State categories, state vision and vernacular woes in Sweden’s language politics

This article deals with the politics of classification in contemporary Sweden. It analyses the language political dispute that has developed over the language political regulation of Övdalsk, a non-standard form of Scandinavian spoken in Älvdalen in northern central Sweden. The analysis focuses on the ways in which a discursive exchange over metalinguistic categories contributes to the efficacy of a state vision of linguistic divisions. In the wake of Sweden’s ratification of the European Charter for Regional and Minority Languages (ECRML), and the language political reforms in which the ratification was embedded, Övdalsk has emerged as a contentious issue. Over three decades (1990s–2010s), the question of what Övdalsk ‘is’—a ‘language’, a ‘dialect’ or something else—has surged repeatedly in political, public and scholarly deliberations (i.e. in expert reports, in policy documents and in scientific publications). Nevertheless, the interests placed in this muddled taxonomic issue have not yet been subjected to any sociolinguistic analysis. Drawing on Bourdieu’s work on the state, the article attends to the ways in which the exchange over Övdalsk has paid tribute to an increasingly entrenched symbolic order. Commenting on Sweden’s commitment to the ECRML more generally, the article accounts for how and why an officialised vision of linguistic division has been rendered symbolically effective. Accordingly, the article argues that a sensitisation to the forms of tacit agreement that underwrite contention is a suitable lens for grasping the maintenance of a political order as legitimate and effective.


Introduction
This article explores an enduring issue in Sweden's language politics. It deals with the dispute that has developed over the political regulation of Ö vdalsk, 1 a nonstandard form of Scandinavian used mainly by a few thousand Swedes residing in the Ä lvdalen municipality in western central Sweden. Classificatory metadiscourses are at the centre of the dispute. Since the mid-1990s, linguists, language policy experts and lawmakers have commented extensively on the question of what Ö vdalsk ultimately 'is': a 'language', a '(Swedish) dialect', or something else. This deceptively ontological 'problematic' is underwritten by language political as well as other interests (see Karlander 2017; see also Silverstein 1996Silverstein , 2015Jaffe 2002;Irvine and Gal 2000;Gal 2006Gal , 2016. It is linked to Sweden's relatively recent language political reforms, in which the European Charter for Regional and Minority Languages (henceforth ECRML) is a key element. Although relatively marginal, the Ö vdalsk question has repeatedly been articulated and addressed in Sweden's language political exchange. 2 As such, it has reiterated several fundamental, long-standing tensions that are bound up with the unification of the linguistic market of the polity.
Engaging with this exchange over Ö vdalsk, the present article seeks to arrive at a comprehensive understanding of the relationship between language political regimentation and symbolic effectiveness for parsimony (Bourdieu 2014: 170). It brings home the point that apparent language ideological contentions may reassert and entrench a unified and unifying language political vision. To this end, it draws on Bourdieu's work on the state, focusing specifically on the self-evidence by which state principles of vision and division impose themselves on the social world. Importantly, the Ö vdalsk language-versus-dialect conundrum is not merely a reflection of a language political order. As much as it has been shaped by regulations of Sweden's linguistic market, it also forms part of the provisions that regiment the regulations of the market (see Bourdieu 1991aBourdieu : 43-57, 1993Bourdieu , 2014Silverstein 1996;Irvine and Gal 2000;Gal 2016). The recent exchange over Ö vdalsk is not merely a universalisation of metadiscourse co-occurring with market unification (see Bourdieu 1994: 8) but an apprehension of this metadiscourse by social agents, who keep reiterating the principles upon which the metadiscourse rests (Bourdieu 1989(Bourdieu , 2004(Bourdieu , 2014; see also Irvine and Gal 2000;Gal 2016). As such, the exchange has rendered the officialised vision of linguistic divisions increasingly dominant, universalised and, hence, effective. Several antagonistic agents have adopted it, recognising it as factual and universally valid. As an effect, the vision 1 Ö vdalsk is also called älvdalska and älvdalsmål (Swedish), and Ö vdalian and Elfdalian (English). In the source texts, the labels vary across authors. I use Ö vdalsk for the sake of parsimony. 2 Exchange refers to the 'thousand little interventions ' (2014: 303) that partake in the construction of a public reality where the state is manifested (Bourdieu 2014). While the present paper only covers a fraction of the exchanges that have entrenched Swedish state power in the field of language politics, it does not limit itself to a specific debate (e.g. in parliamentary politics), a certain institutional sanction (e.g. ECRML monitoring), etc; rather, it aims to cover the exchange over Ö vdalsk exhaustively, not limiting the gaze to a single debate or arena.
has not only been sustained in the exchange over Ö vdalsk, but also by means of the exchange itself.
Unpicking the logic, the present study takes an interest in the discursive labour of agents positioned 'at the intersection of fields where policies are constructed' (Bourdieu 2014: 112). It engages with the textual products of agents who, through their relationship to the field of power, aspire to invoke, shape and reiterate what is tacitly accepted as self-evident. The analytical focus is placed on the body of policy documents, expert committee reports, official statements, opinion pieces and academic texts that have commented authoritatively on the regulation of Ö vdalsk in Sweden's linguistic market. 3 Scrutinising this recent history, the analysis treats the exchange over Ö vdalsk as an element in the formation and consolidation of Sweden's incumbent language political vision. It traces the exchange from its genesis in Sweden's minority language policy overhauls of the mid-1990s (''Beginnings''), via the discursive exchange that preceded its Language Act (SFS 2009:600) (sections ''The Draft Action Programme and the Best Language bill'', ''From guidelines to unifying legislation'') to its political precipitations following this legal landmark (section ''Concord in contention''). Finally, the insights gleaned from these analyses are then summarised. At first, however, I shall further contextualise the case at hand.

