2.1 Constituting Identities

In this paper, identity is construed as multiple and hybrid identities constructed in discourse with different people for different purposes and in different contexts to reflect our representations of ourselves and others’ recognition of who we are (Butler, 1993; Gee, 2001; Hall, 1997; Lemke, 2008, 2009). A student with a physical disability may foreground her advanced conceptual knowledge in an economics class, be recognised as a top-performing debater globally, and actively leverage her achievements to advocate for disability groups nationally.

Identity is constituted and situated in socio-historical and socio-cultural repertoires of practices consistent with a community’s cultural models (Butler, 1990; Gee, 2000; Holland et al., 1998; Lave & Wenger, 1991; Street, 2003; Wenger, 1998). Cultural models are worldviews shared by a community’s members that shape their ways of thinking, speaking and acting (Quinn & Holland, 2012).

A community’s cultural models are underpinned by layers of extended social networks and institutions (Norris, 2011). Therefore, the cultural models of the communities to which an individual belongs, and the social networks and institutions that influence these cultural models, are brought to bear on the individual’s identities. In the aforementioned example, the student’s identities are those of an outstanding member of a localised learning community and an exclusive global club, as well as a vocal and active member of an inclusive community for people with disabilities.

Identities are a resource in service of power, for establishing relational positions and accessing networks (Moje, 2013; Moje et al., 2008). Identity shifts can occur in tandem with the introduction of new forms of cultural resources and participation in new social contexts where identities are recognised, constructed, supported, contested, negotiated, and/or reified (Nasir & Saxe, 2003) such as when the said student graduates to become an economist and discovers new resources and skillsets at her disposal to advocate for disability groups.

First-wave identity theory confines its study to discourse reflecting possibly multiple identity affiliations within a given time and space (Gee, 2001; Holland, et al., 1998; Wenger, 1998) or “identity-in-the-moment” (Lemke, 2008, p. 23). Second-wave identity theory also considers “identity-across-events-and-lifespan” (Lemke, 2008, p. 23) and, more specifically, processes at different time scales embedded in self-characterisations to facilitate an examination of the historical and relational nature of identity construction (Somers, 1994; Somers & Gibson, 1994). Essentially, the second wave encompasses both identity as performativity and identity as narrativity. An exploration of both forms of identity is needed to facilitate an examination of micro to macro linkages that account for shifts in identity (Lemke, 2008).

To illustrate the distinction between identity as performativity and identity as narrativity, consider the example of a student whose first language is Vietnamese and who has developed fluency in the English language. From the perspective of identity as performativity, that student may volunteer to explain a concept in English and Vietnamese to assist his Vietnamese classmates in an English as a Second Language class, thus performing his role as a bilingual mediator and establishing his identity affiliations with both language communities. As for identity as narrativity, the same Vietnamese student may share his journey of learning English, describing the challenges he faced and the support he received that enabled him to persist in learning English. Identity as narrativity brings to light the slow and complex process of developing competency in the English language and the identity affiliations associated with that competency, while also highlighting the choices made to hold onto, forgo, and reconfigure aspects of his Vietnamese identity.

The actions undertaken to project an identity or have an identity ascribed to oneself (Butler, 1990) are the processes by which identity is constructed as individuals immerse themselves in the community and appropriate these practices (Gee, 2000; Lave & Wenger, 1991; Street, 2003; Wenger, 1998). They are often characterised in reductionist terms that allow for easy classification and demarcation of boundaries. Nevertheless, individuals can simultaneously hold different identity positions along several spectrums (Lemke, 2008), such as being positively stereotyped along the more intelligent-less intelligent spectrum and, simultaneously, being negatively stereotyped along a socially proficient-socially less proficient spectrum. Society bundles together practices associated with these simplified cultural stereotypes and rejects deviations from the norm (Lemke, 2008), resulting in the depiction of the intelligent but socially awkward computer geek in a movie such as The Social Network, and the gentle giant with less intellectual acumen but who is good-hearted, such as Rubeus Hagrid in the Harry Potter series.

Multiple affiliations within the complex social networks of contemporary society have led to identities being constantly in flux and becoming progressively fragmented as they are constructed in multiple ways across discourses, that are unrelated, intersecting or antagonistic (Hall, 1996). For example, individuals whose social and moral conservatism is grounded in their religious beliefs may find themselves at odds with others within the same religious community who also espouse neoliberal views, or individuals arguing that the rights of the individual be upheld may find themselves pitted against those who believe this includes unfettered access to guns.

