Varieties of ‘standard accents’ among teachers in contemporary Britain

This paper investigates accent variation within the UK, applied to the professional context of teaching. The purpose is two-fold: first, to discuss a variety of accents in contemporary Britain, demonstrating how accents can be categorised based on the degree to which they mark the speaker’s local origins. Second, by focusing on teachers’ accents within the context of teacher training in the UK, information can be obtainedregardingwhichparticularaccentvarietiesarepre-ferred in a profession in which accent is relevant to professional behaviour. In addition, two aims can serve to provide a better understanding of what are, in effect, suggested ‘standard’ accents within the teaching profession.

reduction might be a reality, as part of a de facto 'standard' , referring to the phonological reality for accents deemed appropriate for teaching. With regard to linguistic standards for teaching, however, discussions within the UK typically only reference the need for teachers to use 'standard English' , which can be spoken in any accent, of course. This paper presents the accounts of five English trainee teachers, for whom pronunciation guidance was provided by their mentors (though one established teacher chose to reduce his accent of his own accord). This focus thus allows for specific phonological information to be provided regarding variety within regional accents; and subsequent suggestions made as to what constitutes implied standard accents within teaching. Thus, the paper reveals how a topic otherwise absent from the official UK discussion of 'teachers' standards' (Department of Education, 2013) is approached within teacher training.

