Code-Switching in the Upper Secondary School EFL Classroom in Sweden

Degree Project with Specialization in English Studies in

Education

15 Credits, Second Cycle

Code-Switching in the Upper Secondary

School EFL Classroom in Sweden

Kodväxling i gymnasieskolans engelskklassrum i Sverige

Moa Torvaldsdotter

Master of Arts/Science in Education, 300 credits English Studies in Education

7 June 2020

Examiner: Chrys Malilang Supervisor: Shaun Nolan FACULTY OF EDUCATION AND SOCIETY

Abstract

Code-switching has been shown to be beneficial for students’ language learning and for strengthening their identities. Despite this, it can be interpreted that code-switching is not encouraged in the syllabus for English in upper secondary school in Sweden. Because of this potential disagreement, this study aims to broaden the knowledge of how upper secondary school teachers relate to code-switching in their different classrooms. Thereby, this study seeks to examine some upper secondary school EFL teachers’ understanding of code-switching as well as the use of code-switching in their different classrooms. In this qualitative study, four upper secondary school teachers of English participated in semi-structured interviews followed by classroom observations. The teachers represent all courses of English at upper secondary school level and they represent schools with different programs and students with different first languages. The results show that the teachers have limited knowledge of code-switching and that they believe that a large amount of target language use in the classroom is favorable. Nevertheless, the results also show that the teachers as well as their students use code-switching both intentionally and unintentionally for various purposes, but none of the participating teachers seem to use code-switching as a strategy to promote long-standing language acquisition. However, the results indicate that the teachers seem to effectively adapt their code-switching behavior to their student group. The findings of this study are in line with earlier research where it, for instance, is argued that code-switching is more suitable for classes with students with lower language proficiency.

Keywords: Code-switching; English language teaching; Multilingual classrooms; Upper secondary education

Table of Contents

1. Introduction ... 4 2. Aim and Research Questions ... 6 2.1. Research Questions ... 6 3. Background: Theory, Policy Documents, and Recent Research ... 7 3.1. Monolingual and Multilingual Education ... 7 3.2. Code-Switching ... 8 3.2.1. Code-Switching in Foreign Language Education ... 9 3.2.2. Difference Between Code-Switching and Translanguaging ... 11 3.2.3. The Place of Code-Switching in the Steering Documents ... 11 3.3. Recent Research on Code-Switching in the Classroom ... 12 4. Method ... 15 4.1. The Participants ... 15 4.2. Materials and Procedure ... 16 4.2.1. Semi-Structured Interviews ... 17 4.2.2. Observations ... 18 4.3. Analysis ... 19 4.4. Ethical Considerations ... 19 5. Results and Discussion ... 20 5.1. EFL Teachers’ Awareness of Code-Switching ... 20 5.2. EFL Teachers’ and Their Students’ Usage of Code-Switching ... 24 5.2.1. Intentional Code-Switching ... 24 5.2.1.1. Focalization on Form ... 24 5.2.1.2. Communication Effectiveness ... 25 5.2.1.3. Relational Purposes ... 28 5.2.1.4. Convenience ... 29 5.2.2. Unintentional Code-Switching ... 30 5.3. Difference in the Use of Code-Switching Depending on the Students’ Language Level ... 31 6. Conclusion ... 34 7. References ... 36 8. Appendix ... 39

1. Introduction

Up until recently, the generally accepted approach to foreign and second language (henceforth FL and L2) education was the monolingual approach, which advocates the target language (TL) being the only language used in the classroom (García & Wei, 2013; Lundahl, 2019; Cummins, 2007). Within this approach, the use of students’ first language (L1) is often seen as a failure and a sign of language deficiency (Cummins, 2007; Liebscher & Dailey-O’Cain, 2005). However, in an article by Cummins (2007) it is stated that there is minimal pedagogical evidence supporting the exclusive use of the TL in FL teaching. And as the world has become more globalized, the monolingual approach has been questioned (García & Wei, 2013). Many researchers now perceive students’ L1 as a resource in both language teaching and learning (Lundahl, 2019; Cummins, 2007). Furthermore, it has been shown that code-switching, the switch between two or more languages within a sentence or conversation, is a natural phenomenon in bilingual speakers’ language use (Poplak, 1980). Using code-switching as an educational approach in the FL classroom is to encourage students to use their whole language repertoire, and it has not only been shown to be beneficial for students’ language learning (Moore, 2002; Liebscher & Dailey-O’Cain, 2005) but also for strengthening their identities (McKay, 2012).

Although much research on code-switching has been done internationally, there is a lack of studies regarding code-switching in English language education in a Swedish context. Therefore, it is difficult to assess the effectiveness of code-switching in this context. Lundahl (2019), however, states that the use of code-switching is dependent on the students’ language level and, therefore, it might not be a beneficial approach in the English language education in countries like Sweden where the students’ linguistic levels are high. This view has also been prevalent within the courses of English in my teaching training, where it has been conveyed that one should only use the TL in the English classroom at upper secondary level. Furthermore, it can be interpreted that code-switching is not encouraged in the syllabus for English in upper secondary school in Sweden as it is stated that “Teaching should as far as possible be conducted in English” (Skolverket, 2011a). However, in the commentary material for the subject English provided by Skolverket (2011b), the National Agency for Education, it is

expressed that on occasion, elements of other languages may occur in the classroom but that it is dependent on the teachers’ professional judgment of the students’ need for support. This clarification is in line with Lundahl (2019) highlighting that code-switching could be used in classrooms with students with a lower linguistic language. In contrast to the negative view of code-switching in the courses of English in my teacher training, it has been viewed more positively within the courses of Swedish as an L2 in the same teacher training. One possible explanation for this is that the syllabus for Swedish as a second language (Skolverket, 2011c) emphasizes the importance of multilingualism more explicitly, and it is, for example, stated that multilingualism is an asset both for individuals and for the society. As such, there seems to be a disagreement in the attitude towards code-switching within the subjects English and Swedish as an L2, even though they are both being taught as an L2 or FL in Swedish upper secondary schools. Considering the above, and the fact that FL classrooms are multilingual as such, there is a reason to examine the approach to code-switching in English in upper secondary school.

2. Aim and Research Questions

The aim of this study is to broaden the knowledge of how upper secondary school teachers of English as a foreign language (EFL) in Sweden relate to code-switching in their different classrooms. Thereby, this study seeks to examine some upper secondary school EFL teachers’ understanding of switching as well as the use of code-switching in their different classrooms.

2.1.

