“I am I”: Self-constructed transgender identities in internet-mediated forum communication

Abstract This article analyses identity constructions and representations of self-identifying transgender individuals on a web-based forum. Although the forum is aimed towards all transgender users, the primary user-group are transfeminine users (intending on) undergoing medico-surgical interventions to align their physiology and identity. The data for this analysis are initial text posts from the forum board used for introductions (i.e. new users of the forum introducing themselves). The article assumes that introductions are the context in which one asserts key identity features; hence, this board is the most pertinent for analysing identity construction. In this article, I use a combination of corpus linguistics and Critical Discourse Studies tools to analyse the use of pronouns and gender-indexical nouns in identity constructions and the representation of social categorisations. This article is an attempt to demonstrate that transgender is not a collective homogeneous identity, and that gender-sex incongruence may not be a salient identity feature for some forum-users. I also examine the ideologies (re)produced in the local forum-communication discourse, and the evaluation of hegemonic practices within transgender discourse and wider gender discourse to further demonstrate the heterogeneity of transgender identity.


Introduction
In broader discussions of gender and identity, including academic research, transgender discourse and identities are considered hegemonically non-normative (e.g. Hines 2006;Horn et al. 2009;Galupo et al. 2014). This article seeks to challenge that notion and attempts to close a gap in current research. I explore the ways in which transgender persons construct identities, represent cognitive models (Lakoff 1987) and ideologies (both individual [Webster 2015] and hegemonic [Gramsci 1971]), and locate themselves within discoursesin terms of social power, and in relation to others (Turner et al. 1987). This article seeks to demonstrate the heterogeneity of transgender identities by critically analysing the discursive strategies used in the construction and representation of identities in internet-mediated communication.
I aim to dispel the myth of homogeneous transgender identity by repudiating the ideological position that there is a singular "type" of transgender identity. I suggest that transgender identities are constructed not as being within a community of singular gendered identity, or social categorisation (Turner 2010(Turner [1982), but as an identity comprising individualistic practices and cognitive models. I do, however, expect some form of identity-relationship within groups of transgender individuals based on concrete communicative practices (i.e. participation in forum-communication groups). The approach I take assumes that a transgender identity comprises individualised socio-psychological identity practices, and that the social categorisation of transgender is not based on a homogeneous identity type.
Additional to practices of transgender in relation to traditional gender discourse are hegemonic ideological social practices specific to transgender discourse, namely: medico-surgical intervention, contentment with gender-sex congruence and politico-legal gender reassignment. The following analysis of discursively represented social practices of transgender, and evaluations thereof, allows for further explorations of socio-political problems within transgender discourse. It also poses questions for further research that may seek to normalise transgender and de-problematise particular social practices within and around transgender discourse. The primary purpose of this article is the normalisation of transgender via the demonstration of the heterogeneity of transgender identities (much like the now-normalised heterogeneity of cisgender female identities).

(Social) identity
Identity is synonymous with sameness (Bucholtz and Hall 2004). Therefore, when a person constructs identity, they are constructing sameness with (or, similarity to) a socio-cultural, -political or -psychological group, attribute or characteristic. Scott (2001) suggests that identity does not "follow predictably or naturally" (Scott 2001: 285) from such social categories as gender or ethnicity, nor from biological categories as race or sex. Instead, it is largely considered true that identity is "generated and reproduced through discourse" (Wodak et al. 1999: 186) and is, rather, the "emergent product" (Bucholtz and Hall 2005: 588) of discursive interaction. Bucholtz and Hall (2004) also claim that identity is constructed through specific semiotic processes: practice, indexicality, ideology and performance (Bucholtz and Hall 2004). These processes are reflected in habitual social action, the association and co-occurrence (of identity practices), organisation of beliefs/values, and "deliberate […] social display" (Bucholtz and Hall 2004: 380), respectively. Each are individually experienced, interpreted and performed. They are, therefore, unlikely to be identical between groups (however small).
Identity is also constructed and represented in the discursive location of the self, or others, within a particular discoursei.e. a social identity. Social identity, then, is a personal "knowledge" of belonging to a social group (Tajfel 1972: 31). Social categorisations are the distinctly divided social groups (Tajfel 1972) with which one may have social identity. "Locating oneself, or another person […] to any social categorization" (Turner 2010: 17-18) is the process of constructing social identity (i.e. self-categorisation [Turner et al. 1987]): whether one belongs, or not, to a collective group identity. For the purpose of this article, self-and other-categorisation will be used interchangeably with identification (see Stets and Burke 2000: 224).
Social identity constructions are an appropriate exploration for demonstrating heterogeneity within transgender discourse, and gender discourses, more broadly: by determining whether discourse participants locate the self or others as within a social group, the ideological effects of social identity construction can be analysed and their social implications illuminated.

