‘Say it like the Queen’: the standard language ideology and language policy making in English primary schools

ABSTRACT
 This article presents an analysis of the standard language ideology within a corpus of school-designed language policy documents from 264 primary schools in England. It examines the processes by which standard language ideological concepts (e.g. ‘Standard English’, ‘correctness’, ‘hegemony’) get textually manifested in school policies, and how these are intertextually and interdiscursively shaped by the broader educational policy context that teachers work in, notably the large-scale curriculum and assessment reforms of National Curriculum 2014. Using tools and methods from critical language policy, I reveal how new meanings emerge in the machinery of the policy-making process and at the contact points between policy levels. I trace how the standard language ideology within government policies gets reconstructed in school policies, with an emphasis on linguistic ‘correctness’ and the near-exclusive requirement for students and teachers to use standardised English in speech and writing. I discuss policies of surveillance, whereby teachers are discursively constructed and positioned as standard language ‘role models’: as powerful and authoritative figures who are granted a license to police, regulate and suppress their students’ language, whilst also having their own language controlled and monitored. Finally, I argue for the place of critical language awareness within the policy-making process at school level.


Introduction
Model accurate talk by addressing grammatical errors […]. Model competent speech and show them the difference between clarity and slang. The way that pupils speak to each other and to staff denotes their character. We expect pupils to speak in full sentences and use standard English. The use of slang and inappropriate language will not be tolerated and pupils will be corrected by staff. We will encourage the pupils 'to leave the street at the gate' and model adult/professional language.
The above excerpts are taken from the language policies of two schools in England, which construct and position teachers as 'standard language role models' who have the authority to police, regulate and suppress students' language. The language discourses found in meso-level, school-produced policies are the focus of this article, which investigates the standard language ideology within a corpus of language (or 'literacy') policy documents from 264 primary schools in England, and how these are shaped by the broader educational policy context which teachers work in, notably the reforms of National Curriculum 2014 (NC2014). Whilst school policies featuring strict 'standard English only' and 'slang bans' have been previously criticised (e.g. Cushing, 2020a;Snell, 2013), this work has focused on individual schools rather than a network of ideologies across policy levels and processes. The current article contributes to existing research by exploring patterns across a large dataset of meso-level policies, to capture how macro-level policies are interpreted and adapted by policy actors in schools. I use concepts and tools from critical language policy in order to answer the following questions: 1. What textual traces of the standard language ideology exist in a corpus of primary school language policies? 2. How are school language policies shaped by the national educational policy context?
It is not my intention to criticise schools, but to critically analyse language ideologies within the machinery of the policy-making process and how they surface within a system where teachers' choices are shaped by government policies and pressures, especially during times of curriculum change (Liddicoat, 2019). Although my focus is on England, the issues concerning language policing and language ideologies are pertinent to schools across the globe (see, e.g., Heller & McElhinny, 2017).

The standard language ideology in policy and practice
Language policy includes all the practices, beliefs and management where issues of language are concerned (Spolsky, 2009), a set of components which are enacted not just through texts but through practices and pedagogies. The critical turn in language policy research foregrounds the political and ideological nature of policy making, whilst focusing on the relations between power, institutions and practices (e.g. Shohamy, 2006;Spolsky, 2009;Tollefson, 2013). Here, language policies are shaped by language beliefs and ideologies, through which individuals can become socialised in during participation in cultural activities (e.g. Ajsic & McGroarty, 2015). Language ideologies can become so entrenched that they come to be seen as 'natural' or 'common sense', masking the structural processes which work to create and perpetuate them. Schools are a key space for this entrenchment, which occurs via language policy 'mechanisms' such as curricula, pedagogies, tests and regulations (Shohamy, 2006). Rosa and Burdick (2017) discuss a range of work exploring how these ideological rituals are at work in schools, which reify hierarchical, monoglossic ways of thinking and doing which are intimately bound up with power and authoritative control.
