Alex Byrne contends that women are (simply) adult human females (calling this thesis “AHF”; Byrne Reference Byrne2020; see also Stock Reference Stock2021, chap. 5). Athough Byrne's argument has already been responded to at length (Dembroff Reference Dembroff2020a; Heartsilver Reference Heartsilver2021), it can be refuted much more straightforwardly, and in a manner that establishes several important broader philosophical lessons. Byrne's argument for AHF is simple. He argues that AHF conforms to the dictionary definition of woman; that one would expect English to have a word that picks out the category of adult human female, and woman is the only candidate; that AHF best explains how we know that an individual is a woman; that AHF predicts the right verdicts in gender-role reversals; and that woman and female are used interchangeably, even in hyperintensional contexts (Byrne Reference Byrne2020, sec. 2). After arguing this, Byrne concludes, “The foregoing makes a strong cumulative case, and AHF is of considerably greater initial appeal than the JTB analysis of knowledge” (3891). This is ironic in two ways: (1) the JTB theory of knowledge is widely thought to have been refuted; and (2) AHF can be refuted in the same way: through simple counterexamples.

This article shows that in ordinary usage, the literal meaning of female is ambiguous, having both a biological interpretation (classifying individuals by reproductive biology) and a gendered interpretation, referring to individuals who satisfy socially constructed gender norms (such as norms for gendered traits and performance, as well as self-identification). In the process, I also show that adult is similarly ambiguous between a biological reading and social interpretation—as an individual can be “fully grown” in a biological sense (namely, biological age) or in a social sense (namely, satisfying phenotypic and psycho-social norms for being “grown up” and socially classified as an adult). I then use these findings to provide clear counterexamples to AHF and to variants of AHF that Byrne might propose in response. In each case, my counterexamples show the same thing: that according to ordinary literal usage, woman is not interchangeable with any variation of adult female except when adult female is specifically disambiguated in a gendered sense. Thus, AHF is false. Women are not simply adult (human) females. Women are adults who are gendered female. Lessons are then drawn, among them that vastly more types of individuals—including trans women, alien women, robot women, and beyond—are women than Byrne allows.

In their article replying to Byrne (Dembroff Reference Dembroff2020a, sec. 2.2.), Dembroff rightly notes (among other things) that Byrne bases his argument on cherry-picked examples. As Dembroff points out, there are “entire linguistic communities where ‘woman’ is explicitly not used to mean ‘adult human female’” (7)—at least not as Byrne interprets “adult human female,” which is in a reductive biological sense (see Byrne Reference Byrne2020, 3785, 3787–88, and sec. 3). However, things are actually much worse for Byrne than Dembroff contends. No competent user of English uses woman to refer (simply) to adult human females, and there are clear counterexamples to AHF. How do I know this? For one thing, Byrne himself notes that woman and female are normally used interchangeably simpliciter when used to refer to adults, neglecting here the “human” component of AHF (sec. 2.6). Moreover, this aspect of AHF is plainly incorrect. As any Star Trek fan knows, in the Star Trek universe there are many planetary species of women: human women, Klingon women, Vulcan women, Cardassian women, and many other species of women besides. Further, if one Googles “Cardassian women,” picture after picture of Cardassian women appear in the search results. So, according to everyday English language usage, Cardassian women are women—though they are clearly not human. So, AHF is false. Woman does not mean adult human female. If it did, then the sentence “Cardassian women have bony facial protrusions”—a sentence that not only makes sense, but is demonstrably true—would assert the nonsensical proposition that Cardassian adult female humans have bony facial protrusions. But no competent user of the English language would take the aforementioned sentence to mean this. Women need not be human to be women.

Byrne might say that he misrepresented his real position here, claiming that he really meant to say that women are (simply) adult females of any humanoid species. Call this thesis AFHS. Alas, AFHS is false too. How do I know this? Well, in the extended Star Wars universe, there are adult female members of the Hutt species—a species of giant slug-like creatures (Star Wars Fandom n.d.). These Hutts are clearly women Hutts. So, something need not be an adult humanoid female to be a woman. To this, Byrne might say that women are simply adult females (that is, of any species). Call this thesis SAF. Alas, SAF is not clearly true. Adult female dung beetles clearly exist—though it is highly doubtful that they are women. Why? The most natural answer, of course, is that they are not members of an advanced species with complex and well-developed gender norms dictating who is a woman—and indeed, who is gendered female—and who is not. Indeed, we can see how plausible this answer is, first, by considering that although no one would normally think it linguistically appropriate or accurate to talk about “women dung beetles,” there is one context in which we would think it appropriate: in cartoons where dung beetles are gendered female in a social sense of broadly conforming to feminine gender norms. For example, in the 1998 animated film, Antz, one of the main characters, Azteca, is clearly identifiable as a woman ant—indeed, as the girlfriend of the male ant, Z. We wouldn't identify actual adult female ants as women, but we do in cartoons. Why? The answer is plain: women ants in cartoons are gendered female, in the sense that they satisfy feminine gender norms.

