Parental discourses of language ideology and linguistic identity in multilingual Finland

ABSTRACT Finland is officially a bilingual country but it is in practice multilingual. In the current study, we examined how mothers and fathers of mixed-language families linguistically identified themselves and others, and how ideological discourses and concepts historically and socially situated in Finland circulated through the parents’ talk. The parents of three families in which at least Finnish, Swedish and English were used on a daily basis were interviewed. A discourse nexus approach showed that the concept of ‘mother tongue(s)’ played a central role and that although all family members were in practice multilingual, there was a strong tendency across the couples to identify themselves and others as monolingual. Bilingualism was identified with Finnish-Swedish rather than other languages and a native discourse expressed bilingual identity as granted by birth rather than acquired later. The discourses could be traced back to official language registration procedures, the educational system in Finland, as well as to parents’ own lived experiences. The study illustrates the intricate relationships between language ideologies and how linguistic identities are created and performed among parents, and it pinpoints the need for further studies on how linguistic identities are passed on to and experienced by children along their life trajectories.


Introduction
Finland is a multilingual country with two equal national languages, Finnish and Swedish. The Sámi as an indigenous group and the Roma, as well as 'other' language groups and users of sign languages, are also acknowledged in the Constitution of Finland 1999. Multilingualism is further emphasised by migration in Finland even though the number of migrants is still small compared with the numbers in other Nordic countries (Honko & Latomaa, 2016). There is a conscious education policy of promoting multilingual skills and identities, as reflected in the recently implemented national core curriculum for basic education (FNBE, 2014), and there is a long tradition of formal language teaching: it is compulsory for schoolchildren to study the national languages (Finnish and Swedish), and in addition to these, at least one 'foreign' language (English), with the option of studying further languages (e.g. German, French, or Spanish).
As a result of its language legislation, Finland has a system of official registration of the linguistic affiliation of its citizens; each individual is assigned an official language referred to as his/her 'mother tongue'. At the moment it is only possible to report one mother tongue; the statistics thus say very little about individual multilingualism and societal linguistic diversity (see Moore, Pietikäinen & Blommaert 2010 for a critical account of numerical representations of speakers). Based on the official statistics, 88.3 % of the population have Finnish, 5.3 % have Swedish, and 6.4 % have other languages as their mother tongue. The largest 'other' languages are Russian, Estonian, Arabic and Somali (Official Statistics of Finland, 2017). Despite the fact that Finland is multilingual, there is a prominent national identity discourse of Finland as bilingual, which refers to Finnish and Swedish (Halonen, Ihalainen & Saarinen, 2015; also Hult & Pietikäinen, 2014).
In the current study, parents of three mixed-language families in Finland who regularly used (at least) Finnish and Swedish were interviewed. During these interviews, parents made frequent references to other languages as well. The focus is on how the parents discursively co-construct and negotiate their own and their children's language and cultural identities, and in doing this, how they circulate official language ideologies. Finnish and Swedish speakers have shared their territory and lived side-byside for many centuries (McRae, 2007). Therefore, the cultural context can be "imagined as a homogeneous space" involving very close cultural proximity, which is often argued to be "unproblematic or at least less problematic than cultural distance" (Piller, 2002, p. 6). In this sense, the study complements prior research about binational family constellations where parents speak languages of high prestige but where the cultures are only relatively close (e.g. Gonçalves, 2013;Piller 2002;Teiss & Perendi, 2017) or are quite distant (e.g. Okita, 2002;Zhu Hua & Li Wei, 2016). This close cultural proximity opens the way for potentially different patterns of family language policy negotiation and experiences of linguistic identity than have hitherto been explored.
The research questions which we examine are the following: 1) How do parents linguistically (and culturally) identify themselves, and negotiate and ascribe identities to each other and to their children? 2) How are the linguistic identities formulated and negotiated by the parents related to their own lived experiences, on the one hand, and to societally situated discourses and ideologies, on the other? The data analysed were semi-structured interviews with the parent couples. In order to identify the discourses and concepts that were foregrounded and negotiated by the participants we applied a nexus analytical approach (Scollon & Scollon, 2004).