One of the most prevalent language ideologies is the standard language ideology (Lippi-Green, 2012;Milroy, 2001); the belief that a language has fixed, easily identifiable forms with a clear delineation between 'standard' and 'non-standard'. The 'standardised form' is constructed by and associated with powerful social groups (western; literate; white; male; middle-upper class), who manage access to opportunities such as employment and education, using standardised language benchmarks as a gatekeeping mechanism. A material consequence of the standard language ideology is that non-standardised forms get subordinated through being constructed as 'deviant' and 'non-compliant', leading to the stratification of language varieties. Institutions such as schools can reproduce these ideologies in their daily activities (e.g. Godley et al., 2007), with language policies often propagating the arbitrary requirement for students and teachers to speak and write in the standardised variety, which can legitimise instances of language policing, punishment and surveillance (see Cushing, 2020a).
Theories of language policy typically use a 'level' or 'layer' metaphor to capture the roles of different actors and power asymmetries, from 'macro' through to 'meso' and 'micro'. I use this metaphor as a convenient starting point rather than a fully representative model of policy organisation, especially in the ways that I consider teachers as agentive arbiters of the policy-making process rather than 'mere conduits of decisions made by lawmakers' (Weinberg, 2020; see also Johnson, 2012;Johnson & Johnson, 2015;Menken & García, 2010). Within this, teachers' ideological and ontological beliefs about language can determine policy, as do the socio-cultural and political conditions in which they work. Nevertheless, teachers' agency is almost always under various constraints, with internal and external pressures affecting the policy-making process. As one way of capturing these dynamic struggles and negotiations, policy enactment theory (Ball et al., 2012) shifts away from the level/layer metaphor towards 'processes' and 'cycles', highlighting the creative work that teachers do in appropriating and recontextualising government policy, especially in accordance with the cultures and contexts of their particular school. Finally, a key concept in exploring the language policy process is intertextuality, where policy documents exist within a network of flows, pathways and relations, reverberating with the ideas of multiple writers. The 'meaning' of a language policy text emerges not in isolation then, but through its textual and discursive contact points with other texts and policy documents of past and present (see Johnson, 2015). Intertextual analyses of language policy help to illustrate historical continuities, how structural and socio-political factors have led to text creation, and where ideas within texts 'come from'.

Language policy making in England's primary schools
Debates about language ideologies and practices in England's schools reach back hundreds of years, and although a detailed history is beyond the limitations of this article, critical descriptions are available, amongst others, in Cameron (2012) and Crowley (2003). For this study, I focus on school policies made after the educational reforms of 2010, triggered by the coalition Conservative-Liberal Democrat government, and from 2015 onwards, under the majority Conservative government. These reforms were major, affecting curricula, assessment, teacher education, funding, school organisation and the role of school inspection bodies. They can be broadly characterised as being underpinned by nostalgic, 'traditional' and conservative educational ideologies, with an overt focus on 'discipline', 'rigour' and 'standards' (e.g. Gove, 2010).
Within the primary school component of NC2014, one major change to curricula and assessment policies was concerned with language, with an explicit (re)emphasis placed on students and teachers to use standardised English in their speech and writing. This was one attempt by the Conservative government to instil 'back to basics' grammar work in schools reminiscent of the 1950s, under the guise that children had become 'ignorant' of 'proper grammar' under the previous Labour government (Gove, 2010), and that explicit language knowledge could be tested in terms of 'right or wrong answers' (DfE, 2011). These shifts work as language policy mechanisms which propagate the standard language ideology and de-value non-standard varieties. Within these mechanisms, 'standard English' is a sanitised and depoliticised construct (see Cushing, 2020b), shielding its socio-political power and disassociating it from any kind of inherent linguistic bias concerned with nationality, race and class. Language discourses of 'correctness' are rife in NC2014, with both teachers and students positioned as policy subjects who are pressured into compliance. For example, teachers are required to 'promote high standards of literacy, articulacy and the correct use of standard English' (DfE, 2013, p. 11), and to teach 'the correct use of grammar' (DfE, 2014, p. 10). Policy conflates written and spoken language, with students under near blanket requirements to speak 'formally' and 'clearly' using standardised English (ibid.), and take compulsory, state-issued grammar tests during which they must identify decontextualised sentences written in standardised English and 'correct' non-standardised constructions. Grainger and Jones (2013) note these policy shifts as a 'resurgence of the socially intolerant "deficit" approach to children's language and communication', which has been criticised by educational sociolinguists since the 1960s (e.g. Rosen, 2017;Snell, 2013). These debates around language in UK schools cannot be disentangled from race and social class, given that standardised English is a social construction which is built and maintained by white, middleupper class, 'literate' and 'educated' speakers. Although non-standardised dialects have long been shown to be just as 'grammatical' as the standardised form (e.g. Trudgill, 1975), these deficit accounts of working class, non-white language still persist in the UK. In her ethnography of an English primary school, Snell (2013) argues comprehensively that one reason for this is that the standard language ideology draws synthetic boundaries between languages and varieties, but in sociolinguistic reality, speakers creatively and rapidly 'mix' language varieties and sets of resources. Schools, a major arena for the playing out of the standard language ideology, are one place where these linguistic boundaries get drawn up, enacted and entrenched, often through the 'correction' of students' non-standardised spoken grammar which can lead to language-based prejudice, often at intersection points with race and class.