A state vision of linguistic divisions
Ö vdalsk has been a recurrent topic in Sweden's language political exchange since the mid-1990s. In broad terms, the struggle over Ö vdalsk is interwoven with the widespread belief that Ö vdalsk is in decline. While representations of Ö vdalsk as threatened by a Swedophone high modernity have circulated for more than a century, such concerns have recently been rearticulated in Sweden's language political exchange. As a language political issue, this representation is bound up with the formation of Sweden's current minority language politics. As such, it has developed in relation to Sweden's adoption of the Framework Convention for the Protection of National Minorities (the FCNM;signed in 1995, ratified in 2000, of the ECRML (signed and ratified in 2000) and of its own unifying legislation on language, condensed in the Language Act (SFS 2009:600) in 2009. In relation to these sanctions, however, its position is obscure. Unlike Sweden's so-called principal language (i.e. standard Swedish), its five national minority languages and the other mother tongues of its immigrants and their descendants, 4 Ö vdalsk lacks any kind of officialisation. It is not mentioned in the Language Act (SFS 2009:600) nor anywhere else in Sweden's unifying legislation on language (see Hult 2012;Salö 2016: 9-11). The same holds for Sweden's other Scandinavian nonstandards-interchangeably labelled 'vernaculars', 'parish vernaculars', '(local) languages', 'genuine dialects' or 'dialects' 5 -which, just as Ö vdalsk, are in use in some of its rural peripheries. However, it is mostly in relation to Ö vdalsk that the distribution of these sanctions is discussed, and sometimes brought into question.
As Heller (2006) has shown, the political idea of linguistic minorities is conjoined with the genesis of the nation-state. Indeed, Sweden's current regulation of its dominated linguistic markets adds to a longue durée of linguistic unification, stretching back to at least the 16th century. At the same time, it goes without saying that these regulations and new modes of intervention both EITHER produce and are produced by OR produce, and are produced by, distinctly contemporary conditions of Sweden's linguistic market. Thus, while Sweden's ECRML ratification marked a definite change in the polity's slowly transforming minority politics, it simultaneously extended and legitimised existing sanctions imposed on the linguistic (sub-) markets of Sweden's 'national' minorities (see Hult 2004;Elenius 2007;Lainio 2015). As this tension indicates, the genesis of a certain language policy is bound up with struggles to conserve or transform representations of legitimate language. Official regulation consecrates such representations, investing them with a degree of symbolic and practical symbolic effectiveness (Bourdieu 2014: 170). The stake of officialisation, in turn, commands practices in struggles over instruments of exercising state power over language (see Bourdieu 1991aBourdieu , 1994Bourdieu , 2014. Such struggles may, conversely, endow state-backed policy with enhanced symbolic effectiveness, making it 'work' through the myriad acts of people who 'believe themselves to be the necessary agents of a necessary policy' (Bourdieu 1996: 383;see also 1994. By the same token, Bourdieu speaks of the 'laws of the (linguistic) market' (e.g. Bourdieu 1993: 78-89), but makes clear that such 'laws' are themselves products of history (Bourdieu 2005: 89ff).
This refutation of legalism-of the compelling autonomy of rules and laws-runs through Bourdieu's programme for a sociology of the state (Bourdieu 1994(Bourdieu , 1996(Bourdieu , 2004(Bourdieu , 2005(Bourdieu , 2014. In this endeavour, the notion of state is closely linked to the notion of field (see Bourdieu 1994Bourdieu , 1996Bourdieu , 2014 and thereby to notions of interest, strategy and contention. The state is not conceived as a self-regulating automaton or an ideological apparatus (Bourdieu 1994(Bourdieu , 1996(Bourdieu , 2014, but rather as an 'ensemble of fields that are the site of struggles [over] the power to constitute and to impose as universal and universally applicable' (Bourdieu and Wacquant 1992: 112). As such, it is bound up with battles between contentious agents, competing for the authority to impose their vision through the state (Bourdieu 2014: 99-101). Thus, while the state has been likened to a 'central bank for symbolic capital' (Bourdieu 1994(Bourdieu : 12, 2014, this simile is only appropriate insofar as we acknowledge that the members of the bank's executive board are likely to nourish many, often conflicting, interests and agendas (Bourdieu 1994(Bourdieu , 1996(Bourdieu : 173-189, 2005(Bourdieu , 2014. Rather than being a unitary entity, the state is a site where different holders of power strive to establish their power as the supremely legitimate one (Bourdieu 1994(Bourdieu , 2014. It is a battleground for agents from different fields, who compete for the power to exercise legitimate control over fields and markets (Bourdieu 2005(Bourdieu : 33, 2014Bourdieu and Wacquant 1992: 111-115). A potent form of state power is, accordingly, the symbolic power to legitimately define and impose principles of vision and division on the social world (Bourdieu 1994(Bourdieu : 7-9, 2014). Yet, this power is not inherent in a given social position, but is located in a struggle between certain socialised agents who, armed with certain inculcated principles of vision and division, occupy certain positions in social space.
In the same vein, Bourdieu stresses that the state exists 'through its effects and through the collective belief in its existence, which lies at the origin of these effects' (Bourdieu 1994(Bourdieu , see also 1977(Bourdieu : 159ff, 1996(Bourdieu , 2014. The state is both an objective structure and a shared representation. As such, it exists simultaneously as divisions of the social world and as principles of division, principles of vision and systems of classification (Bourdieu 1989(Bourdieu , 1991a(Bourdieu : 127ff, 1996(Bourdieu : 271-272, 373ff, 2014; see also Brubaker 2002). Several such schemas of classification pertain to language, recognised as oppositions between 'central' and'local', 'standard' and'vernacular', 'language' and'dialect', andthe like (Bourdieu 1991a, 1993: 78-89;Silverstein 1996Silverstein , 2015Irvine and Gal 2000;Gal 2006Gal , 2016. As part of the unification of Sweden's linguistic market, Ö vdalsk has come to be ordered within such classificatory visions (see Thelander 2011;Karlander 2017). In the wake of Sweden's language policy overhauls, Ö vdalsk has been increasingly bound up with the classificatory logic of the ECRML. While Sweden's adaptation of the ECRML does not extend formally to Ö vdalsk, the language political exchange over Ö vdalsk keeps invoking the ECRML's alluringly mundane system of categories, as made explicit in the first article of the Charter (ECRML 1a; see also Gal 2006: 379). It is instructive to revisit this passage in extenso (excerpt 1).
