Telling Gender: The Pragmatics and Ethics of Gender Ascriptions

1. A Brief Introduction to Speech Act Theory

In order to understand how gender ascriptions organize social space, we must look at them from the point of view of speech act theory, which is the part of philosophy that takes linguistic utterances as actions that impact the social and material world, and analyzes how they function. We begin with a brief recap of speech act theory.

Philosophers frequently write as if the basic function of language is merely to convey information—to describe what’s true and false about the world. As such, they focus on declarative assertions such as “Paris is the capital of France”, or “Metals conduct electricity.” Key to the pragmatic function of such claims is that the standard for assessing their success is whether they are true or not.

Philosophers such as Wittgenstein and J. L. Austin initiated an alternative tradition in philosophy of language, emphasizing that utterances are actions and they can do all sorts of things other than convey information. Orders are not true or false; instead they seek to impose obligations on the people ordered. Invitations welcome someone to an event, and make it permissible for them to go. Marriage pronouncements make people married, changing their legal status. Promises bind the speaker to act in a certain way. None of these function primarily to convey information and none are true or false. We should notice a few features of these non-declarative speech acts.

First, they have concrete material effects. For example, a marriage pronouncement changes people’s tax burden. An order impacts which actions happen next.

Second, the characteristic way that such speech acts causally affect the world is by reshaping social norms. They alter what is permissible, obligatory, socially appropriate, transgressive, legal, and so forth. For instance, a baptism determines what it’s appropriate to call someone; a marriage pronouncement changes the married people’s status as a couple in social and legal space; a promise changes the promiser’s obligations.

Third, most of these speech acts have agent-relative authority conditions. That is, they can only be appropriately and successfully uttered by people with the right kind of social standing. If I am your professor, I am entitled to order you to write a paper by Tuesday, but not to clean your room. If I am not your parent, I can’t name you—one can’t name babies by running down the maternity ward shouting names at them. Trying to utter one of these speech acts without the proper authority will result in a misfire; not only will the speech act be unsuccessful in shifting social space in the way characteristic of that type of speech act, but it will fail to be intelligible as an attempted intervention into social space. Descriptive statements of fact are distinctive in that anyone can make them, but this is not the norm for speech acts.⁴

Fourth, for all these speech acts that are not declaratives, assessing whether they are appropriately uttered is not about assessing whether they are true or false. Rather, their appropriateness hinges on two questions: Whether they were performed with the proper authority, and whether social space should be organized in the way that they organize it. That is, whether an order, promise, or marriage pronouncement is appropriate is a combination of a question about whether it is performed with the proper entitlement so that it can have any effect at all, and a question about ethics, not truth. An order may be entitled or unentitled, and the obligations generated by the order might or might not be normative statuses that one should bring into being, but the order is not true or false.

Important for our purposes is the fact that, as Austin (1975) pointed out, some speech acts have the grammar of declarative assertions, yet they actually serve not to describe truly or falsely, but rather to institute a social state of affairs. Consider, from the chair, “The meeting is adjourned”—this sounds like a descriptive claim, but it in fact institutes the end of the meeting. Or, “I bet you five dollars”—this makes the bet, it doesn’t describe a truth. “That’s true” is an inappropriate response to either of these speech acts. Even an utterance like “It’s really hot in here” can, in the appropriate circumstance, function primarily as a request to open a window or turn down the heat, rather than a truth claim about temperature.

Finally, a given speech act can have multiple pragmatic roles. “Hey! It is really hot in here!”, is often both a description of the temperature, and a request to turn down the heat or open a window (and social context usually makes clear when this is so). One can see this from the fact that one can challenge it by denying the descriptive claim—“No it isn’t. I think you have a fever”—or by dismissing the request with, “I’m busy! Deal with it yourself!” But note that simply agreeing “Why, yes, it is hot” in such a case is a purposeful misunderstanding of the force, in that it misrepresents the act as merely a description and not a request.

2. Gender Ascriptions

Imagine a village with strict kinship relations and no tradition of regular adoption. Conditions in this society so conspire that there are virtually no cases of a child surviving both biological parents. But then one day one baby does survive an accident that takes her parents. The baby is taken in by another family and raised as their own. In time, this situation comes to the attention of the village and the following conversation occurs:

Wait, so these are not your parents?

Yes, these are my parents. (Parents agree.)

No, they are not your parents. They took you in but they are not your real parents.

They are my real parents! I am their child!

Clearly what is at issue here is not a debate about the empirical facts of the case, all of which are agreed upon. What is at issue is what kind of uptake this relationship should be given by others in society and how the child is to understand and narratively interpret her own life. She is insisting on using, and demanding that others use and accept, the norms of parent-child for her relationship. The utterance “I am their child” is a call for a type of recognition as being placed within social space in a particular way, not a denial of an empirical claim.

Consider another example: If I call you my friend, I am seeking to establish a social status between us. I am not telling you what empirical properties you or I have, but rather letting you know how I will treat you and how I expect you to treat me. There is no strict set of necessary and sufficient conditions for how friends should treat one another, any more than there are such conditions for behaving like a man. But in calling you my friend, I am proposing that we occupy relational roles that will give sense to and normatively inflect all the ways we relate. If you say “No, I am not your friend”, then we are not having an empirical disagreement, in the first instance. Rather, your negation is a rejection of my attempt to establish our relationship in social space. If I cite facts about me or you to try to convince you that you are empirically wrong and I am right, I’m missing the point of how language functions here. My speech act was a proposal, not a description, and yours a rejection of that proposal. And even if future science were to discover some complex physical relationship that explained why people become friends, or what their interaction is like when they are friends—maybe something to do with pheromones?—it would be irrelevant to this point. Saying “you are my friend” does not describe this fact. It remains a proposal and the undertaking of a practical commitment.

Our central claim is that similarly, gender ascriptions are not primarily declarative claims, although they typically have declarative grammatical form, and that disagreements over gender ascriptions are not primarily disagreements over empirical facts, but rather social negotiations over how someone will be positioned within social normative space. We claim that gender ascriptions function in the first instance to alter and reorganize social space, not to describe antecedent reality. We are claiming that calling someone a man is more like calling him a friend or saying “I bet you five dollars” than it is like calling him tall. They are best assessed, not in terms of their truth or falsity, but in terms of whether they were performed with the proper authority or not, and whether their effects are ethically appropriate or not. We will take some time explaining and defending this claim.

