Issue No. 8 (July 2004) — Regions of Sexuality
This paper is concerned with the limits of oppositional thinking about the construction of sexed subjects, and with the challenge of relaying knowledge about divergent sex/gender systems to scholars and readers who may never have the opportunity to see for themselves how different cultures operate.  Fieldwork, for most anthropology students, at least until doctoral candidacy, is prohibitively expensive; undergraduates in particular must rely on texts about 'regions of sexuality' other than their own. The paper argues that much of the existing work on cultural systems that incorporate a 'third sex' portray simplistic visions in which societies with more than two sex/gender categories are cast as superior to those that divide the world into just two. I argue that to understand whether a system is more or less oppressive than another we have to understand how it treats its various members, not only its 'thirds'. Glossing over that information impoverishes the information to which scholars unable to (re)visit specific sociocultural locations have access.
While careful anthropology clearly attempts to avoid dichotomous and superficial thinking, there are abuses of anthropological methods and data that we ignore at our peril. Such abuses include asking anthropology to answer for us why we are the way we are, whomever we may be. Lest my concern about this seem to come entirely out of the blue, I should give readers a sense of how I arrived at this observation. In the mid-1990's I 'came out' as an intersexed activist, and—as a graduate student in studies that bridged fields of inquiry in sociology and anthropology—was intrigued by the idea that some of my colleagues in the Intersex Movement put forward: Western culture and medical practice would do well to learn from cultures that had sex categories allowing the recognition of intersex states. I went out, eagerly in pursuit of an answer to the question 'Where can intersexed persons fit in the world?'—a question not entirely dissimilar to the ones that Will Roscoe identifies as informing his research trajectory in anthropology: 'For over two decades now I've been on an odyssey seeking the answers to three questions central to the meaning of being queer: who are we? where did we come from? what are we for?' (Roscoe 2004a).
What I would eventually learn was that recognition of third sexes and third genders is not equal to valuing the presence of those who were neither male nor female, and often hinges on the explicit devaluation of women, as with the Sambia of New Guinea, or on the valuation of female virginity at the expense of valuing female humanity, as in Polynesia.  As someone who was assigned and raised female, and a committed feminist, such models for the recognition of a third sex/gender were not actually feasible for adoption here. Rather than finding as Roscoe has, an answer for who 'we' are (who are we anyway?), my research left me increasingly frustrated and increasingly aware that I wished neither to become one who would remake other cultures' sex/gender systems into my own, nor immerse myself into compliant acceptance of a logic that rested on the vilification of women's bodies as sources of defiling putridity.
Although Western societies of the twentieth century have ossified a 'common sense' understanding of sex and gender, in which male and female are presumed to be the obvious limit of possibility, there are some contemporary, and even more historical, instances of cultures whose sex/gender systems have not been limited by a dichotomous binary opposition of male and female. Anthropologists have been fascinated by social worlds that recognise more than two sexes and/or two genders for, among other things, such symbolic organisations demonstrate that sex is not simply given to us as an obvious biological fact, but that how we apprehend sex is shaped by other cultural, symbolic and structural features. This fascination has a spectacular history that renders it less than benign, shaped as it has been by attitudes that diminish and neglect any world view that does not arrive at the same modern dimorphism of the anthropologists' training and heritage. Fascination then provides no guarantee of protection from intellectual, cultural and political neglect. As Gilbert Herdt explains, 'This neglect [of third sexes and third genders] is largely due to the intellectual, social and morally defined strictures of sexual dimorphism' (Herdt 1994a: 12). Nonetheless, there has been a small explosion of work on cultural systems that include more than just two sexes and/or genders in their symbolic scheme. It is likely that this is due in large part to the conceptual and political changes in (primarily) North American contexts, most specifically the increasing visibility of 'gay culture' in the years following the Stonewall riots of 1969, and the development in the late 1980s of a radically politicised gay civil rights movement arising as a response to the AIDS crisis. Ironically, this rights discourse, because of its appeal to a natural history or untainted past, and popular deployment of anthropological terms, tends to recreate a view of the world organised around sexual dimorphism. 
Transforming anthropological ways of seeing is not a simple matter. The anthropological fascination with cultures incorporating more than male and female sex categories into their symbolic classification schemas has been complicated by the discipline's early ties to missionary activities, and hence, to religious concerns regarding sexuality and sin. Thus, the fascination has not been simply of the sort that would imply a keen interest and relativist approach, but also one of the sort suggesting revulsion and horror and a concomitant desire to 'civilise the natives'. This has resulted in what Gilbert Herdt characterises in his introduction to the Third Sex, Third Gender collection as an ethnographic tradition impoverished by the adherence to a concept of deviance, coupled with a desire to 'Westernise' the populations being written about. Indeed, Mary Douglas has argued that as a further consequence of their religiosity, early theologically minded anthropologists, unable to think beyond their own cultural frameworks, mistakenly assumed that cultures that incorporated 'excess' in ways radically different from Western European cultures confused the categories of purity and impurity, of the sacred and the profane. Early anthropologists believed gender systems that were neither dichotomous nor limited to just two categories constituted a confusion of fundamental (Western, advanced) categories, and therefore classified those cultures as 'primitive'. In Purity and Danger, Douglas demonstrates that Henry Burnett Tyler, Robertson Smith, T. H. Green and Emile Durkheim held strong moral imperatives to define sacred and profane categories as both oppositional and exclusive. Douglas explains that their work relied heavily on a Darwinian theory of evolution, mandating an adherence to a binary system of classification, and positing that civilisations 'progressed' toward systems in which religions would hold the sacred and the profane at a greater and greater distance. Of course, the Judeo-Christian tradition was held to be at the top of the sociocultural, symbolic pile. Douglas argues that this relates to gender divisions in Western 'rationalist' and theological discourses that posit women as creatures who lack the purity of soul to take on important religious duties (i.e. the priesthood in Catholicism) and as beings whose reproductive nature makes them poor candidates for public office and professional life (Douglas 1966: 12–24).
