Dissolving the Gender Category Dilemma. A Deep Dive into Gender Diversity and Feminist Philosophy

Bachelorarbeit , 2021 , 64 Seiten , Note: 1,3

Geschlechterstudien / Gender Studies

What “is” a woman (a man; a non-binary person, etc.)? If one tries to answer this question, one quickly gets into a dilemma. On the one hand, feminist philosophers want to refrain from formulating gender categories at all in order not to fall into gender solipsism – the tendency to consider a small group of privileged people as representative of the entire category, excluding the less privileged people in the process. On the other hand, there is a need for gender categories within feminist philosophy in order to be able to address the structures of oppression they entail. In the course of my thesis, I will discuss this dilemma in more detail, examining its origins and some attempts by various philosophers (including among others Elisabeth Spelman, Catharine MacKinnon, Naomi Zack, Talia Mai Bettcher, Sally Haslanger and Katharine Jenkins) to find a way to dissolve it. My aim, however, is not to single out any particular account as the correct one, but rather to show which methodological approaches are generally advisable when one attempts to define gender categories without running into the aforementioned dilemma.

My aim in this thesis is to explore what an adequate gender theory might look like in general. Thus, I am interested in working out what methodological shifts are needed in feminist philosophy to dissolve the dilemma.
For this purpose, I will draw on various feminist categorization attempts and everyday examples, to illustrate which advantages and problems different approaches entail. Thereby, I will distinguish between i) essentialism, the view that any category is based on a common property, ii) universalism, the view that any category member corresponds to the exact same universal, and iii) realism, the view that the category tracks something in our (social) reality; because this distinction is particularly important to dissolve the dilemma.

Contents

1 Introduction

2 Torn between two target requirements — The gender category dilemma
2.1 The first target requirement: Non-exclusiveness
2.2 The second target requirement: Preservation of real gender categories
2.3 The resulting dilemma
2.4 The loophole of the dilemma

3 The cause of the dilemma — Gender universalism
3.1 The argument from empirical evidence
3.2 The argument from intersectionality
3.3 The argument from normativity

4 The first methodological shift - Gender relationalism
4.1 An introduction to gender relationalism
4.2 Acknowledging the intersectional insights
4.3 The remaining exclusion problem

5 The second methodological shift — Gender pluralism
5.1 The argument from irreducibility
5.2 An introduction to gender pluralism
5.3 Unproblematic normativity
5.4 Specific categories as real categories

6 Pluralistic Relationalism

7 Conclusion

Bibliography

1 Introduction

If you were to ask me what gender I am, I would tell you: “I am a woman”. But what exactly does that indicate? What information did I give you about myself with that statement? These challenging questions also concern feminist philosophy, since there is still no consistent definition as to what it means to have a certain gender. If we look deeper into the matter, we will stumble into an extensive discussion with countless suggestions on how to answer questions like the ones above. From the Beauvoirian position that gender is the social meaning of sex to the position that gender is not actu­ally anything at all, apart from a mere linguistic construct, there is a wide spectrum in which all kinds of different variations are discussed.

The recurring issue is thereby to tread the fine line between too little and too much diversity. If we define gender categories too broadly, we will risk slipping into insignificance, and if we define them too narrowly, we will wrongfully exclude some people that we should actually include. In other words, we need a certain amount of commonality to be able to speak of an actual category and not just of a gerrymandered group, but not too much either, because if we undervalue the diversity of women¹, we neglect the differences between them, which leads to the invisibility of some women*.

In this attempt to reconcile diversity and commonality, the question of an essential property emerges repeatedly. Such a property is characterized by the fact that it is possessed by all and only women*, and it would, therefore, be qualified to constitute the category of women*. The individual suggestions philosophers proceed to make for such an essential property are many and varied. How­ever, none of these approaches succeed in defining gender categories that include all and only the desired people,² as some group is always excluded and/or marginalized from the all-encompassing category. Accordingly, it seems, as if women* in general are too diverse to be captured by such a category. That is why some philosophers reach the conclusion that the whole discussion about es­sential properties is a stalemate that feminist philosophy should finally leave behind, by rejecting gender categories altogether.

Yet, on the other side, philosophers also present good arguments for why gender categories should not simply be abandoned. The most persuasive objection is that having a certain gender has a real impact on people's lives. Gender makes a significant difference in many situations, and this dif­ference should not be overlooked by feminist philosophy. Thus, the debate seems to be stuck in a dilemma where we want to define real gender categories, but cannot do so without wrongfully ex­cluding people. I call this the “gender category dilemma”.

Now, my aim in this thesis is to ascertain how the dilemma can be dissolved; albeit not by pre­senting a full-blown theory, but, more modestly, by exploring what an adequate gender theory might look like in general. Thus, I am interested in working out what methodological shifts are needed in feminist philosophy to dissolve the dilemma. For this purpose, I will draw on various feminist cate­gorization attempts and everyday examples, to illustrate which advantages and problems different approaches entail. Thereby, I will distinguish between i) essentialism, the view that any category is based on a common property, ii) universalism, the view that any category member corresponds to the exact same universal, and iii) realism, the view that the category tracks something in our (social) reality; because this distinction is particularly important to dissolve the dilemma.

During the inquiry the underlying question will always be if and how gender can be theorized in such a way that the theory satisfies both target requirements at once. Therefore, I will start by further examining the two requirements of non-exclusiveness and the preservation of real gender categories, to explain in more detail why neither of them should be neglected. Afterwards, I will proceed to investigate the cause of the dilemma, which lies in an equation of essentialism and uni­versalism, since this equation indicates that the problematic premises of universalism apply to the categorization process as such (section 2). After this realization, the next part of my thesis will deal with the exclusionary nature of universalism in order to find out how to overcome it without losing realistic categories. I will start with the universalistic premise of uniformity, which, as we will see, contradicts the intra-categorical diversity of gender and, hence, establishes problematic normative impactions (section 3). Once we know that, I will argue that such diversity can be better represented through a relational approach, especially if one chooses an intersectionality affirmative property as a basis for categorization. But at the same time, it will become evident to us that this first methodo­logical shift is not yet sufficient to overcome the dilemma, as there is a remaining problem of wrong­ful exclusion (section 4). Next, I will explain that this problem proceeds to exist, because the rela­tional approaches still adhere to the universalistic premise of gender monism. This realization will guide us to the second methodological shift, the adaptation of a pluralistic view, in order to fully overcome universalism and concomitantly the exclusion problem. I will additionally emphasize that both shifts can be carried out, without missing out on the second requirement of real gender catego­ries (section 5). Eventually, relationalism and pluralism will emerge as the two positions which, in combination, allow us to dissolve the dilemma, by providing us with a possible framework to devel­op theories that do not neglect either of the two target requirements (section 6).

2 Torn between two target requirements - The gender category dilemma

The extensive literature published on this topic over the last few decades shows that, up until today, there is still no consensus on which gender theory is the best to work with. But before deciding what form a gender theory should adopt, one must first investigate what requirements the theory should satisfy. Looking again at the literature in question, despite the many differences of opinion, there are yet two requirements most philosophers can agree upon. First, an adequate gender theory should not wrongfully exclude or marginalize any person from or within the theory (see e.g. Martin Alcoff 2006; Barnes 2020; Bettcher 2017a/2017b; Jenkins 2016; Kapusta 2016; Mikkola 2016; Spelman 1988; Stoljar 2011; Stone 2004 & Zack 2005). Secondly, an adequate gender theory should still allow us at least some gender categories³ to operate with (see e.g. Martin Alcoff 2006; Barnes 2020; Haslanger 2012a; Mikkola 2016; Spelman 1988; Stoljar 1995; Stone 2004; Young 1994 & Zack 2005). But why are these requirements so important in feminist philosophy? In order to answer this ques­tion in more detail, both target requirements will now be considered separately; first non­exclusiveness (section 2.1) then the preservation of gender categories (section 2.2). Afterwards, we will see how both requirements end up in the gender category dilemma (section 2.3) and begin our search for a loophole within the dilemma (section 2.4).

2.1 The first target requirement: Non-exclusiveness

The demand for non-exclusiveness originates from the observation that many previous gender theo­ries unjustifiably (and probably also unintentionally) exclude and/or marginalize certain (mostly al­ready marginalized) people. One of the most prominent feminist philosophers to address this prob­lem is Elisabeth Spelman. In her book “Inessential Woman: Problems of Exclusion in Feminist Thought” she attributed the exclusion problem of previous gender theories to white solipsism, the tendency of white feminists to expand their own (privileged) gender experiences as being applicable for everyone else too (see esp. Spelman 1988, 114-116).

This tendency is rightfully criticized by Spelman as an unjustified generalization. For an illus­trative example she quotes Barbara Omolade saying: “Black women are not white women with col­or” (Omolade 1985, 248; ptd in Spelman 1988, 13). Moreover, this line of argument can easily be expanded to being heterosexual, being cis-gendered, being able-bodied, being middle-classed, etc. because all of these are privileges, generating privileged gender experiences which are then “passed off as a metaphysical truth” (Spelman 1988, 186). Consequently, all gender theories that take privi­leged women* as the starting point of their inquiry inevitably pass over the experiences of women* from different social backgrounds, marginalizing them even further as their experience disappears behind the privileged experience (see ibid., 185-187). This is all the more problematic when one con­siders that in addition to their gender experiences, the interests and struggles of the less privileged women* associated with this experience also remain socially and politically unaddressed, because they are not even perceived as an issue.

At this point, the basic idea of intersectionality can already be deduced from Spelman's argu­ment. Intersectionality is a concept introduced by Kimberlé Crenshaw roughly at the same time, and it also directs the focus to the unique situation of multi-marginalized people, which can be identified as the kind of people Spelman has in mind when she talks of less privileged women. First Crenshaw impressively portrays how, as an example, black women are overlooked in the American legal system (1989, 141-152), before she then turns to feminist theorizing and, in agreement with Spelman, argues that the same thing is happening here with multi-marginalized, i.e. less privileged people (see ibid., 154-160).⁴

A gender theory that can be used to elucidate Spelman's and Crenshaw's concerns is, for ex­ample, Catharine MacKinnon's Marxist-inspired view that places sexuality at the centre of gender oppression (see Scott 1986, 1058). Accordingly, all women are sexually submissive, because all of them are victims of sexual objectification by men (see MacKinnon 1989, 113-114). This is, of course, a vastly simplified portrayal of MacKinnon's theory, which offers other possible interpreta­tions as well. However, I will not engage in a discussion about MacKinnon readings, instead, I will use this radical interpretation of her theory, as a striking example of a problematic way to categorize gender⁵. For, if we understand MacKinnon, to claim that all women are sexually submissive, i.e. that this property of sexual submissiveness is constitutive for being a woman, this conception of women is highly problematic, because it disregards the fact that even in our current society not all women are constituted that way (see Harris 1990, 590-601). Afro-American women, for example, are more likely to be targets of “victim blaming”, after becoming victims of sexual violence, precisely because they are not constituted as being sexually submissive (see William & Martinez 2002). Not to mention the variation that occurs when we look at women across different time periods and cultures, because here, too, numerous counterexamples could be provided. Accordingly, not all women are necessarily sexually submissive, and yet they are women; and women who suffer from forms of oppression that are specifically directed against them as women (see Harris 1990, 590-601). It follows that, if one were to adopt this radical interpretation of MacKinnon's gender theory (as e.g. presented by Harris 1990 or Mikkola 2019), the suffering and injustices experienced by a lot of women would go un­addressed or even worse, denied⁶.

What can be derived from this criticism is the resolution to avoid solipsistic tendencies, be­cause otherwise we easily exclude some people which we should (and initially wanted to) include. Moreover, as already mentioned above, the solipsistic line of argument marginalizes the less privi­leged even further by making them invisible. In some cases, exclusion from a gender category can do even more harm, namely, when the people concerned struggle very hard to belong to a particular category; I am talking about trans people⁷.

When trans people get misgendered, hence when they are either excluded from the gender category they identify with, or included but marginalized within that category by an internal hierar­chy, this can cause serious psychological, moral and political harm (see Kapusta 2016, 502). Stepha­nie Kapusta, who introduced the concept of misgendering, gives the fictional trans women Laura as an imposing example: Laura identifies, but does not pass as a woman and for this reason she is often regarded as less of a woman or not a woman at all (see ibid., 505). This circumstance most likely harms Laura in one or more of the three mentioned ways: Maybe she encounters several mi­croaggressions a day which in time cause her to have psychological health issues like anxiety, depres­sion, etc. (psychological harm); or she is restricted to explore her gender identity less freely than oth­er women (like cis women or trans women, passing as women) being also left out of the discourse over what counts as feminine in general (moral harm); or, if she does not legally pass as a woman, she will experience additional discrimination in her everyday life, as well as less access to goods and services designed for women only (political harm) (for a more detailed discussion of the possible harms see ibid., 504-505).

Apparently, a lot of current gender theories entail this sort of misgendering, and they are strongly criticized for doing so.⁸ Following Spelman's criticism of white solipsism one can attribute this problem (of misgendering trans people within feminist philosophy) to the tendency of cisgender philosophers to theorize gender from their own cisgender point of view; I will therefore call this problem “cis solipsism”.

Both problems (misgendering and “classic” exclusion) apply as well, of course, when we talk about men or any other gender, and they are to be avoided by all means in order to prevent the re­sulting harms. Nevertheless, white and cis solipsism do not preclude theorizing about gender in gen­eral, for that would itself be an unwarranted generalization; instead, they only tell us that previous gender theories, which have not taken these phenomena into account, are too short-sighted (see Mikkola 2006 & Zack 2007 for white solipsism and Kapusta 2016 for misgendering). But, if it is so difficult to make assumptions about women, men, genderqueer people, etc., why do we not just vol­untarily abandon these categories and try to get along without them? Or to put it more simply, why do we even need gender categories?

2.2 The second target requirement: Preservation of real gender categories

Some feminist philosophers claim indeed “that the category ‘woman[*]' is a fiction without objective basis and that feminist efforts must be directed toward dismantling this fiction” (Martin Alcoff 2006, 140-141). In the following, I will use “gender deflationalism” as an umbrella term for these kinds of positions, which a) claim that gender categories such as “women” are nothing but a mere fiction, which is kept alive by linguistic discourse; and b) reject these categories as unmaintainable, for ex­ample via elimativism or error theory.⁹

But, as Linda Martin Alcoff points out, this deflationalist understanding of gender is only de- constructive¹⁰ and hence it raises a lot of important questions: “What can we demand in the name of women if ‘women' do not exist and demands in their name simply reenforce the myth that they do? How can we speak out against sexism as detrimental to the interests of women if the category is a fiction?” (ibid., 143). The same questions are also raised by other theorists, for example, Natalie Stoljar asks: “If gender is illusory, how can there be political action on behalf of women, or a con­structive response to women's experience of subordination?” (2011, 44).

The problem that these questions indicate is that we need to have at least some gender catego­ries, in order to address certain forms of injustice that are associated with particular genders — this problem is articulated most prominently by political philosopher Iris Young (1994, 718).¹¹ She em­phasized that the rejection of gender categories is not a good idea, because it leaves us unable to name, understand and therefore work against these forms of gender-related oppression, because without gender categories, all we are left with are separate individuals struggling with seemingly unique problems (see Young 1994, 718). I agree with Young that the rejection of gender categories as such is not a good starting point for feminist philosophy. As Young noted, it is undesirable, be­cause it complicates the path towards gender justice, i.e. a society without any form of gender-related oppression, i.e. the self-declared goal of feminism (undesirability). Besides, it is also not in accord­ance with what empirical evidence tells us; because statistical data also show the impact that gender related oppression has in our society (discordancy).

Let me illustrate both points (undesirability and discordancy) by using the example of sexual violence against women:¹² There is statistically more sexual violence against women than men; that is a fact,¹³ but, if we reject the categories of women at all, we are unable to see this fact any more. Each case of sexual violence would only be considered as one individual case, ignoring the fact that there are underlying structural problems that increase the risk of experiencing sexual violence for some people. So, we need gender categories as our tools to call attention to such cases of gender-related oppression. Furthermore, we need them to take political action against gender-related oppression, because otherwise each affected individual would have to fight for itself, generating much less im­pact.

Adding to the problem is the fact that in some cases, part of the affected individuals would not even know that they are affected in the first place. Take sexual violence again: In a scenario where we are not able to consider sexual violence against women as a structural problem, a lot of women would not even know about their increased risk and would therefore be unable to do any­thing about it. Moreover, even the women who had unfortunately become victims of sexual violence would have limited options to do anything about it. In a worse case they would blame themselves for the incident, perhaps trying to change their external effect on people; and in a better case they would blame the offender, perhaps trying to prevent him from attacking again. However, neither scenario would change the overall situation, as the underlying factors responsible for the statistical effects would still be in place. What we need instead is the category of women as a basis to analyse the underlying factors and further to unite (potential) victims under a common goal. Together, those affected can raise awareness of the form of gender-related oppression they are experiencing and demand relevant changes. In case of sexual violence, women could for example organize a “Reclaim the Night march” as a form of protest, to borrow this real-world example from Katharine Jenkins (2016, 419). I do not see how this kind of activism is possible, if we were to deflate all sorts of gen­der categories.

Especially, since such a deflationalism is additionally linked to radical linguistic constructivism, the view that “everything is constructed, everything is language, or everything is text” (Asta 2011, 51). Thus, according to this definition of the position, gender categories lack of references in reality, they are again regarded as a myth or a fiction that is made up by our discourse. However, as Asta demonstrates, this position of radical linguistic constructivism collapses into “linguistic monism [.] the unhappy view that everything there is [is] a linguistic construction” (ibid.). This view is unhappy, because it is untenable. It suggests that we ourselves are creating gender-related oppression by keep­ing the discourse over these topics alive. Conversely, gender-related oppression should disappear once we stop discussing it, but this obviously is not the case. Sexual violence against women does not disappear when we stop talking about it, it only becomes invisible. So, the conclusion is again, that the degradation of gender categories as mere fiction would not necessarily contribute to the reduction of gender-related oppression; on the contrary, it could make fighting against it a lot harder (ibid.).

