Where two worlds meet: language policing in mainstream and complementary schools in England

We compare language policing in two educational contexts in England: mainstream schools and complementary schools. We draw on a varied dataset (policy documents, in-class observations, interviews) collected from mainstream schools and Greek complementary schools in London. We find similarities in how the two types of schools control, regulate, monitor and suppress the language of school students. Both settings hierarchise standardised and non-standardised varieties in institutional policies that delegitimise the non-standardised varieties. Teachers become vehicles for language ideologies in enacting monovarietal policies drawing on discourses around academic success and the primacy of written language over spoken language, including regional varieties such as Cypriot Greek. Our findings suggest that multilingual and multidialectal students in England who attend both mainstream and complementary schools are exposed to similar kinds of prescriptive discourses across the whole spectrum of their educational experiences, which can have a range of negative effects on their learning and the construction of their self-image. We argue that more links need to be forged between the two educational settings and that these should include the development of integrated pedagogies and policies that legitimise students’ whole linguistic repertoires, encompassing both their standardised and their non-standardised varieties as well as their other linguistic resources. ARTICLE HISTORY Received 6 May 2021 Accepted 7 May 2021


Introduction
This article presents a critical exploration of language policing in mainstream and complementary schools in England. By language policing, we point to the various mechanisms within policies and pedagogies whereby students and teachers have their language controlled, regulated, monitored and suppressed on the basis of language ideologies, especially standard language ideologies. These mechanisms occur within education systems characterised by struggles related to power imbalances, standards and performativity, with the policing metaphor playing out in terms of rules, laws, crimes and punishments.
We consider language policing across policy spaces and layers, from macro-level policy mechanisms through to micro-level classroom interactions (see Amir and Musk 2013), in an attempt to interrogate policing in terms of individual events and broader socio-political structures. Mainstream schooling refers to the typical education experience of 4-16-year olds in England, with students required to attend every working day between roughly the hours of 08:30-15:30 following a timetable which compartmentalises their educational experience into discrete subjects. Primary schools cater for students between the ages of 4 and 11; secondary schools cater for students between 11 and 16 years of age. Complementary schooling refers to the after-hours education that migrant communities organise and offer to their younger members primarily with the aim of teaching the respective community's home(land) language (Blackledge and Creese 2010;İssa and Williams 2009;Li 2006;Lytra and Martin 2010). Education covers a broad range of levels, starting from preschool and reaching preparation for formal qualifications such as A-Levels, taken at age 18. Teaching is provided in an integrated way to all levels of education, meaning that there is no distinction between primary complementary and secondary complementary schools. Teaching normally takes place during the weekend and/or on weekday evenings. In this article, we focus on Greek complementary schools, which teach Greek language and culture to children of England's Greek Cypriot and Greek communities.
Although recent work has looked separately at language policing in mainstream schools (Cushing 2020a) and complementary schools (Karatsareas 2020) in England, we offer an original contribution in the way that we explore the contact points, overlaps and contrasts between language policing in these two settings, which have historically been treated as separate entities within research in educational linguistics and education policy. Within the context of this special issue, this article considers the language ideological processes at work across these settings, with a particular focus on the enshrinement of standard and monoglossic ideologies, which work to further stigmatise non-standardised forms. We have structured the article as follows: we first present the context and settings of our work. We give a brief overview of multilingualism among students in England and discuss recent developments in English mainstream and complementary schools focusing on approaches towards standardised and non-standardised forms of language in the two settings. We then outline the theoretical underpinnings of our study, elaborating on our conceptualisations of language policy, language ideologies and linguistic repertoires. We move on to describe our methods for data collection and analysis. We present our findings on language policing in mainstream and complementary schools, and compare the two contexts before providing our concluding remarks.

Multilingualism among students in England
In 2019/2020, almost 1.7 million students in England were reported to have English as an additional language (EAL), with the overwhelming majority found in state-funded primary schools (Table 1).
So-called EAL students are a highly diverse group in terms of their distribution in the country and in school units; their biographies, including their experiences of mobility; and their linguistic, sociocultural and economic characteristics. In Malmberg and Hall (2015), London as well as Slough, Luton and Leicester were reported as having the highest overall percentages of EAL students. In over half of the schools, less than 5% of the students were classified as EAL, whereas their proportion is under 1% in about a quarter of schools. In one in 12 schools, however, the majority of the students were EAL. Table 1. Headcounts and percentages of UK students who are exposed to a language at home that is known or believed to be other than English (Office for National Statistics 2020). Given the broad definition of the Department for Education ('A student is recorded to have English as an additional language if they are exposed to a language at home that is known or believed to be other than English.'), the group encompasses students who were born in the UK or abroad; come from a diverse range of national, ethnic and cultural backgrounds; are in the UK as a result of different motivations; display varying competences in English and in their other languages and varieties; and, generally have varied experiences of life (Sharples 2020; The Bell Foundation 2019).

