Introduction: Agency and Transgender Voices

In both sociocultural linguistics and contemporary transgender politics, there is a strong connection between agency, power, and ideology. Sociolinguists have, since the field’s earliest days, recognized that speakers have control over at least some aspects of their linguistic practices, and that understanding linguistic variation depends in part on researchers’ ability to tap into these aspects of awareness and agency, as the contributions to this volume make especially clear (particularly Campbell-Kibler, Carmichael, and Babel). Ideology figured prominently in the theorization of sociolinguistic agency from the field’s inception, with a “standard language ideology” (per Lippi-Green Reference Lippi-Green1997) driving people towards more standard speech in contexts of greater awareness or, alternatively, away from that standard as an expression of resistance. One question as yet unexplored is how ideologies about sociolinguistic agency itself might be taken up by speakers as a ground on which to constitute certain kinds of subjectivities. In contrast to the normalization of awareness and control to be found in much sociolinguistic literature, the practice of self-consciously shifting the gendered characteristics of the voice takes on a decidedly different ideological valence in some transgender communities. In the analysis to follow, two perspectives on transgender people’s control over the gendered characteristics of their voices are examined in order to call attention to the variable ways in which agency can be constructed, the ideological implications of those constructs, and the importance of considering these ideologies when producing accounts of speakers’ awareness and control over their sociolinguistic practices.

In transgender communities, agency is a complex and multifaceted issue. Like lesbian and gay activists, trans people often make sense of their gender identities as innate and even biological in origin, in contrast to normative discourses that frame a gender role transition as an unnecessary, even indulgent, deviation from the natural order. This stance, which disavows the ability to choose one’s internally felt gender identity, works to legitimize trans people’s social positionalities through an intertextual relationship with other discourses of authenticity and recasts a gender role transition as a means of realizing an inevitable inner truth. At the same time, however, agency is of critical importance in the discourses about self-definition and bodily self-determination that dominate trans communities in contemporary North American contexts (contrast Besnier Reference Besnier, Holmes and Meyerhoff2003, for instance). That is, a core socio-political goal of many trans communities in the United States has been to secure individuals’ freedom to choose what kinds of gendered body modifications they might wish to pursue and to name the identity categories to which they most authentically belong. This tension, between rejecting control over one’s sexual or gender identity and claiming freedom to act in a particular framework of culture and power (Ahearn Reference Ahearn2001), provides the ideological backdrop for the discourses analyzed in this chapter.

In order to better explicate these ideologies, I investigate the construction of linguistic agency found in transgender speakers’ ethnographically situated discourses and contrast them with those that appear in academic literature on trans people’s voices. Although studying the speech of trans people has promised unique insights for sociocultural linguists, most research on transgender voices to date has been carried out by speech pathologists who have a less reflexive and more scientifically oriented perspective on gender and sex. The discourses that occur in this body of literature reflect and reinforce hegemonic cultural ideologies about the inauthenticity of trans people’s self-defined gender identities, while simultaneously homogenizing the speech of “women” and “men” as macro categories. As I explore in detail below, research on trans voices often has the effect of naturalizing the speech and gender presentations of non-trans (or cisFootnote ¹) women and men, while delegitimizing and pathologizing those of trans speakers. This is accomplished in part by portraying trans people as individuals who are working against their biology in order to imitate the purportedly naturally feminine or masculine voices of cis women or men, respectively. This naturalization erases the tremendous variability that exists in the gendered practices of cis women and men on the bases of class, ethnoracial identity, culture of origin, sexuality, or disability, to name a few. Furthermore, the idea that trans people might have their own ideas about how they would like to sound, might not want to change their voices at all, or might achieve a voice that allows them to “pass” without any special conscious effort is rarely acknowledged. In the case of trans men (i.e. female-to-male trans people), who appear in only a small portion of studies carried out on trans voices, testosterone tends to be emphasized as the primary means of achieving a male voice, although a few more recent studies (particularly van Borsel et al. Reference van Borsel, de Cuypere, Rubens and Destaerke2000; Adler and van Borsel Reference Adler, van Borsel, Adler, Hirsch and Mordaunt2006) have suggested that trans men should seek speech therapy despite the effects of hormone therapy.

Clearly, there are a number of both popular and academic discourses about trans people and their voices that facilitate the assumption that their transition process is driven primarily by individual, agentive choices about how to control their speech and behavior in order to conform to norms for their “new” gender. And, indeed, this kind of self-conscious effort does characterize the experience of some trans individuals. However, my research on the voices of female-to-male transgender speakers highlights a rather different perspective, in which trans people who make use of testosterone orient to self-conscious vocal masculinization as far less desirable than passively allowing hormonal changes to deliver a deeper, more masculine voice. In the metalinguistic discourse analyzed in this chapter, the potential to exercise agency over the gendered characteristics of the voice is consistently disavowed. As the final section of this chapter underscores, this rejection is crucial in producing an account of the sociolinguistic styles found in transmasculineFootnote ² communities. Furthermore, it highlights the complex ways in which power and hegemony factor into linguistic awareness and control.

While transmasculine speakers’ construction of sociolinguistic agency may at times appear contradictory, a closer examination reveals a parallel between these speakers’ discourses about sociolinguistic control and ideologies about personal agency over gender identity and the body. Even as their metalinguistic commentary contests broad transphobic cultural ideologies, however, other problematic ideologies are implicated in their stead as the practice of changing the gendered characteristics of one’s voice is ascribed a negative political and moral value. Specifically, lack of awareness and control is equated with naturalness and authenticity, while the exercise of control is linked to artifice and assimilation.

