The early childhood education and care partnership for bilingualism in minority language schooling: collaboration between bilingual families and pedagogical practitioners

ABSTRACT Educational partnerships occur at the intersection of early childhood education and care (ECEC), families, and the surrounding community, and have been shown to play a significant role in student success rates in education. There is, however, a gap in research on the role and potential of ‘partnership’ in the case of bilingual families with children enrolled in monolingual minority language ECEC. This study aimed to fill this void by examining parental and practitioner discourses on partnership and on obligations, desires, abilities, and competencies involved in acting on a bilingual childhood in the context of monolingual minority ECEC. Parents in nine Finnish–Swedish bilingual families and six pedagogical practitioners at three Swedish-medium minority language ECEC units in Finland were interviewed. We applied ethnographic discourse analysis – nexus analysis – on the interview data and found that the family languages (Finnish and Swedish) did not seem to be given equal importance; Swedish, the minority language in Finland and the language of the ECEC, was foregrounded at the expense of Finnish. The study suggests that partnership is a useful concept and a tool to understand the possibilities and challenges involved in promoting bilingualism in ECEC, especially in the context of a minority language ECEC.


Introduction
The aim of this study is to examine the sociolinguistic possibilities and implications of the concept of an 'early childhood education and care (ECEC) partnership for bilingualism' in offering bilingual children a bilingual childhood within mainstream minority language ECEC. In this case, the language of instruction in the ECEC is a national minority language with a high official status. ECEC partnerships can best be described as formal and informal relations between families and ECEC services and as describing all aspects related to the child in ECEC, for example, questions of secure and balanced growth, warm relationships, and the child's unique personality (Epstein 2011;Kekkonen 2012). The partnership for bilingualism which we examine here lies in the intersection of family language policy (King, Fogle, and Logan-Terry 2008) and the language policy of the ECEC.
Our focus is on mainstreambut at the same time minority language medium -Swedish education in Finland, an officially bilingual country. The educational programme analysed in this study is thus not a bilingual programme in the sense of actively following bilingual pedagogies for bilingual language acquisition (cf. e.g. Baker 2011). The ECEC has no official responsibility for teaching the other national language, Finnish; Finnish is introduced later, in school, where it is taught as an obligatory language subject (see below for a more detailed description of the educational system). The number of children from Finnish-Swedish bilingual families enrolled in Swedish-medium ECEC is high. Our focus in this study is on these bilingual families. We argue that discussing partnerships for bilingualism is equally important in both monolingual and bilingual ECEC.
In order to explore how parents in bilingual families and pedagogical practitioners construct an ECEC partnership for bilingualism, we interviewed nine middle-class families, in each of which one of the parents was a Swedish-speaking Finn and the other a Finnish-speaking Finn, and six Swedish-speaking practitioners.
The questions guiding our analysis were: (1) What major discourses are circulated by the parents and the practitioners when talking about languages and the ECEC partnership? (2) What kind of participatory roles (Törrönen 2013) emerge in these discourses displaying obligations, desires, abilities, and competencies for acting on bilingualism? (3) What kind of change in current ECEC partnership practices is needed, if any?
In the following sections, we will discuss educational partnerships with a focus on ECEC and describe the Finnish context for the study, before we present the study itself.

Educational partnerships
Educational partnerships occur at the intersection of ECEC, families, and the surrounding community. Epstein (2011) describes partnerships as a model of overlapping family, school, and community spheres. Her model accounts for the experiences of parents, teachers, and students as well as dynamic development over time. We have adapted her model for the purposes of our study in Figure 1.
Epstein argues that schools and parents can be seen to have either separate, shared, or sequential responsibilities in the partnerships. In the latter case, the early responsibilities of the home will later become the responsibilities of the school. The greatest overlap of the family and school spheres occurs, according to Epstein, during ECEC and the early elementary years. This makes ECEC Figure 1. ECEC partnership for bilingualism as overlapping spheres of relations in our study. Source: Adapted from Epstein (2011, 32). partnership a fruitful ground for studying bilingualism in the twenty-first century, since it establishes the ground for expectations on how child bilingualism is met and treated at later stages in educational institutions. In the research literature, 'parental involvement' and 'parental empowerment' are often alternative expressions to 'partnership' and they are used interchangeably (OECD 2012). We use the term partnership, as we wish to emphasize the two-way direction of activities and relationships in a partnership, that is, the shared responsibilities.
As for the school, the first sphere in Epstein's model, ECEC partnerships are generally seen in a positive light as they support student success rates in education (European Parliament 2013, 74;OECD 2012). However, it is also acknowledged that what is considered to be an appropriate level of parental involvement in a child's education is dependent on the cultural context. Questions involving partnerships for bilingualism in a wider international perspective are often associated with superdiversity and the challenges encountered when cultural expectations clash in schools (see e.g. Guo and Mohan 2008;Hutchins et al. 2012;Lightfoot 2004;Tran 2014), or with participation in bilingual education such as early immersion or late-exit programmes (Kavanagh 2014;Satterfield Sheffer 2003). A great deal of research is conducted to promote the home learning environment of 'at risk' students, for example, by engaging parents in low-income families or families with a lower level of education to support early literacy among their children.
A particularly interesting literacy study promoting bilingual identity was conducted by Taylor et al. (2008). In their study, culturally and linguistically minority parents in Canada were invited to participate as expert partners in the bi-literacy development of their kindergarten child while the children were collaboratively authoring dual language identity texts. This pedagogical innovation was grounded on the view of literacy as multilingual and multimodal and it gave family literacies a clearer and more active role in the kindergarten curriculum. This study also showed the overlapping and intertwined roles of experiences and ideologies involved in the partnerships, as well as pedagogical solutions.
As for the role of families (the second sphere in Epstein's model), language ideologies, language choices, and practices in families as well as families' choices of type of ECEC and school can all reflect what kind of expectations parents have of the partnerships for bilingualism (see e.g. Schwartz, Moin, and Klayle 2013). There is a growing body of studies on family language policy (e.g. Curdt-Christiansen 2013; King, Fogle, and Logan-Terry 2008;Schwartz and Verschik 2013;Wei 2012), but still little is known when it comes to bilingual families and their experiences of ECEC, particularly when the ECEC is in an official minority language. Our study thus fills an important gap in overall understanding of partnerships.
The third component in Epstein's model is the surrounding community. A bilingual family surrounded by a unilingual majority language community is faced with a different linguistic reality from that facing a family living in a bilingual community where the minority language has a more visible and prominent role. As we will see in the current study, different language environments play a significant role for the bilingual families as well as for the practitioners and affect their views on languages and partnership. Less research has been conducted on bilingual families living in bilingual communities than in unilingual communities (Barron-Hauwaert 2011). Hence, the role of a bilingual community in supporting ECEC partnerships for bilingualism is to a great extent still unexplored.