Notice that what gender we are taken as having inflects nearly every aspect of how we are expected and demanded to negotiate the social and material world. It shapes how we are supposed to hold our body and modulate our voice, what clothes we are supposed to wear, how we are supposed to manifest sexual attraction and attractiveness, where and how we pee, what hobbies and jobs we are supposed to have, who we compete against in sports events and which sports we take up in the first place, what our relationship is to our children, and so forth. Even fetuses, once recognized as ‘boys’ or ‘girls’, are expected to become babies for whom certain nursery and clothing colors and emotions and behaviors are appropriate. Such norms are modulated by race, age, ability, class, body shape, and more—there is not just a single set of norms for each gender, but rather a complex and often contradictory web of norms in which we are differently positioned—but these structures of social significance are inescapable.

Of course, we can violate almost any of these gender norms—people recognized as women can take masculine-coded jobs, people recognized as men can wear makeup. We can try but fail to live up to the gender norms that are perceived as applying to us, or we can defy them. We can even try to change them by resisting from within. But when we do, it is clear that this is socially coded as a transgression—a pushing back against what is expected. Katharine Jenkins (2016) argues that experiencing a set of gendered norms as applying to and making demands on oneself, even if one resists or fails to live up to these norms, is what it is to have a gender identity. This is a metaphysical claim, upon which we remain neutral. We are making the weaker claim that being recognized as having a gender (by oneself or by others) involves being positioned in normative space; the social recognition inserts us into a complex web of normative expectations and pressures. This is so whether or not it turns out to be the case that on the proper metaphysics of gender, having a gender inserts us into norms in this way. Moreover, and crucially, being socially recognized as having a gender normatively positions us in this way, regardless of the underlying metaphysics that anyone might appeal to in deciding whether to ascribe a gender to us, and regardless of whether we in fact have the gender we are recognized as having.

We take the point about the variety and ubiquity of gender norms to be fairly obvious. The important point for us is that what we do when we ascribe gender in language (to ourselves or to someone else) is give communicative recognition to the fact that someone is placed within gender norms, and by doing so, help to insert them into this normative place, and hold them there via mechanisms of social accountability. Even if someone feels like a set of norms applies to them, they will not be socially held to those norms until they are recognized as subject to them, and this is what we use gender ascription speech acts for. This, again, is so regardless of anyone’s underlying metaphysical explanation for why the person is situated within the norms as they are. What holds people in and to these norms is our own and others’ recognition of us as being bound by them. We may feel the pull of a set of gender norms privately, for instance if we are trans but not in any way yet public about that identity (see Jenkins 2016). But we are not held to the norms that we feel the pull of with any kind of concrete accountability until there is some public social recognition of our gender (Shotwell 2011).

Hence gender ascriptions are socially potent speech—just as marriage pronouncements insert people into the norms of marriage, and betting someone $5 establishes that you and they are bound by the norms of betting, so gendering someone in language places them in a gendered social space; this is a performative effect of gender ascription. Telling someone that they have a gender does not merely state a purported fact about them, but rather has complex illocutionary force–it inserts them within this set of norms and expectations, and thereby has concrete consequences for what they can do, where they can go, whether others will accept them, etc. Katharine Jenkins similarly argues that gender ascriptions are “moves in a norm governed activity”; she claims that “utterances such as ‘Alex is a man’ are covert exercitives that enact permissibility facts concerning the appropriate treatment of the individuals in question, namely, that they must be classified as men and not classified as women.” (Jenkins 2021: 25, 26).⁵ We agree, although we think the set of normative status changes instituted by gender ascriptions is richer than just a set of permissions.

Indeed, given the force and persistence with which we are placed into gender norms, and the amount of hard work and disciplinary action it takes to keep people conforming to gender norms, which almost no one does perfectly, one might say that gender ascriptions jam people into gendered social positions, and hold them in there really hard. Of course, no one gender ascription slams us into a complex normative position in a single moment—although gender reveal parties for fetuses do try hard to accomplish this! Rather, each gender ascription recognizes how a person is already (purportedly) entangled in normative space, and in doing so further entrenches and holds them in this space. This is another similarity between “I am a man” and “I am your friend,” since friendship also can’t be instituted with a single speech act, once we leave preschool. It is partly because imposing gender norms is difficult and requires so much force that we need so many repetitions and iterations of gender placement in order for the social positioning to stick, as Judith Butler (1990; 1997) argued.⁶ (This may explain the odd apologetic panic that people manifest when they accidentally misgender someone, even including a baby. Upholding gender positioning takes a lot of work, and mistakes that undo some of that work are not tolerated lightly.)

At this point, someone might object and say: Our gender ascriptions recognize which norms apply to a person, but they do not insert them into those norms; it is biology that makes a set of norms apply to a person, and all our gender ascriptions do is reflect the normative facts that result from biology. But it can’t be our biology alone that inserts us into norms, for two reasons.

First, since the details of gender norms vary substantially from culture to culture and generation to generation, and across the lifespan, biology alone cannot determine which social norms apply to us. Notice that this is a different point from whether biology causes us to have a gender identity, which is a metaphysical or perhaps an etiological claim about which we remain neutral. Regardless of the underlying relationship between biology and identity, biology doesn’t change nearly as much or as fast as gender norms do, so we need something other than biology to hold us in these norms.

Second, even if biology determines some of the gender differences between us, it would determine what we do (or are disposed to do), and not what we should do. It is a straightforward naturalistic fallacy to assume otherwise. The norms that gender ascriptions recognize us as embedded in are social norms. Whether or not biology causes us to behave in various ways or to have various gendered properties, our social accountability for our place within sets of norms cannot be reduced to the biological fact that we are inclined to live up to these norms, if this is a biological fact. We are held to the norms through social recognition, including self-recognition in the case of self-ascription. In this sense they are like marriage norms, tax norms, and other socially enforced norms. Even if our social norms mirror or riff upon biological differences, we still need social recognition to hold us to these norms; biology on its own has no capacity to hold anything accountable to anything.

One might also object to our argument so far by insisting that gender ascriptions insert people into norms by describing them. In saying someone is a man, I am attributing some empirical properties to him—perhaps biological, or perhaps social or psychological. Because he has these properties, he is subject to the relevant norms. Thus gender ascriptions are still declaratives; they are just declaratives with normative consequences. In contrast, we claim that the main thing you are doing with language when you identify someone’s gender is not pointing at an empirical property, but setting social expectations for how they will act and present themselves, shaping the narrative paths open to them as they move through material social space.

Declarative claims about our empirical properties can have normative implications, depending on context. Having a grandparent who is a barber seems like a neutral fact about a person, but in a society in which occupation is rigidly inherited, being known as the grandchild of a barber might be an important marker of the norms and expectations that apply to one. If you compete in a weight-classed sport, then you will gain and lose entry to various events and competitors depending on the empirical fact of your weight, and ascribing a weight to you is declaring this empirical fact, which then has normative consequences in your case. But there are other things that can be ascribed that are normatively inflected all the way down; to ascribe them to someone is to insert them into a location in social normative space. To call me a friend is not in the first instance to call attention to empirical facts about me and then to make normative inferences from those facts, but to impose expectations for what I should do, what my obligations are, how it is appropriate for me to behave, what counts as social success for me.