This chapter, then, concerns the shifting roles played by agency, awareness, and control in the negotiation of gender and the voice among trans men and others on the transmasculine identity spectrum. Much as Silverstein (Reference Silverstein, Mertz and Parmentier1985) describes, the metalinguistic commentary I analyze here demonstrates how ideologies exert an influence over what he calls the “structural realm” – i.e. linguistic patterns of which speakers may not be consciously aware. The trans speakers discussed here show relatively little awareness of the linguistic characteristics that are said to distinguish women’s and men’s voices, but they are far more sensitive to the socio-indexical (pragmatic, for Silverstein) meanings attached to the practice of self-conscious masculinization itself. These ideologies about the undesirability of linguistic control become the means through which agency is exercised on the structural level of sociophonetic style despite these speakers overtly constructing themselves as lacking or rejecting agency over their voices. Agency, from this perspective, is not about an internal sense of intentionality, but is rather the product of discursive practices that bring speakers’ subjectivities into existence (e.g. Butler Reference Butler1990). Silverstein focuses on the ways in which ideology can be a means of exercising linguistic agency, but the present analysis suggests that exercising agency can itself be a tool for ideological work, and that claims on or disavowals of awareness and control may be critically important in our explanations of social and linguistic practice. In the case of the present analysis, the way in which trans speakers’ distance themselves from self-conscious linguistic masculinization provides both an explanation and a legitimation for the non-normatively masculine phonetic characteristics that are evident in many transmasculine people’s speech.

The substance of this chapter begins by characterizing speech pathologists’ investigations of trans voices over the last four decades as a means of illustrating some of the discourses about gender against which trans men’s negotiations of agency are positioned. Next, examples of transmasculine people’s discourses about the voice come from metalinguistic commentary recorded in interviews and conversations with participants in a two-year ethnographic study on fifteen transmasculine individuals in the San Francisco Bay Area during the early stages of testosterone therapy. This corpus of interviews is supplemented by lists of advice known as passing tips that circulate in online communities for trans men. Such commentary, I argue, works to both reflect and partially constitute trans people’s local ideologies that divide an individual’s embodiment from a more authentic inner self. The focus of my analysis is on the potential contradictions that emerge as trans men simultaneously claim certain kinds of agency while disavowing others, highlighting some of the ways in which agency appears to work through “fragmented subjectivities,” in Ahearn’s (Reference Ahearn2000: 13) terms. In resolving these tensions, my goal is to highlight the depth and complexity of speakers’ own engagement with linguistic agency and the necessity of examining that engagement as part of our sociocultural linguistic analyses. Agentive intentionality is not something that can be observed or measured directly, nor should the experience of agency be equated with speakers’ explicit claims about their intentions. What we can observe, alongside the linguistic reflexes of awareness and control, is the way in which language users produce an understanding of agency that has real implications for their linguistic practices; it is on this issue that I conclude with a summary of these speakers’ gendered phonetic styles and the connection between those styles and the ideologies analyzed here. This form of constructed agency is not only exercised over language, but also constituted through language, which serves as a means of doing ideological work with importance that extends beyond the realm of sociolinguistic practice.

Trans People and Sociolinguistic Control in the Speech Pathology Literature

One domain in which we can see the ideology that transgender people must actively attempt to control their gendered presentations – and particularly their speech – is the research carried out by speech-language pathologists, which remains the largest body of work on the voices of trans people. I have presented more extensive analysis of this literature in other spaces (Zimman Reference Zimman2012), wherein I focused on approximately twenty publications from journals and books on speech-language pathology and the speech sciences between 1977 and 2011 (including Bralley et al. Reference Bralley, Bull, Gore and Edgerton1978; Mount and Salmon Reference 276Mount and Salmon1988; Spencer Reference Spencer1988; Wolfe et al. Reference Wolfe, Ratusnik, Smith and Northrop1990; Günzburger Reference Günzburger1995; McFadden et al. Reference McFadden, Pasanen and Callaway1998; Gelfer Reference Gelfer1999; van Borsel et al. Reference van Borsel, de Cuypere, Rubens and Destaerke2000; Gelfer and Schofield Reference Gelfer and Schofield2000; Dacakis Reference Dacakis2002; Gelfer and Mikos Reference Gelfer and Mikos2005; Adler and van Borsel Reference Adler, van Borsel, Adler, Hirsch and Mordaunt2006; Davies and Goldberg Reference Davies and Goldberg2006; T’Sjoen et al. Reference T’Sjoen, Moerman, Van Borsel, Feyen, Rubens, Monstrey, Hoebeke, De Sutter and De Cuypere2006; Owen and Hancock Reference Owen and Hancock2011). In this section, I focus on three of the ideologies I have identified in this body of work that are particularly relevant to the present discussion of agency and sociolinguistic control: the naturalization of gender differences in the voice; the assumption that trans people need agentive interventions to change and control their “natural” voices; and the assumption that trans people want to sound like women or men who embody gender- and heteronormativity – a notion that is itself always inflected by race, class, and other lenses of subjectivity. Although the quotes I’ll use as exemplars may seem extreme or antiquated, the speech pathology literature shows movement away from these discourses only within the last handful of years, while the existing body of work continues to inform both researchers and clinicians who work with trans speakers. Furthermore, these ideologies are simply one instantiation of the wide-spread denaturalization and delegitimation of trans people’s gender identities that continues despite significant gains trans activists have seen in the past few decades. It is worth noting that trans women are the focus of the great majority of the speech-language pathology literature, while trans men’s voices have only recently been the subject of similar investigations. As pathologists’ discourses concerning trans women have spread to the emerging literature on trans men’s voices, ideological clashes become apparent between the language ideologies taken up by researchers and those expressed by transmasculine speakers. Often, these tensions center around the role of agency and control in shaping the gendered characteristics of trans voices. Like speech scientists, transmasculine people who make use of testosterone often naturalize the male voice as a product of hormonal sex, but may also reject the notion that they should put effort into living up to the ideals of cis masculinity. Although the few studies of trans men’s voices that have been published suggest that members of this group could “benefit” from speech therapy (van Borsel et al. Reference van Borsel, de Cuypere, Rubens and Destaerke2000; Adler and van Borsel Reference Adler, van Borsel, Adler, Hirsch and Mordaunt2006), the ideological schisms I identify in this section and the next indicate that awareness and control over the gendered voice is conceptualized quite differently by speech-language pathologists on the one hand and transmasculine individuals who make use of testosterone on the other.