In a recent response to Byrne, Maggie Heartsilver makes a similar point: there is a non-reproductive sense of female according to which to be female is not to be a member of the female biological sex, but rather to be socially and/or legally classified as female, where this is a non-biological classification (Heartsilver Reference Heartsilver2021, 6–7). And indeed, given that Byrne invokes dictionary definitions, it is worth noting here that the Merriam-Webster dictionary gives the following literal definition for “female”:

As we will see throughout this article, this multiple definition clearly corresponds to the ordinary literal usage, as competent speakers standardly use the term in both ways, and in ways that relate in crucial ways to the meaning of woman. To begin to see how, consider the fact that, later in his article, Byrne makes it clear that he means to deny that trans women are women because he thinks they are not female (Byrne Reference Byrne2020, sec. 3.3). Indeed, he asserts that the proposition that trans women are not female is “no doubt true” (3800). Now, Byrne is certainly right than trans women are not female in sense 1(a)(1): trans women do not (ordinarily) have bodies that produce eggs—though it is worth noting that this could plausibly change with future scientific advances (Maron Reference Maron2016). But, in any case, as we have already seen in the cases of dung beetles and cartoon characters, this is just not the sense of female relevant to whether someone is a woman. Woman refers to adults who are female in sense 1(b): adults who are gendered female. But, in this sense, Byrne's claim that trans women are not female is clearly false: trans women are gendered female. How do I know this? Because robot women are gendered female. If one Googles “robot women,” the following kinds of results appear: “Why female sex robots are more dangerous than you think” (Gee Reference Gee2017), “Meet Sophia, the female humanoid robot and newest SXSW celebrity” (Raymundo Reference Raymundo2016), and “Gynoids are humanoid robots that are gendered feminine. They appear widely in science fiction film and art. They are also known as female androids”.Footnote ² And indeed, it's usually easy to identify female gendered androids, at least once one finds out that they are robots: Dolores in the HBO series Westworld is one; so is Ava in the 2014 film Ex Machina.

Are robot women like these biologically female? Well, they are definitely not in the reductive sense that (some) biologists understand this notion, which is in terms of how species “package their DNA into ‘gametes’” (Griffiths Reference Griffiths2020), and/or in terms of more complex biological features (such as “sex chromosomes” and so on; see Bachtrog et al. Reference Bachtrog, Mank, Peichel, Kirkpatrick, Otto, Ashman and Hahn2014). Neither of the robot women whom I just referred to are depicted as having DNA, let alone gamete-producing organs. But, for all that, they are female. How do I know this? For one thing, if you were to ask any competent user of the English language who happens also to be a fan of Westworld, “Who is the lead female android in the series?”, they would respond, “Delores” (okay, maybe they would say “Maeve,” and they wouldn't be far off). Further, if you were to point out that Delores and Maeve are not biologically female (they are robots!), any competent English speaker would presumably agree with that qualification. But notice: it is a qualification. It is, in line with the literal ambiguity of female reflected in the above dictionary definition and ordinary ways of speaking, informative if you tell me you are talking about women who are “biologically female” (such as cis human or Cardassian women), as compared to women who are non-biological females (such as women robots).

Byrne might respond that although women robots in Westworld are not biologically female, they are nevertheless physically or “phenotypically” female in having stereotypical female bodies (namely, “secondary sex characteristics”; see Richards and Hawley Reference Richards and Scott Hawley2011, chap. 9). Further, Byrne might maintain that these phenotypic traits are essential to someone being a woman, or being female, on both interpretations of female discussed herein (that is, the biological and gendered interpretations). However, this proposal is plainly false. Stereotypical physical features have nothing to do with being female in the gendered sense of female relevant to whether someone is a woman. How do I know this? In several ways. First, there are clear cases of girl and women robots and AI who lack anything like a stereotypical female body. For example, in the 2008 Disney film WALL-E, the eponymous protagonist falls in love with a girl robot, Eve. If you were to ask anyone (any adult or child) who has seen WALL-E, “Who is the female robot that WALL-E falls in love with in the movie?,” they would surely say “Eve.” And indeed, character descriptions of Eve clearly identify her as female—as “she” (Corliss Reference Corliss2008). Yet, Eve has nothing like a stereotypical female body. She is a mostly featureless white oval. Viewers identify her as female—as a girl robot—by virtue of her feminine gender features (that is, her name, ways of talking and behaving, and so on—social norms and expectations for identifying girls and women). Similarly, consider the female AI in the 2004 film I, Robot, named “VIKI.” As with Eve, any viewer of the film would surely identify her as a female AI (her name is VIKI (!), her voice is feminine, and so on). And indeed, like Eve, VIKI is universally recognized as female, that is, as “she” (Fandom n.d.). Yet VIKI has no physical “body” at all. She is a disembodied neural network who has, at most, a vaguely feminine-looking digital “face”—and, even before the viewer ever sees her “face,” one can already readily identify her as a female AI. Finally, in season 3 of Westworld, Delores uploads a copy of her mind—that is, herself—into a copy of Martin Connells's body. Yet, despite now existing in a stereotypical male body, upon learning that it is Delores in Connells's body, the viewer surely still recognizes Delores herself as female. Delores—a woman AI—has merely made it such that she now occupies a stereotypical male body. So, one can be (simply) female—in the gendered sense relevant to being a girl or woman—without having anything remotely like a stereotypical female body. What this shows is that, in contrast to how female is used to (biologically) identify female dung beetles (which, again, are very doubtfully women!), female as it is used to identify girls and women is a social category—one that tracks the socially-constructed notion of being gendered female: a property that cis human women, trans women, alien women, robot women, and disembodied-AI women all share.

