The fallacy of bounded identity

The critique of identity-based research on language and sexuality is initially enabled by a reductive representation of the field. Specifically, it ignores the broad body of linguistic scholarship on sexuality and focuses on a narrow subset of research on the linguistic practices of lesbians and gays. The framing of the field in this way requires the bracketing off of research that does not fit this preconceived notion of how sexuality has been studied. Thus, in his 2000 review article in the Annual Review of Anthropology, Kulick reports that the journal's editors had originally requested a review of research on transgender as well as gay and lesbian language use, but that he chose to limit the scope of his discussion to what he calls “gay and lesbian language.” He goes on to represent this topic as constituting the entire current field of language and sexuality.⁷

Such a move allows Kulick to lament the narrow focus of research on language and sexuality, yet simply incorporating the not inconsiderable body of work on transgender language or other issues in the same review would invalidate many of his claims regarding the field's limitations. That is, by his terming the object of his article “gay and lesbian language,” it is Kulick (not the researchers he cites) who in the first instance “reduces sexuality to sexual identity” (2000:246). By focusing only on identity categories, he excludes other integral facets of sexuality. Although Cameron & Kulick (2003a) put forth a broader view of language-and-sexuality scholarship, the critique of sexual identity continues to be directed against research on lesbian and gay language use. We have, of course, no objection to the discussion of lesbian and/or gay language use as one component of language-and-sexuality scholarship – indeed, such work has been integral to the field. Our concern, rather, is that to equate the one with the other is to erase the broad range of work that contributes to this area of research.

The reduction of the field of language and sexuality to “gay and lesbian language” allows critics to misrepresent linguistic research on minority sexual identities as being preoccupied with the search for a “linguistic code” (Kulick 2000:258) – that is, “a distinctive way of speaking and/or writing which serves as an authentic expression of group identity” (Cameron & Kulick 2003a:xiii–xiv). Some studies, to be sure, can be characterized in this way, particularly those from the earliest periods of linguistic research on lesbians and gay men, but the vast majority of current scholarship on sexual identity is informed by feminist and other critiques of such essentialist perspectives, and instead views identity as a variable and indexical phenomenon (see further below). This new wave of scholarship (belatedly acknowledged in Cameron & Kulick 2003a) sees the limitations of previous work not as an indication of the futility of identity research, but as an imperative to theorize the concept of identity more fully.

One reason that the goals of these latter studies have been misinterpreted is critics' failure to realize that terms such as “lesbian speech” and “gay male speech” are not necessarily put forward as lasting technical terminology for a relatively fixed lesbian or gay “code.” Rather, these terms are most often operationalized to describe language use in particular contexts: For example, “lesbian speech” in these discussions is simply a nominal construction of the uncontroversial proposition that “lesbians speak.”⁸

These critiques of identity research assume that, at some level, even theoretically sophisticated scholarship has succumbed to the “quest” (Cameron & Kulick 2003a:xiv) for gay and lesbian languages. They proceed to demonstrate why this quest is impossible by offering a set of principles for how such research would have to be conducted: “Any discussion that wants to make claims about gay or lesbian language must proceed through three steps. First, it must document that gays and lesbians use language in empirically delineable ways. Next, it must establish that those ways of using language are unique to gays and lesbians. Finally it must, at some point, define gay and lesbian” (Kulick 2000:259; cf. Cameron & Kulick 2003a:88–89). In short, if language is to be linked to sexual identities, it must be both shared and distinctive within a clearly bounded group. However, this view of identity has been extensively debunked in both early and recent sociolinguistic research, and especially in scholarship on language, gender, and sexuality. Instead, contemporary studies of identity are founded on the principles of variability and indexicality.

More than three decades ago, variationist sociolinguistics conclusively refuted the notion that language within a speech community must be shared by all members (Labov 1966). Later frameworks of analysis, such as linguistics of contact (Pratt 1987), communities of practice (Eckert & McConnell-Ginet 1992), and acts of identity (Le Page & Tabouret-Keller 1985), have demonstrated that the analyst's imposition of linguistic norms for speech community membership ignores the fundamental heterogeneity of even the smallest social group. Much recent research on gender and sexual identity, inspired by these and similar frameworks, analyzes the diversity of language use in what outsiders may see as a linguistically and socially uniform group (Bucholtz 1999b, Kiesling 1997, Mendoza-Denton 1999, Ogawa & Smith 1997). Linguistic homogeneity should therefore be viewed not as an analytic starting point but as an ideology that may become salient in social interaction across lines of individual difference.

The second problem raised by the critique of identity is linguistic distinctiveness. Critics have objected to researchers' analyses of linguistic features as markers of lesbian and gay identities. They note that other kinds of speakers may also use the same features; thus, such features cannot be said to be distinctively lesbian or gay. This position rests on the fallacy that linguistic forms must be uniquely assigned to particular identities in order to be socially meaningful. A simple solution to this apparent problem is offered by the semiotic concept of indexicality, first introduced into linguistic anthropology by Michael Silverstein (1976). Indexical signs are linguistic structures that point to (or “index”) aspects of the communicative context, such as social positionings. Specific linguistic forms can come to be ideologically associated with particular social identities indirectly by indexing interactional stances, like mitigation or assertiveness, that in turn index identities (Ochs 1992). This kind of indirect indexicality allows for the creation of multiple indexical links to a single linguistic form. Eckert (1999, 2002) illustrates this point with a long list of the varied and even contradictory social meanings that one phonological form, fully released [t], can have for different kinds of speakers of American English.

The insistence on difference as criterial for identity recalls the “sex differences” approach to language and gender. This early research tradition responded to the strongly male normative bias in linguistic research by arguing for the systematicity, functionality, and legitimacy of women's language use despite its apparent differences from male norms. Yet in calling attention to women's linguistic practices in contrast to men's, this body of scholarship opened itself up to charges of ignoring both important linguistic similarities between the genders and differences in language use within each gender (for an extensive discussion, see Crawford 1995). In the late 1980s and the 1990s, within language and gender studies the difference approach was largely supplanted by a deliberately noncomparative perspective that sought not merely to describe how women's language use differed from men's but to document the entire range of women's linguistic repertoires, often with attention to diverse ethnographic contexts (see contributions to Bucholtz et al. 1999, Coates & Cameron 1988, Hall & Bucholtz 1995, Penfield 1987).

This approach is also evident in recent studies of language and sexuality, and especially in research carried out within queer linguistics, which considers the relationship between language and identity to be locally specific. The point of such work is not to compare two groups in order to compute the linguistic and social difference between them, but to examine understudied groups on their own terms rather than as a foil for a more privileged group. Thus, just as descriptions of the linguistic practices of African American women do not inherently imply that these differ from the practices of European American women, African American men, or any other social group, neither do studies that examine the language use of lesbians mean to suggest that their speech is necessarily different from gay men's or straight women's. Such questions are simply not at issue in the vast majority of this research. The problem lies with readers who misconstrue such work as making comparative claims because they assume the priority of an “unmarked” group as the linguistic and social norm, from which all other groups must deviate (Barrett 2002, Hall 2003). To put this point in the simplest possible terms, the discovery that different groups may use similar linguistic resources for identity construction, far from vitiating the concept of identity, demonstrates the robust capacity to create new social meanings from existing linguistic practices.

That said, it is important to remember that the invocation of difference can be a powerful intellectual and political tool. The motivation for this discursive strategy lies in the need to remedy the historical underrepresentation of social groups. Such essentializing moves, known as “strategic essentialism” (Spivak 1995; cf. McElhinny 1996), are temporary overgeneralizations that are often necessary in the establishment of sociopolitical institutions such as research fields and political movements (see also Bucholtz 2003). For example, William Leap's (1994, 1996) early attempts to delineate features of what he calls “gay men's English,” at a time when linguistic research on sexual minorities was both scant and marginalized, may be seen as serving this function.

Another reason why the ideology of essentialism cannot be discarded is that social actors themselves use it to organize and understand identities. Thus, it is theoretically naïve to assert that linguistic practices associated by some researchers with lesbians or gay men are better viewed as

Such a statement ignores the sociopolitical force of ideologically constructed social boundaries that give symbolic meaning to linguistic practice – indeed, the very possibility of linguistic appropriation implied in the above quotation depends on some sense of linguistic ownership. Perhaps even more important, this statement overlooks the fact that speakers in different subject positions have differential access to linguistic resources. Research on language and political economy, inspired by Pierre Bourdieu's (1977) account of linguistic capital, has extensively documented that linguistic resources are unequally distributed across social groups. It is therefore implausible that such class-bound registers as economic or computer discourse may be available to “all speakers.”