Critically orientated work on literacy and language policies in England has examined the consequences of political intervention in school-level policy making and the creation of contested spaces, whereby schools navigate contradictory policy messages which can overly-prescribe the content of local policies. Both Lefstein (2008) and Moss (2009) highlight how such intervention creates tensions between government and schools, especially concerning the use of quantitative data to generate and monitor literacy 'performance' which can pressure teachers into top-down policy compliance. Following two major initiatives by the UK government, first the National Literacy Strategy from 1997 and then the Primary National Strategy from 2003, schools in England have increasingly grappled with multiple language policies, including ones passed down to them by the state, and ones produced through their own policy-making activity. Mills's (2011) work on literacy policy in England shows the interplay between internal policy actors (teachers; head teachers) and external actors (local authorities; private consultants and government agencies). He reveals the struggles that teachers experience in policy making, negotiating a cluster of policy drivers such as league tables, targets, accountability measures, pedagogical beliefs, professional identities and priorities local to the school. Ultimately, policies produced by schools can reveal material traces of policy shifts and trends found at macro-level. These traces, especially those concerned with the standard language ideology, are the key foci of this article.

Data
The research questions were answered by generating and analysing a corpus of schoollevel language policy documents from 264 primary schools in England. This corpus was generated using web scraping software, with parameters defined by relevant keywords ('policy', 'language', 'literacy', 'standard English', 'non-standard English', 'primary school'). I limited my searches to only include policies produced by non-selective, nonfee paying schools in England between 2014 and 2020, in order to capture texts which had been produced following the introduction of NC2014 and its associated reforms. The corpus consisted of policies from the two main types of school in England: 144 were local authority (LA) schools; 120 were academies. 1 All institutions were under Ofsted 2 jurisdiction. All policies were available for download on a publicly available section of each school's website, and these were manually checked before adding to the corpus. Prominent generic features of the policies included various rules and regulations, the name of the school and its logo, any named authors (typically the 'literacy coordinator'), a signature of approval from the head teacher and board of governors, a rationale, and sub-policies concerning areas of literacy, such as oracy, writing, reading, vocabulary, handwriting, spelling, punctuation, grammar and phonics. Some policies included a 'curriculum map', i.e. an overview of the language topics to be taught each year, and these were typically driven by the programmes of study in NC2014 (see DfE, 2014, pp. 20-48).

Analytical procedure
The analytical procedure drew on a number of complementary approaches to critically exploring language policy discourses: corpora and ideology analysis (e.g. Vessey, 2017); critical discourse analyses of policy documents (e.g. Barakos, 2016) and intertextuality in language policy (e.g. Johnson, 2015). Together, these approaches share a commitment to revealing discourses of power and control and how texts are shaped by their socio-political contexts and historical continuities. The process involved me conducting a close reading of all policy documents and sketching out some broad emerging codes before importing the texts into NVivo. Next, text searches across the entire corpus were used to investigate keywords associated with the standard language ideology and refine the broad codes identified in the initial reading of the data. Search terms included metalinguistic nouns and verbs (e.g. 'English', 'grammar', 'speak') and modifiers with subjective connotations (e.g. 'correct', 'articulate', 'accurately'). This enabled me to refine and finalise the coding framework, which was driven by the RQs and anchored to the linguistic patterns of the texts. Top-level codes included: 'representations of standardised English', 'intertextuality across policy levels', 'language modelling and policing' and 'internal/external policy drivers'. Extracts from policies which illustrated these codes were selected for a closer textual analysis, where I carried out critical analyses of language policy discourse (e.g. Barakos, 2016;Johnson, 2015). This provides a systematic procedure for examining how ideologies are encoded in texts, focusing in particular on the lexico-grammatical choices representing language issues, intertextuality, interdiscursivity and the relations between policy documents, and the socio-political and historical contexts in which texts exist.