Ideology and identity in multilingual families
In the case of multilingual families in Finland, the system of language registration imposes parental agency, since the parents choose the official mother tongue of their child and also the primary language of instruction in school. The registered mother tongue can, however, be changed at any time without explanation. The school system has parallel Finnish and Swedish tracks providing education in either of the two languages from early childhood to university level. In the current study, all three families had made an informed choice (cf. Kalaja, Barcelos, Aro & Ruohotie-Lyhty, 2016, p. 19) of Swedish-track early childhood education for their children. They also employed a multilingual language policy and multilingual practices at home. Hence, their family language policy involved explicit and overt planning in relation to language use (King, Fogle & Logan-Terry, 2008) as well as more implicit and covert planning and practices (e.g. Curdt-Christiansen, 2013;Fogle, 2012;King & Fogle, 2017, Palviainen & Boyd, 2013. Beliefs and attitudes about languages and language development play a significant role in mixed-language families. By way of example, on the basis of these beliefs, parents may employ certain language practices which they think will lead to a certain result (e.g. De Houwer, 1999;Piller, 2001;Schwartz & Verschik, 2013).
Ideology, which in this article is seen as "a set of beliefs" and as "the underlying thought system that is expressed through discourse -language in action" (Määttä & Pietikäinen, 2014, p. 8), therefore interacts with agency in the complex web of actions that mediate family language policy (King & Lanza, 2017). Beliefs indicate that individuals or a community accept something as true, and they are at the same time context-dependent, dynamic and possibly even conflicting (Kalaja et al., 2016, p. 10).
'Mother tongue', as mentioned above, is an example of a socially and historically constructed concept (Hult, 2015) that can play out in complex ways as ideological discourse. In one sense, the mother tongue can be very emotionally loaded and closely connected with personal lived experience and identity. In another sense, as Liebkind, Tandefelt and Moring (2007) point out, it can also be merely a bureaucratic term with little content: "people [in Finland] are quite used to the concept of 'mother tongue' because that is a concept that is the expression used in several official registers and documents" (p. 5). The Finnish system, in which everyone is allocated 'one mother tongue', was originally developed in order to serve as a numerical basis for deciding whether a municipality or local authority was bilingual and therefore obliged to provide services in both Finnish and Swedish (or Sámi). As we will argue, however, this system also has profound implications for how parents identify themselves and their children (see also Bergroth 2015).
Identity is constructed and negotiated in a socio-cultural context and refers to "people's understanding of their relationship to the world" (Norton 1997, p. 410). As identities are socially and discursively (co-)constructed (Bucholtz & Hall, 2005), they are also dynamic and shifting and can serve as a site of struggle and negotiation of difference and ambivalence (Block, 2007;Piller, 2002). Piller (2002) contends that identity is not "a matter of labels and categories but rather one of performance" (p. 12); therefore, from a linguistic perspective, she argues that language and social identity are mutually constitutive. We argue that it is important to place the notion of linguistic identity as the focus of analysis and to ask, on the one hand, who has the agentive right to claim a certain linguistic identity as their own in a multilingual family and, on the other hand, who has the agentive rightand on what groundsto deny another's bilingual or multilingual identity within the same family.
Our focus in this article is on the meaning-making processes parents use when constructing their linguistic identity and how these connect with language ideologies. We aim to capture linguistic identities as they are performed and negotiated by parents through their talk. We will also explore how the linguistic identities are constructed as a function of the parents' personal language ideologies and their personal trajectories of language learning, and how these align with larger cycles of public discourses.