Transgender identities (on the internet)
The interplay between trans-and cisgender is relatively inextricable: the very existence of socially conventionalised gender-sex congruence necessitates that those with gender-sex incongruence are a "non-normative" minority. Theories of "minority self-identity" (Eliason and Schope 2007: 4) were produced by Goffman (2009Goffman ( [1963) in Stigma, which suggest that those with "non-normative" identities come to internalise abnormality. Evidence for the same is present in the data for this article. Goffman's (2009Goffman's ( [1963) theory that the stigmatised seek to "correct [their condition]" by "mastery of the area" (Goffman 2009(Goffman [1963: 10) most closely related to the stigma seems to drive much research on transgender (i.e. research on femininity or masculinity akin to that of cisgender females or males, respectively). Previous research has explored the construction of gender identity formation (Gagné et al. 1997) and development (Morgan and Stevens 2008) in transgender individuals. There has also been research into transgender as a shared, "quasi-ethnic", identity (Broad 2002). However, few critical analyses have been conducted in an effort to explore the intersocial identities of transgender practice.
The relatively new phenomenon of the internet, and internet-mediated research, has already coincided with research on transgender discourse and identities. Shapiro's (2010) Gender circuits looks at identity in the "technological age" and, though the focus is neither critically analytic nor specifically on transgender, there is discussion of non-normative gender identity and gender variance in virtual reality programmes and computer-mediated communication.
In their work Transgendering identities (2010), Ekins and King discuss the concepts of "male femaling" and "female maling" (Ekins and King 2010: 25) on blogs and medical resource sites, directly linking transgender to cisgender practices and further reproducing the notion of performativity in gender. However, the focus is on "taboo" (Ekins and King 2010: 28) transgender identities, once again reinforcing and perpetuating the non-normativity and social negativity of transgender identities. Marciano's (2014) Living the VirtuReal looks further, at the use of the internet by transgender people, and looks primarily at what it means to be transgender online; it examines how the online and offline identities within transgender are discursively mediated.
Though the aforementioned works provide initial attempts at analysing transgender identity practices in internet-mediated communication, I aim to go further than their remit. By analysing a corpus of text data written by transgender forum-users from a critical perspective, I intend to illuminate the discursive strategies of identity construction and representation and demonstrate the heterogeneity of transgender identity. The following section outlines the approach used for the analysis in order to achieve this purpose.

Methodological approach
Critical Discourse Analysis (CDA) is "a problem-oriented interdisciplinary research movement" and is primarily focused on analysing "semiotic dimensions of power" and seeking "political-economic or cultural change in society" (Fairclough et al. 2011: 357). The "critical" component of such an interdisciplinary approach to research is derived from critique: the practice of questioning hegemonic ideals. CDA seeks to illuminate social power demarcations and hegemonic practices within a specific discourse in an effort to expose and reconstruct social practice. Hence, a critical perspective on the analysis of transgender identity representations and constructions is appropriate in order to effect this article's aim: to emancipate heterogeneous transgender identities from the hegemonic social practice of homogenising minority identities.
A corpus, then, is a "large and principled collection of natural texts" (Biber et al. 1998: 4) and can be constructed from any communication method and discourse area. Corpora are used for gathering quantitative data and providing statistical analyses, but the analysis of text corpora "often includes qualitative analysis" (Baker et al. 2008: 274). Using corpora in CDA enables not only the sophisticated analysis of large amounts of data that would make manual analysis near impossible, but also provides quantitative support to qualitative arguments.
In analysing the comparative frequency of items in a corpus, I use loglikelihood calculations. Log-likelihood is a calculation that determines the statistical significance of over-/underuse of lexical items in one corpus, compared to another. However, in this article, I am using the same corpus in each calculation in an effort to determine the statistical significance of the overuse of specific lexical items over others within the same corpus data. For the purposes of this article, I consider p ≤ 0.05 to indicate statistical significance; lower p-values (e.g. p < 0.01; p < 0.001) indicate results of higher statistical significance. In addition to analysing the frequency of lexical items in the corpus data, I also examine the lexical items with which they often co-occur and their use in context.

