Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Feminist Philosophy Quarterly Volume 11 | Issue 3 Article 1 Recommended Citation Bettcher, Talia Mae. 2025. “Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression.” Feminist Philosophy Quarterly 11 (3). Article 1. 2025 Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Talia Mae Bettcher Cal State Los Angeles tbettch@calstatela.edu Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 1 Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Talia Mae Bettcher Abstract In her book Ontology and Oppression, Katharine Jenkins defends a framework for the ontology of social kinds that is impressive in both its systematicity and explanatory power. My central concern, however, is that Jenkins seems to merely assume a separation between semantics and ontology—a separation that needs to be investigated and defended more fully. This presumption, I conclude, undermines the ability of her account to answer the question of what a woman is, as well as raising worries about potential misgendering. Keywords: ontology, semantics, social kinds, misgendering, transgender In her book Ontology and Oppression, Katharine Jenkins (2023) defends a framework for the ontology of social kinds that is impressive in both its systematicity and explanatory power. Not only is her position plausible, well argued, and clear, but the book is a great example of how to think about social ontology via analytic methods and how to do it well. In my opinion, Jenkins sets the standard in avoiding some of the common pitfalls of analytic approaches. She does not, for instance, ignore the work of important theorists who are not analytic, such as María Lugones (2010), Hortense Spillers (1987), and C. Riley Snorton (2017)—folks I never would have expected to encounter in analytic discussions of social ontology, let alone to see be taken seriously (Jenkins 2023).1 Additionally, Jenkins is uncommonly sensitive to the on-the-ground political concerns of, certainly, the Gender Recognition Act (2004, c. 7) in the UK and, indeed, perhaps most notably, any possible real-life consequences of her own theory on trans politics and trans lives. In general, it’s safe to say that of the nontrans philosophers doing analytic social ontology today, she is easily one of the most expressly concerned with trans people and the oppression we face. This, for my part, has not gone unnoticed. 1 Jenkins discusses Lugones, Spillers, and Snorton on pp. 124–26, 124–25, and 130– 31, respectively. Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 2 Despite its many virtues, however, I do think that the work is also open to some criticism. My central concern involves an important lacuna in Jenkins’s account— namely, the absence of any discussion of ordinary discourse and its relation to social kinds. To put it differently, I am worried that Jenkins seems to merely assume a separation between semantics and ontology—a separation that needs to be investigated and defended more fully. This presumption, I shall conclude, undermines the ability of her account to answer the question of what a woman is, as well as raising worries about potential misgendering. I begin with a brief overview of Jenkins’s position, and then I develop my concerns over the course of three sections. In the first of these, I introduce one of the objections to Jenkins’s account that Jenkins herself considers—namely, the misgendering objection (roughly, the worry that positing social gender kinds at odds with trans self-identities is to engage in misgendering). I also argue against Jenkins’s contention that the mistaken “ontology-first assumption” (social practice should be determined strictly by social ontology) plays a role in motivating this objection. In the second section, I show how the separation of semantics and ontology leaves Jenkins unable to answer gender-critical feminist arguments in the way she seems to suppose. In the final section, I consider the general question of what role semantics ought to play in answering questions of social ontology. In doing so, I show how Jenkins’s account is confronted with two concerns—namely, the possibility of misgendering through the account’s own theoretical deployment of terms, and the inability to answer the question “What is a woman?” Overview Ontology and Oppression offers a unifying framework for examining whether certain human social kinds—specifically gender and race kinds—are (what Jenkins calls) “ontically oppressive” and whether, in some cases, they are even actually emancipatory (Jenkins 2023, 70–72). Jenkins calls this the “constraints and enablements framework” or CEF (77–115). Ontic oppression is, for Jenkins, a species of what she calls ontic injustice, this latter arising when the constraints and enablements of a social kind inflict moral injury upon its members through the contravention of their moral entitlements (24–26). Ontic oppression is defined as a species of ontic injustice that is systematic, group-based, and subordinating, where subordination includes exploitation, marginalization, powerlessness, cultural imperialism, violence, and communicative curtailment (75–76).2 2 Jenkins’s notion of “ontic oppression” is not far from Robin Dembroff’s (2018, 25) notion of “ontological oppression.” For Jenkins’s discussion of the differences and similarities, see pp. 74–75. Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 3 According to CEF, social kinds are to be understood as explanatory kinds—that is, as groupings according to a property or set of properties where such groupings can serve some set of explanatory purposes (Jenkins 2023, 78). Human social kinds— social kinds with humans as members—are, in this framework, constituted through shared sets of constraints and enablements (83). So, a human is a member of some social kind just in case they are subject to the constraints and enablements constituting that kind (84). Of the many constraints and enablements, Jenkins identifies four types: those concerning the interpersonal behavior of others, those concerning individual psychology, those concerning the body, and those concerning the environment (92). As part of the framework, Jenkins (2023, 88–89) introduces three features— scope, breadth, and granularity—that function like sliders on a sound mixer to enable different ways of viewing human social kinds. Turning to gender and race human social kinds specifically, Jenkins considers three different sorts of social kinds determined through different mixes—namely, hegemonic kinds, interpersonal kinds, and identity kinds (116, 136, 158). Hegemonic kinds, for Jenkins, track large-scale structural phenomena. Think, for instance, of Sally Haslanger’s social class account of a woman as someone who is systematically and for the most part subordinated on the basis of perceived reproductive capacity (Haslanger 2012a, 235). In this case, the scope is very large (the whole world over the last several hundred years, say), the breadth is maximal (that is, all sorts of constraints and enablements are considered), and the granularity is very coarse (Jenkins 2023, 135). Interpersonal kinds concern, by contrast, contextually specific phenomena (Jenkins 2023, 157). Think, for instance, of Ásta’s (2018) account of a social status as contextually conferred on the basis of some tracked property. In this case, the scope is very narrow (context-specific), the breadth is restricted to interpersonal constraints and enablements, and the granularity is fine-grained enough to accommodate intersectionality. Finally, identity kinds roughly concern one’s felt or affirmed relationship to a category. Think, for instance, of Jenkins’s own account of gender identity (as norm-relevancy) and the role it plays in her original response to Haslanger’s ameliorative account of woman as social class (Jenkins 2016). In this case, the scope is medium (between hegemonic and interpersonal kinds), the breadth includes both psychological and bodily constraints, and finally, the granularity is sufficiently fine to include intersectional identities (Jenkins 2023, 167–69). Jenkins concludes that, of these three, while many hegemonic kinds are ontically oppressive, when it comes to interpersonal kinds and identity kinds, some are, some are not, and some can even be emancipatory (Jenkins 2023, 135, 157, 174). The CEF affords substantial benefits, of course. As Jenkins argues in her penultimate chapter, it provides a unified framework for thinking about these three different sorts of race and gender kinds (Jenkins 2023, 176–80). Specifically, CEF allows Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 4 us to understand how the three operate together. It helps provide a way to structure and arbitrate disagreements about the nature of race and gender kinds. And, finally, it helps mitigate objections against one type of kind by placing that type within a larger pluralist apparatus. That said, one might also have some concerns about this framework, and Jenkins turns to these in the penultimate chapter as well (Jenkins 2023, 180–98). The last of them—namely, the concern that any view that allows for gender kinds that do not validate trans identities, a kind of misgendering, is bad for the project of trans liberation—also serves as the subject of the entire final chapter (192–99, 200–240). It is around the topic of this final objection that I begin my own critical engagement with Jenkins’s work. Misgendering and the Ontology-First Assumption. The last objection (or set of objections, rather) Jenkins considers centers on the worry that her pluralist account of gender kinds can lead to the misgendering of at least some trans people. For instance, if we allow for the existence of certain hegemonic kinds or interpersonal kinds, it may turn out that some trans women do not count as members of the hegemonic or interpersonal kind women, and some do; and that some trans men count as women, and some do not.3 The concern is that it seems bad to have an ontology admitting of some kind woman from which some trans women are wrongly excluded and in which some trans men are wrongly included. I’ll call this the misgendering objection, even though it admits of a few variations. The first variant is the worry that because Jenkins herself admits different gender kinds in which at least some trans women would not be admitted as women and some trans men would be, her own account engages in a species of misgendering (Jenkins 2023, 192–93). Let’s call this the basic misgendering objection. Jenkins’s response is to point out that her own account has nothing to say about semantics at all (194). As a consequence, should somebody allege that some trans women are not women insofar as they are not oppressed as women (i.e., not part of the hegemonic kind woman), then Jenkins can say that there is nothing in her view that entails the truth of that allegation. Nor, Jenkins adds, is there anything that affirms the appropriateness of asserting such a thing, even if it were true (194). Nor, even, is there anything that affirms the political wisdom of asserting such a thing, even if it were true and pragmatically appropriate (194–95). For now, I will put the basic misgendering objection to the side, returning to it toward the end of the paper. Instead, I want to take up what I shall call the advanced misgendering objection. In one very strong version of this, the objector wants to insist 3 This objection echoes an important critique of Haslanger’s (2012a) original account issued by Jenkins (2016) herself. Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 5 that there should be no possibility of it turning out that misgendering statements are true. In other words, any viable ontology must guarantee the falsity of all misgendering statements (Jenkins 2023, 195). Understood as a semantic claim about what the word woman actually designates (i.e., the relevant identity kind), there is again no problem, claims Jenkins, since she is not making a semantic claim anyway (197). But if it is to be understood as an ontological claim that restricts gender kinds to identity kinds or that proclaims their fundamentality over the other sorts, then, Jenkins notes, it runs into complete conflict with her own pluralist view (197). Jenkins then goes on to imagine a weaker version of the advanced objection that says simply that an account that does not have this strong ontological guarantee is bad for trans politics and against the aim of trans liberation because it suggests that, at least in some sense, trans women are not women and trans men are not men (Jenkins 2023, 197). Her final chapter is devoted to taking up this worry (as well as to dispelling the background motivation for the stronger version of this). To show that her framework is far from detrimental to the liberation of trans people, but quite useful, and hence that the weak version of the advanced misgendering worry is misplaced, Jenkins focuses on the debate in the UK concerning the Gender Recognition Act. This is an excellent strategy because it affirms that our philosophizing requires a focus on the particular contexts in which we philosophize. Specifically, she argues that underpinning this debate is an “ontology-first” assumption is ultimately bad for trans liberation because it entrenches the position of antitrans advocates and requires that protrans advocates accept terms of debate that are too demanding. The assumption says “that settling questions about the current ontology of gender will automatically determine what shape our gendered practices ought to take” (Jenkins 2023, 201). In the place of the ontology-first assumption, Jenkins suggests we adopt a normative approach in which we consider “the consequences of adopting one or another kind of social practices and to choose the practices that are most conducive to emancipatory efforts” (226). A key move in Jenkins’s response to the ontology-first approach is to recognize that some social kinds can be oppressive. By recognizing the oppressive character of social kinds, says Jenkins, we ought to give up the assumption that current ontology should immediately determine social practices. Here, she draws also from Dembroff’s rejection of (what they call) “the real gender assumption”—namely, that social classification ought to track the membership conditions of social kinds (Jenkins, 2023, 216–17; Dembroff 2018, 6). While the rejection of the ontology-first assumption is plausible, I worry that Jenkins is in fact wrong in supposing that the antitrans and protrans advocates she has in mind really endorse the ontology-first assumption, at least if it is taken as an immediate move from the claims about kind membership within our current ontology to claims about social practice. In such a view, antitrans advocates would presumably Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 6 maintain that insofar as trans women are not women in our current ontology, we ought not accept gendered practices that treat trans women as though they were. Similarly, protrans advocates would presumably contend that since trans women are women in the current ontology, we ought to accept practices that treat them as such. Putting aside quick slogans and trendy hashtags (like “Choose Reality” [Jenkins 2023, 206]), however, I would suggest that both antitrans and protrans advocates also accept some additional empirical and/or ethical principles that are required to make the movement from claims about ontology to claims about social practice. For example, I do not find that those who support preventing trans women from using women’s restrooms do so merely because they view trans women as men. They do so, further, because they take sex segregation to be motivated by concerns about violence against women. They presume that men will tend to behave in sexually assaultive ways toward women. And they maintain that sexual assault is wrong. To better bring this out, consider that often enough, folks are willing to accept the practice of referring to trans people by their preferred names and pronouns in nonintimate contexts. They are willing to do so regardless of their own beliefs about gender ontology, suggesting that they reject the ontology-first assumption. That is, they are willing to “play along.” Notably, then, they may also express concerns about sex-segregated spaces such as public restrooms and changing