As in much research on gender differentiation in the voice, there is a strong tendency in the speech pathology literature to frame the gendered voice as a product of biological differences between the sexes, despite the extensive evidence that gendered phonetic traits emerge during childhood language socialization (e.g. Sachs Reference Sachs, Thorne and Henley1975) and may vary within gender groups as much as between them (e.g. Stuart-Smith Reference Stuart-Smith, Cole and Hualde2007). Testosterone is known to affect the larynges of trans men, creating an often dramatic drop in vocal pitch (van Borsel et al. Reference van Borsel, de Cuypere, Rubens and Destaerke2000; Damrose Reference Damrose2009; Papp Reference Papp2011; Zimman Reference Zimman2012), but studies of trans women’s voices emphasize that feminizing hormone therapy does not raise trans women’s pitch, leading to conclusions like Wolfe et al.’s: “the natural voice pitch of the male-to-female transsexual remains at a lower level, completely at odds with a new female role, unless change is attempted” (Reference Wolfe, Ratusnik, Smith and Northrop1990: 43). Gelfer and Mikos (Reference Gelfer and Mikos2005) claim that “[a]cquiring the voice characteristics of the reassigned gender is a particular challenge for male-to-female transgendered persons, because the vocal mechanism in most cases has attained adult male dimensions, and it is not affected by the administration of female hormones” (545). Notably, it does not appear that this claim has been empirically verified. Even as estrogen is assumed to have no effect on trans women’s voices, phoneticians studying cis women’s voices have claimed (perhaps speciously) that estrogen contributes to greater articulatory precision among women as compared to men (Whiteside et al. Reference Whiteside, Hanson and Cowell2004; Wadnerkar et al. Reference Wadnerkar, Cowell and Whiteside2006).

Based on an understanding of trans women as having male voices by nature, authors in this body of work assume that trans speakers need help overcoming their “natural voices” in order to achieve “female-like speech” (Spencer Reference Spencer1988). A dichotomy is thus constructed between biologically determined gender differences, which give us our “natural voices,” and self-consciously chosen speaking styles that can to some extent mask those natural characteristics. At the same time, the idea of “female-like speech” is homogenized, erasing the variability found among cis women’s voices across communities. Spencer claims that “speech changes which result in a ‘female-sounding’ voice represent deliberate acquisition of specific behaviors rather than the effects of hormone therapy” (Reference Spencer1988: 40). The notion that trans women might sound female without explicit instruction is not usually considered (although Gelfer (Reference Gelfer1999) gives credence to the possibility), which may be in part because trans people fitting this description are unlikely to seek out speech therapy. But trans women who have little difficulty producing normatively feminine voices are erased from the research, reinforcing the idea that trans women are naturally masculine and must exert control over that natural state in order to achieve something that resembles the natural femininity of cis women, who need not put in any effort to sound “female-like.”

A final ideological position evident in the speech pathology literature is the assumption that trans speakers want to sound like stereotypically straight, gender-normative cis people who also embody white, middle-class, American femininity (or, when applicable, masculinity). When Gelfer refers to successful trans women clients as speaking with an “acceptable” female voice (Reference Gelfer1999: 1) and Spencer describes the more feminine-sounding trans women in her study as having “more adequate speech patterns” (Reference Spencer1988: 39), it is against that highly privileged – and not necessarily reality-based – norm that trans women are being compared. This is true if only because scholars’ and speech therapists’ knowledge about gendered characteristics in the voice is informed by studies that often use college students as representatives of “women” and “men” in general. But this assumption is intensified by lack of acknowledgement that, for example, some trans women may prefer to be read as lesbians and/or masculine women, as women of color, as older women, or as working class women. In fact, even where trans women suggest they might have different goals, the insistence on gender normativity may persist; Spencer remarks with wonder that all of her clients were satisfied with their voices despite the fact that not all were perceived as female speakers in decontextualized experimental contexts. She concludes that the trans women who are perceived as men must be mistaken about how they sound: “subjective clients reports of the adequacy of the speech product may not be valid” (Reference Spencer1988: 40).

Trans women remain the primary focus of this research on trans voices – based in part on the demand for speech therapy from this group – but speech researchers who have recently begun to look at trans men’s voices have argued that this group could similarly profit from speech therapy because their voices may continue to diverge from normative expectations for linguistic masculinity – expectations that are not necessarily embodied even by straight cis men. Such arguments, it should be noted, exist in an economically symbiotic relationship with speech therapy practitioners: researchers have the power to legitimize speech therapists’ services with the argument that a segment of the population “needs” speech therapy, while demand for speech therapists’ services in turn justifies the funding of speech pathology research.

Despite the weight of these arguments, trans men on testosterone are often in a position to resist the idea that they require speech therapy or other means of controlling and consciously changing the gendered characteristics of their voices. As the remainder of this chapter demonstrates, this reluctance is driven in large part by trans men’s conflicting relationship with agency over their bodies versus agency over their gender presentations.