Mainstream schools
Teachers and students in mainstream schools in England operate under a dense web of policies, with their work pushed and pulled in various directions by internal and external pressures. These pressures have constructed what Ball, Maguire, and Braun (2012) conceptualise as the 'standards agenda' and cultures of 'performativity', whereby schools are places concerned with high-stakes assessments and performance management under the guise of 'raising standards'. Policy mechanisms and technologies such as the Office for Standards in Education, Children's Services and Skills (Ofsted), Standard Assessment Testing (SATs) and General Certificate of Secondary Education (GCSEs), together with national and international league tables, have created a policy chain whereby schools are under pressure to perform, deliver, and improve. Although issues of standards in language have long formed part of this policy narrative, of particular importance to this article are the standardsbased reforms of post-2010, introduced by the Conservative-Liberal Democrat coalition and then from 2015 onwards, under the majority Conservative government. These reforms, which form the immediate policy context for the discussion of data generated from mainstream schools in this article, were framed using typically conservative keywords such as 'core knowledge', 'order', 'morals', 'discipline', 'rigour', 'standards' and 'tradition', cutting across changes to curricula, assessments, teacher regulations, preferred pedagogies and inspection systems. This bricolage of policy mechanisms works to propagate and concretise language ideologies, notably standard language ideologies, whereby general 'standards' and 'performance' in the classroom become wrapped up with so-called standards in language: speaking and writing 'correctly', 'properly' and 'articulately' (see Cameron 2012). Under this monoglossic ideology, teachers are granted agency to exercise remedial and racialised notions of linguistic correctness, with one consequence being the stigmatising and shaming of non-standardised linguistic repertoires. A detailed discussion of how these policy mechanisms propagate, reinforce and institutionally embed the standard language ideology in mainstream schools can be found in Cushing (2021), but include changes to the national curriculum, state-issued grammar tests, teacher performance measurements, writing assessment frameworks, traditional grammar glossaries and Ofsted inspection methodologies. Throughout these mechanisms, spoken-written grammar is conflated and standardised English is valorised, being discursively constructed as 'correct' English and equated with notions of 'high standards', 'clarity', and 'achievement'. We return to a discussion of these mechanisms in the sections that follow, drawing contact points between standard language discourses in policy, practice and pedagogy. It is important to note that debate and discussion about standardised and non-standardised varieties in schools has occupied a central place across decades of research in educational linguistics (e.g. Crowley 2003). However, these issues have attracted renewed attention in recent years, prompted in particular by a number of schools who have implemented so called 'zero-tolerance' and 'no-excuses' policies which are underpinned by strict systems of control, discipline, intimidation and incarceration (see Kulz 2017). Within the logics of these punitive policies lie attempts to 'ban' and erase regional dialect forms and typical fillers/discourse markers of spontaneous speech, with teachers often making intertextual and interdiscursive references to stateissued mechanisms, with pro-/prescriptive and policing practices often justified on the grounds that school-level policies are subordinated to national incentives and high-stakes assessments (see Cushing 2020aCushing , 2020b. As discussed earlier, although teachers do have power and agency within the policy making process, they are nevertheless bound to a system which can coerce and intimidate them into practices of punitive language surveillance. Language policy making is thus a 'contested space', which is shaped and moulded by internal and external pressures and modes of compliance/resistance.

Complementary schools
Despite the extent of multilingualism not only among students but also among teachers and other staff, mainstream schools in the UK remain largely monolingual and unicultural educational institutions. While teachers are aware of linguistic and cultural diversity, students' languages are not frequently used or even referred to during teaching (Bailey and Marsden 2017), even when teachers' and students' repertoires have (a) language(s) in common (Safford and Kelly 2010; see also Conteh 2018). Schools have also been found to discourage or even prohibit the use of languages other than English in class even in cases where local school policy and guidance stemming from various governmental sources (Department of Education; Department for Children, Schools and Families;Ofsted) would endorse it (Cunningham 2019). English-only policies and practices, which were only reinforced by the stipulation of the Swann Report (1985) that minority ethnic communities should provide teaching in the students' 'mother tongues', mean that mainstream education fails to meet the needs of multilingual students in terms of not only fostering the continuous use of their languages and their maintenance within their communities but also incorporating them in pedagogy and utilising them as resources for learning (Li 2006). Complementary schools are the response of minority communities to this shortcoming (Blackledge and Creese 2010;İssa and Williams 2009;Lytra and Martin 2010). They are also known as supplementary schools, heritage language schools, Saturday schools and community language schools. Each term reflects differing perspectives on the relation between these schools and mainstream schools as well as on the position that the language (s) taught in the schools occupy in the students' linguistic repertoires and their lives more generally (for a terminological discussion, see Ganassin (2020, 4-5) and references therein). We adopt the term 'complementary' as it does not relegate schools to an inferior or secondary role with respect to mainstream schools but suggests that they contribute to the education of their students in a holistic way.
It is estimated that between 3000 and 5000 schools operate currently in England (Evans and Gillan-Thomas 2015). It is, however, difficult to establish how many students attend complementary schools. Together with community associations and community media, schools are among the 'three pillars of diaspora' (Li 2016(Li , 2018Li and Zhu 2013), that is, spaces of community socialisation and networking where 'community identity is preserved, defended, renegotiated and reconstructed in light of discourses circulating within the wider society' (Simon 2018, 4). Although many subjects are taught in complementary schools, including offering support for mainstream school needs, the formal teaching of the language typically associated with the students' ethnic background takes centre stage in their mission, aims and activities. Language is seen as the most fundamental element in constructing the ethnocultural identity of minority ethnic students and in countering the effects of intergenerational shift to English in younger and especially British-born generations, which is often perceived as a danger or threat (Çavuşoğlu 2010).
Complementary schools are spaces in which students can express, explore and celebrate their multilingual and multicultural identities, drawing on their linguistic repertoires. Schools vary with respect to how they manage students' multilingualism both during and outside lesson time. Two opposing, but not mutually exclusive, tendencies have been identified: in separate bilingualism approaches, students' languages are compartmentalised and each language is associated with distinct and non-overlapping aspects of students' identities and ethnocultural backgrounds; in flexible bilingualism approaches, the full range of students' languages and varieties is drawn on in processes of meaning making and identity construction and performance (Blackledge and Creese 2010;Creese and Blackledge 2011). Recent work has, however, highlighted the stigmatisation of non-standardised varieties of students' community languages as a point of convergence across schools in a number of different diasporic communities, especially ones originating in societies with strong diglossic traditions such as Arabic-speaking communities, (Cypriot) Greek-speaking communities, (Cypriot) Turkish-speaking communities, and Banglaand Sylheti-speaking communities (Çavuşoğlu 2019; Georgiou and Karatsareas forthcoming; Karatsareas and Georgiou forthcoming; Matras and Karatsareas 2020;Soliman, Towler, and Snowden 2016;Walters 2011). Schools in these communities promote the teaching of the standardised varieties of the corresponding 'national' languages while at the same time discouraging and even sanctioning the use of non-standardised varieties mostly, but not exclusively, during formal teaching even in students who have very little exposure to the standardised varieties in their homes and within their communities. In many cases, language policing practices of this type echo educational policies and standard language ideologies that have been transplanted from the countries of origin, often incorporating discourses around success in life such as in securing qualifications, admission to university and future employment. Indeed, by only accepting standardised forms in their examinations, the languages GCSEs and A-Levels become mechanisms that shape the de facto language policies that complementary schools implement on a local and community level. In certain cases, they assume this role due to the lack of official language policies. Where such policies are in place for complementary schools, examination requirements reinforce the enactment of standard language ideologies.