The notion of standard accents in the UK
In the UK context, earlier work by Trudgill (1983) provides a pyramid diagram reflective of UK accents, in which Received Pronunciation is placed at the top, thus representing a minority of speakers. Accents at the base, however, exhibit 'considerable amounts of phonological variation' (Trudgill, 2002, p. 173), consisting of a variety of broad regional accents (to be discussed). This provides some background as to what constitutes standard accents, further suggested with Trudgill's reference to a progressively higher status bestowed upon speakers as their accents simultaneously reflect ever decreasing regional variation.
Specific references made to standard accents (Bibby et al., 2017;Evans & Iverson, 2007;Millar, 1997;Trudgill, 1999) often discuss accent in terms of the avoidance of these more local, or broad, sounds (Ramsaran, 2015). Arguably, therefore, it is broad varieties of regional accents that are more likely to be judged negatively, those which retain 'local' features (Millar, 1997). For some interlocutors, such sounds can also be regarded as stigmatised features (Strycharczuk et al., 2020), with Millar (1997) providing some clarity regarding such features, referencing glottal stops and THfronting (Trudgill, 1999).
The classification of a given accent is also broadly reflective of Agha's (2005, p. 38) work on enregisterment, involving 'distinct forms of speech' which become 'socially recognized (or enregistered) as indexical of speaker attributes by a population of language users' . These 'enregistered voices' serve to identify people, which can involve judgements made regarding a speaker's class or profession, for example 'he sounds working-class' . Once such judgement, as attributed to particular speech forms, is conceptualised at the level of a stereotype (Silverstein, 2003) then this constitutes a marked form. However, accents deemed to have marked features can reflect both vernacular speech (Jensen, 2016) and more 'posh' varieties (Blommaert, 2007;Rampton, 2003), and involve stereotypes that are positive or negative. In the context of the Newcastle ('Geordie') accent, for example, Beal (1999, p. 45) regards 'stereotypical pronunciations in key words like "Toon"' , (meaning town) as representing 'a strategy for maintaining the positive aspects of the "Geordie" stereotype: friendliness and a strong sense of regional identity'; on the other hand, Hughes et al. (2012) discuss modern perceptions of RP as referencing snobbery and arrogance.
In terms of the phonological implications for broad accents, salient variants in specific positions can be sufficient to immediately identify the speaker's region of origin in terms of the pronunciation heard in certain words. This is reflected in the work of Labov (1972), regarding the level of awareness that is attached to certain linguistic forms.
For example, the aforementioned stereotypes refer to linguistic features which have a higher degree of awareness attached to them, often as 'a basis for negative comments' (Jensen, 2016, p. 2) due to what can be a stigmatised status.
As I have mentioned, however, this need not be limited merely to one group (for example, working-class) or one type of accent (e.g. broad); nonetheless, for accents deemed to be broad, there can be negative implications for certain professions, such as teaching (Archangeli et al, 2010;Garner, 2013;Baratta, 2017), finance (Moore et al, 2016) and even library work (Lippi-Green, 1997). This further reflects power issues in society, with certain individuals -here, mentors in teacher training -in a position to enforce, or at least suggest, linguistic norms, as part of linguistic capital (Bourdieu, 1991); as Ahearn (2001, p. 111) points out, 'language and power are . . . commonly intertwined' .
RP, at the top of the pyramid, has existed as a de facto standard British accent for some time (Trudgill, 1999;Coupland, 2000;Giles, 1971;Kerswill, 2006;Snell & Andrews, 2016;Milroy, 2000). This is broadly linked to standard language ideology (Milroy & Milroy, 1999), in which certain linguistic features, whether based on accent or syntax, are judged in 'common sense' terms regarding notions of 'right' and 'wrong' , 'correct' or 'incorrect' . Such linguistic ideology has implications for accents within certain professions, as referenced, and can reflect social mobility, which Trudgill (2002, p. 178) references as a catalyst for individuals to make changes to their accent, 'to move themselves up the triangle' . A larger implication is that British people today, despite social mobility (Friedman, 2014;Hey, 1997;Maguire, 1999;Roberts, 2012), may well wish to use their otherwise natural regional accents, but adopt a variety which does not accentuate the more local features (Ramsaran, 2015;Foulkes and Docherty, 1999;Baratta, 2018) -features which would identify the speaker more immediately as being from a specific place of origin. This points toward variation within regional accents, with some varieties deemed more 'fitting' perhaps for certain professions.
Likewise, RP is not monolithic and as with the variety heard within regional accents, there is also variety within RP. Gimson (1980), for example, discussed RP as being conservative, general and advanced, with Wells (1982b) providing further detail regarding the variety within RP, focusing on U-RP, adoptive RP and near-RP, in addition to mainstream RP. A key factor within this variability is public perception, in terms of how accents can create a 'mental image' (Wells, 1982b, p. 279) for the interlocutor, for good or bad. This is of relevance to the teaching profession, as certain accent varieties, if perceived as 'common' , might not be deemed appropriate, in some minds, for teaching. U-RP, for example, is referenced as reflecting 'an elderly Oxbridge don' or 'a jolly-hockey-sticks schoolmistress at an expensive private girls' school' (Wells, 1982b, p. 280). Adoptive RP is the variety spoken by adults who otherwise did not speak with this accent as children, with Wells (1982b, p. 284 citing 'social pressure' to adopt this accent in certain societal contexts, such as work; again, this is evidence of the ways in which one's profession may indeed influence accent (Donnelly et al., 2019).
This need not involve a wholesale change to RP of course, but an attempt to reduce an otherwise regional accent.
Near-RP refers to an accent which, by virtue of its lack of specific regional qualities, would be perceived by many as 'educated' , 'well-spoken' , 'middle-class' , though speakers of the other varieties of RP may or may not agree on this point.
More recently, Cruttenden (2014) speaks of RP as having been influenced by regional accents such as Cockney, again suggestive of a continuum approach, from most to least-RP sounding in this case.
The importance of linguistic contact is also relevant to linguistic variation. Trudgill (1974;1986; and Kerswill (2002) discuss this in terms of dialect levelling, itself a product of speech accommodation, in which 'interlocutors will tend to converge linguistically' (Kerswill, 2002, p. 188), involving both short-and long-term accommodation. Such linguistic contact can of course reflect organic change that occurs naturally within a language, but this need not suggest that accommodating to someone else's speech is otherwise free of conscious effort; thus, we should not rule out agency on the part of speakers to make certain changes to their language in contact situations (Thomason, 1999;Verschik, 2017). In addition, in the context of my study, it was made clear that the phonological changes made by the teachers were indeed consciously enacted, in large part due to the teachers being provided with specific phonological guidance (with one teacher choosing to modify his accent, however, though still providing phonological clarity as to what this involved). This does not suggest of course that instances of accent change were consistently based on conscious decision-making. But given these two points, I argue that in language contact situations, more so when phonological change is prescribed by those in authority, a change in accent can be part of a deliberate strategy.
In the context of the North of England and accent variation, Strycharczuk et al. (2020) describe a more general accent, the implication again being that a removal of broad features leads to this accent's perception as 'general' , as opposed to 'specific' (that is, 'local' or 'broad'). They refer to General Northern English (GNE), focusing on a linguistic spread that incorporates the Northern cities of Manchester, Liverpool, Leeds, Newcastle-upon-Tyne and Sheffield. This is part of a suggested dialect levelling (Kerswill, 2003), in which regional variation might be reduced, more notably amongst educated and middle-class speakers who are the focus of the study by Strycharczuk et  In closing this section, it is argued that UK accents are more likely to be perceived as standard -or a similar judgement such as 'generic' or 'neutral' -when the more immediately recognisable features are absent, the so-called broad or local sounds. The implication of this is that such accents might be heard across a broader geographical area, precisely because the more local sounds -those which would identify the speaker to a smaller, more precise region of originare absent. Therefore, general accents make it more difficult to identify the speakers regarding their specific regional origins, and this latter point has three implications: first, it suggests that avoidance of more local sounds potentially contributes to speakers being less likely to be negatively stereotyped, precisely because the sounds otherwise associated with a stigmatised accent are absent (Coupland & Bishop, 2007;Donnelly et al., 2019;Giles, 1971;MacFarlane & Stuart Smith, 2012); second, if each regional accent in the UK is realised in this manner which seeks to remove more local sounds, then this implies a standard accent variety within each region, whether city-based or, in some cases, a standard that comprises several city regions within a larger area, such as the suggested Standard Northern English; third, there are implications for a speaker's class level, with Trudgill (2002, p. 173) explaining that 'it is usually possible to tell which broad region of the country middle-class speakers come from. And working-class speakers can usually be pinpointed even more accurately as to their geographical origins' . This ties in with the retention of local features, suggestive of working-class speakers, but a more loosely defined accent (such as General Northern English) tied to the middle-classes. Given that teaching is a profession associated, historically at least, with the middle-classes (Apple, 1988;Higginbotham & Weber, 1992), then this is of relevance to the individuals in this study.