Research Questions

1. What is EFL teachers’ awareness of code-switching?

2. What are EFL teachers’ and their students’ usage of code-switching in the classroom?

3. How does the use of code-switching differ depending on the students’ language level?

3. Background:

Theory,

Policy

Documents, and Recent Research

In this section, the theoretical concepts underpinning this study will be unpacked and explained, namely monolingual and multilingual education. Thereafter, the key term code-switching will be defined, unpacked and discussed in the context of the curriculum documents. Lastly, relevant research on code-switching in EFL education will be presented.

3.1.

Monolingual and Multilingual Education

As mentioned above, the monolingual principle has long prevailed in EFL teaching and has been called the English-only discourse (Cummins, 2007; Liebscher & Dailey-O’Cain, 2005; McKay, 2012). Cummins (2007) explains that this principle “emphasizes instructional use of the target language (TL) to the exclusion of students’ L1, with the goal of enabling learners to think in the TL with minimal interference from L1.” (p.223). As it was believed that the languages interfered with each other, Liebscher and Dailey-O’Cain (2005) explain that the monolingual principle, therefore, is based on the idea of “compartmentalizing languages in learners’ minds in a kind of coordinate bilingualism” (p.235). This principle evolved within the notion of learning an L2 through the direct method, which imitates the way a first language is acquired, meaning that the TL is used both as medium and means of the language learning (Cummins, 2007). Support for this approach can be found, as greater use of the TL results in higher achievement in the TL (Cummins, 2007). Therefore, in the context of FL teaching, the use of students’ L1 is often seen as a failure by researchers and teachers (Cummins, 2007).

However, Cummins (2007) argues that there is minimal pedagogical evidence supporting the exclusive use of English in ESL-education and that there is “no empirical justification for any absolute exclusion of students’ L1 from TL instruction” (p.227). As a result of this, it has been a shift in FL research over the years to an approach where the students’ L1 is viewed as a resource and that the goal of language education is to develop bilingual and bicultural identities and skills (Lundahl, 2019). Furthermore,

Lundahl (2019) states that if students’ language ability is limited, it is natural that they use all the language resources they have. Similar claims are made by Cummins (2007), stating that conscious usage of students’ L1 can function as scaffolding to higher language achievement. However, EFL education in a country like Sweden does not fully apply to this approach, as the students’ language proficiency is high (Lundahl, 2019). Lundahl further argues that in this context it is beneficial to retain the English-only principle, as this will help the students develop their language more. Therefore, he concludes that the multilingual approach should be weighed against the importance of using the TL.

Even though Swedish students’ language proficiency in English is high, Lundahl (2019) states that Swedish often is used as a reference when teaching English in Sweden. This could be problematic as Swedish classrooms consist of many students with other L1s than Swedish, and therefore, might be disfavored by such an approach. As such, the teachers’ intended scaffolding might become an additional barrier for these students. In multilingual classrooms where the teacher does not know the students’ L1s, the teacher could instead, for instance, create opportunities for students with the same L1 to support each other with translations (Allard, Apt & Sacks, 2019).

3.2.

Code-Switching

One form of multilingual language use is code-switching, which is defined as “the ability of plurilingual speakers to switch within or between sentences from and to the codes in their repertoire, in order to fulfill communication needs triggered by decisions concerning the communicative context in which they are immersed.” (Corocoll López & González-Davies, 2016, p.69). This alternation occurs naturally in bilingual communities where it requires a large degree of linguistic competence (Poplak, 1980; Lightbown & Spada, 2013). That is, code-switching can be used in conversations where all interlocutors share the same language repertoire (Lightbown & Spada, 2013), which is often seen in bilingual settings such as Spanish-speaking communities in the United States (Poplak, 1980). Code-switching can, in these settings, be used for various purposes, for example to express solidarity, making a joke, or to signal personal identity (Lightbown & Spada, 2013; McKay, 2012). The identity aspect is well discussed in

research on multilingualism where it is argued that multilingualism is a fundamental part of the creation of an identity (Auer, 1988; García & Wei, 2013; McKay, 2011). Despite code-switching being a natural language phenomenon, and requires a large degree of linguistic competence, it could also be a sign of insufficient language proficiency. For example, a code-switching behavior might be the result of insufficient knowledge in the L2 such as a lack of relevant vocabulary (Liebscher & Dailey-O’Cain; 2005; Lightbown & Spada 2013), or a way to avoid expressing oneself in the TL (Lundahl, 2019). It might also be the result of speakers being too lazy to search for the appropriate word (Gardner-Chloros, 2009). Because of this, code-switching has for a long time been a stigmatized practice (García & Wei, 2013).

The concept code-switching, as presented in research (see, e.g. Auer, 1988; Corcoll López & González-Davies, 2016), seem to exclusively cover oral language and thereby excludes writing. Therefore, in conformity with previous research, this study defines code-switching as an oral phenomenon. Moreover, in research, the different codes within the concept of code-switching seem to mainly refer to recognized languages, and therefore, this is also the definition of codes used in this study.

3.2.1. Code-Switching in Foreign Language Education

There are several purposes for using code-switching in the FL classroom. One purpose is focalization on form (Moore, 2002), where the teacher, for example, chooses to use the L1 when explaining certain unknown words or grammatical aspects in the TL (Lundahl, 2019). Bensen and Çavtşoğlu (2013) show that teachers switch languages in these situations for the purpose of clarifying meaning. This has also been shown to be done by the teachers with the purpose of saving time in their teaching (Bensen & Çavtşoğlu, 2013). That is, a second purpose of code-switching is communication effectiveness (Moore, 2002) where it can function as a support for reading, listening, or conversational comprehension (Lundahl, 2019). One example of such code-switching can be found in teacher-student communication where the focus is on content, for instance after reading a text. This can result in the teacher asking a question in the TL, and the students are allowed to answer in the L1 to not break the flow of the conversation in order to ensure that the students have understood the content. Moore

(2002) states that in situations where the focus is on both content and language, code-switching can be especially beneficial as it, for example, can enrich new concepts. As shown, code-switching can be used both by teachers and students in educational settings to facilitate learning. One example of how code-switching can be used to benefit the language learning process is through the learning strategy Pedagogically Based Code-Switching (PBCS) (Corocoll López & González-Davies, 2016). PBCS is described as a strategy designed by the teacher aiming to create informed ways for students to move between languages. This strategy is beneficial, Corocoll López and González-Davies (2016) argue, since the learning of a language does not only involve using the language but also to reflect on language use. Therefore, Corocoll López and González-Davies (2016) highlight that code-switching should be used in FL learning as it can facilitate the recognition of different language aspects.