Speech-language pathology may in some ways seem a world apart from sociocultural linguistics, but I shine a light on this literature because sociocultural linguistic research on trans people’s voices risks the same kind of problematic assumptions about speaker agency if we are not careful to situate our claims about agentive sociolinguistic practices in both locally grounded and broader sociocultural context.

Agency in Transmasculine Discourses about the Gendered Voice

Turning to metalinguistic commentary from transmasculine participants in my ethnographic fieldwork, a different set of ideologies about sociolinguistic control is evident – one that ultimately informs their sociolinguistic practices on a finer level. The overarching theme in this commentary is one of rejecting the idea that vocal masculinization can or should be achieved through self-awareness and control over gendered elements of the voice rather than hormone therapy. On the few occasions that my participants did acknowledge making these kinds of attempts, it was framed as something they had done in the past, before testosterone provided them with a deeper, male-sounding voice, and generally with little success. Of course, starting hormone therapy involves certain kinds of agentive choices, often driven specifically by the desire for a lower voice. Yet the speakers I put into focus do not talk about the influence of testosterone on their voices in terms of personal agency, but rather tend to present it as a matter of biological determination that is, for the most part, out of their hands. As I will argue later in this section, the apparent paradox in making strong claims over certain kinds of agency while simultaneously rejecting others can be understood as part of a broader ideological separation between sexual embodiment and an internally felt and self-defined gender identity that is central to contemporary formulations of trans identity. Importantly, though, other familiar ideologies are simultaneously engaged in this talk, including the naturalization of vocal pitch as determined by hormonal sex, the valorization of bio-medical interventions over behavioral changes, and the idea that femininity is achieved through artifice, while masculinity is characterized by an absence of effort.

There are three primary themes through which transmasculine speakers problematize and reject self-conscious masculinization of their voices: the effects of testosterone on the voice, the high value placed on authenticity, and political objections to linguistic assimilation.

As noted, trans men’s voices are known to drop in pitch when testosterone therapy is administered, and members of this group tend not to seek out speech therapy in part for this reason. Because testosterone is a mainstay of female-to-male medical transition, the received wisdom is that trans men do not need speech therapy in order to sound male. In his review of “transgender and language,” Kulick (Reference Kulick1999) argues that this assumption – that trans women need speech therapy while trans men do not – reflects and perpetuates the naturalization of masculine speaking styles and the treatment of femininity as artificial and achieved only through careful training. Since Kulick’s publication (though probably not because of it), speech-language pathologists have conducted a handful of studies on trans men’s voices and have also begun to question the assumed primacy of testosterone-fuelled changes (most overtly van Borsel et al. Reference van Borsel, de Cuypere, Rubens and Destaerke2000 and Adler and van Borsel Reference Adler, van Borsel, Adler, Hirsch and Mordaunt2006). Transmasculine people making use of testosterone, however, continue to emphasize the effects of hormone therapy on the voice.

Important to this discussion is the idea of passing, or being perceived as a member of one’s self-identified gender, which has a central and yet highly problematized status in many trans communities. On the one hand, trans people who are not perceived in the way they would like often talk about passing as a goal or express frustration at their inability to pass. Since most of my participants were early in their transitions when I started recording them and were thus still negotiating the extent to which they were perceived as men, the term was not uncommon in their conversations. At the same time, when participants in my study talked about passing, they often signaled their awareness of the problems with the term – for example, by gesturing scare quotes with their fingers – as well as in some cases overtly objecting to the notion that they should strive to pass. The main objection to the word passing in this context is that it usually refers to people who are perceived as members of a group to which they do not, in reality, belong. For instance, one might say that people of color at times pass as white despite not being white or that gays and lesbians may similarly pass as straight despite not being straight. To say that a trans person passes as a (wo)man suggests that trans men are not really men and trans women are not really women, only mistakenly perceived as such. Despite objections over the word itself, it is of great importance for many transmasculine people – especially those who have not (or not yet) seen dramatic physical masculinization from hormones – to manage their gender presentations. Indeed, for my participants, the choice to go on testosterone was motivated by a desire to be read as male at least some of the time in their everyday lives. Many also invoked an internal sense of what they feel their body should be like (see Prosser Reference Prosser1998; Salamon Reference Salamon2010 for more on this idea), but the social perceptions of others was an important factor for everyone in my study. Notably, this complex and problematized view of passing is absent from the speech pathology literature.

Because of the interest transmasculine people may have in shaping their self-presentation in order to “pass” as men, a genre of texts known as passing tips has developed for circulation in online trans communities. This information is then further disseminated within local communities. One website, “The FTM Passing Tips Site,”Footnote ³ which contains advice that I encountered in near identical form in online spaces for trans men in the 1990s, features advice on a range of topics from clothing and haircuts to swimming and negotiating men’s restrooms. Some example tips from this site include the following:

• If you live in a cosmopolitan area where there are a lot of butch lesbians then it’s going to be much more difficult for you to pass. One way to help distinguish yourself from them is to dress more conservatively – you might want to leave the leather motorcycle jacket at home for a while.

• All-over crewcuts are also problematic because they emphasize the shape and size of one’s skull and are therefore feminizing (look at Sinead O’ Connor).

• Footwear can be a real problem if you have small feet, although you do save money if you can wear boys’ athletic shoes.

In contrast to passing tips with detailed advice about clothing, haircuts, different methods for binding (i.e. flattening one’s chest) or packing (i.e. creating a bulge in the crotch of one’s pants), relatively little guidance is available on vocal masculinization. This paucity also stands in stark contrast with the both informal and commercial guidance available to trans women for vocal feminization. Like many of the passing tips on this site, the single sentence under the “voice” section is phrased in terms of generalities about how women and men behave. It reads:

• Women tend to use an upward inflection at the end of their sentences, while men tend to speak in more of a monotone.

