Using genre to explain how children linguistically co-construct make-believe social scenarios in classroom role-play

: This paper argues that classroom role-play can be conceptualised theoretically as an oral genre, as de ﬁ ned within Systemic Functional Linguistics (SFL). The work draws on analysis of 15 video-recorded child-led role-plays in which groups of three 4 – 5 year-old children engage in ﬁ ve di ﬀ erent life-like social scenarios. The study is underpinned by SFL register and genre analysis of the children ’ s interactions, and the ﬁ ndings reveal how the children ’ s linguistic choices have a direct impact on the dynamically unfolding role-play, and how imaginary scenarios are construed by the instantiation of individual genre stages, some of which serve to regulate the role-play and others that mimic real life social scenarios. The ﬁ ndings suggest that the two di ﬀ erent types of stages construe two separate, but interwoven contexts, with the make-believe context often being dependent on the regulative context. The paper o ﬀ ers new insights into the ways in which SFL can reveal nuances in children ’ s dialogic and dynamic language in play.


Introduction
It has long been established within research into role-play that young children make deliberate linguistic choices based on their perception of the nature of the social encounter that they are playing (e.g., Sachs and Devine 1976).Research into role-play, a "… social activity in the sense that children in the play group relate to one another in accord with roles compatible with a dramatic theme" (Fein 1981(Fein : 1101)), has considered both the language utterances that are spoken in character and those spoken to organise the role-play (e.g., Blum-Kulka et al. 2004;Stagg Peterson et al. 2020).Alongside is work that explores how these playful scenarios can be explained theoretically including Goffman's (1974) notion of frame (e.g., Gordon 2002), and other with close investigations of language drawing on register (e.g., Hoyle 1998).However, an investigation of the lexicogrammatical choices that can explain role-play as a staged interaction drawing on genre as defined within Systemic Functional Linguistics (SFL) is yet to be fully explored.
In response, this paper to explores 'life-like' dramatic themes i.e., encounters that are reminiscent of typical day to day social scenarios, a doctor's appointment, or sales encounter in contrast to dinosaur land or teddy bears' picnic.Furthermore, it situates the study of role-play within the theoretical framing of SFL (e.g., Halliday and Matthiessen 2014) and draws on genre, a theoretical construct, to describe a social interaction within which 'stages' are co-constructed by, or between, language users in order to achieve expected social goals (Martin 1992;Martin andRose 2007, 2008).
The paper addresses the following questions: 1.How do 4-5 year old children in classroom role-play create life-like scenarios?2. How do the children's lexicogrammatical choices influence the role-plays?
In Section 2, I offer an overview of relevant research into children's role-play before outlining the details of the data collection and the analytical framework in Section 3. In Section 4, the findings are presented and then a conceptualisation of role-play using genre is presented in Section 5.The paper concludes in Section 6 and reflects on the value of understanding children's role-play as an oral genre distinct from other play genres.

