The Metaphysics of Gender and Sexual Difference


“It is certainly true, as nominalists have been concerned to acknowledge, that judgements about kinds are determined in part by human interests, projects, and practices. But the possibility that human interests, projects, and practices sometimes develop as they do because the real (physical or social) world is as it is suggests that this sort of dependence is not by itself an argument against essentialism.”

— Susan Babbitt (1996, 146)

In the previous chapter, I argued for a concept of subjectivity as positionality in order to emphasize the role of variable social contexts in defining what it means to be a woman. What both the poststructuralists and cultural feminists seemed to me to be under-emphasizing was the very concrete and historical social situation in which “one is made a woman.” The cultural feminists had an essentialist conception of female identity that was based on an attachment to positive feminine attributes, attributes which they de-historicized and de-contextualized to such a degree that they could offer no account of how these attributes came into existence or under what conditions they might pass out of existence. On the other hand, the poststructuralists had an anti-essentialist conception that was motivated at least in part by an ahistorical approach to resistance, as if an anti-realism about sexed identity would finally avoid the age-old problem of co-optation where every construction of woman eventually gets turned against us. But they tried to achieve this impervious form of resistance through flying above material particularity into the netherworld of undifferentiated flux or negativity. Thus, I felt that the cultural feminists had de-historicized feminine attributes while the poststructuralists de-historicized resistance. Neither the essentialist nor the anti-essentialist position gave the temporal dimension, or the changeable nature of human characteristics and political contexts, enough significance in the development of their theories of women’s identity.

My idea was that a concept of identity as positionality offered the means to give a content to women’s identity without solidifying that content for all time, since positionality is a content that emerges in relational circumstances that are in constant change as we, and those around us, are engaged in a world that is itself in movement. Like the poststructuralists I agreed we should avoid essentialist definitions that define women in regard to intrinsic, internal characteristics, such as the disposition to nurture or an attentiveness to the corporeal and mundane details of life, as if these proclivities exist prior to our socialization. But like the cultural feminists, I did not want to be left with a mere politics of negation out of the fear of being essentialist, deterministic, or prescriptive. I believed we could do better than to say “I will make demands in the name of women even though I don’t accept the category of ‘women’.”

I believed then and I believe today that we can make many accurate claims about women, as women exist here and now in particular locations, and thus we can make demands that reflect women’s needs. The problem in reality is not an absence of content for the category “women,” but an overabundance and inconsistency of content, given the multiple situations in which women find ourselves in various cultures. The fact that what it means to be a woman varies does not entail that one can say nothing about women, only that one must refrain from universal pronouncements about the nature of women’s oppression or the content of our political goals. Feminist theory needs to locate and limit the scope of its presumed applicability. But women exist, with many common problems across the globe, as well as widely shared needs for freedom from physical violation, for education, for meaningful work, for our rightful share in our society’s political self-determination, and for fair remuneration for all the kinds of work we do whether or not it produces surplus value or shows up in the GDP. I stand by these arguments, and continue to believe that positionality is a helpful way to explain the contextual variability of women’s identity. But positionality is a concept with general relevance to social identity and agency, without special relevance to gender or sexual difference. In this chapter, my focus will be on women’s specifically gendered identity, and its basis in sexual difference.

Fifteen years after that chapter was first published as a journal essay, the debate over essentialism has now become, thankfully, passe. There are two main reasons for this change: First, because almost everyone agrees that the anti-essentialists won the debate, which does not mean that everyone has altered their views but that they must now show that their view is not, and was never, truly essentialist. Second, because the conventional terms of the debate over essentialism are thought by many to have been mistaken in their ahistorical, falsely homogenized account of what essentialism as a concept or a doctrine entails. If the essentialists were guilty of overly homogenizing the category of women, the anti-essentialists were guilty of overly homogenizing the category or idea of essentialism. In actual fact, essentialism can coexist with nominalism and even historicism, since it is a doctrine about essences but not a doctrine about the metaphysical grounds or stability of those essences (at least it has not historically been; see e.g. Babbitt 1996, Battersby 1998, De Lauretis 1990). This does not mean that the anti-essentialists are considered wrong in the substance of their claims, but only in their attribution of these claims to questions of essentialism.

There is also a third reason for the end of the debate, a reason that is given by mainly those drawing from a phenomenological tradition. This reason charges that the debate presented a false dilemma between attentiveness to embodiment and anti-essentialism. These feminists argue that attending to the body does not entail essentialism if we conceptualize it, in Beauvoirean terms, as a “body-in-situation,” that is, as a lived body, whose meanings are dynamic and contingent (see Bauer 2001, Moi 1999, Kruks 2001). Approaching the body in this way is highly plausible, but in some respects it can be categorized under the first group I mentioned above, that is, with those who are still trying to show that their position is not essentialist. Thus, however we deconstruct the terms of the debate, it seems as if the specter of essentialism continues to haunt feminist theory and continues to dictate the baseline conditions of theoretical acceptability.

There is no doubt that there are several ways in which the debate over essentialism was conceptually problematic, e.g. in projecting the Aristotelian version of essentialism as the only possible way to define an essence, but I would argue against those, such as Toril Moi (1999), who think that the debate was generated purely out of these kinds of conceptual confusions. In reality, the debate over essentialism crystalized and synthesized the two principal concerns of feminist theory over the last thirty years, which, despite a wide variety of topics, has ultimately focused on, first, how to address the significant differences among women which are mediated by race, ethnicity, class, nationality, religion, sexuality, age, and able-bodiedness without emptying out the category altogether, and second, how to resist the deterministic and naturalistic arguments that have been used to conceal the coercive conditions under which many women must live highly circumscribed lives as well as to conceal women’s labor as labor and thus justify its exploitation. We are still exploring how the heterogeneity of women impacts both feminist theory and feminist practice, but even while we pursue this question feminists are united in opposing the narrow determinisms by which it is considered legitimate the world over to reject individual women’s aspirations, even the aspiration for a life free of coercive sex or compulsory motherhood. Thus, feminist engagement with the question of essentialism was primarily motivated by these concerns over difference and determinism. Since essentialism, as it was defined, conflicted with difference and invited determinism, everyone agreed it had to be rejected (and those who wanted to retain it had to show how it could accept difference and thwart determinism).

However, I want to argue here that the philosophical core of the debate over essentialism was a debate over the metaphysics of gender, although this was not always made clear (certainly not by me), and despite the fact that the philosophical history of the concept of essences post-Aristotle indicates that essentialism entails no given metaphysics. The danger of homogeneity that feminists were concerned about does not in reality come from essentialism as a concept but rather from a metaphysical account of sexual difference or, in other words, from the claim that the category of sexual difference has a metaphysical basis. It is out of their concerns about difference and determinism that most feminists have fled the very idea of a metaphysics of gender or a politics based on metaphysical claims of sexual difference. Even the term “metaphysics” has become a pejorative, as when Kristeva famously claimed that “the very dichotomy man/woman as an opposition between two rival entities may be understood as belonging to metaphysics.”(Kristeva 1986, 209; emphasis in original) That is, as wrong.

In fleeing from determinism, feminists have fled not only from essentialism and metaphysics but also from realism, naturalism, objectivism, and even the capacity to make truth claims (which I would define as evaluative distinctions based on epistemic and not merely strategic considerations). In this paper I want to follow Susan Babbitt, Sally Haslanger and others to ask whether giving metaphysical content to sexed identity is necessarily determinist, and whether in fact an objectivist (post-positivist) account of sexed identity is philosophically sound. An objectivist account would be one that takes the categorization of human beings by sex as having good metaphysical grounds, irreducible to ideological grounds. I will explain this further in what follows. Existential phenomenological descriptions of women’s embodied situation, which start from the inner world of subjective experience rather than an objective category that is transcendent of the subject, are not inconsistent with an objectivist account of sexed identity, and would in fact be the best approach in my view toward developing a substantively descriptive account of women’s lives as sexed beings. But before we can produce such phenomenological ethnographies of women’s embodiment, we must delimit the category of “women.”

Haslanger has recently defended an objectivist account of gender and critiqued the anti-objectivist feminist position, particularly Judith Butler’s work, and so it will be useful to begin by reviewing Haslanger’s arguments. Though I agree with Haslanger in large measure, there is an important weakness in her analysis of the arguments against objectivism, a weakness which has to do with her interpretation of some of the social constructionist arguments which draw from Foucault. I will consider how detrimental this weakness is to her objectivist conclusion about sexual difference. It is important to consider carefully the strongest anti-objectivist and anti-metaphysical arguments, so before I discuss Haslanger’s critique of anti-objectivism I will give an overview of her strongest opponents which would include not only Butler but also Monique Wittig and Collette Guillamin.

The anti-objectivist position draws from four main claims: (1) a claim about the fluid variability of all categorizations, which means that categories are subject to (even reducible to) ideological manipulation, (2) a claim about the mediated nature of all descriptions, including the descriptions implicit in categorizations, (3) a claim about the inevitably prescriptive effects of description, and (4) a claim that objectivism about sexual difference will serve to reenforce compulsory heterosexuality. Against these four points, I will argue that there are persuasive grounds for an objective rather than totally fluid account of sex categories, that objectivity does not require an escape from mediation in human knowledge or the ability to have “out of theory experiences”(Babbitt 1996,142), and that the tendency for descriptive accounts to become prescriptive is a variable rather than uniform or absolute tendency and can be offset. I will argue that the objective basis of sex categories is in the differential relationship to reproductive capacity between men and women, but that a sexual categorization based on the biological division of reproductive labor does not establish a necessary link between reproduction beyond conception and heterosexuality.