Despite the importance of vocal pitch, it is not treated as something that can or should be consciously manipulated the way that the use of rising versus falling or flat intonation might be. In fact, the only other reference made to gender differences in the voice in this document of approximately 6,000 words is an insistence that testosterone is “the only safe and effective way” to achieve a lower-pitched voice:

• The only safe and effective way to lower your voice, masculinize your body, and grow facial hair is to take testosterone under the care of a doctor. Any other way is ineffective and potentially hazardous to your health.

In this way, the voice is framed differently from other semiotic indexes of gender, such as bodily hexis (Bourdieu Reference 275Bourdieu1984), how much a person talks, or what they talk about. Advice on “body language” is fairly common, with the same website asserting: “Women tend to be less obtrusive, while men tend to take up more space. If you watch commuters on a bus, women tend to sit with their legs crossed and their arms drawn in, and men tend to sit with their legs apart and their arms out.” Other websitesFootnote ⁴ suggest being less talkative, not smiling as much, and avoiding “gossip” in order to pass. Yet, the overall emphasis in transmasculine discourses on linguistic masculinization is on physiological changes brought about by testosterone, such that the voice is listed in conjunction with body hair and the general notion of a masculinized body. Another site goes as far as to instruct trans people not to intentionally adopt a lower pitch range, warning that:

• Deepening your voice on the phone is fine and may be a good idea if you’re on the high side, but don’t try to deepen it when having a conversation in person. You can see it in a person’s face when they’re trying to change their voice.Footnote ⁵

The foregoing discussion of electronically mediated passing tips that circulate within and across trans communities – of which participants in my fieldwork were almost universally aware – leads into the first language ideology taken up by the trans men whose metalinguistic commentary are the focus of the remainder of this section: the idea that testosterone will “take care of … pretty much everything,” as Mack puts it in Excerpt 1 (lines 9 to 10). Notably, Mack’s formulation casts testosterone as the grammatical subject and thematic agent in the process of “taking care of” voice masculinization, grammatically encoding what Duranti (Reference Duranti and Duranti2004) calls the “chain of causality” from testosterone to the voice through strategic manipulations of syntactic structure.

And, indeed, changes in vocal pitch documented for these speakers by Zimman (Reference Zimman2012), as well as similar findings for other trans populations like those studied by Papp (Reference Papp2011), support the conclusion that testosterone generally does help speakers move from a pitch range considered typical for women into a range considered typical for men. One of the speakers in this study, whom I call Mack, characterizes the way in which trans men typically see the effect of testosterone. Mack is a 46-year-old, white, straight-identifiedFootnote ⁶ trans man who grew up in the Bay Area and works as a bus driver for a private charter company. During one of our conversations in his home in the southern outskirts of San Francisco, Mack was telling me about some of his transfeminine friends and how hard many of them work in order to be recognized as women. In Excerpt 1, I ask him whether he thinks trans men tend to put in as much effort into masculinizing their appearance and behavior. Significantly, Mack has been frustrated with the relatively slow speed at which his own voice has changed, yet maintains confidence that testosterone will eventually do for him what it has done for other trans men he knows.

My own observations support Mack’s claim that trans men are by and large confident that testosterone will masculinize their voices to the extent that they see little reason to gain awareness and control that might lead to more normatively masculine speech. It is partly for this reason that trans men and other transmasculine people do not often talk about purported differences between men’s and women’s voices – seemingly rejecting awareness as well as control – aside from pitch and the related feature of intonational patterns. In contrast with the few resources on voice masculinization, trans women have a larger body of knowledge on which to draw, including books and other guides produced by trans women themselves (e.g. James Reference James2012, nd; also Laing Reference Laing1989), which call attention to markers like voice quality, resonance, and particular segments linked to gender (e.g. /s/). However, the important fact here is not simply that trans men’s hormonal regimens affect the voice in ways that trans women’s do not. The more significant point is that even Mack, with his still relatively high-pitched voice, subscribes to the ideology that self-conscious masculinization is both unnecessary and, as the next examples will show, undesirable for transmasculine people on testosterone.

The power of testosterone to bring about changes in the voice is not the only reason transmasculine people on testosterone give for their rejection of agentive linguistic masculinization. The second language ideology expressed by the participants in this study centers around the idea that linguistic self-monitoring is contrary to the ultimate goal of a gender role transition: to express an authentic self that is naturally masculine and/or male. Accordingly, when I asked my participants whether they would consider putting in effort to speak in a more masculine way, I was told by several individuals that constant self-monitoring would undermine their desire to more fully express what they feel as an authentic sense of self. Some of these trans speakers felt it was completely unnecessary to alter their speech, gesture, clothing, or any other aspect of their gender expression because they were already quite normatively masculine. Adam, for example, is a 38-year-old, white, queer-identified trans man who grew up in the New York City suburbs and works as a program director for an LGBT youth organization. He is the only speaker in this study whose voice was occasionally perceived as male even before testosterone, which is a relatively unusual experience among trans men. Adam told me that his self-presentation had been masculine for decades by the time he decided to transition medically, and that he never felt that masculinity required effort; it was always the prospect of subduing his masculinity that presented the bigger challenge. His transition-related anxieties centered less around whether he would be perceived as a man and more around losing his status as a visibly queer person as he began to blend in with straight cis men. But even among the participants in my study whose gender presentations and voices were less typically masculine than Adam’s, it was typical to emphasize the value of authentic self-expression over normative masculinity.

