Multilingual communication under the radar: how multilingual children challenge the dominant monolingual discourse in a super-diverse, Early Years educational setting in England

The operant “one nation – one language”model in Western culture has resulted in linguistic hegemony being almost universally presented as an uncontentious reality. This article accepts Foucault’s challenge to deconstruct this officially sanctioned “truth” by looking at how the educational system in England legitimises the discourse that speaking English is normal, marginalising multilingual practices. Data is drawn from a year-long study of thirty “superdiverse” children in an inner-city school in the north of England. The research presents language portraits in which the children demonstrate they have internalised the rhetoric that English should be spoken in school. These are contrasted with ethnographic observations which provide co-constructed researcher-pupil cartoons representing social interactions in a range of contexts. The illustrations demonstrate how the children challenge the idealisation of English language in school by operating spatial agency in which they seek (and find) opportunities in peripheral and liminal spaces to speak their own language. ARTICLE HISTORY Received 7 May 2019 Accepted 25 October 2019


Introduction
A considerable body of literature identifies the growing presence of the monolingual "English is the language of England" ideology from a post-structural perspective (Pavlenko and Blackledge 2001) and analyses how this has filtered into the domain of schools (Leung and Scarino 2016). To date, however, there has been limited consideration of how children resist this dominant discourse by subverting the linguistic norms of a classroom and communicating "below the radar" in languages other than English.
The aim of this paper is, therefore, to explore ways in which young children, age four to five years, resist the discourse that places English in a position of superiority in relation to other languages in one early years classroom in England. To achieve this, the article first accepts Foucault's challenge to deconstruct officially sanctioned "truths" by examining how the educational system in England legitimises the discourse that speaking English is normal, and the extent to which this discourse marginalises multilingual practices. This is followed by an overview of studies that explore children's strategies to subvert dominant classroom discourses, through employing spatial agency. Vignettes from a longitudinal ethnographic study will then be presented that demonstrate children's resistance to this dominant discourse. Finally, the research presented in this paper extends our understanding of spatial agency by demonstrating how the children use their environment skilfully in order to establish and experiment with their individual identities through their language choices.

English as a dominant discourse in England
For Foucault, dominant discourses are particular conceptions of truth that are officially sanctioned and presented as objective, timeless, facts (Knight, Smith, and Sachs 1990). In Foucault's view of power, the State ensures people act in accordance with such dominant discourses not through punitive measures, but through a much more subtle dispersion of truths relating to health, self-fulfilment and normality, thereby entreating people to regulate themselves in order to be "happy, healthy and fulfilled" (Lawler 2014, 56). Foucault argues that those in power generate a set of truths that govern what is "normal" (and, thus, what is not) by categorising, listing and ranking aspects of the human condition such as development and ability. In this way, the preferred discourse is legitimised and alternative ways of being, acting and thinking are marginalised (MacNaughton 2005). Through the illusion of a rational, objective consensus as to what is and is not normal and desirable, people are persuaded to participate in their own subjugation without the need for external monitoring (Gallagher 2008).
Building on Foucault's ideas, the legitimisation (and suppression) of certain language practices can be viewed through a similar post-structural lens which focuses on how dominant languages are perceived to be "superior", with the concomitant assertion that all other language practices are "inferior" (Scott and Venegas 2017). Multilingualism may be promoted or constrained through language planning that aims to expand or limit linguistic diversity (Liddicoat and Taylor-Leech 2015); and the promotion of linguistic homogeneity in multi-diverse societies can be seen as a tool for encouraging assimilation under the guise of social cohesion (Vasta 2007). Dominant language ideology can be traced through law, policy and popular opinion (Blackledge 2005;Cooke and Simpson 2012). Currently, language policy and planning in the UK appear to support the "one nationone language" model which suggests that each nation state should use one language to identify and unite its people (Cooke and Simpson 2012). However, as Hornberger (2002) points out, it was the rise of nation states in the eighteenth and nineteenth centuries that led to the construction of this "myth" which was then exported globally though colonialism (Woolard and Schieffelin 1994). Similarly, Gramling (2016) details how monolingualism is a "recent historical invention" (28) which has become a dominant discourse in many Western countries.
