TURKISH IS NOT FOR LEARNING, MISS.

(1)

T URKISH IS NOT FOR LEARNING , ^MISS .

Valorizing linguistic diversity in primary education

Proefschrift ingediend tot het behalen van de graad van Doctor in de Taalkunde

2016

door Kathelijne Jordens

Promotor: Prof. dr. Kris Van den Branden Copromotor: dr. Koen Van Gorp

Onderzoekseenheid Taalkunde Faculteit Letteren

KU Leuven

(2)

(3)

D ANKWOORD

Dank jullie wel.

Dit stuk van de thesis is misschien wel het moeilijkste om te schrijven. Dat komt doordat ik verschrikkelijk verwend ben geweest onderweg, door ontzettend veel mensen.

Eerst en vooral dank aan mijn promotoren, Kris en Koen, die het aandurfden om mij aan te nemen, ondanks mijn twijfels over doctoreren, schrijven, academisch Engels en aanverwanten. Die thesis is er dan toch gekomen. Dank je, Joana, om in mijn begeleidingscommissie te zetelen en voor gezellige babbels met onze partners. Dank je, Koen J. en Andrea, dat jullie tijd wilden maken om jullie te verdiepen in mijn thesis en daar op de verdediging jullie licht over te laten schijnen. Thank you for coming over, Andrea.

Hartelijk dank aan de twee onderzoeksscholen, met een hart voor hun kinderen (De Berk in Paal en De Mozaïek in Houthalen). Het is niet evident om een onderzoeker met camera’s binnen te laten, maar ik was altijd welkom.

Dank jullie wel, mijn acht fantastische leerlingen, die in hun echte leven Aysenur, Aysu, Berdan, Bilgesu, Fatih, Ebru, Enes en Mert heten. Dank jullie wel voor jullie tijd, vrolijkheid en vriendschap. Ik kwam altijd heel blij thuis als ik jullie gezien had en ik word opnieuw blij als ik onze opnames bekijk. Jullie zijn fantastisch, ik ben heel trots op jullie.

Ik was gezegend met een team om mij te helpen met het Turks: Büsra, al was het maar kort, maar jij stelde me voor aan Nevim. Nevim, je deed het geweldig. Je hebt niet gewoon vertaald, maar ook steeds veel uitleg voorzien over Turkse uitdrukkingen, gewoontes en liedjes, inclusief links naar Youtube-fimpjes. Je praatte ook steeds over de kinderen alsof je hen echt kende, ook al zag je hen enkel op het scherm. Ook aan Orhan dank je wel om de stukken Turks die in de thesis staan nog eens helemaal uit te pluizen, ik weet dat dat niet jouw hobby is...

Annelies Jehoul, dank je wel voor de uren ploeteren met codes en quotes. Keep up the good work in jouw doctoraat. Dirk Speelman, ook jij ging aan de slag met mijn data.

De grafieken hebben veel inspiratie gegeven, maar ze zijn uiteindelijk uit de hoofdstukken gevallen... iets voor de toekomst.

Lieve CTO’ers, jullie zijn allemaal zo lief en meelevend dat ik soms binnensloop om niet te dikwijls te moeten uitleggen hoe het ging in de laatste periode. Blijf dat vooral

(4)

doen, het is eigenlijk geweldig! Dank aan alle chocoladeschenkers, dat zijn er velen onder jullie! Een specialleke voor de zolderbewoners: Bart – mijn feeder en submitmaatje, Lies – mijn compagnonne de route, en Joke, zonder wie ik het gewoon niet had afgemaakt. Binnenkort is het aan jou, ik zal er zijn.

Lieve Validivjes, jullie zijn ook al zo memorabel. Wat een team! Uitstapjes naar Gent waren favoriete werkdagen (naast uitstapjes naar Limburg natuurlijk). Anouk, Evelien, Evita, Ilona, Iris, Jill, Kirsten, Koen, Lies, Sven en Vicky, het was altijd heel gezellig met jullie. Orhan en Lore, dank jullie wel voor jullie extra begeleiding, aanmoedigingen en opbouwende feedback. Jullie geloofden in mijn academische kunnen (vaak meer dan ikzelf).

Lieve vriendjes allemaal,

die vriendjes die ik veel heb mogen zien: Jeroen H. voor de uren Commerce die ik meteen wil overdoen, Miekemeid (onze maandagavond is terug!) en de Loosjes, ons reservegezin, wat zouden wij nog zijn zonder jullie?

en die vriendjes die ik veel te weinig heb gezien maar met wie het altijd weer zo leuk is...

De Henckies in Herent, de Coolskes in Herent, de Hearinckjes in Gent, de Vandeweertjes in Herent, de Verhaeghetjes in Brugge, de Dumonkes in Neerijse, de Serrientjes in Geel, de Tielenskes in Alken, en alle Spaarstraatburen: deze zomer ben ik thuis en de bbq staat klaar!

Heleen en Olif, Shalini en Lies W., Hilde en Herlinde, dank jullie voor de zeldzame, maar steeds vrolijke lunchdates. Stijn en Thomas, wanneer komen jullie nog es koken?

Marleen en Jan en de Capella Nova, Kurt en de Capella di Voce, Inge en de Cava- madammen, en Patricia: dank jullie om mijn zingende zinnen met mij te verzetten.

Partituren worden weer prioritair! DANSbar, ik kom er weer aan!

Lieve familie,

Francis en Linda, dank jullie ook de heerlijke maandagen, en voor alle uren en uren dat jullie met liefde voor onze meisjes hebben gezorgd als ik weer zat te typen.

Maarten en Kathleen, Robbe en Stiene, bij jullie zijn het steeds zalige pauzes.

Mijn lieve grootouders, die heel erg meeleefden, van jullie heb ik leren volhouden.

Zoals bommake het zegt: ‘altijd ma voorrut menneke’ (denk in sappig Kempisch).

Elisabeth en Johannes (Liske en Jogie), mijn zus en broer, Bart en Karolien, mijn schone broer en zus, en mijn nichtjes Marjolein, Helena en Charlotte: Ik kijk steeds uit naar de vrijdagen, om weer even heerlijk zus en tante Thijne te kunnen zijn.

(5)

Mama en papa, toen ik langsgefietst kwam om te zeggen dat er een tof project was op het CTO, maar wel met een doctoraat eraan verbonden, zeiden jullie meteen ‘doen’.