Theoretical background
Our article is shaped by approaches, concepts and tools associated with critical language policy (e.g. Shohamy 2006;Spolsky 2009;Tollefson 2013) where policy is seen as a major tool used by authoritative bodies to manipulate language behaviour and practice. Research and theories aligned with critical language policy seek to account for the various ways in which speakers have their linguistic choices controlled and managed, whilst uncovering the ideological, political and structural processes at work and how speakers in the policy process (such as teachers) exercise agency and resistance (e.g. García and Menken 2010).
Underpinning this approach is a commitment to conceptualising language as a form of social action, challenging and disrupting hierarchical views of distinct named languages, language varieties and linguistic resources (Avineri et al. 2019). We follow Blommaert (2010) in viewing multilingualism not as 'a collection of "languages" that a speaker controls' but as a 'complex of specific semiotic resources … concrete accents, language varieties, registers, genres, modalities' (102). It follows from this perspective that speaker profiles and linguistic practices that are conventionally described as 'monolingual' may be more akin to profiles and practices that have been viewed as 'multilingual'. We further conceptualise the ways in which people draw on their linguistic and non-linguistic resources to participate in both 'monolingual' and multilingual settings in terms of the notion of linguistic repertoires (Blommaert and Backus 2013;Busch 2015) and in the context of the so-called multilingual turn in applied linguistics (Conteh and Meier 2014;May 2013;Meier 2017). This has a significant biographical dimension as repertoires encompass the full range of (non-)linguistic resources that people acquire along their life trajectories. The social turn in the study of multilingualism (Creese and Blackledge 2011;Jørgensen 2008) foregrounds language as a part of people's everyday practices and identities that needs to be used freely and flexibly even in contexts with a strong monolingual and monovarietal bias such as settings where teaching takes place. Schools, where imbalances in power and authority are notably exercised, are key spaces where language becomes subjugated, marginalised and legitimised. They are spaces where language socialisation processes are at work, in terms of how mainstream and normative assumptions about language get (re)produced and resisted (Corson 1999). Critical language policy work is then, not done at a distance, but involves getting inside the implementational spaces where policy gets done, lived and embodied. We position ourselves as critical ethnographers of language education policy, whereby policy is understood and critiqued in relation to the socio-political contexts in which it operates (Martin-Jones and Da Costa Cabral 2018). As in the approach associated with studies in the sociology of education policy (Ball, Maguire, and Braun 2012), policy is not simply implemented in a linear macro through to micro trajectory, but enacted. That is, policy is conceived of as a process whereby teachers' and other school-level policy makers' voices, agency and autonomy are foregrounded. Johnson and Johnson (2015) conceptualise policy as dynamic processes which are created, interpreted and appropriated in accordance to the local contexts in which they circulate. Policy texts alone do not constitute power; it is the negotiation and enactment of these texts where policy gets done, and where power imbalances and relationships come to be constituted. Giving weight to the agency and autonomy of teachers is crucial then, in understanding how policies get translated into practices. Policy enactment is not simply a matter of top-down imposition, but a multilayered and dynamic series of interactions with teachers in the centre, who 'stir the layers of the language policy onion' Ricento and Hornberger 1996). Traditional micro-macro policy distinctions are therefore problematised, for instance in the ways that they grant too much power to macro-level decisions and documents, or fail to fully acknowledge the complexity of the relationships and struggles between policy layers. At the same time, teachers undoubtedly work within a structure where their agency is under some kind of constraint, often through the imposition of external policy mechanisms such as curricula, tests and other accountability technologies. In the sections that follow, we consider how these constraints, impositions and pressures play a part in what our data looks like.
Language ideologies are crucial mediating factors in language policy enactment, which can be articulated or embodied. Whereas articulated ideologies state a belief about language, embodied ideologies represent a concrete language practice which are directly observable; for example, in written policies or in classroom interactions. As McGroarty (2010) states, whether or not language ideologies are explicit or implicit, they inevitably incorporate evaluations and judgements of language use and language users. In reference to this article, standard language ideologies are particularly important, namely entrenched beliefs that the standardised variety of a language is the 'best' or 'better' form, and that non-standardised varieties are somehow less good or less acceptable, an embodied consequence of this being that speakers of non-standardised varieties are stigmatised and discriminated against (Lippi-Green 2012; see also Rosa and Burdick 2017). Within the policy process, ideologies typically exist as different types of covert or overt devices or mechanisms (Shohamy 2006) such as language tests, curriculum frameworks, assessment criteria, guidance for teachers and school inspection reports. These mechanisms generally work as de facto policies; not declared as language policy as such but influence, create and work as vehicles for the ideologies contained within them.