Literature review
The case to explore children's classroom role-play interactions is justified by the attention role-play has received, for some time, with a focus on how children create and maintain engagement in role-play (e.g., Corsaro 1997;Garvey 1990;Sachs and Devine 1976;Sawyer 1997).In addition and more recently, there has been a turn towards understanding role-play within the context of early years curricula (e.g., Wahyuningsih and Suparno 2018); work that seeks to understand how the 'quality' of role-play can be supported through adults (e.g., Kalkusch et al. 2021); classroom interventions to encourage children to speak (e.g., Meacham et al. 2014;Sawyer and Brooks 2021); to support children's literacy development (e.g., Bluiett 2018;Rajapaksha 2016;Wahyuningsih and Suparno 2018), and in how children learn from each other in role-play (e.g., Bannerjee et al. 2016;Chen et al. 2020;Mukherjee 2016;Stagg Peterson et al. 2020).Future work relies on new in-depth understandings of role-play as a "genre of extended discourse" (Blum-Kulka et al. 2004: 308) with a focus on lexicogrammatical choice across the language of the imaginary social scene and its organisation.The notion of register has been used to explain ways in which young children construe role-play (e.g., Blum-Kulka et al. 2004;Hoyle 1998;Kyratzis 2007) For example, in roles as news-announcer, shop assistant and sportscaster, Hoyle (1998: 51) found that the children used "register-marking constructions" including syntactic reduction and simplified constructions.Thus, language indexical of particular social encounters helps explain role-play, but children must also alter their language in relation to other roles in the play (Hoyle 1998).
Attention to 'co-participants' is highlighted by Sachs and Devine (1976), in their recordings of young children talking to different 'listeners' including role-playing a mother talking to her 'baby'.Differences were found in the way in which the children spoke, with more imperatives used when speaking to the baby than to their (play) mother or peers.Similarly, other research into role-play found that imperatives were used more when addressing someone the children perceived as a lower rank (Mitchell-Kernan and Kernan 1977), and children playing mothers produce more utterances than children playing babies, and that mothers are more likely to use imperatives (Corsaro 1997).Through the presentation of one type of scenario and highlighting lexicogrammatical variation indexical of the register, the utterances and responses point to the young children's understanding of typical social roles and status, and the way in which they can be constructed through language choices.
Role-play is managed in part by children's regulating prompts within and outside the voice of the characters (e.g., Gordon 2002;Sheldon 1996;Stagg Peterson et al. 2020).The notion of metacommunication (first coined by Bateson (1971Bateson ( [1956] ] in observations of animals in play) provided the foundation for Goffman's (1974) concept of Frame and both concepts have been usefully deployed in the work on role-play (e.g., Buchbinder 2008;García-Sánchez 2010;Gordon 2002;Hoyle 1993;Sawyer 1993) to argue that role-play can be conceptualised through multiple embedded play frames to establish the play itself as well as the scenario and precise roles.Within the overall frame of play, metamessages are used both in the frame of the imaginary scenario and outside it, in the literal frame.The idea of 'footing' i.e., "the way in which framing is accomplished in verbal interaction" (Hoyle 1993: 115) through speech styles (Sawyer 1997) or prosodic features differentiating voices (e.g., Anderson 1990;Gordon 2002Gordon , 2006Gordon , 2008) ) indicate to others whether the utterance is integral to the literal frame or the play frame.Hoyle's (1993) work shows register marking constructions and argues that the play affords opportunities to practice different registers "together, in synchrony" (p.66).Thus, the idea of frames and footings, as a way of explaining pretend interactions and the movement between the pretend frames and literal frames without disruption to the overall frame of play is well established in the literature (Garvey 1990;Giffin 1984;Göncü 1993;Gordon 2002;Trawick-Smith 1998).Yet, the use of genre to explain children's playful interactions remains relatively unexplored.
Turning to the concept of genre, most research applying this construct to children's talk has focused on narrative (the telling of a story), often concentrating on talk by a single child.Oral narratives are defined by a series of progressive and coherent stages linked in some way with a particular purpose, thus broadly similar to a role-play and highlighting commonalities between oral narratives and spontaneous role-play.Linguistic analysis has revealed distinct oral genres in children's language interaction both inside the classroom (Hicks 1990) and outside the classroom (Preece 1987).Although they are described as subtle, differentiating linguistic markers are identified that enable Hicks (1990) to propose three narrative genres: narration of events, factual news reporting, and storytelling.Investigating whether different genres could be identified within three 5 year-old boys' conversations in play, Hoyte and colleagues (2014) found that genre, as defined within SFL, could help explain the instantiation of three distinct genres: making together; sharing personal information and story-telling.However, a step-by-step focus of the interaction's in-role and regulative utterances as dynamic and staged remains unexplored.By staged, I mean a close linguistic analysis to investigate the utterances to identify patterns within and across real-life imaginary social scenarios, such as making an appointment, or buying shoes.
3 Data and methodology