Are women robots like Delores and Maeve adult females? Well, again, not in a biological sense. However, if you were to ask any competent language user to distinguish child robots from adult robots (or young girl robots from women ones), they easily would pick out Delores and Maeve as the latter—because they are obviously women robots. By a similar token, fans of the television series My Life as a Teenage Robot readily recognize its protagonist, Jenny Wakeman (also known as “XJ-9”), as a teenage robot girl (Teenage Robot n.d.). So, being an adult female, in the sense relevant to being a woman, isn't biological either: it is a social category that, among other things, enables us to distinguish women robots like Maeve from child robots. And indeed, this ambiguity is latent in standard dictionary definitions of “adult.” First, the Merriam-Webster dictionary defines adult as “fully grown and developed.” Yet, as we see in adult humans and adult robots, there are clearly different ways that a being can be “fully grown and developed.” One can be fully grown and developed in a purely biological sense (having developed biologically from a child), but also (or alternatively) in a social sense. Delores and Maeve, for example, are clearly fully grown and developed in the sense of having stereotypical adult bodies and socio-psychological profiles—and they are both very different in these respects from child robots or human children. Second, the Merriam-Webster definition of adult implicitly recognizes these complexities, adding on the one hand that adults can be understood as “a human being after an age (such as 21) specified by law,” but also, alternatively, in terms of being “mature and sensible; not childish” and “of, relating to, intended for, or befitting adults”—the latter two of which are clearly social classifications. Finally, as further evidence of this, Merriam-Webster gives “grown up” as a synonym for “adult,” defining “grown up” in turn as “not childish or immature”—which is a social (rather than biological) category.

But now if adult Cardassian females are women, and adult robot females (like Maeve and Delores) are women, then trans women are women. How do I know this? Because there is nothing that Cardassian women or robot women have relevant to being girls or women that trans girls and women don't have, and vice versa: they are all (simply) adults who are gendered female. First, as the case of robot women shows, biology and bodily phenotypes have nothing essential to do with being (simply) female, at least with how this concept is understood in relation to womanhood. Indeed, as Delores uploading herself into Martin Connells's body clearly shows, a female person (a woman!) can come to inhabit a body assigned the sex and gender-categories male at birth. Second, as all of these cases show together, female and woman are social categories that can be qualified in all kinds of ways (viz. race, hair color, cis, trans, carbon-based, silicon-based, and so on)—despite members of each identifiable subset ultimately being (simply) women. Third, although there clearly are some differences between trans women and the other kinds of women enumerated above—among them, that trans women were once assigned the sex and gender male at birth, whereas cisgender women were not—it is nevertheless the case that trans women are now gendered female: that is what makes them trans women. Finally, of course, when one chooses to identify subclasses of women in these ways—focusing, in particular, on treating some women fundamentally differently from others (such as whether we should treat Black women differently from White women, or trans women differently from cis women)—we can then ask what those categories are, whether they too are social constructions, and finally, the normative moral question of whether we should focus on those categories and understand them in one way rather than another. But these are all broader issues, far above and beyond Byrne's avowed focus, which is to determine what makes someone a woman. And the relevant points here are these: Byrne's AHF is false. Cis women, robot women, Cardassian women, and trans women are all (simply) women because they are all (simply) adults who are gendered female.

I foresee a number of potential objections. First, Byrne might contend that my argument runs into a fatal dilemma: namely, that my claim that women are adults who are gendered female is either unintelligible or uninformative. If, on the one hand, we treat female as referring to biological sex (a natural kind) and woman as a socially constructed gender category, then my thesis “women are adult gendered females” is incoherent. For, on this reading, the thesis is equivalent to “women (a social kind) are adult biological females (a natural biological kind)”—which is a contradiction. But, of course, I deny that this is how to properly disambiguate female in the context of the present debate. For, as we saw earlier, female is ambiguous between a biological interpretation and a gendered interpretation—and, as we have seen through clear examples, only the latter is relevant to whether someone is a woman. Women are adults who are gendered female, where to be gendered female is a social rather than a biological kind. So, my position is certainly intelligible. Finally, however, it might be argued that if this is how I want to understand “woman,” then my thesis is uninformative, since the concept “gendered female” has not been properly, non-circularly defined.

Is this latter horn of the above dilemma a fatal (or sound) objection to my argument? No. First, the thesis “women are adults gendered female” is non-circular in that to be gendered female is a social practice: it is to be recognized as satisfying relevant social norms (see, for example, Ásta 2013; 2018). So, for example, in I, Robot the artificial intelligence VIKI is recognized as female in part because of her name (“VIKI” is a female name), in part because of her voice (which is stereotypically feminine), and so on. In WALL-E, the featureless robot Eve is recognized as female for similar reasons. Similarly, in the Star Trek universe, women of different alien races broadly satisfy gender norms for their respective planetary race. These social norms specify who is female in a gendered sense, and hence, who is a woman. Second, even if the social practice of gendering is itself in some sense circular—given that it plausibly does hold that being gendered female simply is to satisfy norms for what counts as “gendered female”—I do not think that this is a problem, and for reasons that Louise Antony gives in defending a deflationary account of gender (Antony Reference Antony2020). Just as some epistemologists now take knowledge to be epistemically fundamental (knowledge-first epistemology) and many meta-ethicists take reasons to be normatively fundamental (a reasons-first approach to normativity), I think it is entirely plausible to take gender to be irreducible to anything more basic than social practices determining what these things are (that is, I don't take the property of being gendered female as bottoming out in anything deeper than the very practice defining who counts as “gendered female,” that is, satisfying social norms for what counts as feminine in the sense relevant to being a girl or woman). Thus, if there is any circularity here, it is a non-vicious circularity.

