Exclusion and Erasure: Two Types of Ontological Oppression

4.1. Metaphysical Indeterminacy

Metaphysical indeterminacy is often regarded as mysterious, and this perception may be even more prevalent within social and feminist metaphysics, where there are few (if any) existing connections to the larger literature on metaphysical indeterminacy. For this reason, it is necessary to make a few brief remarks about the nature and logic of metaphysical indeterminacy.

There are various theories on the nature of metaphysical indeterminacy. I specifically conceive of indeterminacy as worldly unsettledness (à la Barnes & Williams 2011). For it to be indeterminate whether P is for the states of affairs represented by P to be unsettled. Metaphorically: we imagine indeterminacy as a kind of absence. It is not as if there is an indeterminate state of affairs (contra Wilson 2013); rather, a state of affairs has yet to have been settled, one way or the other.

I take vagueness to be one type of indeterminacy. Specifically, I reserve the term “vagueness” for cases in which the relevant indeterminacy arises because of considerations involving a quantitative scale of some kind.

Consider the traditional examples of vagueness. How many grains of sand constitute a heap? How blue does a dress have to be in order to be a blue dress? What age does a person have to be in order to be considered old? In cases of vagueness, we find indeterminacy in the borderline cases; we even find it in determining which cases are the borderline cases.

Not all types of metaphysical indeterminacy have this character. For example, the open future is not a kind of vagueness. If the future is open, there is no fact of the matter about whether the sun will rise tomorrow. This has nothing to do with vagueness about the sun or its rising; rather, the facts are currently unsettled with respect to the sun rising in the future.

There are other kinds of indeterminacy, but the most obvious kinds of indeterminacy in cases of erasure are vagueness and open texture.

Vagueness is certainly a factor in erasure. Just consider the fact that colleges have varying policies depending on whether the potential student plans to undergo a gender transition in the near future. But what counts as planning? Merely considering the possibility of a gender transition is not sufficient to count as planning. Of course, if you have already made the relevant social, legal, and medical arrangements associated with gender transition, and you are psychologically certain that you will transition, then this counts as planning. However, there is a range of attitudes and practices in between; vagueness emerges at the borderline cases.

Although vagueness is a factor, I have generally spotlighted the way in which erasure often stems from open texture. In cases of open texture, a concept fails to determinately apply because the concept was not designed for that scenario (Shapiro 2006; Waismann 1945).

Hart (1961) gives a classic example of open texture in law. Imagine a law that says: no vehicles in the park. Hart writes:

There will … be plain cases constantly recurring in similar contexts to which general expressions are clearly applicable (‘If anything is a vehicle a motor-car is one’) but there will also be cases where it is not clear whether they apply or not. (‘Does “vehicle” used here include bicycles, airplanes, roller skates?’) (1961: 126)

The indeterminacy of the statute does not reside in the degree to which something is a car; we do not remove atoms from the car until it is indeterminate whether it is a car. The statute simply leaves open whether or not some things count as vehicles (for the purposes of the statute).

I give open texture an additional metaphysical interpretation: it is metaphysically indeterminate whether the existence of a bicycle in the park is consistent with the law. It is simply unsettled. But it is unsettled because the statute was not designed with certain scenarios in mind.

Open texture is common in the social world, and it has disproportionate effects on people from certain marginalized identities. Some social identities are recognized only to be excluded; e.g., woman, Black, homosexual. Other social identities are often not even excluded; e.g., trans, non-binary, mixed race. In the latter case, social groups and institutions do not have determinate conclusions about individuals with these identities because policy makers often do not imagine the possibility of such people.

There is a growing literature that defends the intelligibility and existence of metaphysical indeterminacy (Akiba 2004; Barnes & Williams 2011; Hawley 2002; Morreau 2002; Parsons 2000; Rosen & Smith 2004; Williams 2008; Wilson 2013; 2016). In the past, it was a foregone conclusion that metaphysical indeterminacy is impossible, but those days are over. I will assume that metaphysical indeterminacy is possible.

What about the logic of indeterminacy? There are various ways to go, but I will be relatively neutral. In my framework, sentences can be determinately true, determinately false, or neither determinately true nor determinately false. I do not assume that truth/falsity just is determinate truth/falsity. Nor do I assume a particular semantics for the threefold distinction I am making.

My claim is: if metaphysical indeterminacy is possible, then it is present in certain cases in the social world. I will not defend the antecedent of this conditional because there is already a wealth of literature defending the possibility of metaphysical indeterminacy. My goal is to show that the putative cases of erasure do involve metaphysical indeterminacy.

