Languagised repertoires: How fictional languages have real effects 1

It is now widely acknowledged in a range of linguistic disciplines that ‘languages’ are sociohistorical constructs rather than ontologically real entities. While this insight has contributed in important ways to challenging the monolingual bias in linguistics, a simplistic dismissal of the notion of ‘languages’ is unhelpful when trying to explain its status and function as a sociocultural, metalinguistic construct. This chapter draws on insights from linguistic anthropology as well as usage-based perspectives on language learning to argue that language use always involves an evaluative dimension linked with sociocultural conventions, and that it is such language use that forms the basis of language learning. It is suggested that sociocultural contexts with a strong discursive orientation to ‘languages’ result in ‘languagised’ individual repertoires that mediate the kind of multilingual language use speakers engage in. The theoretical discussion is illustrated by examples from an interview study with highly proficient adult speakers of Finnish as a second language.


Introduction
In the last couple of decades, it has become widely acknowledged in a range of linguistic disciplines that a conceptualisation of languages as ontologically real entities with clear boundaries is an (at least) problematic starting point for the study of language.Scholars have convincingly shown that such conceptualisations of 'languages' as bounded entities are sociohistorical constructs, shaped and promoted in crucial ways by the emerging discipline of linguistics, and deeply entangled in nationalist and colonialist projects (e.g.Blommaert 1999;Makoni & Pennycook 2005).Together with an increasing interest in multilingualism and new forms of linguistic diversity in sociolinguistics, applied linguistics and related fields, these insights have led scholars to describe 'languages' as 'convenient fictions' (Widdowson 2012: 9) or 'ideological constructions which match real-life use of language poorly' (Jørgensen et al. 2011: 23).Such descriptions suggest that there is little, if any, connection between how language is constructed ideologically and how it is used, thus invoking an oppositional image of 'fictional ', 'artificial' and 'monolingual' ideologies of language and 'real', 'natural' and 'multilingual' language use (also see Canagarajah 2011: 3).
As Jaspers and Madsen (2016) have pointed out, alternative conceptualisations of language use beyond 'languages' should, however, also be treated with caution.While challenging ideologies of bounded 'languages' as well as monolingual perspectives in language studies is still a necessary and important effort, it is problematic to assume that once professional linguistic and pedagogical perspectives are freed from ideologies of separate 'languages', the underlying 'fluid' and 'natural' language use of speakers can simply be discovered or encouraged.In a world where the separation and definition of (standard) languages continues to be a common and valued practice in many contexts, speakers can have valid motives to orient toward 'languages' in navigating their everyday lives (Jaspers & Madsen 2016: 237).
On a more fundamental level, as will be argued throughout this chapter, people's ways of using language are always shaped by sociocultural values, whether these include an appreciation of 'languages' or not (also see Jaspers & Madsen 2016: 236).This aim of this chapter is to explore and further theorise the mediating role of 'languages' as metalinguistic (e.g.Verschueren 2000) conceptualisations from the viewpoint of second language learning and use.It does so by bringing together insights from linguistic anthropology as well as usage-based perspectives on language learning.Linguistic anthropologists (e.g.Silverstein 1993) have long emphasised that language has a metalinguistic or metapragmatic dimension, i.e. a dimension involving evaluation of language use itself that helps contextualise talk and that is inseparable from both linguistic structures and from the way people use language.Consequently, there cannot be any simplistic separation of 'real' from 'ideological' aspects of language use.Usage-based approaches to first and additional language learning (e.g.Eskildsen 2009;Tomasello 2003), on the other hand, have brought forward an understanding of language learning as based not on language 'systems' but on the interaction between general cognitive processes and everyday exposure to and engagement with language.This makes it possible to see language competence as organised not along lines of separate 'languages', but along patterns of linguistic practice that are in turn influenced by sociocultural norms in the environments of language learners.I argue that if we accept that metalinguistic reasoning and metapragmatic activity are necessary parts of language use and that such language use forms the basis of language development in individuals, we might be able to reintroduce 'languages' into the picture: not as an a priori and universally valid category but as a sociocultural construct that can be reproduced on various levels, including on the level of individual language competence.
The chapter proceeds as follows: I first recapture some arguments made by linguistic anthropologists on the necessarily social and evaluative nature of language and then bring these arguments into a dialogue with usage-based conceptualisations of language learning and linguistic competence.Following this, I reflect on some particular features of late additional language learning.Finally, based on data from an empirical study, I explore the role of 'languages' in the experiences of highly proficient adult second language speakers of Finnish.I conclude that, at least for some groups of speakers, it is misleading to conceptualise the ideas of separate 'languages' as an artificial ideology exterior to their language learning and use.Rather, an orientation toward 'languages' and the ability to use 'languages' separately can be seen as part of what constitutes their competence.