A tendency to hierarchically organise cultures according to the division of the sacred and the profane does not begin and end with anthropologists or sociologists. In a tidy summation, Roscoe argues that the imperative to cast some groups as inferior, primitive and hopelessly muddled is traceable at least to the practices of early European conquerors:
Bernal Diaz's chronicle of the conquest of Mexico [describes] the results of a preliminary expedition to the Mexican mainland in 1517. What he and his compatriots reported to their superiors proved to be the essential prerequisites of Spanish conquest. In the temples of the heathens, he wrote, the Spaniards found idols depicting Sodomitical acts . . . and gold. Sodomy and Gold—the two were fatefully linked in the Spanish colonial enterprise. One motivated it, the other justified it (Roscoe 1995: 450).The past two decades have seen a significant turn in perspective from the early anthropologists and conquistadors. This is not necessarily to say that the will to see differently was not operating earlier than the post-Stonewall era, and Douglas' work alone demonstrates this; however, an increasing acceptance of work that falls roughly into 'queer' theory reveals a concentration of publications in the past two decades.  If more recent anthropological work avoids the early finger-pointing behaviours of colonisers and theologically-minded scholars, that does not mean that it avoids the pitfalls of using 'others' to further its own domestic agendas. This may be a case of a shift in vision swinging too far, such that it may address many valid concerns but still continue to accomplish that work at the expense of its subjects. Roscoe's work, for example, is on the one hand insightful for its critique of colonialism, but on the other hand, is so couched in an overtly masculinist, gay civil rights agenda that it is inclined to miss the symbolic importance of anything not concerned with navigating the structural tensions between the threat of gay sexuality to 'proper' masculinity. Carolyn Epple criticises this preoccupation, arguing that 'Roscoe and [Walter] Williams, both proponents of 'gay', associate the 'berdache' or 'alternate gender' with same-sex sexual practices in their terminology and categorical constructs' (Epple 1998: 269–270) and argues further, citing the influence of Robert Padgug on the development of her thought, that anthropologists working on 'alternative' sex and/or gender systems repeatedly make the mistake of treating sexual forms, behaviours and genders as though their meanings and structures were neither especially discrete nor specific when, in fact, they are quite unique and specific (Epple 1998: 270).
Roscoe, for example, continually focuses his attention specifically on anality and mode of dress as though those were the most salient and important features of the berdache when, in fact, it is likely that those features take on a central importance only in a history of Euro-American colonisation. Roscoe's work then, remains paradoxically caught in the very structures that he claims to wish to redress at a broad level: moralist equations of 'sodomy' and anality with subordinate subjectivity that threatens the sanctity of Euro-American masculinity. Richard Trexler explains this paradox in terms of political commitments, arguing that the problems of classification and of understanding 'berdache' are a persistent trouble for anthropologists because of contemporary sexual politics that
puts a premium on the view that individual gays have become what they in truth always were. Stated inversely, gays sometimes fiercely combat any implication that they have become what they are because of some constraint in their upbringing. This means that gays, but not they alone, often resist the notion of social construction in favor of biological determinism. Thus what began as an anti-imperialistic anthropological and historiographical romanticizing of native male Americans has been to some degree converted into a gay view that the large majority of those who became berdaches, in Will Roscoe's words, 'did so entirely of their own volition' (Trexler 2002: 613).Trexler's concern with the epistemic focus on voluntarism in works such as Roscoe's is that it seems to wish to have it 'both ways' and posits third sex and/or third gender categories as both natural and as chosen (Trexler 2002: 613). Such roles are asked then to stand as evidence of apparently transhistorical cultural continuity and as just one of many options in societies far less restrictive than Western sex-gender systems are. For Trexler such a view is not only contradictory but also just as naïve and patronising as are totalising visions of various indigenous groups as peaceful farmers, when they have cultures that are, in fact 'saturated with concerns about blood, violence, and power' (Trexler 2002: 614).
Not all scholars take such a dim view of Roscoe's work as Trexler, and Kath Weston argues that Roscoe's work is among the more theoretically subtle bodies of research available (Weston 1993: 345). However, Weston's own observation that many ethnographers continue to use terms in unproblematised ways applies as much to Roscoe's work as any other on 'berdache' that 'portray[s] male-to-female 'cross-dressing' as a form of 'feminization' perpetuat[ing] Western assumptions about the unambiguously binary character of gender' (Weston 1993: 346). At least some examples from Roscoe's current work demonstrate a theoretical commitment to oppositional gender categories. Roscoe's current website, for example, explains berdache categorically, stating that 'two-spirits specialized in the work of the opposite sex. Male two-spirits had a reputation for excelling in women's arts' (Roscoe 2004b; my emphasis). Roscoe relies on a Western conceptualisation of opposition to categorise two-spiritedness via a gendered division of labour. For anyone who, like me, wishes to question relations of power it is endlessly infuriating to see the definitions of berdache, especially in Roscoe's work on We'Wha, relying on the notion that berdache males were better than women were at 'women's work'. This follows the typical masculinist perception that whatever is done by biological males, however they may be dressed, is de facto superior to the work done by women. Roscoe's acceptance of sex-typed labour is predicated on common sense categories that are questioned by other scholars. Consider, for example, the description given by Weston:
research on American Indian Berdache engages in the fantasy of a society in which homosexuality can be at once normative and transgressive. Berdache is another catch-all term that ethnographers have used to describe males (or, less often, females) who take on at least some of the garments, occupations, and/or sexual partners culturally prescribed for what Anglo-Europeans might call the opposite sex (Weston 1993: 352).Weston's own attempt to redefine Berdache is constrained by dualist language, but draws reflective attention to that problem in the clear recognition that hers is an Anglo-American definition. Clearly, something is being lost in the translation, but at least Weston is aware of that gap between cultures, and across languages and symbolic systems. Obviously then, not all ethnographers are unreflective about terminology and the very fact that they are not raises a serious question: why do anthropologists like Roscoe and Williams continue to cast Berdache in these very binary terms that rest on Western perceptions that privilege masculinity and impoverish femininity?