Data
The research reported in this paper explores data drawing on Clinician Consultations (baby clinics and vet's scenarios) and Service Encounters (cafe, pet shop and shoe shop scenarios).It takes a focused linguistic analysis to explain how the 'script' in role-play is achieved, drawing on the dialogic construction of individual by dynamic 'stages' as defined within SFL.These stages mimic real-life encounters with the children working towards a shared social (imagined) goal.
The data were collected in a co-educational state first school in an ethnically mixed area in the South East of England.From two Early Years (EY) classrooms and in consultation with their teachers, 4-5 year-old children were invited to participate.Participating children were organised into five groups of three, each with a mix of boys and girls.As found in the pilot and noted in other research (e.g., Meacham et al. 2014), three children were found to produce more recordable group talk than four children, who would split into pairs to play.
Reception classes (4-5 years) in England have spaces for role-play, which are themed with props to support 'early learning goals' (Department for Education 2020).
In the school where this data collection took place, an area shared by both classes was dedicated to role-play, and the video-and audio-recordings were made in this space.
One video camera was mounted on a tripod in the corner of the role-play area with a remote microphone, and an audio recorder was positioned in the role-play area to capture language missing from the video recordings.The role-play scenarios recorded were typical of Reception classes across England (Rogers and Evans 2008): a baby clinic, a cafe, pet shop, vet's, and a shoe shop.For each scenario, the area was furnished by the school with different props as per the usual routine: the baby clinic included a doctor's kit, dressing up clothes, dolls, a play telephone, paper and pencils, while the cafe provided a menu, till, plastic food, plates and cutlery.Each scenario was primed by the teacher though a class introduction that explained typical roles and what might happen in the scenario, for example, placing an order at the cafe.However, the resulting role-play was conducted outside immediate practitioner supervision and was entirely child-directed in the moment.
To minimize my involvement, after the recording equipment was in place, the children were left to play.The children sometimes asked me questions about a prop or to settle a squabble.I also was asked to assist with costumes (as noted by Melissa in Table 7).However, I did not influence who would take particular roles (the use or non-use of costumes was for the children to decide) or how the role-play would unfold and did not engage in any of the role-play.The data in this paper are drawn from 15 transcribed video-recordings each of between 15 and 20 min.As the research focus was lexicogrammatical choice, neither participant accent nor prosodic detail is represented in the transcriptions, but the children's emerging grammatical knowledge is accurately recorded.
Design of the consent process was shaped by the ethical considerations in the best practice guidelines at the time of the study set by the British Educational Research Association (BERA 2018) and the British Association for Applied Linguistics (BAAL 2021) and approval confirmed by the Open University Human Research Ethics Committee (HREC/2012/1320/Mukherjee/1).The children's parents/guardians were provided with information by letter and in a face-to-face meeting, and for participating children, signed consent was provided by the children's parents.To gain the children's "provisional consent" (Flewitt 2005: 556), they were provided with information orally by myself, the researcher and their class teacher, and throughout the data collection, the children's non-linguistic behaviour was monitored to ensure, as far as possible, that the children were comfortable with their involvement.

