Why not “ English only ” ? Bachelor’s thesis

(1)

Why not “English only”?

Patterns of code-switching between Swedish and English in

Swedish upper secondary EFL education

Bachelor’s thesis

Author: Johan Westin

Supervisor: Charlotte Hommerberg Examiner: Ibolya Maricic

Term: Spring term 2019 Subject: English

(2)

Abstract

English education in Sweden tends to be viewed as a second language, rather than a foreign language. Therefore, it is generally expected that instruction is performed, and content is taught in English. However, previous research shows that English is generally not the sole language used, even in classrooms with explicit “English only” policies. The following essay has investigated how this translates into classroom practice through observations of classroom interactions between students and teachers, as well as between students and students. Three different teachers were observed at two upper secondary schools. Shorter interviews were also conducted with the teachers. The results showed large differences between observations, but some clear trends were observed nonetheless. Socializing and metalanguage were more prevalent than other code-switches among students, while floor-holding was very rare and in most observations, non-existent. All teachers claimed to use English as much as possible although they also recognized the utility of using Swedish as a tool in the classroom. The view that English should be used as the language of instruction to a great extent was additionally observed in teacher interviews. However, there were no categorical opinions arguing for “English only”. The teachers also largely concurred with the use of Swedish, or other first languages as a language of comparison. This was supported by observations in the classrooms and significant differences existed between teachers. An investigation with a larger cohort in this field could be fruitful for future researchers. The results of the study could be used to further examine strategies to find a balance in the classroom between English and Swedish. The differences between classrooms show that there is no clear consensus on the extent of Swedish use in the EFL classroom. In order to improve teaching it is important to know why students use code-switching and in what situations it might be helpful to them, such as in translation or grammatical rules. However, different classrooms may require different approaches, which highlight the importance of discussing these issues.

Key words: Code-switching, EFL teaching, Language ideology, Equivalence, Reiteration,

(4)

1 Introduction

When it comes to teaching English as a foreign language (EFL) a common view is that

students and teachers should use English to as great extent as possible (Källkvist 2017: p. 27). The curriculum additionally contains formulations that strongly imply a focus on the use of English as a language of instruction (National agency for education. 2011: p. 1). Lin, (2013: p. 211) also describes a similar view from an international perspective. In accordance with this, reluctance by teachers to deviate from English can be observed in literature covering this topic. Källkvist (2017: pp. 28-29) specifically favors claims that the “English only” does not have a sufficient scientific backing. Given this background, further research into the role of other languages than English in the EFL classroom can likely produce interesting results. After all, if teaching is not based on solid evidence, it could jeopardize the overall quality of teaching. The drawbacks to the “English only” approach are for example that certain aspects of second language (L2) teaching such as translation and the study of grammar rules would likely benefit from the inclusion of the first language (L1) as a point of comparison, according Källkvist (2017: p. 29). The usefulness of the L1 when studying vocabulary is also relayed by Lin (2013: p. 206), where the researcher argues that when the teacher uses their L1 to instruct their pupils it makes the learning easier, since the pupils only have to focus on one thing at the time. This strategy is also favored by the EFL researcher Levine (2011: p. 4) who is a

proponent of an approach based on the assumption of multilingual classrooms, which he regards as a more realistic assumption than the monolingual classroom. This is especially relevant in reference to the changing demographics of Sweden (Hult, 2017: p. 277). This study will largely focus on the concept of code-switching which is defined as the use of more than one language in the same conversation, or monologue (Nilep, 2006: p. 16).

(5)

1. Aim and research questions

The aim of this study is to investigate code-switching in teaching practice in EFL classrooms and its relationship to scientific literature in the field. The findings are aimed at being

illustrative of how the local language is used in EFL teaching in the selected schools. The investigation will also attempt to identify reasons for the language use from a pedagogical standpoint through brief, informal follow-up interviews with the teachers whose classes are observed.

The research questions are:

1: What types of oral interactions in the selected upper secondary school EFL classrooms are performed in Swedish?

2: To what extent is the L1 of the students incorporated as a part of the learning of the L2? 3: What are the pedagogical explanations for the use of the first language, or the lack thereof?1

2 Theoretical background

The term switching is initially defined, and after this earlier research of classroom code-switching from student perspectives is presented. This is followed by research focused on the teachers’ use of code-switching. The section is concluded with the introduction of the

ecological approach to second language teaching by Levine (2011).

2.1 Defining code-switching

A large number of studies have been conducted using the term code-switching. However, academics tend to use their own definitions, which are tailored to their own fields. A

laypersons definition of code-switching might be that a “code” is a regional dialect, a social variety or a language. A “code-switch” is then a change between these codes. This essay will follow the socio-cultural view of code-switching, since it follows the socio-cultural

perspective that Levine (2011: p. 24) uses. More specifically this essay will use the definition used of code-switching used by M. Heller, which regards the use of different languages within a single linguistic episode. Additionally, this investigation is only concerned with shifts between Swedish (L1) and English (L2) (Nilep, 2006:p. 16).

(6)

2.2 Code-switching practices in EFL education and its effects on student

learning

There are clear indications in previous research that even in schools where “English only” is the policy, the use of L1 still occurs extensively (Sampson, 2011: p. 294). One such study was conducted by Sampson (2011) who investigated the prevalence of L1 usage in a Colombian school for adult learners. Sampson (2011) analyzed the use of L1 with the aid of a number of code switch categories, which are also used in the present study and these categories are equivalence, metalanguage, socialization, floor-holding, reiteration and L2 avoidance. This study shows that the L1 was mostly used as a communicative strategy to complement the L2 in the lessons observed. Sampson (2011: p. 294) therefore suggests that a total ban of L1 use might inhibit learner fluency. This general conclusion was also reached in 2015 by S’ad & Qadermazi, who investigated the attitudes regarding code-switching in Iranian adult learners. S’ad & Qadermazi (2015: p. 171) additionally imply that a ban of the L1 might negatively impact student motivation, since the students were negative to a complete “English only” approach. Sampson concurs with this conclusion and further argues that code-switching should not be entirely banned in the classroom. He does see difficulty, however, in when L1 use should be encouraged and when it should be limited (Sampson, 2011: p. 302). This issue is also addressed by S’ad and Qadermazi (2015: p. 171), who argue for limited and deliberate use of L1. Bensen’s (2013: p. 78) study more specifically argues that code-switching helps student understanding and increases motivation. The study also emphasizes the context of grammar teaching as especially apt for switching. In contrast, the study argues that code-switching may impede communicative competence in students, if used excessively. In