Data collection and participants
The present study was a part of a larger ethnographically informed research project focusing on mixed-language families in Finland in which there were children between 3-5 years of age attending Swedish early childhood education (ECE). The families were recruited via the ECE centres. As a criterion for participation, the parents had to have different languages (Swedish or Finnish) registered as their official mother tongue. The data used was from interviews carried out with the parents in three of the families, who lived in two different cities. The Pitkänen and the Kuusisto-Lindström couples lived in Weston, an officially bilingual city, while the Kivistö couple lived in Middleton, an officially monolingual Finnish city (see Table 1). In Weston, 69 % of the population had Finnish, 23 % Swedish, and 8 % other languages as their registered mother tongue whereas in Middleton, 95 % of the population were registered as Finnish speakers and 5 % as speakers of other languages (Official Statistics of Finland, 2017.) The interviews followed a structured protocol (Mann, 2016), each one including the same set of questions, which dealt with the participants' language background, languages studied at school and used at work, language choices at home, definitions of bilingualism, and multilingual contacts in the family's daily life. The interviews were between 45 and 60 minutes long and were carried out by the second author, who is multilingual. The participants could choose which language(s) to use in the interviews: the Pitkänen and Kivistö couples used mainly Finnish and the Kuusisto-Lindström couple used Swedish as well as Finnish. The interview excerpts below are provided in the original language as well as translated into English.

Data analysis
The interviews were analysed as a part of a larger set of data with a nexus analytic approach. We see interview material as rich data to tap parental discourses about language ideologies and linguistic identities and a nexus approach makes it possible to disentangle and analyse complex layers of discourse (Palviainen & Boyd, 2013). At the core of nexus analysis is a social action (Scollon & Scollon, 2004), which in this study is the parental co-construction of linguistic identities in talk. A social action is situated at the intersection of the historical body of the individuals, the interaction order which they mutually produce among themselves, and the discourses in place that enable that action (Scollon & Scollon, 2004, pp. 153-154). In our analysis, we therefore carried out mapping and circumferencing discourse analysis of the lived experiences and language learning trajectories of, and beliefs held by, the individual parents (the historical body); norms of interaction and expectations about social rolesorparticipants and their significance (the interaction order); and the situated context in which action took place, including ideologies, socially and historically situated concepts and norms of interpretation (the discourses in place) (Hult, 2015). With the two research questions in mind, we mapped relevant people, places, discourses, objects, and concepts (Scollon & Scollon, 2004, pp. 159-160) in the parental talk and analysed how these were negotiated between the parents. We also identified connections with larger cycles of societal and historical discourses.
Notably, neither of the concepts 'ideology' or 'identity' were mentioned in the interviews by either researchers or parents, yet they are an integral part of the individuals' historical body and of the discourses in place. The concepts also play out in interaction orders and they therefore functioned as tools facilitating our analysis and understanding of the social action at hand. In the following sections, we present the results of the nexus analyses of the three families in turn: the Pitkänens, the Lindström-Kuusistos, and the Kivistös.

Results
The Pitkänen couple Jaakko and Tuija Pitkänen both grew up in Finnish-speaking homes in the Weston area, but whereas Jaakko went to a Finnish-medium school, Tuija attended a Swedishmedium one. Jaakko had studied Swedish, English and German at school, and he was now enrolled in a Finnish-English university programme. Both Jaakko and Tuija used Finnish, Swedish and English in their work. With these backgrounds they were what Kramsch (2009, p. 17) refers to as 'multilingual subjects'. She thus initially stated that she felt that she was neither a Swedish-speaker nor a Finnish-speaker, but then announced that Finnish was the language of her heart because she had inherited it from her own mother. Later in the interview, she ascribed a similar identity to her 4-year-old daughter: bilingual, but more Finnish than Swedish (Excerpt In this exchange, Tuija disqualified Jaakko as a bilingual: he was, rather, a Finnish speaker who knew Swedish and English. Jaakko, who first appropriated bilingualism as something relative (cf [4b]) eventually abandoned this interpretation and aligned with Tuija's conception of bilingualism as a question of either-or. In this way they ended up with a co-constructed identity for Jaakko as a Finnish speaker (even yksikielinen, "monolingual") who had some time later learned and added Swedish and English to his repertoire. The cognitive metaphor that Jaakko used for this process, "glued onto the brain", stands in stark contrast to the emotional heart metaphor Tuija used to describe herself and her daughter. This passage shows that not only is linguistic knowledge interactively assessed (Piller, 2002, p. 103), but so too is who has the right to be called bilingual and on what grounds (cf. O'Rourke, Pujolar & Ramallo, 2015). In this case, it seems as if the spouses agreed that Tuija's childhood, and being born into a family with Finnish-and Swedish-speaking parents and attending a Swedish school, put her in a special position and made it legitimate for her to be called bilingual, whereas Jaakko was not in such a position.