Language as the 'total linguistic fact'
Language is not a neutral, context-free medium of communication, just as language use cannot be seen as 'innocent behaviour' (Coupland & Jaworski 2004: 36).Language always involves an evaluative dimension with regard to both its structures and, more importantly, to the ways people use it to create meaning.Constructing meaning through language usually involves much more than the relatively fixed, denotational aspects of linguistic forms: it relies on indexical signs that contextualise what is said and that are at the same time contextualised through previous usage (Silverstein 2009: 756;also see Gumperz 1996: 365).Indexical signs, in turn, are linked to each other and to the social world through cultural ideologies, i.e. a shared base of knowledge about social and cultural relations (Silverstein 2003: 196-197).Silverstein (1985: 220) has therefore referred to the object of linguistic analysis as the 'total linguistic fact': 'The total linguistic fact, the datum for a science of language, is irreducibly dialectic in nature.It is an unstable mutual interaction of meaningful sign forms, contextualized to situations of interested human use and mediated by the fact of cultural ideology.' Where does this leave 'languages'?One the one hand, a view of language as an 'unstable' process of meaning-making rather than as a fixed system fits in well with the rejection of the notion of bounded 'languages'.On the other hand, if language use always involves an ideological dimension and cultural conceptualisations of language itself are an important part of this dimension, the claim (in its most rigid form) that 'languages' are fictions imposed on and restricting the otherwise 'natural' flow of multilingual language is difficult to maintain.This is, not least, because ideologically mediated norms are what ultimately enables language use (see Piippo 2012: 196) and other kinds of social action, and deliberate deviation from normatively expected behaviour creates endless possibilities for creating meaning (see Agha 2007: 5).For instance, monolingual norms and ideologies of nativeness and separate 'languages' do not simply restrict what speakers can do in a prescriptive fashion, but also enable them to employ such norms and their deliberate crossing in the construction of hybrid, multilingual identities (Otsuji & Pennycook 2010).
Emphasising the inseparability of language and its ideological and cultural dimension is also important with regard to what is learnt when language is learnt.Children do not acquire linguistic forms or grammar 'rules' alone.Language learning crucially involves the learning of norms of language use (Piippo 2012) and socialisation into a particular sociocultural environment (Kramsch & Steffensen 2008;Ochs 1996).From this perspective, what children acquire is different from competence in one or several language 'systems' and closer to what Hymes (1972Hymes ( /2001Hymes ( , 1985) ) calls communicative competence: We have then to account for the fact that a normal child acquires knowledge of sentences, not only as grammatical, but also as appropriate.He or she acquires competence as to when to speak, when not, and as to what to talk about with whom, when, where, in what manner.In short, a child becomes able to accomplish a repertoire of speech acts, to take part in speech events, and to evaluate their accomplishment by others.This competence, moreover, is integral with attitudes, values, and motivations concerning language, its features and uses, and integral with competence for, and attitudes toward, the interrelation of language with the other codes of communicative conduct […].The acquisition of such competency is of course fed by social experience, needs, and motives, and issues in action that is itself a renewed source of motives, needs, experience.(Hymes 1972(Hymes /2001: 60) : 60) The notion of communicative competence offers an alternative to a Chomskyan understanding of competence as the abstract language knowledge of the idealised native speaker.It is, however, important to note that Hymes did not conceptualise the notion as a mere addition to linguistic competence 'proper', but rather based it on a view of language 'beyond any simple distinction, or dichotomy, between grammar and use' (Hymes 1992: 51).
Hymes' definition of competence also states that 'attitudes, values, and motivations concerning language' (Hymes 1972(Hymes /2001: 60) : 60) are a crucial part of what is acquired by a child.His notion of communicative competence thus implies that language is indeed acquired in the sense of the 'total linguistic fact'.
As linguistic anthropologists have pointed out, many, if not most, contexts of language acquisition defy the simple equation of 'one culture -one language', as 'social networks transcend cultural and grammatical systems to create shared interpretive systems beneath linguistic diversity' (Gumperz 1996: 361).Speakers who are socialised into such a context will 'naturally' possess a competence that includes resources from different 'languages'.
However, as argued earlier, multilingual contexts are not exempt from ideologies of and evaluative stances toward language (see e.g.Lehtonen 2015;Møller 2016).Hymes' understanding of competence as communicative competence suggests that this metalinguistic level is an essential part of what it means to know language in a given context.Arguably, then, in contexts where 'languages' are constructed as socioculturally relevant for using and evaluating language, competence has to include the ability to assign linguistic forms or ways of speaking to 'languages'.
In the following section I turn to usage-based approaches to language and language learning to explore more deeply the connection between individual competence and 'languages'.As Blommaert and Backus (2011) have pointed out, recent usage-based reformulations of the anthropological concepts of communicative competence and repertoire offer an alternative to definitions of competence based on 'languages' as independent systems.They argue that what is 'real' are speakers' repertoires, 'biographically assembled patchworks of functionally distributed communicative resources' (Blommaert & Backus 2011: 23), while 'languages' have to be seen as mere abstractions, 'a convenient way to refer to the cumulative inventory of resources shared by most people in a 'community'' (Blommaert & Backus 2011: 8).
However, the question remains whether 'languages', given that they are socioculturally significant constructs, should be seen as external ideologies exerting pressure on 'natural' repertoires or whether they can also be seen as forming part of speakers' competence.