This effort to dismiss the role of identity in lesbians' and gays' linguistic practices by likening them to specialized registers is seen in earlier critiques as well. Thus, Kulick, singling out the jargon of stamp collectors and wine enthusiasts, rhetorically asks, “In what ways is [such hobbyists'] language different? If it isn't different, in what way is the original observation that language constitutes identity not simply a banal platitude?” (1999a; cf. Kulick 2000:270). Similarly, one prominent participant at the 1999 New Ways of Analyzing Variation Conference in Toronto questioned Leap's concept of “gay men's English” by remarking on the ridiculousness of studying “masturbators' English” or “gas-mask fetishers' English.” Clearly, both statements are intended to trivialize any connection between language and sexual identity; while the first compares lesbian and gay identity to upper-class leisure activities, the second suggests that gayness and lesbianism are simply two more deviant sexual practices with no consequences for language use. This sort of dismissive rhetoric, which inhibited the study of language and gender for many years, is likely to have similar chilling effects on the field of language and sexuality. But setting aside the somewhat offensive reductionism implicit in the choice of parallels, such critiques neglect the fact that if “masturbators” or “wine enthusiasts” are indeed salient social categories in a particular culture that are associated with particular ways of speaking, then linguists might indeed do well to study them. In fact, the linguistic practices and associated identity positions of at least one of these social categories have already been investigated: Silverstein (2003) considers the register of oenophiles – wine enthusiasts – as part of the construction of an elite connoisseur “yuppie” identity.

The critiques of language and sexual identity cited above also presuppose that if linguistic forms are tied to social practices, whether hobbies or sexual acts, then they cannot have any interesting connection to identity. Identity, in this view, is construed as an internal essence, not a social phenomenon. But as many sociolinguistic researchers have shown, starting with the ground-breaking work of Penelope Eckert and Sally McConnell-Ginet (1992) on communities of practice, social practice is the very basis of identity. Moreover, while these critiques imply that “gay and lesbian language” is a “mere” register – that is, in traditional sociolinguistic theory, a situation-based variety – recent research in linguistic anthropology has demonstrated that registers index not only social situations but also the social groups associated with those situations (Agha 1998, 2003a, 2003b; see also Bell 2001). Such studies indicate that, regardless of how we want to classify any given set of socially meaningful linguistic practices – as “registers,” “styles,” “varieties,” “dialects,” or “languages” – indexicality works the same way: In every case, language users both draw on and create conventionalized associations between linguistic form and social meaning to construct their own and others' identities.

The problems raised by critics of identity research are thus neither as widespread nor as insurmountable as these writers suggest. Moreover, the proposed replacement for identity – desire – suffers from a far more serious set of problems, both conceptual and methodological.

Difficulties of desire

Any approach to sexuality that privileges desire over identity has substantial limitations; however, the particular version proposed by critics of research on language and sexual identity has special difficulties. These fall into three categories: the definition of the key terms “sexuality” and “desire”; a general adherence to psychoanalysis as a theoretical framework; and analytic assumptions that lead to decontextualized interpretations of linguistic data.

As discussed above, those hostile to identity-centered research on language and sexuality wrongly maintain that such research has restricted the concept of sexuality to sexual identity. However, even though their proposed alternative definition of “sexuality” is represented as recovering “the broad meaning [the term] was intended to have,” in fact it reduces the concept to no more than “the socially constructed expression of erotic desire” (Cameron & Kulick 2003a:4). Besides its omission of identity, such a definition excludes what is unquestionably the most basic element of sexuality: reproduction. Although erotic desire is often attached to reproduction, it need not be. Abortion and birth control, pregnancy and childbirth, kinship, and even the act of reproductive sex itself are not necessarily erotically motivated. For this reason, our own definition of sexuality, offered at the outset of this article, encompasses both eroticized and reproductive meanings of the body as realized within sociopolitical matrices of ideology, practice, and identity.

If “sexuality” is defined too narrowly in these critiques, the other key term, “desire,” is defined too vaguely. The definition of “desire” shifts considerably across various publications. In the first version of Kulick's critique, rather than a definition, the reader is confronted with a series of questions:

The characterization of desire in a subsequent publication as “iterable practices that can be mapped” (Kulick 2003a:130) is no more helpful, for it extends the concept so broadly that it encompasses practically all social phenomena and necessarily includes any use of language. Phenomena as different as agoraphobia and physical hunger are presented as examples of “how desires are socialized” (Kulick 2000:130; cf. Cameron & Kulick 2003a:120–121). These issues are indeed vitally important to sociocultural linguistics, but when desire comes to have infinitely expandable scope, moving from fucking to fear to food in the space of a few pages, it loses its relevance as a theoretical replacement for sexual identity.

What is remarkable about these proposals is that even as they are repudiating identity-based research, the discussion of desire itself relies heavily on the concept of identity. Identity is often an invisible presence in these texts, but at some points it surfaces under the label “desire,” as in the following passage: “Work done on language and transgenderism (for a summary, see Kulick 1999; for examples, see Hall & O'Donovan 1996 [sic], Livia 1995, 1997a) and on language and ‘passing’ (e.g., Bucholtz 1996) highlights how desire (to be, or be seen as, a woman; to be, or be seen as, Hispanic) is structured linguistically” (Kulick 2000:276, n.15). Here, desire is no more than desire for identity. Such efforts to smuggle identity into the desire framework demonstrate that sociocultural research on language and sexuality cannot, as its critics recommend, “leave [identity] behind and forget about it for a while” (Cameron & Kulick 2003a:105).

A second difficulty caused by the privileging of desire is that it leads critics of identity to commit themselves to psychoanalysis as a central theoretical tool for linguistic investigations of sexuality. Various forms of psychoanalysis have extensively theorized desire, and especially sexual desire. Its advocates within sociolinguistics, therefore, look to this theoretical paradigm as a key source of information about “fantasy, repression, pleasure, fear and the unconscious” – in short, “everything that arguably makes sexuality sexuality” (Cameron & Kulick 2003a:106). Building on the psychoanalytic theorist Jacques Lacan's famous statement that “the unconscious is structured like a language” ([1972–73] 1998:48), the desire-centered approach to language and sexuality urges sociolinguists to turn their attention to the unconscious – that is, to the unspoken urges that govern erotic expression. None of these critiques, however, provides an adequate model of how such inarticulateness can be studied empirically. At times, they seem to require sociolinguists, linguistic anthropologists, and discourse analysts to reinvent themselves as field psychoanalysts, ascribing repressed desires to those they study. As shown below, such an undertaking has encouraged interpretations of data that are more accountable to psychoanalytic theory than to the empirical details of the social context.⁹

Another way in which the desire framework sets aside social context, at least in some of its manifestations, is in its reliance on the very “code” view claimed to be characteristic of research on sexual identity. In the most extreme formulation of this view, the linguistic expression of desire is represented as necessarily distinctive in form. Thus, referring to a volume entitled Language and Desire (Harvey & Shalom 1997a) as exemplary of the paradigm he favors, Kulick cites approvingly the stated principle motivating this body of research: “The encoding of desire results in distinct and describable linguistic features and patterns” (Harvey & Shalom 1997b:3; cited by Kulick 2000:276). We are confronted here with a striking contradiction. Whereas Kulick forcefully and repeatedly argues against the possibility of associating any distinctive formal features with sexual identities, he willingly accepts the assumption that distinctive formal features attach to sexual desires. Thus, the interrogation of language and identity cited above can be recast in such a way as to turn the interrogation back on the questioner: In what ways is desirous language different? Our own view, as suggested above, is that language use need not be distinctive to construct sociocultural meanings. Yet the focus on psychoanalysis may blind researchers to the crucial role of social context in creating such meanings.

We now turn to three linguistic studies that have been held up as models for the study of language and desire. In each case, it is only by stripping away social context that the researchers are able to frame desire as having greater analytic interest than identity.