Analytical framework: genre and register from the
Systemic Functional Linguistics (SFL) perspective SFL offers a theoretical and analytical approach to language as a system of meaning based on language choices related to context.Within SFL, linguistic choices can be explored at both clause level and 'beyond the clause' through genre theory, where interactions are "staged, goal oriented social process" (Martin and Rose 2007: 6).These analytic tools are appropriate for investigating how children's language choices construe the nature of the social interaction in detail, dialogically moving forwards towards a shared (pretend) social goal.In addition, it can be used to explain how children move in and out of role-play, as genre reveals and highlights the dialogic and contextual nature of the language realized dynamically through children's lexicogrammatical choices.Lexicogrammatical choice underpins register within SFL (Halliday and Matthiessen 2014;Martin and Rose 2008).Register is instantiated through language resources that are aligned to the variables of field, tenor and mode (FTM) (Halliday and Matthiessen 2014;Martin and Rose 2008).Field is associated with the ideational metafunction and refers to the topic or focus of the interaction or activity and the degree of specialisation.It is realised through language in the choice of participants, processes and circumstances and by the degree of specialist lexis employed.Tenor is concerned with how language users position themselves and build and maintain relationships; users will employ different structures to exchange meanings, thus allowing the speaker to interact with the addressee in a variety of ways.Speech function is the categorisation of utterances (or sentences) by their function in the text, for example, a command or a question, and these functions map on the grammatical possibilities outlined in the function of mood, for example, a declarative, an interrogative.Mode is the register variable aligned with the textual metafunction that links language resources with the role of language and the degree of interaction and spontaneity within a text.
In this study nonverbal and multimodal aspects included were the children's engagement with the props, which may have included writing a note or handing another child something relevant to the play (these engagements are captured in the Notes columns from Table 2 onwards).The terms ancillary and constitutive describe the role that language has in relation to these other semiotic resources.The notions of register and genre can be understood as indivisible as genre is described as "the ways in which field, mode and tenor variables are phased together in a text" (Martin 1997: 12).Martin's approach to register and genre is adopted here for the clear link between his notion of register as realising genre.
Martin's notion of genre builds on Hasan's (1984) earlier work on genre (or generic structure potential), which was developed following her analysis of a face to face spoken sales encounter where she identified obligatory and optional elements of the text.For her, an element is defined as "a stage with some consequence in the progression of a text" (Hasan 1985: 56).Drawing on Hasan's (1985) sales encounter, empirical research on genre in spoken social interactions has been carried out including service encounters (Ventola 1987), doctor-patient interactions (Tebble 1999), two key interactions relevant for the analysis of role-play that mimics real-life.

Analytical procedure
The analytical process began by watching and making notes of the video-recordings and then fully transcribing 15 recordings.Across these transcriptions, I identified three broad 'categories' of language, and a decision was made as to whether each utterance was: i) spoken by the child as an imaginary role, described as in-role; ii) spoken by the child who was in some way referring to the ongoing role-play, described as regulative; or iii) another unconnected utterance, described as 'other'.The initial coding was then made more detailed in respect to the in-role and regulative utterances to include specific aspects of Field, Tenor and Mode (FTM) (Halliday and Matthiessen 2014).In order to explore potential differentiating register features, elements from within each of the FTM register variables were selected: Field: process type, field appropriate lexemes Tenor: mood, speech function, modality Mode: role of language, channel of communication During the register analysis, it appeared that meanings were being created through the dynamically unfolding interaction and co-construed through what appeared to be stages reminiscent of similar 'real-life' social encounters progressing towards a 'goal'.The implication of this new insight was that a synoptic description of the language choices based on register alone no longer accurately captured the children's role-play.It therefore became important to accommodate, theoretically and analytically, the complete text (i.e., from beginning to end of the recording) and to conceptualise and explain the unfolding.
The genre analysis began with an exploration of SFL research into real-life (that is not role-play) adult interactions undertaken in similar social encounters to those in the classroom role-play (Tebble 1999;Ventola 1987).By looking at the real-life research, I wanted to explore if and how the children were appropriating the stages and social goals reminiscent of a similar real-life scenario.In order to analyse the genre stages of the role-play scenarios, first I took the stages identified in the works of Tebble (1999) and Ventola (1987) and 'applied these' to each of the in-role utterances in the role-play data adopting a 'top down common sense' approach (Eggins and Slade 1997).Tebble's stages identified in her work on doctor-patient scenarios were aligned to the baby clinic and the vet's role-plays (and presented here as a Clinician Consultation), and Ventola's service encounter stages to the cafe, shoe shop and pet shop role-play scenarios (Table 1).
In addition to these stages, there were utterances that did not contribute directly to the social goal of the particular encounter.These utterances I termed 'enhancing utterances' as they enhance the children's imaginary scenario.For instance, in one role-play the waiting staff closed the café so they themselves could eat: "Now let's have our own food" (HC-EYC3). 1While fascinating, enhancing utterances are considered outside the scope of this paper.
In line with Eggins and Slade (1997), the early analysis was refined revealing differentiating features in the two types of social encounter (Clinician Consultation or Service Encounter) and in the in-role and regulative language.Typical lexicogrammatical features of each stage initiation across the five different role-play scenarios and the 15 different recordings were identified.In what follows, the paper offers details of the individual stages within the in-role utterances, as the children seek to accomplish a (imaginary) 'social goal' (e.g., buying football boots; making an appointment to see the doctor) and stages where the children speak as themselves to assist the play (regulative stages). 1 Each data extract cited is followed by a reference to the individual recording using scenario type BC Baby Clinic, PS Pet Shop, HC Healthy Café, V Vets, SS Shoe Shop; the class: EYM, Early Years, Masters (teacher pseudonym) or EYC Early Years, Cook (teacher pseudonym), and a number assigned to the group 1, 2, 3.
I turn now to examine the two types of stages in the children's role-plays: i) in-role: the children speaking as an imagined role; and ii) regulative: the children speaking as themselves organising the role-play.I then present how the different types of stages work together by taking the role suggestion stage to illustrate the interwoven nature of the dialogic interaction across the stages.

