Functions of Teacher and Student Code-Switching in an EFL Classroom and Pedagogical Focus : Observations and Implications

Code-switching (CS), the alternation between two or more languages within a stretch of language, is accepted as a valuable strategy of bilinguals in making linguistic choices for communicative purposes. Following different perspectives, CS can be understood either as an attempt to communicate meanings at the macro level (such as identity, solidarity etc.), or to convey intended meaning to the listener within the boundaries of conversational interaction at the micro level. Accepted as a strategy of speakers in bilingual communities, CS in foreign language learning settings is still a contentious matter. However, its potential has been emphasized in recent studies. This small-scale study examines teacher and student CS occurrences in an EFL lesson at a Turkish state university. CS occurrences were transcribed and sample extracts were analyzed through conversational analysis. The results show that teachers and learners apply CS to generate access to language or as a tool in classroom management. This study also reveals that CS can be a learner strategy to avoid L2 when lesson content is of little relevance to learners. In such cases CS cannot fulfil its potential as a means in discourse strategy, and language learning is unlikely to be facilitated.


Introduction
Code-switching (CS) refers to a situation in which a speaker changes between two or more languages within a conversation at inter-sentential or intra-sentential level (Hamers & Blanc, 2004).It is frequently observed in bilingual communities.Even though participants in CS are mostly unaware of the code choice at any given moment, their linguistic choice is not considered arbitrary, but driven by an intended communicative effect (Gumperz, 1982, p.61).The following sections elaborate on CS as a phenomenon in bilingual settings and CS occurring in the foreign language classroom.

Code-switching in bilingual settings
Basically two approaches account for CS in bilingual communities.Following a macrosociolinguistic or 'universalist' approach, social motivation accounts for speakers' choices of linguistic codes.Accordingly, a speaker's CS is related to social and structural factors, such as identity, solidarity, gender, prestige (Myers-Scotton, 1993).A second approach emphasizes the function of CS to address situated concerns bilingual speakers are faced with in the specific context of conversational interaction.Therefore, CS emerges as an outcome of sequential development in talk between bilinguals, i.e. it is a tool in "conversational activity (...) within specific conversational contexts" (Wei, 2005a, p.276).
The universalist account is an identity-related attempt to describe CS within the markedness model, which claims that a speaker's choice is driven by the desire to apply a set of rights and obligations, which refers "to what participants can expect in any given interaction type in their community" (Myers-Scotton, 1998, p.23).Bilingual speakers know which code the expected (i.e.unmarked) code is in a specific contextual interaction; this means they know the indexical value of a code in a bilingual (or multilingual) setting.Then, they carry CS out, either to comply with the unmarked set of rights and obligations, or to impose a different set.Accordingly, CS is an attempt to accomplish communicative purposes by selecting the appropriate socio-psycholinguistic behavior (cf.Harmers & Blanc, 2004, p.268).
From a localist perspective, CS is understood as a strategy in conversational management.Within this framework, sets of functions in conversational CS have been suggested (e.g. Baker, 2000;Auer, 1995), and CS is -at the microstructural level -a bilingual's strategy to exert an effect on the listener during the collaborative attempt of conversation participants to achieve mutual understanding.This means, CS cannot be accounted for without reference to the specific situational context but needs to be examined by means of conversation and discourse analysis (Ustunel & Seedhouse, 2005;Wei, 2002).
Both models, are in fact, main approaches with different contributors setting different emphases (Wei 2005a), and seem to be adversary; it is, however, reasonable to see them as complementary, shedding light on the same phenomenon from different perspectives (Wei, 2005b).To sum up: the first approach is directed to effects of macro level factors like identity, gender or ethnicity on CS, while the second one examines CS at the micro level analyzing conversational factors and discourse management applying conversation analysis.
From whatever perspective CS is examined, there is a common understanding that CS enhances the speaker's linguistic available repertoire, even if it is known that negative attitudes towards code-switching exist when, for example, a language is less prestigious or EDUPIJ / VOLUME 1 / ISSUE 1-2 / SPRING-SUMMER~FALL-WINTER / 2012 unrequested for political reasons in a community.As any other linguistic choice, either conscious or subconscious, CS aims at providing an utterance with an appropriate form so that it can fulfil its function optimally in a given context (Holmes, 2008).