'regional or minority languages' means languages that are: 1. traditionally used within a given territory of a State by nationals of that State who form a group numerically smaller than the rest of the State's population; and 2. different from the official language(s) of that State; it does not include either dialects of the official language(s) of the State or the languages of migrants Policy research on the ECRML (Grin 2003;Nic Craith 2003;Romaine 2013) has tended to focus on the intents and motives underlying the Charter's coming into being, pairing this interest with analyses of its state-level implementations. While this modus operandi is apt for identifying gaps between policy texts and actual institutional practices, it is less suitable for grasping other processes through which a vision of language and sociolinguistic life becomes imposed upon, and is potentially effective in, a given linguistic market. The effectiveness of such vision depends not on the will of the state but on myriad acts in which the state is manifested. It is conditioned by the concord between minds shaped by the state, and thus by structures that the state tends to disguise (Bourdieu 1994(Bourdieu , 2005. As a consequence, the imposition of the state's vision comes with a certain degree of self-evidence. As Bourdieu stresses, a state-imposed symbolic order is maintained as an 'obsequium' (Bourdieu 2014: 34-36) and as a 'spectacle of universality' (Bourdieu 2014: 28), that is, upheld through acts that constitute certain elements of the social world as taken for granted and, thus, located outside the scope of disagreement (Bourdieu 2014: 184-185). Through such practical adhesion to recognised principles of vision and division, an officialised truth becomes established as an omnipresent consensus, to which even radically disagreeing agents subscribe. In this vein, the state is manifested as a kind of naturalisation that makes a large number of social practices, institutions and divisions seem selfevident, to the extent that any explanation of them will appear unnecessary (Bourdieu 2014: 34-36, 105-119).
The language political exchange over Ö vdalsk must also be viewed through this lens. Here, the ECRML's categorical divisions have been rendered symbolically effective through the work of agents who maintain a sufficiently recognised relationship to the official able to move from the descriptive to the normative, thus transforming an observation into a norm (Bourdieu 2014: 34). Tellingly, the antagonistic agents who have invested in Ö vdalsk as a language political 'problem' have all maintained a deep respect for a shared set of principles of vision and division, thereby rendering these principles respectable and legitimate. Both dissenting and consenting commentators have reasserted the officialised categorisations entrenched in the Swedish language political vision. As a sign of this respect, the exchange over Ö vdalsk has revolved around metadiscourses of categorisation, rather than around any other imaginable issues, such as rights, inclusion and policy content. This inculcated preference has, arguably, established the state vision of linguistic divisions as increasingly legitimate, entrenching it through practice as a universally accepted reality. Dispositions for thinking with and acting through state categories are not simply bad intellectual habits of certain state agents (Brubaker 2002: 166). Rather, they are a necessary condition for rendering these categories politically effective.
Bearing this in mind, we turn to the exchange over Ö vdalsk.
The position of Ö vdalsk in Sweden's language politics Beginnings As a language political issue of current relevance, Ö vdalsk is interlinked with relatively recent efforts to further unify, centralise and homogenise the regulation of Sweden's linguistic market (see also Wingstedt 1998;Hult 2004;Stroud 2004;Hyltenstam 2005;Elenius 2007;Milani 2007a;Lainio 2015;Salö 2016). In 1995, concurrent with Sweden's European integration, the Swedish government decreed the formation of the Minority Language Committee, a parliamentary commission tasked with investigating 'whether and in what manner' Sweden should espouse the European Charter for Regional and Minority Languages. Its investigatory work brought together members of parliament, state bureaucrats and legal experts, as well as representatives of the five minority groups that eventually came to be recognised as national minorities. Imposing divisions was central to its mission. Citing the first article of the ECRML (see excerpt 1 above), the government ordered it to 'report which languages that could be considered the minority languages of our country' (Dir. 1995:84). The mission was later expanded to include an investigation of whether Sweden should ratify the Framework Convention for the Protection of National Minorities (FCNM) (Dir. 1996:77). In 1997, the Minority Language Committee presented its findings in a two-volume report. Under the heading Steps toward a Minority Policy, it dealt separately with a prospective Swedish ratification of the ECRML (SOU 1997:192) and of the FCNM (SOU 1997:193). Several themes in the exchange over Ö vdalsk emanate from the discursive labour accomplished in the report's ECRML-oriented volume (i.e. SOU 1997:192). The Minority Language Committee acknowledged that classification was woven into the 'principal questions that it had to consider'. One of these was 'the question of when a certain mode of expression [uttryckssätt] is to be regarded as a language, and when it is to be regarded as a dialect' (p. 59). This question first and foremost precipitated the report's fairly extensive discussion about the distinction between Meänkieli/ Tornedalian Finnish and (standard) Finnish (pp. 106-119; Hyltenstam 1997). Secondarily, it was actualised in reflections on the essence of 'vernaculars' in 'a Swedish context ' (pp. 93-94, 96-97). While brief, these latter passages have been recalled repeatedly in language political exchanges on Ö vdalsk. In its reflection on the ECRML's vision of linguistic divisions (i.e. excerpt 1), the report noted that 'as far as Sweden is concerned, the question arises of how genuine dialects or vernaculars shall be regarded in relation to the Charter' (SOU 1997:192: 94). While claiming that 'vernaculars' such as 'Ö vdalsk, Gutnish and Scanian' could 'be linguistically very different from contemporary standard Swedish', the Minority Language Committee raised doubts about whether the ECRML could be applied to 'vernaculars'. Although 'vernaculars' were deemed to form 'an important part of Sweden's rural and cultural history', the report found it 'dubious' that they 'could be regarded as languages'. 'The general view', it stated, was 'probably that vernaculars should be considered dialects' and that they therefore were exempt from the Charter's definition of regional and minority languages (p. 94).