The Finnish context
Educational system and language policy ECEC in Finland is a part of the welfare state and every child has the right to participate in ECEC. ECEC includes day-care arrangements as well as goal-oriented early childhood education for children under the age of seven (EURYDICE 2015). In 2009, 46.1% of all 3-year-olds and 62.6% of all 5-year-olds in Finland were enrolled in ECEC (OECD 2012). Finland has a system of parallel monolingual education, which means that the administrative language of the ECEC and schools is always either Finnish or Swedish. The dual monolingual system is an effect of Finland being officially a bilingual country: Finnish and Swedish are official national languages with equal status, and both language groups are treated equally (Ministry of Justice 1999). Swedish is the lesser spoken national language in Finland: 89.3% of citizens are registered as Finnish speakers and 5.3% as Swedish speakers (Official Statistics of Finland 2013).
Bilingual Swedish-Finnish marriages are common in the Swedish-speaking population in Finland: around 40% of the children who are registered as Swedish speakers in fact have a bilingual family background (Finnäs 2012, 25;Tandefelt and Finnäs 2007). These bilingual families often choose Swedish-medium ECEC and the Swedish-medium school system as a way to support their children's development in the less spoken language in society (Kumpulainen 2010). The Swedish-medium educational system is commonly seen as essential for maintaining the high status of the Swedish language and Swedish-speaking cultural identity, and for ensuring the continued strength of the Swedish-speaking minority in Finland (Kovero 2011;Lojander-Visapää 2008). Another languagerelated reason for these families to choose a Swedish-medium educational system may be the general language teaching programme in it. Language learning results have been shown to be higher and attitudes towards language learning more positive in Swedish-medium schools than in Finnish-medium schools (Finnish National Board of Education 2015, 24).
It is obligatory for all pupils to study both national languages in Finland. In Swedish-medium schools, the teaching of Finnish usually starts when the children are 9-10 years old (in grade 3 or 4) and bilinguals can opt to read native level Finnish instead of the beginner's course. The teaching of a foreign language (most often English) starts at the age of 10-11 years (in grade 4 or 5) in Swedish-medium schools. 1 Importantly, there is no specific systematic support for the acquisition of Finnish as part of Swedish-medium ECEC.
The language of instruction in schools does not have to be the language of administration, and this makes it possible to have a wide range of bilingual programmes in both Swedish-and Finnish-medium ECEC and schools (Ministry of Education 1998). However, there is no established bilingual programme that is in general use for children who are from birth bilingual in Finnish and Swedish: immersion programmes in Finnish or Swedish, which are well established in Finland, are intended for monolingual children (Björklund, Mård-Miettinen, and Savijärvi 2014). It may be the case, however, that Swedish-medium and Finnish-medium ECEC units are located in the same building. This is the case for the ECEC sites of our study.
There are ongoing socio-political debates as to whether bilingual schools would be a better alternative for Finnish-Swedish bilingual children than the current monolingual ones. In an analysis of these public media debates, Boyd and Palviainen (2015) were able to identify different positions taken (preservationist versus ideologist), depending on what kinds of arguments were used. A closer navigation in these discourses showed that bilingual families were quite rarely represented or discussed in the media and individuals were often treated as representatives of one or other of the language groups, instead of being seen as belonging simultaneously to both. Moreover, practitioners' views on how Swedish-medium ECEC supports or hinders bilingualism are rarely heard. Taken together, these facts show the importance of research giving a voice to bilingual families as well as practitioners in these matters.