Methods and data
In order to explore similarities and differences in the ways in which language policies are enacted across mainstream and complementary schools, we compared two independent studies both of which examined the active, agentive and autonomous role that teachers played in the policy process and how teachers brought their own sets of beliefs and attitudes towards standardised and non-standardised forms of language into policy making within the unique culture and context of their school. Cushing conducted the study focusing on language policing in mainstream schools. Georgiou and Karatsareas undertook the study on complementary education with Greek complementary schools as a case-in-point with the aim of examining ideologies and practices in London's Greek Cypriot diaspora towards Standardised Greek and the non-standardised Cypriot Greek variety. Both studies were ethnographically-oriented, aiming to gain rich understandings from the participants' points of view. They, however, differed in the types of data each collected, the number of sites they collected data from, and the frameworks they adopted for data analysis. The mainstream school study collected data from eleven schools, both primary and secondary.
Only data from secondary schools (ages 11-16) are discussed here due to space limitations. The study closely examined policy documents, which were analysed using a critical discourse analysis approach. The complementary school study focused on two schools that offered teaching ranging from the pre-school level to A-Level in an integrated way, that is, without internal differentiations into primary and secondary. Data were collected from years 5, 6 and the GCSE year. All students in these classes were 11 and above years of age at the time of data collection. Classroom recordings and interviews with teachers and students were the main data collection tools. A discourse analytic approach on spoken interactions was therefore adopted. Although the sample sizes of the two independent studies were different, our aim here is to draw contact points, overlaps and contrasts about language policing in mainstream and complementary education drawing on in depth interpretations rather than comparing results from standard sample sizes. We hope this will be a catalyst for future related work.
In both studies, we use pseudonyms for the schools, students and teachers.

Mainstream school study
This study generated a bricolage of data: contextual data (resources, staffing, buildings, local authority support), interpretative data (attitudes and beliefs of policy agents) and discursive data (language policy artefacts). This included interviews with teachers, institutional-level language policies, classroom observations, government policies, linguistic landscapes and pedagogical materials.
The nature and scope of this data helped to explore the 'messiness' of school life in relation to language policy and the lived experiences of language policing, with threads of data crossing over and interacting with each other akin to a 'ball of twine' (Heller, Pietikäinen, and Pujolar 2018). All data was uploaded to NVivo and organised using a coding framework relating to language and educational ideologies, which was then used to guide the organisation of the analysis section which follows as well as generate illustrative examples from the data. Very broadly, data was analysed using tools from critical discourse analysis (Gee 2014) viewing language as a form of social practice and seeking to reveal the ways in which language policing and control is reproduced in discourse.

Complementary school study
The data for this study were collected over a period of six months (January-July 2018) in two Greek complementary schools in North London, Anemomylos and Gefyri. Recordings of in-class teacher-students and student-student interactions were the main data type, accompanied with fieldnotes, oneto-one interviews with the six teachers and group interviews with the 70 students who were observed across the two schools and the three year groups. Of the six teachers, five were from Cyprus and one was from Greece. Students were for the most part British-born of Greek Cypriot heritage, having standardised and non-standardised varieties of English and Greek as part of their repertoires, English being their dominant language. In-class and interview recordings were transcribed in Greek and English, and a discourse analytic approach (Cameron 2001) was adopted to ensure a fine-gained analysis of the naturally-occurring data, which in turn allowed for the in-depth examination of the role of language in the reproduction of hierarchical ideologies. Physical artefacts including students' work, school documents, policy papers, books and other teaching materials, and classroom displays were also collected.