The implications of 'standard' UK accents
This section focuses on the phonological implications for accents arguably perceived as 'standard' , versus those perceived as broad. This might suggest an oversimplified binary division, though I have in fact argued for a trichotomy approach (Baratta, 2018), a division of accent varieties in terms of broad -general -neutral (thus, standard). This, too, might suggest a rather simplistic approach, especially given the complexities involved with accent variety as discussed previously (Wells, 1982a). However, a trichotomous approach can act as a placeholder concept in the first instance, with the implication being that it is merely a starting point for accent analysis, a means to avoid singular notions of regional accents and delve deeper. Further, the three level classification ('broad -general -neutral') reflects extant approaches in other contexts (Strycharczuk et al., 2020), as well as the intuitions expressed by participants within my study.
Second, we can also expect there to be variability as heard within the suggested distinctive features themselves.
Further, other distinctive variants also occur in words such as baby and go as heard in this region, such as [e:] and [o:], associated with those who are female, young or middle-class (Mearns, 2015). As such, we can conceptualise distinctive features as revealing more than just regional identity and I concede some will take issue with the sounds represented in the table on these grounds. However, there is evidence for the sounds presented to nonetheless be associated with the regions identified, before further identifying factors, such as age and sex, are considered. Class, however, is indeed a factor that will be discussed further, as this impacts on the degree to which individuals are associated with specific city regions in terms of accents deemed broad or not. Finally, my suggestion is that the sounds presented in Table 1 are simply distinctive enough, additional varieties notwithstanding, for the speakers to be identified with the region in question.
The choice of variants above is further explained by first referring to perceptual dialectology (PD) and the work of Montgomery (2012). Montgomery explores perceptions of British accents based on the effects of interlocutors' region of origin, and how this impacts on their ability to clearly identify, as well as how they identify, different regions and by implication, the regions' variety of accents. The relevance this has with the table's contents is as follows. First, the results of Montgomery's study clearly show the extent of participants' awareness of different regions, with accents tied to Newcastle, Liverpool and Birmingham, for example, being very well recognised, perhaps as a result of being firmly established in public perception (Coupland & Bishop, 2007;Trudgill, 1999). Second, there is particular recognition in the study for 'stigmatised dialect areas' (Montgomery, 2012, p. 658), which can tie in with more broad accent features, as referenced and as will be discussed further. Further, if we also consider the results of Coupland and Bishop's study (2007), in which a variety of British accents were evaluated on the basis of prestige and attractiveness, it can be inferred that recognition of an accent -regardless of the perceptions it might carry -can realistically only be aided by recognising its specific sounds, or certainly some of them. The fact that Coupland and Bishop (2007, p. 85) reference 'RP-like accents' is telling, as it is at once vague and yet a clear reference to accents which might be intuitively understood by interlocutors on the basis of lacking more broad accent features; the use of -like also suggests that it is a 'midway' accent, suggestive of an accent which attempts to avoid accentuating either extreme of 'posh' or 'common' accent features (see Foulkes & Docherty, 1999).
Again, however, the more variants are used which, in certain contexts, equate to the so-called 'distinctive features' , the more the accent is perceived as broad; the more such features are removed, the more 'neutral' -or arguably 'standard' -the accent (Baratta, 2016a(Baratta, , 2016b, 2018. Overall, accents thus perceived as more 'standard' tend to exhibit less regional diversity (Wells, 1982a), and here the relationship with class cannot be ignored, with evidence for fewer regional features for middle-class speakers compared with the working-classes (Baranowski & Turton, 2015;Haddican et al., 2013;MacFarlane & Stuart-Smith, 2012). Kerswill (2002, p. 187) also references the influence of dialect levelling, leading to 'the loss of localised features in urban and rural varieties of English in Britain, to be replaced with features found over a wider region' . Of course, the implication here is not that a given regional accent is tied to one of three variants, broad, general or neutral. This is, however, a useful approach to take as an initial placeholder concept, as I had mentioned, and certainly as a means to initially discuss accent variation within a region; this avoids an overly broad approach, such as 'the Manchester accent' , understood as a linguistic monolith. Wells (1982a)  up' one's accent, within an accent termed as 'northern Near-RP' (Wells, 1982a, pp. 352-353), as heard in additional words such as cup [kəp] and brother [brəðə] (Wells, 1982a, pp. 352-353). Wells refers to such pronunciation as involving hypercorrection, but from a sociocultural point of view, it represents upward mobility, or a desire to connote such in one's speech. Again, this reflects a lack of localness in one's accent, to the extent that individuals might not even be associated with their region of origin based on their accent. And from here, 'local stereotypes' can be avoided.
Furthermore, Wells (1982b, p. 354) goes on to explain that, 'there are many educated northerners who would not be caught dead doing something so vulgar as to pronounce STRUT words with [ʊ]' . The suggested 'vulgarity' is not tied to the particular sound per se, but the fact that accent is a proxy for social categories tied to regional, racial and classbased origins. This means that a group who is stigmatised for whatever reason will have such negativity placed onto their language also. In particular, Northern accents are sometimes regarded as reflections of working-class identities (Hey, 1997), with the connotations of being working-class, for some, reflective of being 'common' perhaps (Baratta, 2016a(Baratta, , 2016bBelchem, 2006;Honeybone, 2007;Mugglestone, 2003). Such perceptions might be deemed incompatible in certain professions such as teaching.
To close this section, it is clear that there are implications for accent within the workplace, certainly in professions associated with middle-class aspirations. Research clearly demonstrates that this is still a live issue, in terms of wanting to present an image deemed 'professional' amidst negative perceptions for certain accents within society (Addison & Mountford, 2015;Baratta, 2016aBaratta, , 2016bBaratta, , 2017Baratta, , 2018Cardoso et al., 2019;Donnelly et al., 2019;Friedman, 2014;Garner, 2013;Lawler, 1999;Maguire, 2005). We might then ask what the implications are for the speakers of stigmatised accents in professions such as teaching; however, a more focused question is what are the implications for speakers of broad varieties of such accents in certain professions?

