“Erase/rewind”: How transgender Twitter discourses challenge and (re)politicize lesbian identities

Abstract

Competing views on the in/compatibility of transgender status and lesbian identity is a source of conflict in the ongoing antagonism over transgender recognition. Many individuals with different transgender identities might lay claim to lesbian identity or lesbian discourse(s) more generally. However, this inclusion has been disputed in some circles insofar as it is seen to challenge or contradict characteristics of lesbianism. This paper explores how transgender discourses might challenge and (re)politicize lesbianism and lesbian identities. Given that social media platforms concentrate minority communities in one space and can serve to exacerbate antagonism over identities, I focus in this paper specifically on the Twitter context. This paper uses corpus-informed critical discourse studies to explore how cognitive models of lesbianism are articulated in transgender Twitter discourse/s. Findings indicate that transgender Twitter users (re)articulate sociohistorical narratives in lesbian discourse/s. At the same time, however, they also challenge and (re)politicize the essentialism of sex and gender in relation to lesbian identity and social practice. Hence, transgender Twitter discourse/s reflect potential explanations for contesting transinclusion in lesbianism, which may serve to reinforce transexclusionary claims for retaining lesbianism’s uniqueness as a female space and experience.

Keywords:

• Corpus linguistics
• critical discourse studies
• social media
• transgender
• Twitter

Introduction

Contesting the inclusion of transgender individuals and practices within lesbianism and lesbian identities is nothing new, starting with what Halberstam (1998) called the “border wars” of butch-femme identities in the 1960s and ‘70s and culminating in more recent claims of there being a “postlesbianism” of too-inclusive identities (cf. Forstie, 2020). Jeffreys (1997, p. 64) also cites transgender practices of “playing with gender” as contradicting the characteristics of lesbian feminism. However, it is important to note that some feminists consider transexclusion a fringe movement within radical feminism and lesbianism (see Thomsen & Essig, 2021). Indeed, Williams (2016, p. 254) goes as far as to argue that radical feminism and lesbianism are historically transinclusive and that such inclusivity is hidden in favor of a more “popular media narrative.” Most recently, such public conflicts at the intersection of transgender and lesbian identities have manifested in movements and organizations specifically intended to separate sexual identities from gender identities (e.g., the British charity and advocacy group LGB Alliance). Such movements are billed as responses to the inclusiveness of sexual identities, threats of sexual identities becoming obsolete, and an ignorance of the sex differences that underpin sexualities (LGB Alliance, 2021). Whether for or against transinclusion in lesbian spaces, including the abstract space of identity, arguments invariably rely on differential conceptualizations of the interrelatedness of sex, gender, and sexuality.

Antagonism is predicated on “conflict and struggle over identity” (Walton & Boon, 2014, p. 353) and is manifested in divisions between in-groups and out-groups (see Mouffe, 2013; Thomassen, 2005). The sociohistorical narrative of contesting transinclusion is therefore rooted in an antagonism over who can truly lay claim to lesbian identity and in-group status (cf. Beemyn & Eliason, 2016; Hines, 2020). What is more, there has been a recent proliferation of discourses that contest the inclusion of specific voices on given topics, to which the intersection between factions of transgender activism and lesbian feminism has been no stranger (see Hines, 2019). The argument for transinclusion within lesbian identities, discourses, and spaces is one of prioritizing gender identity and self-identification, rather than sex, as the foundation of sexual identity (see Tate, 2012; Tate & Pearson, 2016). This position argues for a de-essentialization of the relationship between assigned sex and sexuality, citing similarities in gendered self-categorization between transgender and gender-congruent women (Tate & Pearson, 2016, p. 105).¹ The argument for transexclusion, then, cites lesbianism as a uniquely female experience (Jeffreys, 2014). Hence, this position contends that the deeply interwoven and unique sociohistorical narratives, struggles, and political victories of lesbianism over patriarchal and heteronormative structures are erased—or at least diluted—by transinclusion in lesbian experience and the subsumption of lesbian experience under catch-all labels like “LGBT” and “queer” (see Morris, 2016). These polarized perspectives constitute an antagonism over the legitimacy of in-group status—and the right to a voice on such matters—based on either gender identity or sexed experience.