Findings and discussion
This section presents the main findings, split into subsections which are driven by prominent codes which emerged through the coding process. In the first subsection, I look at textual traces of the standard language ideology in the policies, with a particular focus on intertextuality across policy levels, showing ways in which local policies appropriate government ideologies. The second subsection examines how policy texts discursively construct teachers as role models and regulators of their students' language, further entrenching and perpetuating power asymmetries associated with the standard language ideology.

Standard language ideologies and intertextuality
Traces of the standard language ideology in the corpus were markedly shaped by intertextuality, where schools explicitly referenced government-produced policies and/or used policy templates produced by external organisations. 3 NC2014 and its associated mechanisms worked as a powerful policy engine, with schools typically stating that their policies should be read 'in conjunction' with government guidance as well as alongside other school policies. For example, one school listed a selection of extracts from NC2014 before stating: We therefore need a rigorous whole-school English policy which is implemented systematically, with all teachers having due regard to the expectations of both the NC2014 and the Teachers' Standards as teachers of English, whatever their subject specialism. All pupils are expected to have aspirational progress targets so that they are on track to meet or exceed, where possible, expected standards by the end of each key stage. There is now a strong emphasis on the teaching of spelling, punctuation and grammar (SPaG). During inspections, Ofsted will place a stronger emphasis on effective whole-school English and Literacy policies and their successful and systematic implementation across the school. Finally, the new Teachers' Standards require all teachers to 'demonstrate an understanding of, and take responsibility for, promoting high standards of English and Literacy, articulacy and the correct use of standard English.' All teachers must know the English and Literacy barriers for pupils in their classes and adapt their teaching accordingly. Pupils will be tracked half termly in writing and reading. (S17, Local Authority [LA] school) This extract is particularly illustrative of intertextuality across policy levels, where textual features from macro-level policy have been reassembled, recontextualised and recirculated by the school. Through these re-processes, policies come to rely on past texts and discourses, with 'earlier versions either expanded upon, added to, filtered, and/or suppressed' (Johnson, 2015, p. 168). In the S17 excerpt, two macrolevel policies (NC2014 and the Teachers' Standards) are explicitly named and added to, with a quotation directly taken from the latter. Lexico-grammatical choices suggest the school policy is designed to uncritically accommodate the content of macro-level policy and is driven by the standards agenda in schools, namely the pressures to 'perform' to external pressures: 'we therefore need a rigorous […] policy … '; 'there is now a strong emphasis on … '; 'The Teachers' Standards require … ', and 'Ofsted will place a stronger emphasis … '. School policy is here subordinated to government-issued directives and initiatives, with these working as a language policy mechanism which shapes what local-level policies come to look like (see Ball et al., 2012). The overt focus on preparing students for the national grammar tests, the 'tracking' of students' abilities, the issuing of 'progress targets' and the need to meet 'expected standards' indicates the surveillance, assessment and performativity culture of education policy in England, where accountability and measurements drive curricula and pedagogies. Where these kinds of intertextual connections across policy levels are made, they serve to legitimise, reproduce and institutionally embed the macro-level language ideologies within schools, forging interdiscursive relationships across genres, contexts and texts (Johnson, 2015, p. 167).
A further lexico-grammatical feature of interest throughout the corpus was where subjective modifiers were juxtaposed with metalinguistic nouns and verbs (e.g. 'accurate grammar'; 'speak properly'), a construction which is firmly embedded in the history of macro-level policy (e.g. DfE, 1995;DfE, 2014 Policies here imply that (a) it is universally understood what is meant by 'standard English', (b) all teachers and students are speakers of standardised (British) English and (c) and that this is the exclusively legitimate form of the classroom. Such policies fail to distinguish between the complexities of different genres and contexts of talk, with a blanket requirement for the use of spoken standardised English 'at all times'. Despite its textual prevalence, not a single policy offered their own definition of standardised English, and so consequently, teachers are permitted to enact any existing biases they may hold about standardised/non-standardised language in their classroom practice. The phrase 'the correct use of Standard English' appeared 32 times in the corpus, being a direct quote from the Teachers' Standards (DfE, 2013), often used to frame schools' policy rationale and concretising the standard language ideology. These types of normative discourses about language assume that the 'standard' form is the reference point which 'non-standard' forms are judged against, and work to further entrench existing inequalities which arise as a result of the domination of standardised English and its speakers. As previous sections showed, these hierarchical discourses permeate current language policy mechanisms in England, with their material and symbolic power seeming to coerce the writers of school-level policies into reproducing them and further embedding their institutional legitimacy.