The case against sex

Before post-structuralism the most influential anti-deterministic strategy in feminist theory had been developed by Gayle Rubin in 1975, when she argued, based on some extensive empirical cross-cultural research, in favor of making a distinction between sex and gender. Rubin used the term “sex” to refer only to body type and the term “gender” to refer to masculinity and femininity or the substantive aspects of our identity as men, women, and others. Thus Rubin claimed, with some definite echoes from Beauvoir, that “...we are not only oppressed as women, we are oppressed by having to be women, or men as the case may be.”(Rubin 1975, 204) Rubin conceded the naturalism of sex in order to denaturalize gender, thus allowing for the common sense idea that males and females have different bodies, but that these bodily differences are not sufficient to explain the elaborate cultural practices and beliefs surrounding gendered identities, which vary considerably across cultures but share economic motivations centering around, she argued, the ability to exchange women. Thus, Rubin’s distinction actually worked both to resist determinism and also to acknowledge significant differences among women. Sex is about biological reproduction and it is not entirely under our control, but gender is about power and the cultural formations that solidify it, and these vary widely. Contra Shulamith Firestone, who had argued in 1970 that our liberation would require a technological dismembering of sexual differences in reproduction, Rubin argued that gender had to go, or at least be radically transformed, but sex could stay. Rubin thus brilliantly solved our problem without sacrificing theoretical plausibility or committing us to a technologically based form of liberation that many women distrust. We were made free and allowed to be different.

However, Rubin’s view soon came under criticism for its too easy concession to a natural sexual division. Where does this idea of sex come from? Is sexed identity itself—i.e. male/female—so free from culture? All women do not have wombs, all men are not fertile, and thus an individual’s actual reproductive capacity does not determine their position within the division of sex. Moreover, Moira Gatens (1991), among others, argued that the sex/gender divide replicated too closely the androcentric oppositional binaries of mind/body and nature/culture, which are strongly associated with the conceptual justifications given for women’s oppression. The problem with these binary distinctions is not just their hierarchical character, the fact that one term is always superior to the other, or even their nefarious political uses. The main problem is the assumption that one can neatly distinguish the two terms, to be able to say “this is nature over here, that is culture over there,” as philosophers had once said, “this is mind, that is body.” Surely Hegel was correct to portray these distinctions as useful heuristic devices in certain circumstances but not as representations of reality strictly speaking, since in reality everything is quite mixed up together. The nature/culture distinction, which Rubin’s sex/gender distinction did seem to imply, has come under increasing criticism as incoherent, because, in an important sense, everything is natural—including what human beings do and make—and everything natural is actually dynamic and changing—including what human beings can do and can make.1

In the 1980's, on the strength of the alignment that Gatens noted between sex/gender and nature/culture, the sex/gender distinction was replaced with a monistic account in favor of, not sex for once, but gender. Gatens stated that she wanted to avoid entrenching “the historical construction of dualistically conceived sexual difference.”(1991, 139) But she, like other feminists, also wanted to avoid the monistic alternative that favored the masculine as the center or paradigm of universal subjectivity. Thus, if dualism were to be replaced by monism, it would have to be a contingency: that is, what all human beings, both those designated males and those designated females, share is that all that we are, all the way down, is contingent. As a result, the most influential position on the metaphysics of sexual difference and gender became a great refusal not only of the metaphysical basis for gender but also of a metaphysical basis of sexed identity itself. This was the position developed in Judith Butler’s Gender Trouble published in 1989, in Monique Wittig’s 1979 essay “One is not Born a Woman,” and in several essays by Collette Guillamin from the late 70's and throughout the 80's. Each claimed against Rubin that, rather than gender being the product of sex, sex, or sexually differentiated classification, is produced by gender, which Wittig and Guillamin theorized mostly in relation to labor and political economy and Butler theorized primarily as the effect of a “hegemonic cultural discourse.”

For Wittig, “The category of sex is the product of heterosexual society that turns half the population into sexual beings, for sex is a category which women cannot be outside of.”(1992, 7) Wittig’s aim was to contest the idea put forward by previous feminist theorists like Beauvoir and Firestone that “the basis of women’s oppression is biological as well as historical.”(1992, 10) Other feminists such as Adrienne Rich had also contested Beauvoir and Firestone’s view that female physiological experiences, such as menstruation and pregnancy, are necessarily constraining and oppressive, but Wittig’s point was different. Her aim was not to contest the negative valuation or characterization of female biology but to challenge the very categorization of human beings into groups based on biology. Thus she held that it is “dominance” that teaches us from all directions: —that there are before all thinking, all society, ‘sexes’ (two categories of individuals born) with a constitutive difference that has ontological consequences (the metaphysical approach), that there are before all thinking, all social order, ‘sexes’ with a ‘natural’ or ‘biological’ or ‘hormonal’ or ‘genetic’ difference that has sociological consequences (the scientific approach), — that there is before all thinking, all social order, a ‘natural division of labor in the family,’ a ‘division of labor [that] was originally nothing but the division of labor in the sexual act’ (the Marxist approach).... Belonging to the natural order, these relationships cannot be spoken of as social relationships. (Wittig 1992, 4-5 emphasis in original)

Wittig’s materialist feminism held that it is this fundamental ideological lie that works to justify the creation of women as a special class for exploitation. She claimed that “The making of women is like the making of eunuchs, the breeding of slaves, of animals.”(1992, 6) And the categorization by sex, presented as objective and thus prior to “all thinking,” is the necessary starting point for this process.

Collette Guillamin, in writings that in some cases predate both Wittig and Butler, made a similar anti-naturalistic argument against both race and sex categories, developing early arguments about the social construction of race that are today widely accepted. Her aim was to overturn “the basic postulate of all the present forms of thinking” about race and sex which “is the idea that a human group may be physically (or as common sense would put it, ‘objectively’) specific in itself, independently of its relationships or practices.”(1995, 85) She argued instead that these categories are founded on social relations and produced through arbitrary marks and enforced practices, marks that confer symbolic meanings on parts of the body, and practices that constitute race and sex identity as their effect, just as Pascal had argued that prayers produce faith and Butler argues that performances produce gender.

Guillamin’s version of a materialist feminism focused on the materiality of power relations rather than the materiality of embodiment. It is power relations and the “ideological effect: the idea of ‘nature’” that together “reduc[e] women to the state of material objects.”(1995, 179) Guillamin thus analogizes gender to class, as the product of a form of political economy, and invokes the ideas of fetishism and reification to explain the naturalization of gender relations. In this she follows Firestone, but where Firestone accepted Engel’s idea that the biological division of reproduction actually predated women’s oppression, and thus that sexed identity is a natural division rather than merely the ideological effect of private property, Guillamin’s relentless anti-naturalism leads her to hold that what she calls “sexage” is always and only a social category. Guillamin’s main concern, however, was not identity but the appropriation of labor, and, in my view, her best analyses are those that develop insightful characterizations of the way in which women’s unpaid labor is collectively appropriated, for example the ways that not only wives and mothers but unmarried and celibate women are roped into care-work. Her opposition to naturalism was primarily based in her view that it concealed and provided an alibi for this man-made appropriation of female labor.

Thus, the shared view of Guillamin, Wittig and Butler is that sex and gender are fully socially real but not objective or independent of human belief systems and thus not natural. It was Butler who developed the metaphysical implications of this view most fully. In her account, concepts of identity act as forced and enforceable constraints on what is at bottom fluid and inherently unstable, including desire and sexuality. “The challenge for rethinking gender categories outside of the metaphysics of substance,” she said, will be to overcome being by becoming and thus to conceptualize “gender as a doing.”(Butler 1990, 25) And she famously conceptualized this “doing” in performative terms, building from Erving Goffman: “There is no gender identity behind the expressions of gender; that identity is performatively constituted by the very ‘expressions’ that are said to be its results.” (Butler 1990, 25) The trope of performance was meant to eradicate metaphysics from the question of women’s identity once and for all: she says, “That the gendered body is performative suggests that it has no ontological status apart from the various acts which constitute its reality.”(Butler 1990, 136)

In actuality, the idea that gender is a doing rather than a being would preclude a metaphysics of substance, but not necessarily a process metaphysics, a la Whitehead, Spinoza, or Bergson, wherein reality is described as a primordially dynamic, ever-changing present. The metaphysical basis of gender would then not be in a being or substance ontology but in a process, and such a process could be enacted via performance. However, although processes are dynamic, they still have describable form and characteristics, with the fundamental categories changing from “things” to patterns or organizations of movement such as a dialectic. Given the fact that processes are not random flux but organized patterns, processes can then invite comparison with those same regulatory practices that Butler views as the techniques of oppression.2 Since Butler’s main target is not substances but regulative norms, her move away from being and toward becoming and process has created new dilemmas of determinism in Butler’s more recent work. In Bodies That Matter, her 1993 book with the subtitle On the Discursive Limits of “Sex”, she replaces “matter as being” with matter as “a process of materialization that stabilizes over time to produce the effect of boundary, fixity, and surface.”(1993, 9) In this way she seeks to explain the repetitive and recalcitrant features of current society: how certain bodies are made abject, certain identities are reified, and certain relations are stabilized. And the processes of subjectification she describes in The Psychic Life of Power, as I discussed in chapter three, are inescapably oppressive: it is our very entanglement with the process of interpellation that inscribes us within power and dooms us to desiring our own domination.3