(7)

2.3 Code-switching in teacher practice

There has also been research on the role of code-switching in Swedish lower secondary level by Fägersten (2012: pp. 15-17). Fägersten used an observational method to study teacher code-switching. Fägersten’s results indicate that code-switching at this level is primarily used to establish control over the classroom, when students misbehave. This differs considerably from the previously mentioned studies, since adult learners may be less inclined to be

distracted. The study concludes by remarking that, in instances where the teacher corrects L2 use, the general frame of the lesson is broken and this emphasizes the correction (Fägersten, 2012: pp. 15-17). This use of the L1 as a tool of correction can also be found in Lin’s (2013: p. 201) article discussing schools in Hong Kong, were teachers tend to switch to Cantonese when correcting student behavior.

2.4 An “ecological perspective” of EFL teaching

(8)

is based on research conducted in French and Spanish classes in the USA. This may

somewhat impact the ability of the theory to be used in an EFL context, since English has a different role in Sweden than Spanish and French in the USA. English is a core subject in the Swedish education system and is widely proliferated in Swedish society. This differs in the USA where American students can choose their second language and may not encounter the language extensively outside the classroom (Levine, 2011: p. 4).

3 Method and materials

This section will detail information about participants followed by an outline of the collection of data. This is followed by a section on the analysis method used, as well as a table

displaying the categories of analysis. The section is concluded with a discussion on the ethical implications of this study.

3.1 Participants

(9)

Table 1 information regarding observations

Observations Teacher School Class size Language background

Program Lesson time Observation 1 Male teacher

20 years’ experience (Teacher 1) Large upper secondary school in medium sized city 24 70-80% of students have different L1 than Swedish Social Sciences and Humanities program 2018 10:25-11:40

Observation 2 Teacher 1 Large upper secondary school in medium sized city 22 70-80% of students have different L1 than Swedish Social Sciences and Humanities program 2018 10:25-11:40 Observation 3 Female teacher 25 years’ experience (Teacher 2) Small upper secondary school in small town 25 2 students have different L1 than Swedish Social sciences and economy program 2018 13:40-15:00

Observation 4 Teacher 2 Small upper secondary school in small town 14 1 student has a different L1 than Swedish Social sciences and economy program 2017 12:35-13:35

Observation 5 Male teacher 16 years’ experience (Teacher 3) Large upper secondary school in medium sized city 26 2 or 3 Students have different L1 than Swedish Social science program 2018 8:00- 9:15

Observation 6 Teacher 3 Large upper secondary school in medium sized city 26 2 or 3 Students have different L1 than Swedish Economy program 2018 9:30–11:45

Students with other first languages than Swedish were present in all six observations.

(10)

students with other first languages were also present. However, they were much fewer and thus did not affect the teachers’ attitude towards code-switching as significantly. The other first language present in those classrooms was Arabic. It is much less likely that the presence of another first language impacted their teaching as compared to teacher 1.

3.2 Collection of data

This study combines non-participant observations, which are detailed initially, with short, follow–up interviews with the teachers of the observed classes, which concludes the section. The main data was collected through the earlier method, while interview data function to complement the observations and highlight the teacher perspective on the lessons. 3.2.1 Non-participant observation

The method chosen for the observations is what is called non-participant observation (Dörnyei, 2007: p. 179). This kind of observation entails that the observer does not interfere with the content of the observed lesson. Consequently, the observed lessons were interfered with as little as possible. The only other interference with the lessons involved the observer moving around the classroom. Additionally, a request that the lesson contain some oral activity was conveyed in order to evaluate “on task” speaking from students. When observing group work, only one group was observed at a time. This was especially relevant in

observation 5 and 6, where groups were in different classrooms. Groups were observed for ca 10 minutes and then a different group would be observed. Group work was the main part of observation 4, 5 and 6, as well as substantial parts of observations 1 and 2. The form of lesson structure likely had an influence on the result of the observations, which is worth considering. In order to make sure that classroom observations could be reliably analyzed, this study made use of audio recordings. However, two observations were conducted without the aid of

recordings. The reasons for this decision was mainly ethical considerations (section 3.4). This means that total accuracy regarding every word is not attainable in these observations.

Examples from these lessons are therefore referred to as “paraphrases”. Time markings were written down, in observation 1-4, for relevant interactions so that they can be identified and analyzed in depth. A similar method is discussed by Dörnyei (2007: pp. 180-1). However this study used video recording, which is here rejected, since it may be distracting (Dörnyei, 2007: p. 184).

(11)

maintained. Moving on, the observer was placed in the middle of the classroom in following lessons, to get better sound quality. During the initial lesson the observer was placed at the front. The change in placement increases the interference with students, but also provides more clear data. Having a more central position in the classroom, as well as moving around to different students, likely make them more conscious of the presence of the observer and may therefore change their behavior. The most likely change in this study would likely be a tendency for the student to use a more correct language than usual and avoid code-switching. This is what is called the observers paradox and is very hard to avoid completely. However, since the students had been asked to participate long in advance and recording equipment was kept to a minimum, the effect should have been somewhat reduced (Dörnyei, 2007: p. 90). The teacher was also consulted in the interviews about potential deviations from a typical lesson regarding the students’ use of code-switching. Data from the initial observation was used in the final study, since only minor changes were made.