Social media platforms have been used as vehicles for constructing and performing identities, including gendered identities, since their inception (see, for example, boyd & Ellison, 2007). Indeed, each platform has its own restrictions, communicative functions, and normative expectations (cf. Schmidt, 2014), which influence users’ identity formation and performance. Generally, Twitter “enables condensed performances of the self” to public and private audiences of various sizes (Papacharissi, 2012, p. 1990). Such performances on Twitter are driven by social identification insofar as Twitter communication is largely predicated on sharing and finding information relevant to “the people and organizations [users] care about” (Puschmann & Burgess, 2014, p. 47). However, it is prudent to note that these affordances are not inherent in the platform’s capabilities, instead arising from users’ engagement with them (see Tagg, 2015). Hence, the affordances of Twitter—and other social media platforms—will vary across and among user-groups. For example, Tandoc et al. (2019, p. 32) found that Singaporean social media users engage with Facebook for group-formation and -organization, whilst using Twitter for “[sharing] their rants and opinions to a smaller and selected group of friends.” Contrastively, Shane-Simpson et al. (2018) found that U.S. college students use Facebook more for private social bonding with other users also known in the offline context and Twitter for more public self-disclosure. Research on transgender users’ Twitter behaviors indicate similar practices, including both public self-disclosures of sexualized identities and practices (Webster, 2018a) and users sharing opinions about civil rights and personal employment issues (Webster, 2018b).

The social homophily facilitated by identity-driven social media platforms, whereby users converge with one another based on social identity and like-mindedness (Kaakinen et al., 2020), has in many cases led to antagonism between groups. On Twitter, for example, “antagonistic discourses [are] emotionally articulated … to negotiate terms of group identities” (Evolvi, 2019, p. 389) in the same way as offline antagonistic discourses are used to construct conflict vis-à-vis groups and out-groups (see Mouffe, 2013). Indeed, Twitter has proven to be a particularly key site of antagonistic discourses between transgender and feminist groups, representing a space within which offline politics of transgender identity recognition are reflected online (see Hines, 2019). What is more, the social recognition of preferred gender identity and its underpinning regulation is reflected in Twitter’s “Hateful Conduct Policy” (Twitter, 2021). Hence, Twitter is an ideal site for exploring the discourses at the intersection of transgender identity and lesbian feminism. Indeed, given their ubiquity, social media have become technologies inextricable from our daily practices and can therefore be seen as heuristics for a general understanding of social and linguistic behaviors. As such, I use Twitter-mediated discourse in this paper as a vehicle for exploring how lesbian identities and practices are articulated by transgender people in the English-speaking global West, which may contribute to an explanatory reasoning for the antagonism over transinclusion in lesbian discourses. I argue that the articulations of lesbianism and lesbian identity on transgender Twitter reflect potential explanations for the antagonism over transinclusion within lesbianism, reflecting sociohistorical narratives of conflict within lesbian feminism and subsuming lesbian experience under catch-all umbrella terminology.

Methodology

This paper uses a corpus-informed approach to critical discourse studies, which quantifies patterns of meaning in large bodies of linguistic data for subsequent in-depth qualitative analysis (Baker et al., 2008). The data analyzed in this paper comprise the Gender-Variant Online Communication (G-VOC) corpus, containing c.3,700,000 tweets (a total of 60,028,867 words) from 2,882 Twitter users. Users’ data were collected from the follower lists of U.S. and U.K. transgender celebrity and charity accounts on the basis of two essential criteria: (1) the user profile was publicly accessible at the time of collection; and (2) users’ biographies included specific linguistic evidence of transgender self-identification. Transgender self-identification was determined by specific linguistic identifiers indexing gender-variance (see Table 1). Following previous research, which indicates that users with different transgender identities engage in markedly different behaviors, the corpus is also divided into six sub-corpora according to users’ gender-similarity based on the identifiers used in the biography (see Webster, 2018a). This distinction between groups of users may serve to illuminate potential explanations for competing understandings of gender- and sex-based identities that characterize the conflict over transgender-inclusion in the social categorization of lesbianism.