Corpus data
The data comprise a corpus of text posts taken from a forum aimed at transgender users and moderated by transgender individuals. The authors of the web-based data collected and used in the corpus are most frequently those born physiologically male. This reflects research regarding transgender prevalence, which "repeatedly demonstrate[s]" that a higher number of physiologically male persons partake in gender transition practices than do physiologically female persons (Meier and Labuski 2013: 297). This user demographic is indicative of a hegemonic practice within transgender, where transgender-masculine identities and personswho are most often born physiologically femaleare "all but invisible" (Kellaway 2014).
The forum comprises several boards and topics relating specifically to transgender experience (i.e. "coming out", visibility, relationships) and several others that are not specific to transgender experience (i.e. music, television, games). Despite the range of boards available, the text posts used in building the corpus are from the user introductions board: I assume introductory posts will include the most salient identity constructions for posts' authors and, by extension, the most salient materials for an analysis of identity presentation. I also assume authors will represent identities most salient to themselves and locate themselves in terms of transgender as a social categorisation, due to the context and audience of the forum. Only initial introductory posts were used for the corpus; no replies were included due to the aforementioned assumed salience of introductory text posts.
The corpus comprises text data mined from the forum and contains 136,334 tokens. The posting time of the data ranges from April 2012 to December 2014: In order that users' anonymity remains intact, the website and usernames will not be printed, nor will there be direct quotations made from the forum/corpus (see British Psychological Society, Ethics guidelines for internet-mediated research [Hewson et al. (2013)]) that might render users identifiable. Where necessary in the following analysis, I will accurately paraphrase.
6 Data analysis

Pronoun use
Pronouns can indicate number, gender and person. They may also be indicative of social closeness/distance between participants of social practice. Halliday's (2002) interpersonal metafunction describes the "socially meaningful participant relationships" (Halliday 2002: 55) in discourse and places pronouns within the framework of relationship construction: pronoun use can indicate the author's "perception of the relationship between the [reader] and himself" (Feng and Liu 2010: 826). Pronouns are, arguably, the single most appropriate linguistic system for the analysis of social identity construction and representation.