4.2. The Case for Metaphysical Erasure

Many theorists of indeterminacy take indeterminacy to be either semantic or epistemic. So why think that indeterminacy is metaphysical? More importantly: does it really matter whether indeterminacy is metaphysical? One might think that the nature and significance of erasure is independent of metaphysical theses. Against this view, I argue that the metaphysical interpretation of indeterminacy best explains the worldly effects of ontological erasure.

Suppose erasure involves semantic indeterminacy. According to the semantic theory, the source of indeterminacy is our language (e.g., Keefe 2000). “Woman,” as used by Bryn Mawr to determine admissions, does not have a precise meaning. The meaning of “woman” gives determinate verdicts about cis women and cis men, but it does not give determinate verdicts about trans people. The meaning of “woman” is simply undefined, in such cases. (Typically, the semantic theory takes a supervaluationist form, but you do not have to be a supervaluationist to adopt the semantic theory.)

It is plausible that erasure involves semantic indeterminacy. The question is whether erasure is purely a matter of semantic indeterminacy. After all, the existence of semantic indeterminacy is compatible with the existence of metaphysical indeterminacy. If erasure is purely semantic, then determining the membership conditions of B (eligible to be a Bryn Mawr student) solely consists of stipulating that “B” means ϕ, where ϕ is some condition that leaves it indeterminate whether, say, l (Linda) is B. To determine the membership conditions of B, on this view, is to determine the semantic content of “B” in a certain context.

The problem with the purely semantic view is that it does not make sense of the resulting worldly structures that characterize ontological erasure. Stipulating the meaning of a word does not necessarily change the world. I can use the word “smable” to refer to small tables, but this does not change any facts about the world. Similarly, defining “woman” in a way that it is indeterminate with respect to Linda does not, by itself, entail that Linda will be treated in any particular way.

There is a gap, at least in principle, between (a) how we define words and (b) how we structure our institutions and social norms. To bridge this gap, we need to say that our judgments and actions make it the case that certain social structures are created. Trans-erasing gender definitions are not simply harmful because they misrepresent gender; rather, they are harmful because of the trans-erasing social structures that will be built around such definitions. By “social structures,” I am referring to the broad class of facts about agents, their interactions, and the environments they shape.

For example, trans erasure leads to the construction of gendered spaces—most obviously, restrooms—and a set of social norms governing gendered spaces. These gendered spaces are important for the well-being of gendered individuals but they are often designed in a way that trans people cannot safely or reliably access them. If college staff and faculty are instructed to treat students differentially on the basis of trans-erasing gender definitions, then they will be ill-equipped to meet the social, academic, and health needs of trans students. If trans students drop out of college because of trans-erasing gender definitions, then their life chances will be altered; in our society, college degrees partially determine lifetime earnings and social mobility.

The broad way in which legal gender classifications lead to oppressive social structures has been described in detail by Spade:

For trans people, administrative gender classification and the problems it creates for those who are difficult to classify or are misclassified is a major vector of violence and diminished life chances and life spans. Trans people’s gender classification problems are concentrated in three general realms: identity documentation, sex-segregated facilities, and access to health care. (2015: 77)

I emphasize that, strictly speaking, it is the worldly implementation of these classifications that make them especially oppressive. Trans-erasing definitions are oppressive because of the worldly structures and systems they engender, not just the linguistic meanings and mental states associated with them. The semantic view cannot fully capture erasure because it does not capture the worldly aspect of erasure.

For similar reasons, the epistemic view of indeterminacy also fails to capture the worldly nature of erasure. According to epistemicism, indeterminacy (or at least vagueness) stems from our ignorance of the world, rather than metaphysical or semantic indeterminacy (Caie 2012; Hawthorne 2006; Sennet 2012; Williamson 1994; Yli-Vakkuri 2016). The epistemicist is committed to the existence of sharp-cut off lines, but they claim that there are special obstacles to knowing the truth about where these lines lie. So it is epistemically indefinite whether, say, Linda is B. But this lack of knowledge does not, by itself, entail any specific social structures that are oppressive to Linda or other trans women. Ignorance can certainly lead to oppressive outcomes (Medina 2013; Sullivan & Tuana 2007). Nonetheless, it is important to recognize that the oppressive social structures—the non-linguistic, non-mental barriers to the flourishing of trans people—are the main problem.