Monolingualist ideologies govern current UK immigration rules, as demonstrated by the introduction in 2010 of a compulsory English language test for migrants coming to the UK to join or marry their partner. This piece of legislation was justified to "help promote integration, remove cultural barriers and protect public services" (Home Office 2010). This led to an unsuccesful high court challenge (R. (On the Application of Chapti, Ali and Bibi) v. Secretary of State for the Home Department 2011), sparking public debate over the importance of a shared language in England, fuelled by journalists such as Amanda Platell writing for the Daily Mail who asked "What about our right to a common language?" The influence of the press in their promotion of certain, most often the dominant, monolingualist linguistic ideologies should not be underestimated (Blackledge 2002;Lippi-Green 2012;Wright and Brookes 2019).
The consistent promotion of "English as the language of England" continued to be reinforced through the Casey Review (2016) which was undertaken at the request of David Cameron (the then Prime Minister) and Theresa May (the then Home Secretary) into integration and opportunity in isolated and deprived communities. Proficiency in English language was identified as a crucial factor in relation to both integration and economic success. Casey stated that "ensuring everyone is able to speak English enjoys strong public support" (Casey Review 2016, 94) drawing evidence from the 2014 census data on British Social Attitudes which reported that 95% of respondents think speaking English is important for being "truly British" (British Social Attitudes 2014, n.p.). However, the concept of a nation state being a homogenous society is at odds with the socially, culturally and linguistically heterogeneous reality in many parts of England, historically as well as contemporarily (Blackledge 2005).

English as a dominant discourse in the Early Years
Foucault believed that educational institutions are sites where certain knowledge and practices are legitimised in line with dominant discourses (Foucault 1972). Furthermore, he argued that institutional settings are the mechanisms that enable the promotion of certain "regimes of truth" (Foucault 2010, 18) that seek to suppress individuality in favour of uniformity (Ball 2013;MacNaughton 2005). Thus, in the domain of schools, we see a similar enforcement of particular 'regimes of truth' to that presented at a national level, where certain language practices and identities are privileged as social institutions which "hinge on the ideologization of language use" (Woolard and Schieffelin 1994, 56). Thus, in the Early Years Foundation Stage curriculum (birth to five years) (DfE 2017), children's competency in English is assessed through early learning goals which are indicative of the government's wider approach to ensuring that children are achieving the expected standardsan approach which reflects the neoliberal ideology that has dominated the education system since the 1990s (Leung and Scarino 2016).
The Early Years Foundation Stage (EYFS) is the statutory framework in England that provides the standards for learning, development and care for pre-school children. In Foucauldian terms, the EYFS operates as a Panopticon because it provides guidance on curriculum, pedagogy and assessment practices, and specifies the "good level of development" through the achievement of the Early Learning Goals (ELGs) (DfE 2014). The EYFS guidance states that providers have a duty to "fulfil the child's potential" (DfE 2014), and to ensure their security, safety, happiness and access to opportunities in the future. At the end of EYFS children are assessed against the criteria for the ELGs and judged to be meeting the level ("expected"), above the level ("exceeding"), or not quite reaching the level ("emerging") (Standards and Testing Agency 2014). While the principles and recommended practices are seemingly beneficial to children, the EYFS constructs a "typically developing child" through normative measures that must be met if a child is to "fulfil their potential" in order to be "secure, safe and happy". In doing so, the EYFS also establishes what is "good and bad, normal and non-normal" (Dahlberg, Moss, and Pence 2006, 38). Consequently, any child deemed "emerging" is, by definition, perceived to be at a deficit when compared to their peers (Volk and Long 2005), regardless of any relevant social, linguistic and cultural diversities.
With regards to assessment, non-statutory advice in the EYFS profile handbook (Standards and Testing Agency 2019) states that all areas (bar English) may be assessed in the child's home language, yet it is unclear how it is possible to operationalise this given the wide variety of languages and dialects spoken by studentsfor example, the thirty children in the study reported here spoke fourteen different languages (in addition to English). Importantly, however, the statutory framework for the EYFS (DfE 2017) itself makes no such suggestion. Instead, it advises that practitioners should take "reasonable steps to provide opportunities for children to develop and use their home language" (DfE 2014, 9). Importantly, however, the document also places significant emphasis on the providers' duty to: ensure that children have sufficient opportunities to learn and reach a good standard of English language during the EYFS: ensuring children are ready to benefit from the opportunities available to them when they begin Year 1 (DfE 2014, 9) Once again, the policy rhetoric mirrors the Foucauldian perspective of power: if a particular individual matches the "typically developing child" in terms of English language proficiency, then this individual will be "school ready", and thereby able to reap further benefits in Year 1 of the National Curriculum (the first stage of compulsory education in England). Thus the EYFS handbook's recognition of home languages would appear to have limited value unless it is carried through to the ELGs which, at present, it is not.
