Language standards, standardisation and standard ideologies in multilingual contexts

This essay – which also serves as an introduction to the six other articles in this special issue – examines the development of the ﬁ eld of language standardisation studies in recent decades. First, it notes the change in focus occasioned by drawing on the notion of language ideologies, especially standard language ideology. That ideological awakening has, in turn, revealed that standardisation studies have, until recently, been largely ideologically monolingualist. I argue that we must consider multilingualism (broadly conceived) in at least ﬁ ve ways when we study language standardisation: to recognise diaglossia within a single named language; to understand the nature of polycentric standards; to analyse language purism; to appreciate the key role of (foreign/ second) language learning in codi ﬁ cation; and to trace the transmission of ideologies across languages and cultures. The paper gives examples of the ways in which our research can be unwittingly monolingualist in its concepts and methods, and examines the role of the concepts of heteroglossia and translanguaging in challenging that monolingualism. It concludes by setting an agenda for third-wave standardisation studies, with a call for standardisation studies that are enriched both by the ideological turn and by attending to multilingualism.


INTRODUCTION
Language standards, standardisation and standard ideologies in multilingual contexts four areas of language standardisation identified by Haugenselection, codification, elaboration and implementation (Haugen 1983; see also Coupland and Kristiansen 2011, 21) much early research focussed on selection and codification. More recentlyespecially since the focus on language history 'from below' (Elspaß 2005;Elspaß et al. 2007)research also attended to and sought to explain the gaps between prescription and usage, including the extent to which prescriptions might influence usage (see e.g. Rutten et al. 2014, including McLelland 2014 pioneering already in German linguistics were Konopka 1996 andTakada 1998). It has been suggested that some kinds of prescriptions may be more likely to effect change than others, whether for linguistic reasons (e.g. the relative salience of variants in the language of the existing population; see Poplack et al. 2015) or sociolinguistic ones (e.g. top-down language policy), or thanks to differences in the force and/or dissemination of orthographical and morphosyntactic prescriptions. Within such a paradigm of descriptive (historical) sociolinguistics, present-day case studies, besides their immediate practical relevance to contemporary language planning, also served as examples of late standardisations (Deumert and Vandenbussche 2003;Vogl 2012;Rutten et al. 2014), or as updates of 'early' standardisation histories, showing that the teleological process implied by some models does not end with a neat full-stop.
Two developments in the last two decades have changed language standardisation studies, informing our research project and this special issue in particular. The first key development is greater attention paid to the role of ideology in all aspects of language standardisation, a return to the linguistic anthropology roots of language standardisation studies (Haugen 1966 andKloss 1967 both published their work in the journal Anthroplogical Linguistics). Especially in studies of prescriptivism 1 (e.g. Beal, Nocera, and Sturiale 2001), language purism (e.g. Thomas 1991;Jones 1999;Langer and Davies 2005), and stigmatisation (e.g. Langer 2001; Davies and Langer 2006), it has been implicit that standardisation has a 'socio-political' dimension, raising the issues of 'language and power' that already underlie Milroy and Milroy's (1985) notion of 'standard language cultures', in the sense of the 'the beliefs and attitudes, shared practices and discourses which shape and support the historical development' (Deumert and Vandenbussche 2003, 2, 5, 10). Refocussing attention on to standard language ideology makes this explicit, however. The notion of an ideology 'problematizes speakers' consciousness of their language and […] in shaping beliefs, proclamations, and evaluations of linguistic forms and discursive practices' (Kroskrity 2010, 192). Standard language ideology encompasses assumptions about language correctness; belief in 'the one best variety'; and a demotion of all (non-standard) varieties (Vogl 2012, 13). Standard language ideology 'simultaneously shapes and hides many of the actual practices of speakers, especially minorities and migrants' (Gal 2009, 14, cited by Vogl 2012. Recognising such an ideology amounts to 'the final rejection of an innocent, behavioural account of language' (Coupland and Jaworski 2004, 37;cited in Kroskrity 2010, 200). 2 The second fundamental change to standardisation studies is the attention paid to multilingualism in studying language standardisation, as the theoretical lens of language ideologies has helped us to recognise the monolingualist ideological paradigm underlying much work. The initial focus on single delimitable languages in isolation was a corollary of the history of national philologies. Standardisation of one single language per nation-state went hand in hand with patriotism and nationalism (the one language, one nation ideology: Cooke and Simpson 2012, 120), and, by default, assumed monolingual native speakers as the norm. The complexities introduced by acknowledging more than one language at either the level of the state (multilingualism) or the individual (plurilingualism) were largely erased, even if certain exceptions were tolerated, typically Switzerland (e.g. Rash 1998) andLuxembourg (e.g. Newton 1996). Within Modern Languages, research has been undergoing a corrective to this assumption of monolithic cultures of nation states in Europe for decades now. 3 Nevertheless, Auer's (2005) comparative 'typology of European dialect/standard constellations' explicitly precluded multilingual contexts from consideration, except where 'exoglossic standard languages enter the repertoire' (for example, using Latin as the standard for written communication in medieval Europe: Auer 2005, 1). Yet, assuredly, 'it matters whether the context is monolingual or multilingual, and how the different languages are configured in multilingual contexts', as Peters (2017, 355) noted in her epilogue to one of the first collections with a focus on prescription and standards in multilingual contexts (Tieken Boon von Ostade and Percy 2017; see also Wright 2018 for mutilingualism in the history of standard English). Hüning, Vogl, and Moliner (2012) focussed on standard languages and multilingualism in Europe, and Vogl's introductory essay is a valuable account of the tension between standard language ideology and multilingualism within the context of a European language policy that is avowedly supportive of (at least certain kinds of privileged) plurilingualism among its citizens (Vogl 2012, 30). 