Kulick 2000 singles out two studies (Channell 1997 and Langford 1997) as the standard-bearers for linguistic analyses of desire; he describes them as “fine examples of how such accounts might proceed in practice” (2000:276). However, both studies go against the grain of sociocultural linguistic principles in suggesting that language can and should be analyzed independently of social context. Johanna Channell's decision to stay out of what she calls “the identity debate” (1997:144), for example, leads her to analyze the sensationalized telephone conversation between Charles, Prince of Wales, and his mistress, Camilla Parker-Bowles, as just another telephone call between lovers. In her own words, Channell chooses “to ignore the issue of the participants' identity and to deal only with the text, seeing it as a celebration of a close and loving relationship between an anonymous man and a woman” (1997:162). She then uses this decontextualized conversation to identify specific textual features that constitute what she calls “the language of love and desire” (1997:143), such as expressions of emotion, repetition, playful language, and prolongation of the conversational closing. But equally relevant to this interaction is the fact that the lovers in question are members of the extreme upper class whose love affair was illicit and was repeatedly condemned by both the royal family and the British press. Channell takes at face value the press's representation of the conversation as silly, even childish, using it as straightforward evidence for her claim that love talk demonstrates distinctive affective linguistic features. Reading the press commentary as descriptive and not judgmental, she misses the important observation that Prince Charles was in a sense “queered” (that is, rendered sexually deviant; see further below) by the press for not following the normative rules of love and sexuality associated with the royal family. British journalists' over-the-top depictions of Charles's speech as babyish and immature have as much to do with popular disgust at the prince's refusal to “grow up” and embrace state-sanctioned royal kinship as they do with the speech itself. Consideration of such cultural and political specificity – particularly the central role of social class relations in this situation – problematizes Channell's quest for a generalizable language of desire.

The second study held up as a model of research on language and desire is Wendy Langford's (1997) analysis of the creation of alter personalities by participants in romantic relationships. This study too relies on a “code” view of desire, in that Langford strives to develop a theory of alter relationships as bounded “micro-cultures” in their own right, having “their own languages and customs” (1997:170). Calling attention to the use of animal nicknames for loved ones in a variety of texts – specifically, two of Henrik Ibsen's plays, a short story by Virginia Woolf, Valentine's Day messages appearing in the British newspaper the Guardian, and two personal interviews – Langford asks, “How can we account for the fact that adults create ‘in the privacy of their own homes’ such extensive secret cultures, worlds of escape where they play like children, and which are remarkably similar to worlds created by other adults with whom they have no direct communication?” (1997:177). The author ultimately finds her answer in psychoanalysis, proposing that animal alter personas function as a kind of “transitional object” for the couples who use them, in much the same way that in Donald Winnicott's ([1953] 1971) psychoanalytic formulation, a teddy bear functions as a transitional object for a child moving toward greater independence. But Langford's universalizing account of why the adult couples in her study create similar personas of fur and fluff discounts any potential influence of culture. Assuming that “direct communication” is the primary way that cultural knowledge circulates, she concludes that these couples, because they have never met, must be driven by psychoanalytic universals.

But even in her search for psychologically oriented explanations, Langford reveals a class-based cultural bias: “How can we explain why grown adults capable of such sophisticated interests as opera, astrophysics, and advanced linguistic theory should negotiate their love relationships in ‘intermediate areas’ characterised by such cultural primitivism?” (1997:181). Disregarding the classist assumptions embedded in Langford's belief that “primitive” cutesy names are incompatible with higher educational status, we nevertheless must caution against the formulation of a general theory of desire from a highly selective sample. The textual examples used as evidence for her psychoanalytic claim undoubtedly originate in the upper classes, as she herself implies in the question just quoted, yet nowhere in the article are the data acknowledged as class-bound in any way.¹⁰

Although the data in both articles are extremely rich in potential insights for a broad field of language and sexuality, the analytic tack taken in each case prematurely forecloses the most interesting and important aspects of these research situations. This is not necessarily a problem of methodology (for example, the conversation-analytic approach taken by Channell has been extremely fruitful for feminist research; e.g., Stokoe & Weatherall 2002). However, when researchers assert a universality to their data that crosscuts class, ethnicity, gender, sexuality, and nation, as these authors do, they sacrifice the contextual delicacy of sociocultural approaches to language.

A third study, Cameron's (1997) analysis of heterosexual masculinity, is far more sensitive to the central role of identity in the construction of sexuality. The vital importance of identity in the analysis is indicated by the title itself, “Performing gender identity.”¹¹

Yet the 2003 reframing of this article as exemplary of the desire paradigm allows theory rather than data to drive analysis. In this case, data that problematize the theory are literally written out of the discussion. Despite the fact that in the original data the speakers represent their classmate as heterosexual (or at least as flirting with a woman), this portion of the transcript is omitted in the later summary of the article, which asserts that “nothing is said about the ‘really gay guy's' sexual preferences or practices” (Cameron & Kulick 2003a:66). The omission of this information permits the authors to attribute the speakers' performance of heterosexual identity to their fear of “a danger that cannot be acknowledged”: their own homosexual desire (Cameron & Kulick 2003a:122). What is actually “unsaid” here, however, is the data that complicate this claim, for if the classmate is perceived to be heterosexual, homosexual desire is not obviously at issue. The specter that haunts this interaction is not necessarily the speakers' repressed homosexual desire but, more basically, the threat to the gender order. Some sort of social distancing clearly motivates these college students’ display of masculinity, but there is no empirical evidence that it is based in a fear of their own sexual desires and not in the maintenance of gender hierarchy. Although hegemonic masculinity is often constructed through the rejection of homosexuality, any such interpretation must be grounded in the sociocultural context. This grounding is lost when discourse analysts get into the business of ascribing inner states rather than describing social relations.

Feminism and the critique of queer theory

Despite the concerns noted above about the reinterpretation of Cameron's work in terms of desire, her entry into the project of developing the language and desire framework has greatly improved its level of theoretical sophistication. Most important, both gender and feminism figure centrally in the most recent formulation of this paradigm (Cameron & Kulick 2003a). However, in preced ing accounts (Kulick 1999a, 1999b, 2000, 2003a, 2003b), feminism remains a largely “repressed term” in the discussion.¹²

The argument against the linguistic study of sexual identity and in favor of the linguistic study of desire leads its proponents – wittingly or unwittingly – into the very heart of the debate between feminism and queer theory over the nature of sexuality. Where feminism has long been committed to the analysis of sexuality as embedded within structures of power, certain strands of queer theory turn aside from a full analysis of power to focus more narrowly on the subversive potential of nonnormative sexual practice. In this context, the call for a “sexy linguistics” (Kulick 1999a) works ideologically to situate researchers on language and sexual identity as old-fashioned and anti-sex. The diagnosis that language-and-sexuality research has an “unhappy fixation on identity” (Kulick 2000:272) conjures up the age-old image of the frowning feminist talking about un-fun things. This rhetorical trope is also employed in Stephen Murray's (1999) review of the edited collection on language and sexuality Queerly phrased: Language, gender, and sexuality (Livia & Hall 1997a). Bemoaning the fact that “sexual practices” (1999:304) are not the primary focus of a book that includes the word “sexuality” in its subtitle, Murray likens one of its (female) contributors to a “Victorian-era writer” (1999:307) and laments what he sees as the author's naïveté with respect to sexual pleasure.

As feminist critics have discussed at length, this discursive strategy has been used effectively for decades as a weapon against feminist politics and theorizing (see, e.g., Garber 2001; Huffer 2001; Jeffreys 1985; Martin 1996, [1994] 1996; Penn 1995; Soper 1993; Zita 1994). It worked to create the asexual bluestocking caricature of First Wave feminism in the late 19th and early 20th centuries, the prudish if not puritanical “anti-sex” radical of the post-Second Wave sex wars in the 1980s, and the stodgy feminist of the 1990s and today, whose attention to gender spoils the rhetorical fun of a liberated queer theory. In a discussion of the last of these characterizations, Lynn Huffer (2001:8) explicates what she calls the “paradigmatic feminist versus queer opposition” and offers a succinct summary of how queer theory has positioned feminism: “In the new paradigm, the queer performative presents itself as correcting and surpassing a heteronormative feminist narrative about gender; at the same time, feminist narrative becomes the unsexy but ethically superior guardian of a moral order threatened by the libidinal excesses of the queer.” In other words, queer theory restricts feminism to the sexless study of gender identity, while claiming for itself the ground of sexual pleasure.