Varying identity positions along multiple spectrums and multiple identity affiliations across social networks necessitate code-shifting of identity performance (Lemke, 2008) or navigating through hybrid spaces (Moje, 2013) as an identity “lives with and through, not despite, difference; by hybridity” (Hall, 2006, p. 235). These boundaries of difference are constantly being repositioned in relation to the different frames of reference (Hall, 2006) of ideologically disparate communities, leading to individuals downplaying and accentuating different discourse practices and their associated identities in different contexts. Examples of these would be individuals concealing their sexual orientation or their religious affiliations in specific contexts because of the personal, social and financial repercussions that can result from such identifications. With these multiple and hybrid shifting identities adopted across socio-historical and socio-cultural contexts, individuals are engaged simultaneously in processes of identification and othering, necessitating an examination of how the Other is constituted.

2.2 Constituting the Other

The Other is fundamental to self-constitution through differentiation (Butler, 1993; Derrida, 1981). Identities are constructed through a process of distinguishing how they are distinct from the Other, the constitutive outside (Butler, 1993; Derrida, 1981). As such, representations of the self and Other are mutually constitutive as these representations are judged relative to one another and in terms of how the self represents the Other and the Other the self (Coupland, 2010). However, when differentiating the self from the Other, the Other is often characterised as subordinate, morally inferior, abnormal, foreign and distant (Coupland, 2010; Jensen, 2011). This is because the Other is constituted in power relations. Power differentials are founded on the politics of exclusion by ascribing significance to differences (Bucholtz & Hall, 2005; Hall, 1996). Differing values assigned to different sets of discourse practices and positions lead to differences in mobility and access and persistent pressure on the Other to take on dominant discourse practices and positions (Luke, 2003).

Attributes such as “lazy”, “cunning”, “oversensitive” and name-calling such as “snowflake”, “edi/diversity hire”, “hillbilly” and “trailer trash” are ascribed not only to an individual but also typically to an Othered group in the dominant discourse. The corollary of delineating group-based attributes is discrimination and conflict based on group identifications (Coupland, 2010). Given how reductive the portrayal of the Other often is (Coupland, 2010), in identifying the Other, researchers also risk essentialising and reifying the attributes of a group. Nevertheless, a delineation of the Other in terms of group identity can be meaningful because skewed representations, ideological battles and contestations of power hierarchically and horizontally between groups differentiated by their dis/ability, race, gender, class and age still persist (Coupland, 2010; Sandoval, 2000a). Moreover, even shape-shifting individuals cannot conceal differences based on dis/ability, race, gender, class and age. The Other is also a useful analytic concept that would enable researchers to “assess the degree of openness/closure of social identification” (Coupland, 2010, p. 257).

An examination is needed of the prevailing representation of Othered groups within possible sites of oppression, to shed light on the values, assumptions and norms underpinning dominant and marginalised discourses. Precautions are needed to ensure this is not the hegemony arising from speaking for or about the Other, which some scholars like hooks (1990) and Trinh (1989) oppose. It is the concerted effort of speaking with the Other (Kitzinger & Wilkinson, 1996; Liddle & Rai, 1993/2005) or speaking as the Other (Rappa, 2015) through which we gain greater insight into identification and othering, representing the first step towards infiltrating the dominant discourse and bridging the gaps. It is also imperative that we examine how these Othered groups respond to such representations—actively or passively, and embracing, acquiescing or resisting. In identifying discourse practices that lend themselves to othering, new practices can be initiated to alter the dominant discourse. Literature on role-playing with llms will now be examined, highlighting the potential for exploring interactions between self and Other.

2.3 Role-Playing with llms

Researchers have explored the use of a broad range of technological tools to make education more inclusive for targeted populations, and these have yielded substantial benefits for learners; enhancing access to learning content and activities and enabling collaboration and interaction that facilitate active and equitable participation (Navas-Bonilla et al., 2025). While these inclusive technological tools have been crucial for removing barriers to participation, their role has been limited to facilitating access to learning opportunities. Technological tools like llms have tremendous potential not only to enable access for targeted populations but also to enable those playing a critical part in making inclusive education a reality for the marginalised to reflect on the values and beliefs that inform their discourse with the Other through roleplay.