There are social identities and properties that come with various empirical expectations, but for which there is no fixed substratum of empirical properties upon which they depend. To be socially recognized as a DJ, for example, makes it defeasibly reasonable for others to expect that one will keep late hours, love and understand some types of music, and so forth. It makes evaluations in terms of how well one DJs salient. It gives narrative structure and comprehensibility to one’s life choices and practices: Moving from Indiana to Berlin to look for work, say, or working on one’s DJing technique. This identity guides not only our self-understanding and our path through life, but the way society gives us uptake. But there is no determinate set of empirical properties one must have to be socially recognized as a DJ. One can be recognized as a DJ who is out of work, or on vacation, or lacking in talent, or indeed who has never worked as one at all, especially if one publicly claims this as one’s identity. Similarly, we think it was Ernie LePore (with apologies for the misattribution if we are remembering wrong) who pointed out that there are no empirical facts or properties that are necessary for claiming that one is “working on a book.” Being socially recognized as a friend, a DJ, or one working on a book is about bringing a set of social expectations and sense-making resources to bear—it’s about being recognized as occupying a location in normative social space. It’s not about being attributed a fixed set of empirical properties, whether internal or external. (And yet, notice, one can perfectly well lie about being a friend, DJ, or book-writer, so these identities are not simply metaphysically equivalent to their social recognition.)

When it comes to gender, we claim, there is similarly no agreed-upon fixed substratum of empirical properties that anchor our claims to social space or our disagreements over how to position people in social space. Claiming a gender, and receiving social recognition as having one, is more like claiming to be a friend, DJ, or book writer than it is like claiming to have type AB blood or claiming to be in pain. Whether or not there is actually a determinate fact about what gender is and how you get one, there is no general agreement about what properties make up someone’s gender. Gender is taken by different people in different contexts to lie in anything from genitals, to DNA, to reproductive organs, to hormones, to what people feel their identity is on the inside, among other options. What we agree about, when we agree someone is a man, is generally not his genitals, or his DNA, or his inner sense of identity, all of which are typically hidden from us anyhow, but about how he should be positioned in normative space. Conversely, if we find out that someone doesn’t have a penis, say, and we now disagree over whether he is a man, we are not disagreeing about the facts about his genitals (ex hypothesi), but rather contesting which norms should apply to him. Even if you insist that what you are doing with your words when you ascribe gender to someone is making a descriptive claim about which biological, social, or psychological properties they have, the primary social effect of your words is not to inform people of these properties, but to help insert and hold them in social norms. Likewise, if you are pervasively recognized as being a man, then you are subject to the norms of masculinity, as a matter of social fact, regardless of the facts about your psychology or biology.

Two people might in fact disagree over whether someone is a man because they disagree over whether having a penis is necessary (and maybe even necessary and sufficient) for being a man. They might agree that Alfred has no penis, but disagree over whether Alfred’s claims to be a man should be given uptake, because one thinks that having a penis is a necessary condition for occupying the social role of being a man, whereas the other does not. In turn, this might be because they disagree about the metaphysics of gender. But notice that in all but some highly specialized contexts, although they may disagree because they have metaphysical differences, they are not disagreeing about the metaphysics—they are not having a metaphysical argument. Rather, they are arguing about how to give social uptake to Alfred. What their contrary gender ascriptions are doing is competing to position Alfred differently in social space. Indeed, if they move to debating about metaphysics, they will likely move away from gender ascriptions and switch to making general statements about men and body parts, which is a different use of language altogether, and not the one we are analyzing here.⁷ Another clue that the competing gender ascriptions do not constitute a metaphysical argument is that it is completely consistent to privately believe that someone has a particular gender because of their empirical properties such as their genitals, and also believe that social claims to gender positions and identities should be routinely respected even when the gender identity claimed does not match the actual gender. In such a case, the person whose genital-essentialist metaphysics causes them to believe that Alfred is not a man will still use language to ascribe manhood to Alfred, regardless of their private metaphysical beliefs. So, regardless of one’s metaphysical commitments, speech acts that ascribe gender always constitute acts of social positioning. Even when these are based on an empirical judgment and an implicit metaphysics, they generally do not function to litigate that judgment or metaphysics.

We have emphasized that we are not claiming that having a recognized social role as a man is what it is to be a man. We are remaining neutral on the metaphysics of gender. We are, separately from our argument in this paper, committed to the idea that any successful metaphysical account of gender would have to include trans men who are not socially recognized as men as actual men (and mutatis mutandis for trans women and for nonbinary people). Since plenty of trans men do not get this social recognition, either because they are not out or because they do not pass, we do not believe that any form of pure social recognition account of gender can be correct. We mention this just to emphasize that in claiming that gender attributions hold people in gender roles, we are not claiming (contra a possible reading of Butler 1990 and Àsta 2018, for instance) that these attributions constitute gender identities. We are claiming, instead, that it is public recognition of a social role that in fact holds people to the norms attached to that role, regardless of what gender they actually have, or even whether the notion of having a gender is ultimately a coherent one. We leave open the relationship between gendered social positions and genders.⁸

There are specialized contexts in which gender ascriptions do have primarily descriptive force, and there is general agreement as to which descriptive features of people the ascriptions are picking out. Doctors’ offices are a good example. Describing someone as a woman, in that context, generally serves to classify them on the basis of which anatomical parts they have (though not, interestingly, on the basis of their genetic makeup, except in very unusual circumstances). Doctors need to know whether someone needs a pap smear or a prostate exam, and we use ‘woman’ and ‘man’ (or ‘female’ and ‘male’) on forms in order to make such classifications, and talk about “women’s health clinics” and the like.

But notice that although we do use gender ascriptions descriptively in such contexts, arguably we shouldn’t. First, even in such cases, describing is not all we do socially with such ascriptions. This is already clear from the effect it has on trans and nonbinary people. The practice results in their either being misgendered in such medical contexts, or being shut out from appropriate doctors, or having to undo the effects of a form that renders them at least temporarily medically unintelligible. In some contexts, medical classifications by gender have been used to restrict social standing—for example, medical tests imposed by athletic governing bodies that determine who can participate in gendered sporting competitions. So even in cases where the specific interaction is one seeking empirical information, one cannot escape the broader social context and, hence, the richer pragmatic implications of a declaration, even on a medical form, that one is a man or a woman.