For these reasons, I will not pursue the deflationalist approach “that the category ‘woman[*]' is a fiction without objective basis and that feminist efforts must be directed toward dismantling this fiction” (Martin Alcoff 2006, 140-141) any further and instead turn to the positions on the other side of the spectrum. These positions assume that the gender really exists, i.e. that it is not a mere fiction, and that it has an influence on our reality, one that does not just vanish once we start con­sidering gender to be fictional. I will classify these types of positions as “gender realism”. Note, however, that this does not imply that gender must have some biological foundation or that it has to be a rigid concept (one that is established once and remains unchanged thereafter), for as I under­stand realism, realist positions can also consider gender to be socially constructed and capable of changing over time. After all, “collective human activity can create joints in reality, even when that collective human activity isn't tracking anything that's naturally or objectively privileged” (Barnes 2017, 2425).¹⁴ Thus, in this broad understanding of the term “realism”, it is sufficient for a gender theory to imply that gender can and should be categorized in reference to something in our reality, to be considered a gender realistic theory.

Due to these unsolved problems of gender deflationalism, we can conclude that the preserva­tion of real gender categories is as important as avoiding any kind of wrongful exclusion or margin­alization in order to reveal and address gender-related oppression. Whereas “real” means that these categories “explain things, they cause things, we need to understand them to understand the struc­ture of the world” (Barnes 2017, 2425), i.e. they are not mere linguistic.

2.3 The resulting dilemma

Now that we know what requirements an adequate gender theory should satisfy; the next question is how to meet both requirements within one theory. However, this is one of the most challenging question discussed in the current gender debate; for it appears as if a theory can only satisfy one target requirement at the expense of the other (see e.g. Stone 2004; Young 1994 & Zack 2005). The dilemma manifests itself when we try to define gender categories:

The gender category dilemma:

Horn 1: If we do define gender categories (the realist approach), we cannot avoid wrongfully excluding people; thus we fail to meet the first target requirement.

Horn 2: If we do not define gender categories (the deflationist approach), we cannot talk about gender related oppression; thus we fail to meet the second target requirement.

Problem: We either do or do not define gender categories; in both cases we are failing to meet one of the target requirements.

I consider the second horn to be solid, because, as we saw in section 2.2, the rejection of real gender categories is an undesirable and discordant starting point to further theorize gender.

But, once we take a closer look at the first horn, we will see that there is a loophole depending on how we categorize gender. For categorization itself, as the subsuming of individuals based on a common property (or common properties), is not yet problematic. In fact, not even the exclusive­ness is the problem, because we need to exclude something or someone in order to ensure that the category does not dissolve into nothingness. Instead, the problem lies in the wrongful exclusion of people who should be included. Remember for example Laura and the various harms that are in­flicted, if we categorize her as “less of a woman” or even as “not a woman at all” (section 2.1). Or consider a young and single black woman, who is not only at higher risk of sexual violence but also of victim blaming. Let us also give her a character and call her Julia from now on, to make her more tangible. Julia's additional struggle becomes invisible once we do acknowledge her as a valuable member of the category of women. Thus, if we lose focus on her (Julia), she will be seriously harmed as well, because her struggles will become invisible again (section 2.1).

Accordingly, the crucial question is: How can we construct a new realist approach regarding gender, which does not entail wrongful exclusion? Be it misgendering (Laura's case) or “classic” exclusion (Julia's case). But, in order to answer this question, we need to know in advance what caused feminist philosophers to believe in the first horn, i.e. that any kind of categorization would be wrongfully exclusive, in the first place. For, only when we know what caused the exclusiveness of past realistic theories, can we deduce what prevents us from replicating that mistake.

2.4 The loophole of the dilemma

Philosophers who argue against the realist approach often point towards previous realistic gender theories, revealing their exclusionary nature. Thereby the argumentation always proceeds in the same way: First, the argument states that the theory in question postulates a common property which is constitutive for the category of women*. Then, the argument shows that the assumed common property is actually not common to all women; concluding that the theory is therefore wrongfully exclusive¹⁵. This way of arguing is already familiar to us from section 2.1, for it is nothing but a for­mulation of the solipsistic criticism, according to which it is unjustified to consider a single experi­ence (situation/viewpoint/etc.) as universally applicable. But, as we already know, this line of argu­ment can only be used to criticize a specific realistic theory for containing solipsistic tendencies, yet it cannot be used for a broader argumentation against realism itself, because that would be an unjust generalization. Nevertheless, the argument keeps one wondering why so many (if not all) realistic gender theories entail solipsistic tendencies.

That is why, Spelman claims, the problem behind solipsism “is not simply an empirical one that can be easily corrected” (Witt 1995, 328). According to her, there is an underlying systematic error, which caused previous realistic theories to end in wrongful exclusion, and this error can be identified as essentialism. For, women* are highly diverse (see esp. Spelman 1988, 157-158) and therefore essentialism is simply the wrong approach when it comes to theorizing gender.

For essentialism invites me to take what I understand to be true of me ‘as a woman[*]' for some golden nugget of womaness[*] all women[*] have as women[*]; and it makes the participation of other women[*] inessential to the production of the story. How lovely: the many turn out to be one, and the one that they are is me. (Ibid., 159)

Here Spelman takes the idea of essentialism ad absurdum, because (to her) the existence of a common womaness* appears untenable in view of the diversity within the category of women*. But her accusation goes even beyond this argument of absurdity, for she claims that the false promise of essentialism leads one to assume that everyone else is just like oneself. This is an important addition, because it explains why all previous definition attempts had to end in wrongful exclusion: All of them entailed essentialism, as all of them presumed the existence of some property F defining each gender category. This gives us a new, improved argument, based on the diversity of gender:

Argument from diversity:

P1D: Essentialism assumes that the category of women* is constituted by a “golden nug­get of womaness*”.

P2D: Women* are too diverse to share such a “golden nugget of womaness*”.

P3D: If one ignores P2D and nevertheless defines the category of women* based on such a “golden nugget of womaness*”, this approach entails solipsistic tendencies, which inevitably end in wrongful exclusion.

essentialism any further.¹⁶ Within essentialism, the property that is necessary for category­membership is usually called the essential property; but I will prefer to call it the constitutive property, because this term is less theoretically charged than the first one. Only when I want to emphasize the connection to essentialism I will call it the essential property. So, note that all of these terms techni­cally refer to the same thing. For example, if I talk about the category of famous people, I would refer to “being widely known” as the constitutive property, thereby implying that the possession of this property is necessary, i.e. essential to being a member of the category of famous people. Thus, essentialism only postulates that categories are based on some kind of commonality.

Universalism, on the other hand goes back in its origins to Plato's theory of forms, assuming that all individuals instantiate the same form (see Stoljar 2011, 27), hence it postulates not just com­monality but rather uniformity. From here, the position evolved further into a metaphysical debate about types (or universals) and tokens (or particulars) circling around the question, how those two relate (see Wetzel 2018). There are many different views on what a universal is, however, here we will follow Mikkola's interpretation of Bertrand Russell, according to which “a universal will be any­thing which may be shared by many particulars” (Russell 1967, 53; ptd Mikkola 2016, 31). The para­digm example also quoted by Mikkola is the universal “justice”.

Bertrand Russell writes that all just acts must ‘partake of a common nature, which will be found in whatever is just and in nothing else. This common nature, in virtue of which they are all just, will be justice itself, the pure essence the admixture of which with facts of ordinary life produces the multiplicity of just acts'” (Mikkola 2016, 30 referring to Russell 1967, 52).

Accordingly, universalism implies the existence of a universal (e.g. justice), which all and only category members correspond to; bringing them into possession of a particular property F (e.g. the thing by virtue of which all and only just acts are just). Thus, the same F must be instantiated in all category members, to put it in Platonic terms.

A brief clarification at this point: This universalistic position is also referred to as “realism”, both in the feminist philosophy and additionally in the larger philosophical discourse. However, I will not adopt the term “realism” for two reasons. First of all, I have already established that in this thesis the term gender “realism” will be used in a broader sense, to describe the notion that gender categories correspond to something in reality, i.e. that gender is not a mere fiction. Even if this posi­tion is often included in the universalistic reading of realism, the two positions are not the same, and thus it is useful to consider both separately. Second of all, the universalistic reading of “realism” often considers the universal to be something natural, and it is used in contrast to a position called “nominalism”. The term “nominalism” originally goes back to John Locke who held the view that the corresponding universals are “made up by us” (Stoljar 2011, 31), i.e. that they are socially con­structed. However, the justice example already establishes that, in my interpretation of universalism, the relevant universals can also be socially constructed and do not necessarily have to be natural, whatever that means.¹⁷ For these two reasons, I will continue to use the term “universalism” rather than “realism”, to describe the position that each particular corresponds to the same (natural or so­cial) universal.

If we now transfer this universalistic approach to gender, this would mean that there is a uni­form property F that all and only women share, which turns them into women, so to speak. Thus, gender universalism is based on the premise of gender uniformity, as it assumes that all and only the desired people must possess the same monadic property, like being long-haired, being nurturing, being oppressed, etc.¹⁸ Thereby, it is not yet specified whether it must be a biological or a socially constructed property; the decisive factor is that the relevant property is a monadic property, i.e. it is “captured by one- place predicates” (Mikkola 2016, 71), thus, predicates that do not need a second reference.¹⁹ Logically, this universalistic approach can be expressed in a VxFx scheme, where the whole scheme stands for a gender category (VxFx), within which all members x share the same mo­nadic property F.

Now, Spelman appears to equate the two positions, as her essentialism bears the characteris­tics of, what I call universalism. First, this “golden nugget of womaness[*]”, whose existence is chal­lenged by Spelman, strongly resembles a uniform, monadic property; because it sounds like every woman is instantiating the exact same womaness*. Accordingly, Spelman actually opposes against categorization attempts that are inspired by Plato's theory of forms. An observation that is further supported by the fact that at one point she even says: “We are not who or what we are by virtue of being particular instances of some transcendental entity” (Spelman 1988, 70), which is a definite rejection of universalism, but not of categorization as such. Spelman's original intention is rather, to shift our focus towards the diversity of women (see ibid., 171-178); further criticizing that this diver­sity cannot be captured as long as we hold on to the idea that all women (w) are in possession of the “golden nugget of womaness[*]” (F), which is constitutive for the category of women (VxFx). How­ever, this VxFx scheme belongs to universalism, therefore, we can conclude that Spelman is in fact equating essentialism with universalism when she says that “essentialism invites me to take what I understand to be true of me ‘as a woman[*]' for some golden nugget of womaness[*]” (Spelman 1988, 159).

However, Spelman is not the only feminist philosopher who equates the two positions. It is evident, for example, that Alison Stone relies on this equation as well, when she considers anti- essentialism to be the only option when it comes to theorizing gender. She states:

The (false) universalization of claims about women[*] in effect casts particular forms of femi­ nine experience as the norm [...]. Essentialist theoretical moves thereby end up replicating be­tween women[*] the very patterns of oppression and exclusion that feminism should contest.

This point has been pressed particularly forcefully by Elizabeth Spelman. (Stone 2004, 140)

Here, within her premises Stone speaks of the “universalization of claims” and in her conclu­sion she speaks of “Essentialist theoretical moves”. Hence, she also equates essentialism with uni­versalism, assuming that every essentialistic categorization attempts has to carry the problems of universalism as well. Thereby, she explicitly refers to Spelman, saying that she relies on her for this line of argument.

Another, more recent example, in which the equation of both positions is not so obvious, but still evident, is Elisabeth Barnes. She sates: “Identity-based accounts can often sound, prima facie, like a version of essentialism. The claim [.] is easily interpreted as the claim that everyone has an innate sense of gender which is independent of how they behave, how they are socialized, how they are treated, etc.” (Barnes 2020, 6). Here, too, essentialism is portrayed as a position postulating that gender categories have to be uniform, in a sense that every category member (of VxGx) instantiates the same innate sense of gender (G). But yet again, this claim of gender uniformity is made by uni­versalism and not essentialism.²⁰

The above examples are intended to show how the equation of essentialism and universalism is present in several contributions to feminist philosophy, from the 1980s until today. This fact is so striking, because it is exactly this equation that leads us into the gender category dilemma.

On the one hand, we cannot avoid essentialism without losing real gender terms; because we need at least some commonality, to base our category on; for otherwise it is just a gerrymandered group. This would leave us in a place where we are unable to justify why certain people belong to a certain gender and others do not. Accordingly, everyone's gender would be deflated, and we would not have the real categories that we set as our second requirement in section 2.2. On the other hand, we cannot maintain universalism, because universalistic approaches tend to ignore the diversity of gender by presuming gender uniformity and thereby cause wrongful exclusion; so, universalism de­prives us of meeting the first requirement from section 2.1. Thus, if we really equate essentialism with universalism, we merge commonality and uniformity, and then, it appears as if essentialism is unable to represent the diversity of gender, while it is actually just universalism that entails the prob­lematic claim of gender uniformity. Consequently, we can identify the equation as the reason for the gender category dilemma, since it blends the unavoidability of essentialism together with the prob­lematic nature of universalism, implying that categorizing itself, since it is always essentialistic, would inherently be problematic (the first horn).

Fortunately, there is an alternative option for us to categorize gender, because, as I understand both terms, the dependency only exists in one direction: The monadic property F must always be an essential property, for otherwise F could not constitute the corresponding gender category VxFx. However, the essential property need not always be a monadic property, for it could also be a poly­adic property instead.²¹ Arguing differently, one can also say that uniformity is built on commonality, because a uniform property is a common property that everyone possesses in the same manner. This addition is not particularly necessary to constitute a category, since according to essentialism a non­uniform property is also sufficient to do so, as long as it is still possessed by all its members.

Therefore, instead of adopting the equation and getting stuck in the gender category dilemma, we can remove the equation and understand essentialism in a broader sense. This opens up new possibilities for developing theories that take place outside the dilemma, since it is not categorizing itself that is wrongfully exclusive, but it is universalistic categories in particular. It follows that, as long as the constitutive property is something apart from this “golden nugget of womaness*”, i.e. apart from this uniform F, essentialism can be unproblematic.

Consequently, the next step out of the dilemma is to ascertain why universalism causes catego­rization to be wrongfully exclusive. For only then can we ensure to avoid its mistakes in future cate­gorization attempts. At this point, Spelman's argument from diversity comes into focus again. For it does not follow from the dissolution of the equation of essentialism and universalism that Spelman's argument against essentialism is useless; it only means that now it can be understood to oppose uni­versalism instead of essentialism; as it is universalism that assumes a uniform “golden nugget of womaness[*]. Thus, the new, modified argument states:

Argument from diversity':

P1D': Universalism assumes that each gender can be captured by a uniform category.
P2D': Gender is too diverse to be captured by a uniform category.
P3D': If one ignores P2D' and nevertheless defines such a category, this approach entails solipsistic tendencies, which inevitably end in wrongful exclusion.
CD': Universalism is the cause of the wrongful exclusion.

But, for all we know, this argument just rests on Spelman's claim that the idea of a “golden nugget of womaness*”, i.e. a uniform property F is absurd considering the diversity of gender. How­ever, logically it is still possible to be uniform in one aspect and diverse in another aspect (Stoljar 2011, 32). So, Spelman owes us an explanation, first of why gender is too diverse to be reflected by a uniform category, and second of why any approach to define a uniform category must, nevertheless, lead to exclusion.

3 The cause of the dilemma - Gender universalism

As discussed in section 2.4, universalism can be founded in both natural and social universals, de­pending on whether the uniform property is either biological or socially constructed. Especially early approaches of universalism, try to define uniform gender categories using biological properties as the constitutive monadic properties F in the universalistic VxFx scheme. This claim, however, is not advocated but only criticized in feminist philosophy, since it is highly problematic for several rea­sons. First of all, it is a patriarchal conception, which has been used politically as a justification to deprive women of certain rights (Witt 1995, 204-205). For example, women have been denied the right to vote, because, it was argued, their brains were not designed for it (see e.g. Sedgwick 1914, 30). Second of all, the preferred candidate for the uniform F is usually “being female*”, so conse­quently these theories postulate that all women* are human females* (see Stoljar 1995, 268). But this is simply false, since not all women* are female*. “Some women[*] are born intersexed, and some women[*], [...] were first assigned male[*] sexual identity at birth” (Zack 2007, 203). Therefore, one cannot postulate natural (or biological) universals without ending up wrongfully excluding people; since women* are too diverse in their biology, to have a common biological property, constituting their womaness*. The argument from diversity conclusively precludes biological universalism, be­cause this perspective on gender is neither in accordance with reality, for there is no uniform, bio­logical property; nor is it in any way useful to the feminist goal of achieving gender justice. To the contrary, it even provides a foundation for gender-related oppression.

For these reasons, feminist philosophers do not try to construct a biological universalism, but rather try to find a socially constructed property that can serve as a basis for real gender categories. Quite in the tradition of Simone de Beauvoir's famous quote: “One is not born, but rather becomes, a woman” (1956, 273). Nevertheless, several categorization attempts, which are inspired by Beau­voir, still adhere to the universalistic VxFx scheme, by constructing a social property F. For this rea­son, I will continue to call them social universalist.²²

But is this shift from biological to social universalism sufficient to prevent wrongful exclusion; or do the same counterarguments of discordance and undesirability apply in this case? The answer to this question is far more complex, for social universals cannot be as easy refuted as biological universals, since they are less tangible.

For a brief example of a gender theory advocating such a social universalism, think again of MacKinnon's theory from section 2.1. According to this theory, womaness is constituted by the property of “being sexually submissive”, a property, women do not possess by nature, but have ac­quired through social conditioning (see MacKinnon 1989, 128). This definition can be captured in the universalistic scheme; with all women w, who belong to the corresponding category of women VwSw, necessarily possessing the socially constructed property of being sexually submissive S. Now, we already know that this property of “being sexually submissive” is not our “golden nugget” when it comes to categorizing women, because MacKinnon's theory contains solipsistic tendencies, which wrongfully exclude women like Julia (see section 2.3), who are not constituted as being sex­ually submissive. What we do not know so far is, whether these tendencies arise because MacKin­non just chose the wrong F, or whether this is indeed a systematical error due to the underling VxFx scheme, as Spelman suggested.

