“I’m a Learner, I’m Dominantly a Learner”: A Case Study of Code-Switching of an EFL University Teacher Outside Classrooms

Reported here is part of a bigger study to investigate the code-switching behavior among the foreign language teaching professionals. This study is informed by the Markedness Model and identity-inInteraction approach with its focus on the social and discourse motivations of code-switching. An ethnographic design is used to collect instances of code-switching of a university EFL teacher outside his classrooms. The findings suggest the code-switching behavior is an important site of their professional identity construction. Large chunks of switching to English in daily communications with peers are used as a strategy for sustained professional development. The positive attitude to code switching shows the high status of the English language and the profession of EFL in China. In the global context, the frequent code-switching to English in an academic setting in Beijing can be seen as an indication of the hegemonic position of the English language in the world academic discourse.


Introduction
A large number of native English speaking teachers in China teach in the private sectors, while the majority of the English teachers teaching in public schools and public universities are Chinese nationals and native Chinese speakers who have been trained in English programs in Chinese universities. The introduction of Communicative Language Teaching in China since the 1990s and Task-based teaching in the new century has pushed English language teaching professionals to change their approach to teaching from grammar-focused knowledge transmission mostly conducted in L1 Chinese to meaning-focused communicative teaching conducted in the target language of English. Though the communicative target-language-only approach to English teaching in the English as a foreign language (EFL) context (for example, China) has been seen by some scholars as part of "Linguistic Imperialism" (Phillipson 1992) or "Nativespeakerism" (Holliday 2005) that position English as a superior language to other languages, the principle of communicative language teaching is widely accepted as a language teaching approach conducive to the development of students' more communicative abilities. However, L1 Chinese is still selectively used by EFL teachers in the English classes as a teaching strategy. Some studies have also pointed to the pedagogical necessity for EFL teachers to code-switch from English to Chinese in the classroom setting when the teacher is a bilingual and the students are a group of homogenous Chinese speakers (e. g. Auer, 1998;Chen, 1996;Qian, Tian & Wang, 2009). Thus, code-switching in the Chinese EFL community is mostly considered a pure pedagogical issue, a teaching strategy utilized by EFL teachers in the classroom to maximize teaching effectiveness in relation to different teaching goals.
However, the phenomenon of code-switching for EFL professionals does not seem to end in the classroom. Many of them are found to engage in code-switching outside the classroom in their professional work and life setting and beyond (e. g. Gu 2010). Language is a window to the human mind as well as a window into the complexity of the human society we live in. As an interesting linguistic phenomenon, the code-switching among bilinguals has proved to be an important area of research for both cognitively oriented and socially oriented researchers. Through an ethnographic case study of one EFL professional, an overt code-switcher between English as a foreign language and Chinese as his first language, this study is intended to explore the relationship between the EFL professional's code-switching behaviors with his professional identity, including his motivation for and attitudes towards his code-switching, plus the meaning behind his code-switching behavior in his larger professional and life context beyond the classroom. It will have significance in understanding EFL professionals' language alternating behavior and the relations between language behavior and identity work. Hence, individual professionals working with a foreign language as dominant as English, as well as foreign language learners, and even language policy makers have to be prudent when setting certain goals in language learning, because the process and result of learning and using another language is also involved with certain identity work.

Literature a Review
Code-switching is a term used to refer to "the alternative use by bilinguals (or multilinguals) of two or more languages in the same conversation" (Milroy & Muysken, 1995, p. 7). Code-switching between Chinese and English occurs frequently in recent years. Chinese words sprinkled with English are prevalent in advertisement, popular magazines, pop songs, and so on. Code-switching is also common among EFL professionals when speaking to students, colleagues and friends.
Cashman (2008) distinguishes three approaches to codeswitching research: the symbolic approach, the sequential approach and the identity-in-interaction approach. The symbolic approach aims to explain code-switching in terms of communities' socio-historical structures and speakers' rights and obligations as members of those communities, as represented by the Markedness Model by Myers-Scotton. It mainly answers questions like: why speakers choose to code-switch as a marked choice and what changes in external factors result in code-switching as unmarked choice (Cashman, 2008). For example, the type of interlocutors affect significantly the amount of codeswitching (Dewaele & Li, 2014;Dewaele & Zeckel, 2015). It pays much attention to speakers' intentions and it intends to reveal relationships between language varieties and their symbolic meaning. Meanwhile, various types of communicative function have been identified (c. f. Liang, 2006;Y. Zhao, 2012).