3.2.2 Interview data

The interviews were conducted with the three teachers whose classrooms had been observed. The interviews were conducted in Swedish and consequently answers were translated from Swedish to English. This choice was made, since the interviews were conversational and Swedish allowed for a more casual level of discussion. Using Swedish allows for a more easily understood conversation, since some technical language is used. The observer might not understand the intent behind this language if it is only uttered in English. However, the interviews were not recorded, but rather written notes were taken. This choice is further discussed in section 5.1. Interviews were conducted after observations and with a duration of approximately 10-15 minutes and featured questions, which can be viewed in Appendix 2 in both translated and original form. The interviews were of an informal and brief nature and guided by the respondent’s perspective (Diley & Thierney, 2001: p. 92). However, the focus was maintained on L1 use and pedagogical approaches applied by the teacher. The questions were formulated to probe into the teachers’ pedagogical choices in the classroom. As follows, the interviews are included to identify how the teachers view their own, and their students’ use of L1 in the classroom.

3.3 Analysis of data

The main method used here is called a qualitative content analysis, which entails the

(12)

function of the code-switching. This type of categorization is similar to other content analysis and consists mainly of transcription, rereading notes and listening to recordings (Esaiasson, 2017: p. 325). The code-switching are divided into six distinct categories, in accordance with Sampson (2011). In order to clarify the categories used in this study, some examples of language use have been provided in Table 2 below. These categories have a major effect on the outline of the method used, since they are the basis for the observation protocol.

Therefore, this section is placed in the method section, rather than theory. This protocol serves to organize the main material for this study and serves as the main analytical tools. Moreover, the categories also impact how the overall results are presented.

(13)

Table 2. Examples based on categories by Eldridge (1996), via Sampson (2011).

Equivalence is the usage of the L1 in the place of a word, or phrase in the L2. A word or

phrase is uttered in L1, but the sentence otherwise maintains the L2. Equivalence may be used in translation, as in the example below:

T: How do you say Boskap ? (Boskap=Livestock) S: Is it Cattle?

Metalanguage refers to utterances about the task or procedural issues. This can be thought of

as “language about language learning”. This can consist of asking the teacher for instructions about tasks, but also issues such as group divisions or other logistical tasks.

T: Okay before we start the speaking task,-

kan ni sätta er i grupprummet Could you move to the group room S:Visst! Sure

Floor-holding occurs where the L1 is used temporarily to fill place when the speaker cannot

presently recall a particular word in the L2. The speaker often describes the word or does an audible word search. The word or phrase is subsequently reiterated in the L2.

T: What would be your dream holiday if you could choose freely?

S: I would like to… resa…, travel to Japan, because they have an interesting culture.

Reiteration is the repetition of a message that has already been stated in the L2. Reiteration

can be used when students have not understood the message, or when they have simply failed to hear what is said. An example of reiteration is:

T: I would like you all to write 200 words about the plot of the movie we just saw. S: Va? What?

T: Skriv 200 ord om handlingen i filmen. Write 200 words about the movie’s plot

Socializing is a type of code-switching which occurs when students make jokes or otherwise

engage in friendly communication. Sampson (2011: pp. 299-300) argues that this can be used to build group solidarity.

T: Alright, ten more minutes of speaking!

S1: Hänger du med på träningen sen? Will you come to practice later? S2: Nä, jag måste öva till matteprovet? No, I have to practice for the maths test

L2 avoidance is when the L2 of the student could be used, but the L1 is used instead. This

type of code-switching has no clear communicative utility. The following example shows how this looks in the context of a language task:

T: Ok everyone, start by asking your friend how they are. S1: How are you?

S2: Jag mår bra. I’m fine

(14)

Potential drawbacks of using the scheme in Table 2 also exist, since the categories may miss some nuances that another system may have found. One such example may be metalanguage, which is broadly defined and includes both talk about the task at hand and communication about e.g. seating arrangements. L2 avoidance may also be hard to differentiate from

socializing or metalanguage. The differentiation was made, however, by focusing on the issue of communicative purpose. If the communication was deemed not to have a clear purpose, L2 avoidance was chosen, but when a purpose was detected the corresponding category was chosen. In Table 2, the example of L2 avoidance shows a student deviating from the task, seemingly due to distraction. In the socializing and metalanguage examples, however, a purpose to the switch can be identified, i.e. a student is connecting with a friend and a teacher gaining their students attention. There is also the issue of interactions with several functions, which are noted in the data once per category in accordance with Sampson (2011: p. 245), although his entire system is not used. These occurrences are noted in brackets.

3.4 Ethical considerations and practical details

This study has been based on established practice in regards to dealing with human subjects. “Research Ethical Guidelines” (Eriksson, 2019) by Vetenskapsrådet has been the main guide used to make sure the parameters of this research do not breach trust with the teachers and/or the students observed. In order to ensure this, a consent form has been created that includes relevant information about the study for teachers and students, where they give consent to participate the study. They are not referred to by name in any material to make sure that the promised anonymity of the study is upheld. Two consent forms for teachers and students can be found in Appendices 3 and 4. The information on the consent form was also conveyed orally in preparation of the observations.

(15)

recordings in the classroom, since the teacher wished to collect parental consent before any recordings. This was not feasible due to time constraint and a decision was consequently made, in conference with teacher and supervisor, to carry out two observations without recordings. This situation highlights the need to establish communication with teachers early on. It is also vital to be mindful of the teachers’ position and follow the teachers’ instructions regarding observations. This is important even if the general ethical guidelines are met. The final arbiter is still the teacher and/or the school. In order to avoid such difficulties, a thorough consideration of the method is vital. The decision to carry out the last two observations was deemed feasible, due to the experience in using the observation protocol gained by the observer from observations 1-4.

4 Results and analysis

The following section contains the principal results of this study. Initially a general overview is provided in Table 3 & 4, as well as an initial summation of findings based on the numerical trends. Following this is the main qualitative analysis of observations in classrooms. The section is finalized with the analysis of the teacher interviews.

4.1 Overview of code-switching in the full data set

Table 3. Overview of identified code-switching in communication between teacher and students .Bifunctional interactions are showed in brackets

Code-switches Observ.1 Observ.2 Observ.3 Observ.4 Observ.5 Observ.6

(16)

Table 4. Overview of identified code-switching between students.