Second, Byrne and so-called “gender-critical feminists” may take issue with how gender is currently socially constructed, that is, the entire prevailing system of gender norms and classifications according to which trans women are women (see, for example, Jeffreys Reference Jeffreys2014). But this is the crucial point: these philosophers are quite self-consciously advocating for a revisionary conception of gender, one that criticizes what gender is. In their own words, they are “gender-critical.” But to be critical of what is, is just that: contending that what is ought not to be. So, for example, if I criticize calling this article “an article,” or classifying chairs as chairs, then I am clearly doing something revisionary: I am arguing that articles shouldn't be called “articles” or chairs “chairs.” But now if this is how Byrne means to defend AHF, then his argument is predicated upon conflating what gender-critical feminists think woman ought to mean with what it does mean.

Next, Byrne and gender-critical feminists might argue that adult, female, and woman are not truly or unambiguously inclusive in the ways that I have contended, such that cartoon characters, robots, and trans women are not properly termed women. For example, although we may colloquially call cartoon characters and robots “women,” Byrne might say that many of us may be willing hedge these claims, adding (for example) that “robot women” are not really women—that is, that AI or robot “women” are merely artificial women, not real women, where to be a “real woman” requires satisfying some reductive biological definition of female. And indeed, it might even be claimed on this basis that when we talk of “robot women” or “cartoon women,” we are not speaking literally, but rather figuratively, referring to beings who are not literally women as though they are.

Yet it is easy to see how these objections err. First, as we saw earlier, it is clearly infelicitous to talk about “women dung beetles” or “women cows.” We do not call beetles or cows that are biologically female in sense 1(a)(1) women because they are not gendered female. Conversely, all of the individuals whom we do systematically call women—ranging from human women to alien women, to robot women, to cartoon ant women—genuinely share a property in common referred to by the expression that explains this usage: they are all gendered female. Thus, the literal meaning of woman—adhering to “the ordinary construction or primary meaning” of the expression (Merriam-Webster)—clearly refers to being female in a gendered sense. When we talk about robot women, alien women, or cartoon women, we are not merely speaking figuratively: we are literally attributing to them one and the same property—being an adult who is gendered female—that we attribute to cis human women. Finally, we can see that the restriction of woman to include only some adults who are gendered female is linguistically arbitrary. In the 2017 film Blade Runner: 2049, an android woman, Mariette, dismisses a male (android) protagonist's romantic interest in a holographic woman, Joi, by saying, “Oh, you don't like real girls.” Mariette obviously means to imply that she, Mariette, is a real girl, whereas Joi is not. But this is absurd. If Mariette is a real woman (and she is), then so too must Joi be one—because they are equally gendered female. Mariette's dismissal of Joi as not a “real” girl is rooted in nothing more than an arbitrary linguistic prejudice on Mariette's part that androids like her can be “real girls” or women but hologram AIs like Joi cannot. Notice, further, that if one were to tell Mariette that she is not a real woman because she is an android, then she would almost certainly find this to be an arbitrary and offensive denial of what she plainly is: an android woman. But, in that case, by a similar token, the contention at the heart of the objection here—that trans women are not “real women” because android women are not “real women”—is nothing more than a similarly arbitrary linguistic prejudice. Mariette is not a “natural biological woman”—that is for sure! That is, she is not the same kind of woman as a human woman. But still, for all that, Mariette is a woman. How do I know this? Because, as we have seen, all of the types of individuals that the term woman picks out—ranging from cisgender human women to robot women, trans women, Cardassian women, Hutt women, and beyond—have something important in common that explains our classifying them this way: they are all adults gendered female.

We can further see how linguistically arbitrary the above sorts of objections are (for example, denying that hologram women are real women because they are holograms; denying that robot women are real women because they are robots; or denying that trans women are real women because they do not have bodies with particular biological/reproductive characteristics) by considering a variation on this same linguistic move. In the Old Testament, God is said to only “clothe” Adam and Eve “in skin” after they sin by eating the apple of the Tree of Knowledge of Good and Evil (Gen. 3:21). So, it seems, prior to this time Adam and Eve did not have corporeal human bodies at all (with reproductive organs, and so on), but were perhaps more akin to disembodied angels. Suppose, then, for the sake of argument that there were disembodied angels in Heaven witnessing this entire process, and that these angels were gendered, as they seem to be in the Bible (as the angels referred to therein include “Michael,” “Raphael,” “Gabriel,” and so forth). What might a female gendered angel, call her “Angel,” say about Eve once Eve is clothed in human skin? Setting theological claims aside about what angels ought to believe (given their relation to God), one can readily imagine Angel having much the same kind of reaction to Eve that Mariette has to the hologram AI, Joi. Angel might well be inclined to say, of Eve, “Eve is not a real woman. Only we adult female angels are real women. Eve is an artificial woman that Yahweh clothed in skin!” But this is plainly absurd. Why? Because Eve is a woman. Eve is gendered female as much as Angel is—and as much as a robot woman, cartoon woman, alien woman, trans woman, or cis woman is. Cases like this once again demonstrate the point of this article: that when we follow ordinary usage and do not make linguistically arbitrary decisions, we can see that to be a woman just is to be an adult who is gendered female.