In a sense, Butler’s version of process metaphysics continues to operate with a kind of nature/culture distinction, as many have noted, in that she makes matter the effect of power and always post-signification. The real player in the field is then ultimately discourse conceptualized as without material constraint, which is a Derridean as opposed to Foucaultian account of discourse (see e.g. Schrift 2001, Weir 1996, 112-134).4 This means that, in a sense, all is culture, in that there are no objective or human-independent forms, categories, or limits. Butler’s tendency to focus on texts and linguistic terms and her avoidance of directly addressing materiality largely follows from her view that any account of materiality that does not make discourse causally efficacious “in the final instance” would presume an escape from mediation: she asks, “How are we to find the body that preexists its cultural interpellation?”(Butler 1987, 129) Why then theorize bodies, she seems to be saying, if, when we refer to something in the world, we are operating on the basis of discursive boundaries that delimit objects? If there is no possibility of reference without relying on discursively constituted boundaries, then, she argues, there is no possibility of attributing objective status to any entity or type. It is our discourse, rather than the world, that constitutes objects and types.5

Another reason that Butler, Wittig and Guillamin were concerned to free us from a metaphysically grounded sex was because they saw this as a major player in the ideology of heterosexism, that human beings are naturally or properly exclusively heterosexual. A dualistic binary or oppositional system of categorization that divides us into male and female would seem to privilege heterosexual sex as the foundation of sexed identity. Butler rhetorically asks, “To what extent does the category of women achieve stability and coherence only in the context of the heterosexual matrix?”(Butler 1990, 5) Wittig argued that “The body’s sexual responsiveness is restricted through the institutionalization of binary sexual difference.”(Wittig 1992, 135) Thus, the normative weight of heterosexist categories collapses the distinction between description and prescription. Butler worried also that the descriptive binary categories imply a determinism: she says “the presumption of a binary gender system implicitly retains the belief in a mimetic relation of gender to sex whereby gender mirrors sex or is otherwise restricted by it.”(Butler 1990, 6) But ultimately the concern with prescriptive effects and with determinism was based on a prior claim that the description given of categories based on sex is simply wrong in so far as it is presented as objective: privileging reproductive organs over other differences is arbitrary and explainable only by ideological reasons. Butler argued that the heterogeneity within each sexed category cannot justify the homogeneity that the categories imply. And she shares Foucault’s view that “It is the way that anatomy is socially invested that defines gender identity and not the body itself.” (Benhabib and Cornell 1987, 14)

Butler, Wittig and Guillamin pointed to the absence of correlation between actual sexed characteristics and sexed identities, to the variable way in which physical attributes can be grouped and signified, and in Butler’s case to the necessity of linguistic mediation, and so concluded that sex is no less the product of culture than gender. The bottom line was that sexual categories are not based on objective features in the world. Variable discursive formations and variable cultural systems for organizing and exploiting labor can pick out, name, conceptualize, and define the boundaries of the features that are used to demarcate sexual, and not just gendered, identities, and thus sex identity is no more natural or objective than gender. It is the contingent practices of gender that creates sex, and not vice versa.

This view gained further plausibility from (as well as having a major influence on) the power of a new transgendered movement that has swept the United States and some other parts of the global North. There are two elements here, first, a new critique of surgery on transgendered babies and, second, a liberation movement of transgendered adults who have altered their assigned gender in surgical or non-surgical ways. Up to an estimated ten per cent of all babies born have ambiguous, indeterminate, or multiple primary and secondary sexual characteristics, and it has recently become more widely known that U.S. physicians routinely perform surgery on these “naturally” transgendered children in order to force them into our culture’s ideal of anatomical polarity. Such surgery is not always motivated by either a concern for ensuring reproductive capacity or sexual pleasure. Sometimes the children are surgically altered in such a way that they will conform to anatomical norms of appearance but have lost, because of the surgery, their capacity for orgasm or for other sexual pleasures. Critics argue against such surgery, at least until the children are old enough to decide for themselves what they would like to have done, if anything. Moreover, both surgical and non-surgical practices have been developed as voluntary procedures for those individuals who wish to transcend the physical dualisms of male/female or who wish to cross over the divide. Not only is it increasingly common to encounter male-to-females and female-to-males, whose sexed identity must be located somewhere beyond the dualism whether or not they can publically pass as unambiguous, but also to encounter those who have opted for secondary sexual characteristics of both male and female (at least in major cities). Sexed identity has undergone major cultural transformation, and there is no sign that this process will soon stop.

A second support for the denaturalizing of sex has come about through the new reproductive technologies. Artificial insemination, in vitro fertilization, embryo transfer, fertility enhancements, and surrogacy are creating new options in the biological process of reproduction, challenging the belief that there is a necessary connection between reproduction and heterosexuality, and producing a proliferation of legally recognized reproductive roles, from surrogate, to birth mother, to egg-donor, sperm-donor, and so on.6 This development seems capable of soon altering the biological division of reproduction in ways that Shulamith Firestone called for in 1970 as a necessity for women’s liberation. The biological revolution is not merely expanding the opportunities for sexed individuals to reproduce in new ways, it is transforming the ways in which we conceptualize and define the boundaries of biological reproduction, boundaries that have long served as the basis for sexed identities, at least the discursive basis if not the metaphysical basis. As these boundaries are reorganized, so too must our categorization of identities based on reproductive function.

Thus, spurred from both the transgendered movement and the development of reproductive technologies, the naturalism of sex looks to be withering away. Certainly, the categories of reproductive roles, and the categories of sexed identity are subject to variation by cultural practices and forms of political economy.

Anti-naturalism as mastery over nature

However, some feminists have opposed the anti-naturalist tendencies of these recent cultural developments which they see as an indication of the attempt to escape the realm of the immanent or the material(see e.g. Schott 2002). The peculiar dilemma feminism has which complicates its anti-deterministic project is that it has equal reason to pursue nature’s transcendence and nature’s acceptance. Can we be thoroughly anti-naturalist and anti-determinist in regard to gender without finding ourselves in collusion with the Enlightenment narrative of “mastery over nature”? Isn’t the anti-naturalism of some feminist theory in danger of imitating the project of mastery? Diana Fuss, for example, says that anti-essentialist feminists such as herself want a category of women which is a “linguistic rather than a natural kind” so that we can “hold onto the notion of women as a group without submitting to the idea that it is ‘nature’ which categorizes” us as such.(Fuss 1989, 5; see also Babbitt 1996, 141; second emphasis mine).7 Although she expresses this as submission to an idea, it implies a submission to nature or to any natural system that would limit the scope of our categorization. I believe this represents an aspect of much contemporary western feminist theory, that is, that, despite the extensive critiques of reified conceptions of “nature,” there persists a determined resistance to naturalized sex categories and an embrace of linguistic nominalism out of a general desire to avoid submitting to the dictates of nature.

It is without doubt that Butler, Wittig and Guillamin are right to point out that a naturalistic based determinism (as opposed to the earlier theistic based forms of determinism) has been the trump card used since the early modern period to secure women in our pumpkins, but patriarchal ideologies have appealed to nature in this way at the same time that they have maintained a barely concealed contempt and disgust for all things designated natural, which men alone could leave behind in a trail of mathematical formulae. Thus, in the west, women have been oppressed by a two prong attack: by being associated with nature (more closely than men) and by a simultaneous denigration of nature, putting it and us down with the animals and far away from the grace of God. Thinking purely strategically here, if women could manage to achieve transcendence from the realm of nature it wouldn’t matter how nature is regarded. This was the strategy taken by Shulamith Firestone and, arguably, Simone de Beauvoir. But of course, no matter how we come to define the sphere of “nature,” this strategy is unintelligible. Nature cannot be transcended in this way. Everything we make, transform, and build continues to change and decay, as we do ourselves, despite the dreams of genetic regeneration. The floor now under your feet, the metal and plastics which make up your computer or your car, and certainly the soft tissue surrounding your own skeleton, as well as the skeleton itself, are in constant, unstoppable processes of physical transformation, colloquially known as rot. Transcendence—insofar as this means a complete departure from natural processes — exists only in our heads. Given this, we should reconsider the devaluation of the natural from which the desire for transcendence (and mastery) gets its motivation.

Thus, some feminists are concerned that the currently dominant view of both sex and gender as the product of discursive construction is non-coincidentally aligned with the Enlightenment project of overcoming the materiality of our physical selves. Eco-feminists call for rejecting the project of transcendence even while we resist the deterministic justifications of patriarchal ideologies. As Chris Cuomo argues, eco-feminism offers “critiques of the ways in which the social and ecological worlds are gendered,” and articulate “alternative perspectives on the world.”(Cuomo 1998, 19) The project of mastery is an ethical relation to the non-human world, but it is built on a particular metaphysical account of that world and its relation to the human.