The situation is further complicated by a lack of continuity between the EYFS and Key Stage One as children move from Reception into Year One. Here, assessment across the curriculum is conducted in English, and the children's ability to "take part" in the national curriculum hinges on their communication skills in English (DfE 2014). This means that any students who were able to benefit from being assessed in their home languages in Reception will no longer be in this situation once they reach Year One. Indeed, the curriculum available for four and five-year-olds who speak English as an additional language is incoherent, and the lack of clear guidance was noted in the 2016 Rochford Review of assessment for pupils working below the standard of national curriculum tests. This Review recommended that "additional advice or guidance in helping teachers to make assessments accurately or effectively" should be developed (Rochford 2016, 27).
It could be argued that the sidelining of home languages in favour of the promotion of English was originally based on language acquisition logic whereby immersion in (for example) English and the prohibition of other languages was assumed to hasten acquisition in the target language (Nieto 1999;Krashen 1987). This was coupled with a fear that the child's "mother tongue" first language could interfere with their learning of English (Spada 2015;Macaro 1997). The lack of support towards developing a child's first language is also underpinned by the assumption that English is more valuable than minority languages (Al-Azami 2014; Asker and Martin-Jones 2013).
However, more recent research recognises that promoting a balanced approach to bilingualism is linguistically and cognitively beneficial to children (Conteh 2012;García 2009;Leung and Scarino 2016). Thus, Costley (2014) argues that the intense focus on the teaching and learning of English in schools is not about hastening language acquisition but rather an opportunity to mould society by promoting a sense of national identity and pride.
In summary, and notwithstanding the apparent confusion over the status of EAL (Costley 2014;Lamb 2001), the EYFS (and more recently the school census) does provide schools and practitioners with the tools and motivation to assess young children's proficiency in English. However, from a post-structural perspective, such a process of dividing children through assessment and classification is intended to distribute, manipulate and control children according to culturally and historically constructed, normative judgements about a child and "childhood". It is argued that such an approach is problematic as it decontextualises the child and risks losing sight of children and their lives: "their concrete experiences, their actual capabilities, their theories, feelings and hopes" (Dahlberg, Moss, and Pence 2006, 36).

Space, discipline and resistance
Foucault states: "Where there is power, there is resistance" (Foucault 1998, 95) and this has been demonstrated by numerous studies conducted in educational environments where children choose to reject the behavioural norms expected of them within the institution (Forslund Frykedal and Samuelsson 2015; Markström 2010; Van de Kleut and White 2010;McFadden 1995;Giroux 1983). Resistance comes in many forms and can take on a subtle, indirect guise, where children show resistance by testing the ways in which they can bend and stretch the expected social order without explicitly breaking the rules of the institution (Markström 2011). When glancing at a classroom it may appear that children are generally compliant but, by looking beyond the "superficial cooperation", it is possible to uncover acts of resistance, subversion and subterfuge (Halstead andJiamei 2009, 2266). Ethnographic studies are particularly insightful, evidencing how young children create their own cultures and use multiple strategies to assert agency, for example, through sociodramatic play (Corsaro 1993), through silence (Markström and Halldén 2009), or through negotiation (Danby and Baker 1998).
It is common for young children's educational environments, such as the site of the data collection for the research described in this paper, to exercise constant Panopticonlike surveillance which serves to control the children's behaviour (Foucault 1991;Gallagher 2010). Foucault was particularly interested in the importance of space, 1 suggesting: "A whole history remains to be written of spaceswhich would at the same time be the history of powers [both these terms in the plural]from the great strategies of geopolitics to the little tactics of the habitat, institutional architecture from the classroom to the design of hospitals, passing via political and economic installations" (Foucault and Gordon 1980, 149). Furthermore, it is argued that the physical organisation of a particular setting or settings influences the conduct of children in those spaces through a process of governmentality: the children understand what are acceptable ways of behaving in a particular space and govern themselves and others in accordance with these normalised routines (Pike 2008).