Language policing in mainstream schools
We begin with an extract from one school's 'literacy for learning' policy which serves as an initial illustration of how language ideologies and policing are deployed: Policy here propagates, reproduces and concretises the standard language ideology and its idea that there is a single, exclusive way of using spoken language. Schools here are constructed as monoglossic spaces whereby language is tightly regulated, controlled and disciplined. This is the language of expectation and failure, where policy grants teachers a listening licence to remedy, judge and evaluate students' linguistic choices and verbal contributions. The policy targets students' language at phonological, lexical, syntactical and discourse levels, with teachers positioned as the authoritative regulators and role models of language use, whose job it is to 'correct' deficiencies in speech, such as 'basic errors' in non-standard verb forms ('we was'). Students must talk in 'full sentences', suggesting that this policy is basing spoken language on the written form, and drawing a crude relationship between speech and writing which conflates and collapses the two modes together. Vague and subjective modifiers such as 'speak audibly' and 'crisp, clear reminders' police language at a prosodic level, further constructing classrooms as spaces where there are expectations made for certain ways of speaking. Teachers, too, have their language policed, not just in the requirement to use standardised English 'at all times', but in the tightly structured scripts, routines and sentence stems for managing classroom interactions. The beginning of this policy is framed by a number of aims, one of which is to 'raise standards for all students', suggesting the power of policy drivers which impose the 'standards' and 'performativity' agenda on schools under the discriminatory logics of standards-based reforms (Flores and Schissel 2014). There are a number of language policy mechanisms which serve to impose this (see Cushing 2021 for a full discussion), but generally steer teachers towards language pedagogies and policies which are underpinned by standards, correctness and competency. Across the data generated, one of the most pervasive mechanisms explicitly named in interviews and institutional policies was the Teachers' Standards (TS) (DfE 2013), a set of 8 government-produced standards that inand pre-service teachers must adhere to, first in order to gain their teaching qualification, and then to access promotion. The TS are used by management and Ofsted to judge teachers during lesson observations and school inspections, as a benchmark for measuring their general 'personal and professional conduct' (DfE 2013, 3). One of the standards includes the requirement that teachers 'must': demonstrate an understanding of and take responsibility for promoting high standards of literacy, articulacy and the correct use of standard English, whatever the teacher's specialist subject. (DfE 2013, 11, our emphasis) The TS work alongside other mechanisms which propagate standard language ideologies and licence language policing such as the primary school curriculum framework, which states that students 'should be taught to speak clearly and convey ideas confidently using Standard English' and 'should be taught the correct use of grammar' (DfE 2014, 10). Mandatory, stateissued grammar tests at primary school require students to 'correct' non-standardised forms of English (see Cushing 2020b for an extended critique), and a lengthy grammar glossary offers a depoliticised version of standardised/non-standardised grammar with limited, if any, exploration as to how and why language varies and changes according to society and culture. Assessment and curricula regimes in primary and secondary school (e.g. DfE 2014) place explicit emphasis on the requirement that students use standardised English in speech and writing, in order to gain qualifications which then work as gate-keeping mechanisms for entry into employment and further education.
The policy analysed above is not the only example in the dataset where such explicit recommendations about language policing are made, with various other policies encouraging and licencing teachers to patrol, police and discipline language throughout the school. Similar deficit discourses are found within these policies, manifested through word choices such as 'correct', 'error', 'inappropriate', and 'not be tolerated': Extract 2. Teachers should encourage the use of Standard English in lessons and around the school generally. Girls should be corrected when significant errors in grammar and word choice are used in their speech. Extract 3. Staff will challenge students when slang or inappropriate colloquialisms are used.
Extract 4. The way that students speak to each other and to staff denotes their character. We expect students to speak in full sentences and use standard English. The use of slang and inappropriate language will not be tolerated and students will be corrected by staff. We will encourage the students 'to leave the street at the gate' and model adult/professional language.
Written policies, however, only tell one part of the policy enactment process, and over-relying on written policies risks subordinating teachers' views and autonomy, especially given that institutional policies are typically produced by teachers in management positions (head teachers, literacy coordinators, heads of department). Interviews with teachers who did not occupy such positions revealed further ways that these policies were justified and framed. Typically, justifications were geared around a neoliberal, product-focused ontology of language: jobs, finances, employment and academic achievement, as well as a more general discourse of 'maintaining standards' (see Milroy 2001). The need for students to perform in named situations such as examinations and job interviews was a common theme of the interviews. For example: Extract 5. Interview with Cleo, Key Stage 4 teacher, New Urban Academy I do think students need to be able to use standard English in everything they do, and that's why I pick them up on their errors and make sure I'm correcting them when they get things wrong. It's about preparing them for a job interview and the world they'll face when they leave school.
Participants also explicitly named various policy mechanisms such as those discussed above, further suggesting that despite their own power and autonomy in the classroom, their practices are indeed constrained and influenced by external language policy drivers: Extract 6: Interview with Harriet, Newly Qualified Teacher, Green Tree School I am expected to use standard English when I'm teaching and that's been expected of me since doing my teacher training. I've been told to follow the Teachers' Standards, and to make sure I model this myself so that students know they must use it too.
Extract 7. Interview with Alice, Key Stage 3 teacher, Red Bush School I think the whole curriculum is set up in a way that makes non-standard speakers feel kind of picked on, you know? The tests are one part of that, the curriculum, it all works together to make people feel that the way they speak is wrong somehow.
Such intertextual connections up a policy layer provide evidence that language policing in contemporary mainstream schools often occurs as a result of what Loveday (2008) calls the 'tyranny of conformity', the pressures that teachers are under to deliver and perform to meet targets and standards. External policy drivers such as examination boards, inspection regimes, teacher assessment frameworks and national curricula play a significant role in constructing these cultures of pressure and policing.