The sequential approach to bilingual conversation builds upon Gumperz's (1982) notion of the contextualization cue and regards the switch itself as potentially significant to the managements of the ongoing conversation. It does not assume the relevance of social context or the identities speakers "bring along" to the interaction. It follows the "brought-about" social structure, or what emerges through the interaction, in the conversational analytic tradition. This approach can answer questions like "why that now"? The late 1990's and the first decade of the 21st Century witnessed a boom of publication in the conversational analysis of codeswitching. It treats language as a topic in itself rather than as a medium (Li, 2002).
The identity-in-interaction approach is post-structural in orientation (Li 2004). It recognizes that the social order organization is the background for identity-in-interaction. Social organization shapes talk, but it is also shaped by talk. Social identities like nationality, ethnicity, class, gender, being native or non-native speaker are the result of "identity work" in the interaction. They are negotiations of sameness and difference, and involve power relations between participants in interaction. It draws on four semiotic processes to account for how identities are created through language: practice, indexicality, ideology, and performance (Bucholtz & Hall, 2004).
The identity-in-interaction approach can be useful for research in bilingual conversation that attempts to integrate the concerns of both the symbolic and sequential approaches to data analysis. Identity is viewed as an outcome of cultural semiotic code-switching accomplished through the production of contextually relevant sociopolitical relations of similarity and difference, authenticity and inauthenticity, and legitimacy and illegitimacy (Cashman, 2008, p. 291). Identities are constructed and negotiated by co-participants in talk. Codeswitching is found by those researchers to serve both to transgress and to mark group boundaries, to challenge and to reinforce concepts of ethnicity and race (Cashman, 2008, p. 292). For instance, Migge's (2007) study finds that in the Eastern Maroon community in Suriname and French Guiana code-switching is applied to negotiate changes in the nature of the interaction, and more importantly, to manage speakers' self-presentation and relationships with their interlocutors. It points out that these uses are apparent, however, only when the conventional norms are flouted.
As previous studies of code-switching in EFL context mostly focus on code-switching in classroom settings as a teaching strategy, they can be categorized as abiding by the symbolic approach (c. f. Martin-Jones, 1995; Simon 2001). They mainly examine teachers' code-switching from the target language (TL) (English in many cases) to the first language (L1), and less on code-switching from L1 to TL, and mostly on the distribution of the two languages and relevant pedagogical functions. There are very few studies interested in the code-switching as a lens into teachers' identity construction. One such a study shows that codeswitching, "as a way of dramatic expression" (Liu & Feng 2007, p. 137), is a strategy which a teacher takes to express her identities. The current study aims to contribute to the literature in this regard by studying an EFL teacher's codeswitching between TL and L1 outside classrooms and how identity affects and is affected by such behavior. It can also provide a substantial descriptive scenario of a case in EFL context to supplement quantitative generalization of intraand inter-individual variation of code-switchers (c. f.

Research Method
The present study intends to find out the motivation, attitudes and underlying reasons for EFL teacher's codeswitching outside classrooms. It adopts a qualitative case study method (Merriam, 2008) to understand in depth the meaning of the participant's code-switching behavior from the identity-in-interaction perspective. As case studies are usually concerned with the individual, not with the general pattern of a population, the numbers of cases in a case study is usually very small (Zhu & David, 2008). The participant is typically chosen through non-probability sampling approach (Merriam, 2008), and purposive or purposeful sampling in particular. When the researchers intend to discover, understand or reconstruct a certain phenomenon, they must obtain samples capable of offering a large amount of information (Patton, 1990). Among the four types of purposive sampling (cf. Chen, 2000;Patton, 1990), this study mainly draws on typicality sampling.