McKay (2012) argues that it is beneficial to encourage code-switching in EFL classrooms as it gives equal status to the students’ languages and allows students to use their L1 to develop their English proficiency. Further, Liebscher and Dailey-O’Cain (2005) found that students use code-switching both for participant-related functions, such as adapting the choice of language to the language used by the individuals in the conversation, and also for discourse-related functions, such as repeating one’s utterance in another language if it was not being replied to in the first language. These examples could be seen as informed strategies to support the communication, which corresponds to one of the goals for the subject of English: “The ability to use different language strategies in different contexts” (Skolverket, 2011a, p.2). Despite this, Moore (2002) states that many L2 teachers still believe the notion of avoiding the use of L1 in L2 classes as much as possible. Yet, code-switching occurs in the classrooms, especially in those consisting of students with a low linguistic level (Moore, 2002). It is also in these classrooms that teacher code-switching has been found to be most beneficial for students’ language development (Lee & Macaro, 2013). Lee and Macaro (2013) show that even though both young and adult learners benefit from code-switching for vocabulary acquisition, young learners’ gains were shown to be higher and also led to higher vocabulary retention. For the adult groups, the result of vocabulary retention did not differ between a code-switching instruction and an English-only instruction.

3.2.2. Difference Between Code-Switching and Translanguaging

Lundahl (2019) highlights that since code-switching has negative connotations, many researchers use the term translanguaging to refer to various approaches of using the students’ language repertoire to facilitate learning. However, code-switching and translanguaging are not synonymous. Translanguaging is defined by Canagarajah (2011) as “the ability of multilingual speakers to shuttle between languages, treating the diverse languages that form their repertoire as an integrated system” (p.401). Translanguaging, García and Wei (2013) explain, differs from code-switching in that it refers to “the speakers’ construction and use of original and complex interrelated discursive practices that cannot be easily assigned to one or another traditional definition of a language, but that makes up the speakers’ complete language repertoire.” (p.22), whereas code-switching refers to a shift between two or more recognized languages. That is, code-switching is a learning and communicative strategy that is situated within the framework of translanguaging (Corcoll López & González-Davies, 2015). In this thesis, it is, therefore, relevant to use articles on translanguaging where the code-switching aspect is transparent.

3.2.3. The Place of Code-Switching in the Steering Documents

In the Common European Framework of Reference for Languages (CEFR) (Council of Europe, 2001), a plurilingual approach to language teaching is promoted. In this approach, the knowledge of more than one language is not understood as different monolingual competences where the languages are strictly separated in the brain. Instead, plurilingualism is explained as the notion of all the languages an individual knows to be interrelated and together build up “a communicative competence to which all knowledge and experience of language contributes” (p.4). A plurilingual competence, they state, therefore allows alternations of the languages one knows where a person can code switch during an utterance to achieve effective communication. A plurilingual approach enables students to develop enriched identities and increases the ability for further language learning (Council of Europe, 2001).

This positive attitude towards plurilingualism can also be found in the syllabus for the subject English in upper secondary school (Skolverket, 2011a), where it is stated in the

aim of the subject: “Teaching should encourage students' curiosity in language and culture, and allow them to develop plurilingualism where skills in different languages interact and support each other.” (p.1). Despite this promotion of plurilingualism, the syllabus does not give more support to this approach. Instead, it is emphasized that “Teaching should as far as possible be conducted in English.” (Skolverket, 2011a, p.1) which can be interpreted as being somewhat in disagreement with the previous statement. In the commentary material for the subject English (Skolverket, 2011b), some reasons for the English-only approach are presented. These include increased opportunities for input in the TL and that it will encourage students to use the TL themselves. Nevertheless, it is also highlighted that occasional elements of Swedish or students’ L1 can occur in the teaching, if the teacher considers it to contribute to increased understanding.

3.3.

Recent Research on Code-Switching in the

Classroom

Recent research indicates that teachers’ attitudes towards code-switching in the EFL classroom are predominantly positive since they believe it can be useful for various purposes (Allard et al., 2019; Lo, 2015; Simasiku, Kasanda & Smit, 2015; Yuvayapan, 2019; Zhu & Vanek, 2017). For example, it has been shown that teachers believe code-switching to be useful for making students more active in the classroom, making students more actively involved in their learning, giving access to content and ideas, helping low proficiency students, and building bonds with the students (Allard et al., 2019; Simasiku et al. 2015; Zhu & Vanek, 2017; Yuvayapan, 2019). Further, Allard et al. (2019) state that the use of code-switching made the teacher avoid the “infantilizing nature of the English-only classroom” (p.83), meaning that instead of having the students rely on the teacher’s explanation each time they did not understand an utterance in English, they could be more autonomous in their language learning by, for example, translate and confer with their peers in the L1. Moreover, it has also been shown that teachers use code-switching to provide translation, repetition, and to deliver content (Lo, 2015; Zhu & Vanek, 2017). However, some teachers instead prefer a TL-only principle to strengthen their students’ linguistic competence (Simasiku et al., 2015). Similarly, Yuvayapan (2019) found that many teachers avoid using the L1 in the classes, as it interferes with their notion of helping the students achieve a native-like

competence in English. Yuvayapan (2019), further, discusses that this monolingual teaching practice could be the result of factors such as the formulations in the curriculum and the expectations from colleagues or institutions.

In the current research, students’ beliefs of code-switching in the classroom differ. Kocaman and Aslan (2018) show that the majority of the students in their study believed that some usage of L1 in the EFL classroom could be beneficial for instance when explaining grammar, new words as well as differences between the L1 and L2. As the students in this study expressed the necessity of the use of L1 in the classroom, Kocaman and Aslan (2018) conclude that a bilingual approach should be taken, as it can be good for maintaining good relations with the students. However, Allard et al. (2019) found that some students were unsatisfied with their English development due to code-switching in the classroom, and in Simasiku et al. (2015), teachers believed that their students prefer to speak the TL in the classroom.

Despite the varying attitudes towards code-switching in the classroom, its use has shown different results. For example, Allard et al. (2019) show that code-switching is more beneficial when the students use it themselves rather than their teachers using it, and that code-switching is not as beneficial for the students who do not share the same L1 as the majority of the class. However, it has been found that code-switching increases students' activity in the language classroom (Lo, 2015; Rahayu & Margana, 2018; Zhu & Vanek, 2017). Zhu and Vanek (2017) show that the use of code-switching in the classroom enhances the level of student engagement and student talk. Similarly, Lo (2015) shows that classes that adopted a code-switching pedagogy were more interactive and less teacher-centered than classes with a TL-only principle. Rahayu and Margana (2018) also found that code-switching made students more active, but they highlight that this occurs negatively. That is, the students were more talkative but did not speak English. In the compared classrooms, which applied a TL-only principle, the students made more effort to pronounce words correctly as well as to comprehend the meaning of the teachers’ utterances. In these classrooms it was found that the students were more active in using the TL than the students taught through code-switching-based instruction who were instead more comfortable using their L1.