A usage-based perspective on language learning and competence
Usage-based approaches to first (e.g.Tomasello 2003) and second language development (e.g.Eskildsen 2009;Eskildsen & Cadierno 2015;Ortega & Tyler 2016) share the basic assumption that 'languages' and individual language competence are not independent systems whose development follows a predetermined logic.Instead, they are seen as emerging from and continuously developing in actual instances of communication, thus conceptualising them as a 'process rather than an object' (Ortega 2014: 40).Language learning is understood as thoroughly situated, as it unfolds through everyday exposure to and engagement with language: 'The learner, whether an infant or an adult, first gains understanding of individual instances of language occurring in local, meaningful communication with others.Over a history of usage, the learner then gradually creates more abstract, interactive schematic representations, or a mental grammar' (Ortega & Tyler 2016: 2).In this view, language competence is not the static knowledge of a 'language' but is more aptly described as an individual and ever-evolving 'repertoire' in the sense of a dynamic 'inventory of linguistic resources' (Blommaert & Backus 2011: 7).Moreover, from a usage-based perspective, the purely 'linguistic' resources (such as lexical items or grammatical forms) of an individual's repertoire cannot be separated from other communicative means and dispositions (Blommaert & Backus 2011: 7;Eskildsen & Caderno 2015: 6).This perspective ties in well with Hymes' notion of 'communicative competence' as the holistic knowledge of linguistic forms together with their appropriate use in a context.
Because usage-based approaches view language development in terms of experience with language use, rather than in terms of bounded grammatical systems, 'languages' cannot be thought of as 'existing' in the minds of speakers as such.Rather, speakers' repertoires consist of functionally distributed resources (see Blommaert & Backus 2011: 23) that, especially given the increasing diversity of communities, frequently include resources from across 'languages' (as well as regional and social varieties, registers, etc.).Differences in competence can thus not be attributed to abstract speaker statuses (e.g.'native' vs. 'nonnative', 'monolingual' vs. 'multilingual'), but have to be seen as a consequence of individuals' divergent experiences with language (Hall et al. 2006: 230).From this perspective, ideologies of separate 'languages' are indeed in conflict with an understanding of individual repertoires as dynamic and diverse.However, this does not mean that repertoires are not structured or organised in any way at all.They are 'subject to a variety of stabilizing influences that are tied to the constancy of individuals' everyday lived experiences, and more generally, to more encompassing societal norms that value stability' (Hall et al. 2006: 229).In other words, if competence is seen as developing in response to the requirements of contexts of language use, both its dynamic and its more stable features are connected to dynamic or stable patterns of language use (see Bakhtin 1981: 272;Dufva et al. 2011;Sade 2011) in a speaker's environment. 2Lanza (1997Lanza ( /2004) has studied 'language mixing' in infant bilingualism from a sociolinguistic perspective.She argues that, instead of asking whether bilingual children's competence contains one or two language 'systems', 'language mixing' should be studied in relation to the kind of language use children are exposed to as well as the linguistic norms prevalent in the context of socialisation (Lanza 1997(Lanza /2004: 69): 69).Put differently, her study suggests that language use across 'languages' (such as 'language mixing') is neither arbitrary nor governed by linguistic constraints alone, but is the result of socialisation into specific patterns of language use.With regard to adult multilingualism, Canagarajah (2011: 9) notes that the ability to combine resources from different 'languages' is often seen as 'an intuitive capacity for which multilinguals are naturally endowed'.His study of a writing class, however, shows that the students' ability to draw on their multilingual repertoires in their writings was not a given, but developed through practice as well as reflection and discussions during the class (Canagarajah 2011).When interpreted from a usage-based perspective on competence, the insights provided by these examples indicate that ways of using language across 'language' boundaries are not 'natural' but are learned in the same way other kinds of discourse patterns are learned.What this means is that a sociocultural environment that encourages monolingual practices and keeping 'languages' separate will produce speakers with a different kind of competence than an environment with a significant 'plurilingual tradition' (Canagarajah 2009), because the values attached to and patterns of using 'languages' in these contexts will most likely be very different.
In sum, usage-based approaches to language learning defy ideologies of separate 'languages' by viewing language competence (as well as language itself) as an emergent and dynamic system.In this view, it is patterns of language use, and not 'languages', that structure individual competence.However, because patterns of language use are not independent from their social evaluation, ideologies about language have consequences for what resources become part of repertoires, how these resources are organised and what norms for their use speakers internalise.Such ideologies do not have to include the sociocultural notion of 'a language' (which is far from historically and culturally universal), but they certainly can and do in many contexts.From a usage-based perspective, investigating language learning should thus be a holistic endeavour involving not only the study of quantifiable factors and immediate interactional contexts but also of broader discourses and practices of communities as well as learners' subjective understandings of their language learning and use (see Ortega 2014: 47).In the following section I further explore the role of the social and ideological force of 'languages' in environments of language learning, focusing in particular on late additional language learning.