The participant, Daniel (a pseudonym), was finally chosen after three stages of selection. The process of cases selection in purposive sampling has to be based on certain criteria (Merriam, 2008). Daniel met all the criteria made for the study. He had been observed to have code-switched very frequently. More importantly, he met the criteria of accessibility to data. The researchers and the participant were all members of a research institute and participated in a number of research projects. We met on a regular basis in small meetings, seminars and conferences. We also wrote reflections on research projects. The close relationship guaranteed the opportunity to collect data in more informal settings, such as conversations during lunches or informal online chats through instant messengers.
At the time of the study, Daniel was a Ph. D. candidate in his early 30s in a prestigious university in Beijing. He worked as a part time teacher of English at the same time. Before his Ph. D. study, he had been a teacher of English in two well-known universities, one in the East and one in the Northeast of China for ten years. Additionally, he went to a foreign studies university for four years. He received his M. A. in a university in the East. He went abroad for a visit for about a month. He had been observed to be engaged in speaking long chunks of English in some seminars, meetings, or in conferences. There are often instances of complete switching to English.
Both language data and content data are needed in this study to answer the following three research questions: What kinds of code-switching behaviors does Daniel engage in away from his classroom? What attitudes does he hold toward his code-switching behavior? What internal and external reasons are there for his code-switching behavior? The major research instruments comprise of observation, recording and online communications to answer the first research question, and interviews to answer the latter two questions. There are three major data collection methods in this study: ethnographic observation, interviews and online written sources.
Ethnographic observations were used because codeswitching "needs to be described and interpreted as an element of the social and cultural practices of set of speakers, rather than a fixed object existing in nature, to be discovered by an objective observer" (Heller, 2008, p. 149). The actual observation lasted for about two years covering a variety of activities like workshops, meetings, discussions, conference talks. The total time of audio-recording is 2,022 minutes, but only the parts with the participant's utterances were transcribed. An in-depth interview was conducted after the observations. In code-switching studies, interview can be used to obtain linguistic production from bilingual speakers and content data (Codo, 2008, p. 158) and it serves triangulation purposes (Codo, 2008). The actual interview took place in a café and lasted 97 minutes. It was initially semi-structured, but during the process, the participant strongly attributed code-switching to his personal experiences; therefore many more questions were added. It was audio-recorded and transcribed into 14,562 words. The third data type is from online written sources. They include 21 emails of 6,486 words, 5 written reflections on research projects of 3,963 words, and 6 times of online chatting transcripts in 2,538 words. The reflections were about research projects. The real time chat, derived from a software, highly resembles natural conversation and it is considered by Dorleijn & Nortier (2009) as of great value for code-switching research.
Qualitative case studies, in particular, stress the qualitative methods of data analysis and attach great importance to the interpretation of the data, rather than quantitative methods of analysis (Zhu & David, 2008). This study basically draws upon a general set of analytical steps suggested by Gardner-Chloros (2009, pp. 71-72) but does not follow the order. The first step is to describe or classify the speech produced by bilinguals and its patterning from a linguistic perspective. The second step is to relate it to the context in which it occurs. Here context must be construed in the widest possible sense to include the characteristics of the community, the languages' respective status, the presence or likelihood of language change, and the occurrence of code-switching. The third step is to relate bilingual production to the speakers' competence and attitudes, and to the characteristics of the conversation. The present study follows the first step and makes some changes to the other two steps, using two methods of analysis.

Findings
The presentation of findings is guided by and aims to provide answers to the three research questions: features of the code-switching behavior, attitude toward such behavior and motivations behind such behavior.

Features of Code-Switching
The overall pattern of how Daniel code-switches can be shown in Table 1. When Daniel switches to English outside classrooms, his interlocutors are mainly foreigners and close Chinese friends. He spoke English to Chinese people that he felt "close to", such as close friends or colleagues, because they would not mind it and "would enjoy the fun part". He only switches to English with those who has the same language repertoire, for example, those who "understand" typical phrases that he picks up from English TV dramas instead of textbooks. He speaks English at home, though, when he talks to his little son for educational purposes. Daniel is found to be speaking Chinese or English separately most of the time. All his marked switches are from L1 to TL, and no single instance of code-switching from TL to L1 is found. All the three types identified in Poplack (1980) are found to be present in his speech sample (as shown in Example 1), but he can be more accurately described as target-language-dominated code-switcher.