In order to answer this, we will first take a look at P2D', investigating if the social situation of women (like their biology) is generally too diverse to be captured by a monadic property; a claim that is supported by two arguments. The first argument is based on the empirical evidence that there are major differences between people of the same gender category, concluding that social universal­ism is “descriptively false” (Stone 2004, 140), i.e. in discordance with reality (section 3.1). The sec­ond argument states that social universalism is precluded by the intersectional analysis (Stoljar 2011, 38-40), concluding that social universalism is at least undesirable, because it limits our theoretical possibilities for understanding gender (section 3.2). Both objections interact with each other, since the intersectional analysis would explain the intra-categorical difference and the intra-categorical differences would already hint towards intersectionality. Additionally, we will discuss an argument supporting P3D', by drawing attention to the fact that the solipsistic tendencies of both biological and socially constructed universalism emerge from problematic normative implications (section 3.3).

3.1 The argument from empirical evidence

The first indications that social universalism cannot be in accordance with reality are provided by the differences one notices when looking at individuals who belong to the same gender category. After all, what monadic properties do Marie Curie, Frida Kahlo, Laura (from section 2.1) and Julia share? Not much, except that they are all mentioned in this thesis and this property cannot be extended to all women. So, the resulting argument claims:

Argument from empirical evidence:

P1ED: Social universalism assumes that every gender is constituted by a uniform social property F.
P2ED: Empirical evidence shows that people belonging to the same gender category dif­fer strongly, also within their social properties.
P3ED: If P2ED is true, gender cannot be constituted by a uniform social property.

that “every woman is willing to mutilate herself, to satisfy a culturally defined concept of beauty”. This stronger version is empirically indefensible, since not every woman is willing to mutilate herself, to satisfy a culturally defined concept of beauty. Therefore, “voluntary self-mutilation” is surly not the social property that unifies the category of women, i.e. it is not the wanted F, in the universalistic scheme VxFx. Additionally, there are dozens of counter examples, where the differences cannot be reduced to common grounds; see for example the different habits concerning menstruating, which can range from social ostracism to public celebrations to societal indifference (see Mikkola 2016, 55).²³ Therefore, the objection to P2ED is definitely not convincing, as people belonging to the same gender category do differ strongly in their social properties.

However, the argument from empirical difference is still contested, by some advocates of uni­versalism arguing that social universalism and intra-categorical differences do not necessarily contra­dict each other (P3ED). The decisive point in this objection is that the argumentation from empirical evidence only touches the impacts of gender but not gender itself. For illustration, think of the Bud­dhist parable “The blind men and the elephant”, where a group of blind men examines an elephant and, in the end, they all come to different conclusions about what they are examining (see Keown 1996, 1). For all we know, it is possible that people belonging to the same gender category likewise have different impressions of their gender, even though these different impressions all derive from the same source. Thereby, all women* could, correspond to the same universal womaness*, only in different ways, just like the blind men experience the same elephant differently. A similar example is also given by Marie Mikkola, who argues that a rectangular block can be viewed from three different angles and still be the same rectangular block (2006, 86). So, according to this objection, the argu­ment from empirical evidence is invalid insofar as the empirical evidence of intra-categorical differ­ences does not preclude social universalism as descriptively false.

Only by now, decades have passed during which many philosophers have searched for this universal womaness*; and when we return to the elephant parable, it would be expected that, if sev­eral experts exchanged ideas for a while, they would eventually (even blindly) be able to recognize the elephant for what it is. Therefore, even if it is not logically refuted, it is becoming increasingly unlikely that social universalism is really a good option when theorizing gender, considering the in- tra-categorical differences.

This point is additionally emphasized by Stoljar, who compares gender theory to metaethics, where “the observed variation is better explained by the hypothesis that there are many different moral codes, or ‘ways of life,' than by the proposal that our moral perception is systematically and persistently flawed” (Stoljar 2011, 32). In a transferable manner, according to Stoljar (and also to me), the empirical evidence of intra-categorical differences within gender, is also better explained by multiplicity rather than universalism (see Stoljar 2011, 43).²⁴ Thus, the burden of proof has shifted, and it is now up to the proponents of social universalism, to show that their theories are still a desir­able option when it comes to theorizing despite these intra-categorical differences. Additionally, there is a second, even more pressing argument against social universalism, which proponents would also have to refute that is derived from the intersectional analysis.

3.2 The argument from intersectionality

The argument from empirical difference has already shown us how unlikely the existence of a uni­form property F is, which can be unproblematically used as a basis for a corresponding gender cate­gory. However, some feminist philosophers such as Mikkola still argue that social universalism is, nevertheless, worthy of consideration “despite the [.] epistemic difficulties” (2006, 79). In order to refute this claim, we will now draw on the insides from the intersectional analysis, to show that the universalistic scheme is indeed insufficient to reflect the diversity of gender.

The term “intersectionality” was introduced in 1989 by Kimberlé Crenshaw, who is not only a philosopher, but also a lawyer and a civil rights advocate. She originally intended the intersectional analysis to draw attention to the problems of African-American women, who were generally over­looked in the American legal system and, who were therefore often without legal protection.

In a nutshell, the intersectional analysis shifts the focus of social analysis from individual cate­gories to the intersections of several categories, where the single categories²⁵ are in exchange with one another. This shift reveals that if we simply try to add the different categories on top of one another, significant information goes missing (see Crenshaw 1989, 149). Remember Omolades: “Black women are not white women with color” (Omolade 1985, 248; ptd in Spelman 1988, 13) from section 2.1. If one looks at black women like they are white women with color, one is unable to fully understand the sexism these women encounter, as black women are not only oppressed as blacks and as women but also specifically as black women, because sexism can act through racism and the other way around (see Crenshaw 1989, 158-159 & Garry 2008, 100). Crashaw prominently illustrated this issue with a crossroad metaphor:

Discrimination, like traffic through an intersection, may flow in one direction, and it may flow in another. If an accident happens in an intersection, it can be caused by cars traveling from any number of directions and, sometimes, from all of them. (Crenshaw 1989, 149)

Here, the roads coming from different directions are the social categories (like e.g. race, class and gender, to give the most prominent three) and the accident is a case of discrimination (like a sexist or a racist incidence²⁶ ). Consequently, if we only take into account a single road, e.g. the gen­der-road, we drive past the accidents, assuming that they do not concern us: For, what do we care, if there is an accident at the crossroad with race? This must be a problem concerning race. But, the people on the race road think just the same way about gender.

According to Crenshaw, this is precisely what happens in the American legal system, with the result that black women were not protected by antidiscrimination laws. For example employers, who specifically discriminated against black women, could not be sued for racial or sexist discrimination, because they did not discriminate against blacks or women in general (see ibid., 141-150). This ob­servation does not only apply to black women in the American legal system, but concerns multi­marginalized people in general. They are all at risk of finding their interests, problems etc. socially unaddressed, because no one is taking them into account.

But, the insights of intersectional analysis do not only concern these practical implications, they also operate on a deeper level, revealing that multi-marginalized people are usually ignored when it comes to academic investigations of our society. This issue was already addressed by Cren­shaw herself, who criticizes feminist authors for not paying attention to black women either (see ibid., 152-160). From there, intersectionality had a great influence on a variety of disciplines but es­pecially on feminist philosophy. To the extent where “[i]t has become commonplace within feminist theory to claim that women[*]'s lives are constructed by multiple, intersecting systems of oppres­sion” (Carastathis 2014, 304). So, what we should take from this development is that gender catego­ries are always influenced by other social categories as well (see Bailey, 2009 16-20).²⁷

If we use these intersectional insights for our own purposes, they explain, as already men­tioned, why intra-categorical differences can be found in each gender category. Intra-categorical dif­ferences occur, because people do not just belong to gender categories, but to other social categories as well. Hence, these other categories can influence a person's realization of her* gender, turning gender categories into manifold categories. For example, Marie Curie's realization of womaness is different from Frida Kahlo's, because both women belonged to completely different social back­grounds. Marie Curie was also Polish, a scientist, able-bodied, etc. (see McDonald 2014, 37) whereas Frida Kahlo was also Mexican, an artist, disable-bodied, etc. (see Millett-Gallant 2010, 1).²⁸ There­fore, it is only plausible that both women differ strongly in there realization of womaness* and with this in all their social properties, even if both can be categorized as women.

This insight also explains why social universalism is inappropriate when it comes to theorizing gender. For social universalism cannot provide us with the manifold gender categories demanded by the intersectional analysis; and vice versa, the intersectional analysis contradicts the unity, which is postulated by social universalism. Thus, both positions preclude each other:

Argument from intersectionality:

P1I: Intersectionality states that gender, as a social category, intersects with other social categories such as race, class, etc.
P2I: If P1I is true, gender is constantly influenced by these other social categories.
P3I: If P2I is true, there cannot be a universal womaness*.
P4I: Social universalism presumes a universal womaness*.
CI: Intersectionality and social universalism preclude each other.

Here, P3I is the decisive premise, for in contrast to the argument of empirical differences, P3I implies that gender itself is affected by intersecting categories and not just the impacts of gender: If gender itself is constantly influenced by these other social categories, gender itself cannot be constitut­ed by a uniform social property F. This expansion enables the intersectional analysis to go beyond the argument from empirical evidence and fully preclude social universalism. This point is also stressed by Stoljar, who, therefore, rejects social universalism for good (see 2011, 39). Let us consid­er the details:

According to social universalism, all and only women instantiate a universal womaness*, in that VxFx. If this universal womaness* is socially constructed, it follows that womaness* is consti­tuted by a uniform social property F. Consequently, one either possesses F or not — there is no space for other influences, no space for diversity. If one possesses F, one is included in the corresponding category; if one does not possess F, one is excluded from the corresponding category.

This conception is quite the opposite of the intersectional analysis, according to which, wom­en* who differ in their social backgrounds also realize their womaness* differently. That is why, Polish women (as Marie Curie) differ from Mexican women (as Frida Kahlo) differ from Afro- American women (as Julia), etc., and not just in their origin but also in their realization of womaness. The same argument can be made for sexual orientation, bodily condition and social class as well, to name but a few additional social categories. Thus, there cannot be a social universal womaness*, for all there is are different realizations of womaness*. Equally, there cannot be a uniform social proper­ty F either, because F can only be realized in one way, otherwise it would not be uniform.

All gender realist theories, constructed upon a universalistic foundation, have to fail, because they are postulating a non-existing uniformity. Or as Naomi Zack puts it: “[T]he factor that made them [former categorization attempts] so vulnerable to the facts of social differences among women was substantialism,²⁹ an idea there was a thing in women that could constitute them as women” (Zack 2007, 203). Just remember MacKinnon, who thought that “sexual submissiveness” was the uniform property F and thus took on the well-known problems of exclusion. Now we know for sure that not only her choice for F is problematic,³⁰ but her “apparent commitment to gender as a social universal” (Stoljar 2017, 221), and the included presumption of the existence of an F that can be identified in the first place. According to the intersectional analysis, there simply is no social univer­sal, no property F that Marie Curie, Frida Kahlo, Julia and Laura share, for their womaness is influ­enced by so many other social categories, they belong to in addition to the category of women.

This also resonates well with Spelman's initial criticism, for example, when she claims that it is impossible for a white woman to imagine what it is like to be a black woman (see Spelman 1988, 135); or when she talks about herself and black civil rights activist Angela Davis, denying the exist­ence of any “‘woman' substance that is the same in each of [them] and interchangeable between [them]” (ibid., 158). For, even though she does not yet use the corresponding vocabulary, it is now evident that the intersection of race and gender is the foundation of her considerations, because the intersectional analysis reveals that white women* and black women* do not realize their womaness* in the same way.

This realization is all the more important, because we can use it as a tool in the fight against gender-related oppression. Just think again of our example about sexual violence against women from section 2.2. In that particular case, along with many other cases,³¹ the intersectional analysis can provide us with valuable insights. Sexual violence against black women, for example, has a complete­ly different historical background than sexual violence against white women. While white women had to liberate from the notion that their chastity (and basically their sexuality) belonged not to themselves, but to their husbands (see MacKinnon 1989, 171-183); black women never were per­ceived as being chaste (see Crenshaw 1989, 157-160). On the contrary, a myth stated that the rape of a black woman would be impossible (see ibid., fn. 49). Accordingly, black women had to gain sexual autonomy in a completely different way than white women did.

Both backgrounds are, nevertheless, equally important to fully understand how gender-related oppression in the form of sexual violence still works within our current society. Just remember our discussion of white solipsism in section 2.1 and how black women, like Julia, are still more likely to be the targets of victim blaming, because of this historical “rape myth” (William & Martinez 2002). Nevertheless, both black and white women have a higher risk of experiencing sexual violence than men. Therefore, they can still unite against this form of gender-related oppression, for example, by marching side by side in a “Reclaim the Night march”, while still acknowledging their different backgrounds.³²

However, this inter-categorical diversity also raises some concern about the intersectional analysis, which is by no means uncontroversial. Here Mikkola comes into the picture again, because her advocacy of social universalism is based primarily on the concern that an intersectional perspec­tive could not only preclude social universalism, but could lead to the dissolution of gender catego­ries as such (see e.g. 2007/2017), which would result in the second target requirement being missed. Her argument states that, if we take the intersectional insights seriously, we need to define not one, but several categories of women*, like “working-class women[*], black middle-class women[*], poor Jewish women[*], wealthy aristocratic European women[*] and so on” (Mikkola 2017, 171). This condition might appear acceptable at first, but, yet again, it is Martin Alcoff who asks the decisive question: “[H]ow to address the significant differences among women[*] that are mediated by race, ethnicity, class, nationality, religion, sexuality, age, and able bodiedness without emptying out the category altogether[?]” (Martin Alcoff 2006, 153). The problem both Mikkola and Martin Alcoff are alluding to is, that we do not know when to stop differentiating women* into more and more sub­categories until we eventually end up theorizing each individual women*. Now this outcome is un­desirable, because, complementary to the dissolution due to maximum inclusion, the category also dissolves, when it is maximally individualized, for a category with only one member can hardly be called a category (see Stoljar 2011, 44).

That is why some philosophers reject the intersectional analyses, claiming that it would de­prive us of real gender categories and thus from achieving gender justice (see Zack 2005/2007).³³ Meanwhile, Mikkola come to a somewhat different conclusion, saying that the gender debate has manoeuvred itself into a “stalemate” (2017, 176). Of course, this “stalemate” is nothing but another version of the well-known gender category dilemma, where we can either reject the intersectional analysis (like Zack 2005/2007 does), thereby missing our first requirement of non-exclusiveness, or we could acknowledge the intersectional analysis and hence abandon our second requirement of real gender categories (as for example Mikkola 2017, 171-172 interprets Spelman 1989).

However, the first assertion of this objection is already flawed: We do not have to define sev­eral categories of women*, to take the intersectional insights seriously. What we need to define in­stead is a manifold category of women*. This way, “working-class women[*], black middle-class women[*], poor Jewish women[*], wealthy aristocratic European women[*] and so on” (Mikkola 2017, 171) can maintain their diversity while still belonging to the same category of women*. In fact, social categories “must have at least a minimal degree of stable meaning for the [intersectional] anal­ysis to work” (Garry 2008, 107), otherwise there would be no categories to intersect. Therefore, in- tersectionality cannot lead into fragmentation without depriving itself of its own foundation. Thus, the argument opposing intersectionality is built on a false premise; which is once again based on the equation of universalism and essentialism, as we will reconstruct now.

The objection above presumes that the intersectional analysis precludes any all-encompassing category of women. From this, it is concluded that we need to define several categories of women, if we are to take intersectionality seriously. But the intersectional analysis does not really preclude an all-encompassing category of women after all; it only precludes social universalism or more precisely a uniform category of women. There are, however ways to pursue an essentialist approach, i.e. an approach that categorizes gender based on a common property, without this premise of gender uni­formity. For, just because this property is common to all category-members it does not have to be uniform (section 2.4). Instead, it can very well also be a multi-realizable property, constituting a manifold category; and in that case, we can define an all-encompassing category of women*, as long as the multi-realizable properties are able to reflect the intersectional influences.³⁴ Hence, intersec- tionality does not preclude essentialism or categorizing in general, as long as we define non-uniform gender categories. So, we do not have to define several categories to acknowledge the intersectional analysis, thus the concerns are unfounded.

All this considered, we can therefore reach the conclusion that social universalism is not an approach worth pursuing, as it ignores the intersectionality insights and thereby leads to wrongful exclusion, because too much of the social interaction remains invisible. To use Crenshaw's crossroad metaphor: Universalism is precisely what is keeping us straight onto the gender-road, for it makes us believe that only accidents involving ourselves are due to the gender-road; because that is what it would be like, if gender were a uniform category. It is easy to see where this one-way street leads us, namely straight back into the very solipsism, we so desperately want to avoid. So, instead of tacking this known dead end, we should rather acknowledge intersectionality as our newfound tool that we can use to avoid further solipsistic tendencies within our gender theories (see Baileys 2009, 20).

In other words, once intersectionality showed us what we have been missing (the influence other social categories have upon gender) we cannot go back into ignorance; especially, because it was this ignorance, which gave rise to solipsism in the first place. In order to further elaborate this link between the ignorance of gender diversity and the generation of solipsism, we will now turn to the argument of normativity.

3.3 The argument from normativity

At the beginning of this section, we came to the conclusion that the category of women* cannot be based on uniform biological property, because no such property exists. Next, we saw that the empir­ical evidence of intra-categorical differences likewise speaks against a uniform socially constructed property, since it is unlikely that all these differences derive from the same source, the same social universal. Based on this, we then revisited the intersectional analysis, which was already lingering in the back of Spelman's argumentation, and we saw that intersectionality is able to preclude a socially constructed universalism, because the assumed gender uniformity contradicts with the various dif­ferent ways, in which women* realize their womaness*. But why does the ignorance of diversity (be it biological or social diversity) cause wrongful exclusion? To answer this question, we will now draw on Judith Butler's normativity argument.