4 Once we choose to attend to it, it is obvious that multilingualism has played a part in all aspects of language standardisation, in at least five ways: (1) First, taking 'multilingualism' in its most expansive sense of multiple repertoires, the pervasiveness of the 'standard language ideology' has obscured our view of diaglossia, the range of varieties within a (conventionally named) single language, one of the four kinds of 'hidden multilingualism' identified by Vogl (2012, 6) in the standard-language multilingual culture of Europe today. (I return to these below.) 5 As Rutten, Vosters, and Vandenbussche (2014, 13-14) suggested, researchers' accounts of within-language diglossia (i.e. uniform H and L varieties) in the face of actual diaglossia are arguably 'the discursive result of the standard language ideology'. 6 In fact, there is good historical evidence of intermediate varieties in the past, so that seeming 'destandardisation' today is not a new stage beyond some 'natural' end-point of standardisation, but merely a continuation of diaglossia, which Rutten, Vosters, and Vandenbussche (2014) argue has a longer history than Auer's typology suggests. 7 (2) Regarding selection and codification, it has long been recognised that a (conventionally named single) language may have more than one standard, oriented to different centres (pluricentric standards: Clyne 1992Clyne , 1995Clyne and Kipp 2011;Muhr and Clyne 2012;Muhr et al. 2013), as in the case of major European languages, whether in post-colonial contexts (English, French, Spanish, Portuguese), across multiple states within Europe (German), or within single states, as in the case of Irish, for which the more inclusive notion of a 'polynomic' standard has been advocated (Ó Murchadha 2016). 8 (3) Concerning both the elaboration and maintenance of a standard, language purism (defined by Thomas 1991, 12 as 'an aspect of the codification, cultivation and planning of standard languages') is a response to language contact, where multilingual individuals introduce borrowings from one language into another. 9 Language purism was, accordingly, one of the first areas of language standardisation studies to recognise its ideological and discursively constructed nature. 10 (4) Fourth, codification itself often takes place, even in its early stages, less for competent speakers than for the benefit of language learners, beginning at the latest in Europe with grammars for Romans learning Greek. Materials produced for such learners have had a significant influence on language codification (see e.g. Langer 2002;Law 2003;McLelland 2008McLelland , 2012McLelland , 2020. Today, too, an important pressure to codify new standard varieties of minority and/or minoritised languages comes from 'new speakers', who have not acquired the language in early childhood, but on whose successful learning the survival of the language may largely depend. New codifications may, however, leave native speakers marginalised by a standard created for and main- varieties, and contexts has always depended on language contact among multilingual populations. The desire to standardise vernacular European languages in early modern Europe on the model of Latin is one obvious historical example. Today, the studies in Makihara and Schieffelin (2007) show the same process in exemplary fashion in the Pacific region, characterised by 'multilingualism in vernaculars, lingua francas, and colonial and national languages' (Makihara and Schieffelin 2007, 5). Riley (2007), for example, examines 'shifting ideologies' in the relationship between French and indigenous languages. The same is true in Europe, too, as noted above. A standard language ideology newly transferred to regional and minority languages or dialects may reinforce speakers' sense that they do not speak their language 'properly'.

Introducing this special issue: standardisation and multilingualism
Recognising the importance of multilingualism to language standardisation studies underpins the research programme of Strand 2 of the MEITS project (Multilingualism: Empowering Individuals, Transforming Societies, 2016-2020), Standard languages, norms and variation: comparative perspectives in multilingual contexts. This themed issue is its first outcome, based on a workshop in Nottingham in July 2017. While the studies are contemporary in their orientation, they also show the value of historical depth, whether in Huang's study of Chinese in the UK, Kudriatseva's account of the complex history of Ukrainian and its status in relation to Russian, or Denman's tracing of the history of Russian standard language ideology. The six methodologically and epistemologically diverse papers present case studies of the consequences of language standardisation from the UK, Ireland, Italy, Sweden, Ukraine and Switzerland, whether in historically multilingual settings (Switzerland, Ukraine) or in diasporic and immigrant multilingualism (Ireland, Italy, Sweden, and the UK). The languages involvedincluding Chinese in the UKencompass, in Vogl's (2012) typology, instances of 'early' standardisation (Swedish, English, German, Russian), 'middle' standardisation (Ukrainian), and 'late' standardisation (Chinese, Romansh). They also include instances where the process and ultimate outcome of standardisation are contested and, even, a source of social conflict ('disputed' in Vogl's terms): Romansh, Chinese and Ukrainian. Together, the papers constitute a response to the call by Ghyselen, Delarue, and Lybaert (2016, 85) for a 'diversification and integration of methodological perspectives' and, in part, show what quantitative data can offer to studies of language standardisation ideology (see Vogl 2018, 191), alongside more familiar qualitative, often discourse-analytical approaches. I also use the opportunity of introducing the papers to examine critically key concepts that have been used to critique or problematise the historical erasure or invisibilisation of multilingual experience (heteroglossia, translanguaging), as well as methodological factors which may, unwittingly, perpetuate that ideological stance.