It is this false opposition between desire and identity that is of particular relevance to the present discussion. The dichotomy between these concepts that is set up in the desire paradigm of language and sexuality threatens to replicate within our own field the unfortunate polarization between certain versions of queer and feminist theory that began in the early 1990s. Each of these theoretical traditions has its own strengths for the study of language and sexuality. The poststructuralist feminist theorist Judith Butler, whose 1990 book Gender trouble is often viewed as a foundational text of queer theory, has since criticized the territorial separation between feminist and queer theory and situates her own work between these two positions (Butler 1997). Certainly, both perspectives have their share of weaknesses: While heterosexist assumptions frequently underlie theorizing in feminism (and in language and gender studies as well), an unconscious sexism often structures argumentation in queer theory (and, we fear, may do the same in some studies of language and sexuality). But the very project of dichotomizing theory into mutually exclusive “either/or” camps is reductive and yields simplistic analysis. Queer theorists' labeling of desire as a theoretically sexy concept and identity as a dowdy one, for example, leads to inevitable misreadings of how identity is currently theorized. Such is the case when Michael Warner, the reigning czar of queer theory, in an interview with Annamarie Jagose (2000), criticizes what he calls the “unsexy” politics of Urvashi Vaid, former executive director of the National Gay and Lesbian Task Force. Warner accuses Vaid of failing to embrace the perspective of queer theorists like Leo Bersani, who argue that nonnormative sexual practice, not identity, should be the focus of a queer politics.¹³

A focus on desire without feminism is entirely consistent with a strand of queer theory that Donald Morton has identified as “‘queer’ Queer Theory” (1993:139–41), espoused by high theorists such as Warner and Bersani. “Queer” Queer Theory, according to Morton, celebrates sexual practice as pleasure and sets aside uncomfortable issues of power – issues that are often intimately connected to gender. Morton argues that all of queer theory is guilty of this tendency: “Thus Queer Theory in all its variants works to displace gender as a category in an analytical, concept-based materialism … with sexuality as libidinalism so pleasure-saturated that it can disrupt and indeed dispense with conceptuality and analytics … and thereby produce another space of freedom for the bourgeois subject” (1993:141). In our view, Morton's position is too strong, for some versions of queer theory offer important insights for the feminist study of sexuality, particularly the view of queer subjectivities as taking their meaning not from fixed identity categories but from historically and culturally specific heteronormative structures. However, we share his concern that a focus on sexual desire may marginalize issues of gender, power, and agency.

The theoretical framing of sexuality as libido often takes inspiration from psychoanalysis, a paradigm that, as noted above, is antithetical to feminist concerns in many of its manifestations. Within linguistics, this problem is illustrated by the decision to ground the desire approach to language and sexuality in a smattering of grand theories whose patriarchs range from Freud to Lacan to Deleuze. In ignoring other, more contemporary, perspectives – as well as the extensive feminist revisions of psychoanalysis (e.g., Benjamin 1998, Chodorow 1989, Irigaray 1985, Kristeva 1980) – the recommendation to “dump” identity begins to sound a bit Oedipal: Kill the father, marry the mother, or in this instance, kill the mother (feminism), marry the father (Freud, Lacan, or some other celebrant of masculinist theory). At times, the fascination with psychoanalysis seems less concerned with offering a coherent framework of desire than with ousting feminism in favor of some theory, any theory, that restores male-centered sexuality to its formerly dominant position: “We don't have to swallow Lacan's phallus, as it were, but we should respect the fact that he raises it” (Kulick 1999a). Such remarks suggest that the critique of the linguistic study of identity is implicitly a critique of the hesitancy of prudish linguists to embrace the “sexiness” of psychoanalytic theory: Enough of this talk about subjectivity and power, Virginia. Can we please get back to the phallus?

In contrast, feminist approaches to language and sexuality are centrally concerned with locating sexuality with respect to regimes of power and speakers' agency in relation to these. There is little to find titillating in Lisa Ryan's (2002) analysis of the discursive positioning of rape victims in Irish court judgments, Kathleen Wood's (1999) explication of the ways in which heterosexist ideologies shape the production of Deaf and hearing lesbians' coming-out stories, or Anita Liang's (1999) discussion of how lesbians and gay men linguistically shield their sexuality from hostile heterosexuals. Such research, and power itself, have only recently been discussed by advocates of a “desire”-based approach to language and sexuality – a fact that suggests that because these studies are decidedly “unsexy,” they have been seen as not really about “sex” at all.¹⁴

An instructive example in this regard is found in the anthropological literature on sexuality and desire in Brazil. In her book Laughter out of place, Donna Goldstein (2003) critiques the way in which anthropologists have historically represented Brazilian sexuality as transgressive, liberatory, and carnivalesque. Naming this focus “the carnivalization of desire” (2003:231), Goldstein illustrates how scholarship on Brazil's “sex-positivism” has worked to silence other, less playful discussions of sexuality – namely, the feminist critique of normative heterosexual relations. For example, while scholars have tended to explain male homoeroticism in Brazil as derivative of a sexual culture that encourages transgression, Goldstein points out that transgression is itself gendered, with men expected to act as transgressors and women as “boundary-setters.” That women frequently have to protect their daughters from the transgressive sexual behaviors of stepfathers is a recurrent theme in the stories of the Rio de Janeiro shantytown women of Goldstein's ethnography, who because of the local valorization of sex-positive language and attitudes are hesitant to speak critically about violent male sexuality. But their use of sexual humor, though superficially sex-positive, reveals a counterdiscourse of sexuality that works agentively to critique dominant understandings of transgression and desire.

Goldstein's point is multilayered, as she suggests that sex-positivism functions as a restrictive ideology not only for these shantytown women, but also for feminist scholars writing about sexuality in Brazil. What Goldstein calls the “masculinist vision of desire and transgression” (2003:228) that governs scholarship on Brazilian sexuality discourages a sustained critique of the material ways in which dominant discourses of sexuality reproduce unequal power arrangements, to the extent that feminists undertaking such analyses are often dismissed as “sex-negative.” Goldstein's research offers an important reminder that the celebratory representation of sexual practice obscures more troubling aspects of sexuality, such as its relation to power.

Systems of gender, systems of sexuality

Because most theories of sexuality that focus on desire do not engage fully with feminism, they also fail to acknowledge completely the importance of gender in studying sexuality. In linguistics, when proponents of a desire framework do invoke feminism, it is often to justify the marginalization of gender in their theories. For example, in a Derridean moment, Kulick 1999a maintains that the very term “gender” must be placed “sous rature” (‘under erasure’; that is, visible as a trace of a former presence). In support of this position, Kulick points to Gayle Rubin's ([1984] 1993) teasing apart of sexual and gender oppression as interrelated but distinctive phenomena. Although we energetically agree with Rubin's call to develop theories and politics focused on sexuality (this is, indeed, the point of Livia & Hall 1997b), we also want to emphasize the subsequent claim made by scores of feminist theorists – and by Rubin too, even in this early essay – that sexuality, like gender, materializes only within a complex array of social relations. As Rubin notes, “In the long run, feminism's critique of gender hierarchy must be incorporated into a radical theory of sex, and the critique of sexual oppression should enrich feminism” ([1984] 1993:34). In the two decades since Rubin wrote these words, feminists have sought to uncover the power relations specific to race, class, and nation as well as to gender and sexuality. At the same time, they have demonstrated that none of these relations can exist independently of the others.

Rubin's distinction between systems of sexuality and systems of gender relies heavily on Michel Foucault's treatment of sexuality as the product of historical medical discourses. In The history of sexuality, the Ur-text that enables Rubin's theoretical separation and the subsequent development of queer theory, Foucault [1978] 1990 charts the emergence of sexual identity through discourses of sexuality, not gender. Rubin's framework can therefore be problematic for anthropological research on cultures in which the discourses of sexuality Foucault identifies for the Western world simply do not exist and sexuality instead relies on structures of gender. For example, in places like Thailand, as Jackson (2003) points out, sexual behavior, and hence the vocabulary used to describe it, are imagined through a “gender-based eroticism.” Because in Thailand even newly introduced identities such as “gay” are articulated through a discursive regime of gender rather than sexuality, Jackson argues that Rubin's thesis is peculiarly Western-centered. Recent research in linguistic anthropology has documented a number of liminal communities that practice sex through the assumption of eroticized gender roles, including Rudolf Gaudio's (1997, 2001) research on the ‘yan daudu in Nigeria, Niko Besnier's (2003) work on fakaleitı in Tonga, Hall's (1997, forthcoming) studies on hijras and kotis in northern India, and Kulick's (1997) article on Brazilian travestis. The fact that these communities frame sexual desire within a gendered dichotomy of “passive” and “active” or “feminine” and “masculine,” as is the case in many areas of the world, requires a theory that encourages investigating the many ways that gender and sexuality may be discursively intertwined. Such a theory must equally facilitate research on how sexuality interacts with race, social class, and other factors, relationships that are also illustrated in the above studies. Although the possibility of these relationships should always be entertained at a theoretical level, their existence in any local context can be settled only empirically, and usually ethnographically.¹⁵

