The (white) ears of Ofsted: A raciolinguistic perspective on the listening practices of the schools inspectorate

Abstract England has had a schools inspectorate since 1839, first in the form of Her Majesty's Inspectorate (HMI), and since 1992, in the form of the Office for Standards in Education, Children's Services and Skills (Ofsted). The inspectorate, a workforce made up of a majority of white inspectors, conduct regular inspections of all state schools in England, producing reports which comment on various aspects of educational provision, including teachers’ and students’ spoken language. In this article we deploy a raciolinguistic genealogy to examine the listening practices of the inspectorate, drawing on historical inspection reports generated from archival work, inspectorate language policy, and a large corpus of contemporary reports. We show how raciolinguistic ideologies are deeply embedded into the sociopolitical culture of the inspectorate, and how these ideologies translate into systems of sonic surveillance in which the nonstandardised language practices of students and teachers are heard as impoverished, deficient, and unsuitable for school. (Raciolinguistics, schools, language policing, standardised English, Ofsted, England, social class, ideology)

The following section develops these theoretical stances, arguing that whilst it is generally accepted that the standard language ideology is a classed concept in England, this also needs to be understood through the prism of race. We suggest that a raciolinguistic perspective offers a suitable framework to do so. We then outline our methodological approach, describing how a raciolinguistic genealogy (Flores 2021) allows us to explore how standard language and raciolinguistic ideologies sit at the foundation of the inspectorate and continue to shape its contemporary work. Our data is in the form of historical inspection reports from the mid-1800s onwards, a large corpus of contemporary post-2000 inspection reports, and a cluster of key inspectorate policy documents. Our analysis attends to traces of racialised and classed language policing in this data, where we show how both students and teachers are subject to sonic surveillance at the ears of the inspectorate. Ultimately, our argument is that through a raciolinguistic perspective, we can understand the inspectorate's language policing as an intersectional classed and racialised practice which stigmatises bodies whose speech is heard to deviate from standardised English.

S T A N D A R D L A N G U A G E I D E O L O G Y , C L A S S , A N D R A C I O L I N G U I S T I C S
Language ideologies are a 'set of beliefs about language articulated by users as a rationalisation or justification of perceived language structure and use' (Silverstein 1979:193; see also Irvine & Gal 2000). These ideologies are institutionally enshrined, especially in spaces marked by imbalances in power, such as schools. One of the most pervasive language ideologies is the standard language ideology, whereby speakers and hearers come to believe in the existence of a 'standard' language that is 'drawn primarily from the spoken language of the upper middle class' and idealised as the sole 'correct' way of speaking ' (J. Milroy 1999;Crowley 2003;Lippi-Green 2012:67). Within England, the association between 'standard' language, class, and correctness can be traced back to the nineteenth century, which, importantly for this article, coincided with the creation of the inspectorate. It was here that the term 'standard' took on a new meaning, 'signifying a level of excellence to be reached and a quality to be emulated' within speech (Crowley 2003:112). This spoken 'standard' was defined in relation to the characteristics of a privileged group of speakers-the highest classes in London and those who attended prestigious public schools-and was thus 'iconized' (in Irvine & Gal's 2000 terms) as emblematic of educatedness, civility, and superior moral character (Crowley 2003:117; see also Bailey 1997;L. Milroy 1999). Consequently, regional dialects were stigmatised as ignorant, sloppy, and impure through their association with lower class speakers. These ideologies were held not just by privileged groups in society but also by marginalised groups, who could be coerced into accepting the norms of the powerful and regulating their own behaviour accordingly. Language in Society 52:3 (2023) 365 Within public schools, conformity to a set of spoken norms was achieved largely through peer-group pressure (Mugglestone 2003:231); but, within the emergent state system, it was the role of the schoolteacher to intervene. George Sampson, a member of the committee which produced the influential Newbolt Report of 1921, would later explicitly pathologise nonstandardised speech in his declaration that 'the elementary schoolchild began his education with his language in a state of disease, and it was the business of the teacher to purify and disinfect that language ' (1924:28). The result of these attempts at linguistic purification and disinfection worked to enshrine the standard language ideology and create cultures where language shaming and stigmatisation were a normalised part of school life. Teachers felt the repercussions of this too. The requirement that they adopt the ideals of 'standard' speech was underlined in teaching manuals, training colleges, and inspectors' reports (including those in British colonies) where '[r]egional accents, connotative of 'ignorant' rather than 'educated', were increasingly considered incompatible with the office of school teacher' (Mugglestone 2003:243-44).
Awareness of class-based linguistic differences developed apace in the twentieth century, fueled by hostile class relations (L. Milroy 1999:188;Crowley 2003), and class stereotypes continue to dominate linguistic evaluations in the UK today, including in educational policy and practice (Mac Ruairc 2011;Grainger 2013;Snell 2013Snell , 2015Lampropoulou & Cooper 2021). Lesley Milroy compares this with the US, where the standard language ideology has focused on divisions of race rather than class (e.g. Bonfiglio 2002), with African American Vernacular English being 'the most stigmatised linguistic code of all' (L. Milroy 1999:196). US scholars such as Alim & Smitherman (2012) and Baker-Bell (2020) use the term 'White Mainstream English' rather than standardised English to foreground 'how white ways of speaking become the invisible-or better, inaudible-norm' (Baker-Bell 2020:3). Their work within the area of raciolinguistics has shown how the language of racially minoritised speakers in schools is routinely stigmatised for failing to adhere to this white 'norm ' (e.g. Smitherman 2017;Rosa 2018;Baker-Bell 2020). Moreover, these speakers continue to be stigmatised even when using language that would be considered 'standard' when produced by privileged white speakers (Alim 2007;Flores & Rosa 2015:150). Flores & Rosa (2015:150) explain this phenomenon through a theory of racialised language perception, in which raciolinguistic ideologies position racialised bodies as linguistically deficient unrelated to any objective linguistic practice. As with the standard language ideology, they situate raciolinguistic ideologies within the broader history of European colonialism, where racialised speakers were discursively constructed as biologically and linguistically inferior to the European bourgeois subject as part of nation state=colonial governmentality (Rosa & Flores 2017a:623). They argue that the ongoing significance of colonial distinctions between Europeanness and non-Europeanness-and by extension whiteness and nonwhiteness-makes race 'a crucial, indeed global, category of analysis' (2017a:635), and call for greater attention to 'raciolinguistic patterns and particular-ities… across the modern world' (2017a:622).