The Kuusisto-Lindström couple
Kalle Kuusisto also grew up in a Finnish-speaking family in Weston and went to a Finnish school. He described himself as ihan westonlainen and ihan täysin suomenkielinen ("a born and bred Westoner"; "definitely a 100% Finnish speaker").
Jessica Lindström, his wife, also claimed to be riktigt Weston-bo ("a true Westoner") but, in contrast to him, to be fullständigt svenskspråkig ("a 100 % Swedish speaker") It is not clear whether he agreed with Jessica that bilingualism equals being bicultural, but he did express the idea that he had been socialised into many Swedish cultural traditions over time. Discursively, he made a self-correction: instead of saying that he had been ruotsalaistunut ("Swedishized", become a Swede)which refers to Sweden as a nation state and to Swedeshe said that he had been ruotsinkielistynyt ("Swedishspeakerized", become a Swedish-speaker), referring to Swedish speakers in Finland.

Kalle had thus incorporated the norms of interaction of Swedish speakers (in Finland)
and had become successfully assimilated into their linguistic and cultural community.
This negotiation of cultural identities includes elements of pointing out differences and contains discourses of transformation and assimilation (Block, 2007;Kramsch, 2009 In her view, in order to count as bilingual, in addition to the mother tongue (in the singular) you need to know another language sufficiently well. According to her, Tindra met these criteria. Heikki made a mild objection, that it is possible to have more than one mother tongue (Excerpt [13] Whereas Leena's reference to a mother tongue in the singular is a common discourse in the Finnish context due to the bureaucratic system of registering one official mother tongue, Heikki recycled another discourse, where the mother tongue has emotional connotations and is not necessarily restricted to just one. In defining what it takes to count as bilingual, Heikki presented a complex aggregate of different discourses (Scollon & Scollon, p.  German and that there's English and so on, so that's really nice."). Still, she was defined as bilingual (rather than multilingual). The reasoning here reminds us of what we found in the Pitkänen family: one can add languages to one's repertoire as one gets older (cf. Excerpt [5] above), but one is bilingual first and foremost by virtue of being born into it or, alternatively, of having attended school in another language than the one spoken at home.