The self and self-inclusive groups
By far, the most frequent pronoun in use within the corpus data is I, with 6,825 occurrences. Table 1 shows the significance of the use of I in the data comparative to other subjective pronouns and is indicative of the significance of its use in the corpus. Each of the log-likelihood scores indicate a statistical significance to the 99.99th percentile (p < 0.0001, where the critical value is 15.13) that I is overused in comparison with all other subjective pronouns in the corpus data.
The statistically significant use of I over we may be interpreted as purely reflective of the board's purpose: introductions. However, the assumed audience is other transgender individuals, with whom there is some expectation of an identity-relationship. Thus, the vast overuse of I becomes discursively significant. The frequencies of pronoun use in another corpus of posts (90,374 tokens) taken from a different board on the same forumpertaining to issues with transition practicesdescend in the same order with similar frequencies: I (3,798 tokens; 4.2% of corpus), they (591; 0.65%), she (231; 0.26%), he (186; 0.21%), we (162; 0.18%). Hence, the overuse of I in relation to all other pronouns in the corpus data, and the order of significance of the use of each pronoun, is not simply symptomatic of the introductory board's purpose.
A shallow analysis suggests a universal location of the self within transgender discourse: each user is posting in a forum aimed at transgender persons. The use of I within the local discourse of the forum indicates personal belonging and locating the self within the local discourse and, by extension, broader transgender discourse.
However, the significance of I use over any other self-inclusive or selfexclusive pronoun suggests a non-location of others within the local discourse and, by extension, broader transgender discourse; that is not to say that the user represents others as outside of the discourse(s), simply that it does not represent them as within the discourse(s). The non-location of others within the discourse may indicate an initial non-identity with forummembers and, by extension, other transgender persons: users are introducing themselves, thus they are not yet part of the forum-user collective identity, nor are they represented as part of a homogeneous community of transgender identity. The use of we is much more complex than that of I: the referents of we may not remain the same even within directly linked text from the same user. The low frequency of we use implies a discursive interest in individual identity, rather than (transgender) group identity. This implication is strengthened when noting the referents of pronominal we: it is more often used in constructing personal identity-relationships and recounting idiosyncratic personal narratives than in transgender-inclusive group representation and construction. The relative lack of transgender-inclusive group representation and construction implies a lack of transgender community identity, though this may be explained by the users' newness to the forum. Another reading is that there are other, more salient, identity features than transgender status.
The difference between the use of my and our is similar to that of I and we. Used 2,410 and 59 times, respectively, the overuse of my over our is statistically significant in the 99.99th percentile (p < 0.0001, with a log-likelihood score of 2865.57). Again, this seems to indicate that individual identity is more salient than any group identity; hence, there is a further implication that transgender is not a homogeneous collective identity type.
My is used most often in collocation with general practitioner (GP). However, it is also used frequently in collocation with such lexical items as life, name, body and transition. All such discursive collocations are directly linked to gender presentation transition practices, indicating that transgender presentation practices are considered an amalgam of individualistic practices of socio-psychological and sometimes medico-surgically mediated gender presentation. The possessive construction of such discursive strategies indicates the idiosyncrasy of gender presentation practices.
6.1.2 They, she and he They does not carry semantic features of gender and therefore can be used in non-gendered reference to individual persons. However, they can also be used to reference groups of all feminine or all masculine persons and mixed groups of both masculine and feminine persons. As such, they will not be analysed for its non-gender-indexicality, but for its other-categorisation: the use of they in the corpus data and the paradigmatic nature of pronouns (see Wales 1996) suggests a group separate from the self, which also excludes the readerand, by extension, all other forum-users.
As we is used in reference to many self-inclusive groups, they is used similarly. They is often used in reference to medico-legal "gatekeepers" of gender transition: medical professionals and identity documentation service personnel. This is a representation of social distance and of a cognitive model that constructs and represents such groups as opposite, or against, transgender individuals. There is indication of a hegemonic ideology, and shared cognitive model, within transgender discourse that medical/legal professionals and psychiatric diagnoses are obstacles between the individual and congruence in their gender identity and presentation. More simply, she is used in reference to individuals who form part of a personal narrative in the introductory forumposts and is used in reference to many relationships: intimate, medical, familial. He is used in a similar fashion. The gendering of others within the discourse is indicative of a wider social hegemony: binary, cisgender, normativity. There is little assumption that others (outside of personal relationships) identify as transgender or outside of the socially conventionalised gender binary. Table 2 shows the statistical significance of the overuse of they in comparison to its gendered counterparts. Again, there is statistical significance to the 99.99th percentile (p < 0.0001) that they is used more often than gender-indexical pronouns in the corpus data. Similarly, there is statistical significance of the overuse of she over he in the corpus data, with a log-likelihood score of 10.84, at the 99.9th percentile (p < 0.001, where the critical value is 10.83).

You: Between subject and object
The idiosyncratic grammatical and semantic features of you represent "a neutralisation between singular and plural […] [and] of subjective and objective" (Wales 1996: 89). Hence, you was excluded from the above pronoun analyses. Used 905 times, and making up 0.66% of the corpus data, you is most often used as a subjective pronoun and in reference to the paradigmatic transgender individual. The mutually exclusive relationship of pronouns indicates that you is separate and, arguably, opposite from the self (I). Thus, its frequent use in collocation with the paradigmatic transgender individual (e.g. "you have to live full time", "you can get SRS [sexual reassignment surgery]"), and its overuse in comparison with self-inclusive "I am I": Self-constructed transgender identities transgender-group we, is indicative of the non-location of the self in broader transgender discourse relating to the social practices of the "typical" transgender person. This lack of identity with the paradigmatic transgender individual indicates an awareness of the hegemonic ideology of transgender as a social categorisation, but does not imply self-identification with that homogeneous and hegemonic identity. The construction of transgender identity as non-hegemonic and heterogeneous is once again reproduced.

