Unwritten Rules: Code Choice in Task-Based Learner Discourse in an EMI Context in Japan

Abstract

This chapter reports on an exploratory study into learners’ perspectives on the use of their first language during an oral presentation task in a Japanese EMI context. Data included video- and audio-recordings of task-based peer-interaction and stimulated recall interviews collected from first year undergraduate English majors (ten learners in five pairs) in a university in Japan. Qualitative data analysis involved the iterative coding of instances of L1 use according to functions identified in previous research, as well as those emerging from the data. These data were then triangulated with stimulated recall data to identify salient features of L1 use as identified by the learners themselves. Learners also provided their perspectives on the principled use of L1 in L2 interaction and learning, with many expressing support for the ‘English only’ policy of their institution, as well as an indication of how and why they draw on their L1. The results provide evidence that learners in EMI contexts naturally and productively draw on the linguistic resources available to them to complete classroom L2 tasks. In recognition of this, the chapter concludes with suggestions for task-based language policies which take into account learners’ perspectives and the variable cognitive complexity of classroom tasks.

1 Introduction

English-medium instruction (EMI) is gaining in popularity across the Asia-Pacific, especially in countries where English is taught and used as a foreign language (EFL) (Dearden 2014). This is despite there being no agreed-upon definition of EMI, with different interpretations influenced by idiosyncrasies of local educational policy, language ideology and other contextual issues (Hashimoto 2013; Kirkpatrick 2011, 2014a; Tollefson 2015). Some researchers (Ball and Lindsay 2013; Lo and Macaro 2015; Morizumi 2015) equate EMI with content and language integrated learning (CLIL). Others, like Dearden (2014) and her colleagues, distinguish between the two in terms of historical and/or geographical context—CLIL emanates from plurilingual European contexts with no specification as to which language is the ‘second’ language; EMI is focused on the use and learning of English in more generalised EFL contexts. Dearden’s working definition of EMI is “[t]he use of the English language to teach academic subjects in countries or jurisdictions where the first language (L1) of the majority of the population is not English” (p. 4).

The growth of EMI in higher education in such contexts has been attributed to the global spread of, among other things, educational ideologies of internationalisation, international competitiveness (both of universities and economies), and a competitive concern with the measurement of quality (Tollefson 2015), all of which are dependent on interaction and performance in the language of international communication, English. While macro-level policy has been a major focus of research into EMI, investigations into classroom interaction are of major importance in determining the implementation and outcomes of such policy, at the micro-level, on teaching and learning practices in the language and/or content classroom (Chapple 2015; Hamid et al. 2013; Ramanathan and Morgan 2007; Vu and Burns 2014).

This chapter explores one important aspect of EMI in a Japanese university context at the micro-level: the role of a shared L1 (Japanese) in classroom L2 (English) interaction and performance on paired oral presentation tasks. I begin with a brief overview of the state-of-play with regard to EMI in Japanese higher education, followed by a review of literature related to teachers’ and learners’ use of the L1 in L2 classroom interaction. This research has generally involved a focus on teachers’ code choice practices, as often investigated via survey methodology (e.g., Glasgow 2014; Lasagabaster 2013), with comparatively few studies into learner interaction data or the learners’ perspective on their own code choice (Moore 2013; Scott and de la Fuente 2008; Storch and Wigglesworth 2003). I then present data from a study involving micro-analysis of learner interactions leading to the performance of oral presentation tasks, followed by stimulated recall interviews, intended to gain insights into learners’ construals of their own language learning in the context of a Japanese university. As such, the chapter aims to extend research into policy and practice related to code choice in EMI in Asian EFL settings (see studies reported in Kirkpatrick and Sussex 2012; Barnard and McLellan 2013b) by focusing on how languages intersect in the unfolding of a classroom task which is common to content and/or language courses in EMI in higher education.

2 Literature Review

2.1 EMI in Japan

While Japan holds a place in the modern history of international English language teaching methodology (Howatt 1984), and the English language is now taught from primary school (as ‘foreign language activities’; cf. Hashimoto 2011), its population continues to rank poorly on English language proficiency indices (Chapple 2015). As part of its internationalization strategy, the Japanese government, through the Ministry of Education, Culture, Sports, Science and Technology (MEXT), has undertaken initiatives to improve English language capabilities of Japanese students, in conjunction with initiatives at all levels of schooling to: address the decline in the number of university-aged Japanese students by attracting international students to Japanese universities, by relieving the pressure on them to learn Japanese (Chapple 2015; Howe 2009); to improve the comparatively low English language proficiency of Japanese students (Chapple 2015); and to prepare Japanese students to perform in the global economy, through Japan’s ‘Global Jinzai’ (lit. global human resources) program.