Code-switching in the FL classroom
Different from CS in bilingual communities, CS in the foreign language (FL) classroom does not have the positive image of being "an asset and a valuable addition to their [the bilinguals'] array of communicative strategies" (Macaro, 2005, p.63).It is often not desired by educators because it is assumed that code-switching crowds out the target language and therefore has detrimental effects on the learning process (Littlewood & Yu, 2011).Macaro (2005) summarizes the arguments against CS in the FL classroom under five aspects.First, CS means a recourse to L1 which is not desirable because the L2 classroom should aim at maximizing L2 input (given by the teacher) and L2 output (produced by the learners); this is regarded essential in second language acquisition.Second, avoidance of CS warrants the absence of negative interference and follows the maxims of comprehensible input and meaning negotiation, which are also regarded as effective for language learning.Third, the avoidance of CS impedes a return to unfashionable methods such as the grammar translation method, which is seen as an example of a method in which the target language is crowded out and L2 is taught without regard to language use in the 'real world'.Fourth, CS is politically not desired -the background of this claim is the (alleged) wish to employ native speaker teachers for political and economic reasons, i.e. a sort of linguistic imperialism (Phillipson, 1992) is served.Finally, since it is supported by official 'English-only' language teaching policies established by national agencies, avoidance of CS is regarded as an indicator for 'good teaching', i.e. has a control function.
Such assumptions have been criticized severely for several reasons.From a general perspective, Macaro (2005) argued that CS should be allowed in the FL classroom since modern FL approaches and methodologies postulate that the FL classroom should mirror the language use of the 'real world', and CS is an occurrence of the 'real world'.Even though this argument is plausible, it could be objected that in this view language in foreign language classrooms, particularly teacher talk, is regarded somehow artificial, not as naturalistic as language in the outside world.If we accept, however, the FL classroom as "a social context in its own right" (Walsh, 2002, p.4), then teacher and learner talk with a seemingly inevitable occurrence of CS that appears as naturalistic language choice in a given context.In other words, the FL classroom is one discourse setting among others with a specific linguistic choice, and discourse in the FL classroom contains CS due to needs to interact, to initiate or keep communication flowing in the classroom.CS, then, appears to be a tool in FL classroom discourse management as a conversational lubricant (Butzkamm, 1998).
Apart from justifying CS in the FL classroom as a means of discourse strategy, methodologists have offered ways of L1 application in the FL classroom as teaching aids (Cook, 2010;Butzkamm & Caldwell, 2009;Deller & Rinvolucri, 2002;cf. Littlewood & Yu, 2011).Such techniques are used explicitly to achieve language learning goals through crosslinguistic comparison (denoted as strategic use of L1 in contrast to compensatory use by Littlewood & Yu, 2011).Additionally, it has been demonstrated that a bilingual approach in FL teaching and learning allows a dual understanding of both meaning and form, furthers the development of linguistic awareness in learners and leads to an enrichment of activities available in the classroom (Sampson, 2011;Butzkamm & Caldwell, 2009;Macaro, 2005).The next section introduces categorizations of CS in the foreign language classroom.