Feminist theory that has an antipathy to naturalism needs to develop a reflexive awareness of it’s motivations and fears. It may be assuming that all forms of naturalism imply a clear nature/culture division. But the very opposition between a naturalistic and a linguistic basis of sex categories presupposes such a clear division or separation. In contrast, Moi supports Terry Eagleton’s formulation that “...we are not ‘cultural’ rather than ‘natural’ creatures, but cultural beings by virtue of our nature...”(quoted in Moi 1999, 79) I interpret this as implying that our cultural practices and productions occur within a material world, which would eschew both a neat nature/culture divide and the commitment to transcend and master nature. I will return to this idea later in the chapter.

In sum, while I believe that the desire for a transcendent mastery is still manifest in some feminist treatments of sex categories, I also believe that at least some forms of naturalism continue to operate effectively as justificatory alibis for the oppression of women. But neither of these concerns can settle the matter of the metaphysics of sex and gender. We must ask whether this refusal of a metaphysics of sexual difference is persuasive on descriptive grounds and not merely as a political strategy. To what extent are the political concerns about determinism and difference, and the debate over whether we should refuse or embrace naturalism, acting as non-negotiable criteria for a feminist position on the metaphysics of sexual difference, and if so, does this make theoretical sense? If it is implausible to believe that materiality is always, or even, only, the effect of power, what are our real metaphysical options? In other words, what is the metaphysics of sexual difference?

Metaphysics, objectivity, and sex

The idea that there is a metaphysics of sex is associated mainly with the trend some call “difference feminism,” which is based on a post-Freudian approach to the psychic figuration of woman, such as developed most interestingly and comprehensively by Irigaray. Difference feminists postulate that male/female anatomical differences produce, or at least are correlated with, differences in philosophical, psychological, erotic, and linguistic orientations and effects. Their emphasis is on the uncovering and valuation of these differences rather than on the denaturalization of sexed categories. However, the relationship between “difference feminism” and metaphysics is contested: Irigaray, for example, may be drawing on the body and insisting on the real significance of morphological difference, but she has also forcefully critiqued the traditions, methods, and presuppositions of western metaphysics in Plato, Aristotle, Descartes, Locke, and Hegel. Even Nietzsche’s own anti-metaphysics has come under her critical analysis for its gendered symbolic organization and articulation of concepts. The implication of her critique is that metaphysics cannot be an objective description of what exists if it operates through a conceptual imagery that is variegated according to body type but presented as neutral. So, on the one hand, difference feminists are charged with naturalizing sexual difference as a metaphysical category, but, on the other hand, they are associated with the critique of metaphysics as a philosophical pursuit of objective description.

One way to resolve this paradox is to view the differences that difference feminists posit as other than objective or natural: that is, sexual difference might also be an imagined representation of the (physically) real, but a representation that is produced under given social conditions, such as the oppression of women, from which in turn alternative experiences and modes of conceptualization are developed. On this reading, to the extent that Irigaray is doing metaphysics, she is exploring the possible features of an alternative female symbolic and its ramifications for ethics and politics, rather than addressing the question of how to characterize the metaphysical basis for the categories of sexual difference. In other words, she is experimenting with alternative symbolic and conceptual imagery as a way to free us from the stranglehold of masculinist concepts that parade as universal. As interesting as this project is, it does not shed light on the question of whether there is a metaphysical basis for sexual difference, except perhaps as a negative warning against certain kinds of gendered concepts or metaphors that we might use in articulating a metaphysics.

Recent work in the 1990's and beyond, by Sally Haslanger, Susan Babbitt, and others, has reintroduced a different discussion about the metaphysics of sexual difference, not focused on the substance of the difference as much as on the possibility of an objective or natural basis for sexed difference. I want now to consider and assess some of the strongest of these arguments. For Haslanger and Babbitt, working more within an analytic tradition of philosophy, the key idea here is that gender is an objective type, while continental feminist philosophers like Moi and Bauer use existential phenomenology to reassert a category of women that can be rendered with descriptive accuracy as well as positive political effect.

I see these two trends as united in a commitment to a kind of materialist analysis, which I would broadly define as an analysis that maintains the central importance of the material reality of the sexed body. Moi offers the most general explanation for why she believes it is time for a renewed insistence on the body.

The principal thesis of Moi’s book, What is a Woman?, is that the feminist error, starting as far back as Rubin’s work, of separating sex from gender produced a kind of idealist trend in feminist theory, or a neglect of female embodiment. The idealist tendency and over-emphasis on discourse results from two mistaken assumptions, first, that sex exhausts the facticity of the body, and second, that sex has no determinate effect on the practices of gender. The body, then, is reified in a certain sense as well as eliminated from any analysis of gender, a result which is counterintuitive given gender’s consistent manifestation through the body. The resultant theoretical problem of relating sex to gender is analogous to Descartes’ problem of explaining how the mind, which he thought of as a thing without extension, could affect the body, a thing with extension, once he had separated the mind and body in a qualitatively absolute way. The sex/gender division, Moi claims, tends similarly to effect an artificial separation and to reify sex outside of history and social situation, leaving “a gap where the historical and socialized body should be...”(1999, 30) Even with the ongoing danger of determinism, we do not need to build an impenetrable wall between embodied sex and cultural practices of gender, she argues. Like Bauer, she argues for combining gender and the sexed body through using existential categories. Thus, Moi suggests that the argument that gender constructs sex is unnecessary to thwart determinism once we understand the body as historical and social, that it therefore solves only a specious problem, and that it implausibly holds matter accountable to discourse without reciprocity. Moi argues that we can maintain both that the body is not reducible to sexual difference and that the body is subject to natural law as well as human-made meaning systems which are combined in such a way that they cannot be neatly disentangled.(1999, 68-69) There are other interesting arguments against some current feminist doxa in Moi’s book, contesting the claim that all concepts involve exclusion, for example, and to thwart determinism we need only show that biology does not determine social norms, rather than trying to make biology irrelevant to sexed embodiment. Moi claims that to avoid biological determinism without leaving the body behind all we need is to return to Beauvoir’s concept of the body-in-situation, because the latter approach starts with a dialectical account of facticity and transcendence.

But before we can articulate such a dialectical account, I believe we must construe the nature of our facticity itself. What can we say about the body’s relationship to, implication in, even causality of, the categories of sexual difference? Moi’s account of the body-in-situation leaves these questions aside, perhaps because she thinks such questions imply the ability to make a neat divide between what is natural and what is interpreted or mediated, a divide that exisential phenomenology rejects. Haslanger, however, offers an argument about why the repudiation of metaphysics does not require positing an escape from mediation, which I will turn to in a moment. But to begin to theorize sex/gender metaphysically, it will be useful to contrast it with race.

Is sex like race?

Considering for a moment the comparison between sexual difference and race can shed some light here. Both are “visible identities” that are thought to be marked on the body, even that subjectivity emanates from the specificities of bodily difference. The ideologies of sexism and racism is predicated on a claim of causality between physical features and intellectual, moral, and emotional attributes. But there is an important difference between the two. Racism must convince us that biologically insignificant physical attributes such as skin color, the shape of the nose or eyes, or hair type are actually very significant and the signs of fundamental differences in human capacity. Thus Kant can claim with perfect certainty that “...this fellow was quite black from head to foot, a clear proof that what he said was stupid.”(Eze 1997, 57) One might speculate that the recent search for racial genomic differences was motivated by the need to find something more plausible than surface features on which to hang racist claims. But this search has turned out to prove the opposite: that there is no non-trivial genomic difference between socially recognized racial groups.

Thus racism has now turned increasingly to cultural differences to justify its hierarchies, holding that the fundamental differences are located in cultural traditions and practices—the culture of poverty, for example, or the irrationalism of strong religious commitment — rather than what are relatively insignificant human features. On this view, physical features become the sign, not of biological difference, but of cultural difference. However, this change in referent changes the ability of the signs to “work” in reliably racist ways, that is, to circumscribe the potential of individuals. For racism to make use of culture in this way it must impute a stable bounded essence to an entity (culture) which is unstable, unbounded, and without essence. Cultures are simply not immutable. Linking skin color not to biology but to culture, then, cannot justify differential treatment, the withholding of educational resources, or apartheid. Cultural difference can only support racism if it is seen as the effect of a more intractable biological difference, and thus it can only work on the basis of an assumption which is empirically insupportable. Nonetheless, culturalism is operating effectively under the current regime of “official” anti-racism to conceal the racism that motivates the claim of cultural intractability.