Jullie stonden altijd klaar voor alles: vrijdagavondfestijnen, ritten naar Toverhuis en Regenboog, woensdagnamiddagenopvang, acute ziekenboeg (gelukkig niet vaak nodig) en tweede thuis voor de meisjes als ik op conferentie zat. Zonder jullie was het niet gelukt.

Jeroen, dank je wel voor alle hulp bij vorm en inhoud, en dan bedoel ik niet alleen de thesis. Ik ben nog steeds verliefd.

Mijn meisjes, Annabel ‘ik heb een doctoraat voor jou geschreven, mama’ en Maite ‘ik genees jou in mijn doctoraat, mama’, ik kan jullie niet beter verzinnen dan jullie zijn.

Wie ik vergeten zou zijn, wellicht omdat ik het te evident ben gaan vinden dat je deed wat je deed. Dank je wel voor de kleine dingen die ook erg groot zijn geweest.

Laat de zomervakantie maar beginnen.

(6)

(7)

A CKNOWLEDGEMENTS

This work was supported by the Agency for Innovation by Science and Technology (IWT), Flanders (Belgium) under Grant 110008 (Valorizing Linguistic Diversity in Multiple Contexts of Primary Education ‘Validiv’)

(8)

(9)

L IST OF T ABLES

TABLE 0.1: PARTICIPANTS ... 7

TABLE 0.2: OVERVIEW OF DATA COLLECTION ... 11

TABLE 1.1:SUMMARY OF TASKS CHARACTERISTICS ... 27

TABLE 1.2:PERCENTAGE OF TURNS SPOKEN BY THE RESEARCHER/GUEST TEACHER ... 28

TABLE 1.3:PERCENTAGE OF TURNS IN DUTCH,TURKISH, BOTH, AND OTHER FOR EACH TASK ... 30

TABLE 2.1:PERCENTAGE OF TURNS IN DUTCH,TURKISH, BOTH, AND OTHER FOR EACH TASK ... 51

TABLE 4.1: PARTICIPANTS ... 94

TABLE 4.2: OVERVIEW OF DATA COLLECTION ... 96

(15)

L IST OF A PPENDICES

APPENDIX 0.1:THE TASKS ...159

APPENDIX 0.2:GUIDELINES DATA COLLECTION STUDY 4 ...179

APPENDIX 2.1: THE FINAL CODING TREE...187

APPENDIX 2.2:DISTRIBUTION OF FUNCTIONS ...191

APPENDIX 4.1:LANGUAGE PLANETS ...193

APPENDIX 4.2:THE LANGUAGE PORTRAIT OF NURAN ...194

APPENDIX 4.3:THE LANGUAGE PORTRAIT OF OZAN ...195

(16)

(17)

0 I NTRODUCTION

As a remedial teacher in a Flemish primary school, I tutored two multilingual in mathematics. One day, the boys didn’t know how to solve a particular math problem.

They spoke to each other in Berber, just a few sentences. This was against the school rules: they had to talk Dutch, which was the official language of instruction. The interaction in Berber only lasted a few seconds and I just let it be. A few things happened: the boys solved the math problem, and we could move on to the more difficult exercises in a pleasant atmosphere.

Ten years later, the Validiv project was launched. Validiv stands for ‘Valorizing linguistic diversity in multiple contexts of primary education’. The rationale behind the project was based on the tension described above: despite the growing amount of multilingual pupils in Flemish education, the vast majority of schools in Flanders (the northern part of Belgium) adopts a monolingual policy: only Dutch (the official medium of instruction) is allowed at school. While school staff use various arguments to support and maintain this monolingual policy, more tolerant approaches toward multilingual language use are advocated in the recent literature on multilingualism in education (e.g. Cummins 2000; Baker 2010; Creese and Blackledge 2011; Garcia and Wei 2014). The Validiv project aimed to shed light on the tension between a monolingual educational approach on the one hand and an approach valorizing linguistic diversity on the other hand.

As a team member of the project, I got the opportunity to work with (Turkish-Dutch) multilingual pupils, inviting them to use all their languages while doing group work.

That way I could find out for myself whether they would use their mother tongue (Turkish), whether mother tongue use could be of help to them and if so, in what ways.

I also wanted to know how the pupils themselves experienced their language use.

Each of these questions is addressed in one or more of the four explorative studies reported in this dissertation.

In the introduction, we first consider arguments pro monolingual education, and then turn to the counterarguments for multilingual education. We shortly describe the Validiv project and situate the four studies in the larger project. Next, the methodology is described and an outlook on the following chapters is provided.

0.1 Monolingual or multilingual approach to education?

School teams often have good reasons to adopt a monolingual Dutch policy. A first argument for a monolingual policy school teams put forward is that pupils should

(18)

integrate in society, which requires a high proficiency in Dutch (Van Praag 2013;

Kuiken and Van der Linden 2014). Secondly, to make the language policy as clear as possible, there is a simple rule: at school we speak Dutch. The argument is that this simple rule is easy for everyone to understand and follow. A third argument that school teams put forward is that ethnic minority children’s mother tongue skills are

‘too poor’ to function as a resource for learning: their vocabulary is seen as limited and their grammatical development as basic(Ağirdag, Jordens, and Van Houtte 2014;

Sierens and Van Avermaet 2014). Fourthly, a ‘time on task’ argument is raised: the more Dutch children hear and speak, the better they will learn it (Van den Branden and Verhelst 2007; Sierens and Van Avermaet 2014). Lastly, the argument is raised that teachers cannot understand all mother tongues. This leads to a perception of a loss of control undermining good teaching (Berben, Van den Branden, and Van Gorp 2007; Sierens and Van Avermaet 2014).

While these arguments are all valuable, a counter argument for each of them is provided in the literature on multilingualism and education. As for the integration argument, scholars point to the difference between integration and assimilation (Mehan, Hubbard, and Villanueva 1994): one can be perfectly integrated and still speak his or her mother tongue. Thus pupils adhering to their mother tongue should not automatically be regarded as ‘not integrated’. Moreover, Flemish society is becoming super diverse, and so is the linguistic landscape (Vertovec 2007). Since it is an important aim of education to prepare children to become citizens, schools should take advantage of the diversity in the school population to show how pupils speaking different mother tongues can live together in the mini-society of their school, thus preparing the pupils for citizenship in the society outside school.

As for the second Dutch-only argument, authentic language use is not that clear cut.