Mari Mikkola emphasizes that Butler brings into focus that a shared “identity” can never hold as the solidifying ground for a feminist political movement; as “[i]dentity categories [like gender] are never merely descriptive, but always normative, and as such, exclusionary” (Butler 1991, 160; ptd in Mikkola 2017, 172). This statement is the key to our question about the cause of wrongful exclusion, but before we get into that, we will first take a closer look at Mikkola's Butler interpretation, which she presents for instance in the Stanford Encyclopaedia of Philosophy entry on Sex and Gender (2019) and in her contribution to the Routledge Companion to Feminist Philosophy (2017, 169­179).

As Mikkola reads Butler, the normativity argument is used to explain why feminism as a form of identity politics (relying on a common identity) (hereafter FIP) is sentenced to exclude part of the people it initially wanted to represent (see Mikkola 2017, 172-173 discussing esp. Butler 1999, 7-9).

As Mikkola notes, it follows from Butler's normativity claim, that “every definition of ‘woman' will be insidiously normative and thus politically problematic” (Mikkola 2017, 173). She also interprets this as Butler's motivation, to develop her own, well-known theory of gender performativity, which says at its core that rather than being a gender one is doing a gender (ibid.). However, elaborating Butler's own account would take us too far at this point, so for now we will just examine the argu­ment against FIP, extracted by Mikkola, because this is what we can use it for our own inquiry.³⁵

Argument against FIP:

P1FIP: While defining gender categories, one always establishes normative requirements each person has to meet to be included.
P2FIP: There is no basis for an identity category of woman.
P3FIP: FIP, nevertheless, constructs such a category.
CFIP: There always will be wrongful exclusion in FIP.

Within this argument, P1FIP describes the normative implications of categorizing, while P2FIP relies explicitly on the intersectional insights. For Butler herself is indeed an advocate of the intersec­tional analysis, she writes:

“If one ‘is' a woman, that is surely not all one is [.] because gender is not always constituted coherently or consistently in different historical contexts, and because gender intersects with ra­cial, class, ethnic, sexual, and regional modalities of discursively constituted identities” (Butler 1999, 6).

Here, P2FIP is additionally supported by her referencing the instability of the human discourse. For Butler assumes that words like “woman*” do not have a fixed meaning, throughout different discursive practices (see also Mikkola 2017, 172). A convincing addition, since it is quite plausible (at least to me) that for example Plato meant something different by “women” than MacKinnon did, or than Kapusta did, or than Butler herself is referring to.

Now, this instability of “woman” is not a good point to departure for FIP, which tries to speak on the behalf of these very women. This instability, thus, leads to P3FIP, i.e. FIP constructing its own “stable [.] seamless category of woman” (Butler 1999, 7), in order to keep up its political goal; and this is precisely where normative aspects of categorizing becomes problematic. For, when FIP defines the category of women, FIP gets to decide who counts as a woman and who does not. But these self-reliant definition attempts can only entail wrongful exclusion, because there is no uni­ty among women, FIP could rely on (Butler 1999, 7). Indeed, if there were some property F shared by all and only women, FIP only had to use this property F as a basis and the normative implications would not be problematic.³⁶ But there is no such property F, and therefore the normative compo­nent of FIP's categorizing attempts makes wrongful exclusion inevitable. By this point it is evident why the normativity premise can also be understood as targeting universalism. The reason is that both FIP and universalism try to define women along the VxFx scheme.³⁷

To see how the normative premise fits into our argumentation against universalism, we once again have to return to Spelman's initial criticism on solipsism. Remember, her claim that universal­ism is the systematic error behind the failure of previous theories, for it causes them to contain sol- ipsistic tendencies (section 2.4). Now, the normativity premise reveals the underlying power struc­ture and, in combination with intersectional analysis, provides us with the missing link, to explain why wrongful exclusion always has to be the case. For, in the absence of a universal womaness*, the normative claim becomes a solipsistic claim.

Argument from normativity:

P1N: While defining gender categories, one always establishes normative requirements
each person has to meet to be included.
P2N: The intersectional precludes a universal womaness*.
P3n: Universalism tries to define such universal womaness* anyway.
C1N: Thus, every universalistic definition attempt has to end in wrongful exclusion.

Here, the normative premise (P1N) shows us exactly why it is so problematic to ignore the in­tersectional insights (P2N) and yet try to construct a universal womaness* (P3N): Universalism allows us to begin theorizing gender starting by ourselves, because the presumed gender uniformity blinds us to other intersections. But, since the uniformity is only an illusion, those people, who do not share the (assumed) uniform property, are first of all invisible, and second of all wrongfully exclud­ed; because, the uniform property is also the constitutive property and as such it also entails norma­tive impactions. The central point added by the normative premise is, thus, that “any marking of sameness does not only describe something shared by women[*], but also dictates the thing that one must possess in order to be a woman[*]“ (Goldenberg 2007, 144).

This has the additional consequence that people, who already belong to the category, must not lose the constitutive property, because otherwise they also lose their category-membership. This makes universalistic approaches all the more problematic. Take again the MacKinnon example; here the postulation of the constitutive property S (being sexually submissive) is not only problematic because it excludes all the women who do not possess S; but it is additionally problematic, because it makes it impossible for any woman to cast off S, without losing her membership in the category of women.

So, borrowing Crenshaw's crossroad metaphor once again, we can think of the normative im­plications as the road's pavement, determining who has the right equipment to even drive properly on the gender-road. In the case of universalism then, only uniform cars, which possess the constitu­tive property F, get to smoothly drive on the gender road; while deviant cars have a bumpy ride or even break down. Furthermore, it is not surprising, that the people inside the uniform cars dismiss the accidents happening at the intersections as irrelevant to them. For, it seems as if those involved in the accidents simply have the wrong kind of car, i.e. the wrong equipment; when instead it was the road that was poorly constructed to begin with. In less metaphoric words, it is the universalistic framework that causes the wrongful exclusion, because it establishes a normative requirement of either biological or social uniformity that women* as a whole cannot fulfil, so cutbacks always have to be made somewhere.

This explains why we speak of white and cis solipsism rather than black or trans solipsism. Af­ter all, one first cuts back the edges, i.e. the marginal groups, if one wants something to fit inside a form. Moreover, categorizing is always “an expression of power” (Witt, 1995, 337); therefore, only privileged women* get to serve as a starting point for further construction, while less privileged women* get overlooked; for, people defining the category, first make sure that it serves their own benefit. In the words of the metaphor, privileged women* are in a position where they can pay at­tention to what kind of pavement the gender road has.

Remember Spelman's ironic claim “How lovely: the many [women*] turn out to be one, and the one that they are is me” (Spelman 1988, 159). This statement then gave rise to the argument from diversity and the question of how far both (Spelman's statement and the resulting argument) are justified. For, why can universalism not reflect gender diversity? And, why does the ignorance of gender diversity have to end in wrongful exclusion? After our detailed investigation we know now: First, a universalistic approach cannot reflect the diversity of gender, because gender is influenced by too many different variables to be constituted as a uniform category (intersectionality argument).

Second, universalism has to end in wrongful exclusion, because the VxFx scheme is not just an inad­equate form of representation, it also entails an unachievable normative demand for every woman to possess the same uni-realizable property F (normativity argument).

Therefore, we can now close this section by concluding that the argument from diversity is in­deed a well justified argument, which ascribes the cause of the gender category dilemma to universal­ism, or more precisely to the VxFx scheme used by universalism. Spelman is thus correct insofar that this scheme can be identified as the systematic error behind the failure of previous realistic defi­nition attempts. It is this universalistic VxFx scheme that is unsuited to capture the diversity of gen­der and, therefore, leads the universalistic categorization attempts into wrongful exclusion. Only the extended conclusion, drawn in part from Spelman's argument, that categorization itself is problem­atic is too strong, for this conclusion also relies on an equation of universalism and essentialism; an equation that does not necessarily have to be made, since both positions (universalism and essential- ism) are based on different assumptions (see section 2.4). Therefore, it is still possible to categorize gender, as long as we do it in a non-universalistic way that does not repeat any of the mistakes im­posed by the VxFx scheme.

Consequently, we must overcome the idea of gender uniformity, in order to avoid a relapse in­to the gender category dilemma. Instead, we should define manifold gender categories that allow their members to realize their gender in various ways, and, which actually embrace the intersectional insights. Hence, in the following section we will explore what kind of theoretical framework is need­ed to construct such manifold categories.

4 The first methodological shift - Gender relationalism

At the beginning of our enquiry, we noted that an adequate gender theory should define real gender categories, without containing solipsistic tendencies (section 2.1 & 2.2). At first, this seemed like an impossible goal, as both requirements appear to be mutually exclusive within the gender category dilemma (section 2.3). However, when we looked more closely at the cause of the first horn, we saw that it is based on an equation of essentialism and universalism (section 2.4). Moreover, we were able to show that universalism is responsible for the solipsistic tendencies, as it disregards the diversity of gender and thus contains a problematic normativity (section 3). Thus, our loophole out of the di­lemma is to separate essentialism from universalism, in order to develop gender categories that are essentialistic but not universalistic. But what kind of property can henceforth serve as the common property after we rejected universalism?

An answer to this question is provided by Zack, who proposes the idea: “But what if wom- en[*] shared a relation instead [...]?” (Zack 2007, 203). She goes on to elaborating this idea by de­fining gender categories in which the essential property is a relation (see ibid., 203-206). In doing so, she rejects the idea of gender uniformity, because a relation does not have to be realized in the same way in every case. However, Zack is not the only philosopher who tries to categorize gender this way; in fact, many recent categorizing attempts can be understood to categorize gender relationally, as long as the constitutive property is defined as a polyadic property rather than a monadic one.³⁸ In the following I will refer to this position, which abandons the thought of gender uniformity and tries to categorize gender relationally as “gender relationalism”.

To understand the extent of this methodological shift, we have to understand the differences between monadic and polyadic properties. So, the basic difference is that, while monadic properties are captured by a one-place predicate, “[p]olyadic properties are captured by n- place predicates that are expressive of relations like a, relation R, b” (Mikkola 2016, 71). Thus, for example, “being per­ceived as a potential partner for heterosexual men” is a polyadic property, since it consists of a woman w, who is in a specific relation P to being a potential partner for heterosexual men b, whereby the relation P is that w is perceived as b.

Moreover, polyadic properties cannot be reduced to monadic properties because they always need at least two predicates (see ibid.).³⁹ So, within our example, it makes no sense to eliminate the second predicate, just saying that “all women are perceived as” (VwPw), because then one inevitably asks “as what?”. Additionally, it is just as useless to eliminate the relation and proceed as if all wom­en just were potential partners for heterosexual men (VwBw), because the decisive point is that the women are perceived as b, an information that would go missing in the universalistic VxFx scheme. So, instead of trying to squeeze polyadic properties into the universalistic scheme, we should rather use a new relational scheme of the type VxRxy, with Rxy being the polyadic property constitutive for the corresponding category. In that case, all women w can form the category VwPwb only by being perceived P as a potential partner for heterosexual men b.⁴⁰ Consequently, the innovation of relationalism is that instead of possessing a uniform property F, now, one has to possess a polyadic property Rxy, i.e. one has to be in a relation R to another predicate/a second relatum y.

In the following, we will further explore Zack's own approach as an illustrative example of re- lationalism, while seeing that the advantages identified by Zack apply to relationalism in general (sec­tion 4.1). At the same time, however, the discussion of Zack's theory will show us that relationalism alone is not sufficient to overcome the problems of universalism, as Zack's relational approach still struggles with internal marginalization and wrongful exclusion. We will address both problems in more detail afterwards, starting with the problem of intra-categorical marginalization, which can be overcome by other relationalists, who consider the intersectional insights more carefully (section 4.2). After we have discussed this advantage in detail, we will then turn to the remaining exclusion problem and investigate why this problem cannot be overcome by relationalism alone (section 4.3), which will lead us to the second methodological shift.

4.1 An introduction to gender relationalism

Following the section above, the current gender debate can be partly regarded as a great discussion about which polyadic property is best to choose, if one wants to adequately reflect the diversity of gender. Broadly speaking, we can classify the individual attempts into first person relationalism and third person relationalism; whereas the first takes into account whether the person herself* identifies as a woman* or not, while the second depends on whether the person is identified as a woman* by her* social environment.⁴¹ This division is not that sharp, though, as some philosophers also try to construct a hybrid approach, taking both aspects into account (see Stoljar 2017, 216).

But no matter which strategy we consider, all these approaches can be regarded as relational;⁴² because all of them constitute their gender categories based on “identification”; and identification is evidently a relation, as it is part of an n-place predicate, forming the polyadic property Ixy (someone is identified as something; either by oneself or a third party or both). This is precisely what the term “relationalism” is meant to capture: A theory that can be reflected by the relational scheme, such as VwIwy (all women* possess the polyadic property of being identified as women*). Therefore, we can conclude that many recent categorizing attempts can be classified as relationalism.

For further illustration, we will now exemplarily discuss Zack's theory of “women's commu- nality” (2005), as she is the philosopher who most explicitly advocates the position of relationalism. Zack constitutes the category of women based on a polyadic property that “consists of self­identification with the FMP category and/or assignment to it by others” (Zack 2007, 203). Whereby, the FMP category stands for “designated female from birth, or biological mothers, or primary sexual choices of (heterosexual) men” (see ibid.).

According to the classification above, we can take her theory to be a form of hybrid relational- ism; because Zack's constitutive property can either be fulfilled via self-identification or third-party- identification with the FMP category⁴³. Therefore, her relational categorization attempt can be cap­tured in a VwTwm scheme, with all women w being identified T (either by themselves or by others) as members of the FMP category m. Here, it is explicitly emphasized by Zack, that it is the identifi­cation T to the FMP category m that is constitutive for the category-membership, regardless of whether the person actually possesses m or not (see ibid.); after all, this is where the great advantage of relational categorization lies, which will be now discussed in more detail.

The fact that it does not depend on the constitution of the woman* per se whether she is con­sidered a woman* or not indicates that the property Twm is an extrinsic property, as Zack herself declares as well (see ibid., 204). This observation can be expanded to other relational approaches, since polyadic properties always depend on a second reference and hence the constitutive property does not only concern the woman* herself.

But how does this methodological shift towards extrinsic properties as the basis for gender categories improve our theoretical situation? The answer lies in the general constitution of extrinsic properties, because ascriptions of extrinsic properties to something are by definition “not entirely about that thing, though it may well be about some larger whole which includes that thing as part” (Lewis 1983, 197). So, in terms of gender, this means that, if we define an extrinsic property as the constitutive property, it is not determined by the features of the person any longer whether she is a woman* or not. Instead, being a woman* is a complex decision made by the person herself or oth­ers. Even if this is not an advantage of Zack's theory in particular, her categorization attempt nicely illustrates the difference; because, here it is either the woman herself, who identifies with the FMP category, i.e. decides to possess the polyadic property Twm, or someone else takes this decision and thereby attributes the womaness to her (Zack 2007, 204).

It further follows from the definition of extrinsic properties that, if gender is an extrinsic property, a woman* does not acquire her womaness* only “in virtue of the way” she herself is, like the universalist thought, but “in virtue of the way some larger whole is” (Lewis 1983, 197). The im­portance of this difference also becomes evident, when we compare Zack's relational categorization attempt with MacKinnon's universalistic one. According to MacKinnon, a person had to actually be sexually submissive in order to count as a woman, entailing the problem that every woman, who is not sexually submissive, does not count as a woman (any more). Now, according to Zack, a person must only be identified as F (being designated female at birth) or M (being a biological mother) or P (being the primary sexual choices of (heterosexual) men), to count as a woman, regardless if they really are one of those things. Thus, people, who are neither F nor M nor P, but who are, for exam­ple, still identified by others as P “the primary sexual choice of (heterosexual) men”, still count as women. Hence, categorizing via polyadic properties can help to incorporate more women* than categorizing via monadic properties, because more people can provide a constitutive polyadic prop­erty and therefore a wider range of diversity can be included.

These advantages lead Zack to claim that “defining women[*] relationally [...] avoids all of the problems of substantialism” (2007, 204). But this conclusion is drawn a little too hastily, because Zack herself does not manage to fully overcome the problems of universalism, which she calls sub- stantialism.⁴⁴ Her theory reinforces intra-categorical marginalization and, additionally, it is also wrongfully exclusive; thus it misses the first target requirement for non-exclusiveness (section 2.1).

The first problem arises because Zack does not fully accomplish intersectional aims, which makes her theory blind to the diversity of women* (see Pedwell 2006, 367-369). For example, ac­cording to Zack's categorization, Marie Curie and Frida Kahlo are both identified (by themselves as well as by third parties) as all F, M and P. This suggests that both of them are realizing their wom- aness in the same way, when actually their womaness is subject to different social influences (as re­vealed by the intersectional analysis in section 3.2). As a result, the interests and struggles of less privileged women* disappear, yet again, behind those of the standardized heterosexual, able-bodied, white, middle class woman;⁴⁵ because, if we ignore the power structures between different women*, this is enough to contribute to the replication of those same power structures (see Bailey 2009, 14-36 & Pedwell 2006, 367-369). There are, thus, still solipsistic tendencies in place, which do not immedi­ately lead to exclusion, but to the neglect of less privileged women*.

Meanwhile, the second problem becomes visible once we turn to Laura, because in her case it is not even certain that she counts as a woman at all. We know from Kapusta, that Laura's outward appearance still resembles that of a man, in a way that she often “does not ‘pass' as a woman” (2016, 505). Thus, she does not get third-party-identified with either F, M or P. She also most certainly does not identify herself with F or M, so that the last possibility for her, to count as a woman in Zack's theory, would be self-identification with P, but this does not necessarily have to be the case either. So, presumably, Laura does not count as a woman in Zack's theory, and even if she does (in case she actually indeed identifies with P), it would be unjustified to generalize this decision to all trans women. Consequently, Zack's theory has the additional problem of being cis-solipsistic, since her dependence on the FMP category excludes certain trans women* from the category of women.

