While popular claims about the negative effects of pornography (such as inducing sexual violence and emotional detachment) have been entirely discredited, empirical explorations of users' engagement with porn have shown its potential influence on beliefs, attitudes and behaviours (Billard, 2019): audiences use pornographic materials to inform themselves about sexual anatomy, attraction and pleasure, especially when other sources of information are lacking or perceived as judgemental (Albury, 2014:174). In the social construction of sexual desires and identities, porn provides one of the key spaces for articulating peripheral forms that fit poorly within the mainstream (Escoffier, 2011: 279; Rubin, 1992).

Porn has been especially important in constructing public images of trans and other gender nonconforming (GNC) identities. While trans representations in mainstream media are scarce and predominantly derisive (Richardson 2010:131), GNC characters and bodies have been conspicuous in pornography both visual and textual (Escoffier 2011:271-2); for many audiences it represents their first encounter with GNC, and certainly the first that celebrates it. At the same time, porn representations can be problematically loaded with normative assumptions (such as that all GNC people experience dysphoria or aim for surgical/hormonal transition; Jones, 2019; Lovelock, 2017). Steinbock (2017:29–30) argues that trans pornographic films reproduce misogynistic tropes of objectification, hypersexualisation and femininity as submissiveness. However, these claims have not been explored in written pornography.

In this paper, we build on prior work on the linguistic construction of desirability in porn (Baker, 2005; Bolton, 1995; Koller, 2015; Marko, 2008; Morrish & Sauntson, 2007; Motschenbacher, 2010) by exploring patterns of labelling and description of GNC characters. Using a 10-million-word corpus of erotic stories, we identify the range of terms denoting GNC identities and performances and explore their discourse prosodies and textual functions, seeking to identify distinctive ways in which GNC can be construed and appraised. Results suggest that characteristic lexical signals reflect different subcultural manifestations of gender variance. This diversity should be taken into account in order to avoid eliding intersections between gender and other cultural categories.