The report's proposal juxtaposed the Charter's differentiation of 'languages' and 'dialects', and the mythical presuppositions upon which it was contingent, with an already established vision of the 'vernacular', entrenched in 19th century Scandinavian dialectology and its intellectual descendants. Imagined as the 'antithesis' of the nationalised Swedish language (e.g. Lundell 1887: 273), such tropes of the non-standard point to representations of a rural premodernity, temporally and sociolinguistically set apart from the sanctioned language(s) of the unified linguistic market. This chronotopic embedding was latent in the report, which effectively construed 'vernaculars' as traditionalised and largely anachronistic ways of speaking, sidelined by a full-flung high modernity. The report conceded that 'vernaculars used to be relatively widely spoken in various parts of the country' but that 'increased social mobility, urbanisation and the rise of mass media' had confined them to 'a decreasing number of individuals'. This logic conceived vernaculars as 'cultural heritage', the current existence of which appeared to be dwindling, but which, at the same time, appeared unfit for institutionalised 'modern-day' usage (SOU 1997:192: 94).
In the report, the purportedly folkloric essence of 'vernaculars' was presented as the cause of the language political problematic of this class of language objects. On the one hand, 'vernaculars' were construed as something genuinely Swedish, as a class of linguistic objects that always had been spoken by Swedes, emblematised by the rural peasant-type 'folk' of a national creation myth (Crang 1999;Klein 2006). On the other hand, the perceived affinity between 'vernaculars' and Sweden's 'canonized cultural history' of its imagined premodern origins unsettled the presupposed representation of modern monolingual majority Swedes, whose Others the report had set out to regiment. The 'linguistic difference' between 'vernaculars' and 'Swedish' undermined the implicit link between the standard language and the Swedish national identity. By differentiating and concomitantly reclassifying all 'vernaculars' as 'dialects' (without further specification), the Swedes who presumably spoke them were not differentiated, but remained symbolically located within an ethnolinguistically coherent community of Swedes. Through this manoeuvre, Swedish, which already had been contrasted with 'immigrant languages' and 'minority languages ' (i.e. SOU 1997:192), was reasserted as the 'unifying emblem' of a unified linguistic market, and its emblematised-Swedish-speakers were indexically 'swept up into participation in its cultural expression' (Silverstein 1996: 286) regardless of what and how they actually spoke.

The Draft Action Programme and the Best Language bill
In February 2000, Sweden ratified the ECRML, instituting Sámi, Finnish and Meänkieli as minority languages, and Romani and Yiddish as non-territorial minority languages (SÖ 2000:3). The ratification made no reference to Ö vdalsk, nor to 'vernaculars', 'genuine dialects' or any other trope of the non-standard. Arguably, Sweden's ECRML ratification marked a shift in Sweden's language politics. In its wake there followed a period of language political reform work concerned mainly with the Swedish language. Focusing primarily on the regimentation of the institutional use of Swedish in an increasingly globalised national linguistic market, it gained symbolic force through the formation of the Committee for the Swedish Language, a parliamentary commission tasked with constructing an 'action program for the Swedish language' (Dir. 2000 Throughout the 586 pages of the Draft Action Programme's main text, the language politics of 'dialects', 'vernaculars' and other tropes of the non-standard is a minor theme (SOU 2002:27: 173-180, 301-326). The Programme repeated the indexing of 'dialects' as something Swedish established in the Steps towards a Minority Policy report. It loosely differentiated 'dialects' that 'differed so much from standard Swedish' that they could 'practically be considered different languages' (SOU 2002:27: 306) from less divergent, 'levelled' and 'regional' 'dialects' (p. 305). These differentiated and differentially categorised linguistic objects were nevertheless construed as 'dialects of Swedish'. In this vein, 'dialects' were construed as 'linguistically different' from standard Swedish, but were at the same time symbolically tied to the standard. They were addressed as non-standard, differentiated from the standard by virtue of their 'linguistic' properties but were nevertheless considered indexically Swedish (pp. 173,174,178,306). Hence, the Programme placed 'dialects' on a par with the standard, conjoining them as an abstractly variant, ethnoculturally rationalised yet inherently Swedish linguistic whole (p. 305; see also pp. 173, 178-180). 'Swedes' in this view shared 'a long history as a well-established, clearly definable language community, recognised by others' (p. 46). Swedish-a professed totality encompassing both the standard and an array of varied 'dialects'-was imagined as their unifying emblem. Albeit variable and diverse, Swedish was imagined as a cohesive object. 7 In this vein, 'dialects' were apprehended as part of a shared Swedish cultural heritage (e.g. p. 173). They were presented as 'a resource' that 'contributed to the diversity of the [Swedish] language', since they were 'bearers of the local culture, history and traditions of various [rural] areas' (p. 46). In this spirit, the Draft Action Programme saw this 'cultural heritage' as 'threatened'. It stressed the importance of 'improving the dialects' survival-chances' (p. 180) and suggested that measures should be taken to investigate the possibility to include dialects in school curricula (pp. 306-307). Yet, it provided no clues about the scope of such investigation. More prominently, it emphasised the importance of nurturing positive attitudes towards 'the Swedish language as well as [its] different varieties' (p. 177) and that 'everyone should have their language respected, including dialects [sic]' (p. 563).
This discourse on 'dialects' was reiterated in the Best Language bill's 'suggestions and appraisals' regarding the future of Sweden's language politics (Prop. 2005/06:2: 55-56). Just as the Programme did, the bill maintained that 'the dialects should be endorsed'. It stressed that 'dialects represented a linguistically interesting verbal tradition' and that they were 'an important bearer of our common cultural heritage'. While unspecific about its suggestion to 'support' 'dialects', it deplored the 'fact' that 'dialects' were undergoing 'levelling in many places, becoming more similar to the standard language'. In parallel, the bill also expressed a concern that 'increasingly fewer children' of dialect-speaking areas were 7 The parts of this cohesive whole were nevertheless valorised in different terms. The report (SOU 2002:27) traditionalised 'vernaculars' and 'genuine dialects' as 'bearers of culture', but addressed 'socalled multi-ethnic youth language' and 'Swedish with an accent' in a more dubious register (pp. 174-180). It claimed that 'some people' viewed the latter ways of speaking as 'inferior Swedish' while others accepted them as 'dialect -although with an unusual background', seemingly at ease with reporting such flagrantly negative valorisations without further reflection. socialised into speaking dialects (Prop. 2005/06:2: 55). However, the Draft Action Program's vaguely formulated suggestion to investigate the possibility of including 'dialects' in school curricula (SOU 2002:27: 306-307) was omitted from the bill. More prominently, the bill construed 'dialects' as an object of science rather than as a 'societal concern', underscoring the importance of continuously 'collecting, archiving, processing and publishing dialect material'. It favoured 'a scientific basis' for 'raising awareness about dialects on scientific grounds', but limited such practices to language documentation (Prop. 2005/06:2: 55-56). Thus, the Draft Action Programme and the Best Language bill presented a relatively unified vision of the vernacular linguistic objects that they held to be radically 'different' from standard Swedish. This category was addressed in strongly positive terms. It was conceived as a 'valuable' 'cultural heritage', whose non-standard speakers deserved public 'respect' (SOU 2002:27: 417, 455). However, with the exception of a suggestion to allocate negligible funds to dialectological research (Prop. 2005/06:2: 56; 2005/06:KrU4), these considerations were not translated into policy text.