Modality in use
Modality refers to the "'attitudinal' features of language" (Simpson 1993: 47). It can reflect volition, obligation, possibility and certainty, and is directly related to the interpersonal functions of language (Halliday 2002). It is therefore appropriate to explore the use of modality when conducting analyses of social identity construction. In the case of modality use in direct co-occurrence with pronouns, an indication of individual cognitive models of transgender persons may be provided as well as an understanding of why the modal realities are not attributed to other persons/groups (see Webster 2015). Analysing all types of modal construction within the corpus data can provide enough insight for an entire article; here, I will focus solely on constructions including verba sentiendi. Know and feel are verba sentiendi (verbs denoting sentience) indicating epistemic modality, or truth commitment (Simpson 1993). Both co-occur frequently with I in the corpus data (immediately following [in R1]): 147 and 122 occurrences, respectively.
I know constructions in the corpus often appear in discursive narratives pertaining to concrete transgender-specific practicese.g. hormonal and surgical intervention. These co-occurrences tend to have negative semantic prosody (see Louw 1993): (paraphrased) "I know I may always have male traits"; "I know I shouldn't have surgery yet". Such constructions imply an extra-discursive but in direct succession. The construction of "knowing" supposedly concrete concepts in a subjective medico-social "transition" suggests there are naturalised social practices and ideologies that exist within transgender. Such naturalised practices determine schematic expectations for the practice of changing gender presentation (from cultural stereotypes of behaviour pertaining to physiological sex). Such discursive constructions are indicative of an individual cognitive model differing from that which is ideologically naturalised. This variance of individuals' cognitive models from ideological norms of transgender practice may go some way to explain the overuse of I in the corpus data: I separates the user from the transgender community and experience because their cognitive model of what it is to be/do transgender differs from the ideologically naturalised practice of gender transition.
I feel combinations appear frequently with more abstract conceptse.g. emotional states regarding transition, gender identity and health. The colligation of verba sentiendi and personal pronouns constitutes an individual cognitive model whereby there is representation and awareness that others do not share the same cognitive model (see also Webster 2015). The discursive constructions of I feel are representative of a user's personal experience and imply that there are also hegemonic practices of how to feel during gender transition. The use of I in conjunction with feel acknowledges and presupposes that there are others who do not feel the same, and that readers would also know this. The differences between I feel and I know constructions indicate that the expected hegemonic descriptions of feelings during gender transition are not as ideologically naturalised as physical processes of presenting gender transition.
Similar inferences can be drawn from the use of need and want in immediate co-occurrence (R1) with I. Both are realisations of modality; need as deontic (obligation) and want as boulomaic (desire). I need constructions most often relate to socio-/medico-legal stages of transition (e.g. name change, psychiatric diagnosis) and I want constructions used in feelings about "beginning" transition. I need constructions closely relate to I know constructions, insofar as they represent specific practices as being necessary in presenting gender transition and are, therefore, demonstrative of hegemonic ideological practices of transgender. I want constructions are similar to I feel constructions, insofar as they represent specific practices as being ideal in transgender identity.
All of the above constructions suggest that change in gender presentation and gendered practices is an individualistic process and transgender identity an individualistic trait; although there are hegemonic practices, such practices do not necessarily correlate with users' personal experiences. This therefore prevents the users of this forum from sharing a cognitive model and identifying with the hegemonic and paradigmatic transgender community. Hence, the homogeneous transgender collective identity is once again refutedthe hegemony represented, evaluated and defiedand transgender as individualistic identity and practice discursively constructed.