RESEARCH QUESTIONS
Given the suggested evidence, both implicit and explicit, for a variety of regional accents in the UK (Baratta, 2018;Gimson, 1980;Haddican et al., 2013;Ramsaran, 2015;Strycharczuk et al., 2020;Trudgill, 1983), the research aims to address this topic within the context of the teaching profession. This allows for some insight, given a profession in which broad regional accents might be disfavoured, as well as offering specific phonological information as presented by the teachers themselves. Thus, the research can shed light on the implications for certain accents, as well as what this means from a phonological perspective. The research questions are as follows: 1. Which variety of regional accent is adopted by the teachers?
2. What rationale, if any, is provided for this variety being used by the teachers, whether self-imposed or required by others (for example, the senior staff)?

METHODOLOGY AND DATA COLLECTION
The results presented derive from three separate studies conducted from 2014-2016. All three studies were united in the focus on teachers' accents. Specifically, teachers were asked if they had ever changed their accent in any way

Study 1
This study (Table 2), involved 11 trainee teachers in two universities in the North of England, as part of their course of study. Semi-structured interviews were used to collect data, with the interviews conducted in a room on campus where both interviewer and interviewee were assured of privacy, but also felt comfortable. As the teachers were studying as trainees when the interviews were conducted, this allowed for their recollections to be current.
Of note is the way in which T4 described her accent as 'broad' and for several other participants within the interview, the term 'generic' was used to categorise their accents. Moreover, that participants used intuitive terms such as 'broad' and 'generic' (and 'neutral' was also used by some), suggests, in part, that a trichotomy-based understanding of accent is sometimes used by individuals in society though as pointed out, this is not an approach strictly adhered to here but used instead as a means to initiate a discussion into what is otherwise a more complex picture regarding accent variation.

