Abstract
This article investigates moments in social interaction where tacit processes of gender attribution become visible because they are temporarily disrupted and exposed through misgendering. Our data consist of publicly available audio and video-recorded cases of misgendering, mostly from UK and US contexts. Practices of misgendering embody assumptions that map people’s current gender onto their self-presentations and gender histories. Organisational features of social interaction facilitate the reproduction of these assumptions as taken-for-granted criteria for gender attribution. In the current climate of ‘gender panics’, the rise of a norm whereby people’s self-defined gender should be respected clashes against enduring assumptions that uphold a gender order grounded in cisgenderism. The exposure of gender assumptions in moments of misgendering presents a potential for social change, but this potential is also limited by practices that reproduce (rather than challenge) the dominant gender order.
Introduction
Sociology has long challenged essentialist perspectives that treat gender as a ‘natural’ individual trait and has instead respecified it as an assemblage of socially maintained arrangements. This reconceptualisation extends beyond an analysis of systems that privilege members of distinctive gender categories and disadvantage others, to investigate how gender categories are socially formed (West and Zimmerman, 1987; Westbrook and Schilt, 2014). Gender is a pervasive social identity (Chatillon et al., 2018) and a powerful system that constrains possibilities for action at the individual, interactional and macro levels of social organisation (Risman, 2018). The interactional level is a key domain of social organisation in which gender identities are constituted, displayed, validated and rejected (Serano, 2009). We hope to advance understandings of how normative assumptions about gender are manifested and reproduced in face-to-face social interactions.
Conceptual Framework
Drawing on integrative sociological theory (Risman, 2018), we view gender as a system of structures (patterns of organisation and practice) that are reproduced and sustained through material and cultural means at the individual, interactional and macro levels of social organisation. This assemblage can be conceptualised as a ‘gender order’ (Connell, 1987; Matthews, 1984) composed of practical arrangements and ideological beliefs involved in producing inequity.
We view social interaction as a domain where the individual and macro levels of social organisation interlock through people’s situated actions and participation in joint activities (Heritage, 2008; Risman, 2018; Schegloff, 2006). In face-to-face interactions, people display individual membership in gender categories through practices that simultaneously embody and (re)produce macro-level organisational arrangements and ideologies. Social interaction is an arena wherein individual gender identities and macro-level gender normativities are enacted and validated in ways that are mutually reinforcing – but also subject to contradictions and contestations. To examine how gender is constituted in interaction, we draw on several sociological theoretical perspectives including ‘doing gender’ theory, ethnomethodology and conversation analysis.
West and Zimmerman’s (1987) doing gender theory reconceptualised gender from an individual characteristic (something that people are) to an activity – something that people do in social interaction. People’s doings of gender are situated and responsive to local contingencies, making gender ‘an emergent feature of social situations’ (West and Zimmerman, 1987: 126). Rather than a factor influencing people’s conduct in a top–down fashion, gender is constituted in, and through, people’s practical activities. West and Zimmerman (1987) defined sex as an attribution based on socially signified biological criteria (usually carried out at birth through visual inspection of genitals; see Kessler and McKenna, 1978). In nonsexual social interactions, people do not have information about people’s genital configurations and categorise others based on ‘regulated external insignia of sex – such as deportment, dress, and bearing’ (West and Zimmerman, 2009: 113). West and Zimmerman (1987, 2009) referred to this as sex categorisation. In this article, and in line with other authors (e.g. Westbrook and Schilt, 2014), we refer to this as gender categorisation.
Ethnomethodology, as a branch of sociology, has investigated gender as a situated accomplishment since Garfinkel’s (1967) case study of Agnes. Kessler and McKenna (1978) built on Garfinkel’s work to examine how people attribute gender in interaction (how they place others in gender categories; see also Westbrook and Schilt, 2014). Gender attribution relies upon assumptions that uphold the gender order. Here, ethnomethodology converges with other sociological approaches in identifying interconnected assumptions including gender binarism (expecting that individual gender identities fit in the woman/man dichotomy), androcentrism (treating male identities as normative), heteronormativity (establishing heterosexuality as normative) and cisgenderism (privileging gender identities that align with ‘sex’ categories assigned at birth; Ansara and Hegarty, 2013, 2014; Garfinkel, 1967; Hornscheidt, 2015; Ingram, 2019). These assumptions are reinforced by the reification of gender as a trait that is expected to be constant throughout an individual’s lifetime (Garfinkel 1967). Drawing on the interactional sociological perspectives of doing gender and ethnomethodology, we view the (re)production of these assumptions – as well as the organisations of practice that make up a gender order as a whole – as situated achievements. Patterns of interactional conduct create stable configurations over time, and thus, uphold the gender order. However, interaction is also a site of tensions and transformation.
Earlier research conceptualised gender attribution partly in decontextualised ways, without directly observing the interactions in which gender assumptions are mobilised to implement social actions (e.g. Kessler and McKenna, 1978). In this article, we address this by observing interactions in which the inferential work entailed in attributing gender becomes visible. Drawing on conversation analysis as sociological theory (Heritage, 2008), we understand social interaction as an institution organised around interconnected domains of patterned practice including (but not exclusive to) the organisation of participation through turn taking, the enabling of joint activities through the sequencing of actions and the resolution of problems of mutual intelligibility through practices of repair (Hoey and Kendrick, 2017). One of our theoretical contributions is to specify how organisational features of social interaction provide procedural infrastructures for the reproduction of the assumptions that uphold the gender order in society.