rooms. Such a reluctance again suggests that the concern isn’t ontology but, rather, (false) empirical beliefs about men and women and general ethical principles about the wrongness of sexual assault. When a gender-critical feminist claims, then, that trans people ought never be treated in accordance with their gender identity, it is likewise owing to their underlying political and ethical commitments. Similarly, I would urge that protrans advocates who claim that gender identity determines kind membership additionally rely on the view that it is an ethical imperative that gender identity be respected in all cases, except those where concerns about safety might override that imperative. The point, to repeat, is that, in either case, the movement is not from ontology to social practice but from ontology coupled with empirical and normative claims to social practice. What these two sorts of stances have in common, and I think Jenkins would agree with me, is a commitment to the antipluralist view that there are unique answers to questions like what a woman is (Jenkins 2023, 221–22). The difference between our views, then, is that while I think that the ontology-first assumption is a red herring, Jenkins does not. (She does notably think that additional work is required to dislodge the antipluralist stance over and above any work to dislodge the ontology- first assumption. And, of course, an argument in favor of her own pluralist position is an argument against antipluralism (Jenkins 2023, 222). In the section that follows, I take up the question of what motivates this antipluralist position or, at any rate, a Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 7 rejection of the sort of pluralism endorsed by Jenkins. This will bring us to the importance of semantics in determinations of social kinds. Explanation, Reference, and Gender-Critical Feminism The deeper question to which we now turn is just why antitrans and protrans advocates suppose that there is one unique answer to the question of what a woman is (assuming the latter do). Citing her own account of social kinds as explanatory kinds, Jenkins suggests that protrans advocates are expecting gender identity kinds to do all the explanatory work (Jenkins 2023, 214). While she is less explicit in her views about why antitrans advocates expect a unique answer, I will presume that it is the same. And the problem is that this explanation scarcely seems plausible. While Jenkins recognizes that questions about the ontology of gender might be settled by—rather than explanatory considerations—what she calls “some other factor,” no other factor is named and discussed (Jenkins 2023, 215). One evident factor, however, is precisely ordinary discursive practices and the reference of our kind terms. So, for instance, when we ask the question “What is water?” we may be asking what kind the term water picks out, which turns out to be the natural kind H2O. More generally, the thought here is that the way to settle questions of social ontology is to appeal to the semantics of ordinary discursive practices. To put it differently, the question “What is a woman?” consists in the question of what kind of being we’re talking about when we use the word woman. I will call this the discursive approach to social kinds. It seems to me that it is just this approach that grounds the antitrans view that women are adult, female human beings (AFH for short)—namely, that woman picks out the kind AFH. To put it differently, the view that women are adult, female humans depends upon the idea that what we’re talking about when we talk about women, at least ordinarily, are members of the kind AFH. It does not, by contrast, appear to hinge on the claim that this conception of women fulfills all explanatory demands. Unfortunately, it is precisely Jenkins’s elision of the discursive approach that leaves her unprepared to fully engage with those antitrans advocates who adopt this strategy. To develop this point, let me turn to prominent antitrans advocate and philosopher Kathleen Stock. In her paper, “The Importance of Referring to Human Sex in Language,” Stock opens by claiming that woman is typically supposed to designate AFH whereas man is typically supposed to designate AMH (Stock 2022, 25).4 The first thing to notice is that since this position is tantamount to the claim that woman designates something like a natural kind, it is not one that can be addressed by 4 I am grateful to Cyprus Marques for the illuminating discussion we have had about Stock’s paper. Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 8 assuming that woman is a social kind or, even, plurality of social kinds. In other words, Jenkins’s entire framework is irrelevant to a feminist such as Stock. The second thing to notice is that because Stock (2022, 25) takes the designation of this kind to be the effect of ordinary discursive practice, she can then represent any efforts to include trans women as revisionist. On this point, it is worth noting that Haslanger’s (2012a) appeal to the hegemonic kind was originally announced as an ameliorationist position. Indeed, so, too, was Jenkins’s own previous position, which allowed for both woman as a class and woman as an identity (Jenkins 2016). The problem is that once one position is taken to track the ordinary way we speak while the other is taken to reflect a revision in language, the burden of proof is on the latter for endorsing such a revision. Unsurprisingly, then