Further, the above discussion once again highlights an important deeper point: namely, that in claiming that women are (simply) adult human females (in a biological sense), such that this excludes trans women, people like Byrne conflate genuine qualifications that can be meaningfully attached to the category (that is, human woman, robot woman, cartoon bug woman, and so on)—each of which denote a genuinely different type of woman (human women, robot women)—with the meaning of woman itself. For again, if Byrne's AHF were right, then “Azteca is a woman ant,” “Mariette is a woman android,” and so on, would all be literally incoherent, meaning “Azteca is an adult human female ant,” “Mariette is a (biologically-)adult human woman android,” and so forth, such that none of these expressions would express truths. Yet these statements do not express such incoherent absurdities. They express obvious truths: Azteca clearly is a woman ant (at least in her “cartoon world”), Mariette is a woman android, so too are Westworld's Delores and Maeve, and so on—because, again, they all plainly share something in common that explains this: they are all gendered female. So, AHF is false. Woman can be coherently modified in all kinds of ways (robot woman, trans woman, cartoon ant woman)—but all such modifications presuppose the meaning of woman this paper defends: that women are adults who are gendered female.

There is one final set of objections that we should surely address. Consider the first sense of female discussed earlier:

Suppose Byrne were to concede every point defended so far in this article (as he should, since the argument so far is sound). That is, suppose Byrne were to admit what is clearly the case: that on the standard literal meaning of woman, to be a woman is to be an adult who is gendered female, such that alien women, robot women, cis women, and trans women really are all women. That being said, Byrne might contend that there is still another, perfectly intuitive, literal, colloquial meaning of woman according to which robot women and trans women are not women but cisgender human women (and alien women) are women: namely, sense 1a(1)—“being women” in the sense of satisfying a purely biological definition of “female.” Indeed, Byrne might conjoin this objection with the following observation: that when, in Westworld, Delores's mind comes to occupy Martin Connells's body, there is clearly a sense in which “Delores is now a man.” After all, Byrne might say, Delores is quite naturally described this way: as being a woman (in the gendered female sense) in a man's body. Consequently, Byrne might say that even if trans women are women in a gendered sense, they are clearly not “women in a biological sense,” and that this alone is sufficient to deny what many trans-inclusive theorists and activists want to affirm: namely, that trans women are women simpliciter, without qualification.

There are a number of things to say in response here. First, consider again how unnatural (and indeed, infelicitous) it is to classify other biological entities as “women” on the basis of reproductive characteristics. There are, for example, adult female flowers and adult female dung beetles that can produce eggs, but it would be plainly infelicitous to call an adult female rose or geranium a “woman flower,” or again, an adult female dung beetle a “woman beetle.” No competent speaker of ordinary English would speak this way. Why? The most obvious (and correct) answer is that adult female flowers and dung beetles are not gendered in the sense that women humans, aliens, robots, or cartoon characters are—namely, by broadly conforming to socially constructed gender norms qua what counts as feminine in a sense relevant to being a woman (which gender-fluid, gender-queer, and agender individuals violate in various ways, making them not men or women simpliciter, but rather individuals who span these genders or otherwise have another gender or no gender). Flowers and dung beetles are never gendered in this sense, which is why we never call them women. So, woman does not refer to adult females “in a biological sense” . . . unless those individuals are also gendered female—in which case this article's argument still goes through: the explanation for why a particular being is a woman (such as cis human women, Cardassian women, robot women, trans women) is not biological, but rather social, having to do with gender.

I imagine that Byrne and many “gender-critical feminists” might want to reject this explanation, claiming that there is another—as-yet undiscussed—reason to classify all and only adult humans and aliens who are female in sense 1a(1) (but not, say, flowers) as women “in a biological sense” (see, for example, Stock Reference Stock2021), for there is an obvious difference between female flowers and female human beings in biological sense 1a(1), right? In addition to being members of very different biological kingdoms, human women have been and continue to be subjects of oppression based upon their biological features: principally (but not only) on the basis of their reproductive features and exhibiting a “natural” female phenotype that biological males have throughout human history unjustly aimed to dominate and oppress (with behaviors ranging from sexual to domestic violence, discrimination, and so forth).