When my participants did talk about trying to masculinize their voices, it was always framed as something they did at the beginning of their transitions, before their voices changed from testosterone. It was also often framed as unconscious or unsuccessful, each of which distances the speaker from the notion that they have achieved masculinity only through an active imitation of non-trans men. I had a conversation on this topic about halfway through my year of recording Kyle, a 24-year-old queer trans man who grew up in the Bay Area and was working at a camp for children and pre-teens. During an afternoon we spent chatting in my San Francisco apartment, Kyle told me about “subconscious” attempts he had made in the past to “squash” his natural effeminacy (Excerpt 2, lines 02–05). His efforts were motivated by wanting to be seen as male (line 06), but now that he found himself increasingly achieving that goal, he reports becoming more open to reincorporating femininity into his gender expression, for instance by wearing makeup and hot-pants (i.e. very short shorts) to go dancing at a club. The dialogue in Excerpt 2 comes from a conversation about the concern Kyle shares with Adam: that his affiliation with local queer communities will become invisible now that he is being seen as a man and losing his status as a visibly gender non-conforming person. He relates his self-ascribed effeminacy back to this issue of visibility in lines 09 to 11, in which he refers to a conversation he had with his partner, a femme woman who has also struggled with maintaining visibility as a queer-identified person.

In this context, Kyle situates himself as the agent of these shifts away from femininity in his speaking style and other forms of gendered expression, yet he also mitigates the intentionality behind that agency by pointing out that the change has not come about “consciously, but subconsciously” (line 04). Now that testosterone has changed Kyle’s body to the extent that he achieves equal footing with non-trans men when it comes to the so-called secondary sex characteristics, there is room for his “natural,” authentic effeminacy to come to the surface.

Another participant who talked about trying to speak in a more masculine way towards the start of his transition was Dave, whose commentary illustrates the ways in which trans men who are not stereotypically masculine may draw on other elements of their identities as authorization for ostensibly feminine speech characteristics. Dave is a 23-year-old white, queer trans man from an upper-middle-class Bay Area family, and an artist who was unemployed during most of the duration of my fieldwork. Like Kyle, Dave feels more comfortable expressing femininity than he did before he was consistently perceived as a man, particularly because it fits with his identity as a fem queer man.Footnote ⁷ As Dave puts it in Excerpt 3, “now that I read completely as male, […] I’m just gonna sound like a faggot, it’s fine” (lines 09-10). This selection comes from our final meeting, in which I asked him to reflect back on the ways his voice has changed, or not changed, during his transition so far.

Unlike Kyle, Dave does not take any steps towards softening his agentive responsibility for “very consciously tr[ying]” to “swoop less.” However, he also suggests that it was not realistically possible for him to limit his swoopiness given that his attempts to decrease his pitch range or intonational variability were unsuccessful. Dave distances himself from the person he was before he “read completely as male” and emphasizes that he “definitely stopped” monitoring his prosody the way he used to once he had the security of being recognized as a man.

Dave has also developed an overtly political perspective on the passing tips I mentioned above, which he told me about during one of our earlier meetings. As he sees it, the crux of both passing tips and speech therapy is the idea that a person should change themselves in order to meet others’ expectations about how men and women look, act, and sound. When I asked Dave whether he thought people could be successful in self-monitoring their speech – given that Dave told me his own attempts to “swoop less” did not succeed – he told me that he thinks people can be successful, but that “their success is something that [he] find[s] repulsive” (Excerpt 4, lines 03–04). Because he identifies strongly both as a man and as a fem, Dave is especially concerned with making space for male-identified people whose voices fall well outside of the expectations for heteronormative cis masculinity.

Dave’s comment exemplifies the contradiction with which I am concerned: he wants trans men to have what he calls the “privilege” of being perceived as men, if that is their goal, but he doesn’t want that end to be achieved through limiting their self-expression or assimilating to hetero- and gender-normative standards for appropriate masculinity. The least problematic option left for linguistic masculinization, from this perspective, is testosterone.

Agentive Tensions

As I mentioned in the introduction to this chapter, even as the participants in my fieldwork rejected agentive masculinization of their voices, they also valued agency in a number of ways that are central to their communities’ formulations of trans identity. One of these is in the elevation of self-identified gender category as the ultimate determiner of how an individual’s gender should be understood. What determines that a person is truly a man, in such a framework, is the mere fact that he defines himself as one. This emphasis on self-identification runs against dominant cultural discourses about gender and sex, which depend on external authorities (e.g. scientists, doctors, or religious figures) to define what makes a person female or male. It also runs against many academic theories of identity in that gender is seen as a matter of individual self-determination rather than being relational and co-constructed (e.g. Bucholtz and Hall Reference Bucholtz and Hall2004, Reference Bucholtz and Hall2005). Agency is also highly valued in trans communities in discussions of bodily autonomy and access to medical technologies available for the re-shaping of sexed embodiment, which have historically had to be approved by gate-keeping professionals (Speer and Parsons Reference Speer and Parsons2006). Contrary to clinical expectations that a gender role transition must follow a linear set path of hormones and surgery, trans communities have spent decades advocating for the opportunity to pick and choose which aspects of a medical transition fit each individual’s needs, if any.

Going on testosterone is thus clearly understood as a matter of free choice, but the ensuing pitch changes are not framed as agentive shifts in the same way that conscious stylistic changes are. Although this separation of biology and style may seem intuitive, what is most striking is the fact that, for these speakers, even the choice to begin testosterone therapy somehow avoids the threat to authenticity posed by speech therapy. One could easily make the same argument about testosterone that these trans speakers make about agentive style-shifting. Clearly, hormone therapy is as much a tool of potential assimilation as stylistic changes are, and transmasculine people who opt not to make use of hormone therapy may invoke their desire not to change who they are, how they look, or how they speak just so that others will perceive them as legitimately male, extending Dave’s critique about trying to “swoop less” (Excerpt 3, line 03).