Also central to the organisation of interaction is people’s use of membership categorisation devices (Sacks, 1992) to assign themselves and others to social categories. When people place others in a gender category, they concurrently (re)produce conceptions of what counts as a valid gender category and of the qualities and conduct that make an individual a legitimate member of that category (Sacks, 1992). One difficulty with analysing gender categorisation in interaction is that it is largely tacit and only available by inference. People commonly attribute gender without asking (failing to do so can result in having one’s competence challenged). In English interactions, we can infer that people have determined others’ gender whenever they use gendered language to refer to them (e.g. with gendered pronouns such as she and he; Klein, 2011), address them (e.g. madam or sir) or describe them (e.g. woman or man; Stokoe, 2011). Because these linguistic practices are overwhelmingly used without issue, the gender attribution work that they imply largely goes unnoticed. It ordinarily operates under the conversational ‘surface’ (borrowing Jefferson’s (1996) metaphor). For this reason, moments of disruption – when gender attributions fail or are contested – have attracted scholarly attention for their potential to bring some of the assumptions entailed in attributing gender closer to the conversational surface.
In an exception to the absence of observational research on gender attribution, Speer (2005: 71) examined group discussions in which a researcher invited participants to comment on images of people, some of which showed ‘men and women engaged in [. . .] non-traditionally gendered activities (men’s ballet and women’s rugby, for example)’. The participants often exhibited difficulties in making definitive gender attributions, and they explained this by pointing out non-normative gender cues in the visual prompts. Rather than treating the prompts as invalidating gender assumptions, the participants reinstated their validity, ultimately upholding a normative, objective and dichotomous gender order. Therefore, moments in which gender attributions become disrupted are of analytic and theoretical interest. Speer (2005) investigated a research environment that had been engineered to hinder participants’ gender attributions (not unlike Garfinkel’s (1967) breaching experiments). We thus do not know how gender attribution work and its underlying assumptions manifest in naturally occurring interactions. We address this by analysing cases where people’s gender attributions fail or are contested in cases of misgendering.
Misgendering
Misgendering refers to ‘the use of gendered language that does not match how people identify themselves’ (Ansara and Hegarty, 2014: 260), such as when a woman is addressed, referred to or described as a man (Ansara and Hegarty, 2013). Previous research investigated misgendering in academic literature (Ansara and Hegarty, 2013), academic guidance (Ansara and Hegarty, 2014), news media (Capuzza, 2015; Gupta, 2019) and social media (Ingram, 2019). That research showed that practices of misgendering expose people’s reliance on gender assumptions, including expectations that people’s gender aligns with anatomic configurations culturally signified as ‘female’ and ‘male’; and that people’s current gender aligns with their presumed gender history and assigned ‘sex’ (Ansara and Hegarty, 2013, 2014; Ingram, 2019). These assumptions can be seen as forms of cisgenderism, which is understood as ‘the ideology that delegitimises people’s own designations of their genders and bodies’ (Ansara and Hegarty, 2014: 260). Their use marginalises people whose self-designations 1 do not align with normative expectations around ‘gender assignment histories and anatomical configurations’ (Ansara and Hegarty, 2014: 264) including people who may self-identify, or be identified by others, as transgender, intersex, nonbinary and genderqueer (among others; Ansara and Hegarty, 2013; Kapusta, 2016; Serano, 2009). Cisgenderism thus privileges (and naturalises) the experiences of people whose gender self-designation aligns with the ‘sex’ category they were assigned at birth within a binary gender order (and who may self-define or be identified by others as cisgender).
In this article, we study misgendering in face-to-face interactions. When misgendering occurs, processes of gender attribution momentarily become visible at the interactional surface because they fail or are contested. This draws attention to assumptions that tacitly drive people’s determinations of others’ genders. We do not limit ourselves to identifying the assumptions that are (re)produced through misgendering, but also investigate how those assumptions are manifested in interaction. Westbrook and Schilt (2014) showed that criteria of gender determination become the focus of explicit debate in media coverage of legal and political debates relevant to the recognition or rejection of transgender people’s identities. They speculated that gender assumptions are comparatively elusive in face-to-face interactions (see Kessler and McKenna, 1978; West and Zimmerman, 1987). The level of explicitness or implicitness with which gender assumptions manifest is an important empirical problem because it determines the extent to which assumptions become available for possible contestation and revision. We address this by identifying situational contingencies and organisational features of interaction that contribute to making the operation of gender assumptions tacit or visible at the interactional surface.
Materials and Methods
We collected instances of misgendering from publicly available recordings of face-to-face interactions. Limiting our search to interactions in spoken English, we searched open-access platforms with terms including misgendering, transgender, nonbinary, gender, as well as some known personalities (e.g. Caitlyn Jenner). We augmented this collection with instances we came across opportunistically. Admittedly, this convenience sampling returned instances from different types of media formats and interactional activities, including broadcast talk shows, news interviews, podcast conversations and interactions recorded by private users in public spaces (e.g. streets). However, this strategy was consistent with our aim, which was not to investigate specific media or interactional events but to identify cases of misgendering in naturally occurring face-to-face interactions. Data-mining tools are not available for systematic searches of interactional and spoken content (as opposed to texts). Additionally, instances of misgendering are not always labelled as such on platforms and occur through highly contextualised linguistic practices, which are difficult to identify through word searches.