Yet, to see how this objection errs, consider the fact that in many animal species, females (in biological sense 1a(1)) are controlled and dominated by males. This is not only the case in many primate species, such as gorillas and chimpanzees (Van Schaik Pradhan, and Van Noordwijk Reference Van Schaik, Pradhan, van Noordwijk, Kappeler and van Schaik2004), but also in many other species (Smuts and Smuts Reference Smuts and Smuts1993), such as lions—where in lion prides, a single male dominates the entire group, “having the right” to mate at will with female lions in the pride (Bertram Reference Bertram1975). Conversely, in some species, biological females are dominant (Lewis Reference Lewis2018). Still, there is something particularly striking about the scientific literature on these issues: namely, the prevalence of the use of the neutral/ambiguous terms male and female in conjunction with the absence of the use of the terms men and women. For example, in scientific discussions of primate behavior, the discussion centers on “sex differences” and interactions between (biological) “males” and “females” in sense 1a(1)—that is, where male and female are understood in reductive biological terms. Specifically, male lions, chimpanzees, and so on, in this biological sense are identified as seeking to dominate and control (biological) females, principally for reproductive reasons. No one with a reasonable grasp of language or the relevant empirical facts—including trans-inclusive theorists and activists who (correctly) believe that trans women are women—would presumably deny these kinds of facts. Males in many species, including the human species, have aimed to dominate and control females in sense 1a(1): females, that is, in the sense “of, relating to, or being the sex that typically has the capacity to bear young or produce eggs.” Not only that: human males have controlled and dominated human females in sense 1a(1) in very specific ways, some of which may be similar to the ways in which they dominate trans women (for example, harassment and sexual violence), but others of which are very different forms of dominance and control, such as denying human females in sense 1a(1) their basic reproductive freedoms (with abortion laws, and so on), and engaging in distinct forms of violence (namely, transphobic violence) against trans women specifically.

Again, presumably no one with a basic grasp of language and relevant empirical facts would deny any of the above. But here is the key: no one refers to men and women lions or chimpanzees in these contexts. In the scientific literature, the terms male and female are always used instead. Why? At least offhand, one might think that we merely reserve man and woman for human beings simply because they are terms for members of our species—whereas in other species there are analogous distinctions (for example, “boars” and “sows” in the case of swine, “hens” and “roosters” in the case of chickens, and so on). One might suggest, further, that in virtually all of these cases, the relevant terms are synonymous with male and female in the (purely biological) sense. For example, a hen is defined (simply) as a female chicken older than one-year of age, and a sow is (simply) an adult female pig. Alas, although this may seem plausible on the surface, upon further reflection it is entirely clear that this is not how man and woman are understood or used by competent language users. Their meaning is very different. How do I know this? Consider again the series of facts enumerated earlier: competent language users plainly recognize Delores and Maeve as women robots, WALL-E's Eve as a girl robot, Mariette in Blade Runner: 2049 as a woman android, Gardulla Besadii the Elder as a Hutt woman in the Star Wars universe, T'Pol as a Vulcan Woman in Star Trek, and so on. Despite the fact that some of these individuals are robots without reproductive biology, others slug-like creatures, others Vulcans with blue blood and pointy ears, and so on, competent language users readily identify them all as girls or women of the sort they are (for example robot girls, alien women, and so on). This again reveals that, unlike sow or hen—which plainly do refer to members of particular species based solely upon biology—the terms girl, boy, man, and woman do not function this way. There are no “boy” chickens or “woman” swine; rather, there are (simply) male chickens and female swine (in biological sense 1a(1)). But there plainly are girl Vulcans, boy Hutts, android men (such as Joe in Blade Runner: 2049), and women androids, such as Delores, Maeve, and Mariette. The fact that we use the terms girl, boy, woman, and man to refer to individuals who are gendered male or female (respectively) of any biological species (and indeed, to gendered individuals who are not even biological organisms), but not to individuals in species that lack anything like advanced gender norms (for example, chickens or pigs), once again demonstrates this article's thesis: that woman refers not to anything biological but instead to individuals who satisfy the 1(b) definition of female—individuals who satisfy certain socially constructed gender norms, including robot women, alien women, and trans women.

Let us now turn to the remaining part of the same objection: namely, that when (in Westworld) Delores's mind comes to occupy Martin Connells's male body (via uploading her mind into his body), there is a sense in which Delores is now “a man.” Notice, to begin with, that this is on its face a very strange and infelicitous way to speak. No one who watches Westworld thinks Delores has literally become a man simply by occupying Connells's body. She is still Delores, after all. We still recognize her as the woman AI who, in season 1 of the program, was brutally assaulted and repeatedly murdered by the Man in Black, only to be repeatedly “born again” each day to suffer similar horrors as a robot “host” in a Western theme park. The most natural way to describe Delores is as a woman AI who now occupies a male body, a body assigned a different sex and gender than Delores's original sex and gender “at birth” (or rather, at her creation as a woman robot). But, of course, this in turn is precisely why trans women are now (correctly) identified as (trans) women—as women whose gender (as women) does not match the gender they were socially-assigned at birth. So, again, the terms girl and women are gendered terms, and insofar as a person with one gender (Delores) can clearly occupy a (male) body normally associated with another gender (for example, boys and men), it follows once again that trans women, robot women, and alien women are all women—since, as with Delores, individuals in each of these categories can clearly be gendered female.

There is one final objection to address here, which is that in the cases in Delores, Maeve, Mariette, and women Hutts, there is a clear sense in which the individual's psychosocial history is plausibly relevant to whether they are women. For example, at the outset of Westworld, Delores is stereotypically feminine, presented to the viewer as though she is a conventionally attractive female member of the human species—features of her that result in her being dominated and oppressed by men in conventionally sexist ways (including sexual violence). It might be suggested that insofar as Delores “had these characteristics from the beginning,” it makes fine sense to still call her a woman after uploading her mind into Martin Connells's body—since she has the history of a woman and was long treated as one (particularly by men) due to phenotypically-female physical features. But, it might be said, the opposite is true of trans women. Trans women entered the world identified as “male,” may have developed physically “as boys,” and so on—in which case, it may be suggested, they are still boys (or men) regardless of their gendered-feminine features (for example, that may only develop or present later in life).