For this reason, the dichotomy between a sexually “exotic” non-West and the “modern” West is flawed. In this sense, Jackson's critique of Rubin's Atlanticocentrism carries its own share of universalizing assumptions, particularly when he suggests a too simplistic division between sexual desire in the West (as based on discourses of sexuality) versus sexual desire in the non-West (as based on discourses of gender). Jackson makes the provocative claim, for instance, that whereas a male in Thailand is homosexualized by his femininity, a male in the West is feminized by his homosexuality. But Cameron's (1997) analysis of men's gossip discussed earlier shows that males in the West may also be sexualized through gender: It is because of gender practice rather than sexual practice that the speakers she discusses ridicule their classmate as “gay.” That the discursive construction of heterosexuality is often bound up with the discursive construction of femininity and masculinity is by now a familiar finding, reported by a number of linguists working on a variety of diverse communities in both the United States and the United Kingdom (e.g., Bucholtz 1999a, 1999b; Coates forthcoming; Eckert 2002; Hall 1995; Kiesling 2002; Limón [1989] 1996; Mendoza-Denton 1996). If researchers insist that sexuality be analyzed in isolation, whether in Europe, North America, Asia, or anywhere else, they run the risk of reading it through a theoretical lens that may be only partially revealing, at best.

Kulick's (1998) anthropological analysis of the identities of Brazilian travestis, or transgendered prostitutes, illustrates what can be lost in an artificial separation of gender and sexuality, particularly when this separation imposes a single theoretical model onto a situation that demands a more multilayered analysis. The conversations reproduced in the first four chapters of Kulick's book work together to create an extremely complex portrait of travesti identity, one that paints travestis' gendered subjectivity as less about women and men than about the embracing of a sexually “passive” or “active” identity. Travestis do not consider themselves to be “men”; their display of femininity is also a display of their preference to receive instead of give anal penetration, which they see as a sexual practice of “men.” In fact, they speak disparagingly of homosexuals, who “‘hide’ their inherent femininity [i.e., preference to be penetrated] under the masquerade of what travestis consider to be male drag” (Kulick 1998:232). Their talk reveals this gendered perception of self and other in multiple ways: in their understanding of their boyfriends as heterosexual, not homosexual; in their decisions to leave male partners who want to be penetrated; and in their consistent use of feminine reference for themselves and masculine reference for their boyfriends (see also Kulick 1997).

Yet in spite of this rich ethnographic information, Kulick argues that travestis are first and foremost “homosexuals” – a term that misleadingly represents their identity as predominantly structured on same-sex attraction rather than on eroticized gender roles: “Being homosexual is at the heart of the travesti project” (1998:221); “homosexual desire is the current that buoys up and gives meaning to travestis' bodily practices, their affective engagements, and their professional activities” (1998:221); “travestis consider themselves to be homosexual desire in its fullest and most perfect form” (1998:225). Herein lies the contradiction: Kulick analyzes the travestis as homosexual, yet they themselves view their relationships as necessarily involving two different genders, “men” (penetrators) and “not-men” (those who are penetrated) (1998:229). In fact, there is only one example in the book in which a travesti actually uses the term “homosexual” (Brazilian Portuguese homossexual) in reference to herself (1998:222), and she does so to describe how she felt before she adopted a more enlightened travesti identity. The term viado, which travestis often employ in reference to themselves, is not comparable to the English term “homosexual,” for its primary meaning, to a travesti at least, has to do with an erotic effeminacy that is other-directed, not same-directed.¹⁶

Queer linguistics

The discussion above demonstrates the importance of including feminist insights on gender and power in the analysis of sexuality. It is for this reason that we advocate queer linguistics as a framework for linguistic scholarship on sexuality. Three main theoretical bodies of work – queer, sociocultural linguistic, and feminist – contribute to the project of queer linguistics. While these intellectual traditions are closely interconnected, it is useful to focus on the strengths of each and what these can offer to the linguistic study of sexuality.

Queer linguistics follows queer theory in refraining from assigning a fixed, categorical meaning to “queer.” In specific situations, the (temporary and situated) meaning of “queer” emerges at the excluded margins of historically and culturally variable heteronormative systems. This meaning is arrived at analytically through principles established by sociocultural approaches to language, such as ethnography, that foreground the importance of local understandings and contexts. As shown above, feminist theory further contextualizes both queer and hegemonic sexualities in relation to other sociopolitical phenomena, among them gender, race, and social class. The combination of these approaches gives rise to a broad and critically nuanced view of sexuality.

The relational understanding of the term “queer” means that queer linguistics does not restrict itself to the study of homosexualities, but rather considers the full range of sexualized identities, ideologies, and practices that may emerge in specific sociocultural contexts. As Rusty Barrett explains, “The goal … of queer linguistics cannot be the study of the language of a pre-defined set of ‘queers’ (since such a set cannot be defined), but rather a linguistics in which identity categories are not accepted as a priori entities, but are recognized as ideological constructs produced by social discourse” (2002:28). It is useful to draw an analogy with language and gender scholarship. Feminist linguistics is to language and gender as queer linguistics is to language and sexuality: a particular political and theoretical perspective on an area of intellectual inquiry. It follows that queer linguistics is not simply the study of the language of “queers” (however that term is defined), any more than feminist linguistics is simply the study of the language of feminists.

If there is confusion on this point, it may be because not all linguists working on sexually liminal communities – or even all queer theorists for that matter – have employed “queer” in a theoretically consistent way. Part of the problem is that in the United States the term has been embraced by lesbian and gay political activists as a fixed identity designation for groups marginalized by cultural prohibitions against same-sex desire. In these political discussions, the term is most often used as shorthand for the more unwieldy collocation “lesbian, gay, bisexual, and transgender,” a usage that has led some scholars to think that “queer” refers only to these categories (see also McConnell-Ginet 2002). However, as the above example of Prince Charles's extramarital affair indicates, heterosexuals too may be positioned as queer when they fall outside normative structures of sexuality.

The queering of heterosexuals has been pursued mostly in nonlinguistic research on sexuality. The cultural studies scholar Cathy Cohen (1997), for example, remarks that African American women and men who were prohibited by the government from marrying during the period of U.S. slavery were just as “queer” (that is, sexually marginalized) as present-day gays and lesbians, sharing with them an exclusion from state-sanctioned systems of reproductive kinship. Indeed, as Cohen suggests, the contemporary unmarried African American mother is also in some sense “queer,” positioned derogatorily by popular and governmental discourses alike as a “welfare queen” who consumes state resources through unruly and unregulated reproduction. Another example of queer subjectivities is provided by Bonnie McElhinny (2002), who cites the case of a single mother who was charged with child abuse when she sought advice about her feelings of sexual arousal while breastfeeding her toddler. Situations like those described by McElhinny and Cohen reveal “the ways in which processes of exclusion consign certain heterosexuals to a place outside the heteronormative center” (Huffer 2001:17; emphasis in original). What “queers” the subject of both queer linguistics and queer theory, then, is not sexual orientation but sexual marginalization.

The relationality of “queer” is central to another concept that is fundamental to queer linguistics: performativity. This concept has been shaped by all three of the theoretical traditions that contribute to queer linguistics. Butler 1990 adapted the idea of performativity into poststructuralist feminism and queer theory (via Derrida [1972] 1982) from the notion of the performative within the philosophy of language (Austin 1962; for fuller discussion see Hall 2001). Performative utterances – such as We find the defendant not guilty or Class dismissed – do not merely describe the world, but change it; that is, performatives are linguistic social action. However, performatives are effective only when appropriate conditions hold (e.g., the institutional authority of the speaker, the correct execution of the utterance, a conventional association between the utterance and its effect).