ECEC partnerships in Finland
The official language policy in the ECEC sites studied here is monolingual, following Finland's language legislation. The legislation and policy in Finland guarantee people's linguistic rights and give the authorities obligations where languages are concerned. (Ministry of Justice 2003;STAKES 2004, 18). This means that children and parents have linguistic rights in ECEC and the ECEC institution has linguistic obligations. The relations between the parents and ECEC can be characterized by a common trust in working for the best interest of the child. Trust in general is one of the most prominent ethical values in the Nordic countries, and Finnish citizens have great trust in public sector organizations such as educational institutions (Salminen and Ikola-Norrbacka 2010).
Most practitioners in Swedish-medium ECEC in Finland are bilingual or to some extent competent in both Swedish and Finnish, which increases the possibility of the flexible use of both languages by practitioners, children, and parents. There is, however, a lack of research focusing on bilingual pedagogy in mainstream classrooms when a bilingual teacher is available (see, however, Lemberger 1997;Palviainen and Mård-Miettinen 2015). In a British school context, Bourne (2001) showed that bilingual teachers tend to have the role of assistants to class teachers and to act as a link between the school and the home. Similarly, Robertson, Drury, and Cable (2014) found that use of the home language in the classroom was rare because of the expectations placed on the bilingual teacher to help children and parents to conform to the language of the mainstream monolingual classroom (for a related discussion on the Norwegian context, see Hvistendahl 2012). It should be noted that this type of research typically focuses on schools rather than ECEC. Hujala (2004) points out that the role of Finnish ECEC has changed considerably over the years, from a time when families were seen as standing in need of professional help to the modern position, where families are paying clients with the right to demand individual service. Partnerships were highlighted in the national curriculum guidelines (STAKES 2004) as a way of strengthening Finnish ECEC. Both partnership and individual plans 2 for each child are now a legal requirement. The concept of an ECEC partnership is, however, not yet clearly defined in either the Finnish or the wider international context. In Finland, it is described as being 'participation that goes further than cooperation', and it is said that 'the journey from cooperation to ECEC partnership requires mutual, continuous and committed interaction in all matters concerning the child' (STAKES 2004, 3).
In the Finnish context, Alasuutari (2010) examined how pedagogical practitioners position themselves towards parents in contacts and meetings, and found that they can take one or other of two different positions. The 'vertical position' emphasizes the practitioners' professional expertise and knowledge, while the 'horizontal position' gives parents equality in the partnership, which is seen as a two-way interaction in which both parties have relevant information about the child. Alasuutari (2014) also found that the parents' role in partnership discussions was restricted to what concerned the individual child: parents did not have a say in the more general educational practices of ECEC. These findings lead us to wonder if this holds even in partnerships for bilingualism; if so, it would mean that parents do not have a say in the choice of, for example, bilingual practices or pedagogics in ECEC.

Study design
The current study was part of a larger, ethnographically informed research project which had as its goal a deeper understanding of how concepts of language, bilingualism, and bilingual development are both manifested and negotiated in Finnish-Swedish bilingual families and in mainstream Swedish-medium ECEC located in three different geographical sites in Finland. The data included longitudinal observations of bilingual children in the ECEC and in their homes, interviews with children, parents, pedagogical practitioners and ECEC leaders, and written policy documents and current media debates on language policy issues.
The general methodological framework of the study is nexus analysis (Scollon and Scollon 2004), a type of ethnographic discourse analysis suitable for engagement in social issues (Lane 2014). We identified the social actionthe nexusto be examined in detail as the co-construction of an ECEC partnership for bilingualism in Swedish-medium ECEC. The main source of data in the current study was interviews with the parents of nine bilingual children, who were all around three and four years of age at the time of data collection, and six practitioners working in the children's ECEC.

Settings and participants
Data were collected in different regions of Finland. Details of the settings and the participants are summarized in Table 1. In the following section, we describe the participants and settings according to the spheres described by Epstein (2011, see Figure 1): firstly, the community; secondly, the ECEC and practitioners; and thirdly, the families.

The community
The ECEC units involved in the research were situated in different cities with varying degrees of structural bilingualism (for a more detailed description of the system of officially unilingual and bilingual areas in Finland, see e.g. McRae 2007). The first research site was located in an officially unilingual Finnish-speaking city with less than 1% of registered Swedish speakers. The second was located in the capital region of Finland, Helsinki, which is an officially bilingual city with around 6% of the inhabitants registered as Swedish speakers. The third one was located in an officially bilingual city with 23% of its inhabitants registered as Swedish speakers. This city is situated on the west coast, in an area where Swedish is more evident than in other districts of Finland (Tandefelt and Finnäs 2007).
The focus of the current study was on the partnership of the families, on the one hand, and ECEC and its staff, on the other. However, it was clear from the interviews that the surrounding community played an important role for the ECEC partnership. Although the linguistic landscapes varied from one research site to the other, all the interviewees commented on the general dominance of Finnish in the local society. Other potential participatory roles in the partnerships that can be attributed to the community sphere and which were mentioned in the interviews were, for example, the national curriculum guidelines, the library, municipal language coordinators, other nearby ECEC units and their staff, religious communities, universities and other teacher educators, time allocations/ restrictions, schools, and supports for learning. None of the parents was strictly monolingual. The L1 given is based on each parent's own language identification in the interviews. b Parent who did not participate in the interview. c Reported talking mainly Swedish with his children even if he identified himself as a Finnish speaker.

The ECEC units and practitioners
The three Swedish ECEC groups examined were all so-called co-located units: in each case, Swedishmedium groups and Finnish-medium groups were located in the same building. All the children could thus encounter both languages in the common areas like the playground, which were shared by all the ECEC groups. The size of the groups in all the ECEC sites was around 21 children and the teacher-child ratio was 1:7. Six pedagogical practitioners were interviewed, two from each site, all of them working closely with the children of the families included in this study. The practitioners had done their teacher training in Swedish and had a degree from either a university or a university of applied sciences, or at least an upper secondary-level qualification in social welfare and health care. Only one of the six practitioners reported that her teacher education had included anything to do with teaching for bilingualism (cf. Schwartz, Mor-Sommerfeld, and Leikin 2010). The practitioners were all to some extent bilinguals.

The families
Three bilingual families on each site were selected for participation, making a total of nine families. The data included 8 girls and 1 boy within the age range from two years and eleven months to four years and eleven months years at the time of data collection. In the initial phase of the project, we asked the staff at the three ECEC units to select families in which they believed one of the parents spoke mainly Swedish at home and the other parent spoke mainly Finnish. The selection was thus based on the ECEC practitioners' perception of bilingualism in the family rather than on any measurement of individual/family bilingualism or of measured language competences in the child. It later emerged from the interviews with the parents that they considered themselves to be bilinguals to various extents. Most of them believed they could cope in the other language if necessary. (Bergroth 2015). The partners in this study all shared similar ethnic, national, and cultural backgrounds, which minimized some of the cultural differences or clashes usually associated with a partnership in bilingualism and education (Broussard 2003). They also had a similar socio-economic status and a shared ideology of trust, which ought to have had some effect on the partnership.