A key theme that is frequently found within the literature on children's agency and resistance to discipline is their use of space. Thus, throughout the school day, children take part in a range of activities, occupying different spaces within the classroom or the wider setting. From the perspective of Foucault, each of these spaces is the location of a power struggle in that the setting is designed and used by adults in order to increase their ability to monitor and control children's conduct. However, at the same time, the children operate spatial agency by exploiting the peripheries of such monitored spaces in order to avoid surveillance. Pike (2008Pike ( , 2010 adopts an ethnographic approach to examining the socio-spatial interactions occurring in dining halls. Her research demonstrates how the layout of dining rooms inhibited social interaction and facilitated surveillance as the children were encouraged to conform to specific conventions (2008). The children in the study evaded the gaze of the lunchtime staff by occupying the peripheries of the dining hall as they were the least visible parts of the room and, in some cases, utilised gymnasium or theatrical equipment found there as cover or to create a diversion (2010). In parallel, an ethnographic study of outdoor spaces in early childhood education and care settings in Ireland noted that indoor spaces were seen as confining and restrictive, while outdoor spaces were associated with freedom (Kernan and Devine 2010). An in-depth investigation of social practices in a playground by Thomson (2005) revealed how teachers enforced control over children's movement by allowing and denying them access to certain areas of the playground. The children in the study challenged these restrictions by employing strategies such as hiding around corners and acting as look-outs for one another in order to avoid the supervisor's gaze. These studies all demonstrate how space has been shaped to assist the projection of discipline, yet at the same time children beat the adults at their own game by utilising space in ways that enable them to avoid being monitored.

Methodology
With this summary of the current policy framework of English language ideology in England and within schools in mind, the following section will discuss how this ideology plays out in reality. The data are drawn from the lead author's doctoral thesis, which took the form of observations from a twelve-month ethnographic research study of 30 multilingual children, age four to five years, in a Reception class of an inner city primary school in a large city in the North of England. All children in the class were invited to participate and parental consent was gained from all but two of the children's parents; thus the two children whose parents did not wish them to participate were not involved in the research. The researcher worked in the school where the research took place as a teacher in year one for three years and became interested in the children's multilingual communicative practices because of her own enthusiasm for languages. The selected observations exemplify how the dominant "English" discourse impacts the experiences and conduct of young children.
The children were followed through their transition from Reception class (the last year of the EYFS) into Year 1 of the National Curriculum. In addition to English, the children spoke fourteen 'home languages' between them, albeit with varying degrees of fluency. The setting can be categorised as "super-diverse" (Vertovec 2007) because the children came from a multitude of backgrounds in terms of geographical location, reasons for being in England, channels of immigration, transnational links and their experiences, including educational experiences, before joining the school.
Unsurprisingly, in this super-diverse setting the children displayed a wealth of multimodal communicative practices reflecting not only their varying proficiency in communicating, speaking, reading and writing in English, but also their different languages, faiths, cultures and identities.
The research was guided by the following questions: How do the repertoires children learn in out-of-school socio-cultural contexts contribute to children's multimodal communicative practices?
What is the relationship between the immediate contexts of communication and the resources children draw upon to communicate?
The ethnographic approach included visual methods, children's participation and research conversations between the children and the researcher. Written informed consent was gained from children's parents and oral consent was initially gained from the children. The researcher was also mindful of ongoing "ethics in practice" (Guillemin and Gillam 2004, 262) as research presents day-to-day ethical issues, such as a child suddenly appearing uncomfortable when being observed. The researcher accepted that the children's assent needed to be continually sought, thereby respecting children's wishes and opinions throughout the research process (UN 1989, Article 12). It will be appreciated that these methods were ethically sensitive and responsive to the diversities within this setting. For example, because the children and the researcher (the lead author of this paper) did not share a common language (other than English), there was considerable reliance on gestures, visual images, artefacts and other multi-modal forms of communication to help ensure mutual understanding and respect. In order to help overcome the resulting challenges in understanding the messages that the children were imparting, a key part of the research process involved converting observational and conversational data into cartoon strips which were co-constructed with the children who chose to draw and use their own self-portraits. The researcher drew sketches of the written observations and shared these with the participants. The children then drew self-portraits to incorporate into the sketches and commented on the content of the comics (for a detailed description of the method used, see Fashanu 2017). Importantly, the use of cartoon strips opened up spaces for dialogue between the researcher and the children around the observations, thereby integrating and valuing their understanding and perspectives (Brooker 2011). The children chose their own names for the researcher to use as pseudonyms in order to protect their identities.