The author of the first paper, Raphael Berthele, Professor of Multilingualism at the University of Fribourg's Institute of Multilingualism in Switzerland, comes from a research setting 'committed to researching multilingualism and its linguistic, social, political, economic and educational aspects' (to quote the website of his Institute). In the Swiss context, multilingualism at the level of the state is a 'given' and individual plurilingualism is an expectation. That historically relatively stable and predictable multi-and plurilingualism has also been joined by new forms of individual plurilingualism as a result of immigration -Berthele focusses on the case of Portuguese migrants to Romansh-speaking areas. Berthele critiques an ideologically 'positive-celebrationalist stance on linguistic diversity', which, invoked as and when it suits, in different ways for different languages, can mask tensions and differences. Combining content analysis and discourse analysis, Berthele shows how the implementation of language standardisationthe choice of whether or not to promote a standard through education and, if so, which standardhas consequences for the prospects of individual languages in this plurilingual mix. Crucially, the hierarchisation of varieties within a language that results from language standardisation interacts with the hierarchisation among languages. In Switzerland, when Romance-language speaking pupils are taught German, they learn German in its standard High German form, even if that means distancing those pupils from the Swiss German dialect that their German-speaking compatriots use in their everyday lives. A similar standardising policy, when applied to Romansh, has, however, proved problematic. Whilst the policy's intent was to strengthen Romansh to expand into new domains, Berthele argues that in the de facto pecking order of languages, Romansh cannot claim the domains already occupied by German, which sits firmly above Romansh in the Swiss multilingual hierarchy. The well-intentioned policy to position Romansh so that it might expand its functional range is thus doomed, and is, furthermore, distancing the minority language from its function of expressing identity. Here we have, writ large, the differential effects of language standardisation on different kinds of languages with different positions in a multilingual hierarchy, and a powerful demonstration of why language standardisation studiesand the language policy decisions that they may informmust examine their objects in their wider multilingual contexts.
Julia Forsberg et al.'s exploratory study of self-assessments of language proficiency amongst multilingual adolescents in Sweden broadens the perspective by drawing on empirical data as evidence of language ideologies, applied both to 'elite' (the result of formal language learning) and 'folk' multilingualism (the result of circumstance). 11 The interviews that provide the data, originally conducted for the purposes of a sociophonetic study, reveal differences in how the adolescents interviewed assess their abilities across the different languages in their repertoires, when asked to rate their own proficiency on a scale from 'very poor' to 'very good' for each of the 'four skills'in Swedish, English and, often, one or more other home languages. Respondents were most likely to pass judgement on their Swedish, discussing their 'good' and 'bad' use of the language, slang and dialect in ways that they did not do for their other languages. Forsberg et al. also find some evidence of elements of the standard language ideology applied to Englishan important school subject of the 'elite multilingualism' typewhile other home languages ('folk' multilingualism) were not treated in the same way. The authors suggest that for their other languages, the multilingual respondents lack membership of a wider speech community (with a set of shared norms), sustained formal education in the language, and, perhaps, an identified reference group or 'referee group' (in the terminology of Forsberg et al.) whose language is taken as the standard. While not surprising in themselves, these findings are nevertheless valuable evidence of the process by which individuals acquire a speech community's ideologies, and of the differential ways in which the same ideology of correctness may be applied by individuals to their different languages. Forsberg's preliminary findings thus complement those of the large quantitative study by Vogl (2018) of multilingual university students in seven European countries, showing how their beliefs about multilingualism are differently shaped by standard language ideologies.
Kudriatseva's paper demonstrates how different ideologies may be applied to different languages in multilingual contexts. Kudriatseva's methodological approach combines a discourse analysis of public discourse and policy documents (firmly in the tradition of much work on ideology) with quantitative survey data (as in Vogl 2018). Her data demonstrate a tension between 'two views of languageas a marker of identity and as a communication tool' (Kudriatseva, this issue), analogous to the tension identified by Vogl (2018) among her multilingual respondents between language as indexical of identity and language as a skill. Public discourse and recent policy in Ukraine express an ideology of linguistic separatism, requiring that Ukrainian and Russian be kept separate, both because Russian is viewed as a possible threat to Ukrainian and because of an ideology that bilingualism itself is suspect. Yet young people's responses to Kudriatseva's survey in the bilingual Kherson region, which borders the Crimean region annexed by Russia, reflect rather different conceptual frameworks to those found at the level of policy and public discourse. They present 'an essentially positive evaluation of diversity expressed in the multiple-function view of multilingualism, where competing languages are assigned different and clearly defined roles'. In other words, Kherson pupils ascribe different functions to each of the languagesone instrumental, as a tool (Russian), one doing identity work (Ukrainian). That is, significantly, at odds with the recent (2017) Ukrainian Law on Education, which further expands the reach of the ideology of Ukrainian monolingualism, as a 'strategy of resistance to Ukraine's colonial past', by requiring Ukrainian-language instruction in all public schools.