Bilingual speakers tend to mix their languages, which is a very natural way of speaking for them (Romaine 1989; Zentella1997, Jørgensen 2005). So while a one-language- only rule may seem very straightforward for a monolingual person, for a bilingual person it might be unnatural to separate languages that way (Cruz-Ferreira 2010). As a consequence, speaking just one language within the school walls, may be regarded as an unnatural way of communicating by multilingual pupils. Thus while the one- language rule may be clear, the practical implications for multilingual speakers might not be. As Butzkamp (1998) notices in the context of a history lesson taught in English as an L2 in a German school: ‘Code-switching is an integral part of the speech of bilinguals. The mother tongue does not take over but is a necessary conversational lubricant. Even if it was possible to banish it from the classroom, it could never be banished from the pupils’ minds’ (95).

‘Language poverty’ was the third argument raised above against the use of mother tongue use at school. However, Grosjean (2010) warns for the myth that bilinguals should have equal and perfect knowledge of both their languages as bilingual

(19)

speakers show varying degrees of competence in different languages. Moreover, even if some language skills are rather limited, that does not necessarily imply that a learner cannot draw on them (e.g. Gibbons 1998). Prohibiting mother tongue use might mean that a potential cognitive resource (i.e., the full linguistic repertoire of a pupil) that can further learning, is denied to the pupils (Cummins 2007, Baker 2011, Garcia and Wei 2014).

Fourthly, the ‘time on task’ argument is contradicted by the interdependence theory of Cummins (1979a). Cummins states that a bilingual speaker has an ‘underlying proficiency’, as a result of which practicing one language may lead to better skills in the other language as well. Recently, Winsler, Kim, and Richard (2014) found that children’s Spanish language competence at age 4 positively predicted their English proficiency one year later in kindergarten, even after controlling for other background variables. This is in line with previous research showing the importance of strong L1 language skills for L2 development (e.g. Verhoeven 1994; Ordóñez, Carlo, Snow, and McLaughlin 2002; Cárdenas-Hagan, Carlson, and Pollard-Durodola 2007;

Huguet 2014). This indicates that communication in the mother tongue can indirectly contribute to skills in the official medium of instruction, or, at least, that mother tongue use does not undermine the acquisition of a second language.

The final argument, that a teacher cannot know all mother tongues, is valid. In this respect, two projects valorizing multilingualism at school were carried out in Flanders in the past: the bilingual Foyer project in Brussels (Leman 1985) and the ‘Home language project’ (focusing on biliteracy and language awareness) in Ghent (Ramaut et al. 2013). These projects involved, at least partly, teachers speaking the mother tongues of children. However, in view of the high number of different mother tongues that are spoken in some Flemish classrooms today (especially in the bigger cities), this approach is not practicable in every single Flemish school. Nevertheless, it noteworthy that the regular teachers (thus not the Turkish teachers) involved in the Ghent home language project indicated a positive effect of the project on the well- being of the bilingual children, which is an important precondition for learning (Cohen 2006; Dulak et al. 2011). In this respect, Cummins (2001) states: ‘To reject a child's language in the school is to reject the child. When the message, implicit or explicit, communicated to children in the school is: “Leave your language and culture at the schoolhouse door"; children also leave a central part of who they are - their identities - at the schoolhouse door. When they feel this rejection, they are much less likely to participate actively and confidently in classroom instruction’ (19).

0.2 The Validiv project

The Validiv project draws on the idea of ‘functional multilingual learning’ as it aims to valorize pupils’ full linguistic repertoires without enforcing invasive changes in the participating schools’ daily classroom practice (i.e., no introduction of multilingual

(20)

teachers, no extra courses in the curriculum). The Validiv-project had two aims: (1) implementation of multilingual tools (e.g. an E-learning environment, language awareness activities), guided by external pedagogical coaches and (2) investigating multilingual practices and language policies in the primary schools involved. The research component resulted in six PhD dissertations: three quantitative studies, in which 67 primary schools in three regions (Ghent, Brussels and the former mining areas in Limburg) were involved, and three qualitative studies (in the same areas).

Together, the dissertations cover multiple aspects of dealing with linguistic diversity at school: school policy, teacher perceptions and practices, and pupils’ perceptions and practices. The current dissertation is a qualitative study on multilingual peer interaction during group work. It was situated in the province of Limburg, more particularly in the former mining areas, a region with a large Turkish community.

0.3 This dissertation

This dissertation aims to explore what happens if mother tongues are brought into the class with a teacher who does not speak these languages. It focuses on group work involving peer interaction between pupils with the same mother tongue (Turkish).

The research questions were based on preliminary interviews with teachers. They appeared to have questions which were very similar to my own considerations mentioned above. Teachers wanted to know what would happen when Turkish would be allowed: Would pupils talk Turkish? If so, how much and for what purposes? The first two studies of this dissertation address these questions. Two more studies were carried out: one to shed light on the pupils’ intertwined use of Dutch and Turkish as one medium of communication, and one to investigate how pupils experienced and reflected on their own language use. In total, four studies were carried out, based on the same data set, but with different methodological approaches.

To that end, I took two groups of four bilingual pupils out of their regular class, and installed a ‘multilingual space’ where all languages were welcomed. In this multilingual space, each group of four carried out eight different tasks. Also, pupils were interviewed extensively.

The main questions guiding this dissertation are the following:

(1) Which language choices do the bilingual pupils make across tasks? Are there any differences in the proportions of languages used across task performances and can those differences be explained?

(2) If Turkish is used during group work, what functions does it serve across tasks?

(3) Can translanguaging practices (intermingling Dutch and Turkish)

(21)

contribute to task performance? If so, in what ways?

(4) What is it like to be a multilingual pupil speaking a minority language, in particular in a monolingual policy school? How do the multilingual pupils experience the new translanguaging space?

0.4 Methodology

Each study draws on the same data, collected with the same eight participants.

However, the focus, guiding research questions, and method of analysis differs between the four studies. This section describes the methodological principles the four studies have in common. Specific information considering the data analyses is provided in the method section of each study.

0.4.1 Terminology

Before describing the methodology, we consider some terminological choices. In this section, the choices for particular terms (related to multilingualism) used in this dissertation are justified.

We use ‘mother tongue’ and not ‘home language’ to refer to Turkish. In the interviews, the pupils themselves referred to Turkish with the term ‘mother tongue’. When asked about their ‘home language’, they mentioned both Turkish and Dutch.