In sum, we can say that Zack's shift, to take relations as the basis for gendered categories, is a genuinely good move, as it allows for more flexibility by directing the focus from the state of the person to the impression she makes. For now, it is the impression I or others have of me that de­termines whether I am identified as a woman*, i.e. whether I am a woman*. This allows for more genuine diversity within the process of gender categorization, for the impressions can vary from case to case, and we will still reach the same conclusion that the person of concern is a woman*.

However, despite this enormous advantage, Zack does not succeed in portraying the intersec­tional influences that affect womaness*, nor does she overcome the exclusion problem, within her own relational approach. Accordingly, Zack's own theory disproves her statement that relationalism alone can solve all the problems universalism has. Thus, the shift to constitutive polyadic property is not in itself sufficient to solve the gender category dilemma. However, this is not to say that we should immediately give up relationalism; for, in contrast to universalistic categorization, it is not the underlying VxRxy scheme that is defective in the objections above. Instead, it is the choice of the polyadic property Rxy in particular that is causing the two problems. To put it simple, not all and only women have the relevant relation to the FMP category; but this does not mean that there is no relation at all that all and only women share.

Consequently, we have to further investigate whether both problems can be solved by other relational categorizing attempts, before we can say definitively whether or not relationalism offers a loophole to the dilemma. We will now begin with the first problem of intra-categorical marginaliza­tion, which arises because Zack does not fully achieve the second intersectional goal, of reflecting how the various other social categories affect gender.

4.2 Acknowledging the intersectional insights

As discussed above, Zack fails to meet the demand of the intersectional analysis within her categori­zation attempt. That is not a big surprise, considering that she explicitly rejects the intersectional idea (see Zack 2005, 1-23). But, as we have seen in section 3.2, it is important to consider intersec- tionality in order to prevent solipsistic tendencies. This revelation is further supported by the fact that Zack does not manage to fully overcome the solipsistic tendencies with her ignorance of the intersectional insights. But why does Zack reject the intersectional analysis anyhow? Well, like Mik- kkola, she is concerned that intersectional analysis leads to too much individualization of gender categories (see Zack 2005, 2), yet this is precisely what relationalism can prevent. We will discuss this claim later in this section, but first let us quickly recall the outcome of the argument from intersec- tionality, to address the problem of intra-categorical marginalization.

In section 3.2 we established that intersectionality precludes universalism, because the inter­sectional influences of other social categories, which each individual experiences, could not be re­flected by the universalistic VxFx scheme, because here one either possesses F or one does not. From this we concluded that gender categories must be constructed as manifold categories, i.e. cate­gories that are constituted by a multi-realizable property. Thereby, our aims were to allow different women* to differ in their womaness* and furthermore, to also reflect this diversity by revealing which influences different social backgrounds have upon women* as such.

After everything we discussed within this thesis, relationalism is still the best approach, to provide such a flexible property; because unlike monadic properties, polyadic properties can be real­ized in different ways. Thus, relationalism is generally able to allow different women* to differ in their womaness*. Even Zack's theory allows women* to differ in their womaness*, the problem is that she does not acknowledge this. Instead of embracing the manifoldness of her category, she still tries to sell it as uniform, while “attempting to construct ‘universal' foundations for feminism” (Pedwell 2006, 368) by implementing one FMP category, when according to the intersectional analy­sis women from different social backgrounds identify with different Fs, Ms and Ps. It is therefore Zack's choice of the polyadic property, and especially the second relatum y, which disregards the existing diversity of gender. For, even though Zack uses the relational VxRxy scheme, the alleged uniform FMP category as the second relatum y is just too limited to reflect the diversity of gender.

However, besides Zack, there are more intersectionality affirmative approaches, which have chosen a more variable second relatum y. For example, Talia Mai Bettcher provides a first person relationalism⁴⁶ that succeeds in acknowledging the intersectional insights. Bettcher herself only de­fines the constitutive property E as “sincere self-identification” (Bettcher 2017a, 396), but, according to Jenkins, this means:

“[C]laiming, or being disposed to claim in relevant circumstances, that one is a person of a certain gender. [So,] [o]n this account, to have a female gender identity is to be someone who is disposed to reply to the question, ‘What is your gender?', with the statement, ‘I am a woman[*]' (or words to that effect) and to be acting in good faith in doing so (Jenkins 2018, 727).”

So, in this case, a person has to meet two conditions to count as a woman*: One, the person must genuinely declare that she is a woman* (or at least she would do so); and two, the person must also act in accordance with this dictum, i.e. it is indeed not just a question of mere words. If both conditions are satisfied, then the person is a woman*; and she can furthermore declare that she has always been one (see Bettcher 2017a, 396).

Now, the important improvement of this definition attempt in comparison to Zack's is that Bettcher chooses one's individual conception of womaness* as the second relatum y. For I, as a person who identifies as a woman*, am identifying not with the same FMP category everyone else is also identifying with; instead I am identifying with whatever I think it means to be a woman*. Hence, VwEwc (all women* are sincerely self-identifying with whatever they think it is to be a woman*) is an intersectionality affirmative definition that not only allows for intra-categorical diver­sity, but that is, furthermore, well suited to reflect this range of diversity, for we can use it to demon­strate how once social background influences once gender:

Argument from variability:

P1V: Being a certain gender means, self-identification with whatever oneself understands by that gender.
P2V: Once individual conception of what it is to be of a certain gender is influenced by the other social categories one belongs to.

tion of “women”; and it also applies to Laura, who as a trans woman identifies with whatever being a woman means to her. Therefore, all four women possess a different womaness, because all of them have different conceptions of womaness in mind, with which they identify themselves.

Thus, categorization via sincere self-identification covers both aims of the intersectional analy­sis. First, it allows for intra-categorical diversity, because there is no longer any uniform property F any category member has to possess: for, in that place is now a flexible property Rxy that each cate­gory member can realize by taking into account her* own social background, which gives us precise­ly those manifold gender categories, demanded by the intersectional analysis. Second, and this is what sets Bettcher apart from Zack, these manifold gender categories also present themselves as such, since they explicitly indicate that their members vary in their constitutive polyadic property. Marie Curie, Frida Kahlo, Julia and Laura do not identify with the same relatum, they all are identify­ing with their own individual version of this relatum.

This brings us back to Zack's and Mikkola's concern that too much individualization could cause the dissolution of the gender category. For if each person can realize their own version of womaness*, does this not mean that one can no longer speak of a common category? No, it does not; because, as we have already seen in section 3.2, we can give up the idea of gender uniformity without having to give up essentialism as well, as universalism and essentialism do not necessarily have to be equated (section 2.4).

This is what saves us from the dissolution of our categories, because, if we do not equate uni­versalism and essentialism, we are not getting into a situation, in which we have to define several categories in the first place. For, it is not categorization itself that is problematic; the difficulties with intersectionality analysis only occur when we try to define uniform gender categories in particular (see section 3.2). Consequently, if we just reject universalism and thereby the premise of gender uni­formity, we can still theorize gender by defining manifold categories instead; that is, categories that are still essentialistic, because they are based on a common property, while they also take into ac­count the intersectional analysis, thus leaving open the possibility of individual social influences, by abandoning the universalistic VxFx scheme. Our conclusion therefore remains that real categoriza­tion is still possible as long as we base those categories on a multi-realizable property (see section 3.2). Such a property is not uniform but common to all its members and thus constitutes a manifold category that does not run the risk of fragmenting into countless subcategories.

But what is the common property, if any woman* realizes her own version of womaness*? What holds Bettcher's category of women* together? It is the fact that all members, i.e. all women*, identify themselves with such a version of womaness* in the first place, because that is what distin­guishes them from non-women*. Conversely, this means that all people, who realize some wom- aness* in some way, i.e. who possess the constitutive polyadic property, of “sincerely self-identifying with whatever they think it is to be a woman*”, are members of the category of women*. Accord­ingly, Marie Curie, Frida Kahlo, Julia and Laura all belong to the category of women as long as they sincerely self-identify with it in some way.

Consequently, Bettcher manages to define manifold gender categories, which are intersection- ality affirmative, because they allow and reflect individual differences, depending on one's social background, while at the same time they still resist dissolution, because these categories are held together by a multi-realizable property that is essential for category-membership.

This argument can be applied to other relational categorization attempts as well, as long as the authors base their categories on intersectionality affirmative properties. The reason is that the result­ing manifold categories can fulfil two demands at once; something that monadic properties were not able to do. Where uniform categories had to decide between acknowledging diversity and losing real categories or keeping real categories and dismissing diversity, an unpleasant decision indeed (see section 3.2), manifold categories are not even faced with a choice. On the one hand, they provide real gender categories because they track a property possessed by each member (e.g. sincere self­identification), and on the other hand, they also portray the diversity of its members by drawing at­tention to the fact that each of them is realizing the constitutive property in their own individual way.

Consequently, some first person relational theories, like Bettcher's, can satisfy both aims of the intersectional analysis (the allowance and the representation of diversity) and yet do not result in the dissolution of gender categories. Additionally, there are theories that can be classified as third person relationalism and that can satisfy these criteria as well. Haslanger, for example, falls into this category, because according to her theory “women have in common that their (assumed) sex has socially disadvantaged them” (Haslanger 2012b, 239). Thus, all women have the constitutive polyad­ic property of “being socially disadvantaged as assumed females” (VwDwf). Again, the important difference compared to MacKinnon is that Haslanger's theory is not about the actual characteristics of the woman in question (it does not matter if she really is female), but about her being identified as a person in whom such characteristics are to be expected (so it does matter if she is perceived as female). This difference provides room for intersectional influences, as in each case different mark­ers can be of importance as to whether someone is perceived as female or not. Thus, intra- categorical marginalization is evidently not a problem of relationalism in general, but one that pri­marily concerns approaches that, like Zack, use a limited polyadic property, by not defining a varia­ble second relatum.

But what about wrongful exclusion, a problem that has troubled us since the beginning of our inquiry and that is also present in Zack's relational approach. Can this problem also be overcome with the right polyadic property? We will address this question in the next section.

4.3 The remaining exclusion problem

Principally, relationalism can be regarded as a great improvement compared to the universalism we started with (in section 3); for it provides us with the opportunity to constitute gender categories based on multi-realizable properties, by replacing the universalistic VxFx scheme with a more flexi­ble VxRxy scheme. Even if not all philosophers fully exploit the flexibility, some (like Bettcher and Haslanger) actually succeed in developing a gender theory that acknowledges the diversity of gender in matters of social background, i.e. one's race, one's class, one's religion. However, even those in­tersectional affirmative approaches do not manage to fully overcome the exclusion problem. Re­gardless of whether we start from first person relationalism or third person relationalism, there are still some cases in both approaches where certain people are not considered to be women* even though they should be.

Within first person relationalism, the exclusion is caused by the presumption that every person is able to identify with her* gender of choice. This premise is problematic because some mentally dis-abled people might not be capable of such a sincere self-identification act, as it is for example defined by Bettcher (see Barnes 2020, 7). For it is part of Bettcher's definition that the person in question declares herself as a woman* and also acts as such. Consequently, it is necessary for a per­son to communicate their gender in a way that is recognizable to others, to count as that gender. So, people, who do not have these communication skills, are deprived of membership to any gender category.

For illustration, think for example of a 70-year-old woman, let's call her Berta, who has spent her entire life identifying as a woman and who has further expressed this decision both verbally and non-verbally. Now, in her advanced age, Berta gets dementia, which suddenly prevents her from sincerely identifying as a woman any longer (i.e. she cannot declare that she is a woman nor do her actions reveal that she considers herself to be a woman). Thus, if one strictly follows Bettcher's the­ory, Berta no longer meets the necessary conditions to be (considered) a woman. Nevertheless, it seems wrong to exclude Berta from the category of women, once her dementia has progressed. The reasons behind this discomfort are, first, that it contradicts common sense and second, that “[w]e would be denying that women[*] without certain cognitive capacities deserve the label ‘women[*]', specifically because of how their cognitive capacities differ from what is typical” (ibid.). However, such a deviation should not be a reason to deny someone their gender status, especially since men­tally dis-abled women* are also victims of gender-specific oppression (see Barnes 2020, 7).

At this point it is reasonable to think that maybe third-party-identification can overcome this exclusion problem of mentally dis-abled women*. After all, Berta's children and grandchildren are still able to identify her as a woman, so in that case third person relationalism seems to be unprob­lematic. Unfortunately, though, while third person relationalism is able to categorize Berta as a woman, it has its own exclusion problems elsewhere, because, here, certain trans women* are at risk to get misgendered once again.

The decisive point is that trans women* are not always perceived as women* by others. Just remember Laura, and her struggle to be recognized as a woman by society. Even if she might count as a woman in her own community because this is a trans*-friendly environment, she would lose her womaness as soon as she moves outside this community, since trans-unfriendly (or simply unin­formed) people would then no longer identify her as a woman. Thus, if we adopt a third person relationalism, “we risk saying that if people are confused enough by your gender, that's enough to make you not really a woman[*]. And that seems wrong” (Barnes 2018, 587).

So, relational theories, which try to constitute gender categorization via identification, are still wrongfully depriving less privileged women* a membership in the category of women*.⁴⁷ In the case of first person relationalism, it is the presumption of certain mental capacities, which overlooks some women*. In the case of third person relationalism, it is the reliance on external expectations of how a particular gender should be expressed that does not apply to all and only women*. Thus, it can be concluded, even these extremely flexible approaches, in which the relatum can be individually modified, do not manage to represent the diversity of gender in its entirety.

One could try to argue at this point that hybrid relationalism could be the solution, since the women*, who are excluded in one approach, are included in another. But such hybrid approaches have their own difficulties as they are often wrongfully inclusive: A hybrid approach roughly says that all people who identify or are identified as women* are women*. However, people who actually identify as men or non-binary may also be included in this category, because they are identified as women via third-party-identification, and this would also be problematic (see Jenkins 2016, 398­399).⁴⁸

But, why is it so difficult to include all and only the desired people in the corresponding gen­der category? And how can we finally overcome this remaining exclusion problem?

5 The second methodological shift - Gender pluralism

In order to overcome the gender category dilemma, by including all and only the desired people within the corresponding categories, we have to take a step back and return to our examination of universalism, because this is where the exclusion problem originates in the first place. In section 2.4 we saw that universalism is based on the assumption of gender uniformity, a premise that is deeply problematic because it disregards the diversity of gender and thereby reinforces solipsistic tenden­cies, by establishing normative requirements (section 3). Consequently, we rejected the idea of gen­der uniformity and turned instead to relationalism, trying to construct womaness* as a manifold category. This aim is achieved with different degrees of success by the individual theories, depending on how variable the second relatum is. Especially intersectional affirmative approaches enabled us to reduce solipsistic tendencies within current gender theories, because they are able to visualize other intersecting categories, so that stereotypes are not unjustifiably favoured, above less privileged peo­ple.

But, nevertheless, relationalism alone is still unable to fully reflect the diversity of gender, be­cause it still tries to reduce gender to a single aspect (for example self-identification, if we adopt first person relationalism; or social perception, if we adopt third person relationalism⁴⁹ ). I will call this notion of gender as reducible to one aspect, a “gender monism”, because it views gender as a one­dimensional concept. Complementary to this, I will call theories that view gender as multidimen­sional, i.e. that do not assume that the different aspects can be reduced to a common factor, “gender pluralism”.

In the following, we will first address gender monism, to see why this perspective is unable to reflect the diversity of gender (section 5.1), thereby using the irreducibility argument, which can be constructed in parallel to Spelman's argument from diversity. Then we will turn to the general ad­vantages of gender pluralism (section 5.2), and further investigate how pluralistic approaches can overcome the exclusion problem, by entailing unproblematic normative implications (section 5.3). Afterwards, we will discuss why gender pluralism is indeed able to constitute real gender categories, despite possible concerns (section 5.4).

5.1 The argument from irreducibility

In section 2.4 we defined universalism as a position that implies the existence of a universal wom- aness*, which all and only “real” women* correspond to. This thought contains two underlying premises. First, the premise of gender uniformity, postulating that all women* have to instantiate the same universal. Second, the premise of gender monism, postulating that each gender can be cap­tured by a single category; for, if there is only one universal womaness*, there is only one corre­sponding category of women*. In section 4 we saw that relationalism rejects the first premise of gender uniformity, however, it still maintains the monistic perspective, whilst trying to place all women* in a single category.

This premise of gender monism, however, is also responsible for the solipsistic tendencies of realistic categorization attempts, just like the premise of gender uniformity. The reason behind that is that a single property, even defined by intersectionality affirmative relationalism, is still insufficient to cover all the different aspects of gender. So, yet again, it is the ignorance of the diversity of gen­der, which causes problematic normative implications resulting in wrongful exclusion (see section 3.3). Only now is it not the multiple realizability of a gender which is ignored, but it is the plurality of gender itself, which is illegitimately reduced to one aspect.

Accordingly, there is an argument from irreducibility, which can be constructed analogous to Spelman's argument from diversity (see Spelman 1988, 159).

Argument from irreducibility:

P1IR: Current categorization attempts still regard gender as a monistic concept, i.e. they assume that each gender can be reduced to a single aspect, which is then constitu­tive for the corresponding gender category.
P2IR: Gender is too diverse, to be reduced to a single aspect.
P3IR: If one ignores P2IR and, nevertheless, defines gender in a monistic manner, this approach entails solipsistic tendencies, which inevitably end in wrongful exclusion. common ground. Let us take our two familiar aspects (self- and third-party-identification) as an ex­ample:

Evidently, both self-identification and social perception are equally important aspects of gen­der (see Jenkins 2018, 727). On the one hand, it is important for a self-identifying woman* that she self-identifies as a woman*, because this dictum influences how she lives her life. For example, Laura wears her hair long and Julia nurses her old grandmother, because that is what they expect of themselves as women. It is important to note here that this does not mean that the characteristics themselves (wearing long hair or nursing old relatives) are feminine characteristics. It only means that the motivation of the person who performs them (Laura and Julia) can be partly attributed to the fact that they identify themselves as women. On the other hand, it is important for a socially- perceived woman* that she is identified by a third party as a woman*, since this perception influ­ences how other people interact with her and, also, which scope of action is available to her. For example, a self-identifying woman of male stature in traditionally masculine clothing would not be welcomed into a public women's bathroom in Turkey, because the other bystanders would not per­ceive her as a woman (see Asta 2018, 78).