Sexism has more to work with, one might say. The role one plays in the biological division of reproduction, the capacity to sustain an infant entirely on the production of one’s own body, to give birth, to nurse, are much more significant attributes. There is thus a qualitative difference in the significance of these two sets of differences: skin color, hair type, etc. versus role in biological reproduction. Of course, sexism would make female-specific attributes exhaustively and implausibly significant in determining women’s lives, a claim which is losing credibility. Given that female children inherit fully half of their genes from their fathers, as do male children from their mothers, we cannot be that different. There is just a single chromosomal variable by which one’s role in reproduction is determined, which is an implausible determinant over the whole range of human functional capacities as claimed by sexist ideology. Even Plato saw as much; he says “...if it appears that [the male and the female sex] differ only in just this respect that the female bears and the male begets, we shall say that no proof has yet been produced that the woman differs from the man for our purposes, but we shall continue to think that our guardians and their wives ought to follow the same pursuits.”(Republic 454 d-e Book V)

Yet, despite the outlandishness of the claims that have been made for its significance, it remains true that the variable of reproductive role provides a material infrastructure for sexual difference that is qualitatively different from the surface differences of racial categories.8 It would be ridiculous to pursue a research project to study the significance of hair texture on the development of cultural systems; in regard to race it makes much more sense to look at the global political economy and the history of colonialism for an understanding of why and how skin shades gained such ontological significance in recent centuries. Furthermore, empirical evidence suggests that fear and distrust does not automatically arise from perceptual visible difference; something else must instill hostility because looks alone are not sufficient. As I will argue in chapter 7, this does not mitigate against the reality of race, or suggest that racial identity is a chimera, but it is to say that the origin of racialized differences and racial categories is less a metaphysical than a political story.

The physical foundation of sexual difference is another matter altogether. Gender identities in some variation seem to be, unlike race, historically ubiquitous, hardly recent, and based in a set of biological features with more morphological substance. Context and history are necessary to explain the political significance of racialized features; they are not necessary to explain the significance of differential roles in reproduction. The foolhardiness of formulating legal protections based on a presumed sameness between men and women has been well established by now as pregnancy must be labeled disability and maternity leave reforms are stymied by the insistence that they must be absolutely equal to paternity leave. And there is no place for a provision for breast-feeding in a legal discourse predicated on individuals undifferentiated by sex. Whether women want to “overcome” such differences by technological means is a different kind of question altogether from whether we want to overcome the historical and social conditions that make affirmative action necessary to redress racial inequality. One could make an overall point here about the disanalogy between racial/ethnic/cultural identities, on the one hand, and identities such as age, disability, and sex on the other. All are generally visible identities, naturalized as marked on the body without mediation. But the markings that signify age, disability, and sex are qualitatively different in significance from those signifying race, ethnicity, and culture.9 This is not an argument about the virulence or priority of various forms or targets of oppression. It is simply an argument about the quality of the physical basis for sex categories vis-a-vis race categories.

Denise Riley has become famous for her book, Am I that Name? Feminism and the Category of ‘Women’ in History which argues that the category of woman is an indeterminate, unstable category for semantic and historical as well as political reasons. On her view, indeterminacy does not mean that we can say nothing about women or in their name, but that we must acknowledge that the only anchor for the concept of women is in “discursive historical formation.”(Riley 1988, 5) Yet in earlier work, Riley was one of the first to articulate the mistake of repudiating any and all biological aspects of sexual difference. She argued that the problem is not biology but biological determinism, that is, the attempt to reduce “everything to the workings of a changeless biology.”(Riley 1983, 2) And she pointed out that the “usual corrective to biologism” invoked by feminists is a cultural construction thesis so strong that it “ignores the fact that there really is a biology, which must be conceived more clearly.”(Riley 1983, 6, 2) Thus, the cultural construction thesis implausibly “substitutes an unbounded sphere of social determination for that of biological determination.”(Riley 1983, 3)

Again in this earlier work, Riley articulates the feminist project as follows:

The tactical problem is in naming and specifying sexual difference where it has been ignored or misread; but without doing so in a way which guarantees it an eternal life of its if the chance of one’s gendered conception mercilessly guaranteed every subsequent facet of one’s existence at all moments.(Riley 1983, 4)

Perhaps twenty years later it is easier to see that biological facts are neither eternal nor all-determining, and thus to avoid the preemptive conclusion that Riley became influential for later that we must base sexual difference only on history, not on materiality, if we seek an escape from biological determinism. It should not be so difficult, after all, to develop an account of sexual difference that incorporates biology without having biology guarantee “every subsequent facet of one’s existence at all moments.” The question we must ask is whether this biological or material infrastructure for sexual difference yields a metaphysics of sexed identity.

Haslanger’s Objectivism

Sally Haslanger has recently attempted to clarify the issues of debate over the metaphysics of gender, or sexed, identity, and has raised the specter of objectivism in relation to sexual categories, though her defense of objectivism is given abstractly, that is, without specifying what precisely is objective about sexual difference. Haslanger is a rare bridge figure, capable of drawing out the full implications of contemporary analytic approaches and relating these to current feminist and anti-racist discussions occurring at the margins of philosophy. In her essay “Feminism and Metaphysics: Negotiating the Natural,” Haslanger begins with a consideration of the tension between feminism and metaphysics, given that most feminists reject metaphysics and most metaphysicians ignore feminist philosophy. She is critical of both. She argues that feminists who reject metaphysics are operating with an outdated notion of the field, in which metaphysics is still trying to access an unmediated Real and produce irrefutable truths. This characterization is not true of contemporary post-Quinean metaphysics, she holds, which understands itself as operating within a theory-laden field and with a pre-existing set of doxastic commitments that are organized as more of a web than a foundationalist pyramid. But she also argues that post-Quinean metaphysics has been inattentive to the political influences affecting its priorities, its framing questions, and, consequently, its determination of what is central versus peripheral in the organization of the web. Thus, she tries to develop a metaphysical approach to feminist questions that will include a political reflexivity.

Haslanger’s main thesis is to argue that the feminist tendency toward skeptical and nominalist views about the objective basis of gender distinctions is unwarranted both philosophically and politically. That is, feminists have not made sufficient arguments to show that there is no objective basis for gender distinctions, and they have incorrectly assumed that if there were such an objective basis, this could only provide support for patriarchy. Thus, her main argument is in the form of a negative: that the anti-realism about gender is unwarranted. I find most of her reasoning compelling, but in one crucial part of her critique of the skeptical and nominalist position she ignores an important argument. Thus, my concern here will be to see what difference this makes to her realist conclusions.



Haslanger begins by reviewing the feminist critique of metaphysics as androcentric on the grounds that it tends “to draw uncritically on experiences and patterns of thought that are characteristically male or masculine...” Such arguments, and demonstrations of androcentrism, have been made by Iris Young and Merrill and Jaakko Hintikka, among others. But this raises the immediate question of what a better alternative to androcentrism might be, and it turns out that this is not so apparent. Obviously, to theorize from a gynocentric perspective is no improvement over androcentrism if one is trying to make general claims. But the even more difficult issue is to decide what would count as a gynocentric point of view. The basis for most claims about gender has been empirical research on sex differences, research which is now widely recognized as notoriously biased (see Fausto-Sterling). Nor can we simply wait for better research to answer our questions about gender, because, as the critics of sex difference research have pointed out, the problem with the research is in the questions it poses as well as in the answers it gives. Why focus on measuring the variations between genders when the variations within each gender is often far larger, and other differentials, such as culture or age, are more relevant? Moreover, in defining gender through the measurement of gender differences we may in fact be setting up a new regulatory norm or stereotype for “how women (and men) should be” that will exclude or marginalize some or many or even most women. Thus, how can a concept of gynocentrism address the heterogeneity among women and avoid exaggerating the differences with men?

Haslanger argues that this set of problems does not mean, however, that gynocentric approaches are irretrievable. What one must do is retreat from either temporal or spatial universalizations about gender difference and instead make context-based claims. Thus we can make reasonable characterizations of gynocentric experiences within given contexts, as long as we acknowledge that there will still be variety even within a delimited context. This strikes me as an entirely reasonable conclusion. And because the claims are contextual, there is no claim being made about the absolute nature of a gynocentric point of view, and thus no regulatory norm exerting a command into the unending future.

Haslanger points out that even such a modified approach — one that would presume nothing more than the ability to make contextually limited generalizations about women — is still problematic for many feminist theorists, despite the fact that many feminist social scientists pursue just such generalizations. But for many theorists the contextual approach is not really telling us anything about gender, even in context, but only about the discourse of gender in that context.

It is on this point that Haslanger begins her real case against the critics of metaphysics, and to develop her case she first provides an updated characterization of post-Quinean metaphysics. Quine portrayed the empirical claims of the sciences as having a web-like structure rather than a foundationalist pyramidal structure, and argued that any part of the web—even basic empirical observation reports — will be revisable if we are prepared to alter the other parts of the web that are structurally dependent on the part we want to eliminate. And we may well be willing to make those structural adjustments given a particular anomaly we wish to resolve or a particular theoretical project we wish to complete, but these motivations are entirely contingent and variable. Thus, Quine’s approach assumes neither absolute starting points nor access to an unmediated Real, but is non-foundationalist, with a holistic approach to justification, and adopts a fallibilist view of its best claims. Theorizing aims not at purity but at increasing the consistency of our beliefs. Haslanger calls this approach an “aporematic” metaphysics, and explains that it might reasonably be considered immanent metaphysics: the questions, the puzzles, and the proposed answers arise within our thinking in response to current theoretical and practical demands. (Haslanger 2000, 114)

Again, Haslanger is still critical of much of the work in metaphysics that uses this approach for its lack of reflexivity about its priorities and assumptions, but the point is that its anti-foundationalism provides a greater potential openness to seeing the ways in which political realities can affect the discursive context in which questions are formulated and posed as well as the ways in which possible answers are developed and compared.10

According to Haslanger, then, the metaphysical question of gender is whether gender is a natural kind, which is a group that shares a common essence, or whether it is an objective type, which is a unity without an underlying essence, or whether it is neither of these. (I think Haslanger means to be referring to sexed identity here, not to gender in Rubin’s sense, but I will use Haslanger’s terminology while discussing her argument). One could hold a realist view about gender being either a kind or a type, or one could be a skeptic, holding that we simply cannot know, or a nominalist, holding that the basis of either kinds or types is entirely non-objective.