We only refer to L1 (first language) or L2 (second language) when describing research using those labels. The numerical terms suggest some kind of hierarchy, but in the interviews with the pupils, it became clear that they do not agree on the existence of a fixed hierarchy in their linguistic repertoire. Three of them stated Turkish is their L1, three of them answered Dutch, and two of them answered ‘both’.

We use the term ‘bilingual’ or ‘bilinguals’ when emphasizing the fact that the focus pupils can express themselves quite fluently in both Turkish and Dutch. However, sometimes we will refer to them as being multilingual, because they have notions of French, English, German and some other languages. Taking these skills into account the term ‘multilingual’ is more accurate. We considered using the term ‘polylingual’

or ‘plurilingual’, but decided to consequently use ‘multilingual’ for the sake of uniformity. We acknowledge, however, that the added value of ‘poly- or plurilingual’

is in its nuanced reference to the flexible and fluid language(s) use by individual speakers as opposed to their ability to speak two different languages, or the skills of an individual speaker as opposed to multilingual practices of society as a whole.

In the thesis, we refer to ‘multilingual spaces’ (in study 1 and 2) and ‘translanguaging spaces’ (in study 3 and 4). While these are used as synonyms (both referring to a welcoming climate toward all languages of the pupils), we chose to use both

(22)

‘multilingual’ and ‘translanguaging’ space to align with the framing of the particular studies, as pointed out below.

0.4.2 Participants

The participants were pupils of Turkish origin (aged 9-12), who went to a primary school in the former mining area in Limburg which had agreed to participate in the Validiv project. To situate the participants’ background, we shortly sketch the Turkish community in Limburg they live in, and the schools they went to. We then turn to the selection procedure of the participants.

0.4.2.1 Pupils of Turkish origin

The study focuses on pupils in Flemish primary schools with a Turkish immigrant background. These pupils are of particular interest in the Flemish educational context for two reasons. Firstly, the pupils of Turkish origin constitute a relatively large minority group in Flanders, and they have been shown to actively maintain their mother tongue and culture (Leman 1997; Yagmur 2009). Secondly, on average, the academic achievement of pupils with a Turkish mother tongue background is poorer than other groups of immigrant pupils (Verhaeghe, Knipprath, and Mertens 2007;

Belfi et al. 2011). Many teachers tend to perceive the mother tongue use of Turkish pupils as one of the major causes of their academic underachievement (Ağirdag, Jordens, and Van Houtte 2014).

0.4.2.2 Two out of ten project schools

At first, our intention was to carry out an ethnographic study in a class where using Turkish during task performance was allowed. As we did not find such a class (16 schools in the mining area were asked), we reconsidered the design of our study. We decided to create ‘multilingual / translanguaging spaces’ outside the regular class in two Validiv schools with a population of at least 50% pupils of Turkish origin. Out of the 10 Validiv project schools in Limburg, 8 schools fulfilled this requirement. Three schools were invited to participate because they had only one class in the fourth grade. This would make it easy to follow pupils during consequent school years, as classes would not be mixed over two school years. One of the three schools did not wish to cooperate, because they had no extra classroom available for me. As a result, two schools participated.

Data were collected in two Dutch-medium primary schools located in Limburg, a region with a high concentration of pupils with an immigrant background, of which the vast majority is from Turkish origin. In both schools (called ‘green’ and ‘blue’), this was reflected in a school population with more than 50% pupils speaking a mother tongue other than Dutch (mostly Turkish).

At the time of the data collection (2013–2014), both schools had a strict Dutch-only

(23)

policy, compelling pupils to use Dutch in class and at the playground at all times. The reluctance of both school teams to tolerate mother tongue use is illustrated by the fact that they agreed to cooperate with this study only under strict conditions, namely that the use of mother tongue (Turkish) would be restricted to the periods of data collection, during which a group of four pupils was pulled out of the regular classroom to work with me as a guest teacher.

Despite the similar ‘Dutch-only’ policy in both schools, there seemed to be a difference in the way this rule was implemented. In the green school, with a rather mixed population, teachers seemed to be rather lenient. When pupils of the green school talked Turkish, teachers pointed out they shouldn’t, or asked them not to do so anymore. In contrast, in the blue school with almost 100% bilingual children, pupils were often punished for speaking Turkish (e.g. they had to leave the classroom or playground, they were given a remark, they were sent to the principal).

0.4.2.3 Two groups of four pupils

In each school, two girls and two boys (aged 9-12) participated in eight group work sessions, spread over 1.5 years (while they were in grades 4 and 5). They were chosen randomly out of the list of all children who had indicated in the Validiv questionnaire (1) that they used Turkish at home with at least one of their parents and (2) that their Turkish speaking and listening skills were ‘very good’ (at least 4 on a 5-point Likert scale). The self-assessment of ‘very good’ gave an indication that pupils did not just have basic notions of Turkish, but would be able to use it in their task performance.

The information provided by the pupils was checked with their teachers, and informed consent of the school team and the parents was obtained.

The pupils were invited to choose their own pseudonym (see Table 0.1).

Table 0.1: participants

Gender School Pseudonym School Pseudonym

Boy Green Doğukan Blue Ali

Boy Green Ozan Blue Kerem

Girl Green Mira Blue Melisa

Girl Green Nuran Blue Selina

All participants were born in Belgium, except for Melisa (blue group) who moved to Belgium when she was three years old. Since all participants had attended a Flemish,

(24)

Dutch-only school for about seven years at the start of data collection, they had acquired academic skills in Dutch. All pupils had a similar socio-economic background; their parents were either unskilled blue-collar workers or unemployed, and therefore are representative of the vast majority of Belgians with a Turkish immigrant background (Van Craen, Vancluysen, and Ackaert 2007).

0.4.3 Group work

There is a vast amount of studies documenting the value of group work and peer interaction for learning (e.g. Rogoff 1990; Mercer 1996, 2004; Wells 2009). The majority of those studies draw on sociocultural theory inspired by Vygotsky (1978).

The underlying rationale is that the development of individual higher mental functions originates in social experience and interaction (Maybin 2013). Language plays a key role in human cognitive development as it mediates not only our relationship with others but also our own mental activity (Lantolf and Thorne 2006).

This hypothesis has been investigated, among others, in studies of multilingual interaction in which students can draw on their full linguistic repertoire as a resource for task performance and learning (e.g. De Guerrero and Villamil 2000; Swain and Lapkin 2000; Storch and Aldosari 2010; Moore 2013). In all of these studies, mother tongue use in peer interaction appears to serve multiple cognitive and socio- emotional purposes. Therefore, we decided to have pupils perform tasks in groups, to investigate whether they would draw on their full linguistic repertoire.