In research by Kulyk (2007), many Ukrainians surveyedparticularly those who were young, urban, well-educated, and affluentdid not view using Ukrainian as key to their Ukrainian identity.
Many reported mainly speaking Russian, even though they were supportive of policies that support Ukrainian, and might even be more supportive than those who reported mainly using Ukrainian. That is, participants' language practices did not necessarily align even with their own language attitudes and language policy preferencesa complex situation. What happens, then, when people from Ukraine come to a new context where, as migrants, they are viewed externally as a single coherent group? A recent study in Nottingham by Harrison (2019) of the Ukrainian diasporic community in the UK is instructive. 'The' Ukrainian community consists of two distinct waves of migration (post World War II and post-glasnost), who brought with them, at different times, quite different varieties of Ukrainian. The second wave, having lived through a period of Ukraine as a Soviet Socialist Republic, speak what the first wave consider a Russified, impure variety of Ukrainian; the more recent migrants in turn judge the language of the first wave to be a 'backward', 'village yokel' Ukrainian. Different members of the group also have different attitudes to Russian. Widely accepted in Ukraine as a communicative tool, Russian is viewed very negatively by some older migrants and their descendants in the UK, even though all groups share the attachment to Ukrainian as a marker of their Ukrainian identity. Similar tensions are identified by Karatsareas (2018) in the Greek-speaking community of London, where Cypriot Greek is perceived by some as 'villagey' (Karatsareas 2018, 13), and in which Standard Modern Greek is the only variety taught in complementary schools where children are taught their community language. Teachers may correct heritage speakers when they use Cypriot Greek features; the use of Cypriot Greek features is also discouraged in informal settings, including the home. As Karatsareas (2018, 13) points out, it is significant that these differences were not detected in earlier research, where explorations of informants' use of Cypriot Greek were implicitly framed in opposition to English only. The variation within Greek (Vogl's hidden multilingualism #1) was thus invisibilised in the very methodology of that work.
Huang's article in this issue also adds to our knowledge of the social heterogeneity within migrant populations (Vogl's hidden multilingualism #4), in this case the UK Chinese migrant population. Huang presents an ethnographic study of a Chinese complementary school in Birmingham. Such complementary schools are important ideological sites (in the sense of Kroskrity 2010, 199), for the transmissionand sometimes explicit metalinguistic reinforcementof language ideologies. That is the subject of Huang's paper.
It is worth pausing here, however, to reflect on terminology and theory. Huanglike the authors of the remaining two papers in this issueadopts the notion of heteroglossia to examine what in the first three papers (and in Vogl 2012) is discussed as 'multilingualism'. As Huang explains, 'Recently, scholars in sociolinguistics have turned to reapplying Bakhtin's notion of heteroglossia for better understandings of linguistic practices in diverse contemporary societies' (Huang, this issue). Many of these scholarslike Huang herselfcome from a disciplinary background in education, and approach their subject from an epistemological position of social critique, an intellectual relative of critical pedagogy (McLaren 2003, cited in Jaspers 2018. As Bailey (2012, 499) explains, heteroglossia, a translation of the Russian term raznorechie, originally coined by literary analyst and language philosopher Mikhail Bakhtin, 'refers to (1) the simultaneous use of different kinds of forms or signs; and (2) the tensions and conflicts among those signs, based on the sociohistorical associations they carry with them' (my emphasis). The term heteroglossia is, then, conceptually more open than multilingualism. It makes no assumptions about whether the different signs, or systems of signs, exist as, or are labelled as, distinct languages. It is also more self-consciously ideological, for, crucially, it already implies a need to attend to inherent tensions and conflicts between signs or systems of signs. 12 Heteroglossia is thus a challenge to the very notion of the sociolinguist as a neutral observer, whose task is to understand the mechanisms behind language variation without necessarily attending to its social consequences for individuals (the concern of the 'applied sociolinguistics' of Piller 2016; see Piller 2001 for the Bakhtinian contribution to studies of multilingualism from an applied sociolinguistics perspective). The term heteroglossia draws our attention to the 'social and political nature of language' (Bailey 2012, 499).