Categorizing code-switching in the FL classroom
In the previous section, it was argued that the FL classroom is a discourse setting with a specific linguistic choice.Ustunel and Seedhouse (2005) define this setting more precisely as one being characterized by institutional discourse having the overall institutional goal to teach learners a FL.This means that language choice has to be checked against these overall goals.The insight that the FL classroom is a kind of institutional discourse has consequences for categorizing and establishing functions of CS in FL classrooms; those functions are reasonably related to functions of CS in other settings, but have a strong pedagogical focus due to the overall goal.Furthermore, due to different roles that learners and teachers play in educational settings, it is possible to distinguish functions/categories in learner CS from those of teacher CS.Two seminal studies established functions and categories of CS in the foreign language classroom.
The first study on CS reported here was conducted by Eldridge (1996) in a Turkish secondary school.The following specific functions of student initiated CS in the FL classroom were proposed:  equivalence: use or elicitation of equivalent item;  floor-holding: wishing to continue without interruption;  metalanguage: talking about the language or task;  reiteration: reinforcement, emphasis;  group membership: socializing, establishing group identity;  conflict control: avoiding potential conflicts;  alignment/disalignment: assuming the other person in the same/in a different position.
In his study, Eldridge (1996) concedes that those functions are not unproblematic in CS analysis: "The main problem in analyzing code-switching in functional terms is that many switches may be either multi-functional, or open to different functional interpretations" (p.305).
In a further study, conducted in a preparatory program at a Turkish university, Ustunel and Seedhouse (2005) offered, with regard to previous studies, three categories of teacher initiated CS in the FL classroom:  CS for curriculum access: CS is carried out to provide learners access to language;  CS for classroom management: CS is carried out as an attention-focusing device to motivate, discipline or praise learners;  CS for interpersonal relations: CS is carried out to personalize and humanize the classroom by addressing affective factors.
This study aimed at contributing to the understanding of CS in the foreign language classroom by analyzing teacher and learner CS emerging in an EFL (English as a FL) class at a Turkish state university.Through identifying and analyzing conversational functions of CS occurrences, it was sought to evaluate if changes from the target language English to Turkish would match the pedagogical focus of FL instruction.

Methodology
For this study, a video-recorded English lesson carried out at a Turkish state university was used.The lesson took 59 minutes.The lesson was attended by 15 young adult EFL learners (eight males and seven females; mean age: 18y 6m), who were at level A2 according to the Common European Framework of Reference for Languages (Council of Europe, 2001).The students were enrolled in an English preparatory program, with 28 hours of English lessons per week.
The lesson, which took 59 minutes, was video-recorded and then transcribed (see the Appendix for transcription conventions).However, only participants' distinctly recognizable speech was transcribed.There were probably more cases of CS but they were not recognizable because students either spoke simultaneously or to partners, or background noise made parts of the recording inaudible.The clearly recognizable CS occurrences were counted and, if possible, matched to the functions and categories suggested by Eldridge (1996) for learner CS occurrences and Ustunel and Seedhouse (2005) for teacher CS occurrences.
Thus, rather than giving a complete catalogue of all CS occurrences observed in the lesson, this article tries to analyze CS occurrences and to evaluate them according to their pedagogic value.In doing so, the study follows a procedure suggested by Ustunel and Seedhouse (2005).They apply conversation analysis, a flexible procedure that seeks to detect "the evolving relationship between pedagogy and interaction" (p.309) in CS occurrences by paying attention to the sequential structure of an interaction.Turns within an interaction reveal the speaker's understanding of the prior turn; furthermore language choice and indicators such as hesitation or delay show if participants in a conversation agree or disagree with the first part of an adjacency pair: responses are affiliative or disaffiliative.Since we deal with the case of educational discourse, learners' linguistic choices might be interpreted as an indicator of degree of affiliation to the pedagogical focus.Ustunel and Seedhouse (2005) summarize the methodological procedure as follows (p.310): "Another way of presenting the principles of CA is in relation to the question which it asks.The classic question which we must ask at all stages of a CA analysis of data is: "Why that, right now?" Since this study focuses on CS, we modify this question to: "Why that, in that language, right now?"This encapsulates the perspective of interaction as social/pedagogical action ("why that") which is expressed in a particular language in a developing sequence ("right now")."Following this methodological approach, sample extracts of the data set were analyzed.