Data collection
We conducted interviews with the parents from the nine bilingual families and six practitioners. In two of the families, only the mothers participated in the interview. All the interviews were carried out at the ECEC, except for one family that was interviewed at home. The parents could use either Swedish or Finnish, or both, in the interview, whereas all the practitioners were interviewed in Swedish.
The bilingual researcher, and the first author of this article, occupied multiple positions in the interview situations. Her general wish was to avoid any conflicting allegiances with the interviewed participants (Herr and Anderson 2015). In the interviews with the parents, the researcher positioned herself as both a researcher and a mother in a Finnish-Swedish bilingual family. She and the parents could thus be tapping into pools of shared knowledge about issues regarding the choices bilingual families have to make in Finland. In the interviews with the practitioners, the researcher positioned herself as a researcher with experience in bilingual teacher education. In this sense, the researcher was both an insider and an outsider in the observed reality.
The interviews were semi-structured around certain themes: parents were asked questions about language choices in the family, their definition of bilingualism, their contacts with and experiences of the ECEC unit, their linguistic background, and the linguistic landscape of their daily lives. Practitioners were asked comparable questions to those that were put to the parents, with the exception of languages used at home. Instead, practitioners were asked about the language practices in the ECEC and the role of policy documents in their work. Importantly, the concept of a bilingual ECEC partnership was not specifically mentioned in the interview questions, but emerged from the data in response to questions such as 'What languages do you use in your contact with the ECEC when picking up your children?', 'Do you think that you can influence the linguistic practices in the ECEC if you want to?', and 'Have you asked questions about or discussed bilingual growth in the ECEC?'. However, it is clear within the nexus analysis framework that the social action of participating in an interview for scientific purposes is not an optimal way of gaining any deep understanding of another social action, such as the co-construction of an ECEC partnership. The answers that were given were based entirely on parents' and practitioners' selective recall of different events and on accounts in which the partnership was constructed and negotiated. Events discussed included parents' evenings, picking up the children from the ECEC, and the annual meeting that takes place between parents and pedagogical practitioners in which the individual child and its development is in focus. This means that no actual practice was analysed; rather, it was discourses about practice.
The interview data corresponded to approximately 7.5 hours in total and they were transcribed. The examples from the interviews in this article have been translated into English.

Data analysis
We followed nexus analysis procedures, which means that the researcher is first engaging the nexus by recognizing and identifying a relevant social practice, then navigating it by mapping the relevant cycles of people, discourses and concepts circulating through the nexus in order to get an understanding of its complexity, and finally changing the social practice (Scollon and Scollon 2004, 153, 159). We identified the relevant social practice as the co-construction of an ECEC partnership for bilingualism as it was reflected through interviews.
In order to respond to the two first research questionswhat major discourses were circulated by the parents and the practitioners about ECEC partnership and bilingualism, and what kind of participatory roles for those concerned with acting on bilingualism emerged in thesewe distinguished and drew attention to the 'discourses in place … [that] are relevant or foregrounded … for the social action(s) in which we are interested' (Scollon and Scollon 2004, 14;italics in original). In this process, we identified in the accounts of parents and practitioners two major and recurring discourses: the first one was about language choice(s) in ECEC partnerships and the second was about who should support Swedish and/or Finnish and why. Within these, we mapped different types of 'participatory roles' (Törrönen 2013). The participatory roles model provides us with tools to analyse what kind of abilities and competencies for action one can achieve or lose by participating in an action (Törrönen 2013, 90). The four participatory roles which we used as tools in our discourse analysis were obligation ('having to'), desire ('wanting to'), ability ('being able to'), and competency ('knowing how to'). Using these concepts, it was possible to navigate and identify certain values and attitudes in the cultural perceptions of actions, for example, seeing support for bilingualism as either proper or improper.
The analyses led to the final step of the nexus analysis procedure and made it possible to address the third research question: what kind of change in the current partnership practices is needed, if any.

Discourses in place on partnerships for bilingualism: results
In the following section, we will present the two major, recurring discourses that we identified in the accounts of parents and practitioners: the first one was about language choice(s) in ECEC partnerships and the other was about support for both languages for bilingual growth.