Code-switches Observ.1 Observ.2 Observ.3 Observ.4 Observ.5 Observ.6

Equivalence 0 0 7(1) 5 2 1 Metalanguage 5(3) 39(5) 9(2) 11(1) 6 6 Socializing 46(1) 26 29(1) 4 7 1 Reiteration 1 1 0 4 1 1 Floor-holding 0 0 0 3 0 0 L2-Avoidance 8(2) 26(5) 0 6 (1) 3 2 Total 60 92 44 33 19 11

One initial conclusion reached is that there are pronounced differences between observations, which need to be considered. However, some clear trends can, nonetheless be identified, such as the frequent use of socializing code-switching, which is common in 4 of 6 observations. This code switch is not commonly used in communication involving teachers, however. The results also show a sparse use of floor-holding, which is of potential interest. A complete overview of the results can be observed in the following tables. Table 3 regards teacher-student communication, while Table 4 is an overview of teacher-student-teacher-student interactions.

Overall, there are clear differences between teacher to student communication and student to student communication. Teachers used code-switching, comparatively sparsely and

consistently, while student switches were more erratic. Although all observations contained oral tasks, some lessons were more focused on group work and other observations featured individual or pair work. This factor may have impacted the observation result, as well as the timing of lessons, which will be elaborated on in the discussion (section 5). The two following tables show interactions between students (Table 3), as well as between the teacher and

students (Table 4). This division was done in order to separate the teachers’ use of Swedish as a tool and thus closer analyze these interactions.

4.2 Qualitative analysis of observational data

(17)

sections based on the different code-switch categories and the sections include instances from all six observations.

4.2.1 Equivalence

The use of equivalence in the classroom was relatively consistent over the observed lessons. There were generally few occurrences and the teachers were involved in half of them. Equivalence was used by all three teachers when they were referring to a specific place or thing. An example of this language use from observation 1 is provided here:

Example 1:

T: “We are fortunate to have students doing MuT, Musiktillägg, that we have the aesthetic, aesthetics program here, to let us watch all these musical concerts and shows and everything“[3:05-3:25 obs. 1] A noticeable detail from this example is that the specific detail Musiktillägg (additional music) is uttered in Swedish, while the more general aesthetics program is translated into English. This use of equivalence to refer to specific places can also be identified in both observation 3 and 4. An example from observation 3 illustrates this clearly.

Example 2

T:”Would you go to the movies, a movie at Kulturhuset?” (Kulturhuset=local community center) T:”No?”

T:”What would you do then?” T: “Pizzeria Florens?”

S: “Gatuköket!”(Approx. The take away) [48:00-48:20 obs. 3]

This pattern can also be identified in student language, since they use Swedish when referring to locations. This was noticeable in the fourth observation. This use of equivalence is likely used to be less ambiguous, as well as more expedient when communicating. Students were also observed using equivalence code-switching when translating English words to Swedish. In observation 6 the teacher used equivalence to help students translate the word “brat”, but otherwise teacher 3 was not observed using equivalence code-switching in either observation 5 or 6. The students, however, did use equivalence switches a number of times in a similar manner to observations 1-4. An example of this can be seen below, when one of the students switched to Swedish when mentioning the name of the charitable organisation

(18)

Example 3

S1: “You could have given that money to charity, to Barncancerfonden“ [Observation 5] [Paraphrase] The Children’s cancer fund

This use of equivalence is similar to the previous uses of equivalence from both the teachers and students, since the student does not paraphrase the name of the organization, but rather uses the Swedish name.

4.2.2 Metalanguage

The use of metalanguage shifted significantly between observations. This is not entirely surprising since the category of metalanguage is quite widely defined by Sampson (2011) whose definition is followed. This wide definition could explain, at least partially, the relative abundance of metalanguage in observations 2, 3, 4, 5 and 6.

The teachers were sparing in their use of metalanguage, but there were a few noticeable occurrences. Teacher 1 used code-switches from English to Swedish, during observation 1 and 2. These switches were clearly identifiable. This was only observed once each lesson and was seemingly the result of a student needing additional instruction. This type of very

deliberate metalanguage use was used by all three observed teachers. The switch was very clear and the teacher maintained a low sound volume throughout the exchange, which is why only a part of the exchange is displayed here:

Example 4

T:”When we have what’s called the possessive case, genitive, då har vi…”[47:00-47:10 obs. 1] Metalanguage use between the students was significantly more frequent during group and independent student work as compared to teacher-led tasks, which was especially apparent during observation 2. When the students discussed their tasks, they frequently switched over to Swedish when asking each other for help. This shift can be clearly observed in Table 4. One factor that may impact the increased use of metalanguage is the absence of the teacher throughout parts of observation 2. When the teacher left, there was an apparent increase in the use of Swedish metalanguage. When discussing the procedure of one of their tasks the

(19)

Example 5

S1: ”Jag tänkte att jag kommer ta den.” S1: ”Eller att jag måste lämna, förr eller senare.”

S3: ”Då gör man ju det då.” S:2 ”ah…”

S:1 “Nej…” S1: ”Då är det…”

S: 4 “I just wanted to complete the game, so I was like, ok

I thought that I was going to take it. Or that i must leave it sooner or later. Then you do that then.

Uhu… No… Then it is…

[19:10-19:20 obs. 2]

Moving on, teacher 3 prominently used Swedish when discussing shared grammatical features between Swedish and English. The teacher frequently used Swedish for the following minutes of dialogue with the students. This seemed to be a deliberate choice and the shift was rather clear. Students conversed with the teacher in either Swedish or English, which the teacher was seemingly comfortable with. Metalanguage code-witching between students was used

(20)

4.2.3 Floor-holding

Floor-holding was not observed during the first three observations, either from the teachers or the students. However there were some occurrences of floor-holding during observation 4. There was also some floor-holding in observation 5, where the teacher engaged in a short role playing exercise. Further elaborations on the lack of floor-holding can be sound in the

discussion (section 5). An example of floor-holding from observation 4 can be seen here: Example 6

S1: “What’s the word, Dagis…“ (Dagis=Kindergarten) S2: “Daycare?”