0.4.4 Multilingual space – translanguaging space

I took the children from their regular classroom to another one (like the music class, or the home work class they were familiar with). The four pupils were seated around one table. In our first session, I explained that they would be video- and audio- recorded, and that a translator would provide me with the translations of everything they would say in Turkish.

I did my best to foster a classroom climate in which all languages, including Turkish, would be welcomed (I explicitly mentioned Turkish because for the last six years at school they had been asked not to use any Turkish at school). Every session started with an informal chat and a conversation related to speaking/being Turkish. At first, I asked them to teach me some words in Turkish. When I got invited to a wedding of a Turkish friend, I showed them the invitation and asked for a translation (it was written in Turkish) and for advice about presents, clothing, habits and so on. The next time I met the groups they were very curious and we talked about what the wedding was like. Another time, we talked about their visits to their family living in Turkey during summer. This way an atmosphere developed in which speaking Turkish was accepted and ‘normalized’. I made that explicit every time we met: ‘you can speak whatever languages when we are here, Turkish as well’.

(25)

It should be noted that throughout the period of data collection, the monolingual school policy of the two schools involved remained unchanged. Therefore, I considered the multilingual/translanguaging space to be ‘multilingual islands in a monolingual sea’.

0.4.5 My role: researcher and teacher

Originally, I wanted to carry out an ethnographic study, observing language use in a translanguaging space as a non-participating observer. This implied my ‘teacher role’

would be restricted to giving instructions at the start of each task performance and observe the pupils while working. From the very beginning this turned out to be impossible. The pupils immediately asked for my help (just like their own teachers had predicted). Thus, I played a double role: researcher and teacher, which implied a mixed form of observer and participant observer. As all data will be analyzed and results will be discussed from this double perspective, it is important to outline my background and point of view. As Fisher (2009) points out: ‘In actuality, perspective can never be ruled out. We can perceive only from perspectives’ (584).

I am Flemish-born and my mother tongue is Dutch. I speak English as a foreign language and I have basic notions of French, German and Spanish. It is important to note that I do not speak Turkish at all (that is, at the onset of the study, because thanks to the participants I have learned ‘weird’ Turkish vocabulary including names of a few animals and quite a few Turkish swear words). I am aware of the importance of Dutch:

I share the concern of Flemish teachers that pupils need Dutch language skills to function in Flemish society. However, based on both my literature study and my experience as a remedial teacher in a multicultural school, I do not share the belief, omnipresent in many school teams, that Dutch should be acquired at the expense of mother tongues (Ağirdag, Jordens, and Van Houtte 2014).

Clearly indicative of my double role was my ambiguous feeling toward the participants’ use of Turkish. As a researcher, I was happy when they talked Turkish, because I was curious about the role of Turkish in their task performance. As a teacher, however, it was sometimes frustrating and difficult not to be able to understand them.

0.4.5.1 Role of researcher

As a researcher, I conducted all the interviews and guided all focus groups in Dutch. I considered taking my translator with me, because she grew up in the same region and speaks the same varieties of Turkish and Dutch as the participants. Moreover, she would share the pupils’ experiences of what it was like to be of Turkish origin in a Dutch-only policy school. However, due to practical reasons (transport, extra translations) and due to the effort I put into building a personal relationship with the children, we decided it might be best if I supervised all task performances, interviews

(26)

and focus groups.

0.4.5.2 Role of teacher

I am pretty much the prototypical Flemish teacher (female, in my thirties, Dutch- speaking, and speaking no minority language). This contributes to the ecological validity of the study. The children seemed to perceive me as ‘a real teacher’: they were quite surprised that I did not work in a primary school, but at a university.

My role as a teacher was restricted to explaining the task instructions and being around to help them out if necessary. Most of the time, I was sitting behind the teacher’s desk at some distance from the pupils’ table, taking field notes. My intention was to only intervene when they asked me to, or when their noise would disturb the neighboring class. In practice, I intervened more than I had anticipated, as will become clear in the next chapters of this dissertation.

I was a typical teacher in the sense that I could not understand Turkish. I guided them through the work they were to do: I was mostly supportive and encouraging, and only strict when necessary. I was an atypical teacher in three aspects: (a) I was there just for the four of them, (b) I took them out of their regular classroom, and (c) I allowed (and even welcomed) the use of Turkish.

0.4.5.3 Field relationships

During my time with the pupils, I built a warm relationship with them. By the end of the data collection, I was welcomed with hugs. Pupils offered me help to carry my materials, and were keen to come to the ‘translanguaging space’. They shared little tidbits and gossip with me, and trusted me when talking about their teachers.

However, in the latter case, they tended to ask me not to pass on what they had told me (for example tricks to crib during tests). When there were quarrels between pupils, I tried not to take sides, but have the pupils solve the dispute on their own.

Over time, I got involved emotionally with the children. During our last meeting, we agreed to be Facebook-friends, on the condition that they would send me a request.

Six out of eight children did so.

0.4.6 The data set

In total, twelve sessions were organized which delivered the data set for this dissertation. Data for the first three studies (pupils’ language behavior during group work) were collected during the performance of eight tasks. Data for the fourth study (pupils’ perceptions) were collected after each task performance and during three data collection moments at the start, in the middle and at the end of the data collection period. All sessions were audio- and video-recorded and transcribed verbatim.

(27)

Table 0.2: overview of data collection

Session Data collection for the first three studies

Data collection for the

fourth study Grade

1 Language portraits and

language planet focus groups 4

2 Task 1 Dino Interview 4

3 Task 2 The game Interview 4

4 Task 3 Kami 1 - 5

5 Task 4 Kami 2 Interview 5

6 Language portraits and

questionnaires 5

7 Task 5 Vulcano Interview 5

8 Task 6 Bubbels Focus group 5

9 Task 7 Measuring trees Interview 5

10 Task 8 Turkish class Focus group 5

11 In-depth interviews 6

12

Closing session, feedback on main findings and informal

chitchats

6

0.4.7 Eight tasks

As the studies in this dissertation are exploratory in nature, we aimed at covering a range of tasks, eight in total, representing the different subjects in the Flemish curriculum. The original versions of the tasks can be found in Appendix 0.1.