Gender-indexicality, or lack thereof
There are nouns in the English lexicon that carry semantic features of direct gender-indexicality. Similarly, there are their non-gendered counterparts, and those that directly index transgender. All are used within the discourse for varying purposes and can each be considered representative of both hegemonic and individual ideological practices within transgender. Table 3 shows the nouns' frequencies in the corpus data, in descending order.
The manifold alternatives of gender-indexicality distort the apparent significance of the data. By calculating the statistical significance of the use of all non-gender-indexical nouns with their gendered counterparts using loglikelihood calculations, it becomes clear that the overuse of gender-indexical nouns is statistically significant to the 95th percentile, with a log-likelihood of 5.45 (p < 0.05, where the critical value is 3.84). Following this indication of significance, I will analyse the use of gender-indexical and transgenderindexical nouns, respectively, in the subsequent sections.

Gendering the self
Woman and girl are both used when representing the self. They are also used with discursive similarity, both frequently appearing in linguistic manifestations, such as "knowing [they] [are/was] a [woman/girl]". However, there are some differences in the discursive constructions of the self-as-woman and selfas-girl, though both such constructions illuminate similar hegemonic ideologies of transgender as non-normative and inferior to cisgender practice.
Woman is used in constructions denoting "[living/dressing/acting] as a woman". In such constructions, there is social comparison between "transgender women" and "cisgender women". As such, transgender-feminine persons are not locating themselves in the social categorisation of woman. Girl is most often Table 3: Frequency of (non-)gender-indexing lemmata in the corpus data, including subsidiary lexeme frequencies.

Item/lemmata
Total frequency Lexemes (frequency) used in constructing gender-as-fact: "I am a [big/lonely/sensible] girl" (emphasis mine). The use of girl, rather than woman, in asserting the proposition of gender as fact is indicative of the transgender-specific hegemonic ideology that transgender-feminine persons are not women, but are girls. Hence, there is some indication of a shared cognitive model that transgender-feminine individuals have less developed gender identities and presentations: despite reaching adulthood, incongruence between gender identity and physiological sex at birth leads to having to grow up into a woman, from a girl (i.e. re-develop a gendered identity as one does through childhood, adolescence and early adulthood). Additionally, the qualifying adjectives co-occurring with girl reproduce the hegemonic practice that "women" should evaluate themselves in terms of others' perceptionhence, this is somewhat indicative of a representation of the self as located within a typical cisgender discourse of masculine domination over femaleness/femininity. Users construct a duality of identity, as located both within and outside of cisgender discourse. A possible reading of this data is that transgender-feminine persons are more readily accepting of themselves as female/feminine, but that there is a structure consisting of gradable femaleness/femininityand cisgender women are atop the hierarchy of such a social structure. Users' dual identity is also represented and constructed via the use of boy-mode and girl-mode. The two constructions are intended to represent gender presentation as being either masculine-presenting or feminine-presenting. Though male-and andro-(androgynous) mode are used, there is no occurrence of woman-or man-mode. Again, perhaps this is denotive of an ideology that the gender identity and presentation of gender is not fully developed within transgender individuals, especially where gender presentation transition is a new practice for the individual (or is still yet to begin).

Gendering others
The gender-indexical noun used may be determined by the referent's status as a forum-user, or not. Transgender-feminine forum-users are represented as girls and non-users (e.g. the paradigmatic transgender woman) are primarily represented as women. Contextually, the diminutive girls is also used in friendship circles to denote closeness among cisgender women in a friendship group. Women, however, is not typically used in the same way in such groups. The implication is that the forum-communication "friendship group" is a community of practice where like identities participate in conversation, whereas transgender status is not such a community of practice. Although all users consider themselves transgender, the relationship is constituted in the shared use of a particular forum. The combination of this implication and the lack of self-inclusion in hegemonic practices of transgender suggests that identityrelationships between forum-users are rooted in the shared social practice of the local forum-communication discourse, rather than in a shared gender identity as homogeneously transgender.
Women is also used when discursively representing cisgender women However, this usage is infrequent in the data. This may be, in part, due to the forum being specifically aimed towards transgender users. However, it may also be representative of the hegemonic separation of cis-and transgender identities. By virtue of the paradigmatic cisgender woman being represented infrequently, the same is represented as lacking identity with transgender and, by extension, transgender women. Hence, this may be representative of a shared cognitive model between active forum-users: transgender-feminine persons denying cisgender women access to the discourse, a reversal of the naturalised hegemonic practice of transgender-feminine persons being denied access to "women-only spaces". Although perhaps unconscious, this discursive socio-cognitive representation of the social categorisation of woman may constitute a defiant stance toward the hegemonic social domination of cisgender over transgender. It may, however, also have the effect of adhering to the naturalised practice of dividing cisgender and transgender persons.