Second, the terminology is unnecessarily imprecise. What doctors (should) care about is not whether a patient is a woman, with all the potential variation and complication that entails on any account of what gender is, but simply whether they have the relevant anatomy for a test or procedure. Not all women—not even all cis women—have cervixes, and some men and nonbinary people have cervixes. Hence it is simply more precise for forms and questions in doctors’ offices that offer pap smears to ask people if they have a cervix than what their gender is.

Sometimes we talk about gender in order to describe someone’s socially assigned position. If claims like “I am a man” or “she is a woman” serve to assign social positions, we can sometimes describe the positions so ascribed. Thus, one can say things like “Because she is a woman, she was expected to be nicer in the boardroom than others.” One is here describing the gendered social uptake of a person, and using that to explain a form of oppression; these expectations hold because of how she is socially recognized, not because of her (hidden) anatomical parts, her DNA, or even her inner sense of identity. Since what we mean to say in such a case is that the person is treated differently because of how their gender is given social uptake, why not say that? The above explanation is simply made more precise if we say “Because she is taken to be a woman, she was expected to be nicer in the boardroom than others.” Would anyone proffering such an explanation, who believed in a real fact about one’s sex, think that it did not apply in the case of a man who was widely taken to be a woman? It seems obvious that social prejudice functions on the basis of social perception, not hidden realities. Meanwhile, by saying the less precise thing, we also contribute to positioning this person in social space inappropriately.

The argument generalizes. If there is some empirical property of a person that we can, in context, pick out with a gendered ascription, then one can do whatever descriptive and explanatory work one is doing in that context more clearly and without extraneous pragmatic effects, simply by forthrightly and explicitly talking about that empirical property. If you want to talk about genitals, talk about genitals. If you want to talk about DNA, talk about DNA. If you want to talk about social role or psychological senses of self, do so, clearly and explicitly. Eliminate the unusual and always semantically imprecise and pragmatically ambiguous uses of gendered terms for that purpose. And again, the argument for this eliminativism of descriptive gender-talk applies regardless of what facts you think constitute gender. Gender ascriptions always have potent social force and powerfully place people in normative space, which we should not be casually doing every time we want to pick out a descriptive property of a person that is typically associated with a gender.⁹

One might argue that perhaps we should always be substituting gender ascriptions with more precise truth claims about what social role someone is recognized as having. If one reads gender ascriptions as truth claims, then this would seem, on our view, to be simply a less misleading and more precise way of using language. But it would in fact change the pragmatic function of language, and not just precisify its semantics. Inserting people into norms is a different function for language than claiming that they are already in them. Now, one might think that inserting people into gender roles is just plain a bad thing to do, and that we should eliminate gender ascriptions on that basis. (Dembroff & Wodak 2018 and Dembroff & Wodak 2021, for instance, argue for the elimination of almost all gender markers in speech.) Certainly, we argue below, it is a bad thing to do to someone against their will. We are very open to the possibility that in some future utopia, gender roles will be antiquated and merely seen as an oppressive vestige of the past. Surely, in any ideal society, opting out of gender altogether should be an easily accessible possibility. But at least for now, we think, gender self-ascriptions are powerful tools that many people use to gain access to a range of interactions, sense-making resources, possibilities for self-expression, social opportunities, and forms of recognition and permission. Likewise, second- and third-person gender ascriptions that pick up on and give uptake to these self-ascriptions play critical roles in making this access a social reality. Hence to call for their elimination at this moment in history is to sacrifice many people’s well-being for the sake of a possible future.¹⁰

One might also think that although having linguistic tools to insert people into gendered norms is (still?) useful, it is a problem that gender ascriptions have the misleading surface grammar of truth claims, and look like they are just attributing empirical properties. This isn’t a general problem: We don’t have any trouble understanding that although “I bet you five dollars” has the grammar of an empirical truth claim, it is in fact functioning to institute social normative statuses. But it is plausible to think that gender is vexed and potentially oppressive enough, and there is enough obfuscation over the function of gender attributions, that we should just do away with this misleading grammatical form. This is perhaps true; our primary goal is to analyze how language currently functions, but we return to the dangers of the slippery and misleading nature of gender attributional language in our conclusion.

3. Correctness of Gender Ascriptions

If it is true that gender ascriptions primarily serve to place and fix us within social normative space rather than to describe us, then, accordingly, the primary way to assess their appropriateness is not in terms of their truth value. Instead, in assessing the appropriateness of a gender ascription, we need to ask whether the speaker has the proper authority to insert the subject of the ascription into a particular social position, and whether the reorganization of social space effected by the ascription is ethically acceptable. These are substantive questions! There is no obvious answer to the question of when someone would have the proper authority to be entitled to insert someone into a complex social position with their words, nor what the ethics of doing so are.

To begin to answer these questions, notice that it follows from basic norms of self-determination and bodily agency that we ought to have defeasible control over how we are placed within social space and social norms. Of course this is not an undefeatable norm: If someone declares “I am your sovereign!”, even if this is deeply psychologically important to them, accepting such a status would impose unjust restrictions upon others. But in general, assigning other people a tightly controlled social position they have not chosen for themselves is pretty much the definition of social repression—this is why we consider caste systems, hereditary occupations, nonconsensual arranged marriages, Jim Crow laws, and so forth to be fundamental assaults on human rights. We take it as a basic tenet, then, that as long as it isn’t hurting others, and consistent with our capacities, we should be able to determine our social position for ourselves. And gender norms are extremely detailed and control us at a basic bodily level—if they are inappropriately imposed, this seems to be an assault on our core bodily integrity and autonomy. As we pointed out above, people are not just gently nudged into gender norms, but jammed into them and held there hard with the full force of social and discursive power, so that gender ascriptions (typically not singly, but as a repeating social pattern) are no mere trivial intrusions into our self-determination.¹¹ This is so especially but not only if one is trying very hard to extricate oneself from an imposed set of norms, for example because one is transgender or agender.

Gender ascriptions can be in the first-, second-, or third-person voice. If gender ascriptions were declaratives, then they would have the same content and force regardless of voice; they would just state the fact that some person has a particular property. But we maintain that these three types of ascriptions in fact function as substantially different speech acts that are authorized in different ways and have different performative effects. While all three reorganize social space by placing someone within a set of norms, they do so in different ways, and are not all equally entitled, even given the same empirical facts in the background. That is, what entitles me to announce my own gender, and what entitles you to tell me what my gender is, and what entitles someone else to report on my gender, are all different conditions. Moreover, the ways in which they reorganize social space are also different. For the remainder of this paper, we give a more fine-grained analysis of the pragmatic structure of gender ascriptions, distinguishing between first-, second-, and third-person pragmatic voices.