366
Language in Society 52:3 (2023) Little attention has been given to race in discussions of the standard language ideology within education in the UK context. We look to address this by incorporating a raciolinguistic perspective into our critique of the inspectorate, treating race and class as intersectional axes of discrimination. Virdee (2014:8) traces the consolidation of working-class racism to the 1830s and 1840s (with the formation of the schools inspectorate in 1839), as a structuring force in which race was 'constitutive in the making, unmaking and remaking of the working class in England' and in the economic exploitation produced by global capitalism. In the words of Stuart Hall and colleagues, race is 'the modality in which class is lived … the medium in which class relations are experienced' (Hall, Critcher, Jefferson, Clarke, & Roberts 1978:394). Interrogating this tripartite co-naturalisation of language, race, and class is key to the raciolinguistic perspective we take in this article, and how racialised bodies are affected by the 'interlocking systems and structures of linguicism, racism, and classism, which are interrelated and continuously shaping one another' (Baker-Bell 2020:16). We therefore treat standardised English as a racialised as well as a classed concept, one that 'plays a powerful role in maintaining white hegemony' through its association with whiteness and colonial power (Von Esch, Motha, & Kubota 2020:397; see also Pennycook 1998;Chow 2014), and consider the inspectorate to be a state sanctioned tool of enshrining oppressive language ideologies in schools.
Crucial to the raciolinguistic perspective we adopt is the shift in analytic attention away from the speaking practices of marginalised communities and towards the hearing practices of the white listening subject. Rosa & Flores (2017a:627-38) extend their theorisation of the white listening subject to include 'whiteness as an historical and contemporary subject position' that can be inhabited both by 'individuals recognised as white and nonwhite' and by 'nonhuman entities'. Such entities include institutions, policies, and various technologies of surveillance, and we consider the inspectorate and their role as a state listening subject (Pak 2021) as part of this assemblage. By focusing on the listening practice of the inspectorate rather than the speaking practices of individual teachers and students we relocate the problem within state surveillance apparatus and underscore the need for structural change. As Flores & Rosa (2015:167) point out, demanding that individual speakers modify their language practices will not lead to social transformation because 'they are still heard as deficient language users' and 'the underlying racism and monoglossic language ideologies of the white listening subject' remain unchallenged.

I N S P E C T I N G T H E I N S P E C T O R A T E : T R A C I N G I N S T I T U T I O N A L L A N G U A G E I D E O L O G I E S
We interrogate the listening practices of the inspectorate through a raciolinguistic genealogy (Flores 2021), as a way of exploring the ongoing colonial logics of standardised English and how the present is shaped by the past. Flores proposes a Language in Society 52:3 (2023) 367 raciolinguistic genealogy as a method which 'situates contemporary language ideologies and the policies that they work to create within the broader colonial histories that have sorted populations into those deemed fully human (white) and those deemed not fully human (racialized) ' (2021:112). There are three guiding principles to this method. The first, a genealogical stance, rejects a simple chronology in favour of 'examining the grid of intelligibility that produces normative subject positions, or ways of being in the world, to which all are expected to aspire ' (2021:113). Grids of intelligibility (Foucault 1978) describe how language produces conditions in which certain bodies and ways of talking are de=legitimised and how social stratification is both constructed by and through language. The second, a materialist framing, is concerned with locating discourses of language standardisation within broader sociopolitical and economic networks of power, highlighting the role of racial capitalism, Western colonialism, and the exploitation of marginalised communities by white supremacy. The third, a raciolinguistic perspective, places a focus on language ideologies which have 'co-constructed language and race in ways that position racialized populations as inferior to the normative white subject' and 'seeks to denaturalize these raciolinguistic ideologies in the hope of developing spaces of resistance that produce a new grid of intelligibility' (Flores 2021:114). Through a raciolinguistic genealogy which focuses on the listening practices of the inspectorate then, we can trace its role in the construction of governable subjects and the maintenance of white linguistic supremacy in terms of the language ideologies it (re)produces.
Given that genealogies are concerned with how contemporary policies embody continuities with the past, this article draws on a range of data spanning multiple decades. 'Historical' data were generated from archival work at the National Archives in London, which hold elementary and primary school inspectorate reports , secondary school inspectorate reports , and reports from teacher education colleges . We read through approximately 350 reports from the archives, focusing on sections concerned with spoken language and looking for textual traces of language ideologies. 'Contemporary' data were generated through the construction of a corpus of post-2000 Ofsted primary and secondary school inspection reports, which are available on Ofsted's website. We hired a computer programmer to scrape all primary and secondary reports from the Ofsted website, resulting in a corpus of 102,592 reports. Due to the huge size of this corpus, a subcorpus was created by randomly selecting 3,000 reports to make the data manageable for the purposes of this particular article. Future research using the entire corpus is planned. All reports from the subcorpus were imported into the LancsBox software, allowing us to excavate embedded language ideologies using tools such as word searches and collocation patterns (see Vessey 2017). These searches sought to identify the textual traces of language policing by locating clusters of words and phrases deemed to be representative of judgements about language, including where these exhibited similarities with patterns found in the archive data. We focused on metalinguistic tokens such as 368 Language in Society 52:3 (2023) 'non=Standard English', 'errors', 'accent', 'in=correct grammar', and 'full sentences', as well as how words denoting speech were modified, such as 'speak [ing] clearly', 'appropriate speech', 'accurate speech', and 'talk un=grammatically'. We also collected key Ofsted policy documents, including guidance for inspectors, research reports, and subject reviews.