First, there is the insertional code-switching, or emblematic switching (Milroy & Muysken, 1995), involving the insertion of English words or phrases like" hedge" and "paper" in Turn 04 in Example 1. Intra-sentential switching like "say something" and "It may by hedging" occur in Turn 04. The third type is inter-sentential switching involving switching within the clause or sentence boundary, for example, the whole Turn 02 and the underlined part in Turn 04. Inter-sentential code-switching is more commonly found than the other two types. Typically, when other interlocutors switch to English, Daniel's English is "triggered" and will come out naturally.
Example 1 (At a seminar on English writing  , 1993), in such relatively conventionalized speech exchange where the unmarked code choice of using Chinese at a seminar is clear, he wants to call for EFL speakers and listeners to interact in both languages in this EFL speech community. In this event some audience followed suit and used all types of code-switching as well. Code-switching helps the speakers create the social realities for themselves and people involved in the event (Li, 2004).
Besides spoken utterances, Daniel is also found to be codeswitching in the online written medium, including group research reflections, emails and real-time online chat. Example 2 is an extract from a reflection on a workshop on helping teachers to identify problems in teaching. In this example he has three types of switching, because here he finds it easier to express his meaning by using the English terms directly, or switching to English as a strategy to save time and efforts, to save the trouble of translating. The underlying reason might be that EFL professionals have not established their own academic discourse in Chinese, as Daniel explained in the interview.
Example 2 Daniel (Extract of written reflection on a research project). Here I'd like to review the problems raised by teachers: 1. … 2. Ways of given feedback for compositions: In the past (holistic evaluation) was usually applied to score compositions, but now she wants to try a more specific way (analytic evaluation). 3. Ways for post-oral-task assessment: In the past teachers usually give students Judgmental Evaluation, like "Good", but now he wants to try Informational Evaluation, that is, to give students a specific evaluation based on the information they expressed. 4. … 5. School-Based teaching material restructuring: She is the leading teacher in their teaching group, and she hopes to know to integrate textbooks and use textbooks better. Apart from the written reflections, he is found to codeswitch in emails. Among the 20 emails collected, none of them are written in Chinese only. Eleven are written completely in English, and 9 are mainly in Chinese, with a few English insertions of "Dear" to greet and "Best" to close. Such greeting and closing can be seen as an intended sign of solidarity among the EFL research team. In the real-time online chat his record is totally in English. He thinks talking in English is "fun" with close friends whether in face-to-face conversation or online chat. Similarly, Daniel's cellphone text messages the researchers received are also in English, no matter in what language he received the text message. English is the "we-code" (Gumperz 1982) for the smallest EFL community of three people, similar to the finding in Chen (1996). Through such marked code-choice, we evoke and manipulate our EFL speaker identity (Sebba and Wootton, 1998).

Attitude Toward Code-Switching
Daniel's attitude towards his code choice is generally positive, which resembles teachers' attitude towards their students' code-switching in Liu & Hu (2007). He is quite aware of whom he can speak English to and he holds a very positive, though slightly cynical, attitude towards his own code choice. He enjoys speaking English to people who understands and "appreciate the fun part of switching to English". He thinks speaking English was "cool" and "fun". He notices that a lot of people around him do not use any English for fear of being seen as "showy or showing off", because Chinese people, even today, still "regard the Chinese language as an invaluable asset passed down from long history" and "demonstrate a strong sense of loyalty to and an equally strong sense of pride in the language" (Wu 1985, p. 307). But he believes one "got to be relaxed". His attitude to code-switching indicates that as EFL speakers, we have the legitimacy to use the language we are engaged in our profession in domains unconfined to classrooms.
Additionally, he has a strong sense of pride when choosing to speak English. As a teacher of English, he received very positive feedback from students regarding his teaching as "very different". He tries not to speak any Chinese in class and would teach in English very complicated issues such language teaching methodology, the history of language teaching, and second language acquisition. He always talked to foreign passengers in a coffee shop of the airport. He felt proud when he was in the U. S. visiting and teaching short courses on Chinese culture, a lot of American professors were surprised at his "American accent". There were only few times when he felt doubtful and questioned his attitude and identity. But then he told himself that what he cared most was his English proficiency, especially as a teacher of English, as he explained, "I just want to be a professional". Similar to EFL university students, Daniel, the university EFL teacher, embodies the "double cultural identity" (Ren, 2008).