Is it because, as Weston suggests, the Berdache are being called upon to fulfil a Western fantasy of a simultaneously 'exotic' and easily subsumable symbolic order? After all, if we rely on sex-typed work as an internal measure of what defines a 'two-spirit' (and why, if this is a third gender or third sex category, are anthropologists so comfortable using the term 'two-spirit'?), then it seems that the two-spirit category does little to disrupt binary gender categories, and the use of their existence to trace a gay history, or more recently, an intersex or transgender history, becomes suspect.
Caution is necessary when culturally specific symbolic orders are employed to prove a(ny) point about Western sex/gender systems; the notion of learning from 'other' cultures raises serious problems. Testing the validity of the anthropologist's report is impossible. We need to rely on the accuracy of the anthropologist's vision, but how are we to do that when 'women's work' and 'men's work' remain in diametrical opposition in the classificatory schemas used by anthropologists to explain what is supposedly a distinct set of social meanings? What are we supposed to do about the persistence of a two-sex model to define those who are neither male, nor female, who may not, in fact see themselves as superior to women, or not be cast as better partners for men simply because they are not women but can outperform women at their own traditional work?
Weston answers this set of concerns when she charges that information that fails to take these problems into account does so because it is operating through the lenses of fantasy, rather than using clear methods for analysis and interpretation (Weston 1993: 344–346). We move then from colonial ab/uses of 'third' sex/gender categories to contemporary masculinist ab/uses of 'third' sex/gender categories, from the service of the church and state to the service of male-oriented gay studies that leave the dualism of the male-female hierarchy relatively untouched.
Divergent Approaches to 'Third Sex' Categories
Not all work on third sex and third gender categories operates in such a totalising manner, however; if only because of the multiplicity of views held within them, collections such as Gilbert Herdt's Third Sex, Third Gender seek to look with a different eye toward cultures in which sexual roles beyond male and female are an acknowledged, integral, thread of the cultural fabric. Though I find Herdt's own work on 'third categories' to be problematic, the Third Sex, Third Gender collection does not plough women under in an effort to serve a masculinist fantasy of male superiority. Imperfect though the work is, Herdt's preface and introduction indicate a three-fold corrective to the historical blind spots of traditional epistemological approaches to third sex and gender categories: first, scholars must seek to understand the historical and cultural context in which 'third' categories exist; second, rather than focusing solely on labour divisions and ceremonial and kinship institutions, the erotic experiences of the subjects must be recognised; third, scholars must consider the relationship of cultural process and the historical continuance of third sex and gender categories (Herdt 1994b: 18–19). The turn of Herdt's collection then, is away from a tradition in which 'excessive' gender/sex categories are interpreted as confused or impure states that betray the lack of 'civilised' evolution of a culture and toward an understanding of the logic of thirds. The collection also moves away from the tendency to measure 'third' sex and gender categories against a male benchmark.
Herdt aims to bring nuance to the study of complex sex/gender systems, and to avoid the kinds of naïve reconstructive appropriations of Native cultures that I have outlined. Weston singles out Herdt himself for successfully turning away from Western categories when he includes an informant's discussion of the local use-value of being a 'third sex' rather than being 'really' male or female as we do with 'hermaphrodites' in Western societies (Weston 1993: 349). The collection, however, is not perfect; it has its own conceptual blind spots. Herdt's three key directives for fleshing out the range of possibilities and structural positions of 'thirds' within their cultures promises to add a sense of specificity to cross-cultural studies of sex/gender, but lumping all the erotic and symbolic elements of these cultures together under one rubric of 'third sex and gender' categories undermines that goal. The fact that all are organised in the book under the rubric of 'thirds' indicates that many ethnographers still think along a dimorphic axis, permitting the occasional disruption to be entertained, but not actually considering that the so-called 'third' might be a 'first' or even one of any of a multiplicity of possible sex categories. 
As Deborah Elliston indicates in her critique of Herdt's work on 'ritualized homosexuality' the lack of ability to think outside of oppositional categorisations (male versus female; heterosexual versus homosexual) presumes that sexual practices such as semen practices are primary organising principles in Melanesian cultures, when in fact, stratification according to age and gender are more central (Elliston 1995: 848). In other words, the issue is still being cast in terms of a Western preoccupation with centre versus margin, neglecting local cultural places of origin and trajectory for organisation.  A missing context for this genre of error in Herdt and other proponents of supposed 'thirdness' is feminist work on symbolic excess.