Discussion
In this paper, we have analysed linguistic identities as (co-)constructed in parental discourses in three multilingual families. In response to the first research questionhow parents linguistically (and culturally) identified themselves, and negotiated and ascribed identities to each other and to their childrenwe found that all the families and their members were constructed as multilinguals. Although the term 'multilingual' itself was not explicitly employed by any of the parents, it was evident that different languages (Finnish, Swedish, English, German) played significant roles in their own as well as in their children's everyday lives. However, to identify oneself or one's partner as bilingual or multilingual was not straightforward and required negotiation between the parents.
Rather contradictorily, whilst being multilinguals with rich repertoires of inherited as well as acquired language skills, most parents were firm about identifying themselves as a speaker of only one language: olen ihan/aika täysin suominkielinen ("I am definitely/pretty much 100 % a Finnish speaker") or fullständigt svenskspråkig (100 % Swedish speaker). This means that identity was primarily defined in terms of belonging to a single (imagined homogenous) group of speakers of a certain language (cf. the concept of 'speaker-hood' in Moore et al., 2010). Moreover, whereas traditions and habits were put forward by one of the couples (Lindström-Kuusisto) to distinguish Finnish speakers from Swedish speakers, ethnicity was not an issue. This stands in contrast to other studies of binational couples, where ethnicity plays a potentially significant role in negotiations of linguistic identity (e.g., Okita, 2002;Piller, 2002).
As for the second research questionhow the constructed linguistic identities were related to the parents' own lived experiences and to societally situated discourses and ideologieswe found certain concepts that frequently circulated, such as 'mother tongue(s)', 'bilingual(ism)', and 'foreign' and 'learned language(s)'. These are socially and historically situated concepts known from Finnish national language policy and education policy discourses, as well as colloquial discourse. In the interviews, 'mother tongue' was referred to by some of the parents in terms of the official mother tongue, and hence conceptualised in the singular. Others challenged this concept, attached emotional content to 'mother tongue' and distinguished it from 'foreign language', and provided alternative definitions, including the possibility of naming more than one mother tongue as part of one's individual identity.
The parents seemed to conceptualise 'bilingual' as an identity that comes at birth: a child born into a mixed language family had the right to claim and be granted bilingual identity, regardless of their language proficiency or bilingual practices.
Moreover, parents were surprisingly firm that a bilingual or multilingual identity could not be achieved or claimed over time, no matter how much they had studied languages formally, or what multilingual practices there might have been in their family or working life as an adult. Identity in this sense, rather than being dynamic and subject to change over time, becomes a matter of essence: it is something you have (i.e. are born into) rather than something you do (i.e. using or acquiring new languages). There were, however, two exceptions to this ideology of nativeness: Tuija and Heikki were both born into families where Finnish was spoken but they attended Swedish-medium schools. In Bourdieu and Passeron's (1977, p. 42-43) terms, while the primary habitus acquired at home in their early years was monolingual, the secondary habitus acquired at school served to give them a bilingual identity as adults. The importance of going to a Swedish school in Finland for the development of a minority language identity on the national as well as the individual level has also been acknowledged elsewhere (e.g. Lojander-Visapää, 2008).
Despite the fact that all the subjects were in practice multilingual, the main discourses circulating in the parental talk can thus be summarised as ascribing to themselves (and others) a monolingual identity unless they had been born into a mixedlanguage family (or attended a minority language school), in which cases they could be ascribed a bilingual identity. Moreover, the societal discourse equating 'bilingualism' with Finnish-Swedish bilingualism in Finland (Halonen et al., 2015;Hult & Pietikäinen, 2014) was strongly present in all the arguments. This discourse was most likely grounded in the fact that each citizen is identified with one official mother tongue, in combination with a general conception that there are two distinct language groups in Finland. This concept can be traced all the way back to the Constitution of Finland (1919/1999), where it is stated that "the public authorities shall provide for the cultural and societal needs of the Finnish-speaking and Swedish-speaking populations of the country on an equal basis" ( §17) (cf. Ihalainen & Saarinen, 2015). On the whole, although it was obvious that the participants knew several languages, multilingualism was given only a background role in the discussions.

Conclusion
The parental talk around the couples' language backgrounds and what it takes to count as bi-or multilingual revealed hybrid identities and multiple discourses (cf. Gonçalves, 2013;Piller, 2002). Across the couples, we could identify a number of co-occurring discourses based on such aspects as proficiency, order of acquisition, affection, nativeness, inheritance and cultural appropriation. Although revealing a tendency to label and categorise identities and pin them onto themselves as well as onto others, essentially seeing identity as something set from birthhence pursuing a nativeness ideologythe couples also actively negotiated identities and in some cases modified their conceptions while interacting. We further found that all the families applied multilingual language practices within and outside their bilingual homes. However, we also found an equally clear tendency to describe languages learned later as additional, separate pieces glued onto the original identity, rather than being an integral part of it.
Ideologies held by individuals or a community mediateand are mediated throughlanguage use (cf. De Houwer, 1999;Piller, 2002). This study highlights the need to further explore the intricate connections between identities, ideologies and agency (Kalaja et al, 2016, p. 18;King & Lanza, 2017). The study illustrates how systems of social organisation and societal language planning discourses of supporting or safeguarding linguistic rights at a bilingual/multilingual national level can at the same time force monolingual identities onto multilingual individuals. Along with individual lived experiences, societal ideologies and discourses affect how parents define the linguistic identity of their family and its members. We need further research to find out what implications this has for how linguistic identity is passed on across generations. A relevant question to ask in the future is how young childrenlike Eva, Ella and Tindra, who were all given a bilingual identity by their parentswill identify themselves as adults and how that, in turn, will affect their own language practices, ideologies and life trajectories.