From guidelines to unifying legislation
Not surprisingly, the Draft Action Programme's discussion of 'vernaculars' was met with criticism. Dahl (2002), professor of general linguistics, disputed the report's choice of nomenclature for this category. In an expert commentary, he maintained that what the report had labelled 'dialects' were in fact something else. Dahl argued that some of 'the language varieties' that the report 'conceived as ''Swedish dialects'' were not [part of] the Swedish language in any other sense than that they happened to be spoken in Sweden's territory'. 'Some of them', he reasoned, 'differ more from standard Swedish than do the other standard languages of Scandinavia' (Dahl 2002: 5), begetting the question of 'whether they should be regarded as languages, rather than as dialects'.
The overarching rationale for reordering the category, which Dahl conceived as a set of disparate 'Nordic varieties', was a 'linguistic' one. He stressed that there indeed existed places in Sweden where a speaker of standard Swedish would hardly understand the local '(Nordic) variety'. This constrained intelligibility mirrored the significant structural 'distance' between these 'varieties' and standard Swedish. While Dahl claimed that 'the choice to describe a variety as either a language or a dialect was more of a political question than a linguistic one', his argument rationalised 'dialects' as 'continuant' and 'languages' as 'discrepant'. In this view, languages were considered to be relatively discrete objects with internal continua of variability (Harris 1990). While the imagined Swedish 'cultural community' was not directly questioned, he dissociated its standard language icon from a multitude of 'vernacular' linguistic communities (Silverstein 1998: 412-414). Thus, his tentatively proposed differentiation challenged the previously asserted links between 'vernaculars' and standard Swedish. Although Dahl did not discuss, and hence did not differentiate, Ö vdalsk as a specific object or issue, he noted that the 'varieties of Upper Dalarna', among which Ö vdalsk historically has been ordered (e.g. Levander 1925), presented clear examples of structural 'distance' from standard Swedish. It was only after the publication of the Best Language bill in 2005 D. Karlander that this vision of linguistic divisions, which dissociated (some) 'vernaculars' from the standard, was superimposed on Ö vdalsk. It emerged as part of the routine monitoring of Sweden's implementation of the ECRML.
In 2006, the second triennial evaluation report on the Swedish commitment to the ECRML (CoE 2006: 8) recounted claims from 'representatives of the speakers of [Ö vdalsk]' that Ö vdalsk was 'linguistically distinct' from Swedish. The Committee of Experts of ECRML stated that it had received accounts that Ö vdalsk 'should not be considered a dialect of Swedish, and [that it] thus corresponds to the Charter's definition of a minority language'. The report encouraged 'the Swedish authorities to examine the question and return to it in the next report'. The question of what Ö vdalsk was, as queried by the Committee of Experts, migrated into the Riksdag, the Swedish unicameral parliament. Centre Party MP Kenneth Johansson posed it to the responsible minister, asking whether the government planned to commission 'a thorough investigation of the possibilities of recognising Ö vdalsk as a minority or regional language'. Johansson maintained that Ö vdalsk met the criteria of 'being a language', as specified in the Charter: it was 'an old language, spoken in a defined territory', and 'it differed markedly from standard Swedish' (Johansson 2006a). The minister, Social Democrat Jens Orback, answered that the matter had already been settled through Sweden's ratification of the ECRML. He made clear that 'vernaculars, such as Ö vdalsk, could not be considered minority languages in the [institutional] sense intended by the ECRML', but that 'the preservation of Ö vdalsk was important', since Ö vdalsk 'constituted an important part of our shared [Swedish] cultural heritage' (Orback 2006). In response, MP Johansson filed a motion that proposed an inquiry into the possible recognition of Ö vdalsk 'as a minority or regional language in accordance with the Charter' (Johansson 2006b). Underscoring that Ö vdalsk was 'an old language spoken in a demarcated territory', he argued that it 'met the criteria [for being a 'language'] specified in the ECRML', since it 'differed considerably from standard Swedish'. Along with several other motions on minority politics, it was eventually rejected (2007/08: KU13: 38).
These twitches in the field of power were also manifested in an increasing academic engagement in the regulation of Ö vdalsk through explicit policy interventions. In an op-ed , several Swedish linguists, Ö vdalsk language advocates and politicians 8 expressed their support for Johansson's motion. The group called upon the Riksdag to 'apply the Charter to Ö vdalsk, so that it would become recognised as a language in Sweden'. Arguing that Ö vdalsk was 'severely threatened', they stressed that these measures would 'secure the continued existence of Ö vdalsk' and thereby 'preserve a part of the Swedish cultural heritage'. Here, the performative of categorisation was rationalised as a 'fit' between an autonomous object and a purely denotational label. It was not recognised as performing a categorisation but only as conforming, allocating an autonomously existing object to its proper, autonomously existing category.