To understand this potential tension, we need to take into account the ways in which discourses of trans identity separate biological sex from the internal, authentic self (see Valentine Reference Valentine2007 for additional history). In these discourses, a person’s “true” gender transcends their biology, which is what makes it possible for people categorized as biologically female to see their self-identification as male as the more authentic representation of who they are (or vice versa). Particularly instructive here is a final quote from Kam, a 30-year-old white, genderqueer, trans boyFootnote ⁸ from working class Cincinnati who was a graduate student and maintained a distinctly non-normative blend of femininity and masculinity. He also strongly resisted the pull to compromise his self-expression in order to “pass” as male, for instance foregoing chest binding despite the fact that this choice made it more likely that he would be perceived as a woman. Like Dave, Kam recognized that his speaking style was far from the masculine norm because of his pitch range and variability, which was amplified by his frequent use of falsetto voice quality; also like Dave, Kam had no interest in changing this aspect of his self-presentation. When I asked him how much his voice had changed after just five weeks on testosterone, he told me that it’s difficult to tell because he gets excited about his voice being lower, which leads him to start “squealing” out of happiness and brings his pitch back up. The talk from Excerpt 5 is what followed.

Kam describes himself as unconcerned with shifting his intonation patterns to the point that, when he notices himself constraining the way he speaks, he asks himself “what am I doing that for?” (line 05), indicating that he questions his motivations for style-shifting towards more typically masculine patterns. Moreover, he asks this question of himself with falsetto voice quality, suggesting that his internal monologue incorporates this aspect of his voice. Repeating the word squeal to refer to his own voice for the second time in just a few moments, Kam says that he’s just a “giggly, bubbly, like, squealy type of person” (lines 25–6) and that he doesn’t see that changing. In so doing, he iconically links his linguistic style to his characterization of himself as an individual. Despite these facts, Kam does not enact this kind of resistance when it comes to changing his voice with testosterone; in fact, he had told me only a few minutes earlier that his sole reason for going on hormones was to change his voice. Based on his account, this decision to go on testosterone did not conflict with his intent to maintain his gender presentation, which he also described as “femmey” and “faggy.”

There is a critical qualitative difference for transmasculine people on testosterone, then, between agentively choosing to modify the body and agentively choosing to modify other forms of semiotic self-presentation. Yet, this apparent contradiction can be made sense of in terms of the separation that trans discourses of identity draw between the body and the authentic self.

Because the vocal changes that come with testosterone therapy are understood as physiological in nature, they do not inhibit expression of one’s true, inner self, just as transmasculine people see their pre-transition, ostensibly “female” bodies as separable from their internally felt masculine identities. Instead, testosterone simply gives transmasculine people access to a physiological baseline that is comparable with what non-trans men have – almost as if this happens through an accident of nature rather than being precipitated by a conscious decision. Once they can make use of a male-sounding pitch range, transmasculine speakers position themselves as free to avoid changing the stylistic elements of their speech and to thereby continue expressing what they experience as an authentic self, while also tapping into the importance non-trans people typically place on gendered embodiment, including the voice, in the attribution of gender.

Although the ideologies I have described are aimed at unseating transphobic representations of trans people’s inauthenticity, they carry with them a dark side in the form of implicit ideologies that these speakers would be unlikely to overtly endorse. Importantly, the split I have identified between body and mind is not enough to account for the ideologies about agency I have discussed. Transfeminine people, after all, create a similar division between embodied sexual characteristics and self-defined gender, yet they tend to have a very different perspective on making self-conscious vocal changes. This is where additional levels of ideology that are not particular to the context of trans identity come into play.

First, the naturalization of masculinity, discussed above, grants these speakers license to construct their genders as effortless even as many trans women are comfortable acknowledging that a degree of effort is put into producing their femininity. Regardless of whether there is an actual difference in the self-conscious behavioral changes made by trans men as compared to trans women, the cultural context provides men with both greater means for constructing their gender as effortless and the demand that they take advantage of these means if they want to authenticate their masculinity.

Second, even as transmasculine speakers challenge the idea that someone with a deep voice must take on all of the stylistic elements normatively associated with masculinity, they also reinforce a perspective on gender differentiation in the voice that is dependent primarily on biology. The ideologies naturalize sexual embodiment despite the investment trans communities have in contesting this naturalization (Zimman Reference 277Zimman, Zimman, Davis and Raclaw2014). It also creates a system of stratification among trans people; that is, if exposure to testosterone is the most important aspect of whether a voice should be considered female or male, trans women would be understood as having naturally male voices, while the voices of trans men who are not on testosterone would be considered female. In this hierarchy, transmasculine people on testosterone would be uniquely positioned among trans people as possessing a match between vocal sex and gender identity.

Trans people are often the focus of scrutiny over the authenticity of their gender presentations and identities. On the one hand, trans people face delegitimation or violence when they fail to live up to normative expectations for the self-identified gender category. At the same time, trans people who do live up to those expectations may be criticized for being excessively normative and hence reinforcing harmful gender stereotypes. The discourses analyzed in this chapter are one way in which speakers work to balance these demands, while also laying claim on the limited forms of embodied gender privilege to which they have access.