An inductive approach to data analysis, based on constructivist grounded theory (Charmaz 2006), was adopted. The iterative process of immersion in the data, memowriting and coding resulted in the creation of salient theoretical categories, or themes, that guided a re-examination of the data. The lead author's doctoral thesis adopts Rogoff's three planes of analysispersonal, interpersonal and cultural/institutionalto analyse activities (Rogoff 2003). This article draws on a sub-set of the data which contributed to the theme of space within the cultural/institutional plane of analysis: a category that explores how the social construction of space actively contributes to the interactions that occur within that space. Within the broad theme of space, the particular focus in this article is on the relationship between space and the choices children make around spoken language. In order to create the sub-set of data that forms the subject of this paper, the main database was reviewed and examples where the students displayed their understanding of space and how it could be manipulated to their advantage were extracted.

Language portraits
During the first few months of the data collection phase, the researcher became aware of the disparities between children's communicative practices at home and at school. In order to investigate this further, the children were asked to complete "portraits" of their school and home languages. In this task, they were given a preprinted sheet with two body silhouettes and asked to colour these in to represent the languages they spoke at home and at school respectively (Busch 2012(Busch , 2018. They could select the colour they wished to use to indicate each of the languages and there were no rules as to how the children should go about colouring in the silhouette. The children developed the language portraits in small groups of two to four at a time and were simultaneously invited to comment on their language portraits as they coloured them in. This is because analysis of the language portraits can be greatly strengthened by an accompanying biographical commentary that can reveal how the children experience their multilingual repertoires (Busch 2012;Wolf 2014). 29 out of the 30 participants in the study participated in creating language portraits, the one non-participant having moved to another school shortly after the research began. The results of this analysis are summarised in Table 1.
Data gathered from the language portraits task (Table 1) shows that 21 out of the 29 children claimed they spoke English. Out of these children, four said they only spoke English at home and so it is not surprising that they only spoke English at school. What is more interesting is that 17 out of the 29 children who created the language portraits spoke one or more languages other than English at home, and yet they only spoke English at school (see the examples in Figures 1 and 2).
The remaining eight children claimed they spoke both English and another language at school. Of these, one child, Rocky, had recently arrived in the country and, because he did not speak much English, he mixed English and his home language, Oromo (from the North East African region), at school out of necessity. For example, on one occasion Rocky had a sore throat and repeatedly tried to communicate this to the teacher by saying "I have . . . " followed by a word in Oromo. The teacher asked for a student in year four to translate and it became clear that Rocky was attempting to explain he had a sore throat, but lacked the vocabulary in English to express this. The remaining seven children who said they spoke mixed languages both at home and at school offer valuable insights into the use of home languages at school. As previously explained, the children were encouraged to comment on their perspectives of their communicative practices whilst completing their language tasksnot least as this has been demonstrated as strengthening the conclusions drawn by the researcher (Busch 2012). Thus, while colouring in the "school languages" template, all seven clearly explained who they spoke the language other than English to, and in what context. The resultant breakdown is summarised in Table 2: Home languages School languages   Of the seven children who said they spoke English and another language at home and in school: Two Roma children, Darth Vadar and Igor, stated they spoke Roma with each other, but only outside or when the Roma teaching assistant was with them Two Somali children, Asad and Anna, stated that they spoke Somali with each other, but it was also made clear that this only occurred during play time or "Golden Time" Three Arabic speaking children, Aladdin, Afaq and Issa, identified speaking Arabic to each other, but only in the playground and at lunch time, never in the classroom As they explained these "rules", they made it clear to the researcher that their use of home languages at school reflected the particular conditions when they considered it permissible to communicate in a language other than English, for example during free choice or "Golden Time" or in the playground at lunch time. In addition, it appeared that the children believed it to be permissible to speak home languages in the presence of authoritative adults, such as teaching assistants, who shared their language. To emphasise this point, it will be seen that the children's discussions around the parameters of language practices (such as who they spoke with, and in which context) were supported by their careful designation of the amount of the "other language" that they coloured in. Thus, in each of the "school languages" templates where the children claimed to speak more than one language at school, they coloured the majority of their template in the colour designated to English, such as in the example in Figure 3.

Home languages School languages
The children's careful allocation of a tiny proportion of the "school languages" template to other languages, in conjunction with their expression of the self-imposed rules that governed when, where and with whom it was appropriate to speak languages other than English, demonstrate how the children sought to self-regulate their language practices.