This stance was reasserted in a set of papers published under the auspices of the Royal Gustavus Adolphus Academy for Swedish Folk Culture (Hellspong 2007). The heading under which they appeared-'Ö vdalsk: dialect or minority language'-neatly invoked the schema of differentiation utilised in the language political process: 'language' had political entailments, whereas 'dialect' did not. In line with this logic, Professor of Nordic philology Williams (2007: 85) conceived of the classification of Ö vdalsk as a 'language' to be 'largely unproblematic'. In his contribution to the Academy's deliberations, he argued that the matter was settled by the fact that 'the vernacular of Ä lvdalen differed radically from standard Swedish', was 'territorially bounded' and hence was 'an important part of [local] identity formation ' (pp. 84-85). Likewise, Dahl (2007) elaborated on his previous argument about the 'difference' and 'distance' between Ö vdalsk and standard Swedish. He provided '(linguistic) structural', 'functional' and 'diachronic' lines of reasoning that indicated that Ö vdalsk was 'distinct' from Swedish. In brief, Dahl stressed that Ö vdalsk was 'quite unintelligible' for most people outside Ä lvdalen. This relative 'unintelligibility' differentiated it from the otherwise more or less mutually 'intelligible' Scandinavian (standard) languages. Hence, it reflected a distinct structure, distinct patterns of use and a distinct linguistic evolution. Williams (2007: 85), for his part, argued that these criteria merited a political 'recognition' modelled on the ECRML. Given the fact that 'Meänkieli, although being a Finnish dialect, had been recognised as a minority language', Williams argued that Ö vdalsk was more than qualified for similar sanctions. Olle Josephson, at that time director of the Swedish Language Council, Sweden's main national language planning body, was more reluctant. In his reply to Dahl and Williams (Josephson 2007a: 86), he conceded that from 'a structural perspective [Ö vdalsk] was probably different enough from other Nordic varieties to qualify as a language'. However, Josephson (2007a: 86) claimed, this axis of differentiation was not immediately relevant for the implementation of the ECRML. In his view, 'functional' criteria trumped 'structural' criteria, since 'ausbau' had thus far been privileged over 'abstand' 9 in Sweden's minority language politics. Replying to Williams, Josephson argued that Ö vdalsk did not yet qualify as an 'Ausbausprache', whereas Meänkieli, as it was implied, did.
It should be underscored that the ECRML itself provides no definitions whatsoever of the categories it brings into play. Neither 'languages' nor 'dialects', nor the relationship between them, are defined or explained. The contention over Ö vdalsk is a clear example of how an intrinsically muddled scholastic taxonomy likewise remains obscure outside the 'hermeneutically sealed zone of homogenous linguistics' (Hutton, quoted in Harris 1990: 14). Yet, this fully universalised vision of linguistic divisions, inscribed in the linguist's habitus (Bourdieu 1991a: 54; see also Harris 1990), was manifested in a range of authoritative rationalisations of what Ö vdalsk was. As such, they constituted acts of cognition that enforced a recognition of the categories that they bespoke (Bourdieu 1989: 22), that is, a recognition of the universality and legitimacy of a continuous symbolic order. With due precision, they aligned with the dominant principles of differentiation-the categories entextualised in the ECRML-which were gradually brought to bear on Sweden's linguistic market. The exchange over the categorisation of Ö vdalsk, thus, seems to have operated as a 'structured and structuring medium', which successfully imposed 'an apprehension of the established order as natural through the disguised imposition of systems of classification' (Bourdieu 1991a: 169). The struggle over classification legitimised the principles of classification.
Tellingly, Josephson (2007a: 90) argued that Ö vdalsk must be viewed 'in relation to Sweden's entire linguistic situation, and its coordinated politics of language'. Citing the Best Language bill (Prop. 2005/06:2: 55), he stressed that 'one or several shared national languages, manifested in numerous different varieties, dialects and sociolects' were 'commonly viewed as a resource for a country'. Expressing concern over Sweden's lack of a 'politics of dialects'[dialektpolitik], he questioned the rationale for 'reconstructing' Ö vdalsk 'as a minority language' in order 'to save it'. To turn Ö vdalsk into a separate language would mean dissociating 'Sweden's most remarkable dialect' from other diverging, but nonetheless indexically Swedish, ways of speaking, such as 'Gutnish, Pitemål, Burträskmål or Rinkeby Swedish' (p. 89). Similarly, but with less subtleness, Nordic linguist Reinhammar (2007: 83) stated that 'we cannot have any Swedish minority languages in Sweden'. In her view, ethnicity, rationalised in linguistic terms, trumped all other schemas of differentiation. Swedes, albeit linguistically divergent, could not be made into a (linguistic) minority. Josephson (2007b), for his part, argued that the debate over Ö vdalsk was indicative of Sweden's 'obsolete' linguistic legislation. In his view, the exclusion of Ö vdalsk from Sweden's restrictive ECRML-based minority language policies was principally correct (see also Josephson 2007a). At the same time, he noted that Ö vdalsk, just like any other 'dialect', ought to be included in the Swedish language political framework, albeit it not in the guise of a minority language subject to the provisions of the ECRML (Josephson 2007a, b). While he deemed that Ö vdalsk was not the 'language of a minority' (Josephson 2007a: 87), it was the 'absence of a Swedish language legislation', that is to say, policies concerned with the Swedish language, that had caused Ö vdalsk to be left unjustly unregimented (Josephson 2007a: 91, b). 10 However, the Language Act Commission report (i.e. SOU 2008:26), which recommended implementation of a Language Act, did not redeem this situation. It made no explicit reference to Ö vdalsk, or to any 'vernacular', but noted simply that the ECRML could not be applied to 'dialects' (p. 139).