‘Dino’. Task one was a science task with an information-gap eliciting discussion. The goal of the task was to find out which of four different texts contained the most plausible hypothesis for the extinction of dinosaurs. The four pupils had to discuss the different options, reach a consensus and present their conclusion to the ‘multilingual

(28)

camera’. Detailed task instructions (provided on paper) allowed the pupils to work together with little or no interference of the researcher.

‘The game’. In task two, pupils were challenged to play a game on the playground.

During the game two dyads of pupils had to develop strategies to capture the hidden dinosaur eggs of the other dyad. A game board with sixteen squares had to be filled with sixteen cards. Each team put down eight cards: a blue one for water and seven green ones for land. On the back of one green card, there were eggs and on the back of another one a pitfall. The remaining green cards were blank. Dyads had to discuss in pairs where to hide their eggs and where to put a pitfall. They also had to choose a particular dinosaur as their pawn; all dinosaurs had strengths and weaknesses (e.g.

some were able to cross the water, or to cross the game board diagonally). They played the game several times with different partners. After each round, they were challenged to add or change the rules of the game.

‘Kami 1’. The third task was language-oriented: the groups of four were asked to write a story, based on pictures of a kamishibai (storytelling theatre). They chose one out of three picture sets and were asked to write a story that they were going to narrate to a class of preschool children during the next lesson. They were explicitly told that they could choose whatever language or mix of languages to prepare and tell the story, as the activity was part of a project week around ‘multilingual reading aloud’: For one week, all languages were allowed during storytelling time.

‘Kami 2’. During the performance of task four, we went to visit the preschool classes, to tell the story written during the previous task performance. As story tellers the pupils stood behind the kamishibai and they could read out loud their story written in task three.

‘Vulcano’. Pupils were asked to design a volcano out of clay during task performance five, a construction exercise. Materials were provided to decorate the volcano, but there were not enough for every pupil to work individually, forcing them to cooperate.

‘Bubbles’. In task six, scientific experiments exploring ‘bubbles’ were carried out.

Pupils observed sparkling water, described what happened when raisins were dropped in sparkling water, and studied the effect of baking powder in still water. In the final experiment they had to make the volcano (made in task five) erupt. The pupils had to make notes about their observations and hypotheses, in order to present the experiments to the camera.

‘Measuring trees’. For task seven, a mathematics task, we went outside (on the playground) to measure trees using a triangle made of carton. This task was performed in pairs. First, each pair had to estimate the height of a particular tree.

(29)

Then, a step by step explanation was followed measuring the tree. Finally, the pairs filled out a form.

‘Turkish class’. In task eight, a language awareness task, the groups designed a Turkish as a foreign language class for the researcher. The pupils had to cooperate during the preparation of the Turkish class, to make sure there would not be any overlap between their courses.

0.4.8 Data collection to address pupils’ reflections

Besides the interviews that were conducted after each task performance, three data collection moments were fully devoted to the fourth study (moment 1, 6 and 11 – so at the start, in the middle and at the end of the data collection period, colored grey in Table 0.2). During these three moments, more extended focus groups, assisted surveys and language portrait coloring (Busch 2010; Martin 2012) were conducted in order to elicit additional data concerning the pupils’ reflections on the linguistic realities they live in. Preparation documents for these data collection can be found in Appendix 0.2.

During our first meeting, the pupils colored a ‘language portrait’ (Busch 2010; Martin 2012). They were shown a blank portrait (scheme of a human body) and they were asked to color it, according to how they ‘felt’ the languages they spoke were represented in their body. They were reassured that every language would be valued and that they could also represent languages they had just basic notions of (see Appendix 4.2 and 4.3 for colored examples).

Pupils worked individually. When they were ready, they were invited to present their drawing to the ‘multilingual camera’ in whatever languages they chose. Then, a focus group was set up, during which the pupils presented their language portraits to each other. They commented on each other's presentations and boasted about their skills in particular languages. The second part of the focus group was based on a large paper (A3) format on which fifteen ‘language planets’ were drawn (see Appendix 4.1). Each planet represented a context in which language can be used: subjects at school (e.g.

‘mathematics’ and ‘science’), other contexts like ‘homework’, ‘playing’ and

‘quarreling’. For each context presented on a planet, pupils discussed which languages occurred on that particular planet.

The following school year, while they were in the fifth grade (session 6), the pupils drew a second language portrait and presented it to the camera and in a short focus group again. Also, each pupil filled out a questionnaire individually and this was used as a basis for an extensive interview with me afterwards.

The last session (11) took place half a year after the last task performance. During this final session, all the pupils were interviewed extensively. The topics we discussed

(30)

were based on the analysis of the data obtained during the previous sessions, the observations of their task performances and field notes. This interview was semi- structured.

In addition, I visited the schools twice to gather the quantitative data for Validiv- project colleagues (a survey, a reading comprehension test and a science test). This implied that I spent two whole school days in the class of my participants, which provided a lot of contextual information on the atmosphere in the class, friendships, the ethnic composition of the class and so on. Moreover, I replaced the Validiv coach of the school twice in each school, discussing school language policy and the implementation of Validiv. This way I learned more about the atmosphere at school, the school dynamics, the school policy, and the school teams’ discourses, which enriched the data.

The procedures for data analyses are described in each subsequent chapter, as they are specific for every study.

0.4.9 The structure of the dissertation

This dissertations consists of five chapters. The first four chapters represent four studies, the concluding chapter discusses the main findings and implications of the four studies.

Study 1 is exploratory in nature. It addresses the proportions of the languages used by the participants during the performance of eight tasks and it explores the occurrence of Turkish across task performances. Study 2 focuses on three of the eight tasks (the most typically academic tasks – task 1, 3, and 6) to look into the functions of the turns containing Turkish in more depth. Study 3 considers the same three tasks as study 2, but from a different perspective, a translanguaging perspective which considers all languages used as one medium of communication instead of focusing on Turkish.