S1: “No, it’s like a German word…Kindergarten!” [23:50-24:00 obs.4]

This example was one of very few occurrences of floor-holding. In most observations these switches did not noticeably occur. The implications of are further discussed in section 6. 4.2.4 Socializing

The most common use of code-switching in the classroom was in the form of socializing in student-student communication. The socializing code-switching was performed “off-task” and rather quietly in observation 1. The instances are consequently often difficult to record. In contrast, observation 3 contained considerably louder socializing. This type of code switch was common in gaps between tasks, as well as transitions to new tasks. As can be seen in Table 4, the number of socializing occurrences shifted greatly depending on the lesson. This is likely due to differences in lesson structure. However, the prevalence of socializing is a clear pattern in the data. One factor that may have impacted the amount of socializing was the length and time of the observation. One example of this is that observation 3, which was held in the afternoon, contained a lot of socializing. As mentioned, when the students choose groups, especially in observation 2, there was a clear increase in socializing between students. The following example is from an exchange between students during observation 4.

Example 7

(21)

4.2.5 Reiteration

Reiteration code-switching was relatively rarely used in the observed lessons. However, these switches appear throughout all observations in teacher communication with students, even if they are used sparingly. Additionally, there is a clear disparity between the different uses of reiteration between the observed teachers. Teacher 1 was rather sparing in his use of

reiteration switches. The students also made use of reiteration in their communication with each other in observations 1 and 2. This was done when students were translating a word, or a concept. This was rare in all the observed interactions and students seemed to avoid situations where this code switch was necessary. In this example from observation 1 the shift involves a number of speakers and several shifts back and forth between Swedish and English:

S:1 Siege

S:1 I picked that one

S2: Belägring? Siege? S3: it’s like a camp.

S2: Yeah, no.

S1: Vad stod de att det betydde? What did it say, it meant? S:2 Lång besvärlig period.[16:20- 16:40 obs.2] Long troublesome period

Observation 3 differed from observations 1 and 2 since the teacher used reiteration in full class at several points in order to clarify some phrases several times. This use of reiteration was also utilized by teacher 3, but not by teacher 1. Reiteration was used to clarify the meaning of words and phrases, such as in the example below.

Example 8

T: Swedes are tall, blonde Vikings perhaps, that’s a stereotype.

T: Ni känner igen ordet stereotyper? S1: Ah…

T: När man målar upp att en viss grupp av människor har speciella drag.

You recognize the word stereotypes? Uhu…

(22)

4.2.6 L2 avoidance

L2 avoidance shifted quite significantly between observations. In spite of this, several general trends can be identified. It is important to note that many occurrences of L2 avoidance are difficult to classify, since they may also fit another category, such as metalanguage or socializing. A substantial number of the questions put to the teachers were asked in Swedish even if they could have been in English. The teachers tended to answer these questions in English, but they did not ask the student to repeat the question in English in most cases. Only one exception to this was found in observation 5. The example below shows a student

avoiding use of English in order to ask the teacher a question in Swedish from observation 3 Example 10

T: Could you read from the first paragraph?

S1: Ska vi läsa… asså, själva… Should we read… uhu, the actual T: Please read it first. [41:10-41:15 obs.3]

There were also general, off-topic discussions conducted in Swedish throughout all lessons. This is where L2 avoidance and socializing overlap. An example of this is two students discussing the appearance of one of their school computers:

Example 11

S1: Titta Look

S1: Har du sett hur den ser ut på baksidan? Have you seen how it looks in the back? S2: Mh… Uhu… [46: 50- 47:00 obs. 2]

(23)

During observation 5 and 6 the occurrence of L2 avoidance was seemingly rare. During observation 5 the teacher asked students to reiterate in English, which was the only such occurrence that was observed in any lesson. However, the teacher also allowed answers to questions in Swedish later on in the lesson, which is an example of the teacher’s strategy to use Swedish as a tool when comparing Swedish and English. These occurrences have not been categorized as L2 avoidance, since the teacher clearly directed them to using Swedish, when discussing Swedish grammar. These shifts are instead categorized as metalanguage, since they have a clear linguistic purpose.

4.3 Teacher interviews

The following section will detail the follow-up interviews with the teachers whose classrooms were observed. The interviews will be displayed in chronological order. The questions in original and translated form can be found in Appendix 2.

4.3.1 Teacher 1

Teacher 1 kept most strictly to English, which the teacher expressed as a deliberate strategy to establish English as the “work language” in the classroom.

“It is important that the students use Swedish as their work language. It makes it easier for them to be comfortable in their use of English and they get to train using the language more, as well.” [Translation teacher 1 interview]

(24)

4.3.2 Teacher 2

Teacher 2 claimed to have a more contrasting approach, which can be seen in observation 3, where the teacher used code-switching, such as reiteration and equivalence frequently. The teacher claimed the use of Swedish was not only useful, but necessary. The teacher especially emphasized the role of Swedish in teaching grammar structures, which is connected to the teachers extensive use of a textbook.

“The textbook used has a contrastive structure and I use contrasting examples from both Swedish and English, especially when grammar is being discussed” [Translation teacher 2 interview]

This approach has led to a natural inclusion of Swedish as a contrast language, according to the teacher. Teacher 2 claimed to have tried to use “English only”, but deemed it stilted and unnatural. Despite this the teacher argued, similarly to teacher 1, that English should be

maintained whenever possible. The teacher claimed that reiteration is sometimes used to make sure that students fully comprehend, which was also observed. However, the teacher still claimed to use English as much as possible when explaining to students. The teacher also commented on the nature of observation 3, saying that the level of fatigue likely impacted their language use, making socializing switches frequent. This was attributed to the length and lateness of the lesson, which is affirmed by comparisons to other lessons, which contained less socializing (see Table 4).