Study 2
This study involved fifteen trainee teachers from two universities in the South of England, also conducted during their period of training (Table 3). Given the inability to travel to these universities, for what would have been potentially several excursions, a questionnaire-based approach was used to collect data. In this case, a questionnaire was sent to participants via email as a word document, with the participants completing it and emailing it back to me.
Again, we see examples of key words such as 'strong' and 'mild' to describe accent, which would respectively suggest 'broad' and 'neutral' . T18 describing her accent as 'standard English' is surprising, as this is not an accent per se. It could be, however, that she feels she uses an accent that does not accentuate any specific regional origins and thus suggests a certain 'standard-ness' . It is worth pointing out that within the context of teacher training for studies 1 and 2, and TA B L E 3 Data from Study 2

Study 3
Study 3 involved interviewing nine teachers who were already established in the teaching profession (Table 4). While this involved a degree of retrospective discussion with regard to any accent modification, the participants' accounts were detailed, in some cases quite current regarding accent modification and for one participant, T29, very detailed indeed, though involving a discussion from his days as an undergraduate student. Of the nine teachers, one taught at a state secondary school (T29); two taught at a state primary school (T27 and T31); and the remaining six teachers were employed at a private primary school. All three schools are in the Manchester area.
Of note again are the terms used, such as 'neutral' and 'mild' . Moreover, the fact that six participants teach at a private primary school might have implications for accent, for both teachers and students. In a paper focused on the results from this one school alone (Baratta, 2016a), one of the teachers interviewed explained that students from state schools would tend to have Mancunian accents that are more 'in the area' , further suggesting that the local students at the private school did not necessarily have local -broad -accents. The private school does in fact have an elocution class, though optional. But it was reported that within the context of this class, glottal stops are a particular feature which are proscribed in students' speech.

RESULTS
While the studies were originally focused on teacher identity seen from a linguistic perspective, this paper is wholly informed by phonological examples of how teachers modified their accents, whether instructed to do so or chosen by themselves. In terms of this specific focus, I present the accounts of five teachers: T1, T5, T12, T14 and T29. proper speech in a classroom setting is enabled by avoidance of a broad Yorkshire accent, suggesting that his accent used for teaching seeks to avoid 'regional extremes' . T1's own reference to 'regional extremes' is telling, as it simultaneously refers to the use of sounds tied to a given region -here, Yorkshire -but also implies sounds which are very much associated with said region, perhaps immediately in terms of interlocutors readily identifying the speaker's regional origins. For T1, and his mentor, such sounds, even when teaching local children, are suggestive of being somewhat inappropriate. This might be due to the idea that if children use sounds which are not on the extreme end and instead allow for more intermediary sounds, such as that which Wells (1982a) discusses, then it allows for the speakers to be perceived as less working-class, more suggestive of upwardly mobile speech. Though speculative perhaps as applied to children, there is evidence for this, given the recent example in 2013 of Colley Hill primary school in the West Midlands, which banned a list of phrases for its pupils, enacting a 'zero tolerance' approach. While some of the phrases were tied to dialectal expressions (they was instead of they were), some were indeed based on pronunciation (somefink instead of something). The 'ban' made clear that local expressions -and by implication a local accent (in Trudgill's sense, thus meaning 'broad') -have a time and place, but not for more official contexts, such as education. Some parents felt that this was an attack on the Black Country accent, but the school's reference to many pupils being on free school meals makes clear the implication that certain accents can hold students back in the future.

A move to a neutral accent to aid student comprehension
T1 further made clear that teachers need to ensure that the students understand them at all times, referencing the potential for 'misunderstandings' . Having said that, T1 is teaching in his home region, in which we might expect many of the children to have local accents. Once again, however, it appears that the ideal is to avoid too much regionality and replace it with a less broad variety of an otherwise regional accent -here, Yorkshire. As such, a more 'professional' accent is deployed for teaching.