However, it is surprisingly easy to see this line of argument to be unsound—though perhaps just how easy it is has not been fully appreciated due to the standard focus on humans, as opposed to robots, cartoon characters, and so on (as the present article does). For consider what we would doubtless say about a robot (such as WALL-E's Eve) who transitioned from being gendered female to gendered male. Suppose for the sake of argument that future AI programmers developed the ability to replace every feminine trait of Eve's psychology and behavior with corresponding masculine ones. So, for example, in Western societies the following traits are considered stereotypically feminine: affection, cooperativeness, gentleness, humility, nurturance, passivity, sensitivity, supportiveness, sweetness, and warmth (Vetterling-Braggin Reference Vetterling-Braggin1982, 5–6; Stets and Burke Reference Stets, Burke, Borgatta and Montgomery2000; Kite Reference Kite and Worell2001; Windsor Reference Windsor, Smelster and Baltes2015). Conversely, the following traits are stereotypically masculine: assertiveness, independence, leadership, strength (Vetterling-Braggin Reference Vetterling-Braggin1982, 6; Carli Reference Carli and Worell2001; Kimmel Reference Kimmel, Adams, Blumfeld, Castañeda, Hackman, Peters and Zúñiga2013)—not to mention machismo (Morales Reference Morales, Laird and Green1996). Second, there are also clearly norms for gender performance (Butler Reference Butler1988): masculine and feminine forms of dress, speech (prototypically “masculine” voices being lower in pitch, “feminine” voices higher), hair styles, and so on. Finally, there is a person's internal sense of their gender identity—which is known to be informed by a person's sense of their own masculinity or femininity (Bem Reference Bem1974; Wood and Eagly Reference Wood, Eagly, Fiske, Gilbert, Gardner and Jongsma2010; Tate Reference Tate and Miller2014). Importantly, recognizing all of this does not presuppose any kind of gender essentialism—the view that there is an essence of what it is to be a man or woman, nor does it deny that men can be feminine in various ways. It is entirely consistent with the more plausible idea that gender categories are family-resemblance concepts (Wittgenstein Reference Wittgenstein and Anscombe1958, secs. 66–68), such that there is no precise number of masculine or feminine traits that an individual must have to be a man or a woman, respectively—but rather, the more or fewer such traits one instantiates, the more (or less) paradigmatically male or female one is in a gendered sense. Obviously, there are men who are nurturing, passive, sensitive, and so on, as well as women who are assertive and independent. The point is simply that a person's gender will normally be understood—both by themselves and others—as a function of their masculine and feminine traits in conjunction with gender performance and experienced gender identity (the latter of which, in turn, will normally be understood by reference to the former, as a person will typically identify as a man, woman, gender-queer, and so on in virtue of their particular blend of gendered traits and performativity). Notice, finally, that all of this is also consistent with ongoing debates about how gender ought to be understood—such as whether gender should be understood in terms of dispositions, relevant social norms, or self-identification alone (see, for example, Jenkins Reference Jenkins2018), debates that we cannot settle here.

Bearing all of this in mind, let us return to the thought experiment mentioned above: the example of a stereotypically feminine robot having all of her feminine gendered traits replaced by masculine ones. So, for instance, as mentioned earlier, in the Disney film WALL-E, Eve is stereotypically girlish, being a bit bashful and demure toward WALL-E, complete with a high-pitched, female-sounding voice, and so on. Suppose then that Eve entered a “robot gender-transition shop” asking to have these features of her personality and behavior systematically replaced by stereotypically male features: a deep baritone voice, a desire to be called “John” rather than Eve, interests in stereotypically masculine things (American football), stereotypically masculine behaviors, interests, and performance (such as dominating conversations, preferring masculine deodorants to feminine ones, wearing male clothing, and so forth), an internal sense of being a boy robot, and so on. Despite her previous history as Eve, it now seems clear that if Eve has changed so fundamentally in these ways, she will have become a male AI, that is, a boy robot: namely, John. Indeed, suppose that, in addition to the above changes to her psychological and behavioral characteristics, the robot formally known as “Eve”—now John—receives a physical-body makeover to in turn make much of her body stereotypically masculine as well: specifically, in ways that satisfy various (traditional) gender norms and expectations for “how men are supposed to be,” physically speaking. John, let us suppose, has a robot-penis prosthesis permanently attached where before (on Eve) there had been nothing, or perhaps the robot equivalent of biologically-female genitalia. Now it may seem even clearer than before that Eve has become John, a plainly male robot. Eve once was a girl robot—at least in the sense of having been socially conferred this gender (see Ásta 2013; 2018)—but now is a male robot in a gendered sense. This, again, is not because reproductive biology (such as having a penis) objectively makes someone male or female, or indeed, is even necessary for that (since WALL-E is clearly identifiable as a male robot despite having no reproductive organs). Rather, it is because now—in addition to having a masculinized personality and a litany of distinctly masculine behavioral-performative characteristics—John now looks more “like a man is supposed to look” (in a traditional socially defined sense) than Eve did (Stoller Reference Stoller1974), and is able to do characteristically masculine things sexually that Eve could not do (for example, engage in penetrative sexual activities without a temporary prosthesis). Again, John (formerly Eve) needn't have a penile prosthesis to become a distinctively male robot—as the various psychological and behavioral changes would suffice for Eve to transition from distinctly feminine (qua girl) to distinctly masculine (qua boy or man). The point is merely that, in transitioning from the less-masculinized body that Eve had, John would become more masculinized (or “manly”) than before, physically speaking. Finally, though, whether such gender norms are good or just ones for a society to have or for individuals to conform to is another question entirely. It could well be that the most morally defensible (and just) gender norms are ones that define gender (and hence, womanhood) in terms of self-identification alone (Bettcher Reference Bettcher, Garry, Khader and Stone2017a)—though again these are ongoing debates. This brings us to the following.