the antitrans advocates who would have us “choose reality” can claim to have the following on their side: What ordinary people are talking about when they talk about women is the natural kind AFH. That’s the reality. So, to propose a departure from this—to say that a woman is one who self-identifies as one, for instance—is to depart from that reality. My point is that Jenkins’s position does not help in the way she claims it does. For it to help, she would need to argue in favor of the explanatory view of social kinds over the view that privileges the referents of kind terms. And this is not something she does. And since I am not persuaded the explanatory view of social kinds is the way to go anyway, I would propose two other strategies for responding to Stock. The first is to argue against the claim that the kind term woman as deployed in ordinary discourse designates a natural kind in the first place. After all, people can be mistaken about what their ordinary terms mean and what those terms designate. (This is why Haslanger [2012b, 92] distinguishes between the manifest and operative concept.) For instance, people went a very long time using the term water while ignorant of the fact that they were referring to H2O. Or more to the point, consider that probably many folks believe that the term woman designates individuals who are human. Upon closer scrutiny, however, we recognize that woman can designate Vulcan, Klingon, and Kryptonian adult females (Bettcher 2017; Arvan 2023; Bettcher 2025). This means that what ordinary folks think about the referent of woman can be wrong. Similarly, one might argue that the common view that woman designates members of a biological sex is likewise mistaken (Arvan 2023, Bettcher 2025). Instead, one might argue, the term turns out to refer to a social kind after all—a sort of social kind that is wont to pass itself off as a natural one. This means that those who claim woman designates biological females are themselves revisionists about language. It also opens the door for arguing that the social kind in question is an oppressive one that ought to be abandoned. This, in turn, helps motivate a revisionism that secures the validity of trans identities, rather than opposing it. A second strategy consists in simply accepting the contention that woman designates adult female humans—at least for the sake of argument—while arguing Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 9 that various social practices ought to be altered regardless, including discursive ones. After all, Stock (2022) herself provides arguments that moral appeals to the importance of gender identity are not as weighty as trans advocates suppose in motivating a change toward trans inclusion in certain social practices and that the interests of nontrans women in trans exclusion are overwhelming. To counter this, one need only argue that Stock is mistaken on both these counts. Here, one can, as Jenkins herself proposes, adopt a case-by-case approach in making these assessments (Jenkins 2023, 232–33). None of this requires adopting Jenkins’s rendition of social kinds as explanatory kinds, however. Misgendering and the What-Question In this final section, I consider what I regard as the most important questions that this discussion raises. First, as social ontologists, how should we approach social kinds? Should we view them as shared properties that yield explanations? Should we think about them as referents of ordinary-kind terms? Second, in general, what role ought the semantics of ordinary language play in our approach to social ontology? Crucially, it is here that I locate what I take to be the most serious weakness of Jenkins’s account—namely, that there are no compelling arguments in favor of the explanatory-kind view of social kinds. Surely, the discursive position ought to be acknowledged, and the explanatory account of social kinds ought to be defended, if possible, over and above it. The discursive approach to social ontology is scarcely an uncommon view, after all. Instead, the alternative position is simply elided. Worse, there are considerations in favor of the discursive approach that pose further challenges to Jenkins’s account. One concerns the risk of deploying politically contested terms in oppressive ways. (This last reflection will return us to the worry that Jenkins’s account is vulnerable to the basic objection of misgendering that I mentioned earlier in the paper). The second concerns what I shall call what-questions (or w-questions) such as “What is a woman?” and the inability of the explanatory-kind approach to answer them. I consider them in order. Let’s begin with a distinction between two sorts of revision. The first, I will call theory-independent, and the second, I will call theory-dependent. The former occurs when communities of oppressed people develop their own discursive practices in opposition to more mainstream ones. Obviously, community members are typically not social ontologists, social theorists, or philosophers. Usually, these revisions to predominant ways of speaking occur organically and, consequently, are adopted by most members of the community in a way that is largely taken for granted. One obvious example is the revisionary use of the N-word (ending with a) as a reappropriation of the N-word (ending with r). This usage is not one that was originally introduced in some philosophical theory. Rather, it emerged organically in resistant subcultures. Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 10 The latter concerns the reappropriation of a term for theoretical purposes. Here, the theorist themself plucks out some term and gives it either a more precise definition or a different definition altogether. For instance, a philosopher may decide to use the term people in a very precise way—say, to refer beings with the capacity for second-order reflection—even though the ordinary usage of people tends to be used to refer to humans only, regardless of capacity for second-order reflection. While this theoretical usage may be merely stipulative—for the purpose of a paper or talk—it might also be accompanied by the recommendation that other theorists begin speaking this way as well. Indeed, it might even include the recommendation that everybody adopt this way of speaking. Note that, in this case, the revision is not organic so much as artificially determined and then, if particularly ambitious, outright mandated. Now let’s consider the explanatory theorist. Suppose that they have identified the following groupings based upon shared explanatory properties: G1, G2, and G3. Suppose that there is a term W (woman) that, in ordinary English, turns out to name the kind K. The theorist faces the decision of what to call G1, G2, and G3. For instance, they might simply call the groups “G1,” “G2,” and “G3” respectively. Say, however, that the theorist might also call G1 by the more apt expression “subordinated reproductive class,” and that they might call G2 by the more apt expression “individuals who experience certain norms F as relevant to themselves.” It is also possible for the theorist to call either or all these groups by the term W, in departure from ordinary usage. For instance, instead of using W to apply to K, the theorist may use it to pick out G1. The first worry should now be plain. The stipulative decision to apply W to G1 (i.e., the decision to stipulate that woman shall pick out a subordinated reproductive class), is a decision made by the theorist that need not be made. Why not just call this kind “subordinated reproductive class” or “G1”? Unfortunately, Jenkins does speak of the hegemonic kinds man and woman, using the term woman to name G1, and we are left wondering just why she calls these kinds that when to do so is clearly a theory- dependent revisionary move that need not be made at all. She says, “Some trans women are members of the hegemonic gender kind woman, whilst others are members of the hegemonic kind man” (Jenkins 2023, 128). Yet, to make this move actually runs counter to her own attempts to separate social ontology from semantics. To bring my concern into sharper focus, contrast Jenkins’s pluralist account with my own (Bettcher 2013, 2025). In my view, there can be different sets of social ontology-determining discursive and extradiscursive practices such that, in one community, trans people don’t count as men, women, or otherwise in ways that accord with their identities, while in other communities they do. In this case, there is express attention paid to the different semantics of the different systems. And as this is a descriptive project rather than an ameliorative or stipulative one, the charge that Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 11 the account itself engages in misgendering is hard to press, since describing an oppressive discursive system is not the same as using it. That is, while I allow for revision, this revision is theory-independent. To repeat, Jenkins’s decision to call G1 woman is a theory-dependent revisionary move. Since the decision to make that move rests with the theorist, they can be charged with misgendering. To be sure, it is not the theorist’s fault that the social kind G1 excludes some trans women, say. What is their fault is the decision to depart from common usage by calling G1 woman when there was no need to do so in the first place. Indeed, note that by itself, the fact that G1 excludes some trans women is not itself a case of misgendering unless we call that kind “W.” Ironically, then, the only misgendering, in this case, flows from the theorist who has so dubbed that kind woman. This result is by no means inconsequential. Jenkins (2023, 128) makes it seem that by analyzing the hegemonic gender kind G1, we can provide an account of the ways in which trans people are subject to this sort of invalidation. But the analysis of this kind cannot possibly provide an analysis of that, since not only have we decided to put the ordinary discursive practices by which trans people are invalidated to the side, but also the account itself runs counter to such practices. This is proven by the fact that Queen Elizabeth II would also not, in this sense, likely count as a woman. Clearly, any invalidation of Elizabeth’s gender identity is the effect of the hidden decision to dub G1 with woman, not the effect of the everyday social practices that lead to the invalidation of so many trans people. To put it simply: Jenkins’s account would not analyze the right kind, owing to a confusion between G1 and K. Now what likely underlies Jenkins’s decision to call G1 woman (besides the powerful tug of Haslangerian literature) is that, for her, the question is only whether the explanatory kind is in the right vicinity (Jenkins 2023, 117). For instance, certainly this hegemonic kind is in the ballpark of gender, as are interpersonal kinds and identity kinds. All of them “travel under the banner of gender,” says Jenkins (117), drawing on Rowan Bell’s (2022) expression. The problem, however, is that the terms gender and woman operate in different ways in ordinary usage. Gender is sufficiently vague and equivocal that it can be used to refer to gender identity (“Gender is between the ears, not the legs”) and it can also be used to a social role (“Gender is the cultural overlay of biological sex”). Neither of these uses constitute a departure from the common way of speaking. But the same is not true when it comes to the term woman. The use of woman in a Haslangerian sense is clearly revisionary since it entails that Queen Elizabeth II was likely not a woman. That is, while the term gender is sufficiently capacious, because sufficiently vague, to allow that class, status, and identity all qualify, it does not follow that the term woman operates in a similar way. No doubt one of the things that obscures this fact is precisely that, having put semantics aside, we have not paid Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 12 enough attention to usage. But the point is this. While it seems fine to list G1 as a gender kind (i.e., an explanatory kind that would be useful in explaining the operations of things we call gender), it does not seem fine to call it woman. This worry that Jenkins’s account is vulnerable to the basic misgendering can be generalized into a more general worry about the explanatory approach—namely, that in putting semantics to the side such an approach runs the risk of inadvertently using a term to name an explanatory kind in a way that is politically oppressive. To be sure, this is always a risk in almost any endeavor. But the conceit of discussing the metaphysics of social kinds shorn from the use of language—language about social kinds that is typically used in oppressive and/or resistant ways—can easily lead one to overlook the political pitfalls of one’s own theoretically dependent deployment of terms. Let’s turn to the final concern, which I regard as more important because, while the political risks in deploying theoretically dependent terms in oppressive ways are merely potential, the failure of the explanatory approach to answer what- questions is guaranteed. Consider the w-question “What is a woman?” The friend of the discursive approach says that answers to what-questions are largely determined by the ordinary discursive practices within a given cultural framework. By contrast, the defender of the explanatory-kind model of social kinds would presumably argue that answers to what-questions are determined by the explanatory kinds participating in the explanatory work. But the latter just seems wrong. The question is how on earth outlining the explanatory kinds G1, G2, and G3 helps us get an answer to “What is a W?” The first problem is that, obviously, sheer stipulation goes no distance to answering it. That is, identifying some grouping G and then stipulating that we shall call it W clearly does nothing in illuminating what a W is. To be sure, one might adopt a more ambitious ameliorationist project by arguing that this stipulation (using W to refer to G1) is the best way to go for a feminist project, say. That may or may not provide an answer to the what-question. The problem, however, is that ameliorationism has nothing to do with Jenkins’s actual proposal anyway, since we have put semantics to the side. Even the stipulative decision to call (or not call) the various groups W should be regarded as entirely irrelevant. We are therefore stuck with the question what the kinds G1, G2, and G3 even have to do with answering the question “What is a W?” It is hardly obvious. But any answer, it seems to me, is going to have to appeal to actual Ws, where W is deployed in the ordinary way. For instance, one might note that most Ws are G1s and that most G1s are Ws. Such a move, however, already reintroduces the semantics that were supposed to be put aside. And even if we overlook that fact, it isn’t clear how, given that we have now appealed to Ws (in the ordinary sense), learning about this relation to G1s has answered the question what Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 13 Ws are—particularly given that we have abandoned a universal identification of Ws with G1s. One response is to simply give up on answering such w-questions altogether. To do so, however, would be to give up on at least one important sort of project in social ontology: providing an account of the social kinds that are found within a given culture. For example, one might want to investigate the social kind fashion model, banker, or quarterback as they exist in our culture. In such a case, one may be attempting to answer the question of what a fashion model or banker really is—at least in the culture within which said social kind is embedded. To provide an account at this level, however, one would need to include the ordinary discursive application of the relevant terms among the various other extradiscursive practices involved in the constitution of the social kind. In this respect, semantics would be a required topic of investigation as part of a larger investigation into the pertinent social practices that yield the social kind. For example, in answering the question “What is a W?” one might argue that, surprisingly, W picks out kind K and that K is constituted through social practices P1, P2, and P3.5 Now it may be that the sort of project I am