As mentioned in section 3.1 and 3.2.1, it is argued that the use of code-switching should be adapted to the student group as its use differs depending on the students’ language level. This is also found in current research showing that the teacher makes use of students’ L1 more in classes where students' TL proficiency level is low (Lo, 2015). In these classes, Lo found, the teacher mainly used the L1 and code-switched to English mostly to teach English academic language or, for instance, to provide a translation of a student’s L1 utterance into the L2. In the classes with students of higher language ability, the teachers mainly used English, but occasionally made use of the L1 to help deliver the content of the subjects to students. Moreover, in these classes, some teachers switched to L1 for explaining difficult or abstract concepts, whereas others instead elaborated more in the TL.

Even though it has been shown that some teachers adapt the use of code-switching to the students’ language ability, research argues for a more conscious code-switching pedagogy for it to benefit the students (Allard et al., 2019; Yuvayapan, 2019; Zhu & Vanek, 2017). Yuvayapan (2019) states that the teachers’ usage of the L1 in class did not seem to be a conscious strategy to promote a long-standing language acquisition, but instead, it functioned as a tool to reduce the time for clarifications in the classroom. Similar findings are presented in the study by Zhu and Vanek (2017), and they, therefore, argue that code-switching should be used more efficiently as careless use of the L1 gives students fewer opportunities to interact in the TL. Additionally, Allard et al. (2019) also highlight the importance of hearing and using the TL in meaningful ways in order to learn it. Zhu and Vanek (2017), therefore, suggest that code-switching could be optimized by the teacher and the students together deciding on norms about language use in the classroom.

4. Method

To achieve the purpose of this study, which is to examine upper secondary school teachers’ understanding and use of code-switching in Swedish EFL classrooms, a qualitative approach has been adapted. Two sets of data were collected to increase validity, namely: semi-structured interviews with upper secondary school teachers of English, and observations of digital video-based English lessons. The following section describes and discusses the chosen methods, their implementation, the participants and ends with a presentation of the ethical considerations.

4.1.

The Participants

Four teachers of EFL who work at different upper secondary schools in southern Sweden participated in this project. For the participants to remain anonymous they have been given the aliases such as T1 and T2. Table 1 below provides information about the teachers’ age, their working experience, and the English courses they currently teach. The teachers whose classes were observed were T2, T3 and T4.

The participants were selected through both convenience sampling and purposive sampling. Due to the time limitation attributed to this project and the prevailing Corona pandemic that has drastically changed teachers’ workdays, teachers who were already known to the author were directly contacted to increase the likelihood of them participating in the interviews. Nevertheless, Bryman (2016) states that convenience sampling could be problematic as the participants might not be representative of teachers as a whole and that it, therefore, is not possible to generalize the findings. Therefore, a purposive sampling were also applied to ensure that “the sampling is conducted with reference to the research questions” (Bryman, 2016, p.410), which in in this study’s case implies that the interviewed teachers are representing all courses of English at upper secondary school level, that is English 5, English 6 and English 7. Furthermore, it implies that the teachers represent schools with both university preparatory and vocational programs, and also schools where the majority are Swedish L1 speakers and schools where most students have a different L1 than Swedish.

It is also worth noting that T3 is not qualified to teach English. Nevertheless, T3 teaches two courses of English, and therefore, it was motivated to include this participant as T3’s English teaching activities reflect the reality in Swedish upper secondary schools (Skolverket, 2019).

Table 1

Participating Teachers in This Study

Teacher Age Teaching Diploma

Subjects

English Courses They Currently Teach Work Experience Date of the Interview T1 28 History & English English 6 & English 7 2 years 10 April 2020 T2 28 English & Religion English Elementary Level, English 5 &

English 6 1,5 years 14 April 2020 T3 31 Swedish as a L2 & French English 5 & English 6 1,5 years 16 April 2020 T4 30 Swedish & English English 5 & English 6 1 year 17 April 2020

4.2.

Materials and Procedure

As previously mentioned, a qualitative approach was adopted to address this study’s research questions. Since the research questions focus on teachers’ understanding and use of code-switching, the main empirical data is based on the interviews with the four teachers. To increase the validity and reliability of the interpretation of the teachers’ assertions, observations of three of the teachers’ lessons were also made.

4.2.1. Semi-Structured Interviews

To investigate teachers’ awareness and usage of code-switching, the main method for this study is semi-structured interviews. Bryman (2016) states that semi-structured interviews are commonly used when the investigation has a fairly clear focus and when the interest lies in the interviewee’s point of view, which is the case of this study. The interviews were conducted by using an interview guide (see Appendix 1), in which questions to be covered were formulated. The questions are inspired by some of the questions listed in Yuvavapan’s (2019) article, but modified to fit the purpose of this study. During the interview, the order of the questions was then customized to the answers given by the interviewee. Furthermore, follow-up questions were asked to gain rich and detailed answers (Bryman, 2016).

To facilitate the carrying out of the interviews during the Spring 2020 Corona pandemic, the interviews were held via the online video conferencing service Zoom through Malmö University’s server. To manage Zoom’s security issues which became apparent during this time, the following actions were taken: the invitations were e-mailed to the participants with a personal note and it was made sure no one but the interviewer and the interviewee entered the meeting. The video interviews were, furthermore, recorded since, as Bryman (2016) states, this helps the interviewer pay attention to what the interviewee answers and thereby ensures that no utterances go unnoticed and that the right post-interview follow-up questions can be asked if needed. Since Zoom does not allow recording audio only, video was also recorded. However, the participants were informed that only the audio was going to be used in this study, and therefore, the video files were deleted directly after the interviews.

The interviews took between 15-30 minutes and were carried out in Swedish, as this is the participants’ first language and could perhaps therefore result in more detailed answers. Shortly after the interviews were conducted they were transcribed. Kvale (2007) states that the choice of transcription style should be chosen based on the purpose of the research. Since the main interest in this study is on what the interviewees said, and not on how they said it, the interviews were transformed into a written style but it was chosen to include pauses and oral emotional expressions such as laughter, as this could give clues to the interviewees’ attitudes.