Byrne concludes his article with a proposed lesson. He believes that his argument reveals that ameliorative definitions of ordinary terms, such as Sally Haslanger's (a woman is a person “subordinated in a society due to their perceived or imagined female reproductive capacities” [Haslanger Reference Haslanger2012, 8]) only go so far, introducing entirely new concepts and categories. As Byrne writes, “Sometimes a word is replaced or demoted by a synonym . . . but that just makes the old distinction by new means. There is no reason to think that the utility of a word for the category adult human female will markedly decline in the foreseeable future” (Byrne Reference Byrne2020, 3801). However, as we have seen, there is no word for adult human female. That is simply not what woman or any other single word means. How do I know this? Because, again, adult Cardassian females are women (nonhuman ones) and adult robot females are women (despite not being biological beings at all). Byrne, then, is simply choosing to understand a word, “woman,” in a way that it manifestly isn't used or understood by any competent language user of the English language. Why? One can only surmise—though we have seen various possible reasons in this article, each of which is predicated upon some form of confusion. The relevant point for our purposes is that denying any of the above claims is to discriminate against certain kinds of women, deeming them not women when they are. That is to engage in ameliorative definition of woman—a gender-critical definition according to which current gender norms and categories should be rejected, and according to which the concept of woman should only be used to refer to adult human beings who satisfy the dictionary definition of female in biological sense 1a(1).

This article's argument has been primarily descriptive, focusing upon what the word woman literally means, and hence, on which types of beings are women. That being said, my argument has also touched upon a number of normative issues in ways that are plausibly illuminating with respect to ongoing debates over gender-critical versus trans-inclusive feminism, such as which conception of woman ethics and justice require.Footnote ³ To see how, consider again Mariette, an android woman who wishes to deny that Joi (a hologram AI) is “a real girl.” This article's argument reveals just how arbitrary Mariette's exclusion of Joi from this category is. Joi is a “girl” (or rather, woman) hologram just as much as Mariette is a woman android, as they both genuinely satisfy the satisfaction-conditions (or meaning) of girl and woman: they are both gendered female. Mariette's denying Joi this equal recognition as a girl/woman in effect denies linguistically (and socially) that Joi has relevant properties to being one that Joi in fact has: properties that Joi and Mariette both care about, and moreover, that Joi and Mariette genuinely share—namely, being gendered female. Are Joi and Mariette different kinds of women in potentially important ways? No doubt. For one thing, as a hologram Joi has no physical body, so there are things that Mariette can do as an android woman that Joi cannot do as a hologram woman, and ways that Mariette can be physically treated by others that Joi cannot; and vice versa. These differences between Mariette and Joi are real, and may warrant careful consideration as matters of ethics and justice. But, for all that, Joi and Mariette are both women AI. Thus, whatever else ethics and justice may require, denying that Joi is a woman is not only descriptively incorrect (as this article has shown); it is also seriously pro tanto unfair to her, at least in these regards: it denies the literal reality of something true of her that is (or at least may be) central to her existence as a person—namely, that she is gendered female—as well as the further reality that she and Mariette genuinely share this characteristic (irrespective of their many other differences). My argument shows that the same is true of trans women: denying that they are women (simpliciter) denies the literal reality of something true of them and central to their existence as persons—namely, that they are gendered female—as well as the fact that they genuinely share this characteristic with cis women and the other kinds of women discussed herein. Might there be countervailing normative considerations in the debate between trans-inclusive and gender-critical feminism that might justify changing how woman is understood, such that the term should be reconceived in an exclusionary manner, referring merely to “biological sex” rather than to all adults who are gendered female? Much has been written on this important topic, and by authors far more qualified on these matters than I (see, for example, Koyama Reference Koyama, Dicker and Piepmeier2003; Haslanger Reference Haslanger2012; Saul Reference Saul, Crasnow and Superson2012; Bettcher Reference Bettcher, Soble, Power and Halwani2013; Jenkins Reference Jenkins2016; Bettcher Reference Bettcher, Garry, Khader and Stone2017a; Reference Bettcher2017b; Jenkins Reference Jenkins2018; Hines Reference Hines2018; Kirkland Reference Kirkland2019; Dembroff Reference Dembroff2020b; Zanghellini Reference Zanghellini2020. Cf. Jeffreys Reference Jeffreys2014; Bogardus Reference Bogardus2020a; Reference Bogardus2020b; Stock Reference Stock2021). The relevant point for now is that to determine whether the meaning of the term woman should be retained or changed as a matter of ethics or justice, we first need to know what the term means—that is, who women are according to the ordinary literal meaning of the term. This article has shown what that meaning is, and it isn't what Byrne and others who express skepticism about the identity of trans women claim it to be.