The Role of 'Excess' in Binary Analytic Models
Among the first and most recognised theorists to apply the concept of 'excess' to the conceptual structuring of sexual difference is Luce Irigaray, whose conceptual category of 'jouissance' suggests that Western desires for and fear of excess are contained and circumscribed by the spectacular image of the feminine 'other'. Irigaray argues that within Western philosophic and symbolic traditions, 'The feminine is apprehended not in relation to itself, but from the point of view of man, and through a purely erotic strategy, a strategy moreover which is dictated by masculine pleasure' (Irigaray 1991: 178). Irigaray's argument is that the category of the feminine, and the place of women in general as the bearers of that categorical imperative, is always already seen through masculine apprehensions about desire and pleasure. These apprehensions, if one takes them up as signifiers in the symbolic realm, rather than merely as symptoms in the psyche, happen at the borders of difference and threaten to blur the distinctions between categories of meaning. Thus, it is quite possible to read 'the feminine' of Euro-American culture as a repository for various anxieties, most of which hinge upon ambiguity. Mary Douglas argues that ambiguity, seen as a general state of disorder,
spoils pattern [but] also provides the materials of pattern. Order implies restriction; from all possible materials a limited selection has been made . . . [D]isorder by implication is unlimited, no pattern has been realised in it, but its potential for patterning is indefinite (Douglas 1966: 94).It is because the feminine is symbolically unstable, always a repetition of a non-existent antecedent, that it takes up and is posited as the location of excess, dissolution and possibility. As such, the feminine is both a promise and a threat, a position which both contains/limits and produces fear. This fear is always a fear of the dissolution of the masculine/subject, which is the only true subject under the law of patriarchy.  This is how things operate under a two sex, two gender model.
This backdrop is neither transhistorical nor transcultural, but as a context for the operation of Western culture, has had an enormous influence upon the way in which anthropologists have viewed cultural structures outside a Western context. Because, within the Western framework, persons who fall outside the strict binary classification schema have been treated as pariahs, Herdt argues they have been traditionally overlooked and omitted from reports on non-Western societies. Herdt's collection on third sex and gender categories states a self-conscious desire to understand the historical scripting within the anthropological symbolic classification of third sex categories as iconic examples of 'matter out of place,' such that, 'The hermaphrodite, for instance, may become a symbol of boundary blurring: of the anomalous, the unclean, the tainted, the morally inept or corrupt, indeed, the 'monsters' of the cultural imagination of modern Americans' (Herdt 1994b: 17). Herdt lauds Foucault's work on Herculine Barbin for noting the crisis Barbin raised for modern ideology by identifying as neither male nor female (Herdt 1994b: 13). However, Herdt makes the mistake here of trusting Foucault's introduction more than Barbin's own narrative for it is clear in Barbin's own text that although she saw herself in many ways as an exceptional female, she did not perceive herself as necessarily beyond the boundaries of the female.  Foucault conflates Barbin's perception of herself as exceptional with seeing herself as in between male and female or as neither male nor female (Foucault 1980: xiii). Foucault thus fails to see that because the category female already exceeds the category male, recognition of one's exceptional excess does not necessarily violate a sense of oneself as female, though it may well transgress some of the simple, taxonomic distinctions made between male and female. Foucault's use of Barbin's story is then more indicative of his agenda in publishing her manuscript as documentation of the 'happy limbo of a non-identity'—and the masculine, predatory sexual exploits possible within a convent for an individual with a 'non-identity'—than it is of Barbin's perception of herself (Foucault 1980: xiii; Butler 1999: 31–32, 119–135).
If my criticism sounds frustrated, it is likely that as a reader, I am indeed frustrated. Each time I set out to read a new piece about 'thirds' or 'alternate' sexes and/or genders, I head into the material thinking that I am going to read something new and exciting that reveals structural problems in the conceptualisation of difference, perhaps because I too have sought answers regarding 'who we are' as intersexed people. That search has lead to readings of intersexuality that repeatedly fixate on the disruption of masculinity, the frailty of maleness, and repeated uses of 'the feminine' to prove a point about the central importance of masculinity qua masculinity. In short, even those thinkers who contest homophobic attitudes toward 'third' categories do so in ways that leave femaleness as a problematic threat to maleness. For all his radicalism, Foucault remains committed to the notion that it was Barbin's maleness (expressed as superior intellect, sexual assertiveness, and body a markedly different from the 'girls among whom he grew up') that Barbin's female contemporaries, 'blind as characters in a Greek fable', wilfully refused to recognise (Foucault 1980: xii–xiii; my emphasis). Were Foucault's position correct that Barbin profited from being somehow more male than female in the cloisters, it would then be extremely doubtful that Barbin would have killed herself after being assigned male sex and occupation. Yet Foucault's ascription of Barbin's status prevails in Herdt's perception and use of Barbin.
Herdt's acceptance of Barbin as a representative of a 'third gender', especially given his book's mandate to look for more examples of thirds, is caught up in a power play that grants the ethnographer the power to decide whether informants represent third sex or third gender positions. Although the subject-view of the ethnographer is certainly implicated in Herdt's introduction, the transmission of information from ethnographers to readers remains relatively unproblematised, leaving out how each of their active constructions might influence classification. The objectification of 'others' that takes place in these investigations is inadequately considered by Herdt or, arguably, by the other ethnographers in the collection. The problem is an over-reliance on the very conceptual categories that work in this area seeks to contest, even when attempting to work cautiously through those very issues. The most promising essay in the collection is Nico Besnier's work on liminality, though it too has its difficulties.