At the senior secondary level, the initiative involves a proposal, in principle, to teach “English in English” (cf. Hashimoto 2013; Glasgow 2014), and at the university level, the “Global 30” and “Global Jinzai” programs aim to attract 300,000 international students to designated national universities by 2020, and to prepare domestic students for international engagement, respectively (Brown and Iyobe 2014; Glasgow and Paller 2016; Howe 2009). These were followed by the “Top Global University Project,” providing ten years of funding to 37 universities in order to establish their global competitiveness (MEXT 2014). On the whole, these programs involve the creation of whole EMI courses or courses with EMI components (Brown and Iyobe 2014). Interestingly, government policy falls short of labelling English-only courses and activities as EMI, with Hashimoto (2013) arguing that this allows for “facilitating the co-existence of the national language and English without formalising the status of English as a medium of instruction” (p. 18). She adds that “the fundamental aim of Japanese internationalisation [kokusaika] is to promote Japan to the world” (p. 17, cf. also Ramanathan and Morgan 2007), suggesting a struggle between a Japan historically insulated from the outside world and a global context where international competitiveness (in English) is seen as an economic necessity.

In their recent review of EMI in Japan, Brown and Iyobe (2014) note that the Global 30 program had, as at 2013, funded 35 degree programs (with an exclusive focus on international students) across 13 national universities, while the Global Jinzai program (focused on domestic students) had funded 42 programs. Outside these funded programs, several universities have developed their own approaches to EMI, leading the authors to survey how EMI is being implemented in Japanese universities. While 194 Japanese universities had been offering some form of EMI as at 2006 (Brown and Iyobe 2014), and there has been expansion since then, the number of students taking these courses represents a small minority. They identified six approaches to EMI, ranging from “ad hoc,” with foreign language teachers teaching seminar courses, to “campus wide” with (nearly) all classes taught in English. In addition to issues related to the size and range of interpretations of EMI, Brown and Iyobe note issues related to fields of study, and uptake among staff. One participant in their study noted that there are trade-offs between language and content which may be unpopular with faculty specialists: “It’s really hard to tell the academic staff that you can’t teach something at a really high level because the students have to spend more time on the English” (p. 15). They further note difficulty with staffing, partly related to increased workloads, and that staff may be hired under short-term external grants, leading to program instability. The implication of this is the perception that “some models of EMI are becoming part of language teachers’ jobs. As EMI becomes more common, language teachers will be asked to take on more content classes in those programs” (p. 17).

While challenges in terms of teaching and learning content through EMI are highlighted by Brown and Iyobe’s findings, the role of classroom interaction in language development in EFL contexts is also an issue of major interest. Some researchers point to inconsistencies in government language policy and variability in application to institutional policy and classroom practice. With regard to the languages of classroom interaction (assumed to be Japanese and English; cf. Hashimoto 2013), institutional policy statements directing students to “use English only—our official language” may be written in course documentation containing mostly Japanese and accompanied by instructions for teachers to use Japanese at their discretion in order to help students’ understanding. This mismatch between EMI policy and classroom practice has been noted over time and across contexts (e.g., Kirkpatrick 2014a, 2014b; Xu 2014).

2.2 Code Choice in L2/EMI Task-Based Interaction

In EFL contexts like Japan, English language courses, where English language is both medium and content, may be seen as an integral part of EMI. These may be major courses in a degree in foreign languages, or minor courses in other programs. In such programs, as noted by Barnard and McLellan (2013a), “[t]he use of students’ first language has tended to be disparaged by textbook writers, methodologists and educational policymakers in many countries” (pp. 1–2). Kirkpatrick (2014b) further notes:

It is common to find that where English is the medium of instruction, the policy is that only English should be used in the class. It is equally common to find that, in practice, there is frequent use of the L1 in the classroom. (p. 8)