The openness of the heteroglossia concept is matched by, and underpins, another important concept that has come out of the field of language education, translanguaging. As Baker (2012a, 2012b) set out, the notion began as a recommendation for practice in the education of bilingual children, was extended to describe and observe language practices outside the classroom, and ultimately has ideological and, as they assert, 'even political' associations (2012b, 659)a corrective to ideologies of separate bilingualism that have long come under challenge (e.g. Martin-Jones 1988) but remain pervasive. Translanguaging can be helpfully understood as a corrective to the notion of codeswitching. As Bailey (2012) points out, the attention paid by linguists to codeswitching, was, in one sense, welcome affirmation of the experience of bilingual individuals and what they do. And yet, on the other hand, the search for a model to account for 'why' codeswitching occurs rests on an ideology of monolingualism, by assuming that anything that breaks the bounds of a single language is a marked form of communication. Such mixing of signs, in deviating from the assumed norm of one single language at a time, based on a long-established monolingualism ideology, 'must', then, have meaning and be in need of explanation (see Woolard 2004). Outside academia, a preoccupation with codeswitching can reflect a purist fear of harm to the majority language. Davies (2012) observes that the most commented upon characteristic of the language of Turkish-German bilinguals in the media is mixed speech (gemischt sprechen), taken to embody a threat posed by multilingualism to the German standard variety. The concept of translanguaging is, then, an attempt to reposition such language practices as no more in need of explanation than any other communication by individuals using the range of their linguistic repertoires. (However, see Jaspers (2018) for a timely warning on the limits on the concept, both in its most elastic senses and in the ambitious transformative agenda some proponents of translanguaging appear to set themselves.) An ideology of separate bilingualism in pedagogy ('your language is fine, but just not in my classroom/ our school/ this activity now, where we use this other language') may masquerade as, or indeed honestly rest on, the belief that maximum target language exposure is necessary for successful language acquisition. Nevertheless it contributes further to the hierarchisationor 'stratification' in Bakhtinian termsof languages and language varieties that is the inevitable result of the interaction between standard language ideology and multilingualism. That is the theoretical context for Huang's (2016) ethnographic study of a Chinese complementary school, in which she examined the co-existence of three related but competing language ideologies: (i) the standard language ideology that assumes the superiority of a single variety over others; (ii) the ideology of separate bilingualism, which imposes borders between varieties; and (iii) the ideology that accords value to flexible bilingualism and translanguaging. Huang's study illustrates that it is not just mainstream education that seeks to constrain pupils' language practices. In this complementary Chinese language school too, Cantonese teachers in particular impose a Chinese-only regime, explaining to the researcher that 'this is a Chinese school. Plus they don't have other chances to practice their Chinese except for this two-hour session in Chinese school every Sunday' (Huang 2016, 154). However, as in the cases of Ukrainian and Greek, the tension in this Chinese diasporic community is not just in the heteroglossia of English and the community language, but also within the community's language varieties. For the dynamic in the Chinese school that Huang studied must also be seen within the context of the vigorous promotion of Mandarin Chinese (or Putonghua, lit. 'common speech'), both within China and overseas. Huang (2016, 3) explained that most early Chinese complementary schools in the UK originally taught only Cantonese Chinese, 'one of the most influential dialects' whose speakers mainly live in the South-Eastern Province of Guangdong and in neighbouring Hong Kong. In the present paper, Huang refers to the Cantonese 'variety' rather than 'dialect', but in peer reviewing (not for this collection), I have encountered objections to calling Cantonese anything other than 'a language belonging to a different language family' from Mandarin Chinese [i.e. Yue], The reviewer urged the need 'to be careful not to let language policy rhetoric affect scholarly writing'. Historically, Cantonese has dominated Chinese complementary schooling in the UK, but Huang finds a 'hegemonic standardisation towards Putonghua', while Cantonese, previously the 'lingua franca in Chinese diasporas worldwide for more than half a century', is losing its status as the 'standard diasporic language' (Huang, this issue). Huang uncovers the violence done to Cantonese speakers by the exhortation to 'speak Putonghua' in a staff meeting, where the head teacher justifies his expectation of Putonghua with the same statement as the Cantonese teacher cited above: 'This is a Chinese school'. However, whereas for the Cantonese teacher, the sentence justified requiring Cantonese in her classroom, the head teacher repositions Putonghua as 'the' Chinese: Since we are a Chinese school, also because the majority of us here are Putonghua speakers, let's all speak Putonghua to make it easier. I believe that even though some of our Cantonese teachers don't speak much Putonghua, they have no problems with understanding it.
Huang's sensitive ethnographic study of the Chinese school thus captures, at the micro-level, the lived experience that is the underside of the global promotion of Mandarin Chinese at the expense of other varieties. The enforced silencing of those teachers who 'don't speak much Putonghua' is deemed acceptable, a powerful example of how the standard language ideology translates into linguistic inequality in a multilingual context. The imposition of language standardisation here (irrespective of whether it is viewed as the imposition of one variety over another, or the replacement of one language with another) disenfranchises native speakers, who are, moreover, as teachers, paradoxically the authorities in their classrooms. The case study is thus an even more striking instance, and in a very different, diasporic context, of the marginalisation of native speakers, already noted above for speakers of indigenous languages in Europe such as Catalan and Breton.
A central aim of our MEITS research project is to strengthen modern languages scholars' interactions with other disciplines. Accordingly, alongside the papers from researchers based in education and linguistics, the final two authors in this collection, Denman and Burns, have, like me, Modern Languages as their disciplinary home. Denman's study examines a plurilingual diasporic community, part of a larger project on Russian Speakers in the Republic of Ireland (Smyth 2015(Smyth , 2016. The project title matters, for 'Russian speakers' are not just Russians (as we have already seen in Kudriatseva's study of Ukraine). The project matters, too, in making visible the reality of Ireland as a polyglossic society of diversifying immigration since the 1990s. The participants in Denman's facilitated discussions accept the notion of 'the' Russian language, as a unitary standard language, and exhibit purism towards mixing, towards foreign borrowings, and towards stigmatised (heavily obscene) mat language. Yet, in conflict with this ideology of a monolithic pure standard, they must negotiate the reality of their own heteroglossic practices, accepting a disjunct between cultural identity and linguistic repertoire, where proficiency does not accord with assumed-native identity, including accepting that their children may speak a 'foreigner's Russian'. Particularly important in Denman's study is his meta-reflection on the process of research itself, uncovering instances where heteroglossia was erased from the initial transcription record, and had to be effortfully recovered. Research, too, can be a site of unconscious perpetuation of ideology (Armstrong and Mackenzie 2012; cf. also the critique by Milroy 2001).