The in-role stages
Stages as identified by Tebble (1999) and Ventola (1987) appeared in the children's 'in role' language characterized in part by field-appropriate terms evoking technical or specialist vocabulary relevant to the social scenario.Specialist vocabulary relevant to the social scenario is highlighted in the literature (e.g., Sachs et al. 1985) as discussed earlier.However, based on the data in this paper, the analysis revealed that the children used field-appropriate processes in combination with appropriate clause configurations.The short extract in Table 2 illustrates how the stages identified in the real-life service encounter research (Ventola 1987) were mapped against an instantiation of a role-play where two children are buying a rabbit in a pet shop.Although the interaction does not show a 'complete' set of stages (as identified in Ventola 1987), there are enough stages constructed for the interaction to be reminiscent of a service encounter.

Genre in children's classroom role-play
In this short exchange, Nicole and Philip have 'become' customer and shopkeeper respectively through their language choices in the unfolding text.It demonstrates that while earlier literature suggests that children may appropriate relevant registers (Hoyle 1998;Kyratzis 2004) there are also dynamic meanings at the level of the stage that, in this dataset, are contributing to the construal of the scenario.By this I mean that it is not simply the use of particular field-specific lexical terms that ensures the exchange is reminiscent of a service encounter, but the dialogic nature of the children's language where they can be seen to exchange and construe meanings through their language and their use of props stage by stage.For example, both the customer and the shopkeeper instigate the stages expected in a real-life encounter, with Philip in turn 7 instigating a service bid when seeing that he has an opportunity to sell another of his pets, and the customer, wanting to buy the pet, checks the price and hands over the money.Not every service encounter in the data is completed quite so efficiently and not all finish quite so abruptly; it is more often the case that many of the service encounters and clinician consultations are interwoven with the regulative utterances serving to assist the ongoing progression of the individual scenarios and facilitate the playful engagement.