To say that a type is objective is to say that there is some non-random or non-arbitrary basis for its unity. The basis for some unities is weak, such as the unity of things on my desk which at the moment include paper, pens, matches, stones, garbage, tiles, plants, pictures, tissues, keys, candy, cough drops, chewing gum, buttons, and a cat. There is no way to incorporate this debris into one category except in so far as all these things share a temporary location. But some unities clearly have a stronger basis than this, such as the unity of red things, or mammals, or of things which are carbon based. What makes a type objective is that what brings the things in it together into a unity is independent of us. I will address the issue of what “independence” can mean more thoroughly later on, but Haslanger is not putting forward a chimerical language of pure access to the noumena in claiming that the basis of some unities are independent of human practices. She is making a comparative judgement between what are obviously all linguistically conceptualized entities. One needn’t be a positivist, who believes in pure uninterpreted bits of data, to be committed to a theory which holds that the basis for some categories are independent of human beings, such as whether something is carbon based, while other categories are not independent of human beings, such as social structures or literary conventions or even the degree of gender dimorphism in a given time and place.

Gender is, of course, very much a social kind of unity. It has been imbued with cultural values and meanings and it is usually presented as a category of two after which exceptions are forcibly, even surgically, altered to fit. But Haslanger does not think these obvious facts are sufficient to dismiss the possibility that gender is an objective type. What I assume she is thinking here is that the objective type which is gender might be mistakenly characterized, made overly inclusive, and have longstanding crazy cultural associations, and yet the basis of the unity itself might be objective.

To show that this hypothesis—that gender is an objective type—should not be ruled out, Haslanger considers some of the arguments against the objectivity of gender from Wittig and Butler. She discusses their argument that the fundamental importance attached to whether one has a penis or vagina comes after, not before, social conventions, in this case, compulsory heterosexuality. Butler’s variant of this argument relies heavily, and Haslanger thinks too heavily, on the idea that all of our access to reality is mediated. From the latter claim Haslanger argues that Butler fallaciously comes to the conclusion that gender is not objective. Here is how the argument goes.

Butler argues that when we refer to something in the world we are operating on the basis of discursive boundaries that delimit objects. Thus there is no possibility of reference without relying on discursively constituted boundaries, and thus no possibility of attributing objective status to any entity or type. Haslanger is prepared to agree with Butler’s claim that it is our discourse, rather than the world, that constitutes objects and types, if this is understood to be a claim about our language and about our knowledge, but she thinks Butler believes it is also a metaphysical claim about the absence of an independent basis for categories used in language.

To say that it is a claim about language and knowledge is to say that it is a claim about things qua things we refer to. But to say that it is a metaphysical claim is to say that it is a claim about things. Here’s another way to put the distinction: in the first case, Butler is arguing that gender as a concept is discursively constructed, and in the second case she is arguing that gender is discursively constructed. Haslanger thinks Butler has no grounds for the latter claim simply on the basis of the idea that all of our relations with the world are discursively mediated. The ubiquity of mediation itself does not entail that nothing is independent of human beings, nor that our knowledge is blocked by mediation, and here she uses the example of the phone system—an intermediary which improves rather than blocks access to things beyond our reach. Of course, what the phone brings us into contact with is not, under any circumstances, beyond our reach; we could conceivably travel to the person we are trying to communicate with and then communicate directly. So one might argue that, for this reason, the phone system is not analogous to discourse. But all Haslanger is trying to show here is that mediation in and of itself is not sufficient to justify skepticism. Even if we were not able to travel to the person we are speaking to, we might yet be able to communicate very effectively with them using the phone. In this case, mediation does not block knowledge, but in fact aids it.

Moreover, the dismissal of the possibility of knowledge about the world on the grounds that we are bounded by language presupposes very modernist pre-Hegelian bifurcations between phenomenal knowing and a noumenal world. According to Haslanger’s portrayal of much contemporary analytic metaphysics, a pure transparency between belief and world is no longer considered necessary to claim knowledge: we can settle with oblique relations of veridicality, but why deny our ability to reliably claim that the set of things on my desk has a weaker unity than the set of mammals, and that this is a fact independent of human categorical systems? Haslanger says,

There is a temptation to think that if we cannot “get outside” of ourselves to test our beliefs against reality, then there’s nothing further we can do epistemically to regulate belief; we’re left with only political negotiation. But there are other epistemic considerations that can be brought to bear on belief, and provide grounds for claims to truth, e.g., coherence, evidential support, fruitfulness, etc. Oddly, many feminists feel pressed to skepticism about an independent reality because they implicitly endorse a traditional conception that requires certainty or direct access to reality in order to have knowledge of it, while at the same time they often find the traditional conception of knowledge problematic. (Haslanger 2000, 122)

This mistake about the limits of knowledge is related to the mistake some feminists make about analytic metaphysics: assuming that the failure of positivism leads to epistemological skepticism.

Haslanger’s arguments in regard to the implications that follow from the ubiquity of mediation seem to me to be right. This is an important point which, although it has been repeatedly made over the past ten or more years, seems not to have been absorbed into poststructuralist feminist philosophy. Yes, we use concepts to know the world; no, that does not mean that we cannot say anything about the world but only about other concepts.

But where I differ with Haslanger is as follows. She assumes that once she has shown the problem with the argument about mediation, she can then show that categories of gender are not just about the discourse of gender but about gender itself. However, the ubiquity of mediation is not the sole reason Butler gives for the discursive constitution of sex. Butler’s argument also invokes the Foucauldian idea that discourses have identity-altering, materialistic effects. That is, discourses do not merely categorize and rearrange what is in the world but, in some cases at least, create things that didn’t exist previously. I am sure Haslanger would agree that discourse does create some things, like, for example, heterosexuality as a social identity, but the question is whether gender is also one of these things. From Butler’s point of view, gender is performatively enacted on the basis of discursively constituted regulatory norms, but this means that it literally comes into existence as a lived experience and visible phenomena through discourse. That is, gender identity comes into existence, which is not reducible to the possession of a type of genitals. As Pascal said, one kneels and prays and belief comes after. For Butler there is no objective basis of gender, then, in the sense of a basis that is completely independent of human practices.

Are Butler and Haslanger perhaps talking about different things here? One might think that Butler is talking about a much more robust sense of gender whereas Haslanger is talking about a much more minimal sense. And Haslanger allows that the importance of the objective basis of gender is contestable: we can allow that there are prediscursive, objective bases for some of the properties used to demarcate gender, even while contesting whether it is these properties which are truly fundamental to gender in the robust sense, and whether it is political rather than metaphysical criteria in operation here. She says,

The realist can agree with the non-realist that our classification schemes are often motivated by interest-laden concerns, and that we need to look beyond questions of what’s ontologically fundamental to determine how to structure our political lives; these issues are not ones that divide the two sides of the debate. The realists begin to diverge from the non-realists, however, when they claim that in some cases it is important to know what sets are fundamental, e.g., what properties are causally significant, in order to effectively interact with or understand the world. (Haslanger 2000, 123)

I agree with Haslanger on this point, which causes me to part company with Butler who cannot allow such an “objective” criteria of significance in regard to any category. But what Haslanger does not consider, at least in this essay, is the possibility that genuinely causally significant properties can be discursively produced: that things, and not just things qua things referred to, can be discursively produced. Discourse is not simply about the way in which we interpret and organize the world: it can also produce new things in the world, and not just things like “football” or “money” but things with more of a material presence such as inner subjective experiences of inferiority or a significant gender dimorphism or racial purity or the existence of two and only two sexes. Here I agree with Butler that, as Foucault taught us, we need to develop a hermeneutics of suspicion in regard to what looks natural.

If my argument is right, then Haslanger is wrong to say that Butler reaches her conclusions solely through the ubiquity of mediation argument. If one wants to hold onto the idea that gender is an objective type, then, one needs more than a negative argument against the ubiquity of mediation thesis—one needs to address the claim that discourse creates gender which creates sex, and address the issues of practices and not just the issue of naming. One needs to show what objective, fundamental, human independent basis there is for the category of gender or sex, and to show that this basis is not the product of discursive effect. This requires arguments that will go beyond Haslanger. I think we have a good candidate for such a fundamental, human independent basis in the division of labor in biological reproduction.


Sexed identity

Simone de Beauvoir, often credited with inaugurating the idea that “woman” is a social construct, was herself clear about the grounds of the sex distinction itself. She said, “...woman cannot dream of exterminating the males. The bond that unites her to her oppressors is not comparable to any other. The division of the sexes is a biological fact, not an event in human history.”(1989, xxxi) The meanings we confer on and the implications we draw from that “biological fact” are of course subject to free human interpretation, according to Beauvoir, but the division itself is not. But how should we delimit the fact itself?