Burns' article, part of the project Transnationalizing Modern Languages (see note 4), examines multilingual creative writing and migrant writing in Italy in a 'postmonolingual reality' (Yildiz 2012, 2) to demonstrate that creative multilingualism is not just representation, but also a means of acting upon readers' understandings. The heteroglossia of the migrant and transnational writers in Italy whose work Burns examines challenges 'the hegemony of standard language' (Burns, this issue). In the works Burns analysed, Italian is not just influenced by external language matter ('Arabised', for example, in the words of the author Amara Lakhous, who describes himself programmatically both as a writer Arabising Italian and as an Italianiser of Arabic; cited by Burns), but is also 'Romanised, Sicilianised, Milanised'. That is, authors make space for non-standard Italian in characters' translanguaging, for example to give vent to high emotion, and so give expression to the indigenous multilingualism of Italy. Authors treat the standard language as an instrument of creativity, using it as the 'leaky vessel into, through and out of which other languages can be channelled in multiple directions and forms'. The impact of such literary practices of languaging is potentially transformative, for literature is one of the reference points for the standard language ideology, and historically one of the major sites for a standard's promulgation.

A research agenda for third-wave standardisation studies
The studies in this issue demonstrate the value of a consciously multilingual and plurilingual approach to standardisation studies. By contrast, the first wave of standardisation studies reviewed earlier in this essay were, like first-wave sociolinguistics more generally (Eckert 2012), broadly structuralist in approach. They identified and delimited 'standard' varieties in the language systems of individual languages, and developed comparative descriptive frameworks and taxonomies. More recently, what we might call second-wave studies, often coming out of historical sociolinguistics, have yielded more nuanced analyses of the emergence of language norms and standards, and of language use in local networks and communities in relation to such norms (e.g. Deumert 2004, Elspaß 2005Rutten and van der Wal 2014). In acknowledging the use of multiple, intersecting repertoires, such work has increasingly also begun to examine how plurilingual speakers in multilingual contexts may shape, orient to, or be marginalised by standards, whether in one or more of their varieties (e.g. Hüning, Vogl, and Moliner 2012;Langer and Havinga 2015;Rutten et al. 2017;Pahta, Skaffari, and Wright 2018;including Vogl 2018;Wright 2018), but also how different speakers of the same language may use their different repertoires, in relation to the standard, to construct and negotiate their dynamic identities (e.g. Bermingham 2017; O'Rourke and Ramallo 2017). At the same time, as already noted, recent work returning to the roots of the field in anthropological linguistics has re-emphasised that language standardisation has essentially to do with the transmission and perpetuation of ideology. It is, thus, a specific instance of Eckert's observation that in third-wave sociolinguistics, ideology is located 'in language itself' (Eckert 2012, 98). Combining a sensitivity both to multilingualism and to ideology yields a research agenda for what we might then call (in loose analogy to Eckert 2012) third-wave standardisation studies: a research agenda that seeks to understand how language standardisation ideologies and processes are discursively constructed and enacted in multilingual contexts. In the remainder of this essay, I set out research questions for such a programme of research, under four headings.