Language choice(s) in ECEC partnerships
To be able to establish a working partnership, parents and practitioners need to negotiate which language will be used and by whom. Although the focus of this study is on the language chosen for use between parents and practitioners, the practitioners in their interviews often drew attention to the language that they should, and also do, use with the children: Swedish only. Occasionally, depending on the practitioner and the situation, a little Finnish would be used if it is the child's stronger language. These practices are based on the formal role that Swedish-medium ECEC has of guaranteeing the linguistic rights of the Swedish-speaking minority. Such discourses, on national policies and linguistic rights and on local language policy documents, frequently occurred in the practitioners' accounts, evidencing that their choice of the language of communication with the children was driven by obligation (Törrönen 2013): practitioners thought that they were obliged to use mainly Swedish with the children. This both showed how micro and macro discourse cycles circulated through the same nexus and made explicit the different micro and macro layers of policy implementation (Ricento and Hornberger 1996).
Unlike the communication between practitioners and children, communication between parents and practitioners is not defined by obligation, which leaves room for the negotiation mentioned above. Since all the interviewed participants said they were more or less able to communicate in both of the languages, there were numerous comments about these negotiations taking place. Most of these comments on the part of both parents and practitioners pointed out the non-conflictual nature of the negotiations. We asked the practitioners if the choice of language was negotiated or if it 'happened naturally' and they reported that they seldom needed to explicitly negotiate about the language: Practitioner Veronika (Sw): Umm, well it's kind of that you just, you just notice what language they are using and then we speak the language they want to speak. Practitioner Linn (Sw): I think it's rather natural somehow. I'm not really super talented in Finnish so sometimes there are some words and so on that I'm trying to find and I think 'how do you say that', but then they help me out, so we usually solve those problems.
In the examples above, the practitioners described how they monitored the parents' linguistic behaviour and adjusted to it. Common to all the comments about language negotiations between the adults was agreement that the practitioners opted to use the language chosen by the parents. Possible linguistic problems were solved in collaboration with the parents. These comments and observations about language choices seem to suggest a difference between Swedish-medium ECEC in Finland and other international, even bilingual, education. Kavanagh (2014), who studied partnerships in an Irish immersion programme, found that the school encouraged even parents to use Irish on school territory, a practice that caused parents anxiety and was a barrier to their communicating with the school. In the following quote, Anna's father, who had presented himself as basically monolingual in Finnish, humorously described his attitude to the realization that his children would become bilingual and that their language of schooling would be Swedish. This meant, among other things, that any activities arranged by the school would be carried out in Swedish: Anna's Father (Fi): Well, actually, when I realised that my children would become bilingual and she [= the mother] said that of course they will enrol in a Swedishspeaking school, and I thought, yes, that's how it will probably go. But then at some point I realised that the parents' evenings will be in Swedish! Help! Anna's father later explained how well he had managed to communicate with the ECEC, despite his background and his initial fears. This he attributed to the flexible use of language and his right to use Finnish with the ECEC practitioners. Interestingly, the possibility of negotiating only seemed to apply in spoken communication. Written communication was said to occur almost exclusively in Swedish: information sent out by the ECEC was always in Swedish, and the obligation to ensure the linguistic rights of Swedish speakers was highlighted. The ECEC practitioners pointed out that this practice was affected both by minority linguistic rights and by economic and time-related issues. The practitioners working in the unilingual Finnish-speaking city described how they sometimes had problems translating the children's lunch menus into Swedish, and how they needed to translate official letters into Swedish, even if they were not really competent to do so. One of the families regretted this monolingual policy in written communication since the Finnish-speaking parent did not always understand everything; the other families, on the other hand, accepted it, or at least did not comment on it. All three research sites had all extra events officially organized in Swedish, including parents' nights and special events such as Christmas parties. This official policy meant that practitioners used mainly Swedish when they talked to the parents as a group. However, in individual oral communication the choice of language was left to the parents and the practitioners adapted to it. The discourse on language choice was clearly filled with the obligation and the desire to guarantee minority linguistic rights. At the same time, there was a desire to communicate in a non-confrontational way with parents. Since the majority language, Finnish, was usually used with the majority of Finnishspeaking parents if they preferred this, the outcome was a flexible policy. Both practitioners and parents reflected on their partly lacking the ability to use the other language in written or spoken communication, but their joint efforts to communicate seemed to overcome any problems that arose.

Support for both Finnish and Swedish for bilingual growth
The other major discourse that was identified was around the issue of supporting roles. As part of this major discourse, we found two minor discourses, the first of them related to support for Swedish and the other to support for Finnish. We found that the attitude regarding support for Finnish was different from the attitude regarding support for Swedish (Table 2).
We will discuss the differences in these minor discourses in the following two sub-sections.