My claim is not that there cannot be a semantic or epistemic interpretation of the relevant judgments. Rather, my claim is that these judgments eventually result in social kinds that have a tripartite structure, one we see in the case of partially indeterminate kinds. On the metaphysical interpretation, determining the membership conditions of B is not simply a matter of stipulating meanings or acting under ignorance. When Bryn Mawr makes the constitutive judgment that being B means being ϕ, they create a worldly social structure that entails differential treatment depending on whether one is determinately a B, determinately not a B, or is indeterminate whether one is a B. Regardless of whether the judgments are semantically indeterminate or epistemically indeterminate, they contribute to the formation of indeterminate social structures.

Of course, you may think the same work can be done by an indeterminacy-like structure as opposed to a strictly indeterminate one. Instead of there being three states of determinacy (determinately ϕ, determinately ¬ϕ, and indeterminate whether ϕ), there may be three properties that collectively resemble three states of determinacy. Imagine a grocery store that organized its business hours as follows: open (defined as open for the entire day), closed (defined as closed for the entire day), and neither open nor closed (defined as opened for part of the day but not all). Suppose the states of open, closed, and neither open nor closed are treated quite differently by the business, so there are robust worldly patterns surrounding each of the three states. The state neither open nor closed may resemble the property indeterminate whether open, but it does not follow that it is identical to the latter property; nor does it follow that the latter property exists. An opponent of metaphysical indeterminacy could argue that we structure the world in a way that resembles what the world would be like if it contained metaphysical indeterminacy. Maybe it is not genuine metaphysical indeterminacy but actually metaphysical schmindeterminacy.⁸

I am not confident that schmindeterminacy will be good enough to capture ontological erasure; a commitment to metaphysical indeterminacy strikes me as a simpler and better understood option. But I do believe it is more important to think that erasure is metaphysical than it is to think that erasure is metaphysical indeterminacy. Erasure is metaphysical in the sense that it is primarily about worldly social structures. We fail to capture the nature and significance of erasure if we ignore its worldly element.

5.1. Agential Identity, Exclusion, and Erasure

Dembroff and Saint-Croix describe a simple scenario that illustrates the notion of agential identity:

Consider, for example, a young man deciding whether to come out to his parents as gay. For all they know, he is straight. So, he currently occupies the social position of being straight, at least at home. But, having realized that he is gay, he comes to self-identify as gay. This internal identification is crucial to agential identity, but it is the choice to come out, the choice to make that private identity public and allow others to perceive him as gay, that would create an agential identity. (2019: 576)

Call this young man Sam. There are three important features of Sam’s predicament.

First: Sam is perceived as straight by his parents. For this reason, he will be treated as straight by his parents. One’s social position consists in how they are treated on the basis of being perceived to be a member of a certain social group. In his household, then, Sam has the social position of a straight person. This social position exists regardless of whether Sam accepts or identifies with that social position.

Second: Sam initially privately self-identities as gay. His private self-identification exists regardless of whether his parents perceive him as gay; it exists even if he were somehow incorrect about being gay. Sam’s self-identification does not need uptake or validation from others.

Third: Sam wants to externalize his self-identity. To successfully externalize one’s preferred identity is to get others, aside from oneself, to accept or recognize one’s social identity. Sam’s externalization attempt can be as simple as telling his parents that he is gay, or it can consist in engaging in behaviors that are associated with being a gay person. Because Sam intends his self-identity to be publicly accepted, Sam will accept being socially positioned as a gay person.

A person possesses an agential identity just in case they self-identify as a member of a certain group and intend to externalize that self-identification. It is important to note that agential identities do not require uptake by others in order to exist. Sam can have the agential identity of a gay person even if he occupies, and continues to occupy, the social position of a straight person. Sam’s parents, and larger society, can reject his attempt to externalize his self-identity. Sam’s agential identity of being gay could remain, despite this rejection.

The notion of agential identity is useful in making sense of certain kinds of discrimination. In some cases, one is not discriminated against based on self-identity or social position, but instead one is discriminated against on the basis of attempting to externalize one’s self-identity. Dembroff and Saint-Croix (2019) point to Spade (2015: 148)’s observation that “Medicaid provides all of the gender-confirming procedures and medications that trans people request to nontrans people and only denies them to those seeking them based on a transgender diagnostic profile.” So if you self-identify as a woman but occupy the social position of a man, you can get the gender-confirming procedures and medications. It is only when you attempt to externalize your self-identity that you will be denied the relevant healthcare. Call this agential identity discrimination.