4. First-Person Gender Ascriptions

When protestors in the 1968 Memphis sanitation workers strike carried signs reading “I am a man” they were not describing their genetic composition (either as human or as men.) They were not primarily asserting any empirical similarity between them and other more privileged men. They were making a demand for a specific kind of uptake by society at large. Such demands, and related requests, proposals, etc., are common. Someone walks to the front of a classroom the first day of class and says “Welcome students, I am Professor Jones”; one person pleads with another “I am your friend!”. In these cases there are empirical considerations that are relevant to the appropriateness or success of the speech act. One is entitled to take on the role of professor only if one has certain credentials and is duly appointed by the relevant educational institution, and so forth. But the primary function of such an introduction to students is not to assert that these credentials apply. It is to call on the others in the room to give one a particular sort of social uptake, to claim a role. It is a call to the audience to adopt—in this context, for this time—the role of student and to give uptake to the utterances of the speaker as a professor, with all the relevant dimensions of authority that entails. If this stance is adopted, then one will find it normal for the person so taken to begin a lecture, to make assignments, etc. and the invocation of such normative positionality is the primary point of the speech act, not description.

First-person gender ascriptions likewise lay claim to a social position. They function to request—or sometimes to entreat or demand¹²—of others that they take the speaker as held to a set of gendered norms, even if the person self-ascribing in fact wishes to defy, play with, or fight those norms. Likewise, first-person rejections of a gender ascription (including rejections of any gender ascription, in the case of some nonbinary folks) request/entreat/demand of others that one not be taken as subject to a set of gendered norms.

Because we have a prima facie, defeasible right to determine our own social location, as part of our basic right to self-determination, these speech acts are prima facie entitled. We need no special ground for their authority beyond the general insight that people should get to choose how they are positioned in social space—a right that, in the case of gender, is integrally connected to our right to control the motions and boundaries of our own bodies. First-person gender ascriptions may make different sorts of claims depending on their audience: Self-ascribing or rejecting a gender to the world at large functions to claim a place in public gendered social space, whereas self-ascribing or rejecting a gender to a trusted friend confidentially may seek to establish a specific set of normative relationships with that friend, while self-ascribing or rejecting a gender as an act of solidarity with someone else has different pragmatic effects again.¹³ But in each case, the first-person language serves to make a claim on a position in normative space, whether this is public or private space.

Robin Dembroff and Cat St. Croix (2019) coin the term “agential identities” to refer to the identities to which we give public expression, and make available for social recognition. Such identities, according to their account, may or may not correspond to “self-identities”, which are the identities we have regardless of which identities we publicly claim, and “social identities”, which are the identities we are socially recognized as having. On the picture we have been developing, first-personal gender ascriptions, as a matter of pragmatic form, claim agential identities, and in doing so, demand or request the bestowal of social identities. Here, in keeping with our desire to give a pragmatic account that is independent of any metaphysical account, we stay neutral on the relationship between agential identity and self-identity, or the nature of self-identity. But importantly, as Dembroff and St. Croix point out, claiming an agential identity is a structurally first-personal speech act; agential identities are what we claim for ourselves and seek social recognition for. No one else can attribute an agential identity to me; if they try, they will end up attributing a social identity to me instead. Thus, the act of claiming an agential identity is not an act of making a traditional truth claim. It is not a report on a self-identity (as Dembroff and St. Croix sometimes seem to be claiming), since reports have agent-neutral content and anyone can make them. Rather, it is a petition for a certain normative social position, which calls for recognition. The agential identities that we ‘have’ cannot be determined by our internal or social properties, but only by the kinds of first-personal speech acts (broadly construed to include all sorts of communicative performances) in which we engage.

If I say “I bet you five dollars”, my speech act presupposes a number of descriptive beliefs, such as that I will turn out to be right about whatever I am betting about, and that I can afford to gamble five dollars. But even if my factual presuppositions are wrong, I really did make the bet. Likewise, if I announce, “I am a man!”, even if my factual presuppositions are wrong, I really did claim a masculine position in social space, and it is almost always my right to do so. I am the one who gets to express where I belong in social space, when it comes to something as intimate and pervasive as gender. My claim that I am a man, then, is not the kind of thing that can be straightforwardly deceptive, any more than my claim that I bet you five dollars can be, because the speech act helps create reality rather than just reflecting it. My claim might mislead because it may suggest facts about me that aren’t true, but it cannot simply be false. This is so even if I take myself to be making a descriptive claim with my self-ascription; intentions do not dictate the performative force of my speech act. Regardless of my intentions, the primary social effect of my words is to insert myself into social space.

Here we are agreeing with Talia Mae Bettcher (2009), who argues that first-person gender ascriptions are not reports, but what she calls avowals, which are expressions of identity that are immune from straightforward contradiction. Their immunity is not due to their being infallible truth claims (for people can certainly be wrong or insincere about any empirical features of themselves that may constitute their internal identities), but rather grounded in their ethical and political status as expressions of self-determination. We take ourselves, in this section of the paper, to be fleshing out the pragmatics and social functioning of Bettcher’s avowals, where what is avowed are Dembroff and St. Croix’s agential identities.

Such first-person avowals of agential identity should almost always be taken as entitled, given our prima facie right to control our own bodies and social locations. My first-person gender ascriptions should be respected as a matter of ethical and political respect for my self-determination, and not because I always automatically have the facts right about myself and speak sincerely. The point here is not that any of my self-descriptions should be respected on political grounds, even though they might be false or insincere. The point is that what I am doing is not primarily self-describing at all, but instead positioning myself in normative space. So, although I might perfectly well be wrong about any relevant facts about myself, I am not making a claim about these facts in my first-person ascription.

If someone disagrees with my right to ascribe my own gender, their disagreement should be on ethical, not factual grounds—that is, because my claiming of this social position somehow does harm to others, not on the grounds of my having gotten some set of empirical facts wrong. If I am wrong to call myself a man, it is not because I made a false claim, but because I am trying to do something I ought not to be allowed to do. In the case of claiming a gender, it is very hard to cook up cases where this would be so. My claiming a gender identity (unlike my claiming to be a sovereign or a spouse) is no imposition on anyone else.¹⁴ And because of our basic rights to self-determination and bodily integrity are wrapped up in our ability to claim our own position in gendered space, the bar for such harm would have to be relatively high.

5. A Note about Race

Before moving on to second-person gender ascriptions, we need to say something about what might seem to be a parallel case to first-person gender ascriptions, namely first-person race ascriptions. Several people have asked us if our account commits us to respecting all first-personal identity ascriptions, including in particular race ascriptions, even when those self-ascriptions may seem to contradict our social standards for racial classification. We have two responses to this query.