A need for standardised speech
T5 expressed upset at the directive from her mentor to reduce her accent, regarding this as contributing to a lack of (linguistic) diversity as a result. As with T1, T5 was also teaching in her region of origin, broadly speaking, as this involves the North of England. She also shared with me the fact that in the staff room, her otherwise natural accent had been labelled as 'common' , and she explained to me the need to teach 'correct' sounds for phonics. In terms of the 'more formal' , but the mentor's overall justification for the phonological changes was for T5 to adopt more 'standardised speech' . T5 made it clear that this directive was given in relation to her accent, and not her grammar. Therefore, this suggests again the removal, or certainly reduction, of the more clearly identifiable Yorkshire-based variants, with specific phonological guidance provided. It also ties in with notions of standardness with regard to accents, which again demonstrates a move away from more local sounds -thus making one's accent less easy to trace to a specific region.
By removing some of the broader sounds, T5 believed that this resulting 'generic' accent (the term she used) is not reflective of the diversity in society which the teachers are otherwise expected to focus on in the classroom. Another teacher in the study, T9, in fact expressed the idea that what she perceived as a lack of linguistic diversity in the classroom, seen via accent reduction, was not in keeping with the current ethos of exposing students to diversity on several levels. T9 had said that 'if you are going to speak the same as everyone else then you are going to lose who you are and where you're from. Trying to teach children to be who they are . . . but then you say that we've all got to speak the same so it's a bit backwards' . T5 echoed a similar sentiment, explaining that 'it's important that children hear different accents so that when they go other places they can understand what people are saying and don't think the accent's abnormal' . T5 further added that 'there's nothing wrong with the way I speak; there's nothing wrong with my own intellect' .
While these personal perceptions of accent reduction are not the focus of this paper, I have included them here as they help to contribute further to a clearer picture. Specifically, the reference to T5's modified accent as being 'standard' and in her terms reflecting a lack of diversity as a result, suggests three points. First, standard is again being understood as an accent which does not reflect more local speech sounds, the inclusion of which would suggest broad accents. Second, the reference to sounding the same is indicative somewhat of more neutral accents (Baranowski & Turton, 2015;Haddican et al., 2013;Strycharczuk et al., 2020), which indicates that a loss of more local sounds can be perceived as a less regionally identifiable accent (or an accent tied to a broader region of origin, such as GNE). Finally, T5's reference to her intellect suggests that the implications for societal perceptions of broad accents can be negative, implying that her otherwise 'common' accent might be judged as inappropriate for a professional education setting.

Implications for phonics teaching
T12 teaches primary school in the South and was instructed in front of her seminar group by her mentor, a RP speaker, that if she wanted to teach phonics, it was 'best to go back to where you come from' . There was a direct reference to her sounds as in bath and bus: [baθ] and [bʊs], and the need to adopt Southern pronunciation, at least for phonics teaching: [bɑ:θ] and [bʌs]. A compromise was suggested, however, in that T12 was told that she would need to use her teaching assistant to demonstrate these latter sounds for the students. The two sounds that are part of T12's everyday speech are 'distinctive' , albeit on a broad level tied to speakers from the North/Midlands. In other words, T12's distinctive features identify her by who she is clearly not -a Southerner. While there are phonological features more closely linked with her immediate region of origin -the city of Leicester in the East Midlands -T12 only referenced the two aforementioned pronunciations.
On the one hand, does the retention of these two sounds equate to a broad accent? From a Southern perspective, perhaps so. In particular, the vowel /ʊ/ is suggested to be a sound particularly associated with the North when deployed in words such as bus, umbrella and ugly. Within the North/Midlands, however, this sound might also be regarded as broad, with more neutral Northern accents not exhibiting this sound for the most part. I refer again to the earlier quotation by Wells (1982b): There are many educated northerners who would not be caught dead doing something so vulgar as to pronounce STRUT words with [ʊ], but who would feel it to be a denial of their identity as northerners to say BATH words with anything other than short [a]. (Wells, 1982b, p. 354) This suggests that replacing /ʊ/ in words like bus with /ʌ/, for example, is perhaps more acceptable than replacing the front vowel with /ɑ:/, with Trudgill (1986, p. 8) claiming that 'many Northerners, it seems, would rather drop dead than say /dɑ:ns/, the stereotype that this is a Southern form is again too strong' (original emphasis). Combined, the retention of /a/ and the replacement of /ʊ/ in certain words within Northern speech (perhaps replaced with /ʌ/ as in bus and but) can be considered a good example of a neutral accent. Given the accent's removal of a sound in certain positions which has immediate connotations of being a Northerner -or here, being from the Midlands -then clearly this particular sound's removal will make the speaker less identifiable as being from her home region.