Consider the following as a possible objective basis for the category of sexed identity:

Women and men are differentiated by virtue of their different relationship of possibility to biological reproduction, where biological reproduction refers here to conceiving, giving birth, and breastfeeding, involving one’s own body.

By “possibility” here I mean something more than mere logical possibility, something closer to Aristotle’s idea of concrete potentiality, in order to capture the idea that females are expected to have, or have had, the ability to give birth and lactate, whereas males are not. This differential relationship of possibility to biological reproduction remains in place even for women who are post-hysterectomy, women with no desire or intention to reproduce, women who are not fertile, and for both pre-pubescent girls and post-menopausal women. Those classified as women will have a different set of practices, expectations, and feelings in regard to reproduction, no matter how actual their relationship of possibility is to it. That is, even infertile, pre-pubescent, or post-menopausal women, and women who have no intention to reproduce still have a different relationship to biological reproduction than males do. This differential relationship can be the basis of a variety of social segregations, it can engender the development of differential forms of embodiment experienced throughout life, and it can generate a wide variety of affective responses, from pride, delight, shame, guilt, regret, or great relief from having successfully avoided reproduction. But these various accompaniments to a relation of possibility with biological reproduction are themselves culture bound and not simply objective. It is implausible to suggest a one-way linear, causal story from the objective fact of a differential relationship to biological reproduction to the richness of cultural genders; rather I would develop a holistic analysis in which this differential relation of possibility is one objective factor always at play, but one that can be moved about the web, from the center to the periphery, made more or less determinate over the construction of gender depending on cultural context.

An immediate question such a definition raises is its relationship to heterosexism and compulsory heterosexuality. This issue is connected with both the problem of determinism and the problem of difference that were raised at the start of this chapter as key motivating concerns behind the essentialism debate.

However, descriptive claims are not disproven by their pernicious prescriptive effects. I might correctly say to a child, “you have been getting consistently bad reports from your teachers,” a descriptive claim, which may then have the effect of making the child think that he or she is simply “bad” even if I never say such a thing and don’t intend to communicate that idea to the child. Yet current child therapeutic advice persuasively counsels adults to avoid such negative statements, even when they are descriptively accurate about past behavior and not stated as essential truths about the child, because the descriptive claims can have prescriptive effects. Even if I take such advice, however, I may still want to know the accuracy of the descriptive claim, and may want to look into the prejudices of the teachers, whether the reports were really consistently negative or only occasionally negative, and so on. In order to learn how to avoid prescribing the very behavior I want to sanction, I may well need to know as much as I can about the actual events. Prescriptive effects thus do not in all cases obviate the necessity of ascertaining the best descriptions.

Moreover, to assume that a descriptive claim will have a prescriptive effect is to already assume that what is being described is dynamic and changeable. It must be susceptible to suggestion, in other words, and not have a fixed character in regard to the features being described. Another assumption that might be operating here is the assumption that the description’s sole purpose is to prescribe a norm. But it is implausible to claim that any and every descriptive claim is simply a means to enforce norms, so one must look with more care at the kind of claim being made. One might then explore how to avoid pernicious prescriptive effects, and what degree of actual instability or susceptibility to suggestion there is in what is being described.

The claim that sexed identities are objective types based on a biological division of labor in human reproduction does not prescribe compulsory heterosexuality in the sense of mandating heterosexual coupling as the necessary means for the reproduction of children. It may be used to make such claims, but these would not follow logically or empirically. Conception does require heterosexual coupling to the extent that biological material from both a male and a female are necessary, a fact which we are all aware is potentially changeable. Yet it is a significant point that biological material is necessary from both male and female, and thus such a division and the anatomical differences necessary for this process are objective and objectively significant. However, their significance justifies a type categorization consonant with the biological division, that is, it justifies us categorizing human beings into males and females, but it does not justify the claim that heterosexual relationships are the necessary cornerstone to the reproduction of the species. Human reproduction, in any full and meaningful sense, must include a plan for care beyond birth given the feebleness of human infants and the comparatively long period of full maturation, and thus is not reducible to conception or even parturition. Putting biological reproduction as the basis of sexual difference is not the same as putting heterosexuality at the basis or linking heterosexuality with reproduction in a broad sense. Under some contextual social conditions compulsory heterosexuality is demonstrably hurtful for reproduction, providing neither support nor nurturance necessary for a successful pregnancy, sufficient infant care and childcare and all that is necessary for the development of mature and reasonably functional human beings. Compulsory heterosexuality can contribute to the extreme vulnerability of mothers to violence and abuse. What is vital for reproduction is a child’s access to a somewhat stable group of caring adults. Many of us (including me) owe our very survival to the care and support we received from one or more adults that we had no biological relation with, who made it possible to overcome neglect or abuse from our biological parents or simply their absence. Moreover, Firestone’s most obvious error was to argue that biology contributed to patriarchy by creating a period where pregnant and nursing women are dependent on men, but there is no reason that such support must come from a male rather than a female. Reproduction thus does not require sustained heterosexual coupling. Sex may well be important in establishing stable, caring relationships; thus lesbian sex may be necessary to establish such relationships as are needed for some reproductive women.

Compulsory, exclusive heterosexuality, then, is not necessary for reproduction even without the use of adoption, surrogacy, or artificial insemination. Arguably, compulsory, exclusive heterosexuality is not the optimal condition for successful reproduction; many societies with compulsory, exclusive heterosexuality have epidemic proportions of child abuse and neglect. Adult care, some measure of stability, adults who have supportive loving relationships among themselves, and conception, are needed for successful reproduction in the full sense.

The effort to avoid prescriptive effects does not require an avoidance of any and all description. More generally, the effort to overcome deterministic theories about women’s innate limitations and orientations is not advanced through rejecting the idea that sex categories represent real and objective human differences, not differences entirely constructed by patriarchal discourses. I strongly agree with Susan Babbitt, who says the following: “ the extent that feminist theorists have emphasized the significance of processes of investigation and development—for instance, the retelling of social myths and fantasies in a more appropriate way—a commitment to a plausible version of realism is the best expression of both theoretical and political concerns.”(145)

There remains, then, the following question: What, if anything, follows from the material infrastructure of sexual difference in relation to the symbolic and cultural system of differentiating by gender? Feminist resistance to articulating a content to sexed identity is that any amount of content at all is seen as deterministic. But there is plenty of room between the sort of determinism that restricts women to domestic duties and a complete absence of material content.

Marx developed an approach to normative argument which made it anterior to empirical exploration. First we determine as best we can the realistic possibilities that might emerge from a given historical starting point, and then we can engage in ethical debate over which of those possibilities is most desirable and just. In this way one avoids an empty utopianism and can better inspire action and thus enact change. So we must start with ascertaining the facts, but in regard to sexed identity, beyond conception this is not so easy. If we set aside practices and simply look at bodies, currently existing sexual differences cannot be used as proof of much of anything. Large muscle development (promoted by ball sports) continues to be discouraged in girls, encouraged in boys, while the opposite is true of small muscle, or hand, dexterity. Gender dimorphism, or the degree of difference between the average male and female in regard to height, weight, ration of muscle to fat, and so on, is selected for, so that in some societies the average height difference can be as much as six inches whereas in others it can be an inch or even less. Cultural practices and social environments affect everything from testosterone and sperm production, average age of menarch, and life expectancy. The thesis that women are and will always be naturally physically disadvantaged to men is an unproven hypothesis, with some reason to believe it can be eliminated. Though we may never be equal in upper body strength, the potential of lower body strength is already clearly equal, and women generally have balance advantages over men. The question of what our permanent physical differences are should be changed, so that rather than asking what we are, we need to ask what we want to be. Despite all this, the significance of the division of labor in the process of biological reproduction is not unstable or undecidable all the way down. There is much that is variable about it, and social conditions can make pregnancy a true disability, but it will never have the range of variable significance that eye color, skin color, or height can have. It’s objective significance is transformable only by technology.

To categorize human beings on the basis of a biological division of reproductive roles is thus to recognize an objective type. In a sense, this could return us to Gayle Rubin’s original position from thirty years ago, in which a distinction is made between sex as a biological category and gender as a cultural practice. Let me revisit again why this was rejected. First, it was rejected out of a concern that it replicated the nature/culture distinction that presumes that we can clearly tell the effects of one from the other. Second, it was rejected because of the concern that filling anything into the category of sex will become viewed as the determining basis for gender.