There are two ways to define role-playing in education. The first is from the standpoint of the role-play experience itself where the role-play is an instructional method involving a simulation of facets of actual or fictional scenarios depicting an issue or problem that is currently being studied, to facilitate the achievement of learning objectives through learners adopting specific character roles and working at varying levels of agency to complete specific tasks (Winardy & Septiana, 2023). The second definition is from the individual role-players’ perspective as they take on characters with specific attributes and functions and control these characters in such a way that they speak and act in ways they consider their characters would (Williams, et al., 2011). Both definitions are salient to role-playing with llms. However, in this instance, llms assume a character role as well to provide a means for learners to play out their role. As such, user interactions with llms in role-playing scenarios involve both the users and the llms engaging in role-play.

In the research thus far where llms facilitate educational role-play, llms offer practice opportunities to (i) develop specific professional competencies, such as conversations between health professionals and patients, counsellors and their clients, conducting negotiations (Dinnar et al., 2025; Fung & Laing, 2024; Gray et al., 2024; Grundmann et al., 2025; Holderried et al., 2024; Maurya, 2023; Rädel-Ablass et al., 2025), (ii) enhance language competence (God-win-Jones, 2024; Timpe-Laughlin, et al., 2025), and (iii) develop social and emotional competence (Nanda et al., 2025; Yan & Xiang, 2025). Despite the varied roles that have been assigned to llms, and nascent literature on perspective-taking and empathy building, there has been limited research on how llms have been used to facilitate identification with the Other and to examine the processes of identification and othering. There is a critical need to dig deeper beyond a surface-level focus on the development of competencies, as implicit within these role-play scenarios are power differentials wherein the learners in their character roles hold more power than the llms in theirs as excluded populations. While a demonstration of advanced competencies might be indicative of greater identification and less othering, researchers cannot determine this without a conceptual framework to facilitate such analysis.

3.1 The Habitus of Large Language Models: An Overview

Bourdieu’s notions of habitus, field, distinction, capital and profit of distinction underpin this framework (see Figure 1). Habitus refers to an individual’s ingrained and anticipated ways of thinking, speaking, and acting inculcated through frequent interactions within key social networks (Bourdieu, 1990). The habitus is situated within and shaped by one or more fields. The field is a space of practice (for e.g. education, medicine, law, politics) within which institutional power upholds the values, beliefs and practices of those already possessing valued forms of economic, social and cultural capital. People assume power-related positions within a field based on their access to such capital (Bourdieu, 1977). However, each field also reflects conflicts arising from ongoing contests over the legitimacy and relevance of these different forms of economic, social and cultural capital to the field of practice (Bourdieu, 1984). In an inclusive education context where llms are deployed for role-playing, the habitus of the Other is located within the llms’ interface. llms representing the Other could potentially be designed to reflect the complex competing discourses within a field. When individuals encounter the meshed fields of familiar and unfamiliar within the llms, they may feel compelled to think, speak and act in ways different from their habitus of the classroom.

Semiotic, phenomenological and sociological identifications of performativity and narrativity are constitutive of this habitus. These types of identifications are drawn from Lemke’s (2009) work, and concepts and processes associated with these identifications are further developed in the conceptual framework for exploring identification and othering (see Figure 1).

Semiotic and phenomenological identifications can reflect individuals’ sociological identifications or how they align themselves with respect to different social groups. Although individuals have multiple group affiliations complexly related to one another and to the Other, I have simplified these group affiliations into two broad categories to facilitate analysis—Non-Other and Other, as illustrated in the Venn diagram. Tools for analysing semiotic identifications, drawn from the fields of cultural studies (see Hall, 1997) and feminist studies (see Nussbaum, 1995; Sandoval, 2000a), are the processes of othering (see Section 3.3) and revolutionising (see Section 3.4). The two categories of sociological identifications with Non-Other and Other intersect to illustrate the plausibility of individuals at times straddling both groupings.

The continuum illustrates various phenomenological identifications individuals may assume. Their lived experiences during the role-play situated on the continuum may range from a persistent Self-Other binary with no immersion, where semiotic identifications reflect the processes of othering on one end, to Self-Other mergence with full immersion, where semiotic identifications reflect the processes of revolutionising on the other end. Along the continuum between these two poles are variable positions where the individuals’ discourse may be an instantiation of less othering in the non-Other sociological identifications or less revolutionising in Other sociological identifications. Within the intersection of the two broad categories, individuals complexly navigate in and out of these sociological identifications or hybridise their discourse.