imagining is simply of an entirely different sort than the one suggested by the social kind as explanatory kind gloss. The latter seems particularly useful for the sociologist who wishes to posit various structural mechanisms at micro and macro levels (and everything in between) in order to explain various social phenomena. In this case, the sort of social kind deployed is really set by the investigator as a theoretical posit—a tool—for further understanding the phenomena. It is in this sense that Jenkins’s account is pluralist: multiple gender kinds are posited to do sorts of explanatory work. By contrast, the sort of approach I have in mind takes the social kind as the very phenomenon that demands explanation. As such, it is not open to the investigator to choose this or that explanatory property and then say that it somehow constitutes the cultural kind. Rather, what is required is an investigation into to the social practices—and, in this case, crucially, the discursive ones—to see how the relevant kind terms are applied to various humans, to determine what beliefs, true or false, the participants have about both the kind itself as well as their practices. (It is in this sense that my own account is pluralist. I take there to be different sorts of discursive and extradiscursive practices, and so, different answers to the question 5 It is an interesting question whether one ought to include the discursive practices themselves among the social practices listed. Another interesting question concerns the possibility of empty reference—that is, the possibility that kind terms like woman don’t pick out any kind of thing at all. The question here is whether an investigation into the discursive and extradiscursive practices would still count as providing an answer to the w-question “What is a woman?” Feminist Philosophy Quarterly, 2025, Vol. 11, Iss. 3, Article 1 Published by Scholarship@Western, 2025 14 “What is a woman?” as cultural kind). In this case, it may not be important whether the social kind in question is itself explanatory. To be sure, one would expect that the investigator’s own account be in some ways explanatory of the culturally embedded kind we are investigating. And that explanation may itself involve the positing of various explanatory kinds, including ones constituted through constraints and enablements. But an account that attempts to reduce the phenomenon of a cultural kind itself to one of these explanatory kinds may be giving up on the very thing that needs explanation. And that, in my view, would constitute the downfall of the account altogether. References Arvan, Marcus. 2023. “Trans Women, Cis Women, Alien Women, and Robot Women Are Women: They Are All (Simply) Adults Gendered Female.” Hypatia 38 (2): 373–89. Ásta. 2018. Categories We Live By: The Construction of Sex, Gender, Race, and Other Social Categories. New York: Oxford University Press. Bell, Rowan. 2022. “Gender Norms and Gendered Traits.” PhD diss., Syracuse University. https://surface.syr.edu/etd/1541. Bettcher, Talia Mae. 2013. “Trans Women and the Meaning of ‘Woman.’” In Philosophy of Sex: Contemporary Readings, 6th ed., edited by Nicholas Power, Raja Halwani, and Alan Soble, 233–50. Lanham, MD: Rowman & Littlefield. ————. 2017. “Trans 101.” In The Philosophy of Sex: Contemporary Readings, 7th ed., edited by Raja Halwani, Alan Soble, Sarah Hoffman, and Jacob M. Held, 119–37. Lanham, MD: Rowman & Littlefield. ————. 2025. Beyond Personhood: An Essay in Trans Philosophy. Minneapolis: University of Minnesota Press. Dembroff, Robin. 2018. “Real Talk on Metaphysics of Gender.” Philosophical Topics 46 (2): 21–50. Haslanger, Sally. 2012a. “Gender and Race: (What) Are They? (What) Do We Want Them to Be?” In Resisting Reality: Social Construction and Social Critique, 221– 47. New York: Oxford University Press. Essay originally published in 2000. ————. 2012b. “Ontology and Social Construction.” In Resisting Reality: Social Construction and Social Critique, 83–112. New York: Oxford University Press. Essay originally published in 1995. Jenkins, Katharine. 2016. “Amelioration and Inclusion: Gender Identity and the Concept of Woman.” Ethics 126 (2): 394–421. ————. 2023. Ontology and Oppression: Race, Gender, and Social Reality. New York: Oxford University Press. Lugones, María. 2010. “Toward a Decolonial Feminism.” Hypatia 25 (4): 742–59. Bettcher – Social Kinds and Semantics: Notes on Katharine Jenkins’s Ontology and Oppression Published by Scholarship@Western, 2025 15 Snorton, C. Riley. 2017. Black on Both Sides: A Racial History of Trans Identity. 3rd ed. Minneapolis: University of Minnesota Press. Spillers, Hortense J. 1987. “Mama’s Baby, Papa’s Maybe: An American Grammar Book.” Diacritics 17 (2): 65–81. Stock, Kathleen. 2022. “The Importance of Referring to Human Sex in Language.” Law and Contemporary Problems 85 (1): 25–45. https://scholarship.law.duke.edu /lcp/vol85/iss1/3. TALIA MAE BETTCHER is a professor of philosophy at Cal State Los Angeles. She is the author of the recently published Beyond Personhood: An Essay in Trans Philosophy (University of Minnesota Press, 2025) and many articles in trans philosophy, including “Evil Deceivers and Make Believers: Transphobic Violence and the Politics of Illusion” (Hypatia, 2007), “Trapped in the Wrong Theory: Rethinking Trans Oppression and Resistance” (Signs, 2013), and “What Is Trans Philosophy?” (Hypatia, 2019). 113001 Bettcher title page 113001 Bettcher Final Format