Developing competence in a languagised world
Individuals' linguistic repertoires (e.g.Blommaert & Backus 2011;Busch 2012;Rymes 2010) develop in response to the requirements of specific contexts and biographical trajectories.
They are, however, not direct reflections of their environments.We do not pay attention, let alone memorise, everything we see or hear.Conversely, single instances of language use do not reflect the entire potential of linguistic repertoires: we do not say everything we could theoretically say.What our attention is focused on in discourse, what we deem important to remember, and what we feel is appropriate to say is mediated by previous experience as well as sociocultural values (see van Lier 2000: 258), again pointing to the evaluative dimension of language.
The notion of 'competence' itself involves such an evaluative dimension.In its everyday meaning, competence does not so much refer to a speaker's entire repertoire of abilities, but points to how abilities are evaluated against the requirements of a specific context (as in 'being competent' or 'incompetent').With regard to linguistic competence, speakers usually cannot make use of all the linguistic resources available to them, rather, they have to orient to 'what counts as competence in real environments' (Blommaert et al. 2005: 200; emphasis in original).From this perspective, learning trajectories are also shaped by the social relevance and acceptability of ways of speaking, and repertoires appear as sets of 'resources our subject had to accumulate and learn in order to make sense to others' (Blommaert & Backus 2011: 23; emphasis in original).Put shortly, the evaluative dimension of language does not exclude contexts of language learning.Although social acceptability can certainly be based on local norms for language use that do not orient toward 'language' boundaries, ideologies of separate (national, standard) 'languages' are particularly pervasive and powerful points of reference, at least in some contexts.
The learning of an additional language later in life, especially when involving elements of formal instruction, is a case in point.Learning something new necessarily requires going beyond what already comes as 'natural', otherwise, no learning would take place.Contexts of foreign as well as second language instruction are designed to secure exposure to the target language and opportunities to practise communicative skills, but also to make adult learning more efficient by developing raising awareness about knowledge gaps and developing learners' metalinguistic consciousness through focused instruction (see e.g.Nation 2007;Ellis 2011).While a focus on 'monolingual' use of the target language in class is certainly often a practical choice aimed at maximising the effects of instruction, it simultaneously participates in constructing a (standard) 'language' as the legitimate object of learning.
Language learners in contexts of formal instruction are frequently presented with representations of independent language systems (see e.g.Dufva et al. 2011) and of 'nativelike' speech production as the target of their learning (see e.g.Doerr & Kumagai 2009).In order to expand their repertoire (by 'adding' another 'language' to it), learners are typically encouraged to engage in a range of practices aimed at keeping 'languages' separate, e.g.translating from one 'language' to the other, or working toward 'target-like' expressions without 'transfer' from other 'languages'.As Swain (2006) points out, evaluating what is 'correct' or 'appropriate' in the target language is also something learners themselves engage in, as 'languaging', understood as 'the process of making meaning and shaping knowledge and experience through language' (Swain 2006: 98), is one important way in which language is learnt, especially on advanced levels.This orientation to learning 'a language' does not necessarily disappear when learners leave the formal learning environment, not least because it can play an important role in creating opportunities for practice.For instance, in contexts where a lingua franca other than the target language is readily available, as is the case with English for many learners of Finnish in Finland, insisting on 'monolingual' communication in the target language, i.e. deliberately not making use of all linguistic resources available, can be an important or even necessary strategy for learners (see Theodórsdóttir 2011: 205 for a similar observation in the context of Icelandic as a second language).Second language environments do not guarantee access to linguistic resources and often learners have to overcome many obstacles in order to be granted access to interactional learning opportunities (see Pavlenko 2000).Thus, while translanguaging strategies, i.e. meaningful discursive practices across 'languages' (see García 2009: 45) are certainly useful in some contexts of language learning (see e.g.Cenoz & Gorter 2015), in others, the ability to display (sufficient) 'monolingual' competence in the target language can serve as a kind of 'symbolic competence' (Kramsch & Whiteside 2008), securing further opportunities for language learning and use.
From the usage-based perspective introduced in the previous section, such practices and strategies cannot remain without consequences for individual repertoires.If competence develops through instances of language use and this language use is reflexive in that its sociocultural framing is integral to it, we may ask whether it is indeed possible to acquire 'languages': not as ontologically real, systematic entities, but as sets of linguistic resources involving a metalinguistic dimension that associates them with 'languages'.In this view, ideologies of 'languages' appear not so much as detached discourses about language, interfering with multilingual speakers' 'natural' ways of using language, but as a crucial part of competence itself -provided, of course, that 'languages' are socioculturally significant concepts in the context in question.In the next section, I further explore this idea by looking at data from an interview study with highly proficient speakers of Finnish as a second language.I show how an orientation toward 'languages' is not clearly opposed to but rather mediates multilingual language use at the intersection of language use, ideology and repertoire.

An interview study with highly proficient speakers of Finnish as a second language
The examples discussed below are taken from an interview study with 12 highly proficient speakers of Finnish as a second language.Participants were highly educated adults (25-39 years of age) from six different countries (France, Germany, Hungary, the Czech Republic, Poland and Russia) and had moved to Finland 5 to 13 years ago.Their linguistic background was rather similar: all participants had grown up in monolingual families in their country of origin, speaking the national language as their first language (French in France, German in Germany, etc.).In addition to Finnish, they had also learnt other languages to some degree of competence later in life, English being the most common among them.For all participants, the process of learning Finnish involved formal instruction (language courses taken either before or after migration) as well as informal language learning in their social networks.
Their proficiency in Finnish was very high and they used the language in many different contexts of everyday life.The interviews were semi-structured and centred around issues of multilingualism, linguistic practices and identities.The language of the interviews was Finnish, with the exception of the interviews with the three German participants, which were conducted in their and my own L1 German. 3  The following discussion is neither exhaustive of the data nor representative of the whole context of Finnish as a second language.Rather, I highlight examples from the interview data to further illustrate the theoretical perspectives elaborated above.Naturally, interviews are restricted in scope as they do not allow direct inferences about linguistic practices (Codó 2008: 159).However, given the emphasis on the inseparability of ideology and language use in this chapter, interviews are valuable because they provide some insights into the beliefs and motivations that inform the practices that people engage in (Heller 2011: 45).From a usage-based perspective, subjective expressions of beliefs, values and ideologies are important because they help us understand how learners themselves perceive and interpret the sociocultural environment of their language learning.Taking as a starting point my participants' multilingual repertoires, my discussion will touch on their accounts of instances of language 'mixing', on strategies regarding the development of their repertoires, and finally, on statements that suggest that an orientation toward 'languages' can also be seen as a readily 'languagised' embodied disposition.