Table 1. Division of G-VOC corpus into gender-based sub-corpora.

User-group	Number of users	Sub-corpus size (tokens)	Example identifiers*
Transfeminine	992	21,489,758	“mtf,” “m2f,” “transwoman,” “transwoman”
Transgender	615	12,444,491	“trans,” “transgender,” “transperson”
Transmasculine	463	8,107,698	“ftm,” “f2m,” “transman,” “transmasc,” “transguy”
Non-binary	364	11,155,668	“non binary,” “b,” “enby”
Transvestite	277	2,809,300	“TV,” “transvestite,” “crossdresser,” “CD,” “XD”
Transsexual	171	3,021,952	“TS,” “transsexual,” “shemale”

* List not exhaustive, but exemplary.

The first stage of a corpus-informed critical discourse analysis requires the use of quantitative corpus tools, which are used to identify salient topics or themes within the data (see Gabrielatos & Baker, 2008) via an analysis of keywords and collocates. Keywords are the lexemes “most indicative (or characteristic) of one corpus, as compared to another” (Rayson & Garside, 2000, p. 3), which requires a reference corpus. The reference corpus used for comparison in this thesis is a sample of 440,154,502 tweets from the Stanford Large Network Data collection corpus (Leskovec & Krevl, 2014; see also Webster, 2018b). Keyness is measured using log-likelihood ratio, an inferential statistic; the log-likelihood of a keyword must be ±3.84 in order to be deemed statistically significant at the 95th percentile or p < 0.05. Collocates, then, are two words within a corpus between which there is an “above-chance frequent co-occurrence” (Baker et al., 2008, p. 278). True collocates are those that score higher than a conventional threshold of significance on at least two measures (Baker, 2014). This paper uses both mutual information (MI) and t-scores as measures of collocation. MI measures the “probability of observing [two words] together” and has a conventional significance threshold of 3 (Church & Hanks, 1990, p. 23), and t-scores are a confidence measure of the “certainty of collocations” with a conventional significance of 2 (Hunston, 2002, p. 73). As such, thecollocates analyzed in this paper are those that score higher than the conventional significance threshold for both Mutual Information (MI) and t-score.

Following quantitative analysis, the patterns of language identified are explored qualitatively as meaningful conceptualizations of the social world that are shared within an epistemic community. Thus, a qualitative approach qua sociocognitive critical discourse studies (Koller, 2012; van Dijk, 2009) is applied to the collocates of keywords in the G-VOC corpus as a means of identifying the shared cognitive models of transgender Twitter users. Topics and themes are identified in quantitative findings, which illuminate the ideas, interests, and values of the epistemic community under analysis (cf. Van Dijk, 2015). These topics and themes are constrained by the local discourse context of transgender discourse, accounting for a contextualized understanding of the cognitive models at work that take into account individuals’ subjectivities and social positioning (see Webster, 2018a). This sociocognitive approach illuminates the “cognitive interface” that “[influences] social structures” (Van Dijk, 2015, p. 64), offering an initial foundation for an explanatory critique of the relations held between the social categorizations of transgender and lesbian in an age of antagonism over their interrelatedness.

Analysis

The words lesbian, lesbians, and lesbianism are each statistically significant keywords in the Gender-Variant Online Communication (G-VOC) corpus, when compared to the reference corpus of general Twitter use. The same is true for butch*, femme*, and dyke* (see Table 2).²

Table 2. Lesbian-indexing keywords in the G-VOC corpus.

Keyword	Freq (G-VOC)	Freq (Stanford)	Log-likelihood (LL)
Butch*	918	1,021	+1280.57
Dyke*	722	2,583	+210.21
Femme*	4,512	1,975	+1168.31
Lesbian*	7673	7,225	+12965.33

When accounting for users’ gender similarity, it is evident that there are both similarities and differences in the distribution of lesbian-indexing keywords across user-groups (see Table 3). Keywords lesbian*, butch*, and femme* are more frequently used in all the G-VOC sub-corpora, when compared against the reference corpus, each with a statistical significance at the 99.99th percentile (or p < 0.0001).³ The greater use of dyke* is statistically significant at the 99.99th percentile in the transfeminine, transgender, transsexual, and non-binary sub-corpora. Greater use of dyke* in the transmasculine sub-corpus is significant at the 99.9th percentile (or p < 0.001). The word dyke* is statistically significantly underused at the 95th percentile (or p < 0.05) in the transvestite sub-corpus.