Finally, after publicly raising some of our concerns about Ofsted's language ideologies on social media, we had two private meetings with Ofsted representatives. The ethics granted for this research do not extend to describing the content of these meetings, but we can confirm that we as researchers decided not to continue our engagement with Ofsted due to what we felt was-and appropriately given the topic of this article-its institutional inability to listen.
England has had a schools inspectorate of some sort since 1839. Dunford (1998) traces the gradual growth of HMI in its first 100 years, noting the privileged positions of its inspectors: male, white, middle-upper class, and representatives of the church. HMI existed until its absorption into the newly created Ofsted under John Major's Conservative government in 1992. Major's ideologies about education were grounded in white epistemologies and 'colour-blind' language policies (see Gillborn 2005), in which he regularly belittled so-called 'progressive' views on language, class, gender, and race as the antithesis to educational development.
When it comes to education, my critics say I'm 'old fashioned'. Old fashioned? Reading and writing? Old fashioned? Spelling and sums? Great literature-and standard English grammar? … Well, if I'm old-fashioned, so be it. … I also want reform of teacher training. Let us return to basic subject teaching, not course in the theory of education. Primary teachers should learn to teach children how to read, not waste their time on the politics of race, gender and class. (Major 1992) Ofsted have attracted criticism for pressuring schools into enacting certain policies and pedagogies (e.g. Cullingford 1999) and for working as a surveillance mechanism of the state (Page 2017), but the inspection of spoken language has received little attention beyond Mugglestone (2003), Cushing (2022), and Cushing & Snell (2023). According to Mugglestone, it is likely that HMI's cohort of inspectors, who were generally educated at institutions reserved for the white middle-classes (e.g. public schools and then Oxford or Cambridge), would have internalised listening practices that favoured 'standard' speech. She describes how inspectors recommended John Walker's highly prescriptive 1791 Critical Pronouncing Dictionary to schools, and how they made use of manuals such as Daniel Fearon's 1876 School Inspection, which instructs inspectors to listen out for 'provincialisms' and recommends that teacher education courses eradicate regional variations of speech before teachers are allowed into classrooms. Mugglestone notes the social class bias in inspectors' judgements of apparent deficiencies in the language practices of children, for instance: 'the great fault in the lower classes is want of real correctness, good Language in Society 52:3 (2023) 369 articulation, and mastery of the pronunciation ' (2003:251); 'a master should … take pains to correct their incorrect pronunciation ' (2003:243).
We argue that the inspectorate's early work must also be understood in terms of race and racialisation. This is especially so given that the early to mid-1800s was a period marked by aggressive British colonial expansion and racial stratification, as well as a growth in technologies of linguistic standardisation, such as dictionaries, grammars, and elocution manuals, with the concretisation of ideologies pertaining to the idealised native speaker (e.g. Hackert 2012;Stoever 2016;Heller & McElhinny 2017). Four years prior to the formation of the inspectorate, Thomas Macaulay's 1835 Minute on Education had grounded raciolinguistic ideologies in British colonial education policies, and in 1854, the British expanded the role of the schools inspectorate to various colonised territories, as a broader mechanism of its colonial education strategies which continue to resonate today (see Fletcher 1982;Proctor 2014).
Data generated from historical reports through our archival work shows that language policing was a foundational feature of the inspectorate's practice. This work revealed an overwhelming number of evaluative judgements about speech. In the examples we list below and in the remainder of this article, we include the year of the inspection report in parentheses. Inspectors praised schools for paying attention to 'impressive habits of pronunciation' (1864) and 'appropriate articulation' (1910), but criticised schools where it was deemed they were failing to pay due attention to deviant speech, as the following extracts show: 'many of the children are illiterate in regard to patterns of speech' (1867); 'the less able children show particular weaknesses in speaking and proper articulation ' (1947); 'the speech of many children is of insufficient worth' (1949); 'much remains to be done in the cultivation of pleasing intonation and clear speech ' (1950); and 'the children do not speak easily and quite a number are inarticulate but these deficiencies are recognised' (1951). A 1920 report described students' speech as being 'weak', 'ill-constructed', and 'ungrammatical', deploying discourses of poverty and ignorance.
With the exception of two or three of the senior pupils … far too general weakness was shown in the pupils' command of their own spoken language which is … ill-constructed and often ungrammatical, and further there is evidence of poverty of vocabulary and ignorance of the meaning of words used. (1920) These discourses of verbal deprivation also extended to teacher education programmes, such as in a 1931 report of Dudley Training College which described teachers' speech as 'not always clear or elegant' and 'portraying a regrettable poverty of background'. In Extracts from the Reports of Her Majesty's Inspectors of Schools (Privy Council Committee on Education 1852), teachers are described as having 'little knowledge of language, except under its colloquial forms' with students' speech being described as 'slovenly' and 'incomprehensible'.
Post WWI, state-produced manifestos such as the Newbolt Report (Board of Education 1921) framed standardised English as a tool for rebuilding national unity.

370
Language in Society 52:3 (2023) Here, contributions from the HMI's Divisional Inspector for London were geared around 'undesirable' and 'vulgar' speech, illustrating the same deficit-orientated listening practices which characterised the inspectorate's foundational work of the 1800s. they have gone some way towards getting rid of undesirable forms of London speech. … It is lamentable, in a great number of schools, to hear the children habitually mispronouncing words, or mumbling rather than pronouncing them, while their teachers, who may show great concern at inaccuracies where the written word is concerned, seem to accept a pitiably low standard of speech as a thing which must be taken for granted, and scarcely calls for comment. It is emphatically the business of the Elementary School to teach all its pupils who either speak a definite dialect or whose speech is disfigured by vulgarisms, to speak standard English, and to speak it clearly, and with expression. (Board of Education 1921:65) As racial and linguistic heterogeneity in England's schools grew in the decades following WWII amidst the contexts of mass im=migration, so too did the inspectorate's concerns around the maintenance of a standardised language (see Department of Education and Science 1965). These anxieties were reflective of white, post-war attitudes to the speech of racialised communities in society more broadly. For instance, Fryer (2018:280) describes white visual and audible perceptions of West Indian migrants to England in the 1950s, which framed them as linguistically and biologically inferior.