From 'Third' to 'Liminal'
Nico Besnier's essay on Polynesian gender liminality, in Herdt's collection, rejects the term 'berdache' as well as the terms 'transgender,' 'homosexual', and 'gay' for being overly constraining as categorical distinctions and, furthermore, for being out of context as descriptors for Polynesian gender organisation. Rejecting 'Western gay scholarship on the fringe of anthropology [which] clearly buys into a highly romanticised view of the 'Other' ... bear[ing] only a remote relationship to the ethnographic evidence', Besnier instead chooses the term 'liminality' to talk about those persons whose gender status resides at symbolic interstices between male and female (Besnier 1994: 317). Besnier's work seeks to avoid indulging the same romanticising gaze as the Western gay scholarship and draws out at least two points that bear further consideration here. First, Besnier notes that although Berdache roles in North American Native communities are often presented in the literature as prestigious, Besnier (Besnier 1994: 298) argues that the Polynesian 'liminals' have varying social rank of no apparent connection to their liminal status. Second, Besnier reports that involvement 'in homosexual activities is neither a necessary nor a sufficient criterion for gender liminal status' (Besnier 1994: 299). Besnier, who is indeed trying to find out what is being recognised in the liminal person, if not social prestige or erotic behaviour(s), is faced with a problem of what, in fact, the function served by the liminal male is, and argues that interpreting the gender-liminal male as a negative model can be useful, stating that the liminal male serves at least two probable functions within Polynesian kinship structures (Besnier 1994: 304). First, because the liminal male never has sexual relationships with other liminal males but only with 'nonliminal' males, then the liminal male is considered to be 'fair game' in a way that women are not (Besnier 1994: 301). In Tongan society, Besnier reports that the liminal male is capable of providing a sexual outlet for males without presenting the threat of pregnancy or other social disasters that would undo the ideal of marriage in that community (Besnier 1994: 299). Second, taking up on the previous work of Robert Levy, Besnier entertains the idea that the Tahitian Mahu, for example, indicates what non-Mahu males should avoid in their behaviours. Although Besnier rejects this interpretation of the negative-image because of its historical inaccuracy and its failure to recognise the distinctiveness of categories, Besnier does propose that perhaps that Mahu might present a negative-image for younger women. That is to say, the liminal behaviour is not something for men to avoid, but rather, something for women to avoid (Besnier 1994: 308). This is an interesting proposal, which circles back to the protection of the kinship structures around heterosexual marriage in Polynesian cultures. However, Besnier is cautionary about interpreting this function of the liminal as an institutional role because what constitutes a cultural institution is difficult to define and because liminality is by no means a hermetic category, but rather, one which 'leaks at the seams' (Besnier 1994: 309).
Ultimately, according to Besnier, there is no good evidence to support the idea that the liminal person constitutes a recognised 'third sex' or 'third gender' (Besnier 1994: 320); furthermore, there is no reason to presume that the liminal role(s) constitute a challenge to gender dimorphism in Polynesian cultures. This is because of that 'leakiness' which, therefore, creates too much diversity in liminality for it to be recognised as a single category unto itself in which liminal persons resemble one another to the same degree that non-liminal persons do in Besnier's sense of things. Ultimately, Besnier argues, the liminal persons of Polynesia are too changeable, too sensitive to external forces (such as gay discourses of the West) and too diverse within themselves to constitute a clear 'third' sex or gender.
What remains unanswered in Besnier's article is whether or not the liminal gender constitutes a central category, i.e.: would the cultures which recognise liminal persons be significantly or negatively impacted by the loss of that category within the kinship system? If we take the functional arguments put forth by Besnier seriously, then one must suppose that there would, indeed, be a negative impact. Thus, to claim that the liminal set of possibilities is not a clear category may be to do it a disservice. In response to the complication posited by Besnier that the liminal persons present too broad a set of possibilities which are subject to historical change to constitute a clear, third category, one may want to ask, in chorus with Judith Butler, if anyone ever repeats a self-same gender or sex identity. In fact, the very fact that it is re/petition—i.e., performance—gender, according to current theory is always a production of difference/difference. Finally, if this is a category of disorder, then is it a disordering that pollutes or one that sanctifies? If we take up the kinship arguments posited by Besnier, then again, we are faced with a very complex set of possibilities in which, even if it is a negative category, the liminal is that which shores up and stop-gaps the threats and absences in the heterosexual matrix. While all of this may be criticised as an overwhelming subsuming of the liminal person into one's own agenda, it is just as worthwhile to entertain the idea that while this group may not constitute a third sex per se, it does allow a third set of possibilities within the culture and beyond that of the heterosexual. Perhaps it is only a matter of perspective.