We identified 70 instances where people misgender a co-present person or an absent but already known person. Our unit of analysis was the single instance of misgendering rather than the media event or interaction. Restricting our search to English resulted in a collection of instances from western settings – the large majority being from the UK and the USA. We interpret the transferability of our findings accordingly and do not expect them to capture features of gender attribution across cultures or languages. Our aim is also not to investigate differences between the UK and US contexts and report on findings that apply to both sets of interactions. In the time period where the interactions in our sample were recorded (2013–2022), both national contexts have been characterised by an increased visibility of transgender people (Capuzza, 2015). This has been accompanied by ‘moral panics’ (Pearce et al., 2020) in which transgender people are portrayed as threats, especially when discussing access to gender segregated spaces (Westbrook and Schilt, 2014). Misgendering is significant in this context because it is one of several ways in which transgender people’s lived gender is invalidated (Serano, 2009). Additionally, debates about the validity of transgender identities embody a tension between what Westbrook and Schilt (2014) define as ‘identity-based’ criteria for gender validation (ratifying people’s self-defined gender) and ‘biology-based’ criteria (upholding a gender order in which gender maps onto ‘sex’ assigned at birth). Our analyses evidence how these tensions materialise in instances of misgendering in interaction.
Ethical Considerations
Our research process was informed by three ethical matters regarding research on publicly available video data (Association of Internet Researchers, 2012; Legewie and Nassauer, 2018; see also Edmonds and Pino, 2023). The first concern is informed consent. When internet users make recordings publicly available, this cannot be taken as implicit permission to use them for research (Legewie and Nassauer, 2018; Stommel and de Rijk, 2021; Vincent, 2018). Such recordings do however present unique opportunities (Legewie and Nassauer, 2018) such as creating a large collection of instances of misgendering, which would be exceedingly challenging through primary data collection. To minimise the impact of concerns related to lack of informed consent, we only collected instances of misgendering from broadcasts and recordings made in public spaces (e.g. streets) where one can assume that the people featured in them did not have heightened expectations of privacy (Legewie and Nassauer, 2018; Stommel and de Rijk, 2021). We only collected data from open-access platforms, where we can at least assume that those who uploaded recordings had expectations of heightened visibility (Legewie and Nassauer, 2018).
The second concern is the possible harm that our research might cause to the people depicted in the recordings (beyond the risks associated with the public availability of the recordings; Edmonds and Pino, 2023). One risk is recognition (Stommel and de Rijk, 2021). We have minimised such potential harm by not providing the hyperlink to a recording involving a private citizen (extract 5). However, it was not necessary to use a pseudonym because the participant’s name was not featured in that recording. All other cases in this article are from broadcasts where the use of pseudonyms seemed unnecessary. Recognition is further associated with the risk of outing for people whose gender self-designation is independent from assigned ‘sex’ (Ansara and Hegarty, 2014). To minimise this, we only included recordings where people self-identify as transgender or nonbinary in the interactions or had previously done so publicly.
The third concern is the objectification of people whose gender is independent from assigned ‘sex’. Associated with this is a risk of unintentionally reproducing stereotypical and stigmatising views, partly through outdated or generalising language conflating diverse experiences (Vincent, 2018). To manage these risks, we included recordings where (for example) a transgender person was part of a broader scene, but we excluded recordings where the filming selectively focused on a transgender person and involved harassing them. We worked on our language by self-educating and asking colleagues for feedback. We reflected on our positionality as cisgender men working within an anti-cisgenderist framework, and how this positionality inevitably informs how we make sense of the phenomena we investigate (Knott-Fayle et al., 2022). In our writing, we respect people’s gender self-designations and pronouns when made available in the recordings and avoid assumptions when this is not available (Ansara and Hegarty, 2014). We avoided generalising terms (including transgender) to ascribe people to categories across our recordings. We did so to avoid conflating life experiences that are diverse, to avoid rigid demarcations between transgender and cisgender people (Ansara and Hegarty, 2014) and to avoid reducing people’s existence to their transgender identity or status (Vincent, 2018). We nevertheless used these terms for people who self-describe in these ways in the recordings.
Analytic Method
Our method is conversation analysis (CA; Sidnell and Stivers, 2013). In line with CA, we sought data-internal evidence that the participants treat their own or others’ conduct as misgendering someone. In our sample, this happens when misgendering is corrected, complained about or explained. We used established methods in CA to transcribe verbal and vocal conduct (Jefferson, 2004) and embodied conduct (Mondada, 2018) – see Online Appendix.
Using well-established analytic methods in CA, we examined how misgendering is done in each case through ways in which participants design their actions, and ways in which they sequentially and temporally place them (Hoey and Kendrick, 2017). We adopted the constitutive view of gender shared by the ‘doing gender’, ethnomethodology and CA perspectives. We thus investigated how practices of misgendering constitute in situ (rather than merely reflect) macro-social gender arrangements and ideologies. Drawing on advances at the intersection of feminist and CA research (Kitzinger, 2000, 2005; Stokoe and Smithson, 2001), we examined how gender assumptions are often conveyed tacitly in interaction but also examined instances where they become the focus of more explicit articulation. Drawing on well-established analytic and theoretical findings in CA, we identified recurrent situational contingencies and organisational features of interaction that contribute to making gender assumptions more or less explicit. Data extracts for this article were chosen to exemplify patterns, progressing from less to more overt ways in which gender assumptions are manifested.