Table 3. Comparison of lesbian-indexing keywords between gender-based sub-corpora.

	Butch *		Dyke *		Femme *		Lesbian *
	Freq (LL)	Prevalence (% of users)	Freq (LL)	Prevalence (% of users)	Freq (LL)	Prevalence (% of users)	Freq (LL)	Prevalence (% of users)
Transfeminine	259 (+ 370.5)	12%	294 (+ 103)	10%	1487 (+ 4403.6)	24%	2941 (+ 6172.73)	50%
Transgender	303 (+734.39)	15%	184 (+83.99)	12%	1895 (+ 7854.09)	21%	1673 (+ 3419.72)	51%
Transmasculine	133 (+ 264.96)	14%	79 (+ 13)	8%	213 (+ 361.58)	15%	1067 (+2338.8)	46%
Non-binary	190 (+ 383.63)	23%	110 (+ 18.86)	16%	586 (+1611.19)	42%	1201 (+ 2216.45)	66%
Transvestite	6 (- 0.14)	2%	10 (- 3.89)	2%	242 (+ 915.47)	20%	339 (+ 716.49)	28%
Transsexual	27 (+ 29.84)	11%	45 (+ 25.46)	11%	89 (+ 171.27)	15%	452 (+ 1117.73)	38%

These findings indicate that multiple articulations of lesbian identity, whether self-identification or other-representations, are characteristic of all user-groups when compared to the reference corpus of general Twitter use. Regardless of whether users self-identify as lesbians, there is a clear indication that transgender users of all gender-sex configurations have assumed a position within lesbian discourses, which is a contested aspect at the intersection of lesbian and transgender identities (cf. Hines, 2019). A more detailed exploration of the collocates and concordances of keywords, taking into account users’ gender-similarity, illuminate to what extent these articulations challenge or re-politicize lesbian discourses.

Lesbian*

The non-binary sub-corpus has the highest proportion of users articulating lesbian* keywords (66%, or 241 of 364 users), followed by the transgender sub-corpus (51%, or 211 of 615 users) and the transfeminine sub-corpus (50%, or 494 of 992 users).

It may be expected that a greater proportion of transfeminine users would articulate discourses surrounding lesbianism, given the arguably inextricable relationship between womanhood and lesbianism. Indeed, the transgender sub-corpus—which includes users who explicitly self-identify as transgender but do not index either femininity or masculinity in their biographies (cf. Webster, 2018a)—is also likely to include more transfeminine than transmasculine users. That is, demographic research indicates that transfeminine identities are more prevalent than transmasculine identities (cf. Webster, 2018b). The prevalence of lesbian* use among non-binary users is therefore perhaps unexpected, though non-binary individuals are included in some conceptualizations of lesbianism (see Tate & Pearson, 2016). Indeed, the prevalence of transmasculine users’ discursive constructions of lesbianism (48%, or 211 of 463 users) might also be surprising. However, narratives of transmasculine experience indicate “involvement in … lesbian communities” prior to transition (Hines, 2019, p. 145) and inclusive conceptualizations of lesbianism allow that “transgender men can also lay claim to … lesbian community” (Forstie, 2020, p. 1764).

The collocates of lesbian* provide potential explanations for the over-representation of lesbian* in all user-groups. There is an evident consistency among user-groups in the 10 most frequent conventionally significant collocates of lesbian* (see Table 4). Consistent themes identified across sub-corpora can be broadly conceptualized under the categories of: (1) sexuality; (2) gender; and (3) pronouns.⁴ Typical examples of these collocations include strings of often—but not always—hashtagged identifiers, users subsuming lesbianism under a queer umbrella, or equating lesbian and queer identities in some way. Some examples from the data include:

Table 4. Comparison of collocates of lesbian* between gender-based sub-corpora.