View Full Size

A conceptual framework on identifications in the Other’s habitus

Citation: Innovation and Education 7, 1 (2025) ; 10.1163/25248502-bja00012

• Download Figure
• Download figure as PowerPoint slide

Hybridising and navigating represent two possible ways in which resources of a discourse may be employed, and their instantiations are variable. According to Bakhtin (1981), an organic hybrid is a mixture of utterances, styles, languages and systems of meaning-making and beliefs, perpetually evolving with new worldviews. As such, “all our utterances are filled with others’ words [with] varying degrees of otherness or varying degrees of our-own-ness” (Bakhtin, 1986, p. 89). Although Bakhtin (1981) believes the organic hybrid has no formal boundary and is essentially “opaque”, it is plausible to identify these new internal forms by discerning emerging patterns of hybrid language use.

Moje’s (2013) theorisations about navigating could account for both hybrid and non-hybrid language use by role-players. Navigating is a deliberate action undertaken by individuals to position themselves within a discourse. It arises from an understanding of the norms and practices valued by different discourse communities in various social contexts.. People navigate within and across spaces and, in so doing, experience moments of hybridity as they confront the in-between, the discourse that is neither their own nor the Other’s, the practice that they both take up and change. They navigate—or need to learn to navigate—the positions in which they find themselves or which are made available to them in various spaces.

Through explorations of instantiations and the inter-relation of these instantiations of hybridising and navigating (see Rappa, 2015), I have developed an analytical tool particularly for phenomenological identifications. This will be discussed in Section 3.5.

3.2 Semiotic, Phenomenological and Sociological Identifications

In this section, semiotic and phenomenological identifications are unpacked, and their relationship to sociological identifications is explained. In semiotic identifications, identity is constituted within representations with both corporeal and cultural resources being deployed to signify subjectivities (Butler, 1993; Hall, 1996, 2006). The selection of representational markers is deliberate, such that there is an accumulation of meanings across the chosen semiotic resources serving to distinguish or render more clearly the individual’s relationship to different people and groups. These choices occur in conjunction with the recognition and/or desiring of some commonality or disassociation with characteristics and/or beliefs of another person or group (Hall, 1996; Lemke, 2009). To find acceptance in different social circles, individuals might adopt different linguistic forms, such as using more erudite vocabulary or resorting to boasting or putdowns. Their ability to switch to a different set of linguistic resources as they move between social circles is likely to influence how they are received by their peers. As such, semiotic identifications can convey differing permutations of power and status differentials depending on the community in which one positions oneself (Lemke, 2009). The linguistic resources deployed while role-playing with llms are indicative of such semiotic identifications.

Phenomenological identification constitutes lived experiences (Lemke, 2008) and is mediated by others’ phenomenological responses to the individuals’ discourse practices, resulting in a series of exchanges as individuals act to refute or negotiate the assigned identity through their subsequent interactions (Norris, 2011). Lemke (2008) suggests that our identities are constructed in response to desires and fears that begin with corporeal wants and needs and extend to those elaborated by our cultures. Hence, an exploration of learners’ phenomenological identifications would consider how their lived experience within the social space of their role play with llms constitutestheirnegotiation of their identity within power relations. This entails an examination of identity-in-the-moment and identity-across-events or identity-across-the-lifespan through its recurrence (Lemke, 2008).

Both the semiotic and phenomenological can impact the sociological in that we “mobilize these identifications [or cultural capital] to perform the identities we wish to project and have recognized or ratified by others [or increase social capital]” (Lemke, 2009, p. 147) in each community. This is because the ways of knowing, speaking, and acting associated with these semiotic identifications may be labelled in different ways by different communities, for example, confidence or arrogance as opposed to humility or modesty. Such identifications map the trajectory of participation in discursive practices and are, at the same time, circumscribed by the invocation of cultural models through the discursive practices of specific communities (Butler, 1993; Moje & Lewis, 2007; Moje et al., 2008; Wenger, 1998). As such, an examination of sociological identifications involves tracing the constitution of identities in performances and in narratives across social and private spaces. These would include written artefacts in preparation for the role play or following the role play, discourse whilst role-playing with llms, collective and individual reflections post-role play and, more crucially, comparing the constitution of learners’ identities as they shift between role-playing a character and engaging in out-of-character communication with their group members or in their writing and reflections as a learner within and across different spaces.