In an opinion piece in Språktidningen, Sweden's largest popular publication on language and linguistics, Nordic linguist Rosenkvist (2008) criticised the Commission report's erasure of Ö vdalsk. He argued that although Ö vdalsk 'was dying', those who held 'the capacity to reverse the process'-that is, the Language Act Commission, the Swedish Language Council and the Riksdag-'had done nothing', thus 'leaving [Ö vdalsk] to die silently' (p. 86). His protest, too, relied fundamentally on the claim that Ö vdalsk was a 'language'. Listing several qualifications, he stressed that Ö vdalsk 'differed from Swedish on every level: phonetically, lexically and grammatically' and was 'not intelligible for a speaker of Swedish'. It was 'autochthonous to Ä lvdalen', where it was 'perceived to be a language' (p. 85). Affirmative legal sanctions would, hence, not only be just, but also effective in 'supporting the language [Ö vdalsk]'. In the following issue, Dahl (2008) commented on the matter, concluding that 'there were no objective or eternally true answers to the question of whether a certain form of language was a language or a dialect'. Rather, he perceived it to be a 'practical issue', querying instead which classification would best serve the 'continued existence' of Ö vdalsk. 'Maybe', Dahl reasoned, 'it would be best to leave [Ö vdalsk] as a spoken language variety, instead of making it a language with a standardised orthography and [legal] status as a minority language'. Remaining inconclusive, he added: 'but maybe that is exactly the support Ö vdalsk needs'. In a commentary on Rosenkvist (2008) and Dahl (2008), linguist Aktürk (2009) stressed the 'broader' nature of the issue. While 'dialects', in his view, were 'stigmatised and [institutionally] unsupported ', a 'preferential [legal] treatment' of Ö vdalsk threatened 'to create an imbalance between different linguistic varieties' in Sweden. Just like Rosenkvist (2008) and Dahl (2008), he perceived officialisation to be a symbolically powerful intervention that should be handled with due caution.
In spite of these interested interventions, Ö vdalsk was not mentioned in the government bill Language for All-Proposal for a Language Act (Prop. 2008/09: 153), which prefigured the drafting and eventual implementation of the Language Act. The issue was granted some attention in the concomitant government bill From Recognition to Empowerment (Prop. 2008/09:158: 60-62), which presented new political strategies concerning the five national minority groups recognised under the ECRML. The bill dismissed an inclusion of Ö vdalsk in Sweden's language political framework. It clarified that 'the government found no reason to reach a different conclusion than that reached in the ECRML ratification'. Therefore, 'no new minority languages or minorities were currently on the agenda' (p. 62). In relation to this move, Ö vdalsk was categorised as 'an example of a Swedish vernacular'. This class of objects 'could be linguistically very different from contemporary standard Swedish' (pp. 61-62) but, the implication went, was nevertheless something Swedish. Contingently, the government stated that it saw 'a great value in preserving Ö vdalsk as a part of the Swedish cultural heritage, regardless of whether it is a language or a dialect'. It rhetorically downplayed the importance of differentiation, while simultaneously asserting it in practice.

Concord in contention
While the passing of the Language Act in 2009 has been interpreted as an endpoint in debates over the Swedish language (Milani and Johnson 2008;Salö 2014), the exchange over Ö vdalsk has transcended this point of relative discursive closure. Although the passing of the Language Act was the acme of an extended language political process, it did not resolve the argument over Ö vdalsk. Rather, it further sanctioned an already imposed vision of divisions, thereby naturalising its own arbitrariness (Bourdieu 1977: 164). Swedish, brashly undifferentiated, was officialised and universalised as Sweden's principal language. As an act of universalisation, it further legitimised the differentiation of 'language' from 'dialect' and from 'vernacular'. It officialised a finite set of categories and associated them with a finite set of linguistic objects. Yet, this sanction did not bring to a close the exchange over Ö vdalsk.
To be sure, the question of what Ö vdalsk is has continued to be ubiquitous in Sweden's language political exchange. It has been raised repeatedly as part of the monitoring process of Sweden's commitment to the ECRML. ECRML evaluation reports have repeated the plea to the Swedish government to 'clarify' the taxonomic 'status' of Ö vdalsk (CoE 2008(CoE : 7, 2011, and 'clarifications' have duly been provided in Sweden's periodical reports (SPR 2010(SPR , 2013. Ever since the first parliamentary motion advocating an inquiry into the 'possibilities of recognising Ö vdalsk as minority or regional language' (Johansson 2006b) was dismissed by the Riksdag, the question of whether Ö vdalsk should be re-categorised has continually been raised. Centre Party MP Johansson filed another two motions suggesting an official inquiry into 'the possibilities of recognising Ö vdalsk as a minority or regional language' (Johansson 2009(Johansson , 2011. Both motions stressed that Ö vdalsk was 'a language' and that this claim was supported by 'several leading experts on the matter'. Both were dismissed in the Riksdag. In 2015, three parliamentary motions on Ö vdalsk were submitted to the Riksdag. One proposed 'to officially recognise Ö vdalsk as a language' (Jansson and Richtoff 2015), another 'to investigate the possibility of classifying Ö vdalsk as a national minority language' (Malmberg 2015) and a third 'to investigate the possibility of classifying Ö vdalsk as a regional language' (Emilsson 2015).
Notwithstanding the minor differences between the classifications they proposed, all motions made reference to some crucial evidence of Ö vdalsk being a language. Green Party MP Malmberg (2015) maintained that 'most linguists agree that Ö vdalsk is a language in its own right [ett eget språk]. The distance between Ö vdalsk and Swedish is comparable to the distance between Icelandic and Swedish. Its phonology, grammar, syntax and lexicon differ radically from standard Swedish'. On a similar note, Sweden Democrat MP Emilsson (2015) argued that 'many people consider [Ö vdalsk] a language. It is set apart from modern Nordic languages by its Old Norse linguistic features, and it is difficult for a speaker of standard Swedish [en rikssvensk] to understand it […]'. While MP Emilsson granted the existence of some disagreement over the 'classification of Ö vdalsk', his party colleagues MPs Jansson and Richtoff (2015) were less hesitant, straightforwardly concluding that 'Ö vdalsk is a language, because it is so different so that one has to learn it in order to understand it'. The somewhat surprising concord between representatives of the liberal and nationalist ends of Sweden's political spectrum was thus galvanised with glosses from the language sciences. The arguments for classifying Ö vdalsk as 'a regional language', 'a minority language' or simply 'a language' were delivered in a stylised linguist lingo, replete with references to language structure, cross-linguistic and historical comparison, and language learning. Ö vdalsk was a 'real language' and should therefore be endowed with the same 'institutionalised paraphernalia' as any other 'real language' (see Silverstein 1996: 286). Its objective taxonomic status warranted a change of its political regimentation.