Transgendering social actors
Transsexual is often used in conjunction with the indefinite article a -38% of uses of transsexual co-occur (in R1) with a. The term is also used in medicosurgical narratives (e.g. hormone regimens). The statistically significant overuse at the 99.9th percentile (at p < 0.001, with a log-likelihood score of 14.81) of transgender over transsexual indicates a transgender-specific ideology that the paradigmatic transsexual of yester-thought is unwelcome in current transgender discourse. Similarly, the use of transsexual in medico-surgical narratives is reflective of the outdated use of Transsexualism as a psychosexual disorder in medical practice (see World Health Organisation 2015, ICD-10). The use of transsexualism in medical practice is indicative of an institutional power demarcation and may explain the perceived disconnect between transgender persons and medico-surgical practitioners (see Section 6.2).
Transgender is used much differently than transsexual: in reference to the self and specific others. By virtue of its use in such constructions, there is the implication of massive social difference between self-and other-categorisation as transgender and as transsexual. Transgender is used more often because it is socially more acceptable and has not endured the same semantic derogation as transsexual. I am transgender constructions indicate that transgender is individualistic: the constructions do not represent belonging to a group (e.g. "I am a transgender"unlike, "I am a girl"), but represent a state of being.
However, there are also constructions including the use of transgendered. The passivation (see van Leeuwen 1995) of transgendered suggests that transgender is a process without an agenthence, the transgendered person has no agency in their transgender status (i.e. it is not a choice). This may be representative of a cognitive model in which transgender identities are not chosen. By extension, it may represent a cognitive model that shares medical associations' ideology that transgender status is an affliction, due to its hegemonic nonnormativity and the "necessity" of psychiatric diagnosis.

Conclusions
Primarily, there is evidence that transgender is not a collective homogeneous identity, though there are groups comprising transgender members which may hold identity relationships based on concrete social practices of communication (e.g. participation in communication on the same forum). Transgender identities are, rather, an individualised identity and practice.
However, the hegemonic perception of gender identities in society has fostered an environment in which social comparison is necessary in order to demonstrate identity. This is achieved by discursively constructing the paradigmatic transgender person, cisgender woman, and respective hegemonic sociopsychological practices. Such practices are evaluated as correct (hence, they are hegemonic), but are not experienced similarly. Thus, the concept of an ideal transgender experience further marginalises transgender persons.
The use of woman in constructing and representing the self serves as social comparison between transfeminine people and cisgender women whereas the use of girl represents individual identity. Uses of diminutive gender-indexical nouns in reference to the self and other forum-users indicates a cognitive model representation of a lack of maturity in gendered identity and practices and may be reflective of a transgender-specific hegemonic practice of transgender persons having to redevelop gender identity.
There are fewer transgender-only self-inclusive groups and more groups locating the self in other discourses: e.g. Americans. Infrequent use of we in constructing transgender-inclusive groups and more inclusion of the self within other groups suggest that there are more salient aspects of identity than transgender status even in contexts making transgender-identity construction probable (see Stryker 1980). Hence, when transgender is being discussed, and transgender identity is schematically expected to be salient, no homogeneous transgender identity is, in fact, constructed.
Similarly, the term transgender in reference to the self is used much less frequently than girl or woman. Thus, transgender is not salient in the identity of transgender-feminine persons, but identification with femaleness/femininity is. This ideological identity construction is indicative of cis-normative hegemony and may, ultimately, be a destructive identity behaviour (insofar as it is, largely, an unfulfillable identity). However, it is undoubtedly representative of the lack of salience of a homogeneous transgender identity.