For Butler, gender is performative in the sense that it is brought into being through linguistic and other semiotic practices. She argues that gender performances are effective (that is, culturally intelligible) only when they meet socially imposed norms or regulations of gender-appropriateness. The repetition of such performances reproduces the regulation of gender, setting the limits on culturally admissible gender practices. And because gender norms are often tied to sexual norms, the performance of gender has consequences for sexuality as well. Yet some performances may involve a deliberate violation of gender norms, such as drag, or the ironic appropriation of gender practices. This disjunction between norm and performance creates a space for the subversion of gender and/or sexual identity. Butler's theory of performativity has been viewed by some feminists as granting too much agency to individuals to take on different gendered and sexual subjectivities. Such critiques often rest on a theatrical interpretation of “performance.” But in Butler's theory, “performance” denotes not acting but enacting – not a (false) representation of a “true self” but the production of oneself as a culturally recognizable (i.e., gender-normative) subject. As a consequence, other feminists argue that Butler deemphasizes the agency of speakers and hearers. They criticize her for considering subversive performances primarily as they relate to global gender and sexual ideologies rather than as local interventions into these ideological systems.

Several aspects of performativity are useful in the linguistic analysis of sexuality. In particular, the concept challenges the notion that either gender or sexuality is “natural” by maintaining that both acquire social meaning only when physical bodies enter into historically and culturally specific systems of power. Butler's feminist commitment is also vital to the politically responsible study of language and sexuality, because she acknowledges the theoretically fundamental issues of gender and sexual hierarchy in the analysis of sexuality.

At the same time, as Livia & Hall (1997b) observe, Butler's framework requires substantial adaptation for use by researchers of language and sexuality. Most critically, her account of the subject's relation to regulatory systems of power is highly abstract and general. Although the concept of performativity is derived from theories of language, Butler is not interested in analyzing how sexuality emerges in sociolinguistic contexts. To remedy this situation, queer linguistics turns to sociocultural linguistic theories of ideology, practice, and identity, which can be used to anchor the study of sexuality. These concepts, together with sociocultural methods of contextualized linguistic analysis, especially ethnography, give us a much richer picture of how sexuality is produced in language.

Sociolinguistics, linguistic anthropology, and socially oriented discourse analysis emphasize that language is the mediating level between structures of power and human agency (see discussion in Bucholtz & Hall 2003). Language is a primary vehicle by which cultural ideologies circulate, it is a central site of social practice, and it is a crucial means for producing sociocultural identities. Thus, any sociocultural analysis of language is incomplete unless it acknowledges the relationship between systems of power and the ways that they are negotiated by social subjects in local contexts. Sociolinguistic research has demonstrated that ideologies are not only imposed from above but are also built on the ground, in sociocultural and interactional practice. Moreover, the practices of social actors not only reproduce ideologies but also challenge them. The relationship between any given practice and any given ideology must be arrived at empirically, through close attention to its structure and the context in which it is produced. In queer linguistics, these insights allow us to talk about performativity as a phenomenon tethered to local as well as broader sexual identities, ideologies, and practices. We have discussed practice theory above; in the next section we address linguistic theories of ideology at greater length, and focus especially on the sociolinguistic theorizing of identity.

In the broad definition of sexuality proposed at the outset of this article, the ideologies, practices, and identities that make up sexuality are conceptualized as mutually constitutive. However, as discussed earlier, the study of language and sexual identity, even within a queer linguistics framework, has been misunderstood by critics as essentialist at some level. This perception stems from the mistaken belief that identity is necessarily inherent, individual, and intentional. However, when we view identity as an outcome of intersubjectively negotiated practices and ideologies, it becomes clear that such a characterization is inaccurate. Identity cannot be inherent if it is the emergent result rather than the preexisting source of social actions; it cannot be individual if it is socially negotiated; and it cannot be fully intentional if it is produced by practices and ideologies that may exceed conscious awareness.

The remainder of this article demonstrates how such an understanding of identity enriches the study of language and sexuality, and sociocultural linguistic analysis more generally. The view of identity as inseparable from ideology and practice is one component of queer linguistics, but such a view is not limited to that approach. Likewise, the model of identity that we describe below is illustrated using examples from research on language and sexuality; nevertheless, it is equally applicable to any study of social relations, whether along “macro” sociological lines such as gender, class, and race, or with respect to more local and temporary subject positions.¹⁸

Adequation and distinction

The first pair of tactics, adequation and distinction, are processes by which subjects construct and are constructed within social sameness and difference. The term “adequation,” by combining the notions of equation and adequacy, captures the fact that intersubjective similarity is not complete, but sufficient. Thus, adequation is a tactic that establishes what might be called “sufficient similarity”. By contrast, “distinction” is the production of social differentiation. Generalizing from Bourdieu's (1984, 1991) notion of distinction as a strategy used by the bourgeoisie to separate themselves from the lower classes through elite aesthetic preferences, we understand this tactic more broadly as any process that creates social boundaries between groups or individuals.

What unites adequation and distinction as complementary processes of identity formation is their dependence on the erasure of ideologically discordant elements. Erasure is one of three semiotic processes identified by Susan Gal and Judith Irvine (Gal & Irvine 1995, Irvine 2001, Irvine & Gal 2000) in their extremely productive framework for analyzing language ideologies. For them, erasure is “the process in which ideology, in simplifying the sociolinguistic field, renders some persons or activities (or sociolinguistic phenomena) invisible” (Irvine & Gal 2000:38). Here we want to build on this work in several ways. First, we want to bring to the surface a point that is implicit in Gal & Irvine's research: that what is often created through erasure and the other semiotic processes of ideology is precisely identity. For example, erasure is fundamental to the identity relations of adequation and distinction, insofar as the enactment of each requires the effacement of its counterpart. In other words, the construction of sameness demands the obscuring of difference, and vice versa. Hence, while the main focus of Gal & Irvine's work is the semiotics of differentiation, the ideological processes they describe, including erasure, are just as integral to the creation of similarity. Particularly in the analysis of the workings of ideology at the micro level of interaction, it is important to ensure that sameness and difference are equally available as interpretive possibilities.

Related to this, the complementarity of adequation and distinction allows us to recognize a fourth semiotic process of ideology, highlighting, as the converse of erasure. In his work on how professional activities enable particular ways of seeing and knowing the world, Charles Goodwin defines highlighting as a practice “which makes specific phenomena in a complex perceptual field salient by marking them in some fashion” (1994:606). Although he uses this term in reference to physical vision, we extend it here to include any semiotic act that brings to salience some aspect of the social situation, such as identity. These extensions of Gal & Irvine's framework help us to see that identity construction is always rooted in ideology as well as in practice. In identity work, erasure and highlighting often function in tandem to establish interactionally or situationally sufficient alignments and disalignments. Under certain circumstances, these processes can operate so completely as to create naturalized or essentialized identity positions, yet much of the time what is sufficient is merely partial identification or similarity between social subjects. The concept of adequation thus acknowledges that identification is never complete. This view of identity contrasts with the widespread common-sense notion that the foundation of identity is simply the act of equating or likening oneself to others.¹⁹

An example of adequation is found in Robin Queen's analysis of the discourse of a group of lesbians and gay men at a private social gathering. Queen shows that the speakers create a social network by drawing on ideologies and practices that they associate with a common “queer” identity. Although the participants differ along various lines (gender, occupation, race, nationality, native language), these differences are largely downplayed in the interaction in favor of an emphasis on this identity. Adequation with others as fellow queers is accomplished through a variety of strategies for presenting information (about history, stereotypes, practices) as shared knowledge. Queen's focus is on how a common queer identity allows speakers to create a social network; however, she does not merely stipulate this identity a priori but shows how it emerges as an interactional construct. For example, in a hypothetical discussion of how to determine someone's sexual orientation, the participants reject the idea of asking directly and instead invoke a series of tropes that rely on shared knowledge and stereotypes of lesbians and gay men:

The collaborative invocation of such tropes highlights similarities based on sexual identity that span the differences that might in other situations divide the participants. The use of indirect strategies to determine another's sexual identity constructs all participants as knowing how to produce and interpret these tropes. Adequation thus emerges as a result of the interaction, although the resources for this construction of intersubjectivity come from the broader sociocultural context.²⁰

In the joint pursuit of commonality through adequation, differences between interlocutors are temporarily erased or backgrounded while similarities are highlighted or emphasized. Conversely, distinction is the ideological construction of social difference through the erasure of likeness and the highlighting of unlikeness. Distinction therefore differs sharply from distinctiveness. Distinctiveness, as discussed above, entails structural uniqueness; in order for a linguistic feature to count as an index of social identity, its use must be restricted to speakers who share that identity. Distinction (like the other tactics) is not concerned with the formal characteristics of language, but instead with the way linguistic forms gain sociopolitical meanings through ideological processes in local contexts.