Ideological restrictions on language 'mixing'
Other than in contexts where the same multilingual resources are shared by a group of speakers (e.g.bilingual families or communities), the participants in my study can be described as multilingual individuals whose use of Finnish and their L1 is characterised by participation in different geographical, social and linguistic environments.Practices of 'mixing' Finnish with resources from the L1 (or other languages) are thus already restricted by these circumstances: few of their Finnish-speaking friends have knowledge of participants' L1, while family and friends in the 'home country' are not familiar with the Finnish language.There are, however, some 'microcontexts', such as relationships with individual friends with whom participants share the same first language and who speak Finnish as well, in which such language use is possible.Participants reported that they mostly used their L1 in these contexts, but that Finnish resources were also used occasionally.
One participant (Bianka, L1 Hungarian) told me that she and her colleagues often used 'mixed language' (sekakieli): 4

BK: Like that for example. And then we translate into Hungarian, for example, although in Hungarian you wouldn't say […], okay tehdään niin [let's do that]
like Finnish people often say when they agree on something.And then we translated that into Hungarian, but you usually wouldn't say that at all [...]  Bianka mentions that she uses Finnish words when talking about work to her Hungarian colleagues.From a usage-based perspective on repertoires, this is hardly surprising: Finnish is the dominant language of their workplace and lexical resources associated with work probably reflect the local language practices of this specific context more than they are linked with Finnish as a 'language'.Moreover, some ways of speaking atypical of 'normal', i.e. monolingual, Hungarians (a direct translation of the Finnish phrase tehdään niin) have become established in the multilingual group.This, too, can be seen as an indicator of the emergence of language norms beyond a separation of 'languages' in this microcontext.
However, the following excerpt from the same interview shows that multilingual language practices among Bianka's colleagues are also subject to negotiations about acceptability: Example 2 BK: Mulla on kaks unkarilaista työkaverii ja välillä ne puhuu niinku kokonaan suomeksi keskenään.Ja se mua aina ottaa päähän jostain syystä, et haloo, te ootte unkarilaisia, te ootte kahden kesken tai me ollaan täällä kolmen kesken, että niinku miks… ja se mua ärsyttää, sitä me ei tehdä, mut et kokonaisia lauseita esimerkiksi, mut vain just semmonen, mikä puuttuu tai vaikea löytää just joku sana ja sit sen sanoo suomeksi siinä lauseessa.KR: Eli kokonainen lause ois tavallaan jo liikaa?BK: Joo, tai ehkä jos niinku lainataan jotaku-tietysti, että jos keskusteluista lainataan, vaikka työpaikkakeskustelu, silloin toki, mutta ei puhuta keskenämme niinku omista sillai jutuista… BK: I have two Hungarian colleagues and sometimes they speak entirely in Finnish to one another.And that always annoys me for some reason, like, hello, you're Hungarians and there's just the two of you or it's the three of us, so why... and that annoys me, so we don't do, entire sentences for example, only [if] it's something that we don't have [in Hungarian] or it is difficult to find a word and then you say it in Finnish in that sentence.KR: So an entire sentence would be too much?BK: Yes, or maybe if you quote someone... of course, if you quote something from a conversation, for example a conversation at the workplace, then of course, but not if we talk to each other about our own stuff...In this excerpt, Bianka names a few conditions under which she considers the use of Finnish resources among Hungarians acceptable: single expressions are acceptable if the term in question does not exist in Hungarian or if a speaker has trouble recalling the Hungarian word, and full sentences are acceptable if they represent quotes from a Finnish-speaking conversation, but not when talking about their 'own stuff'.Echoing Finnish voices from the workplace in a conversation in Hungarian is, again, a good example of how language use across 'languages' is enabled by experiences with language in a multilingual environment.However, at least from Bianka's perspective, the acceptability of such language use crucially depends on pragmatically distinguishing 'quoting someone in Finnish' from 'speaking with one's own voice in Hungarian', thus linking it back to ideologies of monolingualism, authenticity, and nativeness.With regard to the use of single Finnish words in Hungarian, too, their use can at first thought be explained by convenience (a term does not exist in Hungarian or the speaker does not remember it).At the same time, however, it seems that judgements about which expressions are specific to the Finnish context or difficult to render in Hungarian, are what decides whether practices of language 'mixing' are deemed acceptable or 'annoying'.
In fact, the interviews with the German informants seem to support this reading.All three German participants (Alexander, Sandra and Julia) use Finnish expressions in the interview, as can be seen in the following excerpts:

Well, I am quite, quite uh how should one say uh laiska [lazy] (.) lazy with my studies…
A review of these and other instances from the interviews shows that Finnish resources used by the German informants are most often somehow either specific to the Finnish context (see example 4), the informants had not encountered the concepts earlier for biographical reasons (e.g.work or study related terminology), or the expressions are particularly idiomatic in Finnish (see example 3).This is in line with a usage-based view on language learning: although my informants have comprehensive, 'native' competence in German, they resort to Finnish resources when talking about phenomena that are particularly salient in the Finnish context.However, examples 4 and 5 show that there is also some hesitation around the use of Finnish expressions in a monolingually framed interview 'in German'.In example 4, Sandra uses the Finnish term for a qualification she obtained but the term is framed by hesitation markers ('uhm' and 'uh').Example 5 is different in that Julia uses a very generic Finnish word (laiska, 'lazy') in the context of the German interview.Here, the framing of her use of this word as a struggle to remember appears even stronger: she uses hesitation markers ('uh'), makes a metapragmatic comment ('how should one say'), and takes a clear break before 'translating' the word (that I had of course understood already) into German.This suggests that ideologies of keeping 'languages' separate can be present even when language 'mixing' occurs: the marking of hesitation indicates something out of the ordinary.Moreover, we can assume that the less clear the rationale for using a 'Finnish' word in a 'German' conversation is (as in the case of example 5), the more its use has to be justified through metapragmatic framing.Practices of language 'mixing' can thus be accompanied by an ideologically mediated metapragmatic competence that allows participants to make use of their multilingual repertoires while still constructing themselves as authentic language users.

Shaping repertoires
As shown in the previous section, despite practical and ideological restrictions, participants in the study reported combining resources from Finnish, their L1, and other languages regularly in their everyday language use.This was reported to be even more common with practices engaged in alone, such as thinking, browsing the internet, or writing shopping lists, where many participants claimed to use 'very mixed' (Sophie, L1 French) resources.However, a somewhat opposite concern was the perception of 'declining' competence in their first language.Almost all participants expressed fear of their L1 getting 'rusty' and some intended to actively work against this development, e.g. by reading more books in the L1.From a usage-based perspective, full proficiency in the L1 is indeed not a given but is as much a matter of language use as is L2 learning (see e.g.Opitz 2016).In the following interview excerpt, one participant, (Bianka, L1 Hungarian) describes how she tries to maintain her L1 competence by making conscious decisions about language choice: Example 6 BK: Mä joskus tiedostin sitä, että vaikka reseptejä tai jotain, tai mä kirjoitan niinku kauppalistan… Mä tänään-mun piti niinku oikein kiinnittää huomiota siihen, et mä kirjoitan unkariksi, kun mä olin yksin lähdössä sinne, ei oo väliä.Ja sit mä ajattelin, et hei, että tää menee överiksi, just silleen, että mä halusin niinku unkariksi sen kirjoittaa.Mut kyl se suomi aika lailla valtaa tätä elämää ((nauraa)).KR: Miks sä halusit sitten kirjoittaa unkariksi sen listan?BK: No, mä en halua, et se ruostuu, mä huomaan, et se ruostuu ja sit esimerkiks mä joskus pakotan itteni.Ihan se on just se kauppa jostain syystä, et mä meen sinne ja vaikka mä mietin itsekseni, et mitä pitää ostaa, että mä en luettelis sitä niinku suomeksi vaan ihan unkariksi.BK: I noticed some time that for example with recipes or something, or when I write a shopping list... Today I really had to pay attention to writing it in Hungarian, because I was leaving [for the shop] alone, so it didn't matter.And then I thought, hey, this is going over the top, I want to write it in Hungarian.But my life is quite dominated by Finnish ((laughs)).KR: Why did you want to write it in Hungarian then?BK: Well, I don't want it to get rusty, I realise it is getting rusty and then, for example,