4.2.2. Observations

Since interviews only provide access to the reported behavior of the participants, it is not certain that this behavior is accurate (Bryman, 2016). Therefore, to examine teachers’ actual use of code-switching in a classroom setting, participant observations with three of the four interviewed teachers were carried out, as not all wanted to participate. Table 2 below provides some information about the observed lessons. Table 2

Observed Lessons

Teacher English Course Date of the Observation

T2 English 6 30 April 2020

T3 English 5 22April 2020

T4 English 5 28 April 2020

At the time of the observations, the participants’ schools had implemented distance learning due to the prevailing Corona pandemic and the three observed lessons, therefore, were online video-based lessons carried out on Google Meets. The teachers had both their video and microphone turned on, whereas their students’ turned their microphones on only when speaking. To not interfere with the lesson, the author’s camera and microphone were turned off and the author remained passive throughout the lessons. The observations were carried out a few days after the interviews were conducted, which means that the teachers were aware of the focus of the observation. This, together with the author’s presence in the online classroom, could have had an impact on the behavior of both the teacher and the students.

The observations were carried out in terms of incidents, which Bryman (2016) explains is to wait for something to happen and then record what follows from it. Furthermore,

Bryman states that when the focus of the observation is on a specific question, the observation itself should be oriented to that research focus. As the focus of the observation was on the possibly occurring code-switching in the classroom, the author solely concentrated on what type of code-switching that occurred, when it occurred and with what functions. During the observation, detailed field notes were taken to remember what happened (Bryman, 2016).

4.3.

Analysis

After the data collection, the material was analyzed through a qualitative data analysis. While reading through the material, meaningful units were identified and categorized based on this study’s research questions. The first category regards teachers’ awareness of code-switching and includes the teachers expressed knowledge, or lack of knowledge, about switching. The second category regards the habits of code-switching in which several sub-categories were found. These sub-categories were partially inspired by the categories found in Bensen and Çavtşoğlu (2013) (see section 3.2.1). Lastly, the final category includes to what extent teachers adapt their code-switching behavior to their student group. As the interviews were conducted in Swedish, all quotations are translated into English. The quotations and their original transcription in Swedish can be found in Appendix B.

4.4.

Ethical Considerations

In this study, the four main ethical guidelines stated by Vetenskapsrådet (2002) have been applied in the design, the implementation and the synthesis of the interviews and the observations. All the participants were informed of the purpose of the study and that the participation was voluntary (the information requirement). The participants’ consent was acquired through signed consent forms where the participants were informed that they could cancel or revoke their participation at any time (the consent requirement), they have been disidentified and they have received information about who will be able to read this paper and that the collected material will be stored at Malmö University’s server (the confidentiality requirement). Lastly, the participants have been informed that the material will not be used for anything else but this paper and will be deleted when this thesis has been examined (the requirement for usage).

5. Results and Discussion

In this section, the results are presented according to the research questions of this study. The first question regards EFL teachers’ awareness of code-switching, and the second question regards EFL teachers’ and their students’ usage of code-switching in the classroom. This question is, further, divided into two overall themes named intentional and unintentional usage. The first theme, intentional usage, is then divided into subsections that correspond to the different purposes of code-switching. Lastly, results related to the third research question, namely how the use of code-switching differ depending on the students’ language level, will be presented. These results are subject to analytical discussion throughout this section.

5.1.

EFL Teachers’ Awareness of Code-Switching

The awareness of code-switching differed among the interviewed teachers. Only one of the participants, T3, reported that he had knowledge of it whereas the other three were unfamiliar with the concept prior to the interviews.

T3 remembered that he learned about code-switching when taking courses in Swedish as a L2 during his teacher education. He expressed that code-switching both can concern switching languages if the speaker is multilingual, but that it can also concern “switching [language] styles” depending on the recipient and situation. An example of this switch, T3 explained, can be to “shift between ‘chat language’ and written language when writing an essay versus when writing on Messenger”. This explanation provided by T3 does not fully correspond to the concept of code-switching (see the definition of code-switching presented in section 3.2.) since T3 seems to interpret ‘codes’ as if it also pertains to intra-language variety rather than recognized languages. However, in conformity with code-switching, a switch in language styles requires the speaker to adapt his or her utterances to the recipient and can be used for various purposes such as to signal personal identity (Gardner-Chloros, 2009). Furthermore, T3’s explanation also goes beyond the code-switching framework by including written text. As mentioned in secion 3.2, the concept code-switching seem to exclusively cover oral language and thereby excludes writing (see, e.g. Auer, 1988; Corcoll López & González-Davies,

2016). Accordingly, that T3 has understood code-switching in this way reveals some limitations in his awareness.

In contrast to this, neither T1, T2, nor T4 expressed that they had any knowledge of code-switching before the interviews, and both T1 and T2 mentioned that they searched for the concept before the interview took place. Yet, T1 and T4 vaguely recalled the concept code-switching being mentioned in the English courses during their teacher education but did not remember the purpose of it. Upon the question of whether they believed that code-switching could help students develop English language proficiency, T1 replied hesitatingly: “I do not know how to answer that question” and continued “well, yes, maybe”. T1 further said that one could of course learn English through the use of Swedish by working with vocabulary lists and translation. These findings show that the interviewed teachers have little to no knowledge about code-switching and its potential benefits. By comparison, T3 shows awareness of the potential benefits of code-switching in saying that he believes code-switching could help students develop their English language proficiency. He stated:

“I do not think that there is any teacher who in good conscience could say that they only speak English in class, and I do not think it is the most successful way either. I mean you must be able to switch between explaining the assignment in English and provide clarifications of some parts in Swedish.” (T3)

This utterance is in line with what Lundahl (2019) and Moore (2002) state being one of the purposes for code-switching, namely to make communication more effective by using code-switching as a support for listening comprehension.

Furthermore, T3 and T2 provided similar descriptions of today's students frequently switching between Swedish and English, and that code-switching, therefore, is inevitable and natural. These findings align with the foundations of code-switching expressed by Poplak (1980) and Lightbown and Spada (2013), namely that code-switching is a natural language phenomenon. However, T2 expressed that it occurs “not in a positive sense” in the classroom. One interpretation of this utterance is that T2 believe students’ switch of language to be a sign of insufficient language proficiency. This is in line with Liebscher & Dailey-O’Cain (2005) stating that code-switching could

be the result of insufficient knowledge in the TL, and also with Lundahl (2019) who state that code-switching could be a way to avoid expressing oneself in the TL. T2’s statement could, furthermore, be a result of his lack of knowledge of code-switching and the impact of the syllabus (Skolverket, 2011a). As will be presented further down in this section, T2 interprets the syllabus as a TL-only principle is being promoted, and thereby, it signals that the L1 should be avoided. Another interpretation of T2’s utterance is that he believes that the code-switching that appears in the classroom does not occur in an informed way. This idea is supported in Zhu and Vanek (2017), who found that aimless use of the L1 deprives students of important L2 interaction time. This, furthermore, indicates that T2 gained some knowledge about code-switching from searching for the concept before the interview as he, by expressing that code-switching does not appear in a positive sense, reveals that he understands that code-switching could occur in a positive sense. This is also in line with Rahayu and Margana (2018) who found that allowing code-switching in the classroom made the students more talkative but in a negative way as they did not speak English, whereas the students in classrooms permeated by a TL-only principle were more active in using the TL.