The phrase 'standard English' appeared much more (332 hits) than 'non-standard English' (8 hits), reflecting the reification and emphasise that education policy past and present has placed on this, and how schools' policies held standardised English up as a model. Policies varied in their treatment of non-standardised English, with some making attempts to highlight the social aspect of language variation, and the ways in which language varies according to context. For example: An understanding of the relevance and importance of standard English is fostered as well as a respect for dialect and accent. Children are encouraged to speak in a variety of situations and to use the appropriate form, manner and tone for the respective situation and audience. (S48, LA school) Some policies sought to draw students' critical attention to linguistic variation and explicitly acknowledged the need for local dialects and home languages to be a part of school life, in what might be thought of as an attempt to resist some of the near exclusive emphases placed on standardised English within macro-level policy. These 'moments of resistance' were sparse in the corpus, but represent efforts to craft policies which serve local, rather than state-imposed concerns: We view positively the skills all children have in literacy, regardless of accent, dialect or home language. (S75, LA school) [Students] consider language variation, standard English and dialect, and identify the features of language used in formal and informal contexts. (S89, LA school) Some policies revealed the tension that teachers experience in knowing how to ensure that their students gained awareness of standardised English in adhering to the demands of macro-level policies, whilst balancing a respect for non-standardised forms. This often resulted in somewhat contradictory policy statements, such as that of S103 which sought to 'sensitively explore and celebrate the diversity of regional dialect' but also advocated for an assimilationist pedagogy whereby pupils' non-standardised features should be 'identified and converted into Standard English', and that teachers should use Standard English 'at all times, when interacting with children'.
As indicated in S48 above, discourses of standardised/non-standardised English in schools often occur in reference to 'appropriateness', where policies instruct teachers to teach children to use the 'appropriate' forms at the 'appropriate' times. NC2014 mechanisms use the appropriacy argument extensively, e.g. in DfE (2014:, p. 10): 'students should [… .] select and use appropriate registers for effective communication'. Direct copies and various adaptations of this line appeared throughout the corpus. In his critique of appropriacy-based policies, Fairclough (1992) adopts a language ideology lens, arguing that 'appropriate' simply works as a proxy for 'standard', and that it is dominant social groups who have normalised what counts as 'appropriate'.
He questions the notion that it is always clear what constitutes so-called appropriacy, arguing that the model is based on a reductive version of sociolinguistic reality where social, linguistic and contextual 'boundaries' are neatly delineated. Janks (1997)

echoes this:
'Appropriateness', the concept at the heart of the social in language education, comes under the critical knife because what is appropriate is decided by social norms which, in contexts of power (institutions, prestigious job interviews, the media), are inevitably the naturalised cultural practices of the social elite. (Janks, 1997, p. 243) For Flores and Rosa (2015:, p. 152), appropriateness-based pedagogies conceptualise standardised/non-standardised English as dichotomous, objective categories rather than as raciolinguistic and ideological, and so further empower standardised English speakers whilst reproducing normative practices by expecting language-minoritised students to model their language on their subjugators. In her work on anti-black linguistic racism in schools, Baker-Bell (2020) discusses how standardised 'academic' English works as a proxy for whiteness, and how an 'appropriateness' policy is in effect a covertly racialised practice which reflects and maintains linguistic and social stratification. The use of elite members of society as language models was particularly explicit in one school's policy, which stated: We recognise the need for all pupils to speak, read and write Standard English fluently and accurately, while acknowledging that a pupil's own dialect, or other language is of prime importance. It is our school policy to encourage children to 'Say it like the Queen'. This promotes Standard English. (S67, LA school) The first sentence here is reconstructed from a hybrid policy template of various macro-level policy curriculum documents produced by multiple governments (DES, 2006;DfE, 1995DfE, , 2014. The adverbs 'fluently' and 'accurately' as used to describe standardised English have been present in all iterations of the primary National Curriculum since 1995, and are deeply embedded textual resonances of the 'standards discourse' of the National Literacy Strategy (e.g. Moss, 2009) and the 'literacy hour' (e.g. Mroz et al., 2000). Variations of this sentence appeared throughout the corpus, illustrating diachronic intertextuality within meso-level policy making, and policy lending across different schools. The emphasis on conformity, 'fluency' and 'accuracy' negates any attempt to 'acknowledge' linguistic variation, with the ultimate requirement appearing to be that children should speak in the same way as they write. It is unclear how a policy which encourages children to 'say it like the Queen' would also acknowledge that their own dialect is of 'prime importance', and so teachers here must deal with contradictory and assimilationist messages about language. Ultimately, it raises questions as to why the Queen of England's hyper-formal linguistic stereotypes should be an aspirational target for school students and reinforces the deficit message that their own linguistic repertoires are not suitable for classrooms.