Results
Practices of misgendering reproduce two overarching assumptions; namely, that people’s gender maps onto their self-presentation and their presumed gender history. These assumptions are largely conveyed tacitly. However, some contingencies recurrently contribute to making them the object of explicit articulation. We consider several such contingencies and how, in each of them, organisational features of interaction contribute to making gender assumptions more or less explicit.
Assumption 1: Self-Presentation
One way that participants manage misgendering is through corrections (Jefferson, 1987), which in our data are mostly implemented through repair (interactional practices designed to address problems of speaking, hearing and understanding; Schegloff et al., 1977). With these, participants treat a gendered reference, address or description as incorrect because it misrepresents someone’s gender, and they correct it. Corrections can be implemented by the participant responsible for the misgendering (Self) or by another participant (Other). Corrections implicitly draw attention to assumptions that may have tacitly driven a subsequently revised gender attribution (Ekberg and Ekberg, 2017; Land and Kitzinger, 2005). In most cases, such assumptions are only available by inference. However, there are circumstances in which participants implement additional remedial work, beyond correction, by providing explanations to address concerns with competence, responsibility and ‘face’ (Brown and Levinson, 1987) raised by the misgendering. Those explanations publicly display participants’ situated understandings of what drove the misgendering and usually work to reaffirm the validity of a gender order grounded in cisgenderist assumptions. One of these is the assumption that a person’s gender maps onto their self-presentation (Kessler and McKenna, 1978), which includes a person’s visual and aural appearance (how they look and how they sound), and linguistic identifiers, such as first names and pronouns (Ansara and Hegarty, 2013; Gupta, 2019). The operation of this assumption is mostly evidenced in the misgendering of physically co-present participants.
Extract 1 exemplifies a common occurrence where a participant misgenders someone, self-corrects without being prompted and moves on without explaining the misgendering. The extract is from an episode of ‘BBC Big Questions’ featuring a debate on ‘gender neutral language’. Before extract 1, Emma has self-identified as nonbinary and stated that they use the personal pronouns ‘they/them’. Belinda (Bel) has declared that she refuses to refer to Emma as ‘they’ on the basis that, to her, Emma is a woman – thus, denying the reality of nonbinary identities. As extract 1 begins, a member of the audience (Aud) challenges Belinda’s position by describing it as a ‘belief system’ (line 1) that Belinda imposes on others. The audience member then proposes that Emma also has a ‘belief system’ (alluding to their gender self-designation), which should be respected (lines 3–4, 7–8 and 10–11). When referring to Emma, the audience member initially refers to them with ‘she’ (line 3). The audience member treats this as misgendering Emma by replacing it with ‘he’, to finally settle on ‘they’ (lines 3–4). In the following extracts, misgendering practices and subsequent corrections are highlighted in grey.
(1) DE25 ‘She, he, they’
02
03 you can’t d
04
05 Others: [Heh heh[heh
06 Bel: [Sh:↑e [there we ↑go?]
07 Aud: [They- (b’no) ]that-
08 [that’s ( ) l
09 Bel: [(Are you listen )]
10 Aud: =
11 Aud: Y
The audience member’s self-correction in lines 3–4 (in technical CA terms, done through self-initiated self-repair in the same turn; Schegloff et al., 1977) embodies alignment with a norm that prioritises people’s self-designations of their gender over other considerations (in Westbrook and Schilt’s (2014) terms, she adopts identity-based criteria for gender determination). The audience member voices her commitment to this approach at lines 3, 4 and 10. The self-correction implicitly draws attention to assumptions that may have tacitly grounded the gender attribution implied by the initial use of ‘she’. Before extract 1, the participants’ discussion evidenced that they treat Emma’s visual appearance as feminine (Emma themself commented that they do not attempt to ‘masculinise’ their own appearance; see Edmonds and Pino, 2023). Observers (inside and outside the scene) can thus infer that a reliance on Emma’s visual presentation grounded the misgendering. The misgendering tacitly reproduces the cisgenderist assumption that gender maps onto self-presentation. This reliance on self-presentation operates in the context of displayed commitment to gender binarism – embodied in the initial cycling through the ‘she’ and ‘he’ pronouns (line 3). The subsequent speech perturbations and display of effort (through eye rolling and hyper-articulation of ‘they’, line 4) treat linguistic validation of a nonbinary identity as unfamiliar territory.
The assumptions grounding the misgendering are conveyed tacitly, which arguably contributes to reinforcing their taken-for-grantedness and normativity (Kitzinger, 2005). After self-correcting, the audience member moves on without explaining the misgendering (line 4). When speakers self-correct in these ways, recipients typically treat this as ‘good enough’ and do not invite explanations (Belinda’s comment in line 6 negatively sanctions the audience member’s validation of Emma’s gender self-designation, not the misgendering).
Organisational features of interaction contribute to keeping the self-presentational assumption largely tacit. People use practices of repair to remedy breaks in intersubjectivity (mutual understanding; Schegloff, 2006) including problems of reference, such as the ‘mispronouning’ (Ansara and Hegarty, 2014) in extract 1. Use of repair is nevertheless constrained by a norm of progressivity, which favours the forward movement of interaction through the production of a next item (e.g. a sound in a word or word in a turn; Schegloff, 2006). Because repair operations suspend progressivity, they are usually minimised. For example, there is evidence that participants deploy forms of repair that promote the most economical solution to a break in intersubjectivity (through practices of least collaborative effort; Dingemanse et al., 2015). In cases like extract 1, additional remedial work (e.g. explanations) would further compromise progressivity and might be avoided on this basis.