Ultimately, the purpose of analysing these different types of identifications, performative and narrative, is to examine pathways that learners have taken as they hybridise or navigate their way from self to Other. If the aim is to facilitate identification with the Other, then the curriculum design for role-playing with llms needs to awaken a sense of responsibility or justice among learners. In discerning the needs of the Other and finding ways to rally resources to address these challenges, the learner would then be prompted to move towards the Other (Derrida, 1992, 1997; Levinas, (1972/2006). However, responsibility and movement towards the Other is contingent on the self having certain capacities such as reflexivity, discernment and recognition (Ricoeur, 1994). These capacities would need to be developed prior to role-playing with llms, providing ample opportunities prior to role-playing for learners to learn about and reflect on the lived experience of Othered groups whose identities the llms would emulate. The llms themselves would need to be trained on data on the specific lived experiences of Othered groups to ensure that their representations of these Othered groups are not themselves a form of othering. The following sections describe the tools that researchers can deploy to analyse these identifications when learners role-play with llms.

3.3 Othering Processes in Semiotic Identifications

This section addresses the processes by which groups and individuals are subject to othering. Othering is essentially a societal rather than an individual problem (Galvin, 2003). It involves constructing dichotomies in which one group dominates and shapes definitional understandings of the other (Derrida, 1981). It is a discursive process exercised by the powerful to characterise the identity of subordinate groups in a reductive and unfavourable way to legitimise the power and position of former and, at the same time, disadvantage the latter (Coupland, 2010; Jensen, 2011). Othering is accomplished by means of analytic strategies to simplify and reduce in order to explicate differences. Some of these strategies are discussed in the following paragraphs, namely, stereotyping, homogenisation, objectification, fetishism, binary oppositions, naturalization and silencing (see Coupland, 2010; Hall, 1997; Langton, 2009; Nussbaum, 1995) (See Figure 1).

Stereotyping is an approach in which a person or group is reduced to their supposed essence through labelling or ascribing traits, as exemplified by blacks being stereotyped negatively or positively as Toms, Coons, Tragic Mulattos, Mammies and Bad Bucks (see Bogle, 1973 in Hall, 1997, p. 251). This frequently occurs in tandem with homogenisation. Homogenisation is achieved through selecting and imbuing a mark of difference with an unfavourable meaning and applying it across the entire group. Some examples include a label such as “Islamic fundamentalist”, an icon on a road sign depicting two bent figures and one holding a walking stick to represent the elderly and the pronoun “they” accompanied by an ascribed pattern of behaviour (Coupland, 2010).

In more extreme forms, negative objectification and fetishism occur. In negative objectification, people are considered to lack autonomy and agency, and their experiences and feelings are ignored. They may be treated as a tool, as something owned or interchangeable or violable (Nussbaum, 1995). They may be reduced to their body or appearance (Langton, 2009). An example reflecting a combination of these methods of objectification is the male rendering of female virginity as an elixir of youth and, hence, a tradable commodity in Cambodia (Haworth, 2014). This has resulted in young girls being reduced to their bodies, treated as property, sold for cash, and having no say over their bodies, lives, desires and futures. This example is also illustrative of fetishism. Fetishism occurs when the whole is substituted by a part and involves the fragmentation of the Other and a displacement of these parts (Hall, 1997). For example, women’s genitalia, rather than their whole body, become the focal point in their representation in pornographic materials. These body parts serve as objects of sexual gratification for consumers of such materials. The othering processes I have highlighted in this paragraph illustrate how people are differentiated through a process of dehumanisation.

Differences in corporeal form and practices are then foregrounded through binary oppositions of “good/bad, civilised/ primitive, ugly/excessively attractive, repelling-because-different/ compelling-because strange-and-exotic” (Hall, 1997, p. 229). Such binary oppositions deny the plausibility of overlaps (Galvin, 2003). They are used not only to juxtapose the powerful and the disenfranchised, but also to distinguish from among the disenfranchised. In this sense, binary oppositions can even lead to people who are grouped together in social categories being cast as different in relation to one another (Hall, 1997), resulting in distinctions made, for example, in the skin tones or language varieties of individuals within the same community. Receptivity towards these differences is then established and legitimised by a recognised authority’s tautological statements in the form of aphorisms and maxims (Sandoval, 2000a), normalising (Galvin, 2003) and naturalising or attributing these differences to biological or inherited dispositions to preclude the possibility of change over any time frame (Hall, 1997). It is to render meaning unalterable by securing “discursive or ideological ‘closure’” (Hall, 1997, p. 245). A case in point is that, although the same language is spoken by individuals from the same community, there are different attributes ascribed to speakers of the variety considered “standard” and those of the varieties described as “dialects”. A “regime of representation” (Hall, 1997, p. 232) is established, ritualising these stereotypes and their associated practices and imputing further differentiations based on the initial difference (Goffman, 1963; Hall, 1997). To prevent contestation of these regimes of representation, stereotyped groups are silenced by their internalising these representations (Bourdieu, 1991), through limited access to discourse within these domains or by being rendered invisible in other forms of representation (Coupland, 2010) and by being further labelled as unreasonable or over-imagining or crazy in the event that they might raise their objections (Galvin, 2003; Goffman, 1963). Refer to Table 1 for a summary of the definitions of these analytic processes.