4.3.3 Teacher 3

Teacher 3 claimed to keep mostly to English, often not using Swedish at all. The teacher did make use of reiteration, rather extensively in observation 6, however. Teacher 3 was observed using Swedish when common grammatical principles between Swedish and English were discussed. When asked about this particular use of code-switching, teacher 3 explained that this was a deliberate effort to link the two topics:

“I just want them to understand the links between our subjects. My code switch probably helps them with that.” [Translation teacher 3 interview]

(25)

create links between subjects and explain difficult concepts, while also stressing that English is maintained. Teacher 2 and 3 thus share a more code switch positive approach, with more use of Swedish, while teacher 1 maintains English to a higher degree. These differences were observed in lessons, as well as in the interviews

5 Discussion

The following section will discuss the results of the study in relation to research question 1, which concerns the different types of code-switching used in the observed classrooms.

Following this the last two research questions will be discussed in connection with each other. This is followed by a discussion of the methodology used in this essay and the limitations of this work.

5.1 What kind of code-switching occurs in the classroom?

The present study shows some interesting divergence from previous works, in terms of what code-switching were used. An example of this is that Sampson (2011) found it rare that the L1 was used to avoid using the L2, while this investigation found L2 avoidance to be more

frequent. This difference could stem from the difference in methodology between the studies. However, there is also the possibility that my definition of the categories diverges from that of Sampson (2011). Another finding was that students likely use socializing when they are tired, which fits well with observation 3. Another point of divergence is that the most common code shift in Sampson’s (2011: p. 247) study was equivalence, which does not have an especially prevalent placement in the present investigation. The equivalence code-switching tended to be used when discussing specific things, places, organizations etc. While paraphrases could be used in place of the equivalence code-switching, they would likely be more difficult to understand. Using the L1 instead increases fluency, at the expense of the students practicing accuracy in English, which S’ad Qadermazi (2015: p. 171) mention as a dilemma. It is difficult however, to reach any broader conclusions from these numbers since substantial differences exists in methodology between Sampson (2011) and the present study. However, equivalence is clearly a common code switch in this study as well. Something that caused more surprise was the rare occurrence of floor-holding as opposed to Sampson’s (2011: p. 245) study, in which it occurred with relative frequency. This is one of the clearest findings in this study and raises a number of questions. The clear disparity could be due to a difference in how the categories were coded, since there is inevitably some subjectivity in the

(26)

which case it may be of interest to future research, preferably with a larger cohort and a rigorous method. Lin (2013: p. 212) suggests that a study conducted by the ordinary teacher, following their own students, as well as studying their own practice, may be a fruitful effort for how such future research can be conducted. There is also a great deal more socializing, as well as metalanguage as opposed to the relatively intermediate placement of those categories in Sampson’s (2011: p. 302) observations. Socializing was also expected to be a prominent category and it proved to be so in a majority of observations. The level of socialization seemed to vary between lessons, likely depending on the lesson structure and the time of day, since socializing was more rare in the shortest observation i.e. observation 4. This differed from previous research however, since socializing was not more common than other switches in Sampson’s (2011) study.

5.2 Integration of Swedish in teaching and pedagogical reflections of the

teachers

(27)

“English only” by teacher 2 was not expected, however, and the fact that it had been tried, but proven ineffective is an interesting observation. Moving on, teacher 3 used Swedish as a tool, specifically to discuss Swedish grammar, as juxtaposed to English grammar. This clear and deliberate integration provided an interesting example of how switches can be used in a thematic manner. This choice resonates with Levine’s formulation that teachers should “explore the dynamics of their own situations” and make choices based on local

circumstances. Teacher 3’s strategic use of Swedish, based on a parallel to a Swedish lesson, is a clear example of such a choice, with a clear purpose (Levine, 2011: p. 35).

.An interesting finding in this study is the continued use of English by the teacher even when the students are misbehaving. Fägersten’s (2015: pp. 15-17) study of lower secondary

students showed that teachers at the lower level tend to code-switch to Swedish when students misbehave. However, the observed classes rarely needed correction. This discrepancy may be due to the difference in level between upper and lower secondary school, but since the cohort is small, further research would be needed to confirm this. Another possibility is that upper secondary students are more affected by the observer and behave better when observed. This effect, often called “the observers paradox”, has been considered in this study (see section 3.1.2). However, it is almost impossible to entirely avoid its effects.

Moving on, the teachers claimed to use Swedish primarily in order to clarify concepts or repeat instructions. This fits well in with conclusions from earlier research, such as Bensen (2013: p. 80). In the interview, teacher 1 expressed a desire to counteract “bad English” with “good English”, which likely contributed to the teacher’s avoidance of Swedish. The teacher argued that the English in mass media is usually of poor quality and that it should be

counteracted by the teacher’s correct use of English. Teachers 2 & 3 seemed to put more emphasis on the flow of the lesson and put less emphasis on “correctness”, than teacher 1. Levine (2011: pp. 10-13) argues that language learning should not only focus on teaching language as a skill that is tied to rules and vocabulary. Instead he argues that students should engage more with other languages and contrast and compare to a greater extent than the mainstream suggests. While teacher 1 did express sympathies for the more rule-based view of language, there were also some examples of teacher 1 applying a multilingual approach to his view of language, such as his suggestions that comparisons to the student L1’s other than Swedish could be used. Teacher 2 and 3 very clearly argued for a multilingual approach, with more frequent code-switching than e.g. Teacher 1 and they used it in practice, as well.

(28)

may contribute to better critical thinking (Levine, 2011: p. 13). In conclusion, the teachers all viewed English teaching from a multilingual perspective, but their individual teaching style differed based on circumstances.