Support for Swedish
In our interviews, we asked parents and practitioners if they felt that parents could influence the linguistic practices of the ECEC unit and ask for the use of more or less Finnish or Swedish. The question was often treated as hypothetical by parents as well as practitioners, since the most common response was that they had never felt the need to do so. In other words, the parents seemed to be happy with the current linguistic practices at the ECEC. In the following example, Esther's father is firm about the fact that they deliberately chose a Swedish-speaking ECEC and that only Swedish should be used with the children. However, he also reports that the ECEC has planned to arrange joint bilingual activities with children and staff from the Finnish-speaking section as well, which means a mix of sections and languages. He reports that he has no objections to that: Ester's father (Sw/Fi): [ … ] because we wanted a Swedish-speaking place, so they talk Swedish there and shouldn't speak Finnish there. They have a lot of sections so there are Finnish-speaking sections too. I don't know It is important to keep in mind here that there is no official bilingual ECEC option available within the Finnish educational system, and we did not ask the parents to consider or reflect on hypothetical bilingual alternatives. The only alternative to a Swedish-medium ECEC for children like Ester in the example above would be a Finnish-medium ECEC, which would entail a total switch in the language used. In both Swedish-medium and Finnish-medium sections, a variety of flexible and pluri-lingual pedagogical practices could be used. It seems that the parents in our study are not against these kinds of bilingual practices, but they do not wish to switch the language from Swedish only to Finnish only.
Although there was a clear desire by both parties to actively support the use of Swedish at all times in order to develop the children's Swedish-language skills, parents as well as practitioners acknowledged that it was not always easy to do this (ability). Anna's mother discusses in the following example how the actual physical arrangements in the ECEC might make it difficult to keep Swedish as the main language. In this ECEC, both Finnish-speaking groups and Swedish-speaking groups shared the same playground and the children themselves chose what language to use there for communication. Anna's mother points out that it is difficult to change children's communicative habits, including their language choice: Anna's mother (Sw): When our eldest child was in day care the language of play easily became Finnish. Our daughter played a lot with boys and especially outdoors and we said to other parents that it's a shame [that the language of play easily became Finnish] and everybody thought it was a pity that that was how it was and we talked to the practitioners, but it's so funny, even if you go there and say, 'hey c'mon, talk Swedish', the kids react and they are like 'what?' for a moment and after a while it's switched back [to Finnish], so it's terribly difficult to change it then.
Supporting Swedish seemed basically to mean finding a balance between deliberately supporting, encouraging, and eliciting Swedish and at the same time being positive and receptive to the children's use of Finnish. When we asked the practitioner Henrika if she had ever felt the need to intervene in the linguistic choices the children made among themselves, she replied: Practitioner Henrika (Sw): We're not allowed to do that. Children have to be allowed to speak the language they want to speak.
She was clear that they were not allowed to ask the children to switch language and later on she referred to a local policy paper governing ECEC in her city where this was explicitly stated. She explained that if a monolingual Swedish-speaking child was left out because of the language used, the practitioners would try to explain the situation, discuss it, and negotiate about it with the children, but they could not go further than that and impose one language rather than the other. This means that the children have a say in language choices, too, and practitioners must respect their wishes and choices. Even if parents and practitioners agreed that it was not always easy to have much impact on the linguistic practices children were already used to, parents expressed their trust in the competence of the pedagogical practitioners and believed that they would do their best to find a balance between supporting Swedish and respecting the use of Finnish (competency). The partnerships seem to be built on parental trust that the professional knowledge of practitioners will work for the best interest of the child (Alasuutari 2010). This trust could be seen when Amanda's father told us that he was confident that the practitioners in ECEC were working to promote the Swedish language. For this reason he felt no need to intervene in the linguistic practices of the ECEC: Amanda's father (Fi): I do trust that they do it [=promote Swedish] too because it's a part of their professionalism to care for Swedish since it is Swedish speaking so it must be.
Bilingual development seemed typically to be an issue in partnership discussions only when a bilingual child's development in Swedish was experienced as being slower than their development in Finnish (cf. De Houwer 2009, 47-49). The pedagogical practitioners were rather cautious in the interviews when they talked about bilingual growth and about the advice they gave to parents. They reflected on situations where parents had asked about their child's future schooling, and how they had advised the parents to read a lot to the child in Swedish, and perhaps talk more Swedish at home if possible. In the research site in the unilingual Finnish city, one practitioner was more explicit when we asked her if the parents had asked questions about bilingual growth: The practitioner reported that she would find it more natural if parents talked only Swedish at home because the city itself was unilingual in Finnish. The child in question was more fluent in Finnish and the practitioner felt that the parents should support Swedish more than they did. Interestingly, however, there was a mismatch in conceptions, since the practitioner described the mother as a 'Swedish-speaking Finn', whereas the mother reported herself to be a 'Finnish-speaking Finn' with good skills in Swedish. According to the mother, the family followed the one personone language principle. Rather than opting for more Swedish at home, the family reported that they were looking into the possibility of moving to a bilingual city with stronger societal support for Swedish: rather than adding Swedish at home, the family wanted to add Swedish in the surrounding community. This shows the importance of external strategy for family language policy in the form of a bilingual community factor shaping the partnerships for bilingualism (Schwartz 2013). In this case, it seems clear that expectations about how and by whom Finnish and Swedish would be supported at home and in the ECEC unit had not been discussed in detail in connection with the partnership for bilingualism. Despite her personal opinion, the practitioner made it clear that as a practitioner she had to respect the linguistic choices the parents made. The same receptiveness to parental choices could be heard in other interviews as well. Pedagogical practitioners themselves trusted in their professional knowledge on bilingual growth mainly on the basis of their actual experiences in working life (competency). They pointed out that each child is an individual and that there are no easy answers to parents' questions about bilingual language development.
We asked parents if they had raised any questions about bilingualism and bilingual growth in ECEC when working with practitioners on writing up the obligatory individual ECEC plan for their child. Contrary to what we had expected, parents did not seem to recall any specific conversation about this. What they reported were general comments about Swedish-language development, the child's ability to communicate and to understand, or observations about language mixing: Ester's father (Sw): They said about, or well Ester, she speaks both, so they go a bit like, it's one word Finnish, one word Swedish sometimes, but it's okay, that it will develop later on. Tindra's mother (Fi): They said about Tindra, that she understands Swedish well, but I don't know if we discussed her language skills apart from that. Tove's mother (Fi): Last time, it was not long ago, we talked about the fact that she doesn't speak much yet, and that she talks in short sentences and phrases.
Ester's father told us that they had heard in the meeting that Ester typically mixed languages, but that they should not be worried about it. Tindra's parents noted that language skills had not really been an important issue in the conversation, but rather the focus had been on other matters regarding the child's wellbeing in ECEC. Tove's mother did not recall any discussion about the use of a specific language, but there was, rather, discussion of the general state of Tove's language development. It can be concluded that in supporting Swedish all the participatory rolesobligation, desire, ability, and competency - (Törrönen 2013) are present in the partnership discourses and they are intertwined in complex ways.