It is important to note that the first three studies were framed from two different perspectives. Study 1 and study 2 both focus on the use of Turkish in the pupils’

interactions. As pointed out before, in our multilingual space, all languages (including Turkish) were explicitly welcomed. However, in the analysis of the transcripts, Turkish was approached as the ‘marked choice’ (Myers-Scotton 1983). This was done for two reasons: firstly, the studies are exploratory in nature, as little is known about mother tongue use in Flemish primary schools (especially when it is allowed by exception). Thus a focus on mother tongue use seemed a good starting point to explore multilingual practices. Secondly, in exploratory talks with teachers before the onset of the data collection, they seemed to be very inquisitive about the amount of Turkish pupils would use, and about the functions it would serve. Therefore, studies 1 and 2 are both framed in the literature about L1 use in L2 learning environments,

(31)

and the literature about code-switching in interaction at school. Both studies adopted a perspective that has been called by Bonacina and Gafaranga (2011) (drawing on Gafaranga 2007) ‘an overall order perspective’ in which ‘a more or less explicit distinction is made between the declared medium of instruction and the other language(s) in the classroom. Following this distinction, the medium of instruction is assumed to be the default choice against which the use of the other language(s) is seen as deviance. Thus, in these studies, use of the medium of instruction is seen as unremarkable, while the use of the other language(s) is seen as noticeable and accountable, i.e. as requiring an account’ (323).

This perspective is interesting , as it aligns with the perspective of research context:

As mentioned before, most Flemish schools adopt a Dutch-only policy, with Turkish being considered a deviation from the norm (in this case, Dutch).

While in study 2 and study 3 the data (tasks 1, 3 and 6) was the same, the theoretical and analytical approach was different. Contrary to the ‘overall order perspective’ we adopted in study 2, we we adopted the notion medium of classroom interaction in study 3. The medium of classroom interaction refers to ‘the linguistic code that classroom participants actually orient-to while talking, as opposed to the policy- prescribed medium of instruction’ (Bonacina and Gafaranga 2011, 330). As the medium of classroom interaction can be monolingual or multilingual, this approach can account for stretches of interaction spoken in Dutch, in Turkish, or in a mix of both (ultimately with other languages involved as well). In other words, in study 3, we adopted a ‘translanguaging perspective’, in which we did not focus on the use of Turkish, but instead shifted our attention to how Turkish, Dutch and other ‘socially constructed languages’ (Garcia and Wei 2014) were intertwined in one medium of interaction.

Study 4, aims at giving the pupils a voice. The study explores what they feel it is like to be bilingual Turkish-Dutch in a Dutch-medium school and how they experienced working in the translanguaging space.

The fifth and concluding chapter brings together the main findings of the studies in chapter 1 to 4. This chapter also discusses the implications of the findings and provides an outlook on further research.

(32)

(33)

Toward study 1

Study 1 explores what happens when two groups of four Turkish-Dutch bilingual pupils are exceptionally invited to use ‘whatever languages’ during group work. The pupils performed eight tasks, which were related to different subjects.

Which language choices do the bilingual pupils make across tasks? Are there any differences in the proportions of languages used across task performances and can those differences be explained qualitatively?

‘They won’t use Turkish, they’re not used to it.’ (teacher)

‘Oh no, they will be talking Turkish all the time!’ (teacher)

‘I use more Dutch, because in Turkish I sometimes don’t know the words.’

(Doğucan)

‘In Turkish I can tell more things.’ (Melisa)

(34)

(35)

1 M ULTILINGUAL ISLANDS IN A MONOLINGUAL SEA .

L ANGUAGE CHOICE

PATTERNS DURING GROUP WORK

An adapted version of this chapter is under revision for publication in the International Journal of Bilingual Education and Bilingualism.

(36)

ABSTRACT

This study explores language choice patterns during group work of bilingual pupils (Turkish – Dutch) in two mainstream primary schools in Flanders (Belgium). In each school, a group of four children performed eight different tasks, related to different subjects of the curriculum. During task performance, they were exceptionally invited to use their mother tongue (Turkish), which ran against the rules of the monolingual school. Language choice data in these temporary ‘multilingual islands in a monolingual sea’ were gathered. Results indicate that the proportion of Turkish differed between tasks. However, results show that Turkish was used for purposes related to task performance throughout all tasks. The results challenge the view, currently prevalent in Flemish education, that a submersion approach is the only viable option for bilingual pupils speaking a minority language.

1.1 Introduction

The achievement gap between ethnic minority pupils and pupils with only Dutch as a mother tongue in Flanders (the northern part of Belgium) is one of the greatest among Western European countries (OECD 2006). Many Flemish teachers have been found to routinely associate mother tongues other than Dutch (the main medium of instruction) with ethnic minority pupils’ underachievement, particularly when the mother tongue is Turkish or Berber (Hirtt, Nicaise, and De Zutter 2007; Ağirdag, Van Avermaet, and Van Houtte 2013). To enhance pupils’ development of Dutch, the vast majority of schools adopt a Dutch-only policy, prohibiting the use of other mother tongues at school (Jaspaert and Ramaut 2000; Spotti 2008; Ağirdag, Jordens, and Van Houtte 2014). Although this monolingual approach has become strictly endorsed in the vast majority of Flemish schools during the past decades, the social achievement gap has not diminished (OECD 2009). Still, only a minority of the school teams have reconsidered their current monolingual policy and explored alternative approaches in which mother tongues other than Dutch are used as a resource for learning.

This monolingual approach sharply contrasts with current trends in the literature on multilingualism in the classroom, which emphasize the advantages of validating all languages in class (Cummins 2000; Baker 2011). Using multiple languages in education is claimed to be a ‘communicative and pedagogic resource in bilingual contexts, especially where pupils struggle to understand difficult subject matter whilst simultaneously learning a foreign language, one that is nominally the official medium of instruction’ (Ferguson 2009, 231).

This article reports on an explorative case study that was conducted in two Flemish primary schools with a strict Dutch-only policy. Bilingual children were invited to use their mother tongue (Turkish) while performing different types of tasks during group work during regular class hours, but while being guided by an external teacher (the

(37)

author of this dissertation) who did not belong to the regular staff of the school. The aim of the study was to shed light on the children’s language choice patterns during the task performances in this newly created circumstances. The study reports on the comparison between different types of tasks as far as the children’s language choice patterns are concerned quantitatively (the proportions of languages used), and qualitatively (exploration of language choice patterns).

1.2 Literature

Not only in Flanders, but also in many other regions with multilingual school populations, the use of mother tongues other than the official medium of instruction at school is a contested issue. Numerous studies describe a monolingual norm being applied in schools, in mainstream education where only one language is officially designated as the medium of instruction, but also in bilingual education systems or immersion contexts (e.g. Hornberger 2004; Cruz-Ferreira 2010; García 2009 in the US and Canada; Ncoko, Osman, and Cockcroft 2000; Ferguson 2003; Moodley 2007 in Africa; Gogolin 1997; Jaspaert and Ramaut 2000; Kroon and Sturm 2000; Creese and Martin 2003; Duarte 2011 in Europe).