5.3 Methodological discussion

The use of an observational method led to a work intense study, led to interesting findings, but also had some drawbacks. Firstly, the method of recording could have been improved by a focus on teacher or students interactions. The current approach caused some lacking focus, since notes were taken in two categories simultaneously. As discussed in section 3.4 some additional preparations with regards to the ethical considerations could also have been made. The framework developed by Eldridge (1996) was used in order to provide structure to the observations (Dörnyei, 2007: pp. 180-181). Additionally, using an earlier scheme contributes validity, since the scheme has been tested and showed useful results (Sampson, 2011: p. 302). However, it could have been improved upon, since the category floor-holding was almost unused, and metalanguage may have been too wide of a category. However, with the time allotted, the prioritizing of reliable categories worked well. The results presented here could be an interesting area of research for researchers interested in the field.

Regarding the interviews with the teachers, there were some interesting findings. However, the interviews could have been improved by having more clearly defined questions. This would have allowed for efficiency and clarity. The interviews could also have been recorded, which could have been facilitated. Since the interviews were rather short, the note taking approach works in any case.

5.4 Limitations

(29)

6 Conclusions and implications for teacher practice

This investigation has found that code-switching and overall L1 use occurs in classrooms even at upper secondary level. It is difficult to generalize the information gathered, since the results are not consistent from observation to observation. However, the study shows that teachers do not generally react when their students use Swedish, even if they claim to have English as the language that should be used in class. The study also shows the prevalence of socializing, as well as the clear absence of floor-holding code-switches. The teachers hold a generally reserved view of code-switching, but their degree of usage varied.

This study provides a window into when and why students and teachers switch languages. It is important to know why students switch, so that any corrections do not impact their

language fluency. The results also show how experienced teachers deal with issues of code-switching and related issues in the classroom. While this essay cannot conclusively judge the efficiency of their strategies, their viewpoints could nonetheless be considered. This study highlights the varying approaches used by teachers depending on the classroom, content and other factors. Collegial discussions on how to best use English and Swedish based on these factors is therefore important in order to find working strategies for that particular school, especially in the light of Sweden’s shifting demographic. The curriculum states clearly that multilingualism should be encouraged and the teacher should work towards meaningful interactions between these languages (Natl. ag. f. ed., 2011: p. 1).The argument that research is often too academic and inaccessible to teachers is oft repeated, but it bears repeating again. In order to avoid this disconnect in future research Lin (2013: p. 212) suggests that one should work closer to teachers and take a more in depth look at phenomena like code-switching. This essay has arrived at some interesting conclusions, but as evident, there were also some issues that prevented further findings. Firstly this applies to how the research methodology worked with differing results in classrooms. The formulations of the research questions and the subsequent study based on them resulted in a broad examination of code-switching in the classroom. It provides a comprehensive, in-depth view of the activity in the classroom, as well as the diversity of code-switches used, but it may inhibit more precise conclusions. To

(30)

Reference list

Bensen, Hanife & Çavuşoğlu, Çise, 2013. Reasons for the teachers’ uses of code-switching in

adult EFL classrooms. Hasan Âli Yücel Eğitim Fakültesi Dergisi, Hasan Âli Yücel Eğitim

Fakültesi Dergisi; vol. 10, iss. 2 : 20.pages; pp. 69–82, 2013. Retrieved from

http://proxy.lnu.se/login?url=https://search-proquest-com.proxy.lnu.se/docview/1491096275?accountid=14827 (Accessed 23/3 2019)

Diley & Thierney. Interviewing in education pp. 453-73, Gubrium, J.F. & Holstein, J.A., 2001. Handbook of interview research : context & method, Thousand Oaks, Calif.: SAGE Dörnyei, Z., 2007. Research methods in applied linguistics: quantitative, qualitative, and

mixed methodologies, Oxford University Press.

Eldridge, J.; Code-switching in a Turkish secondary school, ELT Journal, Volume 50, Issue 4, 1 October 1996, Pages 303–311, https://doi.org/10.1093/elt/50.4.303

Eriksson, S. (2019). CODEX - regler och riktlinjer för forskning. Codex.vr.se. Available at: http://www.codex.vr.se/index.shtml [Accessed 5 Feb. 2019].

Esaiasson, P. et al., 2017. Metodpraktikan : konsten att studera samhälle, individ och

marknad fifth edition., Stockholm: Wolters Kluwer.

Francis M. Hult (2017) More than a lingua franca: Functions of English in a globalised

educational language policy, Language, Culture and Curriculum, 30:3, 265-282, DOI: 10.1080/07908318.2017.1321008

Fägersten, Kristy. (2012). Teacher discourse and code choice in a Swedish EFL classroom,

chers' Roles in Second Language Learning: Classroom Applications of Sociocultural Theory,

Teacher discourse and code choice in a Swedish EFL classroom, Information Age Publishing, Editors: B. Yoon, H. Kim, pp.81-98. Available at:

https://www.researchgate.net/profile/Kristy_Faegersten/publication/282731115_Teacher_disc ourse_and_code_choice_in_a_Swedish_EFL_classroom/links/561a39d608ae6d17308962a9.p df (Accessed 5 February 2019)

Källkvist, Marie et al., 2017. English Only in Multilingual Classrooms? Lms - Lingua, 2017(4), pp.27–31. Available at:

https://portal.research.lu.se/ws/files/35428089/English_Only..._Lingua_4_17_.pdf (Accessed 12/2 2019)

Levine, G.S., 2011. Code choice in the language classroom, Multilingual Matters. Channel view publications: Bristol Available at:

https://ebookcentral-proquest-com.proxy.lnu.se/lib/linne-ebooks/detail.action?docID=718009 (Accessed 12/2 2019) Lin, A., 2013. Classroom code-switching: three decades of research. Applied Linguistics Review, 4(1), pp.195–218.