With the benefit of hindsight, we can now see that maintaining a distinction between the objective category of sexed identity and the varied and culturally contingent practices of gender does not presume an absolute distinction of the old-fashioned sort between culture and a reified nature. By “old-fashioned” I mean the idea that nature is immutable and culture, alone, is mutable. In reality, what we put in the category of nature is mutable in several senses: (1) we can alter with technology many “natural” processes, such as the production of breast milk or various inhibitors to conception, (2) human ineptitude is altering the natural environment in catastrophic ways as we speak, and (3), even in regard to the immutability of scientific laws, the one absolutely predictable fact about science is that what we believe today about the universe to be justified by the best methods of science will be superceded in ways we cannot predict, and therefore our very ideas about the immutablity of the most confirmed laws must be held with a grain of salt. Thus, what we set aside as “nature” is in dialectical relation with “culture” in so far as it is altered by human practice and what we know about it is constantly altered as practices evolve, or devolve. Moreover, “culture” is not simply the flexible clay of human will, but its transformations occur within a “natural” context. We cannot, today at least, give a definitive answer to the question of determinism, or the limits of sexual transformation. It would be exceeding hubris, either from sociogeneticists or feminist social constructionists (and both seem to have plenty) to claim to know the limits, or absence thereof, to human biological transformation. Thus we need to set our sites closer to the present. What might we become tomorrow? What do we want to become?

Sexed Identity as Embodied Horizon

Beyond this, the phenomenological and hermeneutic approaches are the best means of exploring how sexed identity is manifest in the lives of particular women in specific social contexts. Phenomenologically one needs to account for the ways in which the body is lived, perceived in the world, presented, and experienced.

Hermeneutics is just as important however. The body is lived as it is lived in part because of the horizon with which it confronts the future. For girls, this horizon will generally include the future possibility of reproduction, even if this turns out never to come to pass and to have been physiologically impossible all along. Knowing that one may become pregnant and give birth to children in the future affects how one feels and thinks about pregnancy and childbirth, sexual relations, familial relations, and various possibilities for paid work or careers. Knowing that one’s mother experienced pregnancy and childbirth under certain conditions has an impact on how one imagines oneself in such situations. These provide the constraints on undecidability and total fluidity for the development of female and male sexed identities. All of these possible experiences are open to vast differences of interpretation, but the point of the hermeneutic account is that they must be dealt with in a way that those who grow up male do not have to deal with, at least not in the same way. The possibility of pregnancy, childbirth, nursing, and in many societies, rape, are parts of female’s horizons which we carry with us throughout childhood and much or all of our adult life. The way these are figured, imagined, experienced, accepted, and so on, is as variable as culture. But these elements exist in the female horizon and they exist there because of the ways in which we are embodied.

Substantive accounts of sexual difference beyond these bare sketches should be, I would suggest, existential descriptions of very particular situations within which groups of women, perhaps large groups, live. Even Irigaray’s imaginative descriptions of female lived experience might be put in this category, as descriptions of the phenomenology of a female anatomy under specified contextual conditions. Also included here would be the work of Emily Martin, Elizabeth Grosz, de Beauvoir, and many others.

Elizabeth Grosz puts brakes on Derrida’s deconstruction of sex in precisely this way, it seems to me. “It is not so easy to see,” she remarks, “how sexuality—in the sense of sexed subjectivity, male and female—can be understood as indeterminate.”(1995, 77) The possibilities of interpretation abound, but they occur within given conditions, conditions which she hints are “somehow ontological but entirely without qualities and attributes.”(1995, 77) She explains:

This is what I presume that the framework of sexual difference implies: that there is an irreducible specificity of each sex relative to the other, that there must be at least, but not necessarily only two sexes. In short, one lives one’s sexual indeterminacy, one’s possibilities for being sexed otherwise differently depending on whether one is male or female. This is not, however, to predetermine how one ‘is’ male or female, but simply to suggest that there is an ineradicable rift between the two, in whatever forms they are lived. Unless such a presumption is made, sexual difference remains in danger of collapsing into a sexual neutrality of precisely the kind Derrida problematizes in Heidegger and Levinas. Derrida’s dream of a multiplicity of ‘sexually marked voices’ seems to me worthy of careful consideration, as long as the question of the limits of possibility of each (sexed) body is recognized. Each sex has the capacity to (and frequently does) play with, become, a number of different sexualities; but not to take on the body and sex of the other.(1995, 77)

I cannot say it better than this.



1 The compatibilist position that has come to dominate the free will debate argues that the belief in incompatibility between free will and determinism rested on the idea that the human decisions which we take to motivate our actions are special kinds of processes, unlike any other processes. But many today agree that decisions are processes like any other—decisions are based on what we know, what we have experienced, what we want, and so on. One can give causal stories about decisions as easily as any other sorts of processes. Thus a certain sort of Davidsonian inspired monism has begun to gain influence in a variety of philosophical realms, and there is increasing interest in returning to previous formulations of monism in the history of western philosophy, such as the writings of Spinoza, by feminists as well as others. Monism would suggest that, to the extent that it presumes a nature/culture divide in regard to sexual identity, Rubin’s sex/gender distinction cannot be metaphysically sustained.

2 Another problem of process metaphysics for Butler would be their naturalism—even Nietzsche’s dynamic flux is arguably a naturalistic account.

3 See Weir 1996 and Ferguson 2004 for more elaborated criticisms of the contradiction in Butler’s work between the assertions of fluidity in Gender trouble and the recalcitrance of organizing norms in The Psychic Life of Power.

4 To give the agency to discourse rather than intentional subjects can produce another form of determinism, it is true, but it always appears less absolute than laws of the universe or the plans of the Gods.

5 To clarify my point here, I do think her account of performativity was misread as voluntaristic, but I don’t think she was misread as holding that a sex/gender distinction or any objectivist account of sexual difference assumed unmediated access to the real, a point that Haslanger puts a lot of weight on as I shall discuss.

6 As Angela Davis (1998) has noted, however, some of this variety in reproductive role is not new but quite old, such as the wet nurses of slavery and feudalism, and surrogate mothers based on class and station from biblical times. Davis urges us to consider why we think (and who thinks) that we have only now overcome the two-person system of reproduction. I would further suggest that the currently escalating globalizations of adoption follow a predictable path of political economy, in which poorer and largely non-white women give up their children to be raised by richer and largely white women. I strongly agree with Davis that reproductive technologies are neither good nor bad necessarily, but their development is structured by privilege and power and they are not in fact subverting structures of privilege. The case is similar for adoption, which is neither good nor bad necessarily, but which is evolving through a global capitalist political economy and in no way challenging to it.

7 To be completely accurate, Fuss does not present herself as an anti-essentialist, and argues that the anti-essentialist mistakenly essentializes the effects and nature of essentialism, an argument with which I agree. The position that she herself argues for, however, is based on the idea that a more consistent anti-essentialist will not be consistently anti-essentialist. I would again agree, but this simply takes anti-essentialism to the meta-level and makes it non-negotiable as a theoretical constraint. This is a deconstructive position I reject.



8 Let me be clear: I am not suggesting that racial categories won’t exist in the future or will naturally wither away. On that issue the jury is still out. Racial identity has two meanings: one is about biology, and is based on specious science, and the other just as common meaning is about culture. Because of history, racialized physical features have become correlated with cultural groups who have real identities. Thus, race is significant because of the cultural or ethnic groups that have emerged out of, in part, racial experiences. Reproduction, on the other hand, is not solely significant because cultures have attached various meanings to it.

9 In regard to disability, this is an ongoing debate. I have been very persuaded by the work of Tobin Siebers (citation needed).

10 As an aside, this has become an interesting pattern in the post-positivist, anti-foundationalist developments of twentieth century analytic epistemology and metaphysics, which has taken various forms through the later Carnap’s pragmatic approach, Neurath’s coherentism, Wittgenstein’s emphasis on games and forms of life, Quine’s holism, Davidson’s Principle of Charity, Putnams’ internal realism, Brandom’s pragmatic semantics, even Bayesian decision theory. The pattern is this: analytic philosophers develop more sophisticated and nuanced understandings of how philosophical reasoning actually proceeds, jettisoning the modernist attachment to substance, essence, pure datum, the thing in itself, and so on, and recognizing indeterminacy, holism, the ultimate grounds of justification in practices and forms of life, and the pluralities of forms of life. But what continually surprises many of us is that analytic philosophers remain singularly uninterested in pursuing the social and political implications of their new models; if they have destabilized the starting point from which philosophical reasoning occurs, they continue to be solely interested in producing rational reconstructions of the later steps of the process rather than becoming reflective about the cultural embeddedness of its initial assumptions. By contrast, continental philosophers have tended to pay sole attention to the less rational and less conscious, more socially situated foundations of philosophical questions, without pursuing the reasoning that occurs after a question has been posed. One might imagine the two traditions could someday be patched together. But the analytic philosopher’s inattention to the social aspects of the philosophical process means that they have often stopped short of facing up to the real implications of their own theses. Lynn Nelson’s book on Quine (1990) shows this most brilliantly, I have tried to show some similar results with Putnam (Alcoff 1995), Naomi Scheman has developed Wittgenstein in these directions, and there are other examples(). The point is not that our feminist readings of these figures are creative adaptations, but that they are simply following the original theory through to its logical conclusions. One could do an interesting piece of sociology of knowledge to diagnose the causes for this pattern of “theoria interruptus.” A generous reading would be that analytic philosophy has simply not yet realized the full implications of the death of positivism, a death that they, more than anyone, brought about. A perhaps less generous reading would be that the unconscious of analytic philosophy cannot face or acknowledge its own maturation, and continues to hold onto a dream world or infantile fantasy life of a parthenogenetic pure reason. They suffer, as John McCumber puts it, from post-positivist depression ().