	Trans-feminine	Transgender	Trans-masculine	Non-binary	Transvestite	Transsexual
Lesbian*	I, gay, lesbian	gay, transgender	gay, I, trans	gay, I, trans	gay, LGBT	I, gay, trans
	trans, lesbians	bisexual, LGBT	new, LGBT	bisexual, you	lesfic, I	out, you
	you, transgender	I, queer, trans	bisexual	not, transgender	hot, here	transgender, like
	my, bisexual	women, my	transgender, out	women, bi	transgender	bisexual, women
	women, who, bi, can, love, woman	Black, gays, bi, out, lesbian, people	can, novel, not, check, reads, women, couple	my, out, queer people, me couple	love, bisexual, more, lesbian, click, free, sexy	hot, check, not, my, love, couples

#LGBT, #lesbian, #gay, #transgender

queer women, including bi and lesbian …

queer/lesbian

lesbian and queer women

The collocation of multiple identifiers may indicate multiple—both contradictory and complementary—understandings of the relationship between lesbian and other identities. For example, listing identifiers alongside one another indicates an understanding of shared interest between several social groups. This is especially the case of hashtagged exchanges, which are often used to generate conversational publics outside of personal follower/following audiences on Twitter (see Bruns & Moe, 2014). Hence, the listing of identifiers alongside one another may not signal conflated identities, but simply like-mindedness (cf. Kaakinen, 2020). On the other hand, queer is variably used as an overarching category subsuming non-heterosexual practice and, separately, as a term interchangeable with or alternative to lesbian. This may reflect arguments that lesbian experience is subsumed by—or, at least, similar to—other identity labels (cf. Morris, 2016).

The collocation of trans and transgender with lesbian* in all sub-corpora is not surprising, given the research context. However, it does indicate a consistent and strong collocation between transgender and lesbian identities or practices within the cognitive model of transgender Twitter, regardless of gendered user-group. Indeed, there is specifically indicated a shared cognitive model of contiguity between multiple non-heterosexual identities (and transgender identities), which mirrors the wider hegemonic collectivization of identities under umbrella concepts of non-normativity (cf. Webster, 2018a). This consistent frequent collocation could arguably be seen to either reject or obscure the specificities and uniqueness of lesbian experience by equating it to gay, bisexual, transgender, and/or queer experience. That is, there is indicated a shared cognitive model between transgender Twitter user-groups that lesbianism is not a uniquely female experience, but a constituent element of wider generic constellations of non-heteronormativity.

Femme* and butch*

Again, the user-group with the highest prevalence of femme* among its users is the non-binary group (42% – or 152 of 364 users), followed by transfeminine (24% – or 243 of 992 users) and transgender (21% – or 131 of 615 users) user-groups, respectively. Butch* is also most prevalent among non-binary users (23% – or 82 of 364 users), though much less than their use of femme*.

The over-representation of transfemininity among transgender identities might also go some way toward explaining the greater frequency of femme* over butch* in all user-groups. The differential representation of femme* and butch* could indicate a shared cognitive model between transgender Twitter users that prizes—or, at least, highlights—feminine identities. Indeed, it may also index a shared cognitive model of transgender hyperfemininity that is a source of antagonism for some critics of transinclusion in lesbianism and feminism (cf. Berberick, 2018).

The difference in the prevalence of butch* among all other user-groups, except the transvestite sub-corpus wherein butch* was underused, is minimal. Nevertheless, the statistically significant use of both butch* and femme* in all user-groups indicates a general self-positioning of users within discourses of historically lesbian-indexing identities. That is, they are articulating their positioning in a discursive space that is contested by others (cf. Hines, 2019). The incorporation of both butch and femme identities arguably also signifies an adherence to historical distinctions between heteronormative lesbian identities (cf. Koller, 2008), which have been the source of some ideological conflict within politicized lesbian communities (cf. Halberstam, 1998). An analysis of collocates of both femme* and butch* offer potential explanations for and implications of the use of these historically lesbian-indexing identity markers (see Table 5).⁵

Table 5. Comparison of collocates of femme* and butch* between gender-based sub-corpora.