Language policing in complementary schools
The teaching of Greek in Greek complementary schools in the UK is guided by curricula produced by the Ministry of Education of the Republic of Cyprus in collaboration with the Cyprus Educational Mission, a London-based unit tasked with supporting and overseeing the schools' operation. Owing to their high degree of institutional and operational independence, schools are not obliged to adopt and apply the curricula. Most schools, however, do draw on them to design and organise their teaching, and make pedagogical decisions. The curricula construct Greek as the most fundamental element in the formation of a Greek identity among students and as a communicative tool that can help them to forge transnational links with their Cyprus and Greece homelands as well as with Greek diasporas in other parts of the world. The most recent curriculum, the official version of which was published in 2019, was the first one to (reluctantly) acknowledge complementary school students as multilingual and multidialectal speakers by including specific references to English and Cypriot Greek as parts of their repertoires. The single reference to Cypriot Greek in the document constructs it as 'a dialect' as opposed to 'the Greek language', which is not an umbrella term that encompasses both Cypriot and Standardised Greek but, rather, an ideologically laden term to refer to the latter (Adger, Wolfram, and Christian 1999;Lippi-Green 2012;Wolfram and Schilling 2016).
Extract 8. The single reference to Cypriot Greek in the 2019 curriculum The use of the Cypriot dialect is accepted and utilised in language learning in a natural, guilt-free and functional way. [The Cypriot dialect] is also used to connect [students] with Cypriot cultural traditions through songs and poems. (7) However, the draft version of the curriculum, which remains publicly available since 2018, included more references to Cypriot Greek and some limited guidance for teachers on how to manage its presence in the classroom. It specified that the variety was only to be accepted in 'oral communication' (16), thus institutionalising two beliefs that are widely held among teachers, students and the wider community: first, that Cypriot Greek can only be spoken but not written, whereas Standardised Greek can and must be written; and, second, that students' using Cypriot Greek features in their production is allowed in the oral component of the GCSE and A-Level examinations but penalised in the written components (Georgiou and Karatsareas forthcoming;Ioannidou et al. 2020). Despite being cast as a formal requirement, the latter belief is not supported by documentary evidence relating to the two qualifications (specifications, mark sheets, teaching and learning materials). Interviews with teachers showed that it is reproduced within the complementary school network as an unwritten but powerful rule. Justification for adhering to it is often given in terms of academic achievement and the instrumentalist use of the language qualifications, which are among the main reasons for students to attend the schools (Karatsareas 2021; Karatzia-Stavlioti and Louca-Crann 1999).
Extract 9. Interview with Christina, GCSE teacher, Anemomylos Greek School. Oh, I don't know. I've heard it from our examiners, I'm not an examiner myself. So, yes, the examiners say that this is what Edexcel says. So, we follow what they tell us.
Extract 10. Interview with Angela, year 5 teacher, Gefyri Greek School. This is what I keep saying to [the students]. You must pass the GCSE because it will help you to gain extra credit. If you do the A-level, it will help you get into university. That is the only thing that will keep them and the only one that does keep them. If you speak with the parents I speak with, they all say I want [my children] to get the GCSE and A-level.
In-class observations in the two schools showed that teachers reproduced prescriptive ideologies, constructing Standard Greek as 'the language' and Cypriot Greek as 'a dialect'. While they recognised that Cypriot Greek was an important part of students' identity, family history and cultural heritage, they variably presented it as too regional, inaccessible for speakers who only speak the standardised variety or other Greek 'dialects', and inappropriate for use in formal contexts such as in teaching. In order to make these notions clear to students, they engaged in a wide range of language policing practices, both explicit and implicit. A representative example is shown in the following extract, in which the teacher is providing feedback on a student's written assignment, singling out the Cypriot form μεινίσκω [miˈnisko] 'to stay' and promoting the corresponding standardised form μένω [ˈmeno] alluding to the notion of unintelligibility.
5 Ms Eleni τζιαι στις εξετάσεις; τι θα σκεwτούμε; and in the exams? what will we think? 6 Alexis Greek would you lose marks for saying Κυπριακά Greek would you lose marks for saying Cypriot The extract suggests that students have internalised the hierarchical ordering of Cypriot Greek and Standard Greek. This was further corroborated by focus group discussions in which students reproduced the spoken versus written binary contrast and the complementary distribution of the two varieties in the written and oral components of the GCSE examination, expressing particular concern about the consequences of using Cypriot Greek in the examination. Their narratives included accounts of teachers' practices of delegitimising and disapproving of Cypriot Greek, recounting incidents of corrections and even cases in which teachers reacted to their use of Cypriot Greek by laughing. In talking about Cypriot Greek, they routinely applied negative labels that are known from the Cyprus context to refer to the variety's most basilectal register, especially χωρκάτικα [xorˈkatika] 'village talk'.
Extract 12. Danai's (year 6 student at Gefyri Greek School) contribution to focus group discussion. We don't talk Greek, it's Cypriot, we talk like villagers.
When asked to elaborate on this notion, they framed their arguments within the schemata of rurality, correctness and properness. At the same time, they recast the opposition between Cypriot and Standardised Greek ways of speaking in terms of binary contrasts with which they are familiar as multidialectal speakers of English, most notably the binary posh versus slang (Karatsareas 2020). They also mentioned that they sometimes found Standardised Greek inaccessible in the sense that they could not produce or understand speech in it easily due to lack of sufficient exposure outside the context of the school.