The regulative stages
While the children's role-plays are primed by teacher introductions, the unique instantiations are created in the moment by the children themselves.It is as a result of this spontaneity that at times the children interrupt a scenario in order to assist the unfolding role-play (as noted above and found in other research, e.g., Gordon 2002;Sawyer 1993).The present analysis also showed that in each of the role-play instantiations, the children use language to initiate and bid for particular roles; comment on the play equipment; and explicitly direct the unfolding play.Linguistic differences at the level of FTM between the three types of utterances mean that this regulative language can also be conceptualised as 'stages' with individual functions (Table 3).
Unlike the in-role stages, the regulative stages do not always logically flow from one to another.Instead, they may be prompted by a child wishing to swap roles, curiosity around a prop or a child seeking to 'manage' the interaction by directing a different outcome.While this characteristic might suggest that they are not stages as typically defined in genre theory, I argue for their categorization as stages for three reasons.Firstly, each stage has a functional goal that relates to, and directly facilitates, the in-role stages (Table 3).Secondly, the frequency of the stages across the five scenarios suggests that they are integral to the role-play interaction (Table 4).
The third point relates to the finding that the stages' identifiable functions and linguistic patterns span the dataset irrespective of whether the interaction has been coded as a Clinician Consultation or a Service Encounter, or the specific scenario within this (e.g., pet shop).For instance, across the data in the role-suggestion stage, the children assign and attempt to assign roles for themselves and for others, and this is achieved through commands realised by declaratives and imperatives.A role proposal by a child for their own role is realised through declaratives, and while these instances could have been coded as statements, it is possible that rather than simply giving information, these declaratives may also be signaling a form of command rather than simply a statement (Torr and Simpson 2003).
In addition, reference to dressing up and the use or approximation of the professional role was also indexical of this stage.For the role-play direction stage, declaratives or imperatives also functioned as commands directing the action, and a key element of this stage is the use of modality and in particular modulation to insist that the action is carried out, e.g., "you have to pay" (HC-EYC2), thus rendering the wish non-negotiable (as viewed by the speaker).Finally, questions realised as interrogatives realise the stage of the 'role-play equipment'.This stage is linked to The children attempt to affect the unfolding instantiation by proposing action.
Pretend you forgot it.

Role-play props
The children explore the props provided in the role-play area.
What's a cross?
Table : Initiation of the regulative stages by scenario type.

Scenario Role suggestion
Role-play direction Role-play props Genre in children's classroom role-play the instantiation of the role-play by the fact that the piece of equipment or word, once understood or identified by the children, can then be incorporated directly into the encounter, or that in some way the exchange around role-play equipment between the children makes something clear that allows the role-play to continue.

