a form of radical internalism about the category of woman, on which whether one is a woman is determined by one’s internal psychological states.

Attempts to develop such a view in detail run into similar problems to those faced by social alternatives to the sex-based account of woman: for example, in requiring dramatic complications to handle simple test cases like infant girls or mentally disabled women. Hence for present purposes, it will be more instructive to consider the fashionable slogan which embodies the pressure to take adult males who say that they are women at their word: “a woman is anyone who says that they are a woman”. When pressed on the possibility of insincerity, those who propound such a slogan may retreat to the fallback: “a woman is anyone who sincerely says that they are a woman”. In practice, however, that qualification fails to make a difference, since no independent mechanism is provided for assessing sincerity. Thus the only interpersonally available criterion proposed for telling whether or not someone is a woman remains “Do they say that they are a woman?”. That is in effect to adopt a radically impoverished epistemology of the category of woman.

Such an epistemology is just as radically misconceived. For reliance on self-avowals of womanhood is far from our only means of assessing who is a woman; indeed, it is hard to see how it could be. To the contrary, independently of the exact details of what “woman” means, it is plausible that the primary way for ordinary humans to assess whether someone is a woman is by whether that person looks and sounds like an adult human female: that is, by exploiting perceptual recognitional capacities for sex (and sexual maturity) in humans. Such recognitional capacities are nearly perfectly reliable, extremely efficient and a standard feature of the human cognitive system, for unsurprising reasons: like other species, humans need to act on information about the sex of their conspecifics in order to reproduce.

Certainly, our recognitional capacities for sex in humans have no built-in special exception which disarms them in the case of fully sex-typical adult human males who say that they are women. Thus the situation we are really in when confronted with such adult males is this: do we trust what our recognitional capacities for sex in humans tell us about whether they are women—or what they themselves do? The simplicity of the sex-based account of woman, especially as compared with the psychological accounts tailor-built to vindicate the slogan “a woman is anyone who says that they are a woman”, helps settle which way to jump.

How could so simplistic a slogan have become so influential, in philosophy as elsewhere? Much of the story will likely have to do with broader social-political developments. But some of it may have to do with squeamishness about the activity of sex. After all, it is often not very nice to attend to the specificities of human male sexuality. Yet the temptation to look away from variant forms of male sexuality may underlie some of the difficulty in understanding why there are sex-typical heterosexual adult human males who not only say that they are women but sometimes make significant personal sacrifices to look the part. More specifically, it inhibits properly addressing a latent challenge for the sex-based account of woman: if such males are unequivocally men, then why do they want so much to be seen as women? For in salient cases, the candid answer turns out to be: because they are sexually aroused by simulating being female. For readers interested in coming to terms with one of the central pressures to forget about sex in philosophy and elsewhere, the basis for that sex-based diagnosis is presented, calmly, in Trouble with Gender.

Daniel Kodsi is completing a DPhil in Philosophy at Trinity College, Oxford, and will be a lecturer at Magdalen College, Oxford, for the 2024-5 academic year.