In cases of agential identity discrimination, one is discriminated against on the basis of one’s agential identity—one’s preferred perceived group membership—and not necessarily on the basis of one’s actual group membership. Strictly speaking, it is people with trans agential identities that are discriminated against. It is a separate question whether those with trans agential identities are “really” men, women, non-binary, or genderqueer. Agential identity discrimination is triggered by the attempt to externalize one’s self-identity, not the self-identity or social position of an individual.

With the notion of agential identity discrimination in hand, I will now argue that exclusion and erasure are both forms of agential identity discrimination. I start with the exclusion case. Calliope Wong had the agential identity of a trans woman when she applied to Smith College. Wong was determinately excluded from being a student because she did not count as a woman, as far as Smith was concerned. Surely this is a form of gender discrimination. The question is what kind of discrimination is at issue.

On the simple view, Smith is discriminating against Wong by failing to classify her as a woman even though she belongs to the social group woman. More specifically, Smith discriminates against trans women, where we assume that, as a metaphysical fact, trans women are women. If we analyze exclusion in this way, the injustice is that Smith fails to treat Wong in the way that is reflective of her real gender. Dembroff (2018) calls this the real gender assumption: that our gender classification practices should mirror the metaphysics of gender. If the real gender assumption applies in this case, then it may appear that we need not appeal to agential identity to explain the relevant discrimination.

The real gender assumption may fail to apply, for two reasons.

The simplest is that the real gender assumption may be false (as Dembroff 2018 argues) Why? Because individuals may not always have a real or univocal gender identity. You may be a nominalist who believe that the predicate “woman”describes the world even though there is no kind woman. Or you may simply be a gender pluralist who thinks there are a variety of contextualized woman kinds like Black woman, trans woman, Southern woman, etc. In either case, you cannot pin down a single gender identity that serves as the basis for discrimination.

While I am sympathetic to the falsehood of the real gender assumption, I think there is a more mundane reason to think that the assumption, even if true, simply does not apply. The reason is this: even if there are real, univocal genders, it will not always be true that individuals are excluded from the membership of a social group because they fail to be a member of a given gender group. For example, Smith may not care about Wong’s true gender identity. Rather, Smith might simply insist that they are not a Smith woman, no matter what the metaphysics of woman tells us.

For example, we can imagine Smith agreeing that trans women are women but nonetheless insisting on their definition of Smith woman, for the purposes of the college. In such a case, Smith treats Wong in a way that is reflective of her real gender; Smith simply does not accept individuals on the basis of their real gender, but on the basis of Smith’s preferred gender categories. In such a case, Wong’s exclusion is an instance of agential identity discrimination. Wong seeks to externalize her self-identity and Smith rejects this externalization. Regardless of Wong’s actual gender identity, Smith prevents Wong’s self-identity from being recognized by the college.

I view these cases through the lens of agential identity discrimination because the central issue, in these cases, is the rejection of one’s attempt to externalize one’s identity. Social groups need not, and in many cases, do not, make grand judgments on the metaphysics of gender or other social kinds. Rather, they make contextual and value-laden judgments about the specific social statuses they prefer to recognize. To be clear: even if these judgments are independent of the metaphysics of gender, they can still qualify as transphobic and oppressive.

For similar reasons, erasure can also be understood as a form of agential identity discrimination. The difference is that, in the case of erasure, Bryn Mawr fails to determinately count Linda as a (Bryn Mawr) woman. Linda takes herself to determinately count as a woman, but Bryn Mawr prevents the externalization of her determinate gender self-identity. Failing to determinately include an individual can be just as harmful as determinately excluding an individual.

5.2. The Injustice of Erasure

Although exclusion and erasure are both instances of agential identity discrimination, the injustice of erasure is importantly different from the injustice of exclusion. Erasure involves a distinctive kind of externalization failure, one that removes erased identities from public view.

Judgment-dependent social groups are governed by publicly available norms and membership conditions. Of course, the degree of publicity will vary from case to case. Formal institutions like colleges distill their norms and membership conditions into publicly available handbooks or websites. In more informal settings, a social group’s norms and membership conditions may be communicated orally or conveyed through the relevant social structures that are built on the basis of the judgments. In any case, these norms and membership conditions must be communicated, either directly or indirectly, to those who are governed by them (or who desire to be governed by them).

These norms and membership conditions are established by the judgments of some individual or group. Smith makes a judgment to the effect that trans women are not women; this judgment becomes the basis for publicly available membership conditions governing Smith College. College faculty and staff make this judgment publicly available, directly or indirectly, because they have to enforce the membership condition that it establishes.