5.4
The need for a 'professional' accent T14, a secondary Art teacher, was told by her mentor that she needed to use a 'professional' accent in her teaching. The specific feature that to him was deemed unprofessional was T14's use of a glottal stop in her speech. T14 described her accent as 'strong' and the use of a glottal stop is arguably a feature that contributes to this perception.
However, where T14 is merely describing the 'intensity' of her accent, her mentor describes it in terms of societal perception. That is, a glottal stop within a teacher's speech might be regarded as, if not unprofessional, then for some inappropriate.
T14's mentor went so far as to instruct her to write the word 'water' with a capital T, in order to avoid T14's pronunciation being realised thus: [woʊʔə]. Furthermore, T14's mentor had written on a feedback sheet the need for T14 to 'use specific learning language IN CLASS (+ OUT OF CLASS)' (original emphasis). Here is again a link to accents being considered more 'standard' because, due to their avoidance of features which might be considered broad -hence 'nonstandard' -any negative perceptions of the speaker are minimised, if not removed.
T14 regarded her accent as bringing 'a touch of reality' to the school by being broad and reflective of her origins in Croydon, or 'the ghetto' as she admitted it was referred to by her students. Thus, an accent which has connotations of 'the ghetto' and associated imagery can be viewed as 'unprofessional' by those who might wish to otherwise rid teachers' speech of such connotations -here seen with the removal of glottal stops. However, this is a feature that arguably helps to contribute to T14's pride in her accent, as well as the region from which her accent derives. As she explained, 'I am very proud I am from Croydon' . Broad accents can reflect more broad stereotypes, and Rivera (2015, p. 263) acknowledges this in terms of work-based contexts and the implications for certain accents, some of which may violate 'the tenets of good polish' . From T14's perspective, however, her accent made her more approachable as it reflected the accents of her students and helped them to relate to her better, in her mind. She further expressed that she had forged 'great professional relationships' with her students, precisely because she speaks differently than the students' 'normal teachers' . In this context, 'normal' can tie in with those who speak with less regional accents, which would imply that a broad variety of a regional accent is the marked form.
From T14's account, we can glean a great deal regarding the ways in which specific sounds carry with them specific stereotypes. Broad accents, given that they are more likely to be perceived as working-class, are thus avoided by some teachers (or a directive to avoid them is provided by mentors). Within the context of this paper's focus on notions of standard accents, it is again clear that an accent deemed as 'standard' is more likely to be considered as such based on the sounds it does not exhibit -those that tie the speaker more readily to a given region -and for T14, a region associated with working-class origins. Thus, such an accent is judged inappropriate by her mentor, who clearly believes that a teacher's accent needs to reflect a 'teacher voice' by means of avoiding stigmatised sounds.

A need to sound 'educated'
T29 is originally from Bristol, but now works in a secondary school in Manchester. He admitted that his accent reduction was something he himself chose to do, a decision made during his undergraduate days at university. He was concerned that an accent perceived as 'country' would be out of place within the teaching profession, perhaps more so given that he teaches outside of his region of origin. In terms of what this means from a phonological perspective, T29 explained to me that he adopted a non-rhotic accent. Thus, by removing the r's from his speech, he believes that he is subsequently removing the negative connotations that come with a rhotic accent in England specifically, certainly if tied to the Southwest: The Bristolian accent has lots of connotations. Normally, things like village idiot, yokel, farmer, you know, friendly but stupid . . . agricultural . . . so because of that once I came into contact with people that had softer accents and accents from elsewhere . . . you have these kind of stereotypes that exist and it wasn't doing me any favours . . . I didn't want to live a life being seen as a thick yokel who lives on a farm. (T29, male) Barras (2015, p. 277) explains that rhotic accents in England can be viewed as 'unsophisticated' , reflective of 'Farmer Giles' imagery. This subsequently leads to potential stereotyping, as indicated by T29 himself, of being friendly, but unintelligent. Clearly, such imagery does not suggest a professional identity to T29 and he believes that by using a non-rhotic accent he avoids negative stereotyping. The reference to identity by T29 contributes to perceptions of standard accents. In this case, T29 chose to remove a sound whose presence was felt to reflect imagery that is far removed from T29's concept of how a teacher should be perceived. In this manner, a switch to a non-rhotic accent equates to standardness.
In Bristol, a rhotic accent is arguably the norm, and is heard throughout the Southwest. However, given that rhotic accents are the exception in England and are tied largely to the Southwest and perhaps parts of Lancashire, then this is indeed a distinctive feature as heard within a broad Northern region whose accents -from Liverpool to Newcastle -are largely non-rhotic. For this reason, a rhotic accent would serve to identify T29's regional origins and associated stereotypes even more; removal of the rhoticity, however, downplays them. This might help to explain why T29 characterised his accent as 'mild' .
T29 further explained that 'when you get to the point where you cannot tell where someone is from . . . you can't hear any regional accent in it, then I would class that as what someone would want the standard accent to be'; he even went on to describe this in terms of 'an RP feel' to accents. If this is the linguistic goal of T29, then he appears to have succeeded by his own account.