Vignettes
Throughout the year-long ethnographic study the researcher made detailed observations of the children's communicative practices. This approach was adopted after an extensive review of the literature that affirmed the potential for rich insights to be revealed by ethnographic studies in early childhood (Konstantoni and Kustatscher 2016;James 2007;Qvortrup 2000); in educational environments (Mukherji and Albon 2010;Siraj-Blatchford 2010;Levinson 2005), in relation to children's resistance (Pike 2010(Pike , 2008Markström and Halldén 2009;Danby and Baker 1998;Corsaro 1993), and in documenting children's communicative practices (Rampton and Charalambous 2016;Martínez-Roldán 2015;Pavlenko and Blackledge 2001;Hornberger 1995).
The field notes were then translated into visual sketches that the children commented on and contributed to in a process that resulted in the co-creation of cartoon strips that depicted vignettes from the observations. Amongst other aspects, an analysis of these  vignettes reveals that, contrary to the children's assertions about their language practices in school through their language portraits, in reality linguistic diversity occurred frequently, albeit "under the radar".

Language choices inside and outside the classroom
The following vignette (Figure 4) shows Darth Vader introducing Igor to the classroom areas. Both children are Roma and lived in Slovakia before coming to England. Darth Vader moved to England two years before, has three older siblings who attend the same school, and speaks English well. Igor has just arrived from Slovakia and does not know much English at all. The teacher asks Darth Vader to be Igor's buddy and show Igor around the classroom during "choosing time". In this vignette, Darth Vader takes his role of being Igor's buddy very seriously. He puts his arm around Igor's shoulder and proceeds to walk around the classroom explaining in English what each of the areas is, what Igor can do in each area, what the rules are for each area and so on. As it is during "choosing time" the other children are engaged in talk and activities all over the classroom, which masks their conversation; and yet Darth Vader continues to conduct his guided tour in English. Soon after it is playtime and Darth Vader, still in the role of buddy, leads Igor to the classroom door. As they step over the threshold into the playground, Darth Vader's language simultaneously converts from English to Romani. He continues to speak to Igor in the same tone of voice, with the same gestures and the same body language, but Darth Vader completes the tour in the language he shares with Igor. The classroom door is a physical threshold to the outdoors, but also represents a metaphorical threshold regarding language choices. Observations conducted outside frequently yielded examples of children conversing in their home languages. This phenomenon was not restricted to the children as parents also spoke with their children in their home languages in the morning while waiting for the classroom door to open, yet as they entered the building, they switched to English.
In this vignette Darth Vader is showing Igor around the setting and simultaneously imparting a knowledge and understanding about accepted language practices in each of the spaces. In doing so, he is unwittingly complicit in the power structures created by wider political ideologies, namely that other languages are not acceptable inside the classroom. The vignette illuminates firstly how Darth Vader is participating in his own  subjugation (Gallagher 2008) by choosing to speak English inside the classroom and, secondly, how regulatory power is dispersed through networks rather than exercised through punitive measures (Lawler 2014). The structured environment inside the classroom has a regulatory effect on Darth Vader's behaviour, while the outdoor space is associated with freedom (Kernan and Devine 2010). That Darth Vader chooses to speak English in the classroom and Romani in the outdoor space is significant as it highlights the constraining effect of the classroom, and the need for free time and free space where children can discover and explore their own identities (Casey 2007).

Hiding in the reading corner
The following vignette ( Figure 5) shows two boys, Ali and Issa, playing in the reading corner. Both boys are refugees originally from Iraq, although Ali lived in Poland for a number of years before coming to live in the north of England. Although both Ali and Issa are new to English, they can communicate basic day-to-day ideas in social contexts. The third child in the vignette is Jason, a boy whose parents are from Kenya and the Democratic Republic of the Congo, but who was born and raised in the city in which the study took place.
In this vignette Ali and Issa are lying down in the reading corner, surrounded by the bookshelves and out of sight of the other children and the teachers. They are giggling and talking to each other in Arabic. Suddenly Jason wanders into the reading corner and begins to look for a book. Jason is not interacting with Ali or Issa directly; however, his presence in the reading corner has an indirect effect on the two boys, who begin to speak to each other in English.
In the language portraits both Ali and Issa stated they spoke Arabic and English at home. However, Ali said he only spoke English at school and Issa said that, although he did speak Arabic at school, it was only with two other children (Aladdin and Afaq), and only in the playground at lunch time. Despite these claims, this vignette offers evidence to the contrary as they are speaking Arabic inside the classroom. Ali and Issa take shelter from the teacher's gaze by lying on the ground inside the reading corner, taking advantage of the book shelf that offers them protection. When Jason enters the space to choose a book he appears quite oblivious to their presence and he does not interact directly with Ali or Issa, yet his physical proximity is enough to cause Ali and Issa to switch into English. In doing so, their concealed use of a language other than English goes undetected. Interestingly, this vignette demonstrates that Ali and Issa are not only hiding from their teacher, but they are also sensitive to the approach of other children. This is substantiated by similar situations that occurred throughout the data collection phase when children spoke in a language other than English and their peers reprimanded them for doing so.