However, despite the hegemony of monolingualism in many educational systems around the world, ethnographic studies show that multilingualism is hard to suppress in education. For one, pupils tend to be very creative in circumventing official language regulations and secretly using their mother tongues at school, also when the language policy of the school explicitly forbids them to do so (e.g. Jaspers 2008;

Probyn 2009).

Individual teachers have been shown to react in different ways to pupils’ subversive behavior. While monolingual beliefs are widespread amongst teachers (Dooly 2007;

Young 2014; Pulinx, Avermaet, and Ağirdag 2015), in practice, teachers often turn a blind eye to pupils’ mother tongue use (Asker and Martin-Jones 2013, Van Praag, Stevens, and Van Houtte 2015; Van Der Wildt, Van Avermaet, and Van Houtte 2015a) or, as Jaspers (2016) remarks, mother tongue use is tolerated in the margins of official business at school.

Still, the prescribed medium of instruction in many schools is the majority language, and the classroom and school interaction are monolingual to maximize the opportunities learners get to acquire the target language.

This kind of educational practice strongly contrasts with the theoretical and research literature on dual-language teaching, two-way bilingual learning, additive language programmes or other forms of bilingual support in which the use of the mother tongue as a valuable resource for learning is stressed (for example Martin-Jones and Saxena 2003; Creese 2004; Duarte 2011 and many others). Recently, a meta-analysis

(38)

on the effectiveness of bilingual programs in Europe reported a small positive effect for bilingual over submersion programs on the academic achievement of language minority children, when their mother tongue was incorporated in regular instruction (Reljić, Ferring, and Martin 2015).

Indeed, as Bonacina and Gafaranga (2011) argue, both the medium of instruction and the medium of interaction can be monolingual or multilingual, especially in the case of pupils with shared mother tongues. Overall, studies on language choice patterns during collaborative group work show that learners draw on their full linguistic resources, including their mother tongue. This is the case in the context of immersion programs and bilingual programs where the (minority) mother tongue of pupils is one of the languages involved, as well as in mainstream education contexts where the medium of instruction differs from the pupils’ mother tongue.

1.2.1 Proportions of mother tongue use

Proportions of language use in several contexts tend to differ across educational systems, and different combinations of languages. Swain and Lapkin (2000) report a proportion of 29% of turns produced in the mother tongue during a jigsaw task and 31% during a dictogloss task, performed by dyads in 8th grade immersion classes. In a two-way immersion program for six-year-olds in the US, Gort (2012) compared writing workshops (WW) in Spanish (native language of most of the pupils) and English, the dominant language in the broader social context. During the Spanish WW, pupils used more code-switching (85% of total code-switches of which 47% to English only and 38% mixed language) than during the English WW (15% of which 2% to Spanish only and 13% mixed language). Also in an immersion context in the US, Reyes (2004) compared the distribution of Spanish and English spoken by 7-year-olds and 10-year-olds in formal (science task) and informal (lunch conversation) situations. She found that the proportion of languages used by the 10-year-olds was similar in both contexts: respectively 67 and 68% Spanish, 15 and 17% English, and 16 and 17% both languages in one turn.

In a mainstream South-African primary school where the language of teaching was English, Setati et al. (2002) compared the use of code-switching (to the mother tongue) by teachers and learners in mathematics, science and English classes. Overall, they observed an increase in the use of code-switching from primary to secondary school, which was related to the increase of group work in teacher practices. They report that ‘the overall pattern of increased use of code-switching conceals important differences across subjects, across levels, and across regions’ (142). In Europe, similar studies were conducted. Planas and Setati (2009) looked at language use of 12-year- olds during mathematics in a Catalan-medium school, where the use of the first language (Spanish) was encouraged. A group of four pupils used Catalan between 5.3% and 15.7% of the time and Spanish between 94.7% and 84.3%. However, in

(39)

whole class interaction, they resorted to Catalan, or remained silent.

In the context of the ‘Køge project’ in Denmark (Jørgensen 2003) several studies were carried out to explore language choice patterns of bilingual Danish-Turkish pupils in Danish mainstream education. Within this project, Esdahl (2003) focused on language patterns of grade 9 pupils performing creative tasks and found 40% Danish-based and 47% Turkish-based utterances and 11% with an intra-sentential code-switch (typically between Danish and Turkish). A similar project was conducted in Ghent (Flanders), encouraging the use of Turkish in primary school. Here, Rosiers (2015) found that bilingual Turkish-Dutch pupils used Turkish in class in 40% of the turns spoken, most of which for on-task purposes.

The common thread running through these studies is that bilingual pupils make extensive use of their mother tongue during collaborative task performance at school when allowed to do so. Code-switching seems to occur more often between turns than within turns. Different proportions of turns in the mother tongues are reported depending on context and subject. Most of these studies were conducted in contexts where the school personnel was able to speak these languages themselves. Notable exceptions are the Køge project and the mother tongue project in Ghent, but in these cases there was a great openness toward mother tongues.

1.2.2 Mother tongue as a resource

In addition, the above-mentioned studies into the use of code-switching during task performance show that the multilingual medium of interaction is used as a resource for task performance. This seems to be the case in language learning tasks, science and mathematics tasks, and more informal tasks (like crafting tasks).

Specifically in the context of language learning, a vast amount of studies report on L1- use as a cognitive resource for L2 task performance (e.g. Swain and Lapkin 2000; Gort 2012). In mainstream education, Eldridge (1996) found that in English courses for Turkish pupils ‘the majority of code-switching in the classroom is highly purposeful, and related to pedagogical goals’ (303). Moreover, case studies into pupils’

performance of science and mathematics tasks report on the ways in which pupils draw on their native language to negotiate meaning and clarifying complex subject matter (e.g. Moskovitch 2007; Planas and Setati 2009). In an immersion context, Reyes (2004) found that children use their native language (Spanish) to negotiate conversational involvement while seeking explanations during science activities (in English). In an ethnographical study, Moodley (2007) reported the strategic use of code-switching by learners fulfilled both social and pedagogical functions, though code-switching to the mother tongue was prohibited. Furthermore, some studies describe pupils’ mother tongue use during the performance of more informal tasks.

For example, in the ‘Køge project’ in Denmark (Jørgensen 2003) several studies were carried out to explore the language choice patterns of bilingual Danish-Turkish pupils

TURKISH IS NOT FOR LEARNING, MISS.