First, we are giving an analysis of gender ascriptions specifically, and we see no a priori reason to assume that the analysis of race ascriptions will be tightly analogous. Indeed, there seem to be important disanalogies between the two cases. These are mostly beyond our scope here, but notice for instance that we typically assign race based on intergenerational features, and likewise, the harms we wish to undo when we think critically about race and engage in anti-racist action are harms that come from generations of sedimented structural disadvantage and compounding lack of access to social goods, which cannot be understood individually. Thus racial ascriptions implicate the identities of more people than just the person to whom they apply, whether they are first-, second-, or third-personal. To return to our earlier point about the limits of self-determination, claiming a race may turn out to be more like claiming that you are someone’s spouse or sovereign than it is like claiming a gender. We do not intend this point to constitute an analysis of racial ascriptions, but just to highlight one among many reasons why the assumption that race and gender ascriptions have analogous pragmatics and ethics is unwarranted.

But, second, we do feel confident (though we don’t argue for it here) that there is a sense in which the two cases are analogous: Race ascriptions, like gender ascriptions, are not fundamentally declarative truth claims whose appropriateness depends on how they correspond with some stable set of empirical facts. Rather, both are normative moves that attempt to claim social space or place someone within that space. Likewise, both should be assessed on ethical and pragmatic grounds: Was this speech act properly entitled, and should this normative move be made? If first-person race ascriptions should not be given prima facie, nearly indefeasible deference the way gender ascriptions should, this is not because they are less likely to be true, but because giving them recognition is more likely to do ethical harm, or because their entitlement conditions are different. Whether someone should get to have whatever racial agential identity they want, and receive social recognition for it, is a normative question about pragmatics, politics, and ethics, not a question about the metaphysics of race (for after all, that metaphysics is arguably even more contested and murky than the metaphysics of gender). Hence in this birds-eye sense, we are willing to say that race and gender ascriptions are analogous, even if the concrete ethical and pragmatic stakes turn out to be quite different in the two cases.

6. Second-Person Gender Ascriptions

First-person gender ascriptions, we have argued, are petitions to occupy a social position; they claim social space. But such positions within normative space require social recognition in order to be solidified as real. If I claim a position in normative space and get no uptake, then nothing holds me in this social place. Thus first-person gender ascriptions require uptake in the form of second-person gender ascriptions and recognitions, and these play an essential role in constituting social identity. Indeed, we have argued in our past work (such as Kukla 2014 and Kukla & Lance 2009) that all speech acts require concrete second-person uptake in order for them to have the norm-shifting force that they strive to effect. Self-identification on its own cannot constitute a gendered position in social space. However, we argue in this section that for ethical reasons, second-person identifications should not on their own constitute gendered positions in social space either.

If I make a claim about an objective matter of fact—“There is a mouse in the room”—and you say “Yes, there is a mouse in the room”, you are agreeing with me. You take up a position in discursive space as my rhetorical ally. Such an act has normative implications—for instance, it is now impermissible to say that my claim is false, or to assert anything incompatible with it. But when I make what is primarily a claim to a position in social space, your agreement or disagreement serves a different function. When you agree with my claim, “I am a man”, you are not primarily committing to treating that statement as true, but to taking the norms of manhood as applying to me, allowing me access to a rich social position, and supporting me in that position by giving me uptake as appropriately belonging in it. And if you disagree with me, then you are rejecting my social demand.

Imagine that in the midst of a heist, I say, “I’m taking charge here”, and you reply, “You are not the leader of this operation!” As in multiple other examples we have looked at, this is not an empirical disagreement. There are no separate facts of the matter that determine whether I get to take charge of the operation. Rather, my claim that I am taking charge is not a description of fact but a social proposal. In turn, your response is a challenge to my proposal—a rejection of my attempt to organize social space and my place within it in a certain way. So too with a second-person denial of a first-person gender attribution. This is first and foremost an act of refusing to engage with the speaker in the social mode that they propose. It is a repudiation of their call for social uptake—a concrete push-back against their attempted self-placing in social space. Likewise, if you respond with, “OK, you’re the leader”, or “Yes, you are a man,” you are not just describing an empirical fact, but giving me social uptake and recognition that places me in the social location that I just petitioned to enter.

From the point of view of speech act theory, if second-person gender ascriptions have these normative effects, we need to ask, (1) who has the authority to impose these norms with their speech, and (2) when is it ethical to do so?

Generally speaking, people need special social credentials to shift social space with their words, and particularly to shift someone else’s status in social space. Only specific people can draft me into the army, establish a new law that turns me into a criminal, or pronounce me married, and likewise only certain people have the social power to remove me from these social locations or deny me these statuses. So we need to ask, what social credentials could authorize someone to impose gender norms on me or deny me a place within them? We also need to ask whether this can be done against my will or without my autonomous acquiescence. Some changes in status require the autonomous participation of the person whose status is being changed and some do not. Someone with the right credentials can draft me into the army or impose new laws on me, but no one (in the United States) can marry me against my will, or enter me into a bet against my will, or consent to sex on my behalf, for instance. So in the case of second-person gender ascriptions, the question is two-fold; what authorizes these speech acts, and can they be authorized without first-person autonomous participation and consent?

In the case of drafting someone into the army or making them a professor, we have clear institutional standards for who has the entitlement to perform the speech acts that institute these statuses. It is notable and ethically crucial that there are no such institutional authorities when it comes to gender credentials. Indeed, when doctors or parents or nosy outsiders take such authority upon themselves, we think, they are helping themselves to a kind of ersatz institutional authority that doesn’t really exist and whose mimicking does harm. We suggest that typically, the only thing that can authorize a second-person gender ascription is a first-person gender ascription. Remember, such second-person ascriptions are not just statements of fact, so they are not just repetitions, but performative acts of giving uptake to and solidifying someone’s claimed position in normative space. Further, all second-person speech acts are speech acts directed at the person in question. Thus, to deny someone’s first-personally embraced gender is not only to seek to impose a social position on them, but to do so to them, to directly reject their self-positioning.

Second-person disagreements with first-person gender ascriptions are, then, almost never justified. There is just nothing that would give a second person the kind of social authority they would need to use their words to overrule a chosen role, and to foist someone into a complex web of social norms, unless invited/demanded/requested to do so by the first person. On the face of it, since being gendered tightly constrains our autonomy and our bodily self-determination and agency, and since we value freedom over our social position, getting to self-ascribe our gender is a basic autonomy right. We would certainly need an argument as to why anyone else could impose something like this on us without our agreement. At least the burden of proof is on someone to show why they have the standing and right to do this, to override my basic social and bodily presumption of self-determination, especially when there are grave social and psychological consequences at stake.