Repair operations also raise considerations about who was responsible for the problem in need of remediation (Robinson, 2006), with a major axis for attributions of responsibility being Self and Other. The occurrence of misgendering generally raises the possibility of a lapse in competence by Self (in accomplishing a correct gender attribution) or Other (in displaying the ‘right’ gender cues in a culture valuing alignment between self-presentation and gender). In extract 1, additional remedial work, especially an explanation that grounded the misgendering in Emma’s self-presentation, could amount to claiming that they are not displaying the ‘right’ gender cues. In terms of Brown and Levinson’s (1987) politeness theory, this would amount to threatening Emma’s positive face, and it might be avoided on this basis. Avoidance of face violations can be understood as an instantiation of recipient design, which refers to the ways in which ‘talk by a party [. . .] is constructed or designed in ways which display an orientation and sensitivity to the particular other(s) who are the coparticipants’ (Sacks et al., 1974: 727).
Participants sometimes engage in additional remedial work, extending beyond correction to include explanations for the misgendering. These explanations manage concerns with competence, responsibility and ‘face’. One circumstance that occasions explanations is correction by Other, as exemplified in extract 2. It is from the British BBC talk show ‘Question Time’. Host Fiona Bruce (Bru) invites a member of the audience (Aud), who has raised her hand, to ask the panellists a question. Bruce selects the audience member through a gendered description (line 1), which the audience member corrects in the next turn (line 3; in CA terms, this is done through other-initiated other-repair). The correction exposes Bruce’s description as misgendering the audience member. Jefferson (1987: 90) observed that exposed corrections can occasion apologies and ‘accountings – which address lapses in competence and/or conduct’. Extract 2 fits the pattern with Bruce apologising (‘I’m so sorry’, line 4, and ‘So forgive me’, line 6) 2 and providing an explanation (lines 4–6).
(2) MP31 ‘Man at the back’
02 (0.9) ((Aud starts to take face covering off))
03 Aud: Um (.) I’m a w
04 Bru: [I’m s
05 your m
06 sh
07 Aud: [It was just me, fair enough.
08 ((Aud goes on to ask question))
The audience member’s correction (line 3) implicitly draws attention to the assumptions that may have grounded the misgendering. Observers can infer that Bruce has tacitly relied upon the audience member’s appearance in the context of physical co-presence in a space characterised by mutual visual access between unacquainted participants (a kind of relationship indexed by the non-recognitional reference term ‘man’ in line 1; Schegloff, 1996). Like extract 1, this misgendering upholds the cisgenderist assumption that gender maps onto self-presentational cues within an overarching adherence to gender binarism. What distinguishes extract 2 is that the assumption is invoked in Bruce’s explanation (lines 4–6). By attributing the misgendering to contingencies that reduced visual access to the audience member’s face (including her face covering 3 and Bruce’s short-sightedness), this explanation alludes to self-presentation as a normative basis for gender attribution.
Bruce’s explanation leaves much unsaid; for example, it does not specify the aspects of the audience member’s appearance that might have led to seeing her as a man in the context of limited visual access. This cannot be explained by reference to a minimisation of remedial work because Bruce does produce an explanation, which delays interactional progressivity beyond the correction. However, it can be explained by reference to the concerns with recipient design we identified in extract 1 (matters of competence, responsibility and ‘face’). By invoking factors that hindered visual access, Bruce implies that she is ordinarily able to determine someone’s gender by their appearance, thus guarding against attributions of incompetence (see Speer, 2005). While Bruce aligns with the norm that prioritises others’ self-designation of their own gender (by accepting the correction), her explanation upholds the reality of a dichotomous gender order in which people’s gender is expected to map onto their self-presentation. The incident is treated as an exception that ‘prove[s] the rule’ (Heritage, 1988: 140). Consistent with the other cases in our data, the explanation does not point out aspects of the other’s self-presentation that might be treated as non-conforming relative to dominant standards – thus, preventing the implication that the audience member might not be displaying the ‘right’ gender cues. Although observers have access to aspects of the audience member’s appearance that might have led to see her as a man in the context of reduced visual access (such as her short hair), those aspects remain unmentioned. This embodies a concern with avoiding threats to the other’s positive face (Brown and Levinson, 1987). The self-presentational assumption is thus invoked, but it is also treated as a largely tacit, unnoticeable aspect of the gender attribution process, whose detailed workings cannot, or should not, be fully articulated.
Another circumstance in which participants provide explanations is when they frame misgendering as intentional rather than an error (Edmonds and Pino, 2023). Several of our cases are from televised debates that promote polarisation for entertainment purposes. Some debates focusing on ‘gender issues’ feature guests who reject gender self-designations that do not align with ‘sex’ assigned at birth. These participants recurrently address, describe and refer to others in ways that invalidate their self-designations, and they explain this by reference to essentialist views of gender (Pearce et al., 2020). As an example, extract 3 is from earlier in the same episode of ‘BBC Big Questions’ as extract 1. As extract 3 begins, Belinda declares that she refuses to refer to Emma as ‘they’ because she considers Emma a woman (lines 1–2). She proceeds to claim that she considers recognition of nonbinary identities an ideological position, which she opposes (lines 3–4).