	Transfeminine	Transgender	Transmasculine	Non-binary
Femme *	I, trans, transgender	en, I, femme	I, trans, butch,	I, trans, my, you
	you, my, femme	me, trans, you	you, gay, ftm,	women, femme
	women	my, guise, women,	stud, lesbian, queer, LGBT	black, queer
	genderqueer, me	transform,	me, women,	love, people
	non-binary, queer more, like, your black	butch, queer, sexy, black, tomboy	people, tomboy, femme	color, like, they, we, out
Butch *	I, femme, my, trans	I, trans, fairy,	I, femme, lesbian	I, trans, women
	women, lesbians	femme, you, handsome	stud, ftm, trans	butch, femme, buff
	lesbian, woman	me, birthday	gay, not	stone, exist, my
	me, you, like, up, butch, stone, she	women, stone, blues, lesbian, benefit, today, lesbians	transman, LGBT, my	me, you, lesbian, blues

Butch* seems to retain its specifically lesbian- and female-indexing foundations in transgender Twitter discourse, given its consistent collocation with woman* and lesbian*. That is, most collocates of butch* fit broadly into two categories: (1) lesbianism; and (2) gender.⁶ However, there is a frequent co-occurrence of trans* and butch* among all user-groups. As such, uses of butch* in context include the expected phrases “butch lesbians” and “butch women” alongside another three-word phrase: “butch transwomen.” The latter suggests there is an alternative use of butch* among transgender Twitter users that is not necessarily associated with sexuality, in much the same ways as femme* has been used in transgender discourses (cf. Webster, 2018a). Indeed, this is perhaps reflective of a re-appropriation of historically lesbian identity markers by transgender users.

Femme* also appears to have been re-contextualized in transgender Twitter discourses, albeit slightly differently. The collocation of femme* with umbrella categories of non-normative gender and sexuality configurations, including genderqueer, queer, and LGBT, occurs among each user-group. Indeed, typical collocations in context include the identifier phrases “queer femme” and lists of identity markers similar to those in uses of lesbian* (e.g., “genderqueer, femme, vegan, fat”). The specific collocation of queer-indexing language with femme* indicates the categorization of femme* as a queer, rather than specifically lesbian identity. Indeed, its general use among transgender Twitter users use is far more closely linked with gender-indexicality than with sexuality.

Subsuming historically lesbian identities under new queer categorizations is exactly what Forstie (2020) claims underpins the postlesbian fear of “too-inclusive” identities. Indeed, the co-opting of butch and femme identities corresponds with Rothblum’s (2010) findings that queer identities correspond with either butch or femme identification. This may have significant implications for butch/femme distinctions that have characterized historical lesbian discourse (cf. Koller, 2008). Instead of simply reinforcing heteronormativity (cf. Beemyn & Eliason, 2016), butch* and femme* identities may be seen to simply be “‘playing’ with gender” (Jeffreys, 1997, p. 64) under generically non-normative sexual and gendered umbrella categorizations. However, the prevalence of femme* over butch* is arguably a reversal of sociohistorical narratives of femme invisibility in lesbian communities (cf. Eves, 2004). This asymmetrical re-contextualization of identifiers may therefore contribute to a perception of transgender Twitter discourse(s) challenging and re-politicizing historically lesbian identities.

Dyke*

Dyke* constructions are also most prevalent among non-binary users (16% – or 57 of 364 users). This is followed by transgender (12% – or 76 of 615), transsexual (11% – or 18 of 171 users), transfeminine (10% – or 103 of 992 users), and transmasculine (8% – of 38 of 463 users) users, respectively.⁷ This may indicate a problematic appropriation or use of a historically lesbophobic epithet, even if used as a reclaimed lesbian identity marker (cf. Jones, 2012), which could easily be considered a potential explanation of antagonism (cf. Mouffe, 2013; Walton & Boon, 2014). Indeed, the collocates of dyke* may be the most directly reflective of the antagonism surrounding claims to lesbian-indexing identities out of the terms identified in the G-VOC corpus (see Table 6).

Table 6. Comparison of collocates of dyke* between gender-based sub-corpora.