Findings
A total of 85 learner code-switches and 44 teacher code-switches could be accurately identified in the data.Learners made most frequent use of CS for equivalence (35 cases), to talk about metalanguage and reiteration (10 occurrences each), while there were no examples for floor holding.Notably, the remark of Eldridge (1996) that assigning CS behaviors to functions is a problematic issue, was confirmed in this study.The overwhelming majority of teacher CS serves to provide curriculum access (35 cases), while CS used for classroom management or related to interpersonal relations occurred five and six times, respectively.One occurrence was attributed to the last two categories mentioned.In the following sections, sample extracts will be analyzed and discussed.

Sample Extracts Extract 1
In extract 1 the teacher is checking sentences in a course book exercise.The CS in turn 1 initiated by the teacher is a frequently observed CS to elicit classroom participation.Although the teacher raises her hand to support her request for hand-raising as a visual classroom procedure, she decides to switch to L1 since the students are looking at the course book, as the video recording reveals.The CS by student 1 in turn 2 is caused by the self-interruption, probably by the insight that speaking without permission might be impolite since the teacher requested hand-raising; the respectful form of addressing hocam supports this interpretation.Not only with regard to the data observed for this study, it can be said that hocam is one of the most frequent switches used by Turkish learners, differently from the example discussed here frequently used to initiate moves (cf.Arikdal, 2006).

Extract 2
In the second example teacher CS is also used for eliciting participation; in the further course of interaction, CS is used to talk about metalanguage.During turn 1, the teacher writes the sentence pair 'They are my friends.this is my friends' house T writes phrase on the board (5 sec) burada bir tamamlama var [tr: there is a modified noun here] (.) bu sizin için bir ipucu [tr: this is a hint for you] (.) this is a hint (.) this is a hint (3 sec) my friends (.) singular or plural?9 S3 plural 10 T plural (.) tekil olsaydı nasıl yazacaktım?[tr: if it were singular how would I have written it?]T writes 'friends'' on the board the first one is singular (.) the second one is plural The teacher switches in turn 8 and 10 from English to Turkish because the metalinguistic explanation in L2 would probably be above the students' level.The translation 'this is a hint' is reiterated but does not elicit a reaction by the students.The pause (3 seconds) might indicate that the teacher expects here a learner utterance.However, she then goes on signaling another aspect by addressee specification ('my friends').In turn 10, the Turkish utterance is not translated.Even though a pedagogical focus is obvious, the strategy applied here does not lead to learner activation.[Turkish pizza (.) best of Turkish pizza While in turn 2 and 4 CS occurrences elicited are given by the teacher, turns 6 and 10 contain those ones which are not requested.Turn 6, which overlaps with the question beginning in turn 5, distracts the teacher from asking the question; the question is then modified to an imperative in turn 9, but student 2 shows disaffiliation by not responding.The following terms reveal a rather uncontrolled calling out of Turkish foods; this and the simultaneous speech in turn 19 demonstrate that the students cannot recognize a pedagogical focus to relate on.

Extract 4
An example of referring to metalanguage initiated by a learner is given in the next extract.The teacher is doing an exercise where students have to find the stressed syllable in a word.Apart from the CS of student 1, the teacher CS in turn 3 is remarkable: since the students did not understand the pedagogical focus in turn 1 (to name the stress), the teacher selects CS as a different strategy which is successful as turn 4 shows.
EDUPIJ / VOLUME 1 / ISSUE 1-2 / SPRING-SUMMER~FALL-WINTER / 2012 passiveness and the repeated responding with CS show remarkable disaffiliation with the teacher, the pedagogical focus and possibly with the content of the lesson.The teacher, thus, applies a new strategy in nominating a good student, Gökhan, in turn 9. His answer in Turkish (turn 10) provokes a strong reaction in S5, whose linguistic choice can be interpreted for sure as marked.It can be assumed that cultural issues (Santa Claus as the representative of a different culture) play a role when the classroom interaction now livens up.In turn 12, the teacher's intra-sentential CS creates an interpersonal relation to student 5, who now contributes using CS in the form of reiteration in turns 13 and 15.Remarkably, in turn 13 the English phrase is reiterated and in 15 the Turkish one through CS.The conversation now takes place between the teacher and student 5.With turn 18, which renders turn 16 but is related to student 5's irrelevant answer in turn 17, student 4 takes turn; summoned by student 6 ('jb' stands for Jim Beam, a whiskey brand as the continuation not documented here reveals), student 4 selects himself using CS at intra-sentential level; apart from the fact that 'eat' is mistakenly used for 'food', the two codes are connected grammatically correct as it is the case in turn 24, where the correct inflection ending added to 'meat' is distinctly intelligible in the recording.Turn 24 as well as 28 are driven by affective factors.The linguistic choice reflects the fact that the focus is not on language learning.Turn 24 functions as in-group identity marker, through which student 4 generates solidarity with the other students, while the teacher emphasizes her disapproval through her sarcastic remarks in turn 28; these remarks would not be comprehensible for the students due to their level if they were uttered in English.The following switch to English ('meat (.) what else Ömer') then shall establish normality -it is an attempt to return to the business of institutional discourse.