Distinction is often the intended byproduct of performance genres that involve parody, in which speakers mock characteristics or practices ideologically associated with particular identities in order to position their own identity as not sharing these traits. Such performances are frequently accomplished through a form of highlighting – the exaggeration of the linguistic practices stereotypically associated with particular groups. This is the case with Hindi-speaking kotis in Delhi, India (Hall forthcoming), a lower-middle-class transgendered identity that is self-defined in opposition to the burgeoning upper-middle-class urban gay community and also in opposition to lower-class hijras. Hijras are a longstanding social group in India who are often referred to as a “third sex” in the anthropological literature because of their self-identification as neither women nor men. Both kotis and hijras have male sexual partners, but the two groups distinguish themselves from each other on the basis of the fact that hijras are often castrated and hence popularly believed to be sexually impotent. Kotis exploit this perception in a performance genre they call “hijra-acting,” in which they parody the hijras' societal role as blessers of newborns at birth celebrations. Drawing on the cultural belief that hijras use foul language to compensate for their inability to have sex, kotis mock the rudeness associated with hijras alongside the prudishness associated with their upper-class patrons. In (2), Mani, playing the role of hijra, is demanding that the mother of a newborn, played by Balli, give Mani financial remuneration for blessings²¹

By contrasting the hijras' use of vulgar Hindi and intimate (int in the transcript) verbal forms – which are considered extremely rude when used between strangers – with the patron's use of English and more cordial familiar (fam) and polite (pol) verbal forms, the kotis develop a parodic critique of the sexuality associated with both upper- and lower-class positions. This act moves the kotis away from both groups by two degrees of separation: Rather than simply assert or display what is ideologically held to be distinctively koti, the speakers perform what is believed to be distinctively un-koti. This is accomplished through linguistic mockery of upper-class sexual repression on the one hand and lower-class vulgarity on the other. It is through the intertextual parody of all that is not koti that the kotis are able to differentiate their sexual identity, positioning themselves as having a desirable (and potent) sexuality in contrast to the prudishness and baseness respectively associated with other class positions. Given the explicit emphasis on sexual desires and practices in this situation, it might seem that a desire-centered approach would be the most illuminating form of analysis. This approach, however, would overlook the mutual dependency of desire and identity, both sexual and class-based. Examples such as (2) support our larger claim that articulations of desire are closely bound up with articulations of identity.

Authentication and denaturalization

The second pair of tactics, authentication and denaturalization, works off the ideological perception of realness and artifice. Whereas authentication concerns the construction of a true or veridical identity, denaturalization foregrounds untruth, pretense, and imposture in identity positioning. We prefer the term “authentication” to “authenticity,” the term more often used in discussions of identity, because the latter often emerges from a static essentialist perspective that holds that identities can be objectively verified or classified as real. The term “authentication,” in contrast, foregrounds the processes by which authenticity is claimed, imposed, or perceived; the concept maintains analytic distance by focusing on how essentialized identities are interactionally activated. By definition, then, the concept of authentication frames identity in nonessentialist terms, while recognizing the importance of essentialism as an identity tactic for social subjects (for further discussion see Bucholtz 2003).

The complex relationship between essentialist and nonessentialist identity construction in the authentication process is exemplified in the situation of phone sex workers discussed by Hall 1995. Hall describes how such workers exploit the resources of stereotypically “feminine” speech as well as racist and sexist ideologies to perform a range of sexualized female identities for callers' consumption, such as the nubile beach bunny, the coy Asian, and the oversexed African American. She argues that, contrary to some feminist perspectives that view “women's language” as powerless and the commodification of sex as a form of female disempowerment, phone sex workers themselves understand their language use as a powerful creative act in which they are able to shape male callers' erotic responses.

At first glance, phone sex may appear to be rooted in desire. In fact, this study has been hailed as another example of the language-and-desire paradigm: “Hall's work is important because it directs us to examine the precise linguistic resources that people use to animate desire. But it does so without reducing desire to identity. Indeed, work like Hall's directs our attention in completely the opposite direction, since the desire emitted through the language of the sex-line workers has nothing to do with their identities” (Kulick 2003a:136; cf. Cameron & Kulick 2003a:129–30). But this perspective fails to notice that what Hall documents is not sexual play for pleasure; it is work. The desire resides, in most cases, entirely with the customer on the other end of the line. Thus, approving remarks about the “disaggregation of desire from identity” in Hall's work (Kulick 2003a:136) miss the crucial point that the performance of desire and the performance of identity go hand in hand in phone sex: The client's desire is not simply for sexy language but for the sexy identity that the worker projects through language. It is difficult to imagine how one could talk about desire in this context without also talking about identity. To be sure, the phenomenon of phone sex undermines essentialism, since the workers are actors who represent themselves in ways that do not necessarily match up with their own biographies, but this is very different from claiming that it undermines identity itself. Thus, while phone sex workers' performances cannot be interpreted by analysts as authentic – given that they cross lines of race, ethnicity, and even gender – their clients are not usually so enlightened. For customer satisfaction to be achieved, the illusion that the language of desire produced by a phone sex worker is authentic and reflects an authentic identity on the other end of the telephone line must be maintained.

Authentication is thus useful for getting out of the bind of essentialism in which advocates of the desire paradigm are trapped. To understand the practices of these sex workers as a “forgery” of both desire and identity (Cameron & Kulick 2003a:129; Kulick 2003a:136) assumes the possibility of a “real” desire or identity. We are less interested in the disparity between real and unreal than in the process by which these linguistic practices authenticate or verify identities. In our view, what needs to be explained is not inauthenticity, as the desire approach implies, but the authenticity with which certain identities become endowed.

A second example of authentication comes from Kulick's (1998) ethnography of Brazilian travestis. Despite the author's later programmatic statements, this book stands as an example of the identity approach to sexuality. The political economy in which travestis live is quite complex. On the one hand, as sex workers they sell their bodies on the market. But they also, crucially, use the money they earn in this way to buy items that will make them more feminine and attractive to men and to purchase presents for their boyfriends. Indeed, Kulick implies that the gifts and financial support that travestis regularly supply to their boyfriends is a way of compensating, quite literally, for not being real women. Here too, desire is not what it seems: “Whatever else travestis get out of their boyfriends, then, it is not sexual fulfillment. The point of having a boyfriend, instead, is for him to help a travesti feel like a woman…. [Travestis] do not want boyfriends for sexual pleasure. They don't get sex from their men – what they get, instead, is gender” (1998:133). The absence of sexual satisfaction that typifies travestis' relationships with their boyfriends indicates that desire is not the sole motivation. Boyfriends provide travestis with validation – authentication – of their feminine identities. Travestis are under no illusion that their femininity is “real,” but this makes authentication more, not less, necessary.

A final and somewhat different example of authentication is found in the widespread phenomenon of special registers (often construed by their speakers as “languages”) spoken by sexual minorities for purposes of both secrecy and solidarity. Such registers include Indonesian Bahasa Gay (Boellstorff forthcoming), British Polari (Lucas 1997; cf. Baker 2002), Filipino Swardspeak (Manalansan 1995), and Indian Hindi-based “Farsi” (Hall, forthcoming). Polari, for instance, as a once secret gay lexicon now falling into disuse, has recently been revived as an authenticating emblem of longstanding queer cultural heritage in the sexually and religiously subversive rituals of the Sisters of Perpetual Indulgence, a queer activist group. Similarly, the lexicon called Farsi has been embraced as a historicizing marker of nonnormative sexual identity by transgendered hijras and kotis in India, who claim an indigenous lineage dating back to the eunuchs of the medieval Moghul courts. Although Koti/Hijra Farsi is unrelated to Persian Farsi, its speakers conceptualize it as the language of the Moghul courts, employing it in the construction of a historically authentic sexual identity that precedes and hence supersedes newly forming urban upper-middle-class gay identities. In both of these situations, social subjects enact authentication by historicizing their identities through claims of linguistic continuity with a valued past.