(1) What, precisely, are the object(s) of standardisation and of standard language ideology, and to what extent are they the same? As Peters (2017, 364) has noted, among discussions of the standard language, even the basic question of 'whether it refers to written or spoken usageor both is not often addressed', despite calls since the 1990s to avoid the error of conflating the two. Language standardisation studies typically concentrate on the written language (the principal focus of historical language standardisation efforts in Europe). For example, Auer's (2005, 2) definition of a standard variety, which would arguably be accepted by the majority of scholars working on the history of European language varieties, specifies that the standard variety is 'looked upon as an H-variety and used for writing', and is 'subject to at least some codification and elaboration' [my emphasis]. In Lippi-Green's (1994, 166) commonly cited definition, however, standard language ideology applies to the spoken language: it is 'a bias toward an abstracted, idealized, homogeneous spoken language, which is imposed and maintained by dominant institutions and which has as its model the written language, but which is drawn primarily from the spoken language of the upper middle class' [my emphasis]. The tension between these two different conceptualizations has yet to be resolved, but will become increasingly relevant as we take into account contemporary aspects of language standards and standardisations, especially the emergence of spoken language standards in Europe, the varying tolerances of variation within languages, 13 and the fact that languages also vary in how much their spoken and written standards differ (Smakman 2016). Looking beyond Europe also forces us to clarify our thinking. In contemporary China, any aspiring teacher or broadcaster must reach a minimum threshold level in a formal test of spoken standard Mandarin. By contrast, aside from the major reforms that have introduced simplified characters from 1956 (in mainland China, but not in Taiwan), there has been and continues to be relatively little consciousness of, let alone stigmatisation of, nonstandard features in written Chinese. That is the reverse of the situation in at least some European language contexts, such as in German. What, too, of norms in the conceptually oral language that characterises much informal online communication? How can our still prototypically monolingual notion of standard language allow for the possibility that code-switching (or translanguaging) is the norm in some settings? (Smakman and Barasa 2017, 33 suggest the use of Swahili and English in Kenya as one such case.) (2) Do speakers of different languages exhibit different degrees of attachment to a standard language ideology? Does that vary according to language function, and do speakers, in fact, even agree on the functions and domains of the standard (Smakman 2016, 6)? What historical, social and political factors explain such differences? To what extent do multilingual speakers apply different notions of the standard language ideology to their different languages, as Forsberg et al.'s data in this issue suggest, and why? How do the language hierarchies that are the concomitant of multilingualism influence these differences? Are some elements of standard language ideology more or less salient, across time, place, and settings, and how are these best accounted for? Vogl (2012, 36), for instance, suggests that a correctness ideology is 'less pronounced' in the cases of 'late' and some 'middle' standard languages than in 'early' standard languages. In accordance with the movement towards transnationalism in our work, we must also take into account language borderlands (see e.g. Langer and Havinga 2015, and ongoing work by Krogull described in Ayres-Bennett et al. 2019, 74-75); the experiences of new speakers (Smith-Christmas et al. 2017); and diasporic communities such as Russian speakers in Ireland, who are, as Denman points out, living at a significant remove from institutional supports for their language community (see e.g. also Clyne and Kipp 2011;Dossena 2012 on diasporic Scots). What differences are there within speech communities in their attachment to aspects of the ideology of a standard language, as in Vogl's (2018) study where students of certain subjects exhibit a greater attachment to 'correctness'? We must, finally, take into account the largely neglected but possibly substantial proportion of people who complete mainstream education literate and able to use a grammar and dictionary, but seemingly relatively untouched by the standard language ideology of their speech community (McLelland 2020). (3) To what extent, in different contexts, does standard language ideology intersect with, or depend on, conflict with, pre-condition, and/or legitimise other ideologies, both non-linguistic and linguistic? Relevant non-linguistic ideologies include national, ethnic, indigenous, and gender identity ideologies that help define social groups, or political ideologies such as democratisation (of different kinds in, for example, the Soviet Union, China and Europe, and in Western Europe of the eighteenth century and today). Relevant linguistic ideologies include the monolingualism and/or separate bilingualism ideologies (Creese and Blackledge 2011); the ideology of dialect (Watts 1999, 67); correctness ideology (Vogl 2012, 36; cf. also Cameron's (2012) 'verbal hygiene'); literacy ideology (the belief that 'proper' languages are written and have a written literate culture); linguistic purism (Ayres-Bennett 2019); and linguistic prescriptivism (Armstrong and Mackenzie 2015). The study of language standardisation can inform a wider comparative historiography of ideologies of the kind advocated by Blommaert (1999, 1), tracing their diachronic change as part of the history of ideas and of society. (4) What are the sites of language ideologies, in the sense of Kroskrity (2010, 199), drawing on Silverstein (1998, 136)? There may be sites of 'ideological production' that express an ideology through more or less institutional rituals or practices, or sites of 'metapragmatic commentary', where normally implicit ideologies may be explicitly rationalised. Public discourse, state and supra-state policy and public institutions are sites for language ideologies in this latter sense, and so too is education (as exemplified for German as a Foreign Language by Lochtman 2017; see also Train 2007 on 'pedagogical hyperstandardisation' in foreign language classrooms). How do language regimes as 'practices of power' (Costa 2019, 6) reproduce and implement language ideologies in social action, in and through such ideological sites? As Huang's study in this issue makes clear, complementary language schools in diasporic contexts are sites for transmitting and transmuting language ideologies and regimes. Language testing is a powerful site for enacting ideologies too, as exemplified by the existence of tests of Mandarin Chinese for speakers of ethnic minority (and minoritised) groups in China, which have no equivalent in European countries that I know of (on these MHK tests, see e.g. Luo Lian forthcoming). It is significant that the latest Companion volume to the Common European Framework of Reference for Languages, so influential in shaping international language testing regimes and, hence, syllabi, now includes rating scales related to plurilingual and pluricultural competences, mediation, and 'facilitating pluricultural space' (CEFR 2018, esp. 157-162).
Finally, our own research is a site in which ideologies are constructed, perpetuated or challenged, through our concepts and methods, as the notions, discussed above, of heteroglossia and translanguaging, show, and, differently, the contested status of non-Mandarin Chinese as 'varieties', 'dialects' or 'languages'. In research design, Karatsareas' (2018) observation in some earlier work of an unwitting blindness to variation within Greek in the UK community is instructive, while Denman's critique (this volume) of erasure of language variability in the practice of transcription shows the potential impact of ideological assumptions even at the micro-level. Literature and other kinds of cultural production are key ideological sites, too, as Burns's contribution to this volume reminds us, and it is imperative that we continue the dialogue with colleagues working in those areas of Modern Languages research. It is one of the benefits of the MEITS project to have enabled such dialogues and interdisciplinary working in a systematic way (see Carruthers and Fisher (Forthcoming)).