Conclusion and Discussion
CS in the FL classroom, as a peculiarity in the wide field of CS, is signified by its own specific features.It is located in the area of conflict between the institutional overall goal -to teach a foreign language -which suggests that L1 should be minimized, and the perception that CS is a natural tendency in bilingual settings with the potential to make conversations run.Conversation in the classroom -classroom interaction -has been clearly identified in its importance for language learning through the insights of social constructivism (Walsh, 2003).Therefore, it is essential to evaluate CS in the FL classroom from the perspective if it contributes to language learning or impedes it.
Similar to the findings of Ustunel and Seedhouse (2005), the teacher-initiated CS in this study aimed to provide learner access to language or to direct their attention to lesson content.However, as illustrated in the extracts, this did not always lead to enhanced classroom interaction.The data analyzed in this study showed that students showed affiliation when they were asked to answer in L1.Responses in L2, however were infrequent, halting and did not go beyond phrase level.Responses in L1 when L2 was expected were frequent, which signals that students showed disaffiliation with the teacher's pedagogical goals.A remarkable amount of the instances of student CS were attempts to avoid speaking in L2 about a topic which had no relevance to them.Extract 5 clearly supports this perception.It seems plausible to assume that the lesson content (New Year) was particularly of little relevance to the students due to cultural distance.Thus, the topic should have been approached differently, e.g.under the aspect of getting to know a special occasion in a foreign culture, instead of asking the student persistently how they celebrated New Year (an occasion that most of the students is at least indifferent to).
EDUPIJ / VOLUME 1 / ISSUE 1-2 / SPRING-SUMMER~FALL-WINTER / 2012 With an awareness of the restricted dataset which does not easily allow for generalization, the results of this study suggest that CS in the FL classroom can be a strategy to avoid communication about lesson content that has little or no relevance.Since it is the aim of FL teaching to have students actively participate in the classroom in order to facilitate learning, the choice of lesson content is of paramount importance.Consequently, CS as a strategy to avoid L2 is not desirable as it contradicts pedagogical foci.
As an implication, teachers should be aware of the effect of teacher CS and modify their own CS behavior accordingly (Sert, 2005;Ustunel & Seedhouse, 2005).For developing awareness, teachers should observe their own teaching through peer observation, video-or audio-recording (Walsh, 2002;Walsh 2006;Harfitt, 2008) and find out about the reflected use of the mother tongue in the related literature (e.g.Butzkamm & Caldwell, 2009).It seems equally important that teachers provide learners with sufficient opportunities to rehearse speech about learner-relevant subject contents during classroom time, particularly in group or pair work where affective filters are lowered.Under such circumstances, CS can fulfil its potential as a language learning aid and means of discourse management in the FL classroom (Sampson, 2011;Eldridge, 1996).
nerede? [tr: where is the stress?](.) afternoon 4 S1 o'da o'da [tr: on the o, on the o] They are my friends'.on the board: As stated above, students most frequently used CS to give equivalence as at the beginning of extract 3: EDUPIJ / VOLUME 1 / ISSUE 1-2 / SPRING-SUMMER~FALL-WINTER / 2012Extract 3