Finnish but in Hungarian [in my mind].
This example shows that Bianka pays attention to the resources she uses not only when talking to others, but that she monitors her language use also when doing something only for herself (like writing a shopping list).It seems that, for Bianka, using Finnish for such activities would be quite 'natural', given that her 'life is quite dominated by Finnish'.Her practice of 'forcing herself' to use Hungarian is then aimed at actively shaping her repertoire according to her beliefs about what her ('native') competence should be like (not 'rusty').
The interview with Sandra, one of the German speaking participants, features another example of conscious intervention and maintenance of the L1 repertoire.Sandra has mentioned earlier that she is part of a book club and uses the German word Lesezirkel [lit.
'reading circle'] to describe it.There is hesitation and some laughter in her voice when producing the word and, given that Lesezirkel is somewhat unidiomatic in the meaning of an informal book club as well as appearing as a direct translation from the Finnish lukupiiri [lit.SA: That's important to me also because of the kids, because I want the kids to learn that when you speak German, you speak German.There's only a few words that I've noticed when I was… well, when I was in these German groups or when I talk to German friends, that all Germans use in Finnish, neuvola [child health clinic] and muskari [musical playschool] and välipala [snack]… Sandra admits that there are indeed some words that 'all Germans use in Finnish'.However, these words are quite specific to the context of being a parent Finland (e.g.neuvola, 'child health clinic', is associated not only with a place but with the whole system of public child health care in Finland) and lack a precise equivalent in German.Apart from the use of these words, she emphasises the importance of being able to use language in a monolingual mode ('when you speak German, you speak German').In her case, the stakes are also particularly high: according to the one-parent-one-language approach (e.g.Barron-Hauwaert 2004) taken by her family, it is her responsibility to pass on the German language to her children, which requires a lot of work from any parent in the position of a minority language speaker (see Döpke 1992).Note that she employs this monolingual mode even in the interview with me, when her children are not present, which suggests that her strategy for ensuring that her children develop a fully functional repertoire in German involves the conscious and continuous shaping of her own linguistic repertoire.Her use of the word Lesezirkel, produced with some difficulty, is not a translation for me (since she is aware that I would have understood the Finnish word as well) but, rather, it appears to be a disciplinary exercise directed toward herself.In a similar way as Bianka with her shopping list, Sandra is forcing herself to 'stick with' German in our conversation in an attempt to actively maintain her own repertoire.Obviously, it is debatable whether the use of the word Lesezirkel in this context represents 'proper' German at all, or whether it is, in fact, a good example of a translingual practice.However, it shows that, whatever the result, Sandra's active monitoring based on an ideology of separate 'languages' (by definition included in the one-parent-one-language approach) plays a least some part in creating this particular instance of language use.With regard to language, it is of course only a small part of linguistic practice that can become the object of conscious reflection or manipulation, and what speakers say they do based on their beliefs about language is often far from what they actually do in practice.
However, it would also be misleading to rule out any connection between conscious beliefs and unconscious behaviour entirely.From a usage-based perspective on repertoires, ways of engaging with language are never without consequences.In this view, deliberate choices and strategies can turn into (local or individual) routines, leading to particular resources becoming entrenched and thus participating in shaping speakers' repertoires.

'Languagised' repertoires
So far I have discussed examples of how multilingual practices are frequently mediated, interpreted or framed by ideologies of 'languages'.In this section, I turn from the ways in which beliefs influence practices and how practices develop repertoires to the less dynamic aspects of competence.Namely, in the case of my participants, it can be argued that their repertoires are already profoundly shaped by experiences with language where 'languages' are kept separate.Researchers have shown how modern developments that promote ideologies of nationalism and separate 'languages' have had profound cultural and material consequences for societies, not least in the European context (see e.g.Bauman & Briggs 2003).Growing up in such a context means frequent exposure to many 'banal' ways (cf. Billig 1995) of constructing (national) 'languages': mentioning 'languages' by their name, studying 'languages' as separate subjects at school (see Dufva et al. 2011), clicking on the icon of a flag when choosing the 'language' of a website, etc.It also means exposure to an enormous amount of text and speech that orients toward monolingual norms of language use.From a usage-based perspective on language learning, such an environment cannot remain without consequences for speakers' orientations toward language as well as the content and structure of their repertoires.
The following excerpt can be seen as an indication of this.I had just asked Marie (L1 French) about ways that she 'mixes' languages in everyday life.She explains that she uses both Finnish and French resources when making to-do-lists for herself on her computer.She shows me an example of a list, which indeed contains elements of both languages.However, talking about the process of making these lists, she explains: Example 8 MA: Mua vähän häiritsee, kun on niinku sekä ranskaa että suomee, niin esim jos otsikoissa.Mut onhan tossa ranskankielisiä otsikoita, mut mua vähän häiritsee.
MA: It annoys me a bit when there's like both French and Finnish, for example in the titles [of the sections].But of course there are also French titles here, but it annoys me a bit.But I'm a bit […] of a pedant, if there's a French title and then a Finnish one and then they don't they don't look the same, so then it just...

KR: So do you sometimes correct your own lists?
MA: Yeah, well at least the titles I change, they can't be in French.
Although Marie makes to-do-lists only for her private use, she reports that she frequently makes the effort of editing a list she has already written in order to keep 'languages' apart at least visually.She then explains her unease with mixing elements that 'don't look the same' by calling herself a 'pedant' in these matters (and other participants seemed indeed more comfortable with this kind of writing).However, it can be argued that it is years of exposure to and practice in monolingual writing (which Marie received e.g. through studying languages in school and at university) in an environment deeply influenced by ideologies of separate 'languages' that enable her to instantly detect language use diverging from monolingual norms and that make her feel uneasy about it.
Another participant (Bianka, L1 Hungarian) tells the following story when asked about her use of different languages in everyday life: BK: Oh yeah, sometimes I tried to write something like that, that you could write some text, some short story, and it could have words in whatever language you like.
[…]And that I think I have sometimes thought about, like writing in whatever [language] comes like the right word [...] in the other language, something like that.KR: Have you tried?BK: I have tried some time, but-I don't-maybe a page.KR: Mmh.BK: I succeeded but that was it ((laughs)).
KR: Was that difficult or easy that you can use the word that comes to mind?BK: Well, maybe that short a text was still easy, but I don't know whether you could write like a whole book like that.Although Bianka's dominant languages are Hungarian and Finnish, her repertoire also contains resources from a number of other languages.In the exchange preceding this excerpt, she told me how she has always thought that some words in a particular language seem to describe a phenomenon much better than any word from any other language, and how those words are fun to use.Inspired by this, she makes the conscious decision to try to produce a 'translingual' text, using whatever words come to mind first.Even though she reports succeeding at it, she never returns to the project and even expresses doubt about whether it would be possible to write a longer text in this way.This suggests that outside of contexts designed especially to assist the development of translanguaging practices (see Canagarajah 2011), individual attempts to engage in such practices, here motivated by ideas similar to the concept of 'translanguaging', may be quickly abandoned by speakers whose repertoires are already thoroughly 'languagised'.This is because multilingual repertoires do not automatically translate into 'fluid' multilingual practices.As the examples in this section suggest, it is not only an active orientation toward monolingual norms that mediates language use across 'language' boundaries, but such norms can also be deeply embodied as part of a speaker's 'natural' disposition, making 'translanguaging' a skill that needs to be learned and practised.