3.4 Revolutionising Processes in Semiotic Identifications

In this section, revolutionising processes are unpacked, pointing the way forward in countering othering processes and facilitating identification with the Other (see Figure 1). The most critical aspect of identifying with the Other, in my view, is that the self needs to engage in dialogue with the Other to internalise the voice of the Other (Ricoeur, 1994). But to do this, one must first hear the voice of this Other, and this voice must not be a subjugated voice but a voice that challenges the dominant discourse. Sandoval (2000a, 2000b) has identified the characteristics of such a voice in the “methodology of the oppressed”. This methodology, also termed “oppositional technologies of power”, stems from a commitment to a consciousness opposing the oppressive power of the dominant discourse and refers to “a set of processes, procedures and technologies for decolonizing the imagination” (Sandoval, 2000a, p. 68). It enables those who are oppressed to discern associations as contrived and arbitrary rather than natural by means of five different technologies.

In line with the purpose of the conceptual framework, I identified three out of the five as revolutionising processes that serve the purpose of unothering, namely, semiotics, deconstruction, meta-ideologizing (Sandoval, 2000a). I also drew on two other linguistic strategies Sandoval describes as ways to counter ideology—zero degree of language and silence. I propose broader definitions of some of these terms to allow broader applications of these revolutionising processes. From a prior study (Rappa, 2015), I identified other processes termed “silencing”, “humanising” and “counter-ideologizing”. These processes are explained in the following paragraphs.

The first oppositional technology of power is “semiotics”, which specifically refers to attending to and interpreting signs used in cultural productions (Sandoval, 2000a, 2000b). I have renamed this “signification” to avoid confusing the terminology with “semiotic identifications”. The following example illustrates how this technology may be deployed. There are assumptions of incapacitation and dependence associated with the wheelchair. To counter how the wheelchair is framed negatively in dominant discourse, new signs illustrating alternative ways in which the wheelchair is used to underscore the freedom and independence it brings, for example, showing someone hurtling down the racetrack on a wheelchair (see Tan, 2006; Figures 2 and 3) or doing daredevil stunts in his wheelchair (see Moser, 2005). These new signs introduce innovations into discourse in the form of metaphors, which can imbue an object with new meanings to counter prevailing institutional usage (Galvin, 2003).

View Full Size

Disabled access wheelchair

Citation: Innovation and Education 7, 1 (2025) ; 10.1163/25248502-bja00012

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Enabled access wheelchair

Citation: Innovation and Education 7, 1 (2025) ; 10.1163/25248502-bja00012

• Download Figure
• Download figure as PowerPoint slide

The second technology is “deconstruction”, which is a process of “separating a form from its dominant meaning” (Sandoval, 2000b, p. 376). This revolutionising process may be used to extract and account for an action undesirably labelled and rendered obscure in the dominant discourse, such as the use of a second or foreign language that reflects affiliations with marginalised groups. I have extended the meaning of deconstruction to encompass the act of making explicit what is tacit to expose the oppressive ideological assumptions and practices of the dominant discourse. This may entail, for example, questioning the practice of ascribing greater value or worth to monolingualism as opposed to bilingualism or multilingualism.

The third technology of power is “meta-ideologizing” which transforms dichotomies by extending conventional categories to include “something else, something beyond them” (Sandoval, 2000a, p. 84). This is not the same as enforcing the use of politically correct terminology. Newly-coined terms conceal oppressive realities and can quickly assume negative connotations (Galvin, 2003). Instead, what I understand by this may be illustrated in the current practices of subverting the meaning of derogatory terms. Subversion occurs when an unfavourable word is appropriated and invested with positive meanings to strengthen, heal and re-educate (Fletcher, 1993; Galvin, 2003), such as in the case of conventionally pejorative terms like “geek”, “crip” and “queer”. These illustrate the “language of revolution” because, in its generative, self-declarative and transparent state, it “transforms reality” (Sandoval, 2000a, p. 107). At the same time, the use of terms may be restricted to the members of a community, such as the Black community’s use of “n---a” as a marker of in-group solidarity and cultural identity, but where its usage in the wider community is considered offensive (Anderson & Lepore, 2013). In these examples, the language does not exclude ideology as claimed by Sandoval. Instead, it tackles ideology head-on. Such linguistic strategies confer new meanings that reposition Othered groups.