Modelling, correcting and policing
As touched upon in the final example above, this section explores moments where policies attempted to model, correct and police the use of language. Many of the policies positioned teachers as authoritative, standard language role models, for example: Staff will constantly model the correct use of Standard English to reinforce children's understanding of it. (S234, LA school) Children need to be using correct Standard English and powerful verbs and adjectives when addressing a group. The use of correct Standard English needs to be modelled by the teacher and the teacher needs to be providing feedback on the type of language children are using. (S174, Academy) Teachers and other adults in school model speaking clearly. This includes clear diction, reasoned argument, using imaginative and challenging language and use of Standard English. (S50, LA school) Standard English must be modelled at all times and pupil's speech corrected […]. Error posters to be in every classroom. These will contain common spoken and written grammatical errors on them as well as the correct form. They must be referred to regularly. (S102, LA school) Teachers here are constructed as powerful managers of language (Spolsky, 2009), who 'need to' and 'must' police students' language. Meso-level policy grants them a licence to do this where language does not 'conform' to standardised English, which is explicitly warranted by macro-level regulations found within the Teachers' Standards (DfE, 2013, p. 11). Some policies included discourses of language modelling and policing as wrapped up in wider behaviour requirements for staff and students, reinforcing indexical associations between language and discipline whereby 'speaking well' is a metaphor for 'behaving well' (Cameron, 2012). For example: Staff should model behaviourie no chewing gum, standard English should be used at all times, dress should be appropriate, mobile phones should not be used. (S18, LA school) Be consistently kind and friendly […] use formal standard English when interacting with staff, parents/carers and other pupils. (S216, LA school) We have the highest standards of uniform, inside and outside the academy. We speak in Standard English. We behave professionally inside and outside the academy. We have excellent manners. (S17, Academy) These are similar to the second excerpt which opened this article, notably the sentence 'the way that pupils speak to each other and to staff denotes their character'. In further reproducing the standard language ideology, policy here frames language as something which denotes human valuation and 'quality' (e.g. Irvine & Gal, 2000) through the process of enregisterment, whereby a linguistic repertoire becomes explicitly linked with social values and personality traits (Johnstone, 2016). The use of 'error posters' that S102 refers to above has been critiqued by others, including Cushing (2020aCushing ( , 2020b, where these material aspects of policy and language surveillance serve as a permanently visible reminder of classroom language laws and the 'undesirable' status of non-standardised English. Other policies in the corpus referred to these kinds of symbols, to 'reinforce what is counted as correct grammar' (S67), with further use of the policing metaphor in the use of 'word jails' (S203) in which 'slang' and other undesirable forms are incarcerated in. As illustrated in the examples below, the requirement for students to 'speak in full sentences' was particularly notable, with 16 policies in the corpus suggesting that this should be the expected practice for classroom discourse: Pupils are exposed to a rich oral language and are encouraged to answer questions in full sentences, using Standard English. (S80, LA school) Make students speak in full sentences and encourage them to use Standard English (not slang). (S122, Academy) Policies which 'encourage' or 'make' students to speak in 'full sentences' reinforce the conflation of written-spoken grammar which typifies macro-level policy and its reductive treatment of oracy. Jones (2017) provides a critique of macro-level policy initiatives which have systematically sidelined the place and value of spoken language in schools (including regional variation), with students having limited opportunities to explore the social aspects of interactive talk, in favour of curriculum content underpinned by formal presentations, recitation and the blanket use of standardised English. Ofsted have praised schools for having 'speak in full sentences' policies (Ofsted, 2013) as well as penalising schools for not doing so in various inspection reports (e.g. Ofsted, 2017, p. 2), and so it may be that schools are designing these in reaction to what they interpret as government requirements, under the promise of rewards and the threat of punishments.