As indicated earlier, whilst the classroom resembles a Panopticon (Foucault 1991;Gallagher 2010), Ali and Issa operated spatial agency by finding a place that is hidden from surveillance, and in this place they resist the imperative that they should speak English in school. In a similar way to the children Pike (2010) observed in the dinner hall, Ali and Issa seek the least visible part of the room and use the apparatus, in this case the book shelf, as a cover. Jason's entrance into their hiding place caused Ali and Issa to regulate their conduct by suddenly speaking English. They were not speaking English for the benefit of Jason, as he made no attempt to interact with them and was simply entering the space in order to choose a book. However, the impact of Jason's presence can be likened to the effect of the Panopticon's surveillance system shining a light on Ali and Issa, resulting in their self-regulation in order to conform to language practices that are considered "normal" in the classroom.

Lining up
The following vignette ( Figure 6) shows two children, Naan and Cinderella, lining up for lunch. Both are from Pakistani families, but they have different backgrounds. Naan moved to England two years before with his mother and younger siblings. His mother does not speak English and Naan learnt English only when he began school. Cinderella was born in England and lives with her parents, who were also born here, and her grandmother, who does not speak English. When asked, both Naan and Cinderella claim to speak Urdu. It is appreciated that diglossia operates in most of Pakistan, where Urdu is the language of school and official tasks, yet families often speak other languages in the home (Ilahi 2013). Therefore, while Naan and Cinderella say they speak Urdu, there is a possibility that in reality they speak different language varieties rather than Urdu, which is typically reserved for formal interactions.
Prior to this vignette, all the children were sitting on the carpet in silence. The teacher asks Naan and Cinderella to get their things and form a line by the door ready for lunch. The other children are asked to join them one by one, which prompts a commotion as the children look for their packed lunches, coats and various other items, much of which requires the teacher's attention, distracting her from her efforts at lining the children up in a quiet, orderly fashion. Naan and Cinderella sense an opportunity to speak in Urdu amongst the hustle and bustle of lining up. Naan calls Cinderella "ganda", to which Cinderella gasps in shock and exclaims "what!?". Naan then clarifies the meaning of "ganda" in English for Cinderella, stating "it means naughty!", but Cinderella corrects him: "No, it means dirty", and Naan agrees: "Oh yeah, khuti means naughty girl". As the hubbub quietens down and the rest of the children are ready for lunch, Naan and Cinderella end their discussion and stand in silence with the other children.
This vignette may be analysed in schematic terms: power relations are produced and reproduced in the classroom through structured activities such as sitting on the carpet and lining up, which can be seen as forms of social control (Bernstein, Elvin, and Peters 1966) and means to increase surveillance (Foucault 1991). Transitions between activities often triggered acts of resistance, such as exchanges between children in languages other than English. However, if we probe deeper, this vignette reveals the complex nature of power relations as a process of continual flux and negotiation between all involved (Flohr 2016). Markström and Halldén (2009) describe how activities rarely fall into the binary categories of conformity-resistance and conclude that pre-school institutions are arenas for children to explore and experiment with power relations by interpreting situations and opportunities to defend personal autonomy. Here, resistance can be seen as a form of "counterpower" (Foucault 1977, 219), rather than a binary opposite to power.
Naan and Cinderella demonstrate they are not just the objects of surveillance; they are vigilant of the teacher's attention and take advantage of the lapse in concentration to challenge the dominant discursive practice of the classroom. Here it is possible to see the process of surveillance is not unidirectional. Indeed, the children were continually monitoring their teacher and skilfully timing their acts of resistance to avoid detection (Halstead and Jiamei 2009). Furthermore, the children employ strategies to distance themselves from the teacher-dominated classroom discourse, while simultaneously fitting in with the institutional power structures that exist (Gutierrez, Rymes, and Larson 1995). The children's activity was not intended to disrupt the status quo or radically alter existing institutional structures; Naan and Cinderella calmly comply with the teacher's request for them to line up. However, despite their physical compliance, the children assert their difference from the role expected of them by speaking to each other in Urdu. This vignette has an added layer of rebellion as the words the Naan uses are actually mild swear words in Urdu. Swear words are defined and sanctioned by institutions, such as schools (Jay 2009); thus the use of taboo words in the classroom setting would be considered highly inappropriate and carries extra weight in terms of defiance and insubordination. Children in this study often explored the boundaries of appropriate social identities by engaging in forbidden activities such as play-fighting, kissing each other or re-appropriating learning materials for their own purposes, for example using the colourful link chains from the maths area as handcuffs. The presence of recurrent acts of subversion highlights how children frequently sought opportunities to express agency within an adult-controlled context (Wood 2014) and how resistance through agency is a counterpower in itself.