Support for Finnish
When we asked if the parents had shown any interest in promoting the Finnish skills of their bilingual children, the answers became more hesitant. Parents reported that they had given their permission for the ECEC unit to arrange joint bilingual activities with the Finnish-medium section in which Finnish would be used together with Swedish. In other words, the initiative for promoting the use of Finnish had come in these cases from the ECEC (desire and ability). Parents generally expressed the view that extensive and non-pedagogical language mixing by the practitioners was undesirable. The practitioners were also rather critical of the unplanned language mixing practices of other staff members. The unplanned mixing of languages seemed to them to symbolize giving in to the stronger, majority language in society, and potentially risking language shift in a place that should be committed to supporting Swedish in early childhood. Interestingly, the only person who expressed a slightly more positive attitude to the more extensive use of Finnish was Ella's mother, the only parent who identified herself as a monolingual Swedish speaker. In the following passage she answers our question about being able to intervene in the linguistic practices of the ECEC unit if necessary. She imagines a situation in which a lot of Finnish was used inside the ECEC unit: Ella's mother, like most of the parents, treated this question as hypothetical 'if I felt'. She concluded that even if she was able to intervene in the linguistic practices in ECEC, she would not necessarily choose to do so, because more Finnish in ECEC might help Ella with her Finnish: Ella was stronger in Swedish and did not normally use Finnish either at home or in ECEC. However, Ella's mother did not say that she would actively ask for more Finnish, and she did go on to say that she might feel different about this when the child was older: she might want to intervene if a great deal of Finnish was used in the pre-primary year, before the start of Swedish-medium school (when Ella is six). Interestingly, the city where Ella is taking part in ECEC is the most balanced in this study in terms of bilingualism. In this city, it is possible, at least in theory, to live monolingually in Swedish. This means that the role of ECEC here might not be as crucial for supporting Swedish in early childhood as it is elsewhere, since the surrounding society is more bilingual than at the other sites. It might be that Ella's mother does not experience Finnish as a threat to good skills in Swedish due to this rather strong community support for Swedish outside the home and ECEC. The desire to acquire the majority language, Finnish, in ECEC was not raised as an issue or even mentioned by the parents of bilingual children. Curiously, parents and practitioners reported that other families, who were monolingual Swedish speakers and who did not use any Finnish at home, wanted to hear more Finnish in ECEC in order to help their children learn it. However, as far as the children of bilingual families and their development in Finnish were concerned, the role of Finnish-language model seemed to be attributed to the home and to the surrounding majority language society, so ECEC did not need to take on that role. This was not a shared responsibility between families and ECEC. In the following example, Amanda's father is confused about the idea that anyone might want to intervene in the current linguistic practices of ECEC and promote the use of Finnish in their ECEC unit: Amanda's father (Fi): […] I started to wonder about the earlier question. If you come here and say, what's the point in that [=asking them to use more Finnish] because it doesn't change anything. Amanda's mother (Sw): No, but because this is a Swedish-speaking unit anyway. Father: Exactly, it doesn't make any sense for them to only speak Finnish.

Mother:
No, think about the direction. If the children were speaking Finnish all the time, the practitioners could intervene and say that we speak more Swedish here.
This example with Amanda's parents shows that they felt comfortable talking about promoting Swedish in ECEC, but her father was confused by the idea that anyone might ask the practitioners to use more Finnish or that anyone would use only Finnish. According to the father, it 'doesn't make any sense' to promote the use of Finnish in a Swedish-medium ECEC. The mother explained that in their situation it would be the other way around, meaning that if the children were using too much Finnish for their liking they could ask the practitioners to promote Swedish instead of Finnish. Promoting Finnish seemed to this family to be unthinkable. Parents, then, did not seem to expect that Finnish would be promoted in Swedish-medium ECEC. In the following example, Eva's mother is telling the researcher that she is thrilled to have just recently found out that bilingual children can have a more advanced course in Finnish in primary school than monolingual Swedish-speaking children (native level Finnish versus Finnish as the second national language, cf. above). When we ask her if she thinks the same system for acknowledging bilingualism would be good in ECEC, she responds: Eva's mother (Sw): In day care? No, somehow I just think that they are just taking care of them here now so they do handicrafts and play with other children so that they gain social [competence], so I haven't been thinking that yet at this stage, or expected that they would, or thought that I'd want them to acknowledge bilingual children in a specific way, no.
According to Eva's mother, the focus of day care for a three-year-old should be care rather than supporting bilingual language acquisition. One reason for this might be that testing and learning assessment are not part of Finnish ECEC (Määttä and Uusiautti 2012). Most of the families did value the learning of Finnish and wanted their children to develop in both of their languages, but this desire was not connected to ECEC. Amanda's family went one step further and reflected on the fact that neither Finnish nor Swedish is a widely spread international language, and the route to English and German is shorter through Swedish than through Finnish. This is most likely due to the fact that Swedish is a closely related Germanic language, whereas Finnish, as a Finno-Ugric language, is not related to the world languages: Amanda's mother (Sw): Yes, we have been talking about it, it was a clear choice that we wanted to enrol in Swedish-medium preschools and all that. Amanda's father (Fi): Swedish is the number one language and it's strong, and Finnish yes, it will eventually come along and about the language of their studies, I don't know, but this is my vision, they will study in Swedish and after that in English. Mother: Yes Father: That yes, of course it's good if they learn to write a little, a short letter in Finnish, and of course to read rather, mm, spoken Finnish is surely a bit easier than to write […] but it's ok if you can communicate and through that you might get interested in the written side of it. Mother: Yes, and I myself feel that the Swedish language is a great asset because it made it so much easier for me to learn English and German and all the other languages whereas Finnish, it's not really terribly, oh it might help you to pronounce Italian, but yes, it will come and I'm sure that they will learn it and they can already speak some Finnish.
In this example, Amanda's parents discussed their high academic expectations arising out of the Swedish language, but their more everyday communicative expectations from the Finnish language. Eva's family also reflected on multilingualism and multilingual competence as economic and social capital rather than bilingualism as a possible goal, if the child was interested in that path. In these discourses, bilingualism through Swedish-medium ECEC could be seen as a gateway not only to bilingualism, but further, to multilingualism. In contrast to the discourses on supporting Swedish, neither parents nor practitioners felt obliged to promote Finnish more than was currently done, nor did they express any desire to do so. Because of this it was not relevant for them to discuss the abilities and competencies needed for promoting Finnish. This means that the participatory roles of obligation and desire are the opposite of what they are for Swedish, and the participatory roles of ability and competency are not actualized.