	Trans-feminine	Transgender	Trans-masculine	Non-binary	Transsexual
Dyke *	I, march, trans, out	march, out, watch	trans	march	pansexual
	NY, Chicago, who	trans, dykes	via	trans	quiet
	dyke, hard	butch, Chicago	stories	you	watch
	they, me, like, not, watch, my	dyke, you		butch, Chicago	out

Across the transfeminine, transgender, and non-binary sub-corpora, there is collocation between dyke* and march. The immediate collocation of the two is in reference to protest marches designed to highlight lesbian visibility (cf. Podmore, 2016) and challenge the “male-focused nature” of gay pride events (Currans, 2012, p. 74). Indeed, typical contexts for this collocation reflect either celebratory (e.g., “My first Dyke March”) or critical (e.g., “anti-Semitic Dyke March”) comments on these specific events. What is more, the collocation of dyke* with trans shows a specific articulation of “Dyke Marches” being explicitly transinclusive (i.e., “Dyke & Trans March” or “Dyke/Trans March”).

“Dyke Marches” have specifically been centered in conflicts at the intersection of transgender and lesbian identity (Hines, 2019). This articulation of transgender inclusion in such practices—or, at least, self-positioning in discourses surrounding them—among several user-groups with a seeming prevalence of “femme” identities (and not by the transmasculine user-group) is perhaps reflective of a contemporary re-contextualization of what dyke represents. From indexing a specifically lesbian identity (cf. Jones, 2012, 2014), dyke now encompasses a postlesbian inclusion of queer femininity (cf. Forstie, 2020). What was a radical confrontation of the heteropatriarchal hegemony and a reclamation of lesbophobic language (cf. Currans, 2012) is now deployed in a form inclusive of wider constellations of sexual and gendered identities. Hence, the use of dyke* in transgender Twitter could arguably be considered demonstrative of the very erasure of the uniquely female lesbian experience (cf. Jeffreys, 2014) that the term was designed to challenge. It is, then, arguable that the shift in meaning of dyke—or, at least, the inclusion of transidentities within protest movements using the epithet—offers a potential explanation for explicitly transexclusionary lesbian discourses. That is, the transinclusiveness of politicized protest leads to an equally politicized reaction of transexclusion.

Conclusion

This paper provides some preliminary evidence that transgender Twitter discourses challenge and (re)politicize lesbian identities in a way that may serve to underpin transexclusionary narratives. That is, because transgender Twitter users appear to articulate lesbian identities and lesbianism in such a way that directly mirrors the fears and issues raised by transexclusionary lesbian discourses, there is some foundational evidence that this antagonism over lesbian identities is not made of straw. The findings of this preliminary corpus analysis indicate that lesbian identities are subsumed under wider generic constellations of non-heteronormativity, which transexclusionary discourses claim contribute to lesbian erasure. Similarly, historically lesbian-indexing identities are asymmetrically re-contextualized solely as gendered, rather than sexualized, with historical discourses of femme invisibility somewhat reversed. This re-contextualization arguably reflects the hyperfemininity and heteronormativity problematized in historical discourses of lesbian feminism. That is, the prevalence of transfemininity among transgender identities contributes to the asymmetrical re-contextualization of historically lesbian identities into normative markers of—largely binary and heteronormative—gender expression. Finally, radical terminology used to challenge heteropatriarchal structures of oppression in the name of uniquely female experience have been re-appropriated as inclusive of—or, at least, the voices of—all feminine identities, regardless of lived female experience.

However, these findings can also be interpreted through a transpositive lens wherein identities are articulated along the lines of historical inclusivity within some radical lesbian and feminist movements. That is, the prioritization of gender-driven, rather than sexed, identities may reflect a de-essentialization of biology and physiology in the construction of or claim to identity. More specifically, claims to both butch and femme identities among transgender individuals reflects a continuation—or, perhaps, evolution—of historically lesbian identities and the similarity between transgender and gender-congruent women in self-categorization, which is the basis of transinclusionary narratives. Similarly, the alignment of lesbianism among wider constellations of non-heteronormative identities and practice may reflect shared political interest and lived experience (e.g., as subjects of patriarchal oppression). This is reinforced by the inclusion of transgender voices in perhaps the most radical of feminist and lesbian spaces, which indicates a presumed intention of transgender users to engage in political practices of resistance against systemic power inequalities that negatively affect both transgender and gender-congruent lesbians.