National Agency for Education. ”Syllabus English 6”. 2013. Available at:

(31)

Nilep, Chad (2006) "“Code Switching” in Sociocultural Linguistics," Colorado Research in Linguistics: Vol. 19. https://doi.org/10.25810/hnq4-jv62 Available at:

https://scholar.colorado.edu/cril/vol19/iss1/1 (Accessed 5 February 2019)

Nilsson, Rickard, Teachers' codeswitching to L1 in the EFL classroom: A comparative study

of a Chinese and Swedish teachers' beliefs and practices, Karlstad University, Faculty of Arts

and Social Sciences (starting 2013), Department of Language, Literature and Intercultural Studies. 2016

Sa'd, Seyyed Hatam Tamimi, and Qadermazi, Zohre. “L1 Use in EFL Classes with

English-Only Policy: Insights from Triangulated Data.” Center for Educational Policy Studies

Journal, vol. 5, no. 2, 2015, pp. 159–175. Available at:

https://doaj.org/article/08256a1667a0454b8dab5b1bece03e7c (Accessed 1/12 2018)

Sampson, Andrew. “Learner Code-Switching versus English Only.” ELT Journal, vol. 66, no. 3, 2012, pp. 293–303. Available at:

(32)

Appendices

Appendix 1 Observation protocol

3_{Bifunctional interactions will be marked with an asterisk ( i.e. *)}

Classroom description

Types of

language use

Categorization of code-switching in the classroom

(33)

Appendix 2 Post-observation interview questions

Fråga 1: När använder du svenska i undervisningen? Varför görs språkbytet?

(Translation) Question 1: When do you decide to switch to Swedish in an educational situation? Why do you switch language?

Fråga 2: Var lektionen typisk gällande användning av Engelska, respektive svenska? Om skillnader finns, hur brukar språkanvändandet se ut?

(Translation ) Question 2: Was this lesson typical of the language use of a regular lesson? If not, in what way did it differ?

Fråga 3: Har elever i klassrummet andra första språk än svenska? Påverkar detta hur svenska och engelska används i klassrummet?

(Translation) Question 3: Are there students with other L1: s in the classroom ? Does that affect the use of Swedish?

Fråga 4: Används svenska som ett verktyg i engelskaundervisningen? På vilket sätt används svenska?

(Translation) Question 4: Is Swedish used as an educational tool in the classroom? How is Swedish used?

Fråga 5: Anpassar du språket du använder till undervisningsnivån? T.ex. används mer svenska i klasser med mindre förkunskaper?

(34)

Appendix 3 Deltagande i observationsstudie

Studiens huvudsakliga syfte är att studera förekomsten av engelska och svenska i

engelskaundervisningen på gymnasienivå. Detta görs för att identifiera vilka funktioner svenska, respektive engelska, har i klassrummet. Målet med studien är alltså att ge en överblick över hur klassrummens användning av engelska, respektive svenska, ser ut i praktiken.

Deltagande i denna studie är frivilligt och kan avbrytas när som helst. Detta innebär att du har rätt att självständigt bestämma om och hur länge du önskar medverka. Du kommer att kunna vara helt anonym i studien och alla uppgifter som används i studien kommer att anonymiseras. Under mina observationer kommer jag att spela in ljud under delar av lektionen för att få en helhetbild av kommunikationen. Ljudfilerna kommer att bevaras säkert och efter studien kommer de att raderas. Dessutom kommer alla insamlade anteckningar under mina observationer bearbetas och undersökas, så att inga deltagare kommer att kunna identifieras. Ditt och elevernas deltagande kommer alltså vara helt anonymt. De uppgifter som insamlas kommer inte användas för annat syfte än forskning och slutresultatet av underökningen kommer efter studiens genomförande att publiceras på databasen Diva. Observationen genomförs för att få en bild av hur ett vanligt klassrum fungerar gällande användandet av engelska, samt svenska. Syftet är inte att bedöma vare sig elever eller lärare. Det är därför viktigt att lektionen inte avviker i stor grad från en vanlig lektion. Tack så mycket för ditt deltagande!

Härmed intygas att jag väljer att delta i denna studie och att observationerna får användas som material för analys. Jag har erhållit skriftlig information angående observationsstudiens syfte samt läst och accepterat ovanstående information.

(35)

Appendix 4 Observation i ditt klassrum

Hej!

Mitt namn är Johan Westin och jag kommer att besöka ett antal lektioner i engelska på din skola. Jag skriver en uppsats som handlar om hur engelska och svenska används i engelskaklassrummet. Jag kommer därför att observera lektioner i några skolor för att se hur elever och lärare använder engelska och svenska.

Deltagande i denna studie är självklart frivilligt och kan avbrytas när som helst. Detta innebär att du har rätt att självständigt bestämma om du vill medverka. Du kommer att kunna vara helt anonym i studien och alla uppgifter som används i studien kommer att anonymiseras. Jag kommer inte ens att skriva vilka skolor jag besökt. Under mina besök kommer jag att spela in ljud under delar av lektionen för att kunna analysera lektionen senare. Ljudfilerna kommer att bevaras säkert och efter studien kommer de att raderas. Dessutom kommer alla insamlade anteckningar under mina observationer bearbetas och undersökas, så att inga deltagare kommer att kunna identifieras. Ditt deltagande kommer alltså vara helt anonymt. De uppgifter som insamlas kommer bara att användas i min uppsats

slutresultatet av underökningen kommer efter studiens genomförande att publiceras på databasen Diva. Diva är en databas för forskning och studentuppsatser.

Observationen genomförs för att få en bild av hur ett vanligt klassrum fungerar. Syftet är inte att bedöma vare sig elever eller lärare och mina observationer kommer inte att ha någon påverkan på ditt betyg. Jag kommer inte att delta i undervisningen och kommer endast sitta och göra anteckningar. Ni behöver alltså inte ändra på något under lektionen utan ni kan ha lektion precis som vanligt. Har ni några frågor kan ni fråga efter observationerna eller på mejladressen nedan.

Tack så mycket för ditt deltagande!

(36)

Appendix 5 Interview excerpts in Swedish

1. ”Det är viktigt att eleverna har engelska som sitt arbetsspråk. Det underlättar för dem att bli bekväma med att använda engelska, samt att de får mängdträning”. [Teacher 1 Interview]

2. ”Boken som används är kontrastiv och exempel från både engelska och svenska används, särskilt när vi pratar om grammatik” [Teacher 2 Interview].

Why not “ English only ” ? Bachelor’s thesis