4.3
The unfolding of a social encounter through interwoven inrole and regulative stages By taking an example from a service encounter and a clinician consultation, this section explores how the in-role and regulative stages are interwoven in the unfolding role-play and discuss how they blend in order to facilitate the unfolding scenario.An example from a service encounter is shown in Table 5 with Alex, Alfie and Isla-Rose playing a shoe shop.They are at the point of getting their first service encounter underway and the role of shop assistant is the starting point of this excerpt with Alex suggesting the role of shop assistant to Alfie (turn 6).Alex's utterance "Pretend you were this now" (turn 6) together with passing the toy till to Alfie serves as the role-suggestion stage.The till is a semiotic resource, which here functions as a signal for the role of shop assistant and in doing so, Alex positions himself in a different role.The role-suggestion stage can be seen here to do more than simply assign a role, it instigates a shift into the service encounter.Alfie's subsequent utterance (turn 7) simultaneously agrees his role as shop assistant and initiatives the service encounter by addressing Isla-Rose as the customer in-role through a stage of Service bid (although Isla-Rose declines the pair of shoes offered).
Having successfully placed Alfie as the shop assistant, Alex (turn 9) then requests the role of customer for himself and assumes this role immediately through his instigation of a second Service stage.The role-play continues to unfold and Alex, through his negotiation of roles, continues to make an impact on the unfolding scenario.
The role suggestion stage has the potential to open a space for the negotiation of the roles, but the commands as realised by imperatives, "pretend you were this now" and "And pretend I was another customer" in Alex's turns 6 and 8, offer little room for negotiation and thereby appear to assist the acceptance of the roles by Alfie and Isla-Rose.The data also demonstrates that while the role-suggestion stage progresses the role-play, it is not enough on its own, i.e., it is the take up from the children accepting roles and employing in-role stages following the regulative stage that begin to instantiate or continue the social encounter.Alex's further contribution (turn 13) supports the social encounter (justifying his request for football boots), and while it does not contribute directly to the 'social goal', it builds a more developed construal of the individual stage.
In summary, an initial request in the regulative stage by one child, which is taken up by another (through their own role suggestion or acceptance), motivates a move into 'in-role' through the confirmation of all roles taking place through the in-role stages.This link provides evidence of a strong relationship between the regulative and the in-role stages, and I argue that it is productive to view these utterances as discreet stages in an interwoven staged interaction.An example from a clinician consultation demonstrates that this point works across the two different types of social scenarios.
In Table 6, Daniel, Yusra and Meggie are in the Baby Clinic, and the extract starts with Daniel and Yusra negotiating the nurse and doctor roles.In turn 1, the dual purpose of Daniel's utterance is to propose his preferred role and implicitly assign other possible roles to the rest of the group.Yusra (turn 2) responds agreeing with Daniel's role, and she proposes her own as nurse.Daniel (turn 3) continues to establish the roles through a command specifying something that Yusra, as the nurse, would have to do.These two roles are then confirmed immediately by Daniel and Yusra in their initiation of the in-role stages (Stating/eliciting problem) and by the acceptance of the third member of the group as the parent with the sick baby.Again, the impact of the speech function and choice of mood is relevant, as these have a direct influence on how the two types of stages blend.The commands realised by declaratives (in this example) in turns 2 and 3 are powerful positioning utterances that are effective in gaining a coveted role.In contrast, role proposals or requests that are seen in the data realised by interrogatives are less effective than a declarative in terms of securing a role.The presence of an interrogative opens up a space for negotiation by inviting a response that may be rejected or at times ignored completely.This finding is exemplified in the baby clinic where Ishaan requests the doctor role on six separate occasions and his request is flatly refused (as exemplified in lines 102, 104 and 106 in Table 7).
Differences in the children's language choices in relation to tenor variables of speech function, mood, interpersonal grammatical metaphor and modality influence the enacting of the social roles and contribute to variations in social roles, status and distance.Diagnosing facts Passes an instrument to Daniel

Discussion
A theoretical conceptualization of classroom role-play as an oral genreI have proposed that the peer-led interaction of real-life scenarios in classroom role-play is organised through a number of stages that can be split into two distinct types, those that mimic real-life, reminiscent of stages identified in research into similar real-life encounters by Ventola (1987) and Tebble (1999), and those that are regulative, which organise and manage the role-play.While such regulative utterances have already been identified in research into children's play (e.g., Gordon 2002;Sutton-Smith 1997), and it is widely agreed that these utterances' importance for children's role-play (as discussed earlier) have not been considered as genre stages previously.
In arguing that role-play can be defined as a staged interaction I propose that children's classroom role-play can be conceptualised as an oral genre illustrated in the synoptic model in Figure 1.
Three regulative stages span different role-play instantiations and types of scenarios.The in-role stages are formed through utterances that construe a life-like social scenario within a Service Encounter or a Clinician Consultation.Enhancing utterances, while out of scope in this paper, may show their own genre stages, genre switching, or genre mixing (Ventola 1987).The different functions and linguistic realizations of the two types of stages, I suggest, instantiate two different contexts of situation within the same overall interaction (classroom role-play), summarized through register in Table 8.
The individual stages have recognisable and predictable stages, yet characterized by different instantiations of different Field, Tenor and Mode (Table 8), highlighting subtle shifts in language and props as integral to the scenario construal.
While the findings have shown how one stage can motivate the next, there is unpredictability due to the spontaneous, co-constructed and oral nature of the a Melissa had previously asked me to help her with the doctor's tabard.
Genre in children's classroom role-play interaction, which has also been found in Ventola's (1987) work on spoken genres.Obligatory stages, as those that define a genre (Hasan 1985), were identified by Ventola (1987) to be unhelpful and that certain stages may still be left out and the social goal still accomplished.Ventola's observations resonate with classroom Mode Channel: face to face spoken interaction Role of language: mainly constitutive but may at times be ancillary to ongoing play Channel: mainly face to face spoken interaction, but may include telephone calls and the exchange of written documents a created as part of the role-play.Role of language: mainly ancillary Channel: mainly face to face spoken interaction with some written.Role of language: mainly ancillary.
a These were scripts created by the children themselves, for example a shopping list or an appointment note.role-play as a genre, and I argue that the more predictable flow of stages that may characterize other genres, particularly written genres (Martin 1992), is not required.Instead, the presence of regulative stages (which occur in all the role-plays across the dataset), and the attempts of in-role stages provide evidence of the role-play genre being instantiated, and that the stages function together to support the playful context dynamically.