Far less examined than such authenticating actions are the ways in which identities are denaturalized, thereby making them literally incredible. Researchers have tended to set aside such disruptions of the authentic as marked or inconsequential, yet they are far more prevalent than their relatively scant documentation in linguistic research suggests. Rusty Barrett's (1999) study of African American drag queen performances provides a dramatic example of how ruptures are fundamental to the process of identity construction. Instead of maintaining a seamless identity like the phone sex workers discussed above, the drag queens of Barrett's study deliberately create dissonance between a “flawless” physical self-presentation as a woman and a heteroglossic linguistic performance that combines features ideologically associated with white women, African Americans, and gay men. In (3), for instance, one drag performer studied by Barrett draws on several of these features in rapid succession and thus denaturalizes what could in other circumstances be a straightforward feminine physical self-representation:

Gender theorists like Butler have drawn on drag queen performances like this one to argue for the denaturalization of hegemonic conceptions of gender. Although we agree that such performances carry this potential, our own understanding of denaturalization is much broader, encompassing not only the rupture of dominant ideologies of gender and sexuality, but also the creative power of individual identity work that is often best arrived at via ethnographic research. For Butler, drag queens are foot soldiers at the vanguard of the gender revolution, combating the perception that gender is biologically constituted. But drag is practice as well as theory, and as Barrett so aptly illustrates, drag queens may be at least as interested in constructing their own identities as in challenging the identities of others. In (3), the performer combines a stereotypical feminine abhorrence of indelicate language with an explicit articulation of gay desire.²²

Denaturalization may also occur when the authenticity of an identity is challenged or questioned because a rupture of that identity has been perceived. Example (4) illustrates this type of denaturalization. The speakers, Christine and Claire, are European American high school girls in the San Francisco Bay area who are both self-described “nerds.” In (4), they are discussing two of their friends, whom, moments before this excerpt, Christine has described as extremely nerdly and naïve. The girl and boy under discussion, Gabby and Danny, have recently entered into a romantic relationship²³

Nerdiness in this and many other American high schools is associated ideologically, and often in practice, with a relative disengagement from what Eckert 2000 calls the “heterosexual marketplace” (Bucholtz 1996). Nerds' deliberate separation from cool, trendy, sexually active students is signaled in large part through language. Christine's use of superstandard English (Bucholtz 2001), as manifested in careful articulation of alveolar stop segments in fainted (line 6) and assumed (line 33) and formal-register lexical choices such as privy (line 32), exemplifies this nerdy speech style. Nerdy students are less likely than cool students to become romantically involved, and when they do, their relationships may have little or no (publicly displayed) sexual dimension. The evidence that Christine's and Claire's friends are sexually active is discursively represented by Christine as heartbreaking as well as horrifying because it is inappropriate for these people – nerds. If Gabby and Danny are no longer sexually “naïve”, by definition they are no longer completely “nerdly”. Christine's story denaturalizes her friends' formerly authentic nerd identities by evaluating their sexual activity as a tragic loss of nerdy innocence. This denaturalization is also enacted linguistically through the contrast between the representation of Gabby's speech as high-pitched, giggly, and silly and Christine and Claire's own deeper-pitched expressions of disgust. This contrast ties the inappropriate display of gender to the inappropriate engagement in sexuality.

Authorization and illegitimation

The final pair of tactics we wish to describe, authorization and illegitimation, primarily foregrounds relations of power. Authorization is the use of power to legitimate certain social identities as culturally intelligible, while illegitimation is the revoking or withholding of such validation from particular identities. These tactics are often associated with large-scale or institutional workings of power, but they may also be more local. This pair of tactics often works in close conjunction with the tactics of authentication and denaturalization, for issues of genuineness are frequently linked to issues of legitimacy. However, it is useful to separate these phenomena conceptually because they can operate independently of one another; the defining criterion of authorization and illegitimation is the central role of institutionalized power structures.

In the context of sexuality, authorization and illegitimation are tied to the cultural recognition of some sexual identities and practices but not of others. Indeed, this issue is the impetus for the project of queer theory, in that structurally marginalized sexual categories are placed at the very center of intellectual and political analysis. This move in itself is a kind of authorization via the institutional legitimation of queer sexualities as a topic of academic research. Queer theory is therefore a response to the more typical forms of sexual authorization – heteronormative and homophobic discourses – in which straight sexual subjects are called into being or, to use Althusser's (1971) term, “interpellated” at the expense of queer subjects.

A demonstration of the role of language in the authorization of normative heterosexuality is offered by Celia Kitzinger's analysis of out-of-hours telephone calls to health service physicians on behalf of an ill family member, friend, or neighbor. These physicians are not the patients' regular family doctors, and hence information about the caller, the patient, and their relationship must be provided within the telephone interaction. Through conversation analysis, Kitzinger shows that both doctors and callers take for granted a normative notion of family as nuclear, biological, cohabiting, legally recognized, and heterosexual. The inferences associated with this normative structure allow the interaction to proceed without difficulties. In (5), the caller's phone call is returned by the doctor²⁴

Elsewhere in Kitzinger's data, mention of nonnuclear and nonfamilial relationships such as “boyfriend” or “neighbor” often gives rise to additional explanation; by contrast, the “husband” relationship invoked by the recipient of the doctor's telephone call in (5) requires no further comment. In addition, the speaker displays (and is treated by the doctor as having) privileged access to the patient's medical situation (e.g., lines 26–27). Finally, the on-call doctor's inquiry, Has- Have you: seen: (.) your own doctor at all¿ (line 20) presupposes that the caller and the patient share a family doctor, a presupposition that is not in force when the caller is not the spouse or parent of the patient under discussion. As Kitzinger notes, such mundane and casual reinforcements of heterosexual kinship reproduce a narrow hegemonic concept of family structure.

Authorization focuses on identities that receive institutional sanction; illegitimation, in contrast, calls attention to identities that are denied such recognition. The structures of power at work in these tactics may be broad in scope or may be specific to particular communities. Thus, while the above example illustrates how large-scale cultural institutions like kinship and medicine may yield power asymmetries in favor of a restrictive heterosexual norm, institutions may also reproduce power inequities even within sexually marginalized groups. In this way, some sexual minorities may be double victims of illegimation – by the dominant culture and by other sexual minorities. This is the case in Livia's (2002b) study of personal ads in a French lesbian magazine. Livia found that the majority of such ads explicitly exclude as prospective sexual and romantic partners those lesbians who display a masculine style, as in (6):

Although, as Livia points out, the magazine's editors often criticize this narrow view and promote inclusiveness in its readership in some ways, they too may illegitimate lesbian identities that appear “too masculine.” By permitting and even producing anti-butch discourse within its pages, the magazine disseminates a particular hierarchical perspective on which lesbian identities are acceptable. That is, the magazine, as an institutional authority for a French lesbian “community,” patrols the borders of gender and targets some lesbians for illegitimation on this basis, both in its publication of exclusionary personal ads and in its editorial decisions.

Interactions between tactics

Although we have highlighted a particular tactic in each of the above examples, intersubjective relations are far more often multiple than singular, and more than one tactic is relevant to all of the studies we have discussed here. A final example, taken from Rudolf Gaudio's (2001) research, illustrates how tactics may work together in identity construction. In discussing talk about sex between a gay European American anthropologist (Gaudio himself) and black African 'yan daudu (men who act like women) in Nigerian Hausaland, Gaudio reports the construction by all the speakers of an essentialized “difference” in sexual practice across cultural boundaries. Yet he also identifies moments when some speakers challenge these boundaries and the sociopolitical realities that underlie them:

What Gaudio found startling about this response was that Aliyu himself was a dark-skinned man, yet in the U.S. context the term “chocolate queen” refers exclusively to nonblack gay men. For this reason, Gaudio argues that Aliyu's appropriation of a Western sexual label is not merely a statement about desire; he notes that a friend of Aliyu had traveled to the United States as the boyfriend of a wealthy white man, and so Aliyu was well aware of both the racial and economic injustice experienced there by people of color. Aliyu's claim to the label “chocolate queen” is thus also a claim to an identity category that in Western contexts is withheld from him as illegitimate because he is black. Throughout this interaction, various tactics shape the identity positions of participants. First, Gaudio's claim to the term “chocolate queen” creates an intersubjective relation of distinction between black African and white American homosexualities where desire is tied to the racialized identities of self and other. Aliyu's response, in turn, uses the same term to assert a shared identity with Gaudio, an act of adequation. It is here that Gaudio realizes that his own identity claim reproduces a racialized history of colonialist sexuality by authorizing whites' desire for black bodies like Aliyu's. The effect of Aliyu's comment (at least on Gaudio) is to illegitimate white-dominated systems of sexuality that would rigidly locate him – regardless of his own sexual self-positioning – as desired object rather than desiring subject.

As this and the other examples show, tactics of intersubjectivity may work together or against one another as discourse unfolds, creating an ever-shifting matrix of identity relations – both imposed and chosen, both deliberate and accidental – in specific sociocultural contexts.