This status of “being a woman*” and at the same time “not being a woman*” is due to the fact that the two aspects (self-identification and social perception) constitute different categories of women*. When we rely on self-identification as the fundamental aspect of gender, we are choosing “self-identification as a woman*” as the essential property, constituting the corresponding category of (VwJwy). When we rely on social perception as the fundamental aspect of gender, we are choos­ing “third-party-identification as a woman*” as the essential property, constituting the correspond­ing category of (VwKwy). Both categories are admittedly congruent to a large extent, but they are not completely the same. Like two circles that overlap each other to a large extent, but still leave a narrow edge protruding on each side. These edges symbolize cases, in which a woman* only fits into one of the two categories. This mostly concerns less privileged women* like Laura or Berta, but sometimes privileged women* too, as the bathroom example demonstrates. Consequently, neither of the two categories can be reduced to the other category without wrongfully excluding people from the category of women*. Or in the words of the metaphor, whenever we choose a circle, we lose the protruding area on the other side.

Therefore, the P2IR is justified: Gender is too diverse to be reduced to a single aspect. Moreo­ver, if we ignore this diversity (P3IR), an exclusion problem arises, because of the normative side of categorizing. For as we saw in section 3.3, the normative requirements become solipsistic once we try to define properties that are possessed by all and only women* (i.e. we cannot get the two circles to fully overlap). Thus, we end up cutting back the edges (again) as we try to maintain one all­encompassing category, and this is the reason why we have a remaining exclusion problem. After all, one will always draw the circle in such a way that oneself is included in its set, i.e. one will define gender categories in which oneself is categorized “correctly”, possibly even without noticing that the category is not an all-encompassing one, because most of the time there is a large overlap with other aspects of gender.⁵⁰

Optionally, one could try to solve this problem by constructing a third circle that contains the other two circles, i.e. one could try to define an overarching category, saying that all persons are cat­egorized as women* who either identify as women* or are identified as such (see e.g. Stoljar 2017, 215-216). But this would be a sort of hybrid approach and, as we already saw in section 4.3, a hybrid approach is not really a solution, since such an overarching category would then also categorize some trans men* and non-binary persons as women*, which is just as problematic as wrongful ex­clusion (see again Mikkola 2016, 65-70).

Consequently, the best option to overcome this remaining exclusion problem is “to employ a distinction between gender as a third-person social status and gender as a first-person identification” (Stoljar 2017, 216), so that both categories can coexist as equally important. However, to achieve such a plurality we have to stop regarding gender as a one-dimensional concept, because as long as we try to reduce gender to a single aspect, we favour this aspect above the others. Thus, the remain­ing exclusion problem can again be traced back to universalism, only this time it is the premise of gender monism that must be rejected in order to represent the diversity of gender in its entirety. Therefore, our way out of this problem is provided by gender pluralism.

5.2 An introduction to gender pluralism

It has been shown that gender monism is not a good starting point when theorizing gender. Either one favours one aspect and thereby dismisses another aspect, or one chooses the overlap of all as­pects as constitutive for the corresponding gender category and thereby excludes all protruding parts, or one tries to create an all-encompassing category and then also includes unwanted parts. All three scenarios are undesirable. Now, the alternative is to abandon gender monism and to shift to­wards gender pluralism instead, but what are the advantages of this second methodological shift?

The most important advantage is that with a pluralistic perspective we regard gender as a mul­tidimensional concept consisting of equally important aspects. Thereby, each aspect can constitute its own independent category, so each gender is not captured by a single category but rather by a category complex. This means that in a pluralistic approach, the category of women* is further dif­ferentiated into more specific sub-categories like self-identifying women* and socially-perceived women*. There are probably even more aspects of gender (and therefore more sub-categories) than those two; it is only that self-identification and social perception are the two most prominent aspects in the current literature, which is why, here, we are focusing on them.⁵¹

Some theories do even contain a disclaimer, in which the authors point out that they do not claim to define gender as a whole, but only to illuminate certain aspects. Nevertheless, the authors are usually interpreted in such a way that they at least consider the aspect under discussion to be of particular importance. For example, Jenkins says of Haslanger: “[A]lthough she acknowledges that gender terms are used in multiple ways, she takes the idea of a social class based on presumed sex to capture the sense of gender that is most important for a critical feminist analysis of gender” (Jenkins 2016, 397).

If we look at the original text then Haslanger says at the beginning: “I do not want to argue that my proposals provide the acceptable ways only to define [...] gender; in fact, the epistemologi­cal framework I employ is explicitly designed to allow for different definitions responding to differ­ent concerns” (Haslanger 2012b, 222). But later she admits that: “On the account I've offered, it is true that certain females don't count as “real” women[*]; and it is true that I've privileged certain facts of women[*]'s lives as definitive” (ibid., 240), which again leads to the marginalization⁵² of women*, who do not possess these “privileged certain facts” (see section 4.3).

Therefore, it is not enough to recognize the other aspects of gender as co-existent, one further has to consider them as equally real: A socially-perceived woman* is not “more woman*” than a self-identifying woman*, and a woman* who falls into both categories is not “more woman*” than a woman* who falls into only one of the two categories. Social perception and self-identification are simply two separate aspects of gender. This does not mean that they cannot influence each other, but at least they cannot be preferred over each other.

Jenkins also achieves a similar conclusion, for, after identifying Haslanger's approach as being cis solipsistic (i.e. after saying that it passes over certain trans people (2016, 397-402)), she goes on to modify Haslanger's theory to the effect that it has two equally important “branches”, namely “being classed as a woman [*] and having a female [*] gender identity” (ibid., 415-416 original emphasis). Although there are still legitimate reservations about Jenkin's particular approach (see Andler 2017 & Bettcher 2017b),⁵³ the underlying development towards more plurality is nevertheless a good one.

The decisive point is that a woman*, who belongs into more than one category, just is a wom­an* in more than one aspect and nothing more. For, within gender pluralism there is no “real” womaness*, (no ultimate circle, to get back to the metaphor); all there is are different ways in which one can be a woman* (i.e. many different overlapping circles to which one can belong or not).

Now, this makes a huge difference in the way we portray gender; and in order to demonstrate this, it makes sense to come back to Marie Curie, Frida Kahlo, Julia, Laura and also Berta. First of all, regarding gender pluralism all of these women can be legitimately considered to be women. So, even if this means different things for some of them, none of them is wrongfully deprived of their gender. Moreover, we can say about Marie Curie, Frida Kahlo, and Julia that they are all self­identifying and socially-perceived women and thus simply belong to two categories, realizing two different kinds of womaness (i.e. they are located in the overlap of the two circles). Again, this does not make them “more women” but it only makes them women in more kinds of ways. Conversely, it does not make Laura any less of a woman that she is a self-identified woman but not a socially- perceived woman, it merely reflects her reality. For Laura is indeed denied access to the category of socially-perceived woman in many situations. So, while it is problematic to say that Laura is not a woman, it is not problematic to say that Laura is not a socially-perceived woman. On the contrary, this statement can be used to claim that maybe all self-identifying women* (or at least all self- expressive women*⁵⁴ ) should also belong to the category socially-perceived woman* (for further elaboration see Asta 2018, 77-79). In the context of the circle metaphor, this would mean moving the circle of self-identifying (or self-expressive) women* to the point where it is included in the cir­cle of socially-perceived women*.

Similarly, to Laura, Berta is also not less of a woman because she is a socially-perceived but not a self-identifying woman. Again, this categorization only reflects Berta's reality, as Berta is not a self-identifying woman, at least not in Bettcher's sense of self-identification, but she is nevertheless socially perceived as a woman, with all the consequences that this entails. Nevertheless, this situation is different from Laura's in that it is not desirable to make all socially-perceived women* self­identifying women* as well, because then we arrive at wrongful inclusion again. So, just because the circle of self-identifying (or self-expressive) women* can be contained in the circle of socially- perceived women*, this does not mean that the two circles are congruent. The circle of self­identifying (or self-expressive) women* can simply be a smaller part of the circle of socially- perceived women*.⁵⁵

Thus, if we treat gender as such a category complex rather than a single category, we can final­ly meet our first target requirement of non-exclusiveness, because we are no longer under the sword of Damocles, of having to define “some golden nugget of womaness[*] all women[*] have as wom- en[*]” (Spelman 1988, 159). In addition, we are able to show the relationships between the individu­al aspects of gender - something that was also not possible with former monistic approaches.

Previously, we always had to commit to a single property; and, even if this property was an in- tersectionality affirmative one (like self-identifying as a woman*) this monistic perspective always led to the privileging of one aspect of gender (like self-identity) and thus a certain type of womaness* (like self-identifying women*), thereby wrongfully excluding or marginalizing less privileged women* as not being “real” women* (like mentally-disabled women*). Now, there is no longer such a com­mitment, instead all the different aspects of gender can coexist, and we can investigate how they interact with each other.

But why can gender pluralism overcome wrongful exclusion? To answer this, we will now take a closer look at the normative implications of these new “specific categories⁵⁶ ”, as previously, these implications were the reason for the solipsistic tendencies and the unjustified exclusion (see section 5.1).

5.3 Unproblematic normativity

Before we get to gender pluralism in particular, let us first recall the problem of normative implica­tions in general. Basically, every case of categorization entails normative implications. Take for ex­ample the category of women*:⁵⁷ Not anybody can belong into the category of women*; otherwise, the category itself would dissolve. So, instead of just including anybody, one has to possess a consti­tutive property to be a category member. Conversely, there will be some people who are excluded from the category (and some always will be), because they do not acquire the constitutive property, to be a category member (see section 3.3).

This establishes, according to Mikkola's Butler interpretation, that the realist project is des­tined to fail, because “every definition of ‘woman[*]' will be insidiously normative and thus political­ly problematic” (Mikkola 2017, 173). Although, without gender categories we lose the ability to talk about gender related oppression, thus, it seems, once again, as if the rejection of universalism, more precisely universalism's premise of gender monism, has led us back into the gender category dilem­ma, where we are unable to satisfy both target requirements at once. But, yet again this relapse only occurs if we equate universalism with essentialism: Yes, all categorizing is in some way normative, and yes, every universal categorizing about gender is problematic due to its normative implication, which leads into exclusions. But the reason why all universal categorization attempts are problematic is the presumption of gender as a monistic concept (see section 5.1). Therefore, there can be un­problematic categorizing about gender, despite the normative implications, as long as we acknowledge gender in its diversity and consider it to be a pluralistic concept.⁵⁸ In fact, Mikkola also reaches this conclusion, that normative implications alone are not sufficient to preclude the realistic project, albeit for different reasons. She claims: “After all, we can revise gender notions in a mean­ingful way, which demonstrates that normativity per se is not the problem—misguided and exclu­sionary normativity is” (ibid., 174).

Now, the decisive change of gender pluralism which sets unproblematic normativity apart from this “misguided and exclusionary normativity” is that the normative requirements, no longer apply to an entire gender category but only to specific categories, reflecting certain aspects of gen­der. In doing so, pluralism manages to be normative in an unproblematic way. Let me elaborate this point:

If we are defining specific gender categories, all that we claim is that, for example, all socially- perceived woman w *⁵⁹ share the constitutive property of being socially perceived as a woman* (i.e. the polyadic property w D f) and therefore they all meet the normative requirement (they all are third- party-identified as women w D f) to belong to the category of socially-perceived women* (w V w D f). Hence, we only talk about the aspect of gender as a class, i.e. about what it is to be a socially- perceived women*, and not about what it is to be a “real” woman*. So, the people outside the cate­gory of socially-perceived women* (i.e. the ones who do not possess the required constitutive prop­erty w D f of being third-party-identified as women*) are not considered to “not be women*”, but only to “not be socially-perceived women*”, and since this is in fact the case, for otherwise they would be in the category, the normative requirement is not wrongfully exclusive (see the example of Laura section 5.2). Thus, there are all and only the desired people in our category of socially- perceived women*, because the desired people no longer consist of “women*” but “socially- perceived women*” in particular. So, in that case, “the hope is that the account's implication would not reinforce but would help undermine the structures of sexual oppression” (Haslanger 2012b, 228-229). Since, due to this new pluralistic framework, we can start a howl new debate on whether it is acceptable that women like Laura are not socially perceived as women, and how we could then change this fact.

In sum, we can say that the second methodological shift leaves us with a plurality of specific categories and without any overarching category of just “women*” or “real women*” or whatsoever. This solves the exclusion problem because finally all and only the desired people belong into the corresponding category. Admittedly, there can still be solipsistic tendencies, for example when phi­losophers favour a certain category by saying that this is the “true category” or something like that. However, we can now call out such incidents with the help of gender pluralism; and we can enter into a discourse about whether it makes sense to focus on a specific aspect at a certain point, with­out saying that this is a more fundamental aspect. This idea also resonates with Mikkola as she ex­plains that “different metaphysical theories of gender can do different work for us” (2017, 178).

But, does this idea of plurality not contradict Mikkola's concerns about the intersectional anal­ysis? After all, we now have exactly what Mikkola feared: several categories of women*.

5.4 Specific categories as real categories

Gender pluralism reveals that the all-encompassing category of women* does not exist; and that gender is instead a multidimensional concept whose diversity can only be fully represented with the aid of several categories, each of which focuses on a specific aspect of gender. This plurality has the advantage of not being wrongfully exclusive, as people who are not included in one category can still belong to another category (like Laura and Berta — see section 5.2). Nevertheless, opponents might argue, as stated above, that this specific categories fail to meet our second target requirement of real gender categories. Similar to the objection to intersectional analysis (see section 3.2 & 4.2), this line of argument might bring to attention that such a plurality entails the danger of too much individua­tion, so that the categories lose their significance.

However, these objections can be easily resolved; for, the initial worry was that “we end up dissolving groups into individuals” (Mikkola, 2017, 171) and this is not the case with specific catego­ries. Of course, one can debate about how best to define the category of socially-perceived women, but this does not break the category itself apart. If one should indeed come to the conclusion to further differentiate this category, this decision would have to be well justified, like the asset of some additional information.⁶⁰ Furthermore, it does not follow that, just because gender is seen as a plural­istic concept, one can just randomly define gender categories for any fairly common property. The category still has to capture a fundamental aspect of gender, meaning that it must not be reducible to another aspect and, furthermore, it should also neither reinforce existing misogynist structures nor produce new ones.

One could still argue that these more specific categories are, nevertheless, insignificant, not because they lose themselves in individuation, but, because they only inform us that the category­members possess the property to which they owe their membership and nothing more. However, this objection can be refuted as well. Yes, the category-membership does not determine its mem­bers, but that does not mean that the membership has no influence on the members at all. As we saw in section 5.1, some people are denied the use of public bathrooms because they do not belong to the category of socially-perceived women. And this is just one of many examples where certain aspects of gender are linked to certain privileges (for more examples see Asta 2018).

These relations should be of particular interest to us, as they can also help us to “name, un­derstand and therefore work against these forms of gender-related oppression” (section 2.2). For further illustration, recall the opening example of sexual violence against women (also section 2.2). As Jenkins also notes, “‘[sexual] violence against women' is a form of oppression that operates both through gender as identity and through gender as class” (Jenkins 2016, 420),⁶¹ so, if we want to take political action against this form of gender-related oppression, gender pluralism is a good approach to differentiate, who is actually affected. Of course, self-identifying women are particularly vulnera­ble to sexual violence, even or especially if they are not perceived as “real” women by their social environment.⁶² But the same applies for people who do not identify as women but are socially per­ceived as women, for example people who identify as non-binary but who were designated female at birth and who possess the associated body features (see ibid., 419-420). According to this, both gen­der self-identification and third-party-identification have an influence on someone's vulnerability to sexual violence. It is therefore not possible to adequately address the problem by focusing on a sin­gle aspect, rather by exploring all relevant aspects equally well.

Moreover, if one is following Jenkins' real world example, and organizes a “Reclaim the Night march” that is “women-only” (ibid., 419), to raise attention to the issue of higher sexual violence risks, a gender pluralism is a better theoretical foundation than gender monism. Because, adopting a monistic approach, one is forced to reduce gender to one aspect, i.e. one category only. Thus, one has to decide between self-identifying women and socially-perceived women, who of them are “real” women. Either way, there are people wrongfully excluded from the march, even though they are victims of this type of gender-related oppression, simply because they do not belong into the catego­ry of choice (see Jenkins 2016, 419-420).

With a pluralistic approach, in contrast, one can keep both aspects of gender visible by simply operating with two categories instead of one; all in view of the fact that there is no “real” wom- aness*, the two categories can be reduced to. A similar conclusion was reached by Jenkins, as well, for she sees gender as class and gender as identity as equally important, and so she decides that: “The march is open to all self-defining women. If you do not define as a woman* but experience discrimination because you are perceived as female, you are also welcome to attend” (ibid., 420). So, yet again, a pluralistic perspective allows for a more detailed examination, and is hence anything but insignificant.

However, pluralism alone is also not enough to fully understand the problem of sexual vio­lence against women; one still has to keep the intersectional insights in mind; and not only in this example but in general. After all, the category of self-identifying women, for example, is still shaped by the fact that every member identifies with what is considered feminine in their social environ­ment. The previous argument in favour of pluralism is, therefore, to be understood in such a way that the specific categories are also manifold categories; for pluralism and relationalism must be combined, to ensure that no wrongful exclusion or marginalization takes place.

This is not to say that pluralistic relationalism is the ultimate way to theorize gender, but that it is currently the most preferable option to operate with. Because, while all known alternatives inevi­tably lead to the gender category dilemma, pluralistic relationalism opens up the possibility of satisfy­ing both target requirements at once, and therefore it dissolves the dilemma rather than getting stuck in it.

6 Pluralistic Relationalism

In the first methodological shift (section 4) we saw, that relationalism is a better point of departure when theorizing gender, than universalism, because relationalism opens up the possibility of consti­tuting gender categories on the basis of a multi-relational property. This is a general advantage be­cause multi-realizable properties can meet the intersectional aims and make the diversity of women* visible to viewers. Then, in the second methodological shift (section 5) we saw, that the pluralistic perspective on gender is superior to a monistic one, because a pluralistic perspective allows us to focus on specific aspects of gender without denying the others. This is also a general advantage, in our aim to dissolve the gender category dilemma, because in allowing for multiple categories of women at once we circumvent the problem of wrongful exclusion, by acknowledging that women* can be women* in more than one way.