In conformity with research (see, e.g. Yuvayapan, 2019; Allard et al., 2019), T3 argued that it could be favorable to “use it [code-switching] consciously and clarify to the students [...] when it is efficient and not to do it”. T3 further explained that even though the goal is to speak English in the classroom, he does not believe in a classroom culture where students are forbidden to use Swedish as this might result in students not daring to ask questions or ask for clarifications in class. Instead, T3 believes that one should lower the students’ bar to switch between the languages. This idea is supported in both Zhu and Vanek (2017) and Lo (2015) who found that an adoption of a code-switching pedagogy made the students more active in the classroom and it enhanced the level of student talk. This is, furthermore, relevant to the communicative classroom as it is stated in the aim for English in upper secondary school (Skolverket, 2011c) that students’ should be given opportunities to interact in speech and to produce spoken language.

All the interviewed teachers expressed that the TL should be used as much as possible in the language classroom and provided different reasons for that. For example, T2 explained that code-switching is not promoted in the guidelines by Skolverket and he

interprets the guidelines as one should abide by English as much as possible in the classroom. This finding partly agrees with the statements in the syllabus for English in upper secondary school (Skolverket, 2011a) where it is state that “teaching should as far as possible be conducted in English” (p.1), but does not take the expression of the promotion of plurilingualism, expressed in the same syllabus, into account. T2’s explanation is also supported in Yuvayapan (2019) who proposes that, despite the potential advantages of code-switching, a TL-only practice could be the result of the impact of the curriculum. A similar view as T2 presented was communicated by T3 who said that one should mainly use the TL in language education in school. Although T3 did not explain this statement further, it is interpreted as he was referring to the same passage in the syllabus as T2 did. Moreover, T1 said that he almost exclusively uses English in the classroom and believes that “it generally does not impair the learning in any way”. Here it can be seen that T2 relates to the commentary material for the subject English (Skolverket, 2011b) where it is expressed that occasional elements of L1 could be used in teaching if the teacher considers it as helpful to increase understanding. Consequently, these ideas of using the TL as much as possible in the language classroom expressed by both the syllabus and the participating teachers are in line with Cummins (2007) and Lundahl (2019) who state that a greater use of the TL can result in higher language achievement. Similarly, that T1 expressed that a TL-only practice did not impair learning in any way, is also supported by the findings by Lee and Macaro (2013). They show that adult students’ vocabulary acquisition did not differ between a code-switching instruction and a TL-only instruction. In Lee and Macaro’s (2013) study, adult learners refer to students at university level, and they state that proficiency levels often accompany age differences. Since T1’s reports that his students are relatively proficient in English (see section 5.3), a TL-only principle in T1’s classrooms seems to be well-motivated.

Furthermore, T4 explained that she tries to use English for the most part in the classroom to “create an English education which is as authentic as possible” and T2 mentioned that he assesses his students based on the idea that an English speaking person should understand the students’ production. Therefore, if the student switches between Swedish and English too much, the English speaking person would not understand the message. This notion of using the TL as much as possible also corroborates the findings by Simasiku et al. (2015) and Yuvayapan (2019) who found

that teachers prefer a TL-only principle to help students strengthen their linguistic competence to acquire a native-like proficiency. This goal of reaching native-like proficiency is not stated in the syllabus (Skolverket, 2011a). Instead, it is expressed that students should develop knowledge of English “so that they have the ability, desire and confidence to use English in different situations and for different purposes” (Skolverket, 2011a, p.1) in order to increase their opportunities to participate in “different social and cultural contexts, as well as in global studies and in working life” (p.1).

5.2.

EFL Teachers’ and Their Students’ Usage of

Code-Switching

Although most of the interviewed teachers lack knowledge of the concept of code-switching, and they all expressed that a large amount of TL-use is favorable, all participating teachers presented examples of code-switching in their teaching. Some of these occurrences of code-switching were intentionally planned for whereas others happened unintentionally.

5.2.1.

Intentional Code-Switching

The examples of intentional code-switching found in the material regard the use of code-switching to focalize on form, for communication effectiveness, for relational purposes, and due to convenience to avoid using the TL.

5.2.1.1.

Focalization on Form

All teachers but T2 spoke of using code-switching to focalize on form. For example, T1 expressed that he occasionally switches to Swedish to explain something grammatical that he believes his students need to revise or if he, for example, “needs to explain the difference in what the student says and what he instead should say”. That is, T1 clarifies the inaccurate use of English in an utterance by providing an example of how the same utterance would sound in Swedish. This finding is in line with what Moore (2002) and Lundahl (2019) express as being one of the functions of using code-switching in the classroom. Furthermore, this idea of using code-switching for a form focus is also supported by Kocama and Aslan (2018) showing that students find it positive to make use of the L1 for explaining grammar or for highlighting differences between the L1

and the L2. T1’s use of code-switching for this purpose is also in line with the commentary material for English (Skolverket, 2011b) where it is expressed that occasional elements of the L1 can occur if the teacher considers it to contribute to increased understanding.

Moreover, T3 expressed that he uses code-switching when talking about difficult words and their meaning. In those cases, he said, he can make quick comparisons between different Latin languages since “many of the difficult words in both English and Swedish have their origin [there]”. Furthermore, he tries to link difficult words to as many languages as possible for the students to see connections between the languages. In correspondence with the result presented above, T3’s examples of code-switching are also in line with theory (Lundahl, 2019; Kocama & Aslan, 2018; Moore, 2002) where it is expressed that code-switching can function as an effective tool for explaining new words. Also, this use of code-switching can be promoted both by the desirable plurilingual approach expressed in CEFR (Council of Europe, 2001) and in the syllabus for English (Skolverket, 2011a) where it is stated that teaching should encourage students’ curiosity in language.

5.2.1.2.