Comparative discussion
Discourses of deficit were evident in both mainstream and complementary schools. In both educational settings, standardised varieties tended to be legitimised and non-standardised varieties tended to be stigmatised. In mainstream schools, varieties were mainly hierarchised through policy documents that constructed standardised English as the only acceptable way of using English, excluding all other varieties. In complementary schools, policy documents played a more secondary role. Hierarchies were mostly reproduced through the practices of teachers who actively discouraged students from using Cypriot Greek in line with the curriculum that reinforced the idea that the non-standardised variety of Cyprus was 'a dialect'. In both settings, policy documents and teachers' practices as key policy mechanisms silently acknowledged the fact that students have expanded linguistic repertoires. Both, however, treated students' languages and varieties as discrete and bounded entities, and only promoted standardised varieties, thus creating few or even no opportunities for students to capitalise on their full repertoires during teaching.
In both types of schools, a distinction was made between spoken and written forms of language, a binary contrast that was reproduced through teachers' language policing practices in both contexts. In mainstream schools, policy documents elevated written language to the status of a model that spoken language had to follow and instructed teachers to expect from students to speak in the same way as they are expected to write. Features of spoken language including nonstandardised forms were labelled as errors that had to be eradicated from students' language use, at least in the context of teaching and learning. In complementary schools, the spoken versus written language distinction was made by reproducing what is constructed as a piece of widely shared and commonsensical knowledge that echoes the sociolinguistic and educational context of Cyprus, namely, that students may use the 'unwritable' Cypriot Greek 'dialect' to some extent in speaking but never in writing. This compartmentalisation of students' repertoires and the suppression of their non-standardised and/or spoken parts can engender negative attitudes towards their linguistic identities and prevent them from freely participating in classroom exchanges as shown in Ioannidou (2014).
The focus on performance and the exchange value of standardised varieties was found in teachers in both settings. In mainstream schools, policy documents were powerful in reproducing the performativity discourse by emphasising the qualifications that students needed gain. Teachers built on this narrative to make further links with employability opportunities, further education and students' succeeding in demanding and decisive real-life situations such as job interviews. In the same vein, complementary schoolteachers supported the promotion of the standardised variety with reference to the requirements of the GCSE and A-Level examinations. The two qualifications were presented as the key that would lead to students' success in terms of being admitted to university. There was also an assumption that a high grade should be easily achievable in the Modern Greek GCSE and A-Level given the students' background such that failing to attain this target due to the use of Cypriot Greek in writing would equal to missing a great opportunity. Such views of language as a commodity risk widening the social inequality gap between students who have access to the commodity and students who do not as in most of the times the latter, participate 'into a system of linguistic evaluation that works against them' (Snell 2013, 21).
While teachers in both mainstream and complementary schools reproduced hierarchical ideologies and engaged in language policing practices, at the same time they conceded that school policies about standardised and non-standardised language, whether overtly or covertly communicated, did not always accommodate their students' learning needs. Teachers in mainstream schools recognised that policy documents put some students in a marginalised position and insinuated that they were ready for changes in the curriculum. Similarly, teachers in complementary schools acknowledged and appreciated the central role Cypriot Greek plays in forming students multilingual and multicultural identities, and expressed positive views towards a relaxing assessment requirements. These findings suggest that teachers do not blindly follow policy documents but can and do consciously problematise well-established ideologies through their own beliefs, which could be used as a powerful mechanism to bring about change at a micro policy level .
There were, in contrast, two main differences between the two contexts: the types of policy mechanisms employed in each case and the degree of explicitness in the ways in which language policies were performed. In mainstream schools, there was a heavy focus on policy documents. In complementary schools, teachers' practices were more powerful in sustaining language ideologies. This may be explained by differences in the degree of schools' operational independence. Mainstream schools run under a much more centralised system than complementary schools and are required to adhere to a large body of policies, guidelines and standards set by government for a vast number of school units. Complementary schools, on the other hand, constitute community initiatives that are established from the bottom up on a very local level and have more autonomy in making decisions relating not only to pedagogy but most, if not all, aspects relating to their day-today and long-term running. While in some cases complementary schools do have links with national governments of the countries of origin of the communities they serve, these are often loose as overseas governments take on a more supportive rather than regulatory role. This is the case of Greek complementary schools in the UK and the links they have with the government of the Republic of Cyprus via the Cyprus Educational Mission. Against this backdrop, teachers take centre stage in the shaping and enactment of language policies. These differences account also for the varying degrees of explicitness across the two school types. The centralised nature of mainstream schools fosters the production and use of powerful policy documents, which 'fix the message in stone' (Lippi-Green 2012). In contrast, the high degree of operational autonomy of complementary schools and their loose dependence on centralised government creates conditions for more implicit enactments of policy such as through communicating expectations and requirements through word of mouth.