Conclusion
This article draws attention to the extent to which the dominant discourse that "speaking English is considered to be superior to all other language practices" is visible from the EYFS onwards. This discourse is apparent at a national level in law, education policy, the media and popular opinion. Language policy in schools is similarly focused on English as the sole priority, as articulated through the statutory curriculum. However, the evidence presented in this study demonstrates there is a clear tension between the nation state's goal of being a homogenous society, and the reality of social, cultural and linguistic heterogeneity (Grillo 2005), particularly in super-diverse communities (Vertovec 2007).
Data drawn from the lead author's doctoral thesis demonstrates how children are aware of the dominant discourses around acceptable language practices in the classroom, as evidenced by the language portraits, the accompanying commentary where children explain rules surrounding when it is and is not permissible to speak languages other than English, and their concealment of clandestine language practices in the classroom. It is clear that the children self-regulate their linguistic heterogeneity in favour of conforming to the idealised "English-speaking" model that is set out by the EYFS, in keeping with dominant discourse and "regimes of truth" (Foucault 2010;Knight, Smith, and Sachs 1990).
Throughout the research it also became evident that space in which children's interactions occurred is not just a backdrop; it is "intimately tied" to the events that unfolded (Warf and Arias 2008, 4). The most "free" space that children occupy in the school day is the playground and here the children believe it to be acceptable to express themselves in whatever language they choose. The classroom, however, was a space in which English was the only acceptable language and any communication conducted in other languages had to be camouflaged. The children in this study are skilled spatial agents, working with the physical space to seek when and where surveillance is at its weakest. In these pockets of space at the peripheries of the classroom structure and organisation, the children in this study take advantage of blind spots in order to break the rules and regulations that typically govern the space (i.e. that English ought to be spoken when at school). Speaking in languages other than English thereby becomes a clandestine activity, relegated to the borderlands, hidden from the teacher's gazebut one that the children undertake on a daily basis, even though they know it is "wrong". Importantly, this practice reveals how children actively choose to experiment with power relations (Markström and Halldén 2009), challenging the dominant discourse that exists in the school setting regarding legitimised and forbidden language practices.
It is recognised that this research could be extended beyond the binary contexts of home and school to explore other environments that children occupy. For example, the children could be asked to identify different spaces they occupy within school and outside school, and further investigation through language portraits and ethnographic observations could be conducted into their communicative practices in each environment.
The evidence presented here extends the findings of previous studies that show how children employ spatial agency to challenge restrictions on their physical movement (Thomson 2005) and behaviour (Pike 2008(Pike , 2010 in educational settings. In this study children exercise agency to express their individual identities through language choices in an environment that seeks to reduce linguistic heterogeneity in favour of a more homogenous, "English speaking" model. Note 1. The "spatial turn" refers to an academic movement that gained momentum in the last half of the twentieth century emphasising that "space" is not just "a backdrop against which life unfolds sequentially, but rather, is intimately tied to lived experience" (Warf and Arias 2008, 4). Foucault was among the theorists, such as Henri Lefebvre, Gaston Bachelard, Michel de Certeau, David Harvey, Saskia Sassen, Edward Soja and Iris Young who reinvigorated inquiry into the importance of space.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the Economic and Social Research Council [140218037].

Notes on contributors
Christina Fashanu is an ESRC funded PhD student at the University of Sheffield. She is currently researching the communicative practices of young children in super-diverse, educational environments. She has a particular interest in developing participatory methods to assist research with young children who speak English as an additional language.
Elizabeth Wood is Professor of Education, specialising in early childhood education. Her research focuses on play and pedagogy, critical policy analysis, and teachers' and children's use of digital play in education settings.
Mark Payne is a Lecturer in Language Education. His research focuses on language policy and planning, applied linguistics and multilingual and multicultural education more broadly. He is currently focusing on the educational engagement and achievement of the Slovak Roma community in Sheffield.