In the classroom context where the maximal use of the target language is encouraged (Cheng 2013), and the marked switching from TL to L1 is called for only in difficult subject and mainly regarded as a pedagogic and communicative strategy (Qian, Tian & Wang 2009). Outside classrooms the marked switching is usually from L1 to TL and people's attitude can be mixed, either seeing the speaker as being "showing off".

Motivation for Code-Switching
Besides factors relating to his learning and teaching experiences, Daniel mentioned several other factors that can help account for his motivation for switching. And these factors relate closely to language confidence and language identity.
After years of learning and teaching as an EFL speaker, Daniel has become confident in his English. He thinks his English is more impressive than his Chinese. He speaks more flatly in Chinese and cannot use fancy words to talk about wonderful things. He even pointed out explicitly that he to a certain extent suffers from L1 attrition. He mentioned that he still spoke Chinese but did not usually come up with so much variety as in English.
The second category of motivation has to do with the EFL professional identities he intends to construct through codeswitching (Le Page & Tabouret-Keller 1985). When he asked himself if speaking a lot of English outside classrooms is "good or bad", he assured himself that he was "a professional", and it was "wonderful", in fact. He looked at others' similar questioning in the same way: Extract 1 (Interview). "I'm not sure. Professionally, as for me, professionally as a teacher, lot of…lot of guys would say 'Hey Daniel, you're losing your Chinese, shame on you'. But I think, professionally, professionally, for me as a teacher, for my career, this is wonderful, I think, I think it is responsible for me to do. I think it can take a lot of guys, ages to get to where I am. If I am, I am not too proud. You know a lot of Chinese guys, they only study English in China, it will take them years to get where I am. I mean the…the ease I have in coming out English. A lot of guys, they don't have that. You know what I mean? " As a matter of fact, any opportunities to speak English were his opportunities to learn and to use English, and that is the way he learns English in a Chinese context. He emphasized in the interview, "I'm a learner, I'm dominantly a learner", and "All is my learning mode". Using English whenever possible is his learning strategy.
His desire for becoming a competent EFL professional is so strong that he feels he really wants to be identified with English and with the culture at a high level. He pointed out that some people who might criticize the English education in China would "use" him as a way to show the negative effect of putting too much stress on English education. But he argued that he was a professional. He has to be really good at the language that he learns and teaches. Besides, he wants to teach the Chinese culture and everything Chinese to the world. It was his dream. Another dream is to be a "global citizen". He hoped that one day he would be able to find a job anywhere, not just in China, as a professional, either teaching English or Chinese.

Discussion
Different from earlier studies which showed that switching from English to their native Chinese was the norm for EFL teachers in the classroom context, this study found that Daniel engaged in large chunks of switching from L1 Chinese to English in contexts away from the classroom. Such non-classroom contexts include research discussions with peer researchers, informal conversations with close friends, and in written emails and cellphone messages with colleagues and friends.
The in-depth interview with Daniel shows that he is cognizant of his code-switching behavior and holds very positive attitude toward such behavior despite the potential negative views on such behavior in the large Chinese society. His personal traits such as being open-minded, having high global proficiency values, and high education level as identified in Dewaele & Li (2013) are all related to his positive attitude towards code-switching. The positive attitude toward his code-switching into English in the often public spaces can serve as an indication of the high status of the English language and the EFL profession in the Chinese society. This shows the connection between people's codeswitching as linguistic phenomenon with the larger sociocultural context (Li, Milroy & Pong, 2000;Wu, 1985). From the late 1970s, China began its open-up policy in a constant effort to reach out to the rest of the world, and English became an important tool for such an out-reaching effort. In the interview, Daniel recalls the nightly news program on his family's first 17-inch black and white TV set in the mid-1980s. He noticed the interpreters for national leaders and was "deeply impressed with the work of interpreters, believing that their job was very important and admirable", with "high prestige and high expertise." Daniel also mentioned his experience studying English at university, a popular major in the 1990s in China. It was a university specialized in foreign language education, where "everyone was supposed to speak good English", and "if you don't, you feel shameful", so he got pressure. And for a long time, after he gained confidence, he was trying to further improve his English. He had the ambition of speaking like a native speaker and formed the habit of switching to English at times. This experience of Daniel's also speaks to the high status of English in China and the high-stake in a good proficiency in English.