Some schools were explicit in what they meant as non-compliance, listing various 'misconceptions' which listed non-standardised constructions as those which ought to be particularly policed, and further framing teachers as role models of 'accurate' and 'competent' language. For example: Reinforce the importance of accuracy in spoken or written languagefor example, correcting 'we was' in students' speech. (S03, Academy) To understand and use Standard English forms instead of local forms, e.g. 'We were … ' instead of 'We was … '; e.g. 'I was given … ' instead of 'I got given … '. (S40, LA school) Model accurate talk by addressing grammatical errors (would have, NOT would of), fillers (er, um etc.) and unnecessary repetition (like, OK) […]. Model competent speech and show them the difference between clarity and slang -(nah, that's wet init). (S259, Academy) Should my child speak grammatically? We are expected to teach children to speak and write 'Standard English'. This is the language of business that people use in the workplace. It is helpful if children know the correct past tense words: . I brought (not I brung); . I drew a picture (not I drawed) . We were singing (not we was singing) There are also some Leicester ways of talking that are non-standard, which we correct pupils on: . Can I go toilet? -Can I go to the toilet? (S165, LA school) The language of these policies reproduces deficit notions of non-standardised grammar, denying speakers of their right to their own linguistic resources and further perpetuating the standard language ideology. They reveal misconceptions about the grammaticality of non-standardised language, especially S165, where in asking the question 'should my child speak grammatically?' and listing a selection of 'incorrect' variations implies that non-standardised forms have ungrammatical status. Ironically, S165's 'Leicester ways of talking' policy works as linguistic bias against the language practices of the community it serves (and is created by), with policy targeting specific local forms of the city's dialect. Textually, these types of policy statements both create and legitimise language discrimination, where non-standardised usage is subordinated through its portrayal as 'deviant ' and 'non-compliant' (see Lippi-Green, 2012, p. 70).
As discussed in the preceding section, there are power dynamics at work here in terms of the trajectory of language ideologies from macro-to meso-level policies. Teachers are granted a licence to 'make' students speak in certain ways, and this has explicit support from political figures such as Michael Goveoften referred to as the architect of post-2010 educational reformswho has repeatedly championed schools where grammar is 'rigorously policed' (e.g. Gove, 2013). S165 justifies its suppressive policy as something that they as an institution are 'expected' to do by government, suggesting that their policy is driven, at least in part, by top-down pressures. The policing of was/were verb variation as found in the policies of S03, S40 and S165 link intertextually to macro-level policy, with the standardised form being a construction which is listed as a 'statutory requirement' which students must use (DfE, 2014, p. 77). Previous versions of the curriculum (e.g. DfEE, 1999, p. 45) have listed non-standardised constructions, including was/were, as ones for teachers to be particularly 'aware of'. Levey's (2012) work focuses on this variable in children's speech, highlighting its abundant use and making the argument that teaching children to simply substitute standardised variants for non-standardised ones is unlikely to raise standard language proficiency. Instead, students who experience these macro-meso policies are penalised for their 'failure' in having a language variety which happens to be different to the one that carries linguistic capital in school, whilst at the same time, the position of standardised English is bolstered through the structural suppression of non-standardised forms.

Conclusions, implications, recommendations
This article builds on work which is challenging language ideologies within current education policy in England (see Cushing, 2020a;Grainger, 2013;Snell, 2013). It expands on this by examining a large dataset and employing a range of tools from critical language policy, drawing attention to the importance of intertextuality and how new meanings emerge at the contact points between policy histories, layers and levels in order to interrogate policies of surveillance and compliance. It has examined textual manifestations of the standard language ideology in macro-and meso-level education language policies, arguing that deficit notions of 'correctness', 'accuracy' and 'appropriacy' are present across these policy levels, which serve to legitimise language policing and prejudice. Throughout these policies, standardised English is reified and concretised: uncritically and apolitically framed as an aspirational variety to which teachers and students should adhere with in order to comply with the discourse expectations of the school and broader society. Teachers are constructed as language role models and regulators of their students' language, granted power to police classroom discourse in ways which potentially marginalises and stigmatises speakers of non-standardised forms.