Notes 1. Ayres-Bennett (2019) draws a useful distinction between the neutral term codification and prescription, where the latter encompasses 'a notion of what is right and wrong, correct and incorrect' (Ayres-Bennett 2019, 5), in effect following Milroy and Milroy's (1985) distinction between the two. However, Ayres-Bennett still appears to use prescription and prescriptivism interchangeably. I prefer the term prescriptivism to refer to the ideology (and accompanying activities) that places value judgements on prescriptions and to the belief in the necessity for their widespread implementation. In my sense, then, and following common usage, prescription is synonymous with codification, but choosing the term prescription implies that the codification is interpreted prescriptively. At its most neutral, a codification (whether viewed as a prescription or not) might simply be a verb paradigm; prescriptivism may be observed in the metalanguage present, but also in absences, in the discursive erasure or invisibilization of variants, as when a paradigm does not give all possible variants. 2. It is useful here to clarify the overlaps and differences between attitudes and ideologies. Crucial to the notion of ideologies is their status as internalised hegemonic seeming-truths, serving the interests of a privileged group and excluding others, yet typically taken for granted as commonsensical, also by those subject to the authority or power benefitting from the ideology. Attitudes research and studies of ideology thus come from different research paradigms, although both language ideologies and language attitudes have to do with the evaluation of language. However, in discussing empirical data, the distinction between evidence of ideology and attitude is not always clear in practice (see Vogl 2018, 190-191;Harrison 2019, 84-95). Harrison (2019, 94-95) summarises the overlaps and differences in the concepts as follows: 'Language ideologies are a "macro-phenomenon", shaped by and shared across societies; language attitudes are held by individuals. Language ideologiesas a societal phenomenonprecede and influence language attitudes. Language ideologies are long-term and enduring across a (section of) society.' Language attitudes, while earlier research saw them as stable and durable, are potentially more dynamic and may change according to an individual's circumstances. 3. The shift is exemplified by AHRC-funded projects such as Transnationalizing Modern Languages (2014Languages ( -2016 and The history of the French language in Russia (2012-2014), a major contribution (among many others) to the recognition of historical multilingualism in Europe. The four AHRC-funded Open World Research Initiative projects (of which our own MEITS project is one) all illustrate this shift too. https://ahrc.ukri.org/research/ fundedthemesandprogrammes/themes/owri/ (accessed December 2019). 4. It is worth repeating Vogl's point that support only goes so far. Dialects are explicitly excluded from support in the European Charter of Regional or Minority Languages (1992), something which has encouraged reconceptualizations of the dialect-language boundaryand, in turn, the application of the standard language ideology to new codifications of 'regional languages'. 5. Vogl's other three 'hidden' forms of multilingualismnot ruled out, so much as discursively erased (Gal and Irvine 1995, 973-75) or invisibilised in discourse and/or in deed (Langer and Havinga 2015) are: (1) multilingualism acquired through formal language learning; (2) multilingualism arising from migration; and (3) diversity within populations that share a language, for example the importance of religion as a 'marker of ethnic identity, despite the dominant role of standard languages in self-identification (and partly hidden by their dominant role)' (Vogl 2012, 34). As will become clear below, each of Vogl's four 'hidden' forms of multilingualism is represented in the case studies in this special issue. 6. A diaglossic repertoire is characterised by intermediate variants between standard and (base) dialect. See Auer (2005, 28), for whom diaglossia is one of four types in his 'European dialect/standard constellations', 'a relatively late phenomenon, usually of the late nineteenth or early twentieth century'. 7. Davies and Langer (2006) rightly resist the term destandardisation itself, which mistakenly implies a moment of 'completed' standardisation. 8. Advocating a polynomic standard is an ideological stance which 'aims to treat linguistic variation as an intrinsic good and seeks to identify the social and cultural conditions in which linguistic diversity is maintained or reduced' (Ó Murchadha 2016, 3, with reference to Jaffe 2008). 9. Thomas's distinction between rational-democratic and xenophobic purism, for example, anticipates in some ways the framework of Geeraerts (2003), who outlines rational and romantic models for language standardisation, drawn on by both Denman and Kudriatseva in this issue. 10. For work on the discursively constructed nature of linguistic purism, see e.g. for German, Jones (1999), Stukenbrock (2005); for English, Beal (2012). In the current issue, both Denman's data on the purist attitudes of Russians in Ireland, and Kudriatseva's work on the status of Russian and Ukrainian in Ukraine are relevant. 11. I prefer these terms as the most neutral of the labels discussed by Vogl (2018, 6). 12. As Bailey (2012, 499) explains, one tension among signs is 'between pulls toward a (national) standard form of language -Bakhtin referred to these forces as "centripetal" forcesand forces that push toward various local, non-standard, or demotic forms, which he called "centrifugal" forces'. 13. Vogl (2012, 13) notes, for example, that in Spanish/Castilian, variation within the spoken standard is comparatively more accepted than in the case of French.