As an EFL teacher, Daniel's code-switching behavior becomes a self-imposed approach to sustained professional development, a constant effort to brush up his English skills in daily communication when English is not a natural medium. The code-switching behavior also serves as an important site of Daniel's professional identity construction. Given the high status of English again, Daniel is not shy about his EFL professional identity, instead, he takes pride in this identity and continues on a daily basis on improving his professionalism, reflected in the excellent conversational ability.
But at the same time, Daniel is aware of and very much alert to the hegemony of the academic discourse formed in the English language. He is very much cognizant that the dominant English academic discourse marginalizes and suppresses the growth of the discourses in other languages. In the EFL academic spheres we do not have our own academic discourse, he said, "We have to borrow the Western discourse" when "we are talking about academic stuff". We had to think "in the parameters set up, outside China, from the West". He also cited the example of many Chinese commodities having "foriegnized" brand names, which conveys the ideas of inferiority of Chinese, Chinese culture and Chinese language.
For this reason, he is cynical about his code-switching behavior as well, knowing that this behavior may serve to reinforce the hegemonic position of English. But at the same time, he tends to believe that it is important to avoid a purely nationalistic perspective on this issue, but should embrace a global stance, which is "a more liberal perspective on the whole thing". He explained that "in the whole process of globalization, we lose something, we lose but we gain more". He noticed that at the moment of the study there were many discussions about the English education during the two congress meetings in China. Some people said that we were "making too big deal" of the English education. There were drawbacks of "giving too much stress, too much attention to English education". According to Daniel, this may be a narrow view.

Conclusion
Different from earlier studies which showed that switching from English to their native Chinese was the norm for EFL teachers in the classroom context, this study found that Daniel engaged in large chunks of switching from L1 Chinese to English in contexts away from the classroom.
Such non-classroom contexts include research discussions with peer researchers, informal conversations with close friends, and in written emails and cellphone messages with colleagues and friends. "I'm a learner, I'm dominantly a learner" is what the participant emphasized for multiple times. The sentence can account for the feature, motivation and factors for code-witching in an EFL context.
Language and discourse have an influence on individuals' identity construction, as language learning inevitably involves the socialization into a certain linguistic and cultural community (Duff 2007). The data in this article seems to show that code-switching can serve as an interesting lens into EFL teachers' professional identity construction. Codeswitching from Chinese to English at any available opportunity in daily communication and in professional settings is used as a learning strategy for Daniel to maintain and to improve target language proficiency which is key to his professional standing as an EFL teacher.
In an immensely popular novel published in the middle of the last century in China (Qian 1947), there was a sarcastic description of the code-switching behavior of a comprador in Shanghai: "Due to frequent contacts with foreigners, he developed a habit in speaking, that is, he likes to insert some nonsignificant English words in between. It is not like he has novel meanings that are difficult to express in Chinese and thus he must borrow from English. Their value is incomparable to one's golden teeth, as the golden teeth are not only decorative, but also functional. They are rather like the meat stuck between one's teeth after a meal, whose only value is to demonstrate that he has eaten well. Other than that, there is no other use." (p 15) The above description in a way shows the negative Chinese attitude toward code-switching into English in the middle of the last century in the colonial context where China suffered a 100 years' humiliation partially as a result of Western invasions and interference. More than half a century later, the Chinese attitude toward code-switching to English can still be mixed, but the most important finding in the article is that individuals like Daniel begin to exert the agency to choose and negotiate their own identity desired.
Case study has an innate inadequacy of generalization, but this study has important implications for EFL education policy. Language learning will inevitably influence a learner's cognition and behavior, especially when they want to achieve supreme high level of proficiency in the target language and target culture. When we implement certain language education policies we must be highly aware of its possible consequences, both the intended and the undesired (cf. Spolsky 2011, R. Zhao, 2012).