Language ideologies, policing and prejudice in schools need to be understood in terms of the intersections between coercion, power, agency and policy levelsthese issues represent a complex network and structure, rather than an individual 'event' or policy statement. Intertextual analyses of school-level policies revealed some of these structural processes at work, especially where policies were in textual contact with macro-level policies and other external policy drivers, such as national curricula, statutory tests, league tables, Ofsted and corporate policy templates. This work then adds to the argument that schools are under pressure to 'deliver', and in doing this performative work can often juggle a panoply of policy demands (Ball et al., 2012). Further work is required to understand how language policies get enacted and lived at 'ground level' in spaces throughout schools.
Finally, it is important that schools and teachers develop critical language awareness if they are to engage in policy making and pedagogies which resist some of the hegemonic language discourses found to be present in the policies analysed in this study. Many of the policies in the corpus toggled between what Baker-Bell (2020:, pp. 28-31) calls 'respectability' and 'eradicationist' approaches, in her work on African American language in schools. Respectability pedagogies seek to affirm and validate all forms of a language, whereas eradicationist pedagogies seek to remove and replace non-standardised forms with standardised equivalents. Although respectability policy-pedagogies may appear to be well-meaning and emancipatory, Baker-Bell is critical of their long-term goal which she sees as simply using non-mainstream repertoires as a 'bridge' to learn standardised English, and in doing so, fails to adequately challenge the standard language ideology. Janks (2010) notes that teachers face a paradox here: if they provide students 'access' to dominant forms (i.e. standardised English), this helps to maintain the dominance of these forms. But, if students are denied access, then any societal marginalisation they may face risks being perpetuated. Both Baker-Bell and Janks offer critical literacies as alternative pedagogies; as a way of not just 'teaching standard English', but teaching about standardised English, including its political status and its intersectionality with race, class and power. Similarly, Bacon (2017) suggests that: […] it is arguably more necessary to disrupt the SE [Standard English] myth among those who are socialized to believe themselves to speak "correctly." In neglecting to problematize the historical realities by which a dialect becomes the supposed standard, the decontextualized teaching of SE serves to institutionally legitimize the SE myth and the language prejudices embedded within this ideology. (Bacon, 2017) Central to his recommendations for how policymakers can mitigate the risk of language-based stigma and shame, Corson (1997) insists that teachers and students must become critically aware of language to resist the coercive power of the standard language ideology, including creating curriculum spaces where issues of language, ideology, power, class and race are explored. Despite efforts by critical linguists in the macrolevel policy-making process (see Cushing, 2020b), one of the reasons that the standard language ideology persists in education is that it benefits established hierarchies and dominant groups, and so there is perhaps little interest from macro-level policymakers in changing this. As Johnson and Johnson (2015, p. 223) write, the language ideologies of dominant groups can 'act as a template with which policymakers justify policies that restrict educational access and privilege particular ethnolinguistic groups'. Instead then, critical energy at the meso-and micro-levels of the policy-making process is perhaps one way of building teachers' own critical language skills and awareness. For whilst teachers are partly constrained by macro-level policy, they nevertheless retain a creative licence in how they appropriate and recontextualise policies for their own needs. Critical linguists collaborating with teachers in the unique context of their schools would be a useful activity in developing local-level policies which promote the use of speakers' whole linguistic repertoires rather than the simple 'grafting on' of standardised English, ensuring critical attention to the language ideological struggles which play out in staffrooms, corridors and classrooms. Notes 1. Local authority (LA) funded schools are under local council control and must follow the National Curriculum. Academy schools are run by private companies and do not have to follow the National Curriculum. Both types of school must enter students for national assessments, such as SATs, with teachers working under the guidelines of the Teachers' Standards (DfE, 2013). 2. Office for Standards in Education, Children's Services and Skills. Ofsted inspect schools and grade them in one of four ways: 'outstanding', 'good', 'requires improvement' or 'inadequate'. 3. One of these was TheSchoolBus.net, a company who produce generic, decontextualised policy templates and models which claim to ensure 'curriculum compliance'. Schools pay an annual subscription in order to access to an 'ever-growing list' of these.