I should emphasize that “public” does not necessarily mean: easily accessible and transparent. Consider the law. The laws of a country may be publicly available, but in practice, we often learn about those laws after we face material consequences for breaking them. Police, prisons, stop signs, border walls, restroom gender icons—all of these are ways of making public our judgments about what is, and is not, socially permissible. Publicity can merely consist in the availability of worldly social structures that are built on the basis of our judgments.

In the case of Wong, Smith refuses to externalize Wong’s identity as a trans woman. Wong’s exclusion is a matter of college policy. Nonetheless, Smith does externalize the identity trans woman. By determinately excluding trans women from their college, Smith externalizes the identity trans women but fails to externalize its instances. It is as if Smith says, “There may be a concept or kind of trans women, but the individuals in that category are not women. Or at least, they will not be considered women at Smith College.” The identity trans woman is publicly available because Smith makes clear (albeit unsatisfactory) judgments about it.

In contrast, Bryn Mawr fails to externalize the category trans woman because Bryn Mawr fails to make judgments about the category. Linda is ontologically erased because it is indeterminate whether trans women can attend the college, and it is indeterminate whether trans women can attend the college because Bryn Mawr failed to make a judgment, one way or the other. The absence of judgment leads to an absence of publicity. Decisions will certainly be made about individual trans women who apply to the college. However, the identity of trans woman is obscured from public view because there is no judgment to be made about the social kind as a whole.

I claim that erasure results in an absence of publicity. However, this fact is not a metaphysical necessity. We can imagine a possible world in which the erasure of trans women is just as publicly visible as their exclusion. Absences can be conspicuous. So one might wonder why ontological erasure is bound up with social invisibility.

My hypothesis is that the act of making (or failing to make) a social distinction—a distinction in ways the social world might be—is an indication of whether or not a distinction is socially significant. In cases of erasure, then, the failure to make a social distinction is also a failure to recognize the significance of a social distinction.

Let me explain. Humans have many properties but only some are regarded as socially significant. Having exactly such-and-such many hair follicles is not a socially significant property, but having such-and-such sex characteristics is. Both properties may carve out distinctions in nature, but in the latter case, the distinction is regularly and publicly made in the course of regulating our social lives.

In cases of ontological erasure, then, there is a failure to import the significance of a social distinction. The kind Bryn Mawr woman is gappy with respect to the kind (or agential identity) trans woman in the sense that: if x is a trans woman, it is indeterminate whether x is a Bryn Mawr woman. (This is an idealization since the kind probably does not perfectly fit the gaps of Bryn Mawr woman.) Within the context of the college, Bryn Mawr does not make a social distinction concerning trans women. Because they fail to make the relevant distinction, they indicate that the distinction is insignificant (in that context).

Putting it another way: erased kinds fall through the cracks of our classification practices, and our classification practices dictate what kinds are socially significant. This leads me to conclude that the injustice of erasure differs from the injustice of exclusion because erasure negatively affects the publicity of oppressed social identities. This lack of publicity has harmful epistemic and political consequences.

Epistemically, erasure contributes to hermeneutic injustice, which is “the injustice of having some significant area of one’s social experience obscured from collective understanding owing to a structural identity prejudice in the collective hermeneutical resource” (Fricker 2007: 195). When individuals are erased, their social experiences are obscured from collective understanding, much moreso than in cases of exclusion. Furthermore, erasure is not accidental or unsystematic. Trans women were erased in the context of Bryn Mawr, but they were erased in the context of other women’s colleges and social institutions more broadly.

In the case of erasure, the lack of collective hermeneutical resources is partly due to how the social world is structured. There are worldly facts about the lack of determinate inclusion or exclusion of trans people with respect to gendered spaces. There is a gap in hermeneutic resources because there is a gap in what the social world is like.

Politically, the lack of publicity undermines the likelihood of political change. Wong’s rejection from Smith served as a catalyst for political change because Smith drew clear lines in the sand. The college decided, unequivocally, that trans women were not women, and this decision was the subject of controversy and a subsequent political battle. The Wong case led the public to pay greater attention to the policies of other women’s colleges.

In a way, erasure is more dangerous than exclusion. Oppression thrives best when it is invisible, and erasure, by its nature, renders oppressed identities invisible. We overlook the injustices that live in the metaphysical gaps of social groups and institutions. The current analysis spotlights these injustices so that we can (a) have a more complete typology of metaphysical injustices and (b) someday abolish them.