DISCUSSION
From these teachers' accounts of phonological change for their teaching, it is suggested that broad accents are disfavoured. Given that many of the participants within the studies overall indeed referred to accents as 'neutral' , 'general' , 'broad' -and used related terms such as 'strong' , 'mild' and 'standard' -it again appears that there is a suggested societal scale in mind regarding standard accents. The implications for the differing realisations of regional accents within certain contexts and settings -here, teaching -suggest that neutral varieties of a given accent are preferred, and regarded as 'standard' and thus largely free from the negative associations that broad accents can carry.
My conclusions from a sample of five teachers cannot be generalised very widely, though I do take into consideration further insights from the studies. These included discussions provided by other teachers which, while not referencing specific dialectological information, nonetheless indicated a preference for accents which could not be described as 'broad' . One such example concerns an individual from Rossendale in Lancashire, part of a preliminary study I conducted prior to Study 1. At a job interview for a teacher training programme, the person conducting the interview told him that the interview would be stopped unless he changed his accent. The rationale was again a need to be understood by the students, but with the added point made that this individual was not speaking 'standard English' and yet wanted to train to be an English teacher. The person conducting the interview explained that if this individual were to teach with his otherwise natural accent, parents would complain, questioning why he had been hired; the interviewee was also told that he would need to 'correct' students if they spoke like him.
That 'standard' English was used to refer to an individual's accent, and not his use of grammar and vocabulary, once again suggests that UK society may indeed have some kind of notion of what a standard accent sounds like. This might be even more relevant in certain professions, such as teaching, in which the ideal may be to avoid broad accents at all costs. This is not about changing one's accent per se, but simply about reducing it. This makes for a less recognisable accent in terms of its region and as I have made clear, it also suggests that the potential for negative stereotypingas in this case -is reduced. In terms of phonological clarity, this particular teacher only referenced that his modified accent involved adding h's where they were not otherwise needed, reflecting a case of hypercorrection. Another point to make pertains to stereotypes (Labov, 1972;Silverstein, 2003). The teacher from Rossendale readily categorised his accent as 'working-class' which in his mind marked him as being 'hard-working and straightforward'; to the person conducting the job interview, however, this working-class accent marked the teacher as 'unprofessional' .
Whether or not trainee teachers need to be better informed about UK accents is a topic for a future discussion, and a topic which might make for enlightening discussion indeed. This can help to explore a variety of views on accent within the teaching profession, with views provided from mentors and teachers (and possibly those of the students themselves). Just what is a 'standard' accent in purely linguistic terms? And do these terms need to be spelled out for teachers, as they were for the sample of teachers provided in this paper? Or, is it sufficient to leave teachers and mentors to address this issue between themselves, without any further discussion needed (though potentially involving a linguistic tug of war)?
In the meantime, I have argued that there is more than one standard accent -essentially a standard for all the various UK regions. However, given that neutral -hence standard -varieties of regional accents lose much of their regionality, this in turn means that standard accents might best be understood as spanning regions at times (for example, General Northern English), and not always individual cities.

CONCLUSION
Based on the results, and in terms of addressing the research questions, it is strongly suggested that more neutral accents, those less broad, are favoured and encouraged by mentors and/or by the teachers themselves. This is based on the avoidance of specific sounds, sounds which in some cases serve to identify individuals as being from outside their region of teaching (in the case of T12 and T29) and can also serve to identify teachers, even if teaching within their region of origin, as having an accent perceived as broad. The suggested broadness is based again on salient variants (or 'phonological giveaways'), which mark accents as being more identifiable from a given city. The fact that such accents can be regarded as incompatible within teaching contexts might reflect social mobility, in that accents which rid themselves of the more broad features, may also reflect a desire on the speakers' part to rid themselves of negative class-based stereotyping as a result (Addison & Mountford, 2015;Lawler, 1999;Maguire, 2005).
The rationale for the modification, whether chosen by the teachers or not, largely rests upon the desire to use accents which are seen as 'professional' . This broad term necessarily encompasses a great many perspectives referred to by the teachers' mentors/senior staff. For example, the need to be understood by speaking 'properly' (T1), the use of a 'standard' accent (T5) and the avoidance of negative connotations attached to certain accents (T14, T29) might collectively suggest 'linguistic professionalism' . The account of two Midlands teachers (T7, T12) regarding their accentbased directives implies a certain professionalism by virtue of using the expected pronunciation for phonics teaching in the South. This has implications for what exactly the phonics standards are within the UK and more immediately, I suggest that the use of non-Southern segmental sounds in phonics teaching in the South may well be viewed in more absolute terms as 'incorrect' . A future study is planned, aiming to capture the views of 1,000 teachers in all parts of the UK, from both primary and secondary levels of teaching. The purpose is again to obtain insights into the role of accent in the teaching profession, notably those instances in which accent reduction is practised and understanding the 'how' and the 'why' of this practice. Until then, it is hoped that the approach described here has helped to more accurately discuss the diversity of accents, seen with varieties within a variety, and how this plays out in professional contexts in which language is indeed a relevant aspect of professional behaviour.