Zero degree of language occurs when individuals can circumvent ideological terms by connecting language to an object through action rather than a person (Sandoval, 2000a). For example, instead of referring to “a homeless man who lives in a tent”, we can refer to “a man making his home in a tent”. The person is transitively linked to “home”, and the statement is not about the individual but describes the action. Sandoval (2000a) argues that such a material response liberates the individual from ideology. In a role-play, this entails an examination of whether the emphasis is placed on the material act.

Another strategy is silence, which seeks liberation from ideology through active resistance (Sandoval, 2000a). However, ascribing intentions to silence can be tricky because silence can assume a submissive posture where one cowers in the face of oppressive ideology. As such, it is more accurate to characterise silence as non-cooperative. An example of non-cooperative silence may be found in a memoir by Carly Fleischmann, who refused to type out a response on her laptop to communicate with her parents and caregivers because she felt that the behaviourist techniques they adopted to train her were dehumanising to her (Fleischmann & Fleischmann, 2012).

“Silencing” as a revolutionising process is the equal and opposite reaction to silencing as a process of Othering. These may include statements in the imperative mood commanding the oppressor to stop the bullying or to get lost. The goal is to stop the propagation of prejudices embedded in the dominant discourse by preventing their articulation. “Counter-ideologising” refers to an attempt to marginalise the oppressor in the same way the oppressor has sought to alienate the oppressed. Both silencing and counter-ideologising are controversial because they are confrontational. They do not eliminate power differentials but, instead, seek to transform the discourse by shifting power from the oppressor to the oppressed.

More reconciliatory technologies would be those of “meta-ideologising” (discussed earlier) and “humanising”. “Humanising” is a revolutionising process aimed at countering attempts to dehumanise the Other. In this process, the Other gives vent to raw yet relatable emotions to emphasise shared humanity and to elicit empathy from those who are otherwise ignorant, afraid and prejudiced. Life stories of the Other are told as a way to make this Other known and understood. These life stories are a means to underscore the shared experiences and a way to account for differences. The discourse seeks to dispel fear and ignorance by conveying the message “I am like you”.

These revolutionising processes demonstrate how language can be reformed at the word, symbol, semantic and discourse levels. These reformations are then circulated within the discourse. However, such processes should not be confined only to those who are marginalised. They should be applied to those who hold positions of power to examine how they deploy semiotic, phenomenological and sociological identifications with the Other and whether instances of innovations occurred in their own discourse, which reconstituted their identifications. Refer to Table 2 for a summary of the definitions of these analytic processes.

3.5 Hybridising/Navigating Phenomenological Identifications

The analytical tool in this section depicts various phenomenological identification processes. This framework shifts identity studies away from identity as typology and towards identification processes, extending it beyond a list of assumed positions and identities to developing metacognitive understandings of identifications and facilitating comparisons with the identity of the Other projected by llms.

The framework drew on Goffman’s (1981, p. 128) notion of “footing” referring to the “alignment, or set, or stance, or posture, or projected self” of the role-players, which can be located in an utterance in the form of “codeswitching” or “pitch, volume, rhythm, stress, tonal quality”. Goffman (1981, p. 128) emphasises the need for a continuum to illustrate obvious to subtle changes in footing with the “new footing having a liminal role, serving as a buffer between two more substantially sustained episodes”. Discourse moves are incorporated into this notion of footing to illustrate how learners positioned their identity relative to the identity of the Other as depicted by the llms, highlighting changes in footing relative to that projected by the llms during the role-play. The terms to describe such discourse moves and their definitions are provided in Table 3.

The learners’ footing sequences along the self-Other binary – Self-Other Mergence continuum illustrate changes in footing in terms of differences from or similarities with the identity of the Other as they locate themselves on different parts of this continuum for different purposes (see Table 4). A continuum was chosen to illustrate that shifts between these discourse moves can be quite seamless. These discourse moves only provide insights into the individuals’ stance in relation to the Other as depicted by the llms. This framework does not assume uniformity amongst Othered groups and individuals and acknowledges the need to analyse the unique configuration of phenomenological identifications of specific Others.