Although I find much of Roscoe's work deeply troubling, his article in the Herdt collection avoids some of the masculinist bias in his imposition of sex-typed hierarchisations, but continues to navigate rather uneasily the issues of dress, work and sexuality. Developing an interesting complexity of thought, Roscoe's article, 'How to Become a Berdache' begins with one of the principles I discussed earlier in this paper, namely that what people write about third sex categories 'reflects more the influence of existing Western discourses on gender, sexuality and the Other than on what observers actually witnessed' (Roscoe 1994: 330). Yet Roscoe continues to use sex-typed behaviours to report that Berdache are written as being men who take on woman's work and woman's dress in their everyday lives (Roscoe 1994: 330). Roscoe recognises the conceptual limitations of such understandings and argues that this approach to Berdache is severely impoverished by a lack of social and linguistic concepts through which to interpret the lives of berdache, who have been referred to as hermaphrodites, sodomites, transsexuals and homosexuals (Roscoe 1994: 331). Roscoe points out that at the very least, transsexual and homosexual refer to gender and to sexuality respectively, indicating confusion about what is being marked in using these terms. I would add that the use of the term 'hermaphrodite' indicates a perplexity on the part of writers regarding the morphology of the Berdache and that the use of the term 'sodomite' refers possibly to sexuality, possibly to desire, and probably has more to do with the various writers' moral frameworks than with anything else being described. The fact that Roscoe recognises the term 'berdache' and employs it in his writings, even though it is a blanketing term applied by external forces upon a wide variety of cultures, suggests that in spite of a declared desire to avoid a Westernising gaze, some terms retain a certain hegemonic utility to continually re-set Roscoe's conceptual stage for thinking about sexuality in Native North American cultures.  Moreover, Roscoe's continued use of 'he' to locate the subject place of most Berdaches indicates that Roscoe is not so much concerned with where 'we' have come from as 'queer' people. Roscoe is careful, however, to note that his use of the term 'berdache' is separated by feminist critiques of biological determinism from the initial deployment of the term to denote younger males in homosexual relationships (Roscoe 1994: 331). What is actually rather odd about this article is that, published in 1994, it is more conceptually nuanced than his essay of the following year, 'Strange Craft, Strange History, Strange Folks' (Roscoe 1995), which clearly relies on the idea that Berdache are those males who have engaged in a gender crossing. Perhaps it is a matter of discursive and political convenience for Roscoe to decide to nod toward feminist thought in one place and to neglect it in another.
According to Roscoe, there is a consensus that there are three intrinsic markers common to both male and female berdache: first, and most importantly, one must have productive specialisation in tasks generally thought to be the work of one's 'opposite' sex; second, there must be supernatural sanction and the adoption of a special religious function within one's culture; third, and least importantly, there will be an observable gender variation (Roscoe 1994: 332). However, on the point of gender variation, although cross-dressing is widely reported as being common to the berdache, as well as the most obvious form of gender-crossing, Roscoe reports that it is not a reliable marker and that, in fact, Berdache usually adopt a form of dress which is distinctive and neither male nor female (Roscoe 1994: 332). Here then, Roscoe is presenting his first evidence that, in fact, the Berdache are poorly understood, if at all, if they are taken to represent a crossing over between two genders; in fact, he argues that Berdache is quite distinctively a separate and third sex/gender category. Because in the different language groups one finds terms which bear nothing of either male or female markers to indicate berdache, combined with the fact that many Berdache did not engage in any mimicry of an 'opposite' sex, Roscoe is able to posit that anthropologists would be better off to conceive of Berdache as a distinct category of sex and gender. Taking such a viewpoint refuses the notion of individual pathology present in typical anthropological approaches to berdache—namely, that the role allowed weak men who were 'inadequate' as males to serve a functional role in their cultures in spite of their weaknesses. In fact, Roscoe states that the Mohave, for example, have:
the initiation ceremony for male Berdaches . . . not to effect a transformation in their 'personal habits' but merely to acknowledge them. This ceremony follows the common pattern of rites of passage as outlined by Victor Turner, with the phases of separation, liminality and incorporation . . . The passage, however, is not from male to female, but from boy to alyha•, a transition of both gender status and age (Roscoe 1994: 362).This, among other examples of similar distinctions, and not the traditional interpretations of Berdache as a liminal state in between male and female or as a crossing over of one gender to another, leads Roscoe to posit three new criteria for determining, describing and approaching berdache. Similar to the hierarchical first three criteria outlined above, Roscoe posits firstly that the Berdache division of labour and system of prestige exists within a cultural context in which women may be economically viable subjects unto themselves; secondly he states that there must be a system in which gender is not held to be pinned to biology or a necessary correlate to the sex of one's body, or alternately, there must be a system of thought which understands the physical body as changeable and nondichotomous; and finally, there must be historical imperatives or significant persons that facilitate taking on and creating/shaping gender identities for oneself (Roscoe 1994: 371).
Roscoe's goal in outlining the three imperatives above is to note that the possibility for third sex and gender status is not a phenomenon unique to North American Native cultures, but occurs in different temporal and geographical sites worldwide. Roscoe thus concludes that it will be the task of anthropology to correctly interpret such cases not as indications of pathologies in any given society, but as markers of diversity that move away from a tyranny of biology and binarisms over subjects.
Binaries and Excess Revisited
In her pivotal essay, 'The Traffic in Women,' Gayle Rubin argues that: 'Sex is sex, but what counts as sex is . . . culturally determined and obtained. Every society also has a sex/gender system—a set of arrangements by which the biological raw material of human sex . . . is shaped by human, social intervention and satisfied in a conventional manner' (Rubin 1975: 165). In the essay, Rubin rejects the conceptualisations of 'mode of reproduction' and 'patriarchy' as synonymous and/or adequate identifiers for sex/gender systems because, she argues, the structure of a sex/gender system is not simply the locus of patriarchal hierarchy or of reproduction alone. Instead of accepting these categories as sufficient descriptors, Rubin links kinship structures and gift-giving together to explain that most kinship systems involve a process in which women are transacted objects passed between individual males and the clans they represent. That some cultures have what appear to be 'extra' gender categories which may enter into the field of exchange was noted in the early years of anthropology, and as I argued at the beginning of this paper, led anthropologists to maintain that these cultures were confusing their categories of sacred and profane relationships. That Euro-American cultures also transact women such that 'the initiatory performative, 'It's a girl!' anticipates the eventual arrival of the sanction, 'I pronounce you man and wife'' (Butler 1993: 232), pre-scripted the set of relations that anthropologists would and would not assess as being 'normal' mappings of appropriate and inappropriate sanctions.