Conclusion
Previous research has usefully established that children draw on indexical register features and speech styles to indicate a play or a literal frame (e.g., Blum-Kulka et al. 2004;Kyratzis 2007); it has also explained role-play through the concept of frames and scripts (e.g., Hoyle 1993;Sawyer 1993).The research reported in this paper departs from these established concepts and has aimed to reveal detailed lexicogrammatical patterns through SFL genre linguistic analysis across different imaginary real-life social interactions.It thus extends previous work that has identified indexical register features in single scenarios (e.g., Gordon 2008), or home themed role-plays (e.g., mother-baby talk, Sachs and Devine 1976).The paper has extended the findings in other work (e.g., Marjanovič-Umek and Lewnik-Mušek 2001) to show how props function alongside children's utterances to construe individual stages both in the regulative and in-role stagesthis is evidence that explains the children's utterances and engagement with props to show how they construct a shared social (imaginary) goal.
Both the Clinician Consultation and the Service Encounter stages are seen to be realised by appropriate lexicogrammatical choices with particular speech functions, appropriate mood constructions as well as technical lexemes reflecting and facilitating what is needed in order to accomplish the stage and progress to the next.Across the dataset of 15 video recordings of five different social scenarios, the lexicogrammatical patterns across FTM are seen at the level of the context (the in-role or the regulative contexts); the two different genre sub-types (Clinician Consultation and Service Encounter) and at the level of the stage vis-à-vis the overall progression of the unfolding text towards the social goal i.e., buying a pair of shoes.
These in-depth understandings reveal what the children understand about the world and achieve using language in the in-role context as well as the regulative stages: the children ask each other for support in understanding props through questions (Section 5.2), thereby supporting their own learning; stages appropriated are effectively modelled for the benefit of the whole group e.g., diagnosing facts (turn 6, Table 6); literacy is supported as the children are motivated to write, simultaneously, showing an awareness of how written texts are integral to some of the stages (e.g., making a list).An understanding of classroom role-play as an oral genre, and thus a complete language interaction different to other play genres identified (e.g., Hoyte et al. 2014) or language in play, reinforces the value of the regulative contexts as supporting the in-role contexts (Gordon 2002), in turn affording peer-led role-play in classrooms without the need for adult players in the role-plays (cf.Kalkusch et al. 2021;Rajapaksha 2016).
The present findings offer a greater understanding about the precise ways in which young children collaboratively construe pretend social scenarios which imitate different kinds of real-life experience, and reinforces children's role-play as a fascinating and rich site of research into children's language and understandings of the world.

Table  :
'Real-life' stages as captured in research by Tebble () and Ventola () and the role-play scenarios.

Table  :
A complete service encounter in the pet shop (EYM).
a pas in Pounds and Pence, the currency in the UK.

Table  :
Regulative stages, description and sample realisation.

Table  :
In-role and role suggestion stages in the shoe shop (SS-EYC).
a Emboldened text denotes 'in role' stages.

Table  :
In-role and role suggestion stages in the baby clinic (BC-EYC).

Table  :
The regulative and in-role contexts.