What the two shifts have in common is that they both lead us in the direction of more diversi­ty without becoming insignificant, and this is decisive for dissolving the gender category dilemma. So now we will see how a combination of both provides us with a possible framework to develop theories that do not miss either of the two target requirements. Hence, the point of the following (section 6) is not to present a full-blown theory, but to explore how both shifts lead us to a place where we are able to develop theories that take place outside the dilemma.

If we combine pluralism and relationalism, the gender categories become both specific and manifold categories: a) specific, because they only cover a specific aspect of gender (see section 5.2); and b) manifold, because these specific aspects are additionally influenced by the social background of the category-members (see section 4.2). In this way, the high diversity among women* can finally be represented by, first, several categories that represent the different ways in which one can be a wom­an*; and second, these categories also being inherently diverse, due to their constitution by the rela­tional VxRxy scheme.

Accordingly, a specific, manifold gender category, for example, the category of self-identifying women* can be compared to the category of objects in my garage: When we look at all the objects currently located in my garage Vo, “being located in my garage” is the constitutive property Bog for the corresponding category VoBog. Here, the multi-realizability of the constitutive polyadic property Bog ensures that the objects in my garage are not determined any further. (I can have a number of different objects in my garage, which have nothing in common other than their current location. For example, my car, my old roller skates, a ladder, some tools, my sister's rocking horse, etc.) What the property Bog defines is merely the membership in the corresponding category “objects in my gar­age” VoBog (Thus, as soon as my sister gets her horse back, it no longer belongs to VoBog). Addi­tionally, we are not considering all my belongings to be members in the category VoBog, but only those objects, which are currently located specifically within my garage.

Transferred to the category of self-identifying women*, here relationalism ensures that the category is constituted by a mutirealizable polyadic property that leaves enough room for intra- categorical differences by not determining these women* any further; while pluralism, causes us to specify our target, so that we no longer have to reconcile all aspects of gender.

According to Witt the category of “objects in my garage” is only a collection, i.e. a mere arbi­trary group of things (see Witt 2011, fn. 1). But I do not agree with her, because all of these objects still share a common property, the property of “being located in my garage”, therefore they are not 53

mere arbitrary. Moreover, belonging to the category VtBog still has important consequences for its members. For example, if I have a garage sale and I put all the things in my garage up for auction, it makes a big difference whether my sister's rocking horse is still located in my garage or not.

The same applies to specific, manifold gender categories, because if a person belong to the group of self-identified women*, this property still distinguishes her* from non-category-members, and she* will experience some consequences because of this distinction. So, for example, Julia has a sense of duty to nurse her old grandmother which her brother lacks, because Julia understands this solicitude as part of her womaness*. So, Julia's membership in the category of self-identifying wom­en is a fact, which does not determine her (her sense of duty can change, whilst she still self­identifies as a woman, i.e. still belongs to this specific category) but it influences her (for now she has this sense of duty because it is currently part of her concept of womaness). This also underlines again, that gender is not a mere fiction, as all of its various aspects (like once gender identity) have impacts on our reality.

Thereby, pluralistic gender relationalism shows us a way, to successfully dissolve the gender category dilemma, by meeting both requirements at once. The first methodological shift provides us with the possibility of reducing solipsistic tendencies and ensuring that no one within the gender categories gets marginalized and the second shift further adds to this reduction of solipsistic tenden­cies, by emphasizing that these manifold categories are only meant to capture certain aspects of gen­der (like self-identifying women*, socially-perceived women*, etc.). Therefore, the combination of both shifts (i.e. pluralistic relationalism) liberates us from the remaining exclusion problem.

So, if we carry out both shifts, we can proceed by defining gender theories which need not end in the gender category dilemma: The first shift ensures that the constitutive property is multi- realizable (so once self-identification, for example, can be individually influenced by intersecting categories); while the second makes it easier to settle for a constitutive property for now we only try to define specific gender categories instead of all-encompassing ones.

Nevertheless, the specific, manifold categories can remain real categories, because they display real social processes, which influence our lives in various ways. “[T]hey explain things” (Barnes 2017, 2425), like why a masculine-appearing woman is not welcomed in a public Turkish bathroom; “they cause things” (ibid.), like Laura letting her hair grow long; and “we need to understand them to understand the structure of the world” (ibid.), without these categories we cannot fully under­stand how, for example, sexual violence against women functions as a form of gender-related op­pression. As Barnes so eloquently concludes, “they are real because we made them” (ibid.).

7 Conclusion

We began this inquiry by asking which methodological shifts can help us to dissolve the gender cat­egory dilemma. Yet, in order to find a loophole, we first needed to know what caused the dilemma in the first place. So, we inspected the two target requirements and the resulting dilemma, and we found out that the first horn of the dilemma is indeed contestable, because, it is not categorization per se that fails to meet the first requirement of non-exclusiveness, but only certain ways of catego­rizing do so. Hence, we asked ourselves, what makes categorization wrongful and in response we exposed universalism as the systematic error, which ensures that all former realistic categorization attempts entail solipsistic tendencies, and thus wrongful exclusion. So, because universalism was the reason for the dilemma, we concluded that by overcoming universalism we would also overcome the dilemma. Thus, we localized our loophole in finding a way of categorizing that is essentialist and realistic but not universalistic. An aim that is theoretically possible since all three positions must not necessarily be equated.

So, now that the framework had been made a bit more explicit, we focused on the reasons why universalistic categorization must result in the gender category dilemma, with the underlying aim of overcoming the respective reasons and thus detecting which methodological shifts are gener­ally needed to develop a theory which circumvents the dilemma. In this context, we discussed the argument from diversity which could be further divided into the arguments from empirical evidence and intersectionality. The result was that the universalistic premise of gender uniformity is precluded by the fact that there is not one uniform womaness*, and that womaness* is rather realized in vari­ous different ways, depending on the remaining social environment. Moreover, we identified the ignorance of this diversity as the underlying cause of the solipsistic tendencies; because the process of categorization also has normative side to it, by choosing a certain property to be the constitutive one. Thus, if this property is uniform F a lot of people fail to possess it, due to the intra-categorical diversity. Here we additionally realized, since categorization is always done from a position of pow­er, it not surprising that it is mainly the less privileged people that are wrongfully excluded.

In order to overcome this idea of uniformity, so that we do not establish or reinforce prob­lematic normative implications, we performed the first methodological shift towards gender rela- tionalism. At this point, the intersectional insights had already shown us that it is desirable to define multi-realizable properties as being constitutive, and this is exactly what relationalism enabled us to do. For in relationalism, the associated categories are based on polyadic properties and this offers the possibility for the scope for intersectional influences. However, we also saw that relational cate­gorization attempts do not necessarily have to be intersectionality affirmative, and we concluded that it is further important to ensure that the constitutive polyadic property is not only multi-realizable but also represents this multi-realizability externally. At this point we also encountered some con­cerns that the category of women* could get lost in too much individuation, if we take intersection- ality too seriously. Yet, this concern could be easily refuted, because within relational categorization attempts the polyadic property is still also the essential property that constitutes one (and not sever­al) category(ies) of women*. The only difference is that the category of women* is no longer a uni­form but a manifold category; and therefore, we have succeeded in overcoming one premise of uni­versalism.

Consequently, the first methodological shift towards gender relationalism has already brought us closer to resolving the gender category dilemma, because the possibility of intersectionality af­firmative properties allows us to formulate theories with less solipsistic tendencies and to represent a greater diversity of womaness*. However, on closer inspection we had to admit that this first shift, on its own, is still insufficient to dissolve the dilemma, because even relational categorization ap­proaches, which do define an intersectional affirmative property, still struggle with wrongful exclu­sion.

Therefore, our next aim was to find out why this wrongful exclusion is still in place, and yet again we could trace this problem back to universalism. We discovered that, in order to fully over­come universalism, we must not only reject the premise of uniformity, but also abandon the monis­tic perspective, which is also presupposed by universalism. So, to reveal the underlying reasons in this case as well, we constructed an irreducibility argument analogous to the argument from diversi­ty. This argument revealed that gender simply has too many equally important, independent aspects to be reduced to one of them; because, every time we choose an aspect over another, we establish normative implications that cannot be fulfilled by all women*.

We concluded that we are in need of a second methodological shift towards gender pluralism, to overcome this monistic perspective as well, because with a pluralistic perspective we are no longer forced to choose a single aspect, a single category, a single constitutive property. On the contrary, the pluralistic perspective allowed us to consider several aspects of gender as equal and thus to de­fine several specific categories for one gender, so women* were no longer represented by one cate­gory but by a whole category complex. We further elaborated that these specific categories still entail normative implications; only these implications are not problematic, because we rejected the idea of an all-encompassing category and these smaller specific categories actually can contain all and only the desired people. Possible objections that the pluralistic view would be lost in too much individua­tion or insignificance could also be refuted. After all, the specific categories are not constituted ran­domly, instead, careful consideration must be given to which aspects are really irreducible and do provide us with valuable information about gender, as these aspects are meant to reflect the differ­ent influences that gender has upon our reality.

Consequently, the second methodological shift towards gender pluralism allowed us to over­come the second premise of universalism, by opening up the possibility to constitute multiple gen­der categories. Thus, we concluded that pluralistic theories are generally in a better position to avoid the gender category dilemma, because they can represent the different aspects of gender without favouring one aspect above the others. Moreover, pluralistic theories enable us to actually discuss how all these specific aspects interact with each other and how we might want to influence these relations.

So in the end, we resonated that both methodological shifts are needed in order to dissolve the gender category dilemma; because both reject one of the underlying premises of universalism which caused the dilemma in the first place. The first shift towards relationalism rejects the underly­ing premises of gender uniformity and the second shift rejects the underlying premise of gender monism. Yet, approaches which carry out both shifts still remain essentialistic, because the resulting, specific, manifold gender categories are constructed based on a common property; and they also remain realistic, because these categories also refer to processes occurring in reality. Thus, pluralistic relationalism enables our theories to satisfy both target requirements at once, by helping us to tread the fine line between diversity and commonality. Because if we define gender categories in a rela­tional and pluralistic manner we can include and reflect all the different inherently diverse aspects of gender, while at the same time we are still able to maintain enough commonality to speak of real gender categories and not just of gerrymandered groups.

So philosophers who try to develop an adequate gender theory, one that does not result in the gender category dilemma, should first shift from categorization via constitutive monadic properties to constitutive polyadic properties; and then shift from a monistic to a pluralistic perspective, as these two methodological shifts away from universalism are what it takes, to be able to embrace the diversity of gender (i.e. the different ways one can be a woman* and the different possibilities one can realize these different ways) without losing the ability to construct real gender categories.

Thus, if you were to ask me what gender I am, I would first have to ask you back, in which aspect of gender you are interested in. The conversation could then proceed in such a way that you would ask me which gender I identify with, and then I could give you a specific answer and tell you that I identify as a woman. What exactly I mean by this statement and how me self-identifying as a woman differs from Marie Curie, Frida Kahlo, Julia or Laura self-identifying as women, are ques­tions that need to be discussed in future debates.

All I can (and want to) contribute at this point is the realization that we must shift towards pluralistic relationalism within our categorization attempts, to ensure not to end up in the gender category dilemma. Within this framework of pluralistic relationalism questions such as “What does it mean to self-identify as a woman*?” can be discussed without the expectation that the answers will provide information about what it is to be a woman* per se. So, we minimize the risk of wrongfully excluding or dismissing anyone, without reducing gender to a mere fiction.

This is not to say, that the discussions inside the gender debate are any less important or less delicate than before. On the contrary, the discussion becomes even more complex, as we have to investigate whether specific aspects such as self-identification are really just one aspect or several, how the aspects interact with each other and how individual women* realize these aspects in differ­ent ways. Hence, this openness to more diversity offers enrichment for the gender debate itself.

Bibliography

Allen, Sophie: “Properties“, in: Fieser, James & Dowden, Bradley (eds.): Internet Encyclopedia of Philoso­phy, https://iep.utm.edu/properties (Accessed: 17.03.21).

Andler, Matthew S.: “Gender Identity and Exclusion: A Reply to Jenkins*”, in: Ethics 127 (2017), 883­895.

Bailey, Alison: “On Intersectionality, Empathy, and Feminist Solidarity: A Reply to Naomi Zack”, in: The Journalfor Peace and Justice Studies 19 (2009), 14-36.

Barnes, Elizabeth: “Realism and Social Structure”, in: Philosophical Studies 174 (2017), 2417-2433.

Barnes, Elizabeth: “The Metaphysics of Gender”, in: Rosen, Gideon; Byrne, Alex; Cohen, Joshua; Harman, Elizabeth & Shiffrin, Seana V. (eds.): The Norton Introduction to Philosophy (2nd edition). New York/London: W. W. Norton & Company 2018, 581-594.

Barnes, Elizabeth: “Gender and Gender Terms”, in: Nous 54 (2020), 704-730.

Beauvoir, Simone de: The Second Sex (transl. and ed. by H. M. Pashley). London: Jonathan Cape 1956.

Bettcher, Talia M.: “Through the Looking Glass: Trans Theory Meets Feminist Philosophy”, in: Garry, Ann; Khader, Serene J. & Stone, Alison (eds.): The Routledge Companion to Feminist Philosophy. New York/London: Routledge 2017a, 393-404.

Bettcher, Talia M.: “Trans Feminism: Recent Philosophical Developments”, in: Philosophy Compass 12 /11 (2017b), e12438.

Butler, Judith: “Contingent Foundations: Feminism and the Question of ‘Postmodernism'”, in: Praxis International 11 (1991), 150-165.

Butler, Judith: Gender Trouble: Feminism and the Subversion of Identity (2nd edition) . New York: Routledge 1999.

Carastathis, Anna: “The Concept of Intersectionality in Feminist Theory”, in: Philosophy Compass 9 /5 (2014), 304-314.

Chodorow, Nancy: “Family Structure and Feminine Personality”, in: Tuana, Nancy & Tong, Rose­marie (eds.): Feminism and Philosophy. Boulder: Westview Press 1995.

Crenshaw, Kimberlé: “Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics”, in: University of Chicago Le­gal Forum (1989), 139-167.

Dembroff, Robin: “Beyond Binary: Genderqueer as Critical Gender Kind”, in: Philosophers' Imprint 20 /9 (2020), 1-23.

European Union Agency for Fundamental Rights (FRA): Violence against women: an EU-wide survey — Main results. Luxembourg: Publications Office of the European Union 2015. doi: 10.2811/62230.

Friedan, Betty: Feminine Mystique. Harmondsworth: Penguin Books Ltd 1963.

Garry, Ann: “Intersections, Social Change, and ‘Engaged' Theories: Implications of North American Feminism”, in: Pacific and American Studies 8 (2008), 99-111.

Goldenberg, Maya J.: “The Problem of Exclusion in Feminist Theory and Politics: A Metaphysical Investigation into Constructing a Category of ‘Woman'”, in: Journal of Gender Studies 16 /2 (2007), 139-153.

Harris, Angela P.: “Race and Essentialism in Feminist Legal Theory”, in: Stanford Law Review 42 (1990), 581-616.

Haslanger, Sally: “Feminism in Metaphysics. Negotiating the Natural”, in: Haslanger, Sally (ed.): Resist­ing Reality. Social Construction and Social Critique. New York: Oxford University Press 2012a, 139­157.

Haslanger, Sally: “Gender and Race(What) Are They? (What) Do We Want Them to Be?”, in: Haslanger, Sally (ed.): Resisting Reality. Social Construction and Social Critique. New York: Oxford University Press 2012b, 221-247.

Haslanger, Sally: “Ontology and Social Construction?”, in: Haslanger, Sally (ed.): Resisting Reality. Social Construction and Social Critique. New York: Oxford University Press 2012c, 83-112.

hooks, bell: Feminist Theory: From Margins to Center (2nd edition) . London: Pluto Press 2000.

James, S. E.; Herman, J. L.; Rankin, S.; Keisling, M.; Mottet, L. & Anafi, M: The Report of the 2015 U.S. Transgender Survey. Washington: National Center for Transgender Equality 2016.

Jenkins, Katharine: “Amelioration and Inclusion: Gender Identity and the Concept of Woman*”, in: Ethics 126 (2016), 394-421.

Jenkins, Katharine & Garavaso, Pieranna: “Introduction to Feminist Metaphysics”, in: Garava- so, Pieranna (ed.): The Bloomsbury Companion to Analytic Feminism. London: Bloomsbury 2018, 45­57.

Kapusta, Stephanie J.: „Misgendering and Its Moral Contestability”, in: Hypatia 31 /3 (2016), 502-519.

Keown, Damien: Buddhism: A Very Short Introduction. Oxford: Oxford University Press 1996.

Kristeva, Julia: “Woman Can Never Be Defined”, in: Marks, Elaine & de Courtivron, Isabelle (eds.): New French Feminisms. New York: Schocken 1981, 137-141.

Lewis, David: “Extrinsic Properties”, in: Philosophical Studies 44 (1983), 197-200.

MacKinnon, Catharine A.: Toward a Feminist Theory of State. Cambridge: Harvard University Press 1989.

Martin Alcoff, Linda: “Cultural Feminism versus Post-Structuralism: The Identity Crisis in Feminist Theory”, in: Signs 13 /3 (1988), 405-436.

Martin Alcoff, Linda: Visible Identities: Race, Gender, and the Self. New York: Oxford University Press 2006.

McDonald, Linda L.: “Marie Curie”, in: Great Scientist. Chemists. New York: Cavendish Square Publish­ing 2014, 37-41.

Mikkola, Mari: “Elizabeth Spelman, Gender Realism, and Women”, in: Hypatia 21 /4 (2006), 77-96.

Mikkola, Mari: “Gender sceptics and feminist politics”, in: Res Publica 13, (2007) 361-380.