To reiterate from my own introduction, within 'our' cultural system, to become an appropriately gendered member of our kinship system is, primarily, to turn away from one's desire for the mother. Within our particular binary structure, the mother is structurally and symbolically imbued with a threatened negation of the child who must, against that threat, become a sexed subject, a gendered 'I' who, if she is to be 'female', will either be given away by her father to another man, or taken from her father by another man, or, if the subject is to be 'male', will refuse the desires of and for his mother, such that he will neither resemble her as an exchange-good or as one who desires men. To fall beyond this scope is to fall into the margins of Western culture, it is to be a 'deviant,' not quite female, or not quite 'male'— in short, it is to become impure.
Of course, the Euro-American kinship system is no more eternal or static than any other system; it is neither a once-and-for-all-time thing, nor a system of identical copies, but rather, a constantly reforming and reconstituting set of resemblances. Thus, although women are still significantly less powerful within our social institutions, not the least of which is capitalism and its mode/s of production/reproduction, women are not always already transacted into marriages, and families rarely have total say over whom one will or will not marry. Along with these changes, we see that a refusal to play along the heterosexual axis does not necessarily or irrevocably label the dissenters (lesbians, prostitutes, single mothers, unmarried women, gay men, bisexuals, the polyamourous, etcetera) as 'deviants', although they may still be seen as 'marginal'. However, there is a further margin beyond the realm of the categories of gender and deviance, and that is where our notion of 'excess' or 'impurity' in the sex/gender system now lies: on the shoulders of the transsex and transgender communities. This does not, however, mean that 'the feminine' no longer represents a threatening excess, but merely that trans-communities and feminine excess/jouissance can be mapped into an overlapping liminality. At this time, we do not have a recognised 'third sex' category.
If I take Besnier, Herdt and Roscoe at their word, and consider their assessments against the other theorists considered in this essay, Douglas, Epple, and Trexler, then it appears that there is a relationship between the relative power of women and their central position as autonomous subjects, and the degree of legitimate recognition of 'third sex' or 'third gender' categories. For instance, whereas Roscoe argues that Native American cultures are traditionally more egalitarian, recognising women as inherently capable, viable and productive members of their cultures' economies, Besnier argues that the various Polynesian cultures approach female production and reproduction as subordinate and dependent fields within their economic structures. Thus, where Roscoe sees less gender stratification in Native American cultures, he also sees a legitimate and central 'third category' as a logical correlative. Within the same logical paradigm, Besnier argues that where women's roles are given less value, the 'liminal' subject can only be seen as a negative image, not as a central, sanctioned 'third sex' or 'third gender'.
Whether or not this model of gender stratification and female subordination—being directly related to the degree to which gender 'liminality' is either central or marginal—can be borne out in all cultures is not clear. However, the theorisations of these two broad categories of Polynesian and North American Native cultures—each of which contains multivalent cultures within these 'short-hand' nomenclatures, are consistent with what is evident in Euro-American culture: that where we recognise only two sexes, male and female, and where we have both productive and reproductive hierarchies valuing the male term over the female, we also largely refuse recognition of any sex or gender possibility beyond that scope. In addition, persons who do fall outside the terms 'male' and 'female', or 'man' and 'woman' are referred to as transsexual and transgendered, respectively. These labels, which suggest a crossing from one possible term to the only other available term, are not recognised as a 'third', or even fourth, possibility in themselves, but are rather considered to be temporary stopping points along an axis of pathology in which one is seeking, or being made to rectify, the situation of 'being in the wrong body'. This is more similar to the Polynesian range of gender identification and liminality than it is to the Native systems in which being a Berdache has nothing to do with being in the wrong body; in fact, it is about being in the absolutely correct body: one which is required to complete the kinship structure and spiritual requirements of one's community.
It seems then, that until a society does away with a stratified sex/gender system, those things residing outside the accepted and central terms will continue to be perceived as impure states. Perhaps, therefore, the measure of a society's civil liberties comes partly through the measure of its sex/gender system. 