Communication Effectiveness

All the participating teachers use code-switching for communication effectiveness. For example, both T3 and T4 explained that if their students are uncertain of a specific word or want to ask a question in Swedish, they encourage them to try to explain the word or question in English, but if it does not work they allow them to use Swedish instead. These findings corroborate the findings by Lo (2015) and Zhu and Vanek (2017), who found that teachers use code-switching to provide translation in order for the students to, for instance, access content. In the situations explained by T3 and T4, one could also argue that the use of code-switching is a result of the lack of relevant vocabulary (Lightbown & Spada, 2013). However, as the students await the teachers’ approval to use the L1, it appears to be an informed strategy that students are able to use in the classroom, which corresponds to one of the goals for the subject of English: “The ability to use different language strategies in different contexts” (Skolverket, 2011a, p.2). The idea of letting students code-switch is further supported by Zhu and Vanek

(2017) who show that code-switching encourages more student interactions, which is relevant for promoting the communicative English classroom.

Furthermore, T4 explained that she shifts to Swedish on occasions where it is particularly important that the information is getting through. For example, because of the prevailing distance learning, she has sent e-mails to her students in Swedish to ensure that important information gets through. This is in line with Allard et al. (2019) who found that teachers believe code-switching to be useful for giving access to content. Another example of code-switching in T4’s classroom is that T4 and her students have agreed on her switching to Swedish when explaining or talking about the knowledge requirements since the students must understand them. This was also shown in the observed class of T4 where she explained a new assignment in English and suddenly shifted to Swedish:

T4: Och eftersom det handlar om betygsättning går jag över till svenska And since this is about grading I will move on to Swedish

This finding is an example of an intentionally planned code-switching and thus is what Zhu and Vanek (2017) suggest as an optimized use of code-switching. Zhu and Vanek (2017) state that the teacher and the students should agree on norms about language use in the classroom, which is exactly what T4 has done in her classroom. She has consequently given the students possibilities to influence the teaching and has adapted the teaching to the students’ conditions and needs, which are two aspects required from Skolverket (2011c).

In the same class, it was also shown that T4 uses code-switching to translate specific words within a sentence:

T4: Do you need props, rekvisita, for your movie

props

Here, in contrast to the focalization on form (see section 5.2.1.1), the purpose of this utterance by the teacher was to convey a message, that is, to explain an assignment, and not specifically to teach the students this new word. Thereby, according to the author’s

interpretation, the purpose of this code-switch was to make the communication more effective. This finding is in line with Lo (2015) and Zhu and Vanek (2017), showing that teachers in their studies used code-switching for the same purposes.

Similar occurrences of code-switching were found in T3’s lesson. To exemplify, T3 gave the students the task to “write the headlines in abbreviations” and one of his students’ immediately asked, in Swedish, what they were supposed to do. In this particular case, T3 explained the task again, but this time in Swedish. This finding corroborates the findings by Allard et al. (2019), showing that teachers use code-switching to give access to content. Furthermore, Yuvavapan (2019) also shows that teachers use code-switching to help low proficiency students, which the example from T3’s class could be an example of. Other examples of code-switching for communication effectiveness were when Swedish was used by the students to answer questions, which is illustrated by the following example:

T3: What is a noun?

S: Substantiv Noun

In cases like these, T3 explained after the observation, where T3 asks a relatively simple question to see if the students are following, he finds it more efficient to let students answer in Swedish instead of having them trying to explain what a specific term is in English. That is, code-switching is used to make the communication more efficient, which is also found in Moore (2002).

Lastly, T2 declared that in his classes of English at the elementary level he “has to rely on Swedish to, for example, explain assignments”. He explained that when speaking English in class he has to repeat and explain in Swedish what he just said so that his students who “do not understand English” will get the message. In these classes, student-coaches who speak the students’ L1s, which are not Swedish, aid T2 in explaining the given assignments in the students’ L1s. Besides this, T2 also expressed that he asks his students to use the online dictionary Lexin where they can use their L1 to understand an assignment or translate a word. These examples of code-switching have a content focus and functions, as Moore (2002) and Lundahl (2019) state, as a

support for listening, reading, or conversational comprehension. That is, code-switching is used for the students to be able to carry out assignments and the teacher adapts his choice of language to the pupils so that they can develop their English, which is in line with what McKay (2012) states being beneficial for the development of their English proficiency. As T2 explains, his students are not that proficient in English, which means that this finding is also in line with the findings in Lo (2019), showing that L1 is used to a greater extent in classes where students’ TL proficiency level is low.

Throughout this section, several examples of the use of Swedish in the English language classrooms are presented. Although the participating teachers seem to adapt their teaching to their specific group of students, both Allard et al. (2019) and Lundahl (2019) highlight that a shift from L2 to L1 in the classroom is not as beneficial for the students who do not share the same L1 as the majority of the class. In all the participating teachers’ classes, there were students with other L1s than Swedish. As such, these students might be disfavored by the approach where Swedish is used as a reference when learning English. Nevertheless, as the participating teachers are not proficient in the students’ L1s, they cannot use those languages when giving instructions. Instead, when they notice that the students do not understand what is being said in English, they switch to Swedish. As such, they make use of the languages they know, intending to fulfill communication needs. Also, the teachers provide other communicative support for the students who might not be benefited by this code-switch. One example of this is T2’s classroom practices where student-coaches who speak the students’ L1s are able to assist the students, and that T2 encourages students to use translations.

5.2.1.3.

Relational Purposes

Both T1 and T4 mentioned that they use code-switching for relational purposes. T1 pointed out that if a student comes to him after class and explains that he or she is not feeling well or is feeling stressed, he switches to Swedish; “if I notice it [that the student is not feeling well] and still would continue with English, it just feels annoying”. Instead of talking to the students in English with the purpose of developing their skills in those situations, T1 wants to “talk to them for real”. Similar examples were provided by T4 who explained “I get a completely different relationship with them [the students]

when I speak Swedish”. This was also shown in the lesson observation with T4, where one of the students stayed in the virtual classroom while the rest of the class logged out. The conversation that took place between T4 and the student, in which the student was crying and told T4 that she was stressed and T4 tried to support the student, was conducted solely in Swedish. These results corroborate the findings by Yuvayapan (2019) who shows that the teachers in her study believed code-switching to be useful for building bonds with the students. This is furthermore an example of where the teachers step out of their professional roles as teachers, where their main goal is to teach students the TL, and enters a more informal role where their main goal is to maintain a good relationship with their students.

5.2.1.4.