Discussion
An ECEC partnership for bilingualism emerges in the nexus of multiple discourses and can therefore be interpreted from different viewpoints. The descriptions of partnerships given by both parents and practitioners in our study showed that bilingualism was considered to be the natural state of affairs and that the language choices between practitioners and parents were open to negotiation. Furthermore, the idea of the best for the child in the form of loving care (Puroila, Estola, and Syrjälä 2012) was generally considered to be more important than any strict observance of official language policies. Such attitudes allowed language mixing practices between all stakeholders and gave some linguistic freedom even to the children. However, the parents and practitioners alike considered the extensive use of Finnish with children to be undesirable. This view arose both from the need to secure the linguistic rights of Swedish speakers on a general level and also from the need to develop strong Swedish-language skills in all the children so that they can attend school in that language. Our findings confirm earlier findings by Robertson, Drury, and Cable (2014) that the presence of a bilingual practitioner is not enough for bilingualism to be actively promoted in mainstream classrooms. In this study, the bilingualism of the practitioners allowed parents to choose which language to use with the practitioners, but it did not greatly affect the linguistic or pedagogical practices used with the children. It did, however, also give the children the possibility of using Finnish if they so wished.
It could be observed that all the stakeholders agreed that supporting Swedish was of greater importance than supporting bilingualism, despite a generally positive and friction-free picture of bilingualism in these ECEC units. Swedish-medium ECEC was seen as an effective way of supporting the less spoken national language by creating a clear space for Swedish-speaking interactions. The existence of 'Swedish-speaking spaces' (Sw. 'svenska rum') cannot be taken for granted in settings where Finnish is the dominant, majority language. The discourses analysed in this study suggested that Swedish-speaking spaces offered bilingual children an important counterweight to the predominant linguistic environment in which they lived. This accentuates the importance of analysing the community sphere of partnerships for bilingualism in greater detail than has been done in this study.
The strong support for minority language rights that emerged in our data simultaneously placed the other home language, Finnish, in a tightly sealed box somewhere out of sight in the ECEC partnership. The acquisition of Finnish was taken more or less for granted, and it did not seem to be supported in any systematic way in ECEC. In other words, it seemed as if the acquisition and development of Finnish occurred mainly through exposure to it as the dominant language outside the ECEC. In fact, the partnership for bilingualism referred to Finnish only in cases when Finnish was the child's stronger language and more support was needed for the acquisition of Swedish. Although some Finnish was encouraged through joint bilingual activities for children from the Swedish-medium and Finnish-medium ECEC groups, these were not explicitly designed with bilinguals in mind. Hence, in these partnerships for bilingualism there seemed to be no parental desire for the promotion of a broader view of bilingual Finnish-Swedish child culture, for example, by offering songs, rhymes, and stories in Finnish in addition to Swedish.
It would be worth thinking more deeply about the apparent absence of any parental desire to support Finnish. Is it simply a question of there being no other ECEC alternatives or models available, is it about 'locked conceptions', or is it a conscious choice by parents to support Swedish only in ECEC? As researchers in bilingualism, we also need to ask ourselves whether there is a need to add support for majority language Finnish in minority language ECEC for bilinguals, and if so, what kind of language support that would be.
Our findings confirmed the need to continue to work on identifying the underlying conceptions around bilingualism in minority language-medium education, since they seemed to be in something of a mismatch with the public discourse demanding more bilingual options for bilinguals. It would also be worthwhile to explore in detail whether or not practitioners themselves feel that their work experience and prior education have given them the competency and ability to actively support Finnish for bilinguals in ECEC.

Conclusions
Our findings indicate that it is imperative to continue exploring partnerships for bilingualism between practitioners, parents, and community in order to understand the possibilities and challenges involved in promoting bilingualism through monolingual ECEC in the twenty-first century. Despite the cultural similarities between Swedish-speaking and Finnish-speaking Finns and a generally positive attitude towards bilingual language acquisition in Finland, the languages in question did not seem to be given equal importance when parents and practitioners were working together for the best interest of the bilingual child. Additional studies on partnerships for bilingualism are needed both in situations where the focus is on a possible linguistic or cultural conflict and in situations like our study, where bilingualism seems less conflicted. In our study, equality of language status in such partnerships seemed to be considered neither relevant nor desirable by either parents or practitioners, since the relative weight of the two languages was reversed in the surrounding Finnish society. This reveals a need for further studies, as it seems that the community sphere of partnerships has an important role in partnerships for bilingualism, perhaps a more important role than has hitherto been appreciated.
Social change is an important aspect of the nexus analysis conducted in our study (Scollon and Scollon 2004). With regard to changing the nexus of co-construction of an ECEC partnership for bilingualism, it is crucial to raise awareness of any languages placed in 'tightly sealed boxes'. All these boxes need to be opened up; it is important to discuss desires, obligations, abilities, and competencies in relation to all languages on an equal basis in order to find the best bilingual solution for each bilingual child. This would not necessarily mean any radical changes in the linguistic practices of ECEC, but it would provide new information and increase understanding of child bilingualism for both practitioners and parents. This in turn could help support each individual bilingual child in his/her bilingualism in the best possible way.

Transcription key
[…] Some content left out.
[ ] Content added for clarity.
[= ] Explanation for unclear references in citations is given inside square brackets. Notes 1. In Finnish-medium schools, most pupils study English as their first foreign language and start to study Swedish by the age of 12 years (in grade 6). Although native-level Swedish is an alternative for bilingual students in Finnishmedium schools, it is rarely offered in practice and much less so than native-level Finnish in Swedish-medium schools. 2. According to Act on the Status and Rights of Social Welfare Clients (Ministry of Social Affairs and Health 2001), the family and the ECEC jointly draw up a plan agreeing on how the care will be carried out. In all decisions made, the family's interests come first. Both family and the ECEC unit should be committed to the plan.

Disclosure statement
No potential conflict of interest was reported by the authors.