They saw them as heathens who practised head-hunting, cannibalism, infanticide, polygamy, and 'black magic'. They saw them as uncivilized, backward people, inherently inferior to Europeans, living in primitive mud huts 'in the bush', wearing few clothes, eating strange foods, and suffering from unpleasant diseases. They saw them as ignorant and illiterate, speaking strange languages, and lacking proper education. (Fryer 2018:380) Modes of perception shaped by prescriptive ideologies are further evidenced by two major HMI surveys of primary and secondary schools in 1978 and 1979 respectively (Department of Education andScience 1978, 1979). Department of Education and Science (1978) makes no explicit reference to 'Standard English', but frames 'desirable' speech with a cluster of labels which construct an idealised speaker -'precision', 'accuracy', 'competence', 'clarity', and 'fluency'-positioning teachers as language police who have a licence to make subjective judgements on audible language 'quality'. Department of Education and Science (1979) is more explicit in how teachers should police their students' language, praising schools where standardised English is normalised and describing how 'the best' teachers would lead their pupils to a 'surer command of language itself'.
Most of the language of classroom talk and of textbooks was in standard English, and it was a part of the concern of teachers to help pupils to acquire this form of English through talking as well as through the related activities of reading and writing. The best teachers were sensitive to differences in language and led their pupils discreetly and by a variety of means towards a wider range of language use and a surer command of language itself. (Department of Education and Science 1979:99) It is not fully clear what 'sensitive to differences in language' refers to here, but given how standardised English is equated with having a 'command' over Language in Society 52:3 (2023) 371 language, the implications are that these 'differences' are references to racial and class heterogeneity, which requires homogenising and erasing by teachers via a 'variety of means'. As well as constructing ideologies of a uniform and empirically audible standardised English, the 1979 Department of Education and Science document continues to criticise teachers who had 'adopted features of the language of pupils' and the 'superficial' nature of this, whilst commending a 'posh' talking teacher who was 'anxious to maintain linguistic standards ' (1979:99).
These HMI surveys were published between the Bullock Report (Department of Education and Science 1975) and the Rampton Report (Department of Education and Science 1981), both of which argued that a child's home language practices should be accorded value in school (although we reject 'home' and 'school' as a dichotomy), whilst acknowledging how institutional racism leads to Black children's underachievement. Between 1984Between -1989 and framing the introduction of the first nationally imposed curriculum in 1988, HMI published a series of seventeen booklets under the heading 'Curriculum matters'. One of these, English from 5 to 16, includes a lengthy passage on 'Standard English', which includes the assertion that it 'must be central to the work of all English teachers', and not to do so would be to 'deny some of their rights as citizens' (Department of Education and Science 1984:36). In the same document, HMI also claimed that no accent is 'inherently superior to any other' and that children's home language practices should not be criticised, belittled or proscribed ' (1984:15). However, this is negated not just in our own discussion of HMI reports so far, but in the same document by HMI themselves, in their claim that 'pupils should learn to speak clearly and intelligibly; and if their accent is difficult for those outside their speech community to understand, they should be able to modify it when necessary ' (1984:15). HMI again locate the problem within the speaker and require them to modify their language, whereas a raciolinguistic perspective suggests that the problem is located within the inability of the white listening subject to hear minoritised bodies speaking appropriately (Rosa & Flores 2017b:284). We continue our discussion of these ideologies in the following section, where we turn our attention to post-2010 education policies. students as 'workers' and governable subjects whilst perpetuating the meritocratic ideology that economic privilege relies on a capacity to use standardised language. The new government explicitly instructed Ofsted to police speech in its inspections, with a 2011 training document describing how 'inspectors might expect to see a more formal approach to correcting basic errors extended to pupils' speech' (Ofsted 2011a:33). There was no overt mention of 'Standard English' or oral corrections in the 2013 inspection framework (Ofsed 2013a), and so the policing of language was not part of its de jure remit during the lifespan of this policy, but was nevertheless still enacted by inspectors and used as a factor in judging schools, which we return to in the sections that follow. Between 2011-2021, Ofsted published six reports under the label of 'Languages and literacy'. Across these, standardised English is a recurring theme, such as in Ofsted (2011b) which drew together deficit perspectives of classed and racialised language (using proxies such as 'inner-city' and 'free school meals' 1 ) whilst commending schools for any presences of sonic surveillance, including where students were selfpolicing their speech.
In another inner-city school, serving a high proportion of pupils known to be eligible for free school meals, staff paid close attention to the difference between standard and non-standard English in spoken language. Pupils were quick to correct themselves when they used words such as 'ain't' and 'gotten' in their speech when responding to questions from teachers. They explained to inspectors how teachers and assistants taught them to use standard English by reminding them constantly during conversations and in lessons. (Ofsted 2011b:22) A 2013 research report namechecks the 'correct use of standard English' clause of the Teachers' Standards (Department of Education 2011) as a mechanism by which school leaders should 'monitor rigorously how effectively teachers are developing pupils' literacy skills' (Ofsted 2013b:8) and provides examples of how teachers insisted on 'full sentences' in spoken discourse. This example comes from a school in Hackney, East London, who in 2013 imposed a policy where 'colloquial language' was 'banned'. Around the same time, media reports described how an Ofsted inspector had told a teacher to modify her accent and 'sound less northern' (Garner 2013). Whilst Ofsted's response was that such comments 'should form no part' of its inspection activities, our data shows that instances like this cannot simply be dismissed as individual acts of language policing, but represent a larger, structuring logic which has long shaped the inspectorate's mode of perception.