Conclusion
An 'artefactualized image of language' (Blommaert 2010: 4) as the basis of the study of language has come under heavy scrutiny in the past few decades.However, what appears now as a distorted view in the realm of professional linguistics, is also part of how people encounter language in many everyday contexts where speaking of 'languages', debating 'language' boundaries, and assigning linguistic forms to 'languages' are common practices.In this chapter, I have suggested that the opposition of 'artificial' ideologies of 'languages' and 'natural' multilingual language use is somewhat misleading.I have argued that if we take seriously the idea that language always has an evaluative dimension linking it back to sociocultural conceptions and ideologies, it seems unreasonable to exclude the cultural construct of 'languages' from this dimension in principle.I have also argued that if we further subscribe to an understanding of language learning as based in experiences with situated language use, thus certainly involving this evaluative dimension, it is equally inconsequential to exclude 'languages' from experiences of language learning and thereby from learners' trajectories and repertoires.
Obviously, whether and how exactly 'languages' form part of different contexts, and whether and how (much) they interact with other linguistic processes is a matter of empirical investigation.However, the possibility that 'languages' represent, in certain contexts, a central concept around which language use and language learning are organised should not be precluded.This implies that sociolinguists and applied linguists need to be aware of the 'contingent constructedness' of 'languages' (Makoni & Pennycook 2005: 142) while, at the same time, taking seriously the ways in which people use 'languages' for constructing meaning and identities.The arguments brought forward in this chapter offer one possible perspective for understanding how 'languages' are 'fictions' at the same time as they are 'real': by becoming and being part of speakers' 'languagised' repertoires.
ich war total deprimiert und dann auf einmal quasi wars hatt ich einen einen riesen onnenpotku.And I felt really down and then, all of a sudden it was like, I had an enormous onnenpotku [stroke of luck].Example 4 SA: Ich hatte vorher auch schon angefangen mich dafür zu qualifizieren, ich hab […] dieses ähm opettajan S2-perus ja aineopinnot äh hab ich hab ich gemacht… I had already started to get the necessary qualifications, […] I did this did this uhm opettajan S2-perus-ja aineopinnot uh [basic and intermediate studies in Finnish as a second language for teachers] ... Example 5 JL: Also ich bin relativ, relativ ööh wie soll man sagen ööh laiska (.) faul halt im [Studium]…

'
reading circle'], I begin to suspect that her hesitation indicated a struggle to access a German term for it.Further along in the interview, I ask her about this directly: Example 7 KR: Wie ist es mit so Wörtern, die… ich bin mir nicht ganz sicher, aber ich hatte fast das Gefühl, also du grade von deinem Lesezirkel erzählt hast, dass du zuerst an das Wort lukupiiri gedacht hast.SA: Ja ja, genau.KR: Genau, aber du hast es ja trotzdem dann sozusagen übersetzt.[…] Jetzt nicht extra für mich, aber du hast dich trotzdem dafür […] entschieden das zu übersetzen.SA: Ja.KR: Ist das… achtest du irgendwie dadrauf oder also du… SA: Ja, das ist mir wichtig.KR: Okay.SA: Das ist mir wichtig auch wegen der Kinder, weil ich möchte, dass die Kinder lernen, dass man Deutsch spricht, wenn man Deutsch spricht.Es gibt ein paar Wörter, also wo ich auch festgestellt habe, als ich in diesen… also wenn ich hier in diesen deutschen Gruppen bin oder auch wenn ich hier mit irgendwelchen deutschen Freunden spreche, die alle Deutsche auf Finnisch benutzen, neuvola und muskari und välipala… KR: How about words that… I'm not quite sure but I almost felt like when you just told me about your book club you first thought of the word lukupiiri.SA: Yeah, that's right.KR: Right, but then you still kind of translated it.[…] Like not especially for me but you still decided to […] translate it.SA: Yes.KR: So is that… do you somehow watch yourself there or do you… SA: Yes, that's important to me.KR: Okay.
in Hungary.That's maybe the two ways.KR: But you say that in your circle of friends now.BK: Yes [...] I'm sure it sounds really strange to a normal Hungarian.But we're used to saying that now.