Concluding remarks
In this article, we have explored similarities and differences in terms of the policing of non-standardised language across mainstream and complementary schools in England. Previous research has identified several aspects of what can be termed a disjointed coexistence between the two settings, which operate largely independently of each other despite the fact that together they compose the full educational experience of a non-negligible number of multilingual students. Kenner and Ruby (2012), for example, found that mainstream school teachers, despite being aware that their students had rich linguistic repertoires, knew very little about their students' fluency in their community languages and doubted whether or how students' languages could be relevant to their learning (they were, however, keen on exploring this possibility), while few teachers had visited a complementary school. From their part, multilingual students have also been found to perceive the two settings as disconnected from each other and fundamentally different in their nature, scope and approaches to learning (Archer, Francis, and Mau 2009). In the light of these and other similar findings, there have been calls for more and closer links to be established between mainstream and complementary schools with the aim of improving the educational experiences of students in multilingual classrooms and countering the assimilationist effects of monolingualising discourses and ideologies (Conteh 2015;Robertson, Drury, and Cable 2014;Sneddon 2007Sneddon , 2014. Such calls often argue in favour of developing multilingual pedagogies that draw on students' sociocultural and linguistic knowledge across the mainstream curriculum and on the expertise and experience of complementary school teachers, many of whom work in the mainstream sector as well. Our findings highlight a point of convergence between the two settings and have pedagogical implications for both mainstream and complementary schools. Despite their differences and the culture of separation between them, both educational systems independently enshrine the hierarchisation of standardised and non-standardised varieties in institutional policy and share a strong orientation towards the ideological valorisation of standardised varieties and the delegitimisation of non-standardised ones. Teachers in both systems enact such monovarietal policies by drawing on discourses around academic success and the primacy of written language, and by engaging in language policing practices that stigmatise, ban or sanction the use of non-standardised language on behalf of students. This suggests that multilingual students who attend both mainstream and complementary schools are exposed to similar kinds of prescriptive discourses across the whole spectrum of their educational experiences, which can have negative effects on their learning and the construction of their self-images as speakers of their languages and varieties. Apart from sites of multilingualism (Blackledge and Creese 2010;Lytra and Martin 2010) and safe spaces for multilingual and multicultural students (Chatzidaki 2019; Conteh and Brock 2010; Creese et al. 2006), complementary schools may also become sites of tensions that suppress parts of the linguistic repertoires and competences students possess. We would therefore join the calls for more connections and cross-fertilisation between mainstream and complementary schools and the development of integrated pedagogical approaches and policies that draw on the students' whole 'funds of knowledge' (Moll et al. 1992; see also Conteh 2015). However, we would argue that these need to legitimise non-standardised varieties and treat them as valuable linguistic resources and valued parts of students' life trajectories, their everyday lives and their identities as speakers. The benefits of such inclusive approaches have been identified independently for mainstream (Godley, Carpenter, and Werner 2007) and complementary schools (Ioannidou et al. 2020;Matras and Karatsareas 2020). We advocate in favour of bringing these two strands of work together in partnerships that will challenge linguistic prejudices and discourse of linguistic deficit across the educational spectrum, whilst at the same time, also join calls within educational linguistics to interrogate the listening practices of white authoritative bodies rather than simply arguing for the 'celebration' of linguistic diversity (Rosa and Flores 2017).
A limitation in our analysis stems from the fact that the two studies we compared were conducted independently of each other and were based on different types of data and on different sample sizes. In order to better and more fully explore the effect language policing has on multilingual students who have both standardised and non-standardised varieties in their repertoires, future work will need to examine the same students in all the educational settings in which they participate, both mainstream and complementary. The complementary schools study provided evidence that multilingual students construct the relation between standardised and nonstandardised varieties of their community language with reference to notions, ideological schemata and labels drawn from the English part of their repertoires and the stigmatisation of nonstandardised English varieties as 'slang' (Karatsareas 2020). This supports the idea put forward by Kenner and Ruby (2012) that the two worlds in which multilingual students find themselves are interconnected and that students are the connecting points whose full educational experiences need to be better understood.

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

Notes on contributors
Ian Cushing is a Lecturer in Education at Brunel University London. His research examines the ways in which standard language and raciolinguistic ideologies get transformed into policies and practices, and how certain speakers and bodies are placed under surveillance in schools.
Alexandra Georgiou is an Adjunct Lecturer in the School of Education at the University of Nicosia. Her research focuses on inclusive education and multilingual teaching and learning drawing on sociocultural perspectives. She has been involved in numerous research projects working with multilingual and multicultural families and communities in the UK as well as with community-based organisations in Europe. Alexandra worked as a primary school teaching in London's mainstream schools and Greek complementary schools.
Petros Karatsareas is a Senior Lecturer in English Language and Linguistics at the University of Westminster. He specialises in the sociolinguistics of multilingualism in contexts of migration and diaspora with a focus on the languages of minoritised communities in the UK. He is interested in language ideologies, attitudes towards non-prestigious linguistic varieties, and community language education.