The latest version of the inspection framework includes no direct references to 'Standard English', but is accompanied by an overview of research (Ofsted 2019b) which includes multiple traces of standard language and raciolinguistic ideologies. This includes detail of how inspectors will be listening for a 'language-rich environment' (2019b:18) through encouraging schools to address so-called 'word gaps': a construct based on widely critiqued research from Hart & Risley (1995) which frames the home language practices of low-income, racialised speakers as unsuitable for school. Others absences, and missing parts are underpinned by a raciolinguistic ideology, and we argue they also represent a continuation of the foundational work of the inspectorate, but rebranded under a benevolent guise of what Ofsted call 'research-informed' policy. The 2019 inspection framework is also accompanied by an 'equality, diversity, and inclusion' statement which describes the work of the inspectorate as having 'due regard to the need to eliminate discrimination' (Ofsted 2019c:3). Despite this, our analysis suggests that the inspectorate is failing here in regard to discrimination on the grounds of language.

T H E S O N I C S U R V E I L L A N C E O F T E A C H E R S
We now focus on data generated from the corpus of school-level inspection reports, beginning by examining the representation of teachers. Searches revealed how the inspectorate commend teachers who enact aggressive listening practices, a discursive pattern which works to normalise and legitimise the policing of speech. This is shown in the following examples. These descriptions of language policing are congratulatory in nature, drawing up images of combat and conflict in the 'struggle' against nonstandardised practices, echoed in some earlier reports where teachers were commended for their 'fight against the bad standards of speech ' (1943). They also contribute to inspection reports rated as 'good' or 'outstanding'-indeed, where the inspectorate did hear teachers policing the language of students, this was extoled and taken as an indicator of high-quality teaching.
Our searches also revealed how the inspectorate commended practices which sought to erase nonstandardised features completely. For example, a 2013 report included how 'in the best lessons, teachers reference the need for standard English and students are provided with a list of 'banned' words, to remind them', whilst a 2021 report described how 'misconceptions' arise 'when pupils use non-standard forms of grammar in their speaking and writing in lessons'. In a London school serving a majority low-income community of students from racially minoritised backgrounds, the efficacy of a whole-school language policy was praised for 'almost eradicating' the word 'like', despite this simply being a feature of spontaneous speech (see also Snell 2013Snell , 2015Snell , 2018aCushing 2020;Cushing & Ahmed 2021 for an extended critique on whole-school language bans, including their classed and racialised nature).
In all lessons, in all subjects, students are challenged to write and speak accurately, without slang or colloquial language, and at length. The academy's progress in almost eradicating the use of 'like' as a sentence connective by students is one example of this hugely successful focus. (2014) 374 Language in Society 52:3 (2023) As has long been the case in the inspectorate's policies, teachers are constructed as linguistic role-models who are under state-level pressure to both use and promote standardised English, and who occupy positions where they can enact language policing in hostile ways. One of the key criteria the inspectorate use in its work is government-issued standards for teachers, which have historically included references to 'Standard English'. For example, the standards designed by Labour and used in their tenure between 1997-2010 (Department of Education and Employment 1997) are underpinned by the standard language ideology, with teachers required 'to communicate clearly and grammatically in spoken and written standard English ' (1997:14, 43) and ensure that 'pupils have ample opportunities to listen to well-spoken, standard English ' (1997:22). The standards used within post-2010 reforms (Department of Education 2011:11) include the requirement that teachers 'take responsibility for promoting high standards of literacy, articulacy and the correct use of standard English'. There were numerous intertextual traces of these clauses within the corpus, with teachers praised for upholding and modelling standards in line with state-mandated expectations.
Some examples brought together teachers' modelling of standardised English with ideologies of correctness, such as 'teachers model standard English well and readily correct incorrect use of English' (2016), and 'teachers model standard English continuously and help pupils to communicate and enunciate correctly' (2019). Within the institutional logics of Ofsted, good role models are good language police, with praise handed to teachers when they were heard to police 'slippages' and 'lapses' into the local dialect. One 2015 report described the good practice of teachers in which they would 'paraphrase speaking in standard English when pupils lapse into the local dialect', whilst a 2002 report described how 'teachers have brought about considerable improvements in the vocabularies used by their pupils although the local dialect is still a powerful influence on the spoken word'. Another 2002 report highlighted how teachers 'made pupils aware' of standardised English when they 'revert to colloquial expression'. A 2018 report of a Bradford school with a community of primarily Pakistani and Roma students from low-income backgrounds was praised for how 'adults teach children to speak in standard English' and 'speak clearly and encourage children to repeat what they say'. These examples conjure images of classrooms where the policing of spoken language is normalised and highlighted as good practice by the inspectorateespecially for schools in areas of economic disadvantage serving largely racialised students.