Of course, there are limitations to the conclusions that can be drawn from this preliminary and exploratory study. Indeed, as with any study, I do not claim that the dataset used in this paper is exhaustive of all possible transgender articulations of lesbian identity and lesbianism. For example, Twitter users who identify as transgender but do not explicitly indicate as much in their biographies are not included in this dataset, which may well have some impact on the findings herein. Similarly, the very nature of social media platforms means that so-called “trolls” or fake accounts may be included in the dataset. However, it is not within my gift to determine who is and who is not truly transgender based on the language they use and there are no reliable measures as yet for identifying “trolls” via user biographies alone. As such, this paper is intended simply as an initial attempt at exploring the long-standing antagonism over transgender inclusion in lesbian discourse(s) as it relates to lesbian obsolescence. In doing so within the limited scope and scale of one paper, there are nuances of context that cannot be fully considered by the corpus-driven methods relied upon herein. Further research should explore the nuances of transgender articulations of lesbianism using more in-depth qualitative analysis of texts-in-context to identify how these articulations are manifested and legitimized discursively as representations of actors and their actions.

I do not claim to have the answers to resolving the long-standing issue of this antagonism at the intersection of lesbianism and transgender identity. However, such articulations of lesbian identities as identified in this paper may provide either the foundations of an explanatory reasoning for transexclusionary narratives or evidence of the inextricability of trans and lesbian experience, depending on the interpretation taken. They certainly have implications for the question of lesbian obsolescence, especially at the intersection between transgender and lesbian identities. Preliminary evidence points to the notion that the conceptualization of lesbianism (and, perhaps, specifically lesbian feminism) as a uniquely female experience of sexual and politicized identity is obsolete when transinclusive. However, the term lesbian and its associated identity markers—whether transinclusive or not—are not obsolete in the sense that they are used as heuristics for indexing both feminine and non-heterosexual identities.

Contesting transinclusion in lesbianism and lesbian identities is not likely to end anytime soon, insofar as antagonism offers no compromise and neither “side” is likely to lay down their arms. However, research that explores potential explanations for sociopolitical issues can provide the foundations for reforming the battlegrounds upon which claims to recognition are fought. Only from a thorough understanding of both sides of antagonism can change be theorized. This paper simply aims to lay one stone in the endeavor.

Declaration of interest statement

The author declares that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The research and data collection underpinning this publication were conducted under the auspices of Lancaster University as part of a larger-scale doctoral project, though this paper was later written whilst the author was employed at Manchester Metropolitan University.

Funding

The work conducted under the auspices of Lancaster University was funded by Lancaster University’s Faculty of Arts and Social Sciences Postgraduate Studentship.

Additional information

Notes on contributors

Lexi Webster

Dr. Lexi Webster is a Lecturer in Linguistics at Manchester Metropolitan University. Lexi’s research focuses primarily on corpus-driven critical discourse studies, focusing on the implications that identity construction/s and cognitive models have for actors, institutions, and social structures.

Notes

1 The term cisgender is not used in this paper, due to its contested nature.

2 The asterisks on butch*, femme*, and dyke* refer to truncation, indicating the inclusion of both the singular and plural form of the terms. Lesbian*, as it is used in the remainder of the paper, includes each lesbian, lesbians, and lesbianism.

3 Log-likelihood ratios of ±3.84 are significant to the 95th percentile, ratios of ±6.63 to the 99th percentile, ratios of ±10.83 to the 99.9th percentile and ratios of ±15.13 to the 99.99th percentile (see Rayson et al., 2004).

4 There is also indicated a sexualization of lesbianism in the transvestite and transsexual sub-corpora via sexy and hot (cf. Webster, 2018a), though an exploration of this phenomenon is beyond the scope of this paper.

5 There were no significant collocates for either femme* or butch* in the transsexual and transvestite sub-corpora.

6 Indeed, the inclusion of stone and blues as across the sub-corpora may refer to Stone Butch Blues (Feinberg, 1993).

7 Dyke* is statistically significantly under-represented in the transvestite sub-corpus.