Some people are genuinely unsure of their gender identity and have made no solid first-person ascription. In this case too, a second-person ascription such as “You are a woman!” is inappropriate. It lacks the direct hostility of a rejection, but retains a morally indefensible presumption. It is no one else’s place to announce to someone how they are to be positioned in social space.

This is not to say that gender self-identifications must always be accepted without question. But even in the rare cases where one does have good reason to push back second-personally against someone’s gender identification, flat-out denial is not a decent or entitled response. If we had good reason to believe that someone was going to be miserable pursuing the life of an athlete, if it was incompatible with other deeply held self-identifications, some careful push-back by the right person in the right way and in the right context could be called for. But even here, it is hard to imagine “No! You are not an athlete!” being a decent response. Perhaps a reasonable response might be, “Have you thought about how miserable you are when you train? Is this really consistent with other aspects of your current lifestyle that mean so much to you?” Where a simple denial is a usurpation of core autonomy, a probing and questioning response can—if delivered in the right way, by the right person, in the right context—invite the person to rethink whether it is a good idea for them to petition for that form of uptake. Similarly, in the case of gender, one might engage in a process of helping someone think through what they want, and what will grant them a flourishing existence, although such pushback against people’s expressed identification should be very rare, since generally speaking people ought to feel free to experiment with any gender position they want to try on. But in any case, mere denial is a kind of second-personal denunciation of their status that brings a dimension of discursive and social violence that is hard to justify.

The main point here is that shifting someone else’s status in social space generally requires special credentialling, formal or informal, that gives the person the authority to do this. But generally, people have not stopped to even wonder by what authority someone else could impose a gender status on me or take one away from me. If we interpret gender ascriptions as declaratives, then it can seem like such denials are just disagreements over facts, and anyone has the authority to dispute a purported fact. But we have been arguing against this interpretation, and if gender ascriptions are not declaratives then the question of their authority conditions becomes vivid. Moreover, the ethical effects of unentitled second-person gender ascriptions are pernicious. Denying someone’s right to self-position in social space has consequences for their self-determination and bodily agency, and is a display of basic disrespect for their first-person authority.

7. Third-Person Gender Ascriptions

If someone tells a third party, “Sam is a man,” the core pragmatic function of this speech act is not to give Sam uptake or to engage with Sam. But it also does not serve only or primarily to declare empirical facts about Sam. Rather, it calls upon others to give Sam uptake as a man. The ethical and pragmatic evaluation of such a speech act has to function against the background of Sam’s own first-personal gender ascriptions. In the rare cases where Sam has neither implicitly nor explicitly made such an ascription, we take such third-person ascriptions to be presumptuous in much the same way that second-person ascriptions are, if not as interpersonally aggressive.

But there are pragmatic dimensions to a third-person ascription that are different from the second-person case. A second-person agreement—“I am a woman!” “Yes, you are a woman”—enacts a kind of solidarity. It is a way of giving uptake to the initial call in the way aimed at by the initial speech act, and a commitment to engaging with the speaker in the ways implied by that. As such it ratifies a sort of social arrangement, much as “Of course I’m your friend,” “Yes, you are my teammate,” and “Welcome, comrade!” all do. A third-person gendered ascription, by contrast, engages first and foremost with someone else. Thus, if Sam sees himself as a man, and Joe says to Axel, “Sam is a man,” this is a call from Joe to Axel that petitions Axel to give a set of social uptakes to Sam, or take it as appropriate that others do. It is a kind of holding of Axel to the proper social response to Sam’s self-identification. Likewise, if Joe insists, “Sam is not a man,” he is not only withholding a set of social uptakes from Sam, but petitioning that Axel withhold them as well. These sets of petitions for recognition are also embedded into the use of gendered or gender-neutral pronouns. Moreover, audience, here again, matters. For instance, there are important pragmatic differences between using he/him pronouns for Sam or announcing his manhood in a public forum, which constitutes a generalized call to hold him in a particular socially position, versus doing so privately with one other person, which constitutes a specific second-person call to that person to recognize Sam as a man.

Notice that there is a pragmatic difference between saying “Sam is a man,” and, “Sam takes himself to be a man.” This is so whatever metaphysical story we end up telling about the conceptual or causal relationships between these two claims. The latter is typically a psychological description, and it also suggests or implicates that there is reasonable debate as to whether Sam is correct in taking himself to be a man. We have argued that in almost every case, it is reasonable to say, “Sam is a man”, when and only when Sam in fact makes clear that he wants to position himself as a man. But despite the coextension of these situations, the two speech acts have different force. One proposes giving uptake to Sam as embedded in norms in a way that positions him as a man, while the other states a psychological fact, and is properly assessed in terms of its truth value. The ethics of making such a declarative assertion are complex, because in practice it calls into question the correctness of Sam’s self-identification. It comes across as withholding judgment as to who we should take Sam to be: Sam takes himself to be a man, it tells us, but it does not (without more context) demand that anyone else take him to be one.

Second-person gender attributions are entitled only in the case of agreement with first-person attributions, we argued. But in the third-person case, we are often not in a position to ask someone directly, or to have access to their self-ascriptions. We routinely say, “There was a man waiting for the bus this morning” and the like, and it seems strong to claim that such speech acts are unentitled because we don’t have entitlement passed on directly from the person we are talking about. Generally speaking, even if someone has not directly self-ascribed a gender, we can make a pretty good guess about their gender identity based on their social and physical presentation, which are also ways of “telling” us their gender, typically (though not always). We think that such guesses can be relatively harmless if they are based on reasonable evidence, as they are typically just picking up on and reinforcing a position in social space that has already been claimed. But they need to be treated as provisional and highly open to revision, and should be avoided when there is any ambiguity at all in someone’s gender presentation. We also think they don’t need to be as common as they are; gender is not relevant to our third-person speech nearly as often as we include it.