So far, we have described how the inspectorate approve of teachers who reinforce standard language and raciolinguistic ideologies and attempt to erase any audible traces of nonstandardised patterns. Conversely, where teachers failed to police the language of students, this was construed as poor practice, with schools punished through negative comments and gradings in reports. For example, a 2003 report of a school in Birmingham listed speech as a factor which 'required improvement', in how there was a 'significant proportion of children who do not know or use Standard English', and where 'there is insufficient follow-up to brief, grammatically incorrect or otherwise imperfect spoken responses'. This 'significant proportion' were multilingual, racialised students from low-income backgrounds, with the report stating that 90% of pupils were from minority ethnic backgrounds and speak English as an additional language, with an above average number living in poverty. A 2009 report of the same school described the speech of students as 'limited', whilst a 2017 report of the school praised teachers for ensuring children 'speak using full sentences and standard English', declaring that 'this is why children overcome their communication difficulties'. A 2014 report of a different school stated that students were not given 'enough opportunities to articulate their learning', and when they were, 'teachers fail to help them speak correctly and in standard English'. Ironically, if pupils are routinely corrected on the way they speak when they do offer classroom contributions, then this will curtail opportunities to 'articulate their learning', as students are publicly shamed for the way they speak and made to feel conscious about their own language (Snell 2013(Snell , 2019Snell & Cushing 2022). Another report described how if adults 'do not model correct spoken English' then this 'reinforces misconceptions' and leads to 'uncertain explanations', and to remedy this, teachers needed to self-surveil their language. not all teachers are careful enough about grammar and spoken English and this can lead to uncertain explanations … They do not always pick up enough on pupils' errors so that pupils can improve. For example, one teacher failed to correct a pupil, when he started a sentence, 'Me and James went…'. (2001) In a 2019 report, Ofsted instructed management that they 'must ensure that all staff have the knowledge, skills and understanding they need to model the correct use of standard English', placing pressure on senior leaders to introduce school-level policies underpinned by standard language ideologies as determined by the state (see Cushing 2021). Another report noted the audible absence of language policing, writing of how 'only rarely … do teachers correct pupils' oral use of standard English or model spoken English to emulate ' (2003), representing a deficit perspective in which teachers' nonstandardised language patterns were deemed to have a damaging effect on the development of students. In a 2003 report of a London school in an area of high economic deprivation and serving largely Black African and Black Caribbean students, many of whom were recent immigrants, teachers were criticised for not 'drawing attention' to the presence of nonstandardised 'we done' and 'we was', whilst 'a very good lesson, involving Year 1 pupils with speech and language difficulties, focused on grammatically correct sentences'.
This framing by the inspectorate shifts responsibility onto teachers for policing both the language of themselves and minoritised others, adding to institutional pressures concerned with linguistic performance which are historically embedded in England's schools. The 'ungrammaticality' of teachers' language was a widespread 376 Language in Society 52:3 (2023) concern in reports, reflecting the failure of inspectors to recognise teachers' linguistic dexterity. One 2018 report included how teachers were not 'careful' enough to use standardised English, with three further examples revealing inspectors' deficitorientated descriptions of teachers' language which equates 'unskilled speaking' with nonstandardised patterns.
Teachers do not generally demonstrate a wide range of vocabulary and sometimes talk ungrammatically. For example, one teacher was heard to say, "I don't want no more shouting out". As a result, the pupils make slow progress in their learning of speaking skills. (2003) At times, adults use slang and are satisfied with one-word replies to their questions, rather than helping children to extend words into full sentences. (2016) Some teachers model incorrect grammar in their spoken English. (2018) Through both praise and punishment, teachers are represented in inspection reports as school-level agents of language policing who surveil language on behalf of the state. The following section examines how students' language is heard by the inspectorate.

T H E S O N I C S U R V E I L L A N C E O F S T U D E N T S
The speech of students is also placed under sonic surveillance, and in this section we show how the inspectorate described students' speech along a continuum of judgements from good, to bad, to a complete absence. The inspectorate offered praise where speech was heard to conform with standardised English, and criticism where it was deemed to deviate. Where students were heard to speak in standardised English, this was typically taken to be an indicator of quality, with such comments working as a process of iconization (Irvine & Gal 2000:37-39), where language practices become discursively associated with various social factors. These relations included high-performing academic ability, intelligence, confidence, maturity, complexity, neatness, a willingness to learn, and good manners. For example, reports described how 'more able pupils use standard English fluently' (2001); how 'higher attaining pupils are very articulate: they have a … good command of standard spoken English, and they respond willingly to questions ' (2004), and how 'children quickly learn to speak in whole sentences, which are increasingly grammatically correct' (2017).
Contrastingly, students who were heard not to use standardised English were equated with low academic ability, weakness, incompleteness, intellectual inferiority, and as articulating language practices not suitable for school. For example, a 2004 report of a school in an economically deprived area of Manchester concluded that 'their speaking skills are poor; many have a limited vocabulary and weak grammar, and some have difficulties with articulation'. 'Successful' and 'unsuccessful' students were sometimes directly contrasted against each other based on how 'articulate' and 'clear' inspectors had rated them, such as in a 2002 report Language in Society 52:3 (2023) 377 which described how 'a small minority of pupils are highly articulate, expressing their thoughts very clearly and persuasively', whilst 'a substantial minority speak briefly, often with insufficient attention to accuracy in standard English'. A 2001 report described 'unsatisfactory' teaching as being marked by 'incorrectly spoken words such as, 'twenny' for 'twenty'', whereas 'better' teaching happened in 'clear communication'. The inspectorate's criticisms of students were located within their unsuitable language practices, directing blame towards individuals who had been deemed to fail to meet the linguistic requirements needed for school. This was often done in a way which dichotomised standardised English with 'colloquial' English, with the former being framed as 'correct' and 'grammatical', and the latter being framed as 'incorrect' or 'ungrammatical'. For example, 'pupils have … insufficient understanding of the difference between their own colloquial English and grammatically correct Standard English' (2004), and 'because of their use of local dialect, also have problems in using appropriate Standard English … this lowers their attainment ' (2002).
In similar ways to how teachers were encouraged to self-police their own language practices, students were also encouraged to police the language of their peers and teachers, some of which was done with apparent 'hilarity'.