Convenience

All the participating teachers provided examples of code-switching due to convenience in their classrooms. Based on these findings, it seems like this occurrence of code-switching is due to students’ insecurity. For instance, T1 explained that a few of his students show resistance towards speaking English in front of the class. Likewise, T2 said that his students prefer to speak Swedish or Arabic in the classroom because he believes they “prefer to speak what they are more comfortable speaking”. Similarly, T3 said that most of his students prefer to speak Swedish in the classroom because “in general, they do not speak it [English] outside the classroom”. These examples of code-switching could be a sign of students’ lack of proficiency in the TL. As Lundahl (2019) states, a code-switching behavior could be a way to avoid expressing oneself in the TL. Or it could also be the result of the lack of relevant vocabulary (Lightbown & Spada, 2013). Another interpretation of this could be that the students are lazy seeing that Gardner-Chloros (2009) states that bilingual speakers themselves report that they sometimes use code-switching when they cannot be bothered to find the appropriate word in the language. In these cases of switching due to convenience, code-switching does not function as an informed way to reach higher linguistic competence, but it could potentially function as a way to signal personal identity, which Lightbown and Spada (2013) and McKay (2012) state is one of the functions of code-switching. These findings are also examples of using code-switching to avoid expressing oneself in the TL (Lundahl, 2019).

Furthermore, T2 mentioned that some of his students do not dare to express themselves in English because they are ashamed of their language. T2 believes that the students do not want to expose themselves to say something or pronounce a word incorrectly because they are afraid that the class will laugh. A similar description was also shown by T4, where she expressed that in her English classes, most students are Swedish L1 speakers, but some are English L1 speakers. In these classes, T4 has noticed that some Swedish L1 students do not dare to speak English to the same extent, as she believes they would have done if there were no English L1 students in the class.

5.2.2.

Unintentional Code-Switching

The examples of unintentional switching found in the material regard code-switching due to forgetfulness. All the participating teachers explained that both their students and themselves sometimes tend to forget to speak English. T1 described how if a student comes up to him outside of class and asks about something, he answers in Swedish because he is not “in an English mode”. Also, both T2 and T3 said that they occasionally forget to speak English themselves in the classroom, and T2, therefore, sometimes receives complaints from his students for doing that. The example from T2’s classroom is in line with what Allard et al. (2019) found in their study, namely that some students are unsatisfied with their English language development due to the use of L1 in the classroom. It is also in line with what Simasiku et al. (2015) found, namely that teachers believed that their students prefer to speak the TL in the classroom. In the study by Zhu and Vanek (2017), they also warn against L1 overuse in these types of settings but argue for a need to optimize the use of L1 to ensure comprehensible input. Moreover, T2, T3, and T4 all provided stories of their students forgetting to speak English in the classroom. For example, T2 reported that many of his students switch to Swedish unconsciously while speaking and, for example, frequently add the Swedish filler word “liksom” while speaking English. Furthermore, T3 said that his students tend to forget to speak English while working in class and that it usually is enough for him to walk around in the classroom for the students to realize that they are speaking Swedish and immediately switch to English again. A similar report was given by T4 who showed an understanding of the tendency to forget to speak English as she expressed that it is unnatural to speak another language with someone whom you know speaks your L1.

Therefore, she pointed out that “one must force ones’ brain” to focus on the L2 and that is why she finds it important that everyone in the class tries to solely speak English. An example of forgetting to speak English was observed in T4’s lesson:

S: Vilka filmer har blivit tagna? What movies have been taken?

T4: Can you say that in English?

S: Oh, yes, which movies are already taken?

In this example, the student asks a question in Swedish upon which T4 reminds her to speak English. The student reacts with an “oh” which indicates that she had forgotten to speak English and immediately after she repeats the question in English.

These examples of unintentional code-switching could be explained by the nature of code-switching, namely that it occurs naturally in bilingual settings and that it can be used in conversations where all interlocutors share the same language repertoire (Lightbown & Spada, 2013). As shown in this section, code-switching seems to happen frequently in these settings where the teacher shares the same L1 as the students. That is, a code-switching behavior is functional, yet to some extent undesirable, in the participants’ classrooms. This is, for example, the case with the students’ use of the Swedish filler word “liksom”. This filler word does not have a corresponding word in English, and therefore, code-switching, in this case, does not seem to be a sign of insufficient language use. Instead, it seems to be consistent with the nature of code-switching as it is used to fulfill communicative needs (Corocoll López & González-Davies, 2016).

5.3.

Difference in the Use of Code-Switching Depending

on the Students’ Language Level

The third research question seeks to answer how teachers’ choices of code-switching in the classroom differ depending on their students’ language level. As illustrated by the interviewed teachers, the use of other languages than English seems to differ between the different English courses as well as their student groups. One exception to this is illustrated by T4 who said that she does not adapt her language at all regardless of the English course since she tries to “have this genuine English in the classroom always”.

As previously mentioned, T1 practices an English-only principle in his classrooms and motivated this by explaining that an exclusive use of English will make the students learn more of the language. T1 explained that his students are “generally very, very good at speaking English” and that some of his students have English as their L1 whereas other Swedish L1 students speak English as if it was their L1. T1 said that “I would say that all my students keep up with everything we do, even though I am only speaking English” and instead of switching to Swedish, he adapts his language to different levels. He gives an example of when the students are working with a project about, for instance, the climate and they are supposed to use difficult vocabulary he tries to “take it down to a less abstract level and talk about what the terms mean instead of blurting out the terms and thereby lose the students”. Another example of adaptation to the students’ levels was provided by T2 who explained that he is much more strict with the English-only principle in English 6 than he is in English 5. He described that “in English 5, if they happen to say something in Swedish or if they maybe ask how something is pronounced or so, one can overlook it in English 5, but not in English 6”. These findings are in line with previous research and theory (Lo, 2015; Lundahl, 2019). For example, Lundahl (2019) argues that the usage of code-switching should be dependent on the students’ language level, which both T1 and T2 seem to have adapted. Similarly, in the study by Lo (2015), it was found that the teacher makes use of students’ L1 more in classes where the students TL proficiency level is low. In this study it was found that in classes with students of higher language ability, as is the case for T1 and T2’s classes, the teachers occasionally made use of the L1 as a support, but in some cases the teachers instead elaborated more in the TL. Moreover, both T2 and T3’s code-switching behavior is in line with the recommendations from Skolverket (2011c), namely to adapt the teaching to the specific students’ needs.

T3 said that he speaks more Swedish in the course English 5 than in English 6 because the language requirements are lower in English 5. Furthermore, T3 explained that the students taking that course are in a vocational program, and “their goal only has to be to get an E in the course to get their school leaving certificate”, and therefore he puts more effort into pushing them over that finish line. On the other hand, in the English 6 course, T3 explained that “the idea is that you should be able to study at university afterward and be able to have course books in English and lectures in English, so therefore one does one's students a disservice if one adapts too much there”. These findings, similar to