One complicated case is babies, who are not yet in a position to first-personally claim a gender identity. It follows from our account that we should withhold gender ascriptions from babies until they are old enough to start claiming a place in gendered space for themselves. Our early gender ascriptions to babies and young children are far from innocent, after all. In ways that feminists have thoroughly documented from Simone de Beauvoir onwards, these early placings in gendered space radically constrain how a child grows up to move, look, act, experience, and have possibilities open and closed to them (De Beauvoir 2010; Young 1980; Bartky 2015). However, we suspect that this is too radical a switch in practice to get wide traction at this moment in history, so for now we would merely urge that third-person gender ascriptions of babies be as pragmatically gentle and provisional as possible. Once a child starts self-ascribing their gender (or refusing to do so), disagreement with or contradiction of their self-ascription is an unjustifiable and unentitled act of social violence.¹⁵

Elizabeth Barnes (in press) argues that having a gender cannot depend entirely on first-person claimings of gender identities, because this would have the consequence that cognitively disabled people who are not able to claim a gender identity do not have a gender. Our argument is not in any direct tension with hers, since we are not weighing in on what it takes to have a gender, which is a metaphysical question. It is perfectly consistent with our position that someone who does not and cannot claim any position in gendered social space has a gender. However, we are indeed claiming that someone who cannot claim a gender identity should not have a gendered social position thrust upon them. So we would be opposed to treating severely cognitively disabled people as gendered, which we suspect violates the spirit though not the letter of Barnes’s concern. But we do not see having a gendered social position as a success of any sort, or treating people as gendered as any sign of respect. Indeed, a big part of the impetus of this paper is to make room for people to refuse binary gender roles altogether, and one of our underlying premises is that, while it offers expressive and other possibilities to those who want them, gender roles are on balance used to socially restrict and oppress more than to liberate. We have argued for taking very seriously the idea that imposing a gender role is an assault on self-determination. We think that our account better respects the dignity and self-determination of cognitively disabled people in virtue of not imposing gendered expectations and restrictions on them that they did not ask for or choose. We thus see second-person and third-person gender ascriptions for people who have not made any first-person ascription as unentitled, including for cognitively disabled people.¹⁶

Our concern here is with which speech acts are entitled and what those speech acts do in social space. We are not talking about what beliefs about empirical facts people are right to hold. I may well believe that someone is a man, based on what I know about his genitals, his looks, his behavior, his DNA, or something else. Even if I don’t have a well-developed metaphysical theory of gender, my belief may be a sensible inference based on the fact that others with similar body parts or behaviors routinely turn out to be men. But this doesn’t mean that I am necessarily entitled to ascribe a gender to him in public speech, as doing so does more than convey facts; it also inserts him into and holds him in normative space. The bar for our entitlement to affect someone’s social status is higher than the bar for our entitlement to form a private belief about them. The criteria for when we can do so are ethical, not merely descriptive.

This is an important point about the nature of speech acts. Because speech acts are actions, they affect the social and material world. Hence there are always separate ethical and entitlement questions about whether any speech act should be performed, regardless of our private beliefs. Even in the case of declarative descriptions, speech acts do more than merely mirror private thoughts. But in the case of a socially potent performative like a gender ascription, there is a larger and more substantive pragmatic and ethical gap between the justification for a private belief, which is true or false, and the justification for a performative that shapes social reality. John Corvino similarly points out that there is an ethical gap between what gender someone has and how we should address and describe them (2000: 179). However, since he doesn’t distinguish between descriptive and performative functions of language, his point can make it sound as though it is fine to misrepresent reality in order to protect people’s feelings (which maybe sometimes it is, but that’s a separate issue). We, on the other hand, are not advocating making false claims about reality for strategic purposes. Instead, we are urging that since gender ascriptions have performative effects other than description, what entitles them is not our descriptive beliefs.

Most of the literature on gender ascriptions has focused on first-person ascriptions (e.g., Bettcher 2009) or on third-person ascriptions (e.g., Jenkins 2021). We hope to have shown along the way that the relationship between first-, second-, and third-person ascriptions is complex, and that understanding the pragmatics of gender ascriptions requires looking at all three types of speech acts, and the differences in their entitlement conditions, pragmatic form, and normative effects. We lose some of the complexity of each of these types of speech acts if we do not attend to how they work in relationship with one another, whether this relationship is mutually supportive or antagonistic.

8. Conclusion: Why Do Gender Ascriptions Work?

Our arguments imply that first-person gender ascriptions, and refusals of gender ascriptions (including refusals of all gender ascriptions) are almost always entitled, and second- and third-person ascriptions that contradict first-person ascriptions almost never are, regardless of the social, biological, or psychological facts about the person. They are also unethical impositions on someone’s basic self-determination, more akin to pronouncing someone married against their will than to describing them incorrectly.

This raises a puzzle: If second- and third-person gender attributions that do not inherit their authority from first-person attributions are not entitled, then why do they have a social impact at all? Why aren’t they just misfires? After all, if I order my student to clean their room, I don’t impart an obligation to them; I have just misused language. And again, if I shout names at random babies, I do not succeed in naming them. How do people succeed in imposing gender norms upon one another through language, without the entitlement to do so?

This trick is based on a kind of masquerade. Gender ascriptions pass themselves off as doing mere descriptive work, as just attributing descriptive properties to people. As such, anyone would be entitled to utter such speech acts, and then we could debate their truth or falsity. But in fact, the normative effects of the speech act are smuggled in and given uptake, and thereby solidified. Thus, gender ascriptions, even when they are not entitled, do their performative work by way of what we might call constitutive misrecognition.¹⁷ We act as though in attributing gender we are just reporting on pre-settled facts, while in doing so we actually constitute social reality. We are here agreeing with Judith Butler that gender ascriptions are citational as well as iterative—we constitute gender positions by recognizing them, and this recognition takes the form of recognizing something as a repetition of what was already there (Butler 1997). This is a kind of a pragmatic enactment of the naturalistic fallacy; we use speech to establish how things ought to be by acting like we are just reporting on how they already were. (Compare: Of course he is noble. Do you not know who his father is?)

The logic of constitutive misrecognition is built into the grammar and the ideology of gender. As we pointed out earlier, using the surface grammar of a declarative to perform an exercitive or a speech act with a different illocutionary force need not lead to ambiguity or confusion. When I say to my newborn child, “You are named Jeremy,” I am not confused over the fact that my speech is constituting rather than describing Jeremy’s name. But in the case of gender roles, these constitutive speech acts are situated within an ideology that pushes hard for essentialization and naturalization, casting our social position as an inevitable destiny. In this context, the grammar of gender ascriptions helps to hide the equivocation between the constitutive work of language and its putative descriptive work. In other words, part of the way that constitutive misrecognition gets hidden in our grammar is through our systematic ambiguity, in discourse around gender, between describing social reality, describing intrinsic empirical facts about individuals, and doing the constitutive work of placing people within normatively articulated social positions. We certainly need not be conscious of this strategy; indeed most people take themselves to be making merely descriptive claims when they attribute gender. This is what leads them to act as if their calling someone’s gender self-ascription into question is just a matter of having a healthy debate over facts. We hope to have shown why these apparent ‘healthy debates over the facts’ are in fact substantial and robust assaults on people’s self-determination, and so why the practices encoded in that grammar call for revision.