Ostensibly, the gist of the parliamentary motions may lend support to the idea that 'the language/dialect distinction is a highly perspectival issue and essentially not a linguistic one […]', as Fishman (2003: 97) wrote, claiming that 'whenever this distinction is made (particularly in sharply contested individual cases), political and self-interested concerns, rather than scholarly interests, are normally decisive'. Yet, the politics and the linguistics of classification are fundamentally entwined. Metalinguistic categories cannot be grasped from one single point of view, since they are neither purely political nor purely philosophical, but are at once deeply entrenched in both these enterprises (Bourdieu 1991b(Bourdieu : 51, 1996. They circulate on a wider linguistic market, on which the regulation of language continuously unfolds. Significantly, as the case at hand shows, there is a certain division of labour in a certain economy of legitimation (Bourdieu 1996: 384-385). Poised against the whole history of exchanges over the sanction directed at 'vernaculars', which eventually metamorphosed into an exchange over Ö vdalsk, it is clear that parliamentary motions have fed off and into an already dominant vision of linguistic division. Taken together, all historical articulations on the categorisation of Ö vdalsk do not constitute a unique and fully unified discourse. They have unfolded as two more or less distinct stances on institutional sanctions of the linguistic market, either supporting or dismissing an officialisation of Ö vdalsk with reference to its rationalised taxonomic position. These discursive acts have enunciated an increasingly legitimate vision, imposing it on the linguistic market (see Bourdieu 2014: 142ff). In the struggle over the institutional 'status'-that is, the legitimate categorisation-of Ö vdalsk, the contending agents may have quarrelled over established categories, but nonetheless apprehended fully the principles of categorisation foundational to these categories. It is clear that lawmakers, government commissions, policy experts and linguists alike have championed the ECRML's linguocentric schema of differentiation. Most strikingly, agents, who in opposition to the established sanctions have favoured an officialisation of Ö vdalsk, have stressed that Ö vdalsk is a 'language'. In relation to this strategy, they have formulated a range of arguments aimed at proving the 'actual' nature of Ö vdalsk as being that of a 'language ' (e.g. Johansson 2006b;Williams 2007;Rosenkvist 2008). However, such arguments have been met with similar rhetoric from supporters of the established sanctions, according to whom the 'actual' nature of Ö vdalsk was not that of a 'language ' (e.g. Orback 2006;Reinhammar 2007;SPR 2010). As forcefully formulated in a Swedish ECRML periodical report to the Council of Europe (SPR 2013: 18), these agents maintained that the 'general view in Sweden' was that Ö vdalsk 'is a dialect'. While both these stances have been mitigated (e.g. Josephson 2007a, b;Dahl 2008), they correspond to the dominant vision, according to which 'language' and 'dialect' are oppositional, but mutually constitutive, linguistic essences.
Under Sweden's current regime of language, each category carries different implications and potentially also different political entailments on the linguistic market. An officialised, state-supported reclassification of Ö vdalsk as a 'language' could perhaps rejig the regimentation of Ö vdalsk, possibly yielding new institutional entailments for Ö vdalsk speakers. However, it is not yet clear whether an official reclassification of Ö vdalsk as a 'language' would remain a merely 'symbolic' act, or whether it would coerce changes in some set of institutional practices directed at Ö vdalsk. 11 To be sure, the particularities of any such, so far potential, arrangements are yet to be discerned. Nonetheless, such sanctions would, regardless of their content, only constitute an inflection of an already established, deeply entrenched, vision. By way of concluding, we shall dwell on the selfevidence of this way of seeing.

Conclusion
Every society has its own shifting inventory of legitimate social problems 'worthy of being debated, of being made public and sometimes officialised' (Bourdieu 1992: 236). In Sweden, classificatory metadiscourses have come to form part of this inventory. In relation to Ö vdalsk, questions pertaining to categories and categorisation have, arguably, been the most privileged way of talking about language politics and language political interventions. Significantly, the lingering question about the correct categorisation of Ö vdalsk has overshadowed, if not obstructed, less abstract but nevertheless language politically relevant discussions about rights, inclusion and policy content. As a strategy, this display of respect might not preclude a changed regimentation of Ö vdalsk. However, it remains to be seen whether it will be effective to this end.
More importantly, the nearly exclusive interest in Ö vdalsk as a matter pertaining to categories and categorisation has lent symbolic effectiveness to the dominant vision. Indicative of this efficacy, all contention over Ö vdalsk was secondary to an immediate agreement with the established principles of vision and division for which 'even the most critical, the most anarchistic, the most subversive [social agents] displayed a pure respect' (Bourdieu 2014: 34). What is interesting about the struggle over Ö vdalsk, as it has manifested itself thus far in Sweden's language political exchange, is not the antagonisms that it has encompassed, but the antagonists' immediate and tacit agreement on the rules of engagement, which has rendered them unconsciously committed to an increasingly entrenched symbolic order (see Bourdieu 1977Bourdieu : 163-171, 2014). Their practical disagreement over 'correct' classificatory measures has legitimised the principles underlying the measures, as well as the implicit universality of these principles. In this way, the exchange has universalised the categories it has rendered politically meaningful. All agents, regardless of the position they occupied in the exchange, have acted in defence of the state-sanctioned vision. While stances on the regimentation of Ö vdalsk have shifted, the vision of division that legitimately separates 'languages' from 'dialects' has remained unquestioned. Thus, the state has not merely been manifest in the imposition of laws or sanctions, but has been brought to bear on the linguistic market as an incremental universalisation of the dominant vision of linguistic divisions. A minor adjustment of sanctions is unlikely to undermine its dominance. If any of the current parliamentary motions on Ö vdalsk should be approved, the dominant vision will be legitimised. If they should not be approved, the dominant vision will be legitimised.
In this vein, historically produced conceptions about language seem continuously to saturate both scholarly agendas and political activities (see Gal 2006: 377). They are not merely a metadiscourse about the social world, but are inscribed in the social world and thereby come to co-constitute the social world, (Bourdieu 2014: 183). As I have argued here, such exchanges, which have effectively constituted the protracted language political 'problematic' of Ö vdalsk, have partaken in imposing a legitimate vision of sociolinguistic divisions. The struggle over categorisation has engendered numerous obsequious acts of state, with all involved agents venerating the principles of vision and division that have guided them in the battle. Their exchange has not yet transcended its fondness for categories.