(3) DE24 ‘Clearly a woman’
4
02 here=who’s
03 that would be <im↑po:sing> an ideological
04 .h[h p o l
05 Hos: [How is she clearly a w
06 Bel: =↑i:s. If-
07 she is a ↑woman.
Belinda upholds (rather than corrects) her description of Emma as a woman (line 1). She explains this choice with the parenthetical remark ‘who’s clearly a woman’ (line 2). The recurrence of these explanations in our sample suggests that these actions are treated as accountable because they violate the norm that prioritises people’s self-designations of their own gender. Belinda’s explanation invokes Emma’s visual appearance (especially through ‘clearly’ and pointing towards Emma with extended hands, appearing to invite viewers to judge Emma’s appearance for themselves).
Belinda does not specify which aspects of Emma’s appearance ground her treatment of them as a woman (lines 1–2). To explain this, we note that Belinda’s action is adversarial and as such, it embodies a departure from otherwise normative practices in social interaction that ‘suppress discordant actions while promoting solidary actions’ (Clayman, 2002: 229). 5 There is evidence that such departures expose a participant’s conduct to scrutiny – others may inspect it for flaws that can be leveraged to undermine their position (Clift and Pino, 2020). In this polarised context, any grounds for considering Emma a woman based on appearance could be treated as defeasible and make Belinda’s position vulnerable to objections. Notably, Belinda resists providing such grounds when the host invites her to do so (line 5). Belinda rather invokes self-presentation as a self-evident, and thus unnecessary to explicate, basis for a gender attribution (as something that ‘everybody . . . can see’; lines 6–7) – a cisgenderist assumption. Reflexively, this constitutes the self-presentational assumption as a tacit and taken-for-granted basis of gender attribution, whose detailed workings cannot, or should not, be articulated.
Our analyses so far return a complex picture in which: (a) features in the organisation of interaction converge in making the self-presentational assumption largely tacit; but (b) recurrent contingencies also contribute to evidence its operation at the interactional surface.
Assumption 2: Gender History
Other instances of misgendering imply persistent reliance on people’s former gender designations (particularly, their ‘assignment histories’; Ansara and Hegarty, 2014: 264). In our data, this assumption is mostly manifested in the misgendering of already known, but absent, people. Several of these cases entail discussion about transgender and nonbinary people who are public figures. These individuals have changed their gender self-designations and associated gendered linguistic signifiers including pronouns, first names and titles, to better reflect their gender identity (Gupta, 2019). However, their former gender designations are also publicly known. Use of those former designations tacitly upholds a cisgenderist assumption that treats gender as a stable, enduring reality.
Extract 4 involves discussion about Chelsea Manning (who is not present). Manning became famous in a whistleblowing case leading to her incarceration by the US military at a time when she was publicly known as a man. She publicly identified as a woman in 2013 while in prison (Capuzza, 2015) – an event that attracted media attention. In 2017, Obama (then US president) commuted Manning’s sentence, leading to her release. In extract 4, a journalist from outlet TMZ (Jou) is soliciting Judge Judy’s (Jud; known for a courtroom TV show) opinion about Obama’s decision. Before the extract, the journalist invited Judge Judy’s view with: ‘Did she serve enough time?’ Judge Judy declined to comment (data not shown). In extract 4, the journalist pursues her view with a question (lines 1–2), expanded through the increment ‘To- (.) to- commute his sentence’ (line 2). The latter features the gendered possessive pronoun ‘his’ (line 2) with reference to Manning.
(4) DE05 ‘His sentence’
02 leakers of information. (.) To- (.) to- commute his sentence.
03 (0.4)
04 Jou: H
05 (1.2)
06 Jud: I didn’t think it was a wise judgement.
Notably, the journalist uses ‘his’ for Manning, despite having previously referred to her as ‘she’ in the same interaction (data not shown). Judge Judy does not immediately respond in line 3, and a gap emerges (Sacks et al., 1974), which the journalist uses to self-correct ‘his’ to ‘her’ (line 4; in CA terms, through a self-initiated self repair in the transition space; Schegloff et al., 1977). Consistent with the findings in our previous section, a lack of explanation for the misgendering embodies participants’ treatment of the self-correction as ‘good enough’, thus prioritising interactional progressivity over additional remedial work. The journalist therefore does not articulate the assumptions that may have grounded the misgendering. However, the gendered pronoun in line 2 (‘his’) matches Manning’s former gender designation. Therefore, inferentially available is the possibility that the misgendering was grounded in the assumption that people’s current gender maps onto their gender history – and thus, that gender is a stable feature across time (although the correction implies an overriding of these considerations in line with the norm that prioritises people’s current self-designations). This aspect remains unarticulated and contributes to establishing it as a tacit basis of gender attribution.
One circumstance in which explanations for misgendering emerge is when a participant makes their position towards someone’s gender ambiguous. As an example, extract 5 is from a videorecording uploaded by an advocacy group. It features campaigners knocking on residents’ doors in a US neighbourhood and engaging them in discussion about transgender rights. In extract 5, the participants establish that someone in the resident’s family is transgender (lines 1–4). In response to the campaigner’s (Cam) invitation to elaborate (line 6), the resident (Res) identifies the relative with the gendered kinship term ‘nephew’ (line 8).
(5) DE11 ‘Nephew . . . niece’
02 (0.4)
03 Cam: So- is transgender?=
04 Res: =Mm hm.
05 (0.2)
06 Cam: U:m ↑cool. How are you:: how are you rel
07 (0.4)
08 Res: It’s my nephew. [
09 Cam: [Your ] n
10 (.)