A telling example is the 'grammar police', when pupils take turns to signal when classmates and teachers speak ungrammatically or use banned dialect words such as 'yous'. This is done with some hilarity, but underlies an earnest desire to 'speak properly'. (2014) The inspectorate praised various classroom activities where students substituted nonstandardised language with standardised equivalents, with many of these activities including proxies for race. For example, 'writing a rap song into standard English' (2002), 'translating a poem written in patois into standard English' (2013), and '[turning] non-standard English into standard English when writing a letter ' (2003). Activities such as these, especially when praised by inspectors, enshrine raciolinguistic ideologies in which unsuitable language practices are not just criticised, but erased and replaced (Irvine & Gal 2000). Where students' home languages were valued, these were typically framed within additive approaches which emphasise racialised discourses of appropriateness (Flores & Rosa 2015), whereby standardised forms are heard as more appropriate for school, and the converse for nonstandardised forms. Yet, white middle-class speakers are often afforded greater flexibility in this regard, able to deviate from forms idealized as appropriate without censure, 'while racialized people can adhere to these idealized linguistic practices and still face profound institutional exclusion' (Flores & Rosa 2015:165).
In addition to hearing speech as being poor quality, reports also framed students' language as being deprived, restricted, and in some cases, entirely absent. These discourses of language lacunae and impoverishment have a long history in deficit ideologies of minoritised speakers' language (e.g. Johnson et al. 2017), in which linguistic features that are not heard to align with white, middle-class communication patterns are evaluated as impeding academic development. For example,

378
Language in Society 52:3 (2023) 'speaking skills are restricted by the poor command of standard English … most pupils have very poor language skills on entry' (2019), 'language and communication skills are underdeveloped' (2017), and 'many pupils have a poor command of standard English and have only a limited range of vocabulary and advanced speech structures ' (2000). One 2003 report claimed that whilst 'many lively speakers are apparent', a 'lack of standard forms of English is still evident and general vocabulary often remains restricted'. Some of these descriptions of students' language were particularly hostile, such as a 2005 report which focused on gaps, absences, and restrictions.
[P]upils' speech is poorly constructed and very ungrammatical. They use a limited range of vocabulary and regular employ 'thingy' and 'whatsit' in the gaps for words that they would like to use, but cannot recall (or do not know). They mimic the speech pattern they hear and employ slang and colloquialisms as if this is the only way to speak. (2005) A 2004 report of a school in an economically deprived area of Hampshire went to some length in its deficit descriptions of language, as well as conflating speech and writing and illustrating a misunderstanding of what language variation is and how it works.
Throughout the school, one of the main barriers to pupils' achievement is their lack of descriptive vocabulary either verbally or in written form. Many pupils struggle to answer questions in full sentences and often revert to phrases, one-word answers or gestures. The spoken English of the majority of pupils does not conform to Standard English with words like 'of ' and 'have'; 'was' and 'were'; 'is' and 'are' being inter-used. This, together with pupils' limited use of descriptive English is hindering their creative efforts. Many pupils do not have the confidence to move into an imagined world because they feel they do not have the language to support their creative ideas. Too frequently, this results in pupils taking the simple alternative. (2004) In this report, the students in question are primarily working-class, with demographic information describing how 'a large percentage' are entitled to free school meals, and how the school serves 'very few families from a professional background'. For these children living in poverty and from working-class homes, their language practices are described by the inspectorate as limited, deficient, and basic. Their apparent failure to conform with standardised patterns is perceived to be 'barriers' to achievement, a 'hinderance' to creativity, and obstacles to the construction of imagined worlds. These deficiencies are not just framed by the inspectorate as 'poor' or 'restricted' language', but a complete lack of language at all, in what Rosa (2016) describes as an ideology of languagelessness, which we return to in the following section. Similarly, a 2000 report of a school in one of the most economically deprived areas of Manchester and serving a majority racialised community of students (where, according to the inspectorate, 'most come from homes where Patois or Creole are spoken') drew direct links between low academic ability and the presence of classed and racialised features of speech.
By the age of eleven, many pupils have fallen behind, and are not achieving satisfactorily, particularly the boys. The more able pupils are mainly speaking Standard English in school, with sound Language in Society 52:3 (2023) 379 pronunciation and good sense. A few pupils lack clarity in their speech which results in some confusion in the way they say 't ' and 'th', as 'd' or 'f'. (2000) We end this section by re-drawing attention to the fact that the textual traces of language policing we have uncovered in the corpus of contemporary reports should be seen as a direct genealogy from their foundational work and has been part of its institutional culture for almost 200 years. These ideologies are historically embedded as a normalised aspect of the inspectorate's activity in schools, with racialised students living in poverty particularly likely to be targets of hostile language policing.

T H E R A C I O L I N G U I S T I C L O G I C S O F T H E I N S P E C T O R A T E
The inspectorate claims that it 'report [s] what they see, rather than judg[ing] against pre-set criteria' (Lee & Fitz 1997:43). However, in this article we have shown that the inspectorate does not just report on what it sees, but what it HEARS, making auditory judgements against the imagined criteria of so-called 'Standard English' and how far speakers are heard to deviate from the language patterns of the white speaking subject. The raciolinguistic genealogy we have deployed in this article (Flores 2021) reveals how since its establishment as HMI in 1839, the inspectorate must be understood as state sponsored language police. Their work has enshrined raciolinguistic and standard language ideologies in schools, in which the social and colonial construct of 'Standard English' is heard as the exclusively legitimate code of the classroom. We have shown how the inspectorate's ideologies about language, filtered through predominantly white and economically privileged ears, work to normalise school cultures where certain bodies and ways of talking are commended, whilst others are silenced-with those that speak in 'Standard English' constructed as normative and an auditory benchmark 'to which all are expected to aspire' (Flores 2021:113). These structures of language policing were particularly marked for schools serving racialised speakers living in poverty.