M. Morgan Holmes has been researching the intersection of sexuality, gender and medicine since 1993. She has published in queer theory anthologies and most recently in the journal Sexualities. Morgan is an Assistant Professor in the Sociology department at Wilfrid Laurier University, Waterloo, Canada and researches primarily in the areas concerned with the social importance of embodiment and the medicalization of difference. E-mail: email@example.com
 I wish to thank Wendy O'Brien and Iain Morland for their patience and encouragement, my anonymous reviewers for their helpful comments on the first draft of the paper, and the support of my colleagues at Wilfrid Laurier University. [return]
 See Gilbert Herdt (1994c) for a full discussion of the Sambia vision of women as a dangerous source of defilement that can be the undoing of maleness. I deal with the valuation of female virginity at the expense of female humanity more fully below, when I discuss Nico Besnier's work on gender liminality. [return]
 See Raven Kaldera (1998) and Berdache Jordan (2004) for examples. [return]
 The reasons for increasing acceptance of 'queer' work in academic publishing and conferencing exceeds the scope of this paper. I mention it here only for the benefit of readers who may be coming to the material for the first time, so that they may have a rough sense of what makes it 'suddenly' more feasible to have an academic career focused on issues of sexuality, yet not grounded in a concern to 'explain' 'deviance'. [return]
 The Siberian Chuckhi, for example, have at least seven discrete gender categories evident in their vocabulary. [return]
 Although some of the papers of the collection challenge the notion of a third category, they have, nonetheless, accepted its terms enough to profit from the prestige of publication in the volume. [return]
 Space considerations prohibit an elaboration on this point which is derived mainly from the Lacanian conceptualisation of the Law of the Father: the role and position of agency which is symbolically apprehended only through the language/power of the Father who intercedes in the Oedipal scene, laying down the Law (Lacan 1999). [return]
 Barbin states her place as a stranger in the world, but not as a stranger in the convent, and this is one of many examples (scattered throughout the entire memoirs) to suggest that Barbin did not see herself as out of place in the convent (Barbin 1980: 3). It was not the convent that was Barbin's problem; it was the prejudices of the larger world. [return]
 In fact, Roscoe notes that 150 Native North American cultures have been noted to have 'berdache' roles (Roscoe 1994: 330). [return]
 Future work in this area might profit from incorporating an analysis of the sex/gender organisation of oppression into an inquiry regarding the degree to which Western forms of feminism welcome a reciprocal political and academic relationship with gay and lesbian civil liberties movements as well as with transsex and transgender movements. For while the present climate indicates similar goals and frequent alliances between feminists and les-bi-gay activists, persistent forms of essentialisms on both sides maintain a certain animosity between feminists and trans movements. It may be that although women are generally subordinated to men in Western culture, they are reluctant to share space with trans movements due to a jealous fear of losing what power they do have. [return]
Barbin, Herculine. Herculine Barbin: Being the Recently Discovered Memoirs of a Nineteenth-Century French Hermaphrodite. Trans. Robert McDougall. New York: Pantheon, 1980; 1978.
Berdache Jordan. 'This 'Other': A Thumbnail History of Myself.' HealthyPlace (2004). http://www.healthyplace.com/communities/gender/intersexuals/about_me.htm.
Besnier, Nico. 'Polynesian Gender Liminality Through Time and Space.' Third Sex, Third Gender. Ed. Gilbert Herdt. New York: Zone, 1994, 285–328.
Butler, Judith. Bodies That Matter: On the Discursive Limits of 'Sex'. London and New York: Routledge, 1993.
——. Gender Trouble: Feminism and the Subversion of Identity, 2nd edn. London and New York: Routledge, 1999.
Douglas, Mary. Purity and Danger: An Analysis of the Concepts of Pollution and Taboo. London and New York: Routledge and Kegan Paul, 1966.
Elliston, Deborah. 'Erotic Anthropology: 'Ritualized Homosexuality' in Melanesia and Beyond.' American Ethnologist 22.4 (1995): 848–867.
Epple, Carolyn. 'Coming to Terms with Navajo Nadleehi: a Critique of Berdache, 'Gay', 'Alternative Gender', and 'Two Spirit'.' American Ethnologist 25.2 (1998): 267–290.
Foucault, Michel. 'Introduction.' Herculine Barbin. Herculine Barbin: Being the Recently Discovered Memoirs of a Nineteenth-Century French Hermaphrodite. Trans. Robert McDougall. New York: Pantheon, 1980; 1978, vii-xvii.
Herdt, Gilbert. 'Introduction: Third Sexes and Third Genders.' Third Sex, Third Gender: Beyond Sexual Dimorphism in Culture and History. Ed. Gilbert Herdt. New York: Zone, 1994a, 21–81.
——. 'Preface.' Third Sex, Third Gender: Beyond Sexual Dimorphism in Culture and History. Ed. Gilbert Herdt . New York: Zone, 1994b, 11–20.
——. 'Mistaken Sex: Culture, Biology and the Third Sex in New Guinea.' Third Sex, Third Gender: Beyond Sexual Dimorphism in Culture and History. Ed. Gilbert Herdt. New York: Zone, 1994c, 419–445.
Irigaray, Luce. The Irigaray Reader. Ed. Margaret Whitford. Cambridge, MA: Basil Blackwell, 1991.
Kaldera, Raven. 'Agdistis' Children: Living Bi-Gendered in a Single-Gendered World.' Looking Queer: Body Image and Identity in Lesbian, Bisexual, Gay, and Transgender Communities. Ed. Dawn Atkins.New York: Harrington Park, 1998, 227–232.
Lacan, Jacques. Encore: On Feminine Sexuality, The Limits of Love and Knowledge, 1972–1973. Trans. Bruce Fink. Ed. Jacques Alain Miller. New York and London: W. W. Norton, 1999; 1975.
Roscoe, Will. 'How to Become a Berdache: Toward a Unified Analysis of Gender Diversity.' Third Sex, Third Gender: Beyond Sexual Dimorphism in Culture and History. Ed. Gilbert Herdt. New York: Zone, 1994, 329–372.
——. ''Strange Craft, Strange History, Strange Folks': Cultural Amnesia and the Case for Lesbian/Gay Studies.' American Anthropologist 97.3 (1995): 448–453.
——. Changing Ones (2004a),
——. 'Slide Show', Changing Ones (2004b),
Rubin, Gayle. 'The Traffic in Women: Notes Toward a Political Economy of Sex.' Toward an Anthropology of Women. Ed. Rayna R. Reiter. New York: Monthly Review, 1975, 157–210.
Trexler, Richard. 'Making the American Berdache: Choice or Constraint?' Journal of Social History 35.3 (2002): 213–236
Weston, Kath. 'Lesbian/Gay Studies in the House of Anthropology.'
Annual Review of Anthropology 22 (1993): 339–367.