Ideology into Practice

A summary of the sociolinguistic styles employed by the speakers I have been discussing is now warranted as a means of exploring how ideologies about agency shape sociolinguistic practice. The sociophonetic arm of this project was focused on documenting change in the voices of the fifteen participants during their first and/or second year on testosterone. As would be expected, all of these speakers saw a decrease in vocal pitch, although to different extents and at different rates of change. A few speakers never reached a normative male pitch range during the time I recorded them; Mack from Excerpt 1 is one example of a speaker whose mean pitch decreased only from approximately 220 to 200 Hz over the course of the year. However, most dropped to an average well below 140 Hz. Much more variability was found, however, in articulatory indices of gender like the spectral qualities of /s/. Zimman (Reference Zimman2012, Reference Zimman and Milani2015) describes the enormous variability in the mean frequencies of /s/, with these fifteen speakers covering the entire range typically reported for American English-speaking men and the entire range typically reported for their female counterparts (i.e. as low as 4,500 Hz and as high as 10,000 Hz; see Flipsen Reference Flipsen, Shrilberg, Weismer, Karlsson and McSweeny1999). While some speakers did undergo statistically significant changes in their articulation of /s/, most did not. Among those who did see some change, a few actually shifted upwards, away from the norm for men, which aligns with speaker self-reports about being more comfortable expressing femininity as their pitch settled into a male-sounding range. The speakers with the lowest pitch voices in the study, including Dave from Excerpts 3 and 4 (with an average below 115 Hz), also had among the highest mean frequency ranges for /s/ (averaging above 9,000 Hz). Indeed, in interviews and ethnographic interactions with nearly 100 transmasculine people in three metropolitan areas in the United States over the last ten years, I have found it not at all uncommon for trans people on testosterone to have low-pitched voices paired with articulatory characteristics commonly associated with femininity or gay male identities (see Zimman Reference Zimman2013 for more on perceived gayness among trans men).

One potential interpretation of these facts is that transmasculine speakers really don’t need to change features of their voice other than pitch in order to be perceived as men on the basis of their voices. However, early results from a perceptual pilot study using modified guises of these participants’ read speech suggest that both formant frequencies and /s/ contribute significantly to how low a speaker’s mean pitch needs to be in order for their voice to be perceived as male in an experimental context. That is, a speaker with a very high frequency /s/ will need to have a lower mean pitch to be perceived as male than another speaker who has a lower frequency /s/. Another potential interpretation of these facts is that trans men are not very good at masculinizing their voices and may be more successful with this process if they worked with a speech therapist. However, once we take into account the rejections of sociolinguistic agency articulated in the discourse presented above, a third explanation emerges in which ideology and sociolinguistic style align. These speakers’ ideologies about sociolinguistic agency are for this reason an essential part of understanding why they employ the particular combinations of sociophonetic features they do. In rejecting control over their voices, trans people on testosterone construct styles that both index an ideological stance and constitute a particular kind of subjectivity invested with moral and political value. Here, stylistic practice and ideology mirror and mutually reinforce one another. Regardless of these speakers’ potential ability to control their voices – a question that is worth exploring further in its own right – the agency they claim for themselves is limited.

Conclusion

The discourse analyzed in this chapter highlights the complex ways in which ideology, power, and hegemony factor into sociolinguistic awareness and control. Silverstein (Reference Silverstein, Mertz and Parmentier1985) proposes that the effect of ideology on structure (i.e. linguistic forms) is one way agency is exercised over language. People need not be explicitly aware of the linguistic details of a change in the formal realm – which in this case stands in for the gendered elements of the voice – in order to exercise power and control over that change. Rather than explicit linguistic awareness, agency is realized through ideologically motivated shifts in the indexical realm, which by extension has an effect on linguistic forms themselves. For transmasculine speakers, ideologies about hormonal sex and gendered authenticity assign particular indexical meanings to the practice of shifting towards gender-based linguistic norms, which then work to shape the actual styles found among trans speakers. Paradoxically, ideologies about the undesirability of linguistic control become the means through which agency is exercised on the structural level of sociophonetic style, even as these speakers overtly constructing themselves as lacking or refusing agency over their voices.

Silverstein focuses on the ways in which ideology can be a means of exercising linguistic agency, but my own analysis suggests that the construction of agency itself is a tool for ideological work, and that claims on or disavowals of awareness and control may be of great importantance for our explanations of social and linguistic practice. While power is often gained through claims on agency, for these speakers it is through the disavowal of sociolinguistic agency that they contest transphobic ideologies about the voice. This is not to say, however, that the discourse analyzed is wholly subversive. Even as they challenge transphobic ideologies about gendered authenticity, they simultaneously lay claim on certain forms of male privilege and reinscribe the naturalization of hormonal effects on the voice in ways that are derived directly from cis-normative systems of oppression and inequality.

A final lesson to be taken from these speakers’ complex and at times seemingly contradictory engagements with agency and control is that these notions are sociocultural constructs that carry weighty implications for systems of power. This point is of particular importance for sociocultural linguists who hope to understand how speakers exercise sociolinguistic agency because it demands that we consider our field’s ideologies about awareness and control and how our speakers’ perspectives might be erased in our analyses in favor of our own. Rather than assuming we know what agency means for our speakers, we could learn a great deal by pairing our sociolinguistic analyses of awareness and control with closer attention to the discourses within which these practices take place. Such attention will challenge linguists to confront the inherent complexity of agency itself much in the way anthropologists have done – to recognize it not as a psychological property of the individual, but as a culturally grounded, ideological experience at the core of what it means to be a speaking subject.

Transcription Conventions

[

overlapping speech

(( ))

non-linguistic action (e.g. laughter, coughing)

=

latching speech (i.e. continued from previous line with no pause)

(.)

brief pause

(word)

uncertainty regarding transcription

(xxx)

indecipherable speech

^word^

falsetto voice quality

“word”

stylized reported speech

?

rising intonation

,

continuing intonation

.

falling intonation

:

lengthened phone