Mikkola, Mari: “How Essential is Gender Essentialism? Comments on Charlotte Witt's The Metaphys­ics of Gender“, in: Symposia on Gender, Race and Philosophy 8/2 (2012), http://web.mit.edu/sgrp (Accessed: 09.01.21).

Mikkola, Mari: The Wrong of Injustice. Dehumanization and Its Role in Feminist Philosophy. New York: Oxford University Press 2016.

Mikkola, Mari: “Gender Essentialism and Anti-Essentialism”, in: Garry, Ann; Khader, Serene J. & Stone, Alison (eds.): The Routledge Companion to Feminist Philosophy. New York/London: Routledge 2017, 168-179.

Mikkola, Mari: “Feminist Perspectives on Sex and Gender”, in: Zalta, Edward N. (ed): The Stanford Encyclopedia of Philosophy (Fall 2019 Edition), https://plato.stanford.edu/entries/feminism- gender/ (Accessed: 04.03.21).

Millett-Gallant, Ann: The Disabled Body in Contemporary Art. New York: Palgrave Macmillan 2010.

Omolade, Barbara: “Black Women and Feminism”, in: Eisenstein, Hester & Jardine, Alice (eds.): The Future of Difference. New Brunswick: Rutgers University Press 1985, 247-257.

Russell, Bertrand: The Problems of Philosophy. Oxford: Oxford University Press 1967.

Scott, Joan W.: “Gender: A Useful Category of Historical Analysis”, in: The American Historical Review 91 /5 (1986), 1053-1075.

Sedgwick, William T.: “Feminist Revolutionary Principle is Biological Bosh, Prof. William. T. Sedg­wick, Noted Biologist, Riddles the Pretensions of Supporters of the New Doctrine for Women and Says if the Movement Succeeds It Will Throw the World Back a Thousand Years”, in: The New York Times (1914), 30.

Spelman, Elisabeth: Inessential Woman: Problems of Exclusion in Feminist Thought. Boston: Beacon Press 1988.

Stoljar, Natalie: “Essence, Identity, and the Concept of Woman”, in: Philosophical Topics 23 /2 (1995), 261-293.

Stoljar, Natalie: “Different Women. Gender and the Realism-Nominalism Debate”, in: Witt, Charlotte (ed.): Feminist Metaphysics. Dordrecht: Springer 2011, 27-47.

Stoljar, Natalie: “The Metaphysics of Gender”, in: Lippert-Rasmussen, Kasper; Brownlee, Kimberley & Coady, David (eds.): A Companion to Applied Philosophy. Chichester: John Wiley & Sons, Ltd. 2017, 211-223.

Stoljar, Natalie: “Gender and the Unthinkable”, in: Garavaso, Pieranna (ed.): The Bloomsbury Companion to Analytic Feminism. London: Bloomsbury 2018, 123-143.

Stone, Alison: „Essentialism and Anti-essentialism in Feminist Philosophy” in: Journal of Moral Philoso­phy 1 /2 (2004), 135-153.

Sveinsdottir, Asta K.: “The Metaphysics of Sex and Gender”, in: Witt, Charlotte (ed.): Feminist Meta­physics. Dordrecht: Springer 2011, 27-47.

Sveinsdottir, Asta K.: Categories We Live By. The Construction of Sex, Gender, Race, and Other Social Categories. New York: Oxford University Press 2018.

West, Robin: “The Difference in Women's Hedonic Lives: A Phenomenological Critique of Feminist Legaltheory”, in: Wisconsin Women's law journal 3 (1987), 81-145.

Wetzel, Linda: “Types and Tokens”, in: Zalta, Edward N. (ed): The Stanford Encyclopedia of Philosophy (Fall 2018 Edition), https://plato.stanford.edu/archives/faU2018/entries/types-tokens/ (Ac­cessed: 10.01.21).

White House Council on Women and Girls & Office of the Vice President: Rape and Sexual Assault: A Renewed Call to Aclion. Washington DC 2014, https://www.knowyourix.org/wp-

content/uploads/2017/01/sexual assault report 1-21-14.pdf (Accessed: 04.01.21).

William, H. George & Martinez, Lorraine J.: “Victim Blaming in Rape: Effects of Victim and Perpetra­tor Race, Type of Rape, and Participant Racism”, in: Psychology of Women Quarterly 26 (2002), 110­119.

Witt, Charlotte: “Anti-Essentialism in Feminist Theory”, in: Philosophical Topics 23 /2 (1995), 321-344.

Witt, Charlotte: “What Is Gender Essentialism?”, in: Witt, Charlotte (ed.): Feminist Metaphysics. Dor­drecht: Springer 2011, 11-25.

Young, Iris M.: “Gender as Seriality: Thinking about Women as a Social Collective”, in: Signs 19 /3 (1994), 713-738.

Zack, Naomi: Inclusive Feminism: A Third Wave Theory of Women's Commonality. Lanham: Rowan & Little­field 2005.

Zack, Naomi: “Can Third Wave Feminism Be Inclusive? Intersectionality, Its Problems and New Directions”, in: Alcoff, Linda M. & Kittay, Eva F. (eds.): The Blackwell Guide to Feminist Philosophy. Malden: Blackwell 2007, 193-207.

[...]

---

¹ Of course, this argumentation does not only apply to women, but to any other gender as well. Yet, since the common literature focuses mainly on women, I will also, for the sake of better readability, address women as my standard ex­ample. In the following, I will use the genderstar to show that the argumentation can be expanded to all genders.

² By gender categories that include all and only the desired people I mean, for example, that the category of women* includes all and only women*.

³ Some philosophers speak of gender terms rather than of gender categories. But, since one cannot use a term without referring to the corresponding category and vice versa, I will use both terms as implying each other.

⁴ A more detailed discussion of Intersectionality will follow in section 3.2.

⁵ A similar interpretation of MacKinnon can also be found in the Stanford Encyclopaedia of Philosophy article on Femi­nist Perspectives on Sex and Gender (see Mikkola 2019).

⁶ Another prominent example, is Betty Friedan's view that women in general are oppressed because they are reduced to being housewives (see Friedan 1963), because it disregards the fact that some women, for example poor women, single mothers, women with a non-western background, etc., are forced to take a job in order to earn money, and thus are not reduced to being housewives (see hooks 2000, 1-3 & Spelman 1989, 8).

⁷ By trans people I mean all people who identify with a gender that was not assigned to them at birth.

⁸ For explicit criticism see e.g. Andler (2017); Bettcher (2017a & 2017b); Jenkins (2016) & Kapusta (2016).

⁹ For more on deflactionalism in combination with gender see Barnes 2017.

¹⁰ As examples of such a purely negative (or as I would say deflationalistic) conceptions of “women” see e.g. Martin Alcoff (1988, 418-422) problematizing Kristeva (1981).

¹¹ Young (1994) continues by developing her own theory of gender as seriality, however, her theory also entails wrongful exclusion, and, hence, she fails to meet both target requirements at once (see Stone 2004, 144-146 & Mikkola 2016, 54­58).

¹² I am choosing this example, because it is quite a prominent example in feminist literature (see e.g. Asta 2011; Jenkins 2016 & Mikkola 2016). However, I do not want to indicate that there is no sexual violence against other genders as well; nor that there is no gender-related oppressing directed against the other genders.

¹³ For the actual statistics see e.g. Surveys of the White House Council on Women and Girls & Office of the Vice Presi­dent (2014) and the European Union Agency for Fundamental Rights (2015).

¹⁴ I will not go into further detail about the relation between social construction and realism. I regard it as a really fasci­nating topic, however, one for a separate paper. I merely assume here that social construction and realism are not nec­essarily mutually exclusive. For the basis of this assumption and further detailed discussions of gender realism, see Barnes (2017); Haslanger (2012a/2012c); Mikkola (2006/2017) & Witt (1995).

¹⁵ See e.g. Spelman (1988, 80-113) criticizing Chodorow (1995); Harris (1990, 590-607) criticizing MacKinnon (1989) and West (1987) or Stone (2004, 144-146) criticizing Young (1994) for this line of argument. CD: Essentialism is the cause of the wrongful exclusion. But this argument only applies to certain forms of essentialism, namely those that actually as­sume a “golden nugget of womaness”, and such approaches are then no longer only essentialistic; they also incorporate universalism, as well. In the following, we will first look at both positions (es- sentialism and universalism) separately, to then ascertain how Spelman equates the two in her line of argument. Essentialism is “the thesis that certain of an object's properties are necessary to it” (Witt 1995, 321). This thesis can be further differentiated into kind essentialism, when the object's property(ies) is(are) necessary to it as a member of a distinct category (or kind) (see Witt 2011, 13), and individual essentialism, when the object's property(ies) is(are) necessary to it as being that precise object (see ibid., 14). Even though both versions originate in the Aristotelian philosophy about species, they should be regarded as “conceptually independent of one another” (ibid., 16), since the presence of one does not guarantee the presence of the other and vice versa. Moreover, as individual essentialism plays a minor role within feminist philosophy, from now on I will use “essentialism” as a reference to kind essentialism only and not investigate individual

¹⁶ For more information on individual essentialism see Charlotte Witt's own gender theory (2011) and for a detailed discussion of it, see Mikkola (2012) as well as Stoljar (2018).

¹⁷ Indeed, there is a debate about whether the Lockean distinction between nominalism and realism is even useful in feminist philosophy. For more on this see Goldenberg (2007) & Stoljar (1995).

¹⁸ This list is for illustration only and has no deeper purpose.

¹⁹ Accordingly, “being perceived as a potential partner for heterosexual men” is not a monadic property, because it con­sists of a relation “being perceived as” and a second reference “a potential partner for heterosexual men”. Further­more, this property cannot be captured in the VxFx schema, which is due to the fact that this is a polyadic property. We will discuss those in section 4.1.

²⁰ Other examples can be found in Stoljar (2011); Witt (1995) & Zack (2005).

²¹ As mentioned above, we will discuss polyadic properties in section 4.1.

²² The term “social universal(ist)” is also used by Mikkola (2016) & Stoljar (2017). CED: Thus, social universalism is descriptively false. If this argument is justified, social universalism is descriptively false, because it is in discord­ance with the empirical facts. Yet, before we can reach this conclusion, we have to examine its premises in more detail, because both P2ED and P3ED are contested by advocates of social universal­ism. At first, P2ED is challenged by the assumption that these intra-categorical differences in social properties can be reduced to a common ground. For example, Charlotte Witt explains how “cosmet­ic surgery and foot-binding, while certainly very different procedures in very different cultures, nonetheless can both be classed together as practices in which women mutilate themselves in order to satisfy a culturally defined concept of beauty” (Witt 1995, 329). But, even if we agree with her that women's social properties might not differ as strongly as expected, this does not state that these social properties do not differ at all. To establish that, Witt would have to extend her claim, saying

²³ Besides, properties concerning menstruation are always problematic, simply because menstruation cannot be assigned to a specific gender.

²⁴ Stoljar herself argues for resemblance nominalism (see 1995 & 2011), a position that can be classified as relationalism rather that universalism, as in her account a woman has to resemble a woman paradigm to be categorized as a woman and this act of resemblance is a relation (see Mikkola 2016, 71-72). For more information on relationalism see section 4.

²⁵ Crenshaw also refers to them as axes, but I will keep just calling them categories, because I want to stick with one set of vocabulary.

²⁶ The wording is intentionally vague, as discriminatory forces can be expressed in various ways, such as actions, words, thoughts, etc.

²⁷ For an extensive discussion of the contribution intersectionality has made to feminist philosophy, see for instance Bailey (2009); Carastathis (2014) & Garry (2008).

²⁸ Both lists could be extended with several other social categories.

²⁹ This is precisely what I refer to as a constitutive monadic property.

³⁰ Even though, it is certainly problematic in itself.

³¹ As for example difference in payment, chances for higher education, political representation.

³² To get a more mature idea how such coalitions could work, see Stone (2004, 152-153).

³³ We will discuss Zack's own approach and her criticism of the intersectional analysis in more detail in section 4.1 & 4.2.

³⁴ We will continue this discussion of multi-realizable properties in section 4.2.

³⁵ Therefore, in the following discussion of the argument against FIP, I will rely on Mikkola's Butler interpretation and not look at it impartially.

³⁶ More on unproblematic normative implications in section 5.3.

³⁷ Mikkola's Butler interpretation ends at this point, so the move against universalism is my own application of the nor­mativity premise.

³⁸ For example Haslanger (2012b) and Bettcher (2017a) can be understood as defining gender relationally, both examples will be discussed in this section 4.2.

³⁹ Theoretically, polyadic properties could consist of even more than two predicates, but in this thesis I will only address polyadic properties that have exactly two predicates. For more information on polyadic properties see Allen https://iep.utm.edu/properties.

⁴⁰ Again I am only using the property “being perceived as a potential partner for heterosexual men” as an illustrative example. I am not saying that all women actually are perceived as a potential partner for heterosexual men.

⁴¹ Similar distinctions are also made, for example, by Barnes, Dembroff and Stoljar even though they do not speak of relationalism, but instead of externalist vs. internalists approaches (see Barnes 2018, 585 & Dembroff 2020, 4) and first-person performance vs. third person social status (see Stoljar 2017, 212).

⁴² I would go as far as saying that even the classic formats of predicate nominalism (Stoljar 1995 & 2011) and resem­blance nominalism (Stoljar 2011), can be incorporated into this classification. However, a detailed explanation would take us too far; see therefore Mikkola (2016, 71-72).

⁴³ Again, assignment can be treated as third-party-identification.

⁴⁴ See footnote 29.

⁴⁵ This is not a closed list.

⁴⁶ Of course, Bettcher does not call it relational herself, but since her categories are based on the polyadic property of identification, we can take her approach to be relational. CV: Thus, once gender is influenced by other social categories. So, the quintessence is that categorizing via sincere self-identification resonates well with the intersectional analysis, because the thing one identifies with “can mean very different things to dif­ferent people” (Barnes 2020, 6) and Bettcher's realtionalism can actually acknowledge that (see Bet- tcher 2017a, 402). For further illustration take, for example, Marie Curie and Frida Kahlo again. In section 3.2 we established that both differ in their womaness, because one is a Polish, able-bodied, scientist and the other is a Mexican, disable-bodied, artist. Now categorization via sincere self-identification can re­flect these differences, because the constitutive polyadic property “sincerely self-identifying as a woman*” can be realized differently. It simply means something different to Marie Curie, to identify as women, than it means to Frida Kahlo, since both understand the concept “women” differently, according to their social backgrounds, so both are indeed identifying with different things. The same argument applies to Julia, who as a young Afro-American woman identifies with yet another concep-

⁴⁷ If you know a counter example, please let me know I would be happy to read it.

⁴⁸ For an illustrative example of a criticism of such a hybrid account entailing wrongful inclusion, see Mikkola (2016, 65­70).

⁴⁹ As I see it, third-party-identification and social perception are mutually inclusive, so in what follows I assume that the aspect of social perception is defined by the polyadic property of third-party-identification as a woman. CIR: The presumption of gender as a monistic concept is causal for wrongful exclusion. Similar to the argument from diversity, P2IR postulates that monistic categorization attempts are generally incapable to reflect gender in its entirety; while P3IR stresses that this insufficient reflec­tion leads to wrongful exclusion, because it turns normative claims into solipsistic claims (section 3.3). So, in short, the irreducibility of gender stems from the fact that gender as a concept, consists of a plurality of aspects, with some of them being so different, that they cannot be reduced to a

⁵⁰ To return briefly to Spelman: Since monism is a part of universalism, which Spelman equates with essentialism, her criticism of essentialism could also be interpreted as opposing this monistic perspective, since in gender monism again “the many turn out to be one, and the one that they are is me” (Spelman 1988, 159).

⁵¹ For example, it might be worth discussing whether Robin Dembroff's distinction between critical gender kinds and non-critical gender kinds (2020) represents another fundamental aspect of gender.

⁵² Here I speak of marginalization rather than exclusion, because Haslanger herself says that there are other ways to be categorized as a woman, so it is not really exclusion. However, one gets the impression that women, who do not pos­sess the property of third-party-identification as a woman, are less important, which is why marginalization is still a problem.

⁵³ Both argue that Jenkin's intention of a trans-inclusive approach fails.

⁵⁴ In fact, I would argue that there is an important difference to be made here, but this would be a topic for another paper. Only this much: I think that it makes a difference whether one only identifies as a woman but keeps this to one­self or whether one identifies as a woman and expresses this outwardly. And I also think that because of this distinc­tion, it is also worthy of discussion, whether the aspect of self-identification should not be further subdivided into two separate aspects.

⁵⁵ Remember that this is only a hypothetical scenario, because currently it is not the case that all self-identifying women* are also perceived as women*, see for example Laura.

⁵⁶ I will generally refer to categories like self-identifying women* or socially-perceived women* as specific categories as they are reflecting a specific aspect of gender.

⁵⁷ Of course, there is no such category according to gender pluralism, but let us assume it for the sake of the argument.

⁵⁸ That is precisely why so many relational categorization attempts struggle with problematic normative implications as well; they reject the idea of universal categorization, but retain the monistic perspective.

⁵⁹ The discursive way of writing is meant to indicate that we are dealing only with self-identifying women* and not with women* as such.

⁶⁰ For example, one could discuss whether the specific category of socially-perceived women* should be divided into smaller categories, depending on the social context in which one is perceived as a woman (is this a queer-positive space or a standard workplace environment). (This suggestion is based on Bettcher 2017a/2017b.)

⁶¹ And maybe to other dimensions of gender as well, but this would be a question for a separate paper.

⁶² See e.g. the report of the 2015 U.S. Transgender Survey (James et al. 2016).

• Hochschule: Rheinische Friedrich-Wilhelms-Universität Bonn (Philosophie)
• Veranstaltung: Feministische Philosophie / Gender Theorie
• Note: 1,3
• Autor: Joline Kretschmer (Autor:in)
• Erscheinungsjahr: 2021
• Seiten: 64
• Katalognummer: V1363044
• ISBN (PDF): 9783346887634
• ISBN (Buch): 9783346887641
• Sprache: Englisch
• Schlagworte: Gender, Feministische Theorie, Theoretische Philosophie, Intersektionalität