11 Res: [n
12 Cam: [Or ] [niece. ] Niece. Okay.
13 (0.3)
14 Cam: Okay.
15 Cam: So she: (.) um
16 (.)
17 Res: He was [born] a bo[y (but)] he wants to be a ↑girl.
18 Cam: [(mm-)] [e:h ]
19 Cam: ↑Okay_(0.2) Has- how much has she talked to you (0.2) about
20 that at all.
The resident self-corrects ‘nephew’ to ‘niece’ (lines 8 and 11), thus establishing the former as misgendering the relative. The ‘or’ preface (line 8; Lerner and Kitzinger, 2015) and appended ‘whatever’ (uttered after the resident starts to revert to ‘nephew’ but cuts it off at the first syllable; line 11) embody an ambivalent commitment towards the niece’s gender. Meanwhile, the campaigner settles on ‘niece’ (line 12) and uses the ‘she’ pronoun when inviting the resident to elaborate (line 15).
The campaigner prompts the resident to elaborate through an incomplete turn construction (‘So she: (.) um’) rather than a full question – a design that may embody caution in the context of ambiguity about the resident’s position towards the niece’s gender. The prompt works as an invitation to resolve (and possibly explain) that ambiguity. The resident’s response provides this by describing the niece as someone who ‘was born a boy (but) wants to be a girl’ (line 17). The focus on gender assigned at birth makes available the resident’s reliance on the niece’s former gender designation – an approach further embodied in the use of ‘he’ in the same turn to refer to the niece (line 17).
In the previous section, we presented evidence that participants’ explanations for misgendering a co-present person omit details of that person’s self-presentation that might be regarded as incongruent with their gender. By contrast, explanations for misgendering absent persons more liberally invoke those persons’ former gender designations (extract 5). This can be partly explained by the absence of the misgendered person, which may relax ‘face’ considerations. However, it is also possible that participants treat others’ gender histories as a less delicate topic compared with their self-presentations. This is likely grounded in the experienced ‘reality’ of institutionally sanctioned gender ascriptions (especially when presumably encoded as ‘sex’ at birth). Explanations that invoke someone’s former gender thus naturalise the cisgenderist assumption that gender is an enduring reality expected to remain constant throughout someone’s lifetime.
Explanations are also produced in contexts of persistent incongruity between participants’ positions towards somebody’s gender designation. Extract 6 is from an interview with Zinnia Jones (Jon) and focuses on Chelsea Manning (who is not present). The interview aired in 2013, on the day that Manning publicly self-identified as a woman. The interviewer, Jake Tapper (Tap), has introduced Jones as a friend of Manning; discussion before the extract has focused on how Manning used to confide to Jones problems she was experiencing when serving in the military (data not shown). As the extract begins, Tapper asks about Manning’s and Jones’ ‘statuses’ (lines 1–4), which alludes to their transgender identities. Referring to a time when Manning was publicly known as a man, Tapper uses her deadname, 6 ‘Bradley’ (line 2).
(6) DE07 ‘Bradley’
02 as was- u:h Bradley, (0.2) uh but you <
03 to: (0.2) uh the:se (.) s:
04 right?
05 (0.9)
06 Jon: Y
07 evidence tha:t sh
08 as two thousand t
09 Tap: ○.hh○ And when you refer to (0.2) uh Bradley Manning uh- (0.2)
10 because at the time (.) certainly he was- (.) he was
11 Br
12 people were ha
13 being a g
Tapper’s use of ‘Bradley’ and ‘man’ are instances of retroactive misgendering, ‘a form of misgendering that disregards people’s current descriptions of their genders when describing time periods prior to those self-identifications’ (Ansara and Hegarty, 2014: 266). This is evidenced in Jones’ answer (line 7), where she refers to Manning with ‘she’ (in CA terms, this is an embedded correction [Jefferson, 1987], a form of correction not involving repair practices). Tapper continues referring to Manning as ‘Bradley’ (line 9) and uses the gendered pronoun ‘he’ (line 10). This creates a persistent incongruity between the participants’ positions. Tapper addresses this with a parenthetical explanation (lines 10–11), which grounds his choice to refer to Manning as man in the consideration that ‘at the time’ Manning ‘was Bradley’ (lines 10–11). Consistent with the finding exemplified with extract 5, Manning’s former gender designation is liberally mentioned. Tapper’s explanation normalises the persistent reliance on a former designation as basis of gender attribution. It further evokes related cisgenderist assumptions, such as the idea that Manning ‘was’ indeed a man at the time of the events under discussion – as opposed to the possibility that she always identified as a woman but did not publicly identify as such because of societal oppression.
Discussion
When misgendering occurs, the operation of gender assumptions is momentarily visible at the interactional ‘surface’. By adopting a constitutive view of gender (West and Zimmerman, 1987), our analyses show that interactional practices of misgendering embody and reproduce cisgenderist assumptions that map people’s gender onto their self-presentation and onto their gender histories. Because our data mostly contain UK and US interactions, they do not capture the full spectrum of practices through which people are misgendered. 7 With this limitation in mind, we observe that the gender assumptions evidenced in our data uphold a gender order that systematically disadvantages people whose appearance does not conform to culturally dominant standards of gendered self-presentation, and people whose gender self-designations do not align with their gender histories and assigned ‘sex’ (Ansara and Hegarty, 2013, 2014; Serano, 2009).