We call on the inspectorate to engage with its lasting legacies of colonialism which continues to shape its contemporary practice, to bear responsibility for its role in the marginalisation of classed and racialised communities, and to reflect on its role as a powerful tool of sonic surveillance. With future work on our corpus planned, our analysis is just the beginning of a larger project in exposing the institutionalised language ideologies of the inspectorate. We have focused on how speakers who are heard to deviate from 'standard' language patterns are shamed and delegitimised. For teachers, this is branded as unprofessionalism and a failure to help their students make educational progress. For students, this is branded as a failure to meet the academic standards set by school as well as a clear indication that their home language practices are not welcome. For schools, this potentially contributes to a damaging inspection report, which carries material consequences in terms of local reputation, parental trust, funding, recruitment, 380 Language in Society 52:3 (2023) performativity measures such as school rankings, and in some cases, the threat of closure and=or takeover.
As is a general logic within state organisations who propagate standard language and raciolinguistic ideologies, the inspectorate insist that inequity is solved through modifying individual language behaviours as opposed to dismantling institutional structures which exacerbate racial and class inequalities and hegemonic regimes. This is especially pertinent given that Ofsted increasingly draw on discourses of social justice in its policies. For instance, Ofsted (2019b:8) argue how disadvantaged children require access to 'cultural capital', including standardised English, which frames schools as spaces in which minoritised, racialised children are required to model their language practices on the white speaking subject (see also Nightingale 2020). This meritocratic myth that if marginalised speakers simply modify their language practices to conform with benchmarks set by white listeners, then this will solve social inequality is a core tenet of a raciolinguistic ideology, despite the fact that the white listening subject-such as the inspectorate-will continue to hear these students' language as insufficient (Flores & Rosa 2015). For us, social justice is only possible through the dismantling of structures and the redistribution of power, rather than the model of linguistic assimilation which underpins the inspectorate's approach.
As per our discussion above, ideologies of linguistic and biological impurity extend even further-with some speakers being described as lacking language or aspects of language completely. For these speakers, their language practices are erased through an ideological process of languagelessness (Rosa 2016). Ideologies of languagelessness do not just stigmatise practices heard to be nonstandardised, but 'call into question linguistic competence-and by extension, legitimate personhood-altogether', which renders certain bodies as 'incapable of producing any legitimate language' (Rosa 2016:163). These ideologies have long been a feature of education policy in England, such as in the Newbolt Report which stated that 'the first and chief duty of the Elementary School is to GIVE its pupils speech-to make them articulate and civilised human beings' (Board of Education 1921:60; our emphasis). Conversely, through the praising of teachers and students whose language practices are deemed to conform with the expectations of white middle-class ears and standardised English, the inspectorate work to construct the idealised speaker in terms of linguistic completeness and academic competence.
Finally, we argue that when understood through the prism of standard language and raciolinguistic ideologies, the inspectorate's aural perceptions of language means not only that they exacerbate class and racial inequalities, but also that they fail to focus on the issues that DO matter with regards to classroom talk. In policing nonstandardised patterns, the inspectorate overlook the social role that language plays in effective teaching, classroom management, and social relationships. One example from our discussion above is particularly illustrative of this-a 2003 report in which an inspector heard a teacher say "I don't want no more shouting out", for which they received criticism around their supposed lack of vocabulary and ungrammatical talk, which was considered as negatively impinging on students' own speech development. However, rather than hearing their language as deficient as the inspectorate do, we would argue that teachers simply draw upon nonstandardised patterns in order to preserve solidary relations with students and a congenial learning environment (see Ioannidou 2009;Snell 2018b). Further, the inspectorate's superficial focus on 'correctness' and 'standards' overshadows more productive approaches to language, which can help young people both to develop their linguistic repertoires and engage in meaningful classroom learning. For example, Ofsted fail to make an important distinction between 'talk for performance' and 'talk for learning' (Snell 2019). When developing the skills necessary to give speeches and presentations and participate in structured debate-that is, 'talk for performance'-it might be seen by teachers as reasonable to encourage students to minimise fillers (such as 'like') and avoid stigmatised grammatical forms (unless used deliberately for rhetorical effect, see e.g. Moore 2019). It is crucial, however, that this should be done with explicit discussion of how these forms are not 'incorrect' but have acquired negative social value through their association with social groups who have typically lacked power in society. However, when it comes to talk for learning, the aim is to think aloud and contribute spontaneously to an evolving dialogue in the classroom. There is no reason why such contributions should be made in standardised English, because it is possible to express complex ideas in a variety of linguistic forms and styles (Resnick & Schantz 2015). What is important is that students feel able to contribute to classroom discussion, given growing international evidence that stimulating classroom discussion is crucial to students learning and cognitive development (Resnick, Asterhan, & Clarke 2015;Alexander 2020). As Resnick & Schantz (2015:447) point out, '[t]his discussion space can accept, "Um like um like if the um-wait, what were we talking about?"' However, speculative talk and thinking aloud are less likely to happen in classrooms where 'correct' forms of expression are valued over ideas and where language is aggressively surveilled and policed.
The inspectorate may be tempted to dismiss our criticisms on the grounds that instances of language policing represent individual mistakes or are representative of policies which no longer carry official remit. Despite Ofsted's recent claims that 'equality, diversity and inclusion are at the heart of our work' (Ofsted 2021:57), we firmly reject this and argue that our research has exposed how deficitorientated listening practices are in fact institutionally embedded into the very heart of the inspectorate's ideologies and policies. These practices are historically entrenched-as Tomlinson (2008:112) shows, there is a long history of school inspectors having zero or inadequate training in race, institutional racism, and linguistic diversity. We conclude then, by reaffirming the importance of the raciolinguistic perspective we have taken in this article, placing critical attention on the actions of the listener rather than the speaker. Instead of asking students and teachers to adjust their language practices to conform with standardised English and the 'appropriate' patterns expected of them by the state, we suggest that the inspectorate 382 Language in Society 52:3 (2023) modify its listening practices to begin undoing its classed and racialised modes of perception. As Rosa & Flores (2017b:284) write, this shift in focus allows us to ask what if the problem is not the speakers' language practices but the inability of the white listening subject to hear minoritised bodies speaking appropriately, or, as we would add, even at all?