3.6 Eliciting Sociological Identifications

As mentioned previously, individuals’ sociological identifications are complexly linked to multiple communities with similar or competing cultural models, situating these individuals at the intersection of these different group affiliations. These communities’ cultural models can help or hinder learners’ sociological identifications with the Other during their role. At the same time, it is critical to acknowledge the biases inherent in llms (Huang & Huang, 2025; Tao et al., 2024). After conducting a comprehensive examination for cultural bias in five widely used llms across 107 countries and territories, Tao et al. (2024) found that these llms favoured self-expression values, that is, the notion of subjective well-being, of English-speaking and Protestant European countries. They observed that specifying a specific cultural identity in the prompt reduced cultural bias. In line with this finding, scenarios foregrounding a contemporary issue (e.g. youth homelessness) for the role play may not explicitly identify the Othered group (e.g. young people who have a low socioeconomic status). Therefore, there is a need to make explicit the identity of the llm as an Other or a member of an Othered group, as well as reflect the complexity of multiple group affiliations of both the learner and the llm for role-playing to minimise cultural bias and move beyond perfunctory interactions between learners and llms.

In preparation for the role play, the parameters for the llm could be defined by the researcher through instructions on how the llm should respond and the resources that it should draw on to minimise cultural bias. A role-playing session could then begin with the learner in a specific role responding to a series of questions put forth by the llm about how they self-identify, their group affiliations, and their beliefs, values and experiences regarding a contemporary issue before commencing the role play to bring to the fore the sociocultural and economic factors shaping their views and to encourage reflexivity. Upon analysing the learners’ group affiliations and cultural models, the llm could be instructed to (i) always provide a response grounded in facts to respond to the queries of learners, (ii) customise responses to make the Other more relatable to learners by connecting with one or more of the learner’s group affiliations to encourage sociological identification and, once a bridge has been established, (iii) offer different or countering beliefs, values and experiences to challenge the learners’ assumptions about the Other. It is hoped that these steps can assist in identifying the learner’s existing sociological identifications and facilitate their sociological identification with the Othered group.

3.7 A Hypothetical Example

The hypothetical scenario presented in this section exemplifies how the conceptual framework and the continuum might be deployed to analyse learners’ othering of or identification with the Other. I have selected youth homelessness as the issue being addressed in a role play, with the learner taking on the role of a policymaker and the llm assuming the identity of a homeless young person.

Prior to role-playing, the learner might complete a brief survey on their views on youth homelessness issued by the lms, which could reveal that they belong to a lower socioeconomic class but live in a stable home environment. To distinguish themselves from homeless young people, they might express the sentiment that young people who are homeless are unwilling to work and prefer a free ride (sociological identifications). The learner’s views may or may not be evident in the interaction data during the role play. Therefore, the researcher needs to code both the self-description data and the interaction data for stereotyping and homogenisation (semiotic identifications). To counter the stereotype, the llm uploaded with the conceptual framework could autonomously respond to the learner by sharing systemic barriers young people who are homeless face, using the revolutionising processes of signification and humanising (semiotic identifications).

Along the self-Other continuum, some learners may echo the llms’ language, acknowledging how difficult it is to find work without a stable address, others may reconstitute that identity, emphasising the resilience of these young people instead of positioning them as victims, and yet others may subvert that identity by questioning the authenticity of the narrated experiences (phenomenological identifications). The researcher maps these moves on the continuum, noting shifts towards or away from self-Other mergence. The researcher might observe that those who express a strong affinity for high-performing and ambitious young people might navigate their discourses such that they express more empathy during interactions with the llm and express less empathetic views outside of that discourse. In contrast, those who participate actively in youth advocacy groups might hybridise their discourse, blending their personal experience as a youth advocate with the llm ’s narrative (phenomenological and sociological identifications). As such, an analysis of learners’ discourse would provide insights into the what, how and why of learners’ identifications. To encourage reflexivity, a co-reflection activity could be undertaken with the llm to unpack their discourse in line with the learning goals of the activity, and this could be followed up with a debriefing session with the educator or researcher and a counsellor to ensure the well-being of learners.

In summary, in presenting comprehensive analytical processes for semiotic, phenomenological and sociological identifications with or othering of the Other, the conceptual framework proposed in this paper addresses the different discourse processes and moves learners might deploy when interacting with the Other as depicted by llms during role-plays, and also the different discourse processes and moves llms could potentially be trained to respond with that could deepen learners’ experience of engaging with the Other.