Our empirical focus on misgendering enabled us to investigate gender attribution through direct observation of naturally occurring interactions – something that had not been done previously. Our unique theoretical contribution is to explicate how organisational features of face-to-face interaction provide procedural infrastructures for the reproduction of gender assumptions as assumptions – that is, as unnoticed and normalised understandings about gender (Kessler and McKenna, 1978; West and Zimmerman, 1987). When people correct misgendering, they often do not verbally articulate the assumptions that may have driven a gender (mis)attribution (extracts 1 and 4). We accounted for this by reference to a minimisation of remedial work in favour of interactional progressivity. When people do articulate their assumptions, other organisational features converge in minimising the extent to which those assumptions are detailed. People withhold pointing out aspects of co-present participants’ self-presentation that may have grounded a gender misattribution to avoid threatening their positive ‘face’ (extract 2). In adversarial interactions, withholding such details protects participants against possible challenges (extract 3). Across the situational contingencies in which misgendering occurs, organisational features of interaction promote an implicitness of common assumptions. These findings are consistent with an understanding of social interaction as a system that ‘maximises inference’ (Levinson, 2000: 29), thus freeing participants from spelling out the assumptions that they rely upon to construct meaning. This has important reflexive consequences: by treating certain assumptions as not requiring articulation or detailing, participants contribute to (re)producing them as unquestioned, normative criteria of gender attribution in face-to-face interactions (Kitzinger, 2005). At the interface between the interaction order and the gender order, we thus find procedures that help maintain macro-social arrangements and cultural ideologies as unquestioned foundations of social life.
The increasing visibility of misgendering as a social phenomenon (Capuzza, 2015; Gupta, 2019; Ingram, 2019) and the recurrence of instances in which it is corrected or called out, reflect the rise of a norm whereby people’s self-designation of their own gender should be respected. Building on Meyerowitz (2002), Westbrook and Schilt (2014) observed that the increasing adoption of identity-based criteria of gender determination (at least in nonsexual spaces and in gender-integrated spaces) reflects a ‘liberal moment’ in the constitution of gender. Our analyses further identify interactional contingencies that lead to exposing gender assumptions. This includes instances where misgendering raises concerns about competence, responsibility and ‘face’ (extracts 2 and 3), or more fundamental issues of intersubjectivity – including ‘reality disjunctures’ (Pollner, 1976), that is, radical incompatibilities between participants’ positions towards someone’s gender (extracts 5 and 6). In such cases, gender assumptions become the object of more explicit articulation, and this may prospectively offer opportunities to question them, thus de-normalising their use. 8 This could gradually transform the gender order by reducing people’s reliance on cisgenderist assumptions, moving towards a more just system where people ask – rather than assume – how others self-identify (this could be conceptualised as a ‘redoing’ of gender; Connell, 2010; West and Zimmerman, 2009). Alternatively, it could entail transition to a system where people largely abandon gendered pronouns, address terms and descriptors in favour of gender-neutral language (this could be conceptualised as a partial ‘undoing’ of gender in interaction; Connell, 2010; Risman, 2009; West and Zimmerman, 2009).
The rise of identity-based criteria of gender determination (Westbrook and Schilt, 2014) and the increasing sanctionability of misgendering (Edmonds and Pino, 2023; Kapusta, 2016) raise, at least theoretically, the possibility of ‘redoing’ or ‘undoing’ gender. However, we also found that when misgendering disrupts gender attribution in interaction, participants deploy explanations that treat momentary gender misattributions as exceptions that ‘prove the rule’ (Heritage, 1988: 140). They work to prop up (rather than dismantle) the cisgenderist assumptions that uphold the gender order (see Capuzza, 2015; Speer, 2005). These maintenance practices are reminiscent of macro-social trends in which attempts to authenticate people’s self-designations of their genders and bodies engender ‘a moral panic around the breakdown of conventional notions of sex/gender’ (Pearce et al., 2020: 682). Our study thus highlights an unresolved tension in which opportunities to transform the gender order are offset by practices that sustain it. Future research should investigate how these lines of tension can be leveraged to promote a shift towards more just systems that are less reliant on cisgenderism as a foundation of gender attribution in social interaction.
Supplemental Material
sj-pdf-1-soc-10.1177_00380385241237194 – Supplemental material for Misgendering, Cisgenderism and the Reproduction of the Gender Order in Social Interaction
Supplemental material, sj-pdf-1-soc-10.1177_00380385241237194 for Misgendering, Cisgenderism and the Reproduction of the Gender Order in Social Interaction by Marco Pino and David Matthew Edmonds in Sociology
Footnotes
Acknowledgements
We are grateful to Sam Hope and David Wilson for their generous feedback and suggestions in the first stage of the project that led to this publication. We would like to thank Onni Gust for their invaluable comments on an earlier version for this manuscript. We are grateful to Magnus Hamann for bringing to our attention some instances of misgendering, including extract 2. Finally, we want to express our gratitude to two anonymous reviewers for their very helpful comments and suggestions during the review process.
Funding
The authors received no financial support for the research, authorship and/or publication of this article.
Ethics statement
The first author’s university considered the study exempt from ethical approval because of the public nature of the recordings. The second author’s university provided ethical approval (Human Research Ethics Committee of the University of Hong Kong, Reference Number: EA210446).
Supplemental material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