As part of what have been termed the social and bi/multilingual “turns” in applied linguistics (cf. Block 2003, and Ortega 2013) there has been ongoing interest in the role of the L1 in L2 use, learning and teaching. Related research has established that, during peer L2 tasks, learners draw productively on their first language for social purposes, such as negotiating disagreement, and cognitive purposes, such as discussing the grammar of the L2 (Antón and DiCamilla 1998; DiCamilla and Antón 2012; Moore 2013; Storch and Aldosari 2010; Swain and Lapkin 2000; Villamil and de Guerrero 1996). With regard to recommendations for classroom pedagogy, the debate has moved from whether or not L1 should be ‘allowed’ in the classroom, to the development of principles for incorporating valid L1 use (Levine 2011; Macaro 2009; Moore 2013; Swain and Lapkin 2013; Swain et al. 2011). Swain and Lapkin (2013), drawing on Vygostky’s sociocultural theory and a review of related research, offer three general guiding principles for L1 use in classroom interaction: L1 should be permitted in collaborative dialogue or private speech to mediate learners’ understanding of complex concepts (languaging) in the production of L2 texts, though the reliance on L2 mediation should be encouraged as L2 proficiency increases; teachers should make their expectations clear and work to create a supportive classroom environment; and “use of the L1 should be purposeful, not random” (p. 123). Moore (2013) further notes that “any attempt to influence L1 use in the L2 classroom must take into account that L1 use arises naturally and productively in L2/bilingual discourse” (p. 251) and that the demands of the specific context, including task and participants, must be taken into account when deciding what kind of L1 use might be planned for or predicted.

The research above is mostly based on researchers’ analysis and interpretation of transcribed discourse, recorded while learners work on language learning tasks. While researchers have found that the L1 can play a productive role in L2 learning, learners may or may not perceive, or even agree with these benefits, based on their experience of peer interaction (Moore 2013; Storch and Wigglesworth 2003). Their perceptions may be influenced by factors such as interpersonal relationships with peers (Philp et al. 2010), what they think teachers or researchers expect of them (Storch and Wigglesworth 2003), or other contextual factors which may not be evident from analysis of transcribed discourse of peer interaction (Moore 2013). The remainder of this chapter reports on an investigation into learners’ perspectives on their use of L1 (Japanese) and L2 (English) as they collaborate in pairs to develop and perform oral presentation tasks in a Japanese university context. Reflective stimulated recall interviews, conducted immediately after task performance, were conducted to provide the best possible chance of learners recalling events and providing insights into their own use of L1.

3 The Study

3.1 Research Questions

• 1. How do learners in an EFL/EMI context in a Japanese university draw on their available linguistic resources to complete a pair oral presentation task in English?

• 2. What is the learners’ perspective on their own use of L1 and L2 to complete the task?

3.2 Context and Participants

The data for this chapter were collected from first-year (second semester) students in the Faculty of Foreign Languages of a private university in Japan, where they studied courses in English, other languages, and/or international communication. The university had EMI courses at both undergraduate and postgraduate level, with the undergraduate program funded under MEXT’s Global Jinzai program (see above). As is common in such contexts in Japan, English language (and some other) classes and some common areas in the university were covered by an ‘English only’ policy. Data collection occurred outside regular classes in a research office, and students were invited to participate via email and participant information sheets which were distributed in their classes, as per ethics approval for the study. Ten students (three male, seven female) agreed to participate in the study and each was paid a nominal sum (JPY2000) for their participation. Participant data is provided below in Table 1 (pseudonyms used).

Table 1 Participant data

3.3 Data Collection

Data collection sessions lasted approximately 90 min, and involved the following activities:

• 1. Introductions, explanation and equipment testing (5 min)

• 2. Task description and explanation of materials (5 min)

• 3. Preparation for presentation (20 min)

• 4. Presentation (3–5 min)

• 5. Stimulated recall (30 min)

All ten interactions were recorded with digital video cameras. Immediately after the task, I showed each dyad the video recording of their interaction and conducted a stimulated recall session (cf. Gass and Mackey 2000), which lasted approximately 30 min. I asked students questions about their language use during interaction, and they were invited to pause the recording if they wanted to make comments. These sessions were generally conducted in English, but participants were free to comment in Japanese if they desired.

• The oral presentation task

Participants were presented with task instructions, including a topic, seven pictures and summarised content information.

The task instructions are reproduced below:

Presentation Task instructions

Presentation task

Planning time: approximately 20 min

Length of presentation: maximum 5 min

In pairs, create a presentation using the seven pictures and information provided. Share planning and presentation time equally. The topic of the presentation is “The Great Wall of China.” In planning your presentation, please think about:

• (1) structure (introduction, body, conclusion);

• (2) grammatical accuracy;

• (3) use of voice, eye contact and gesture; and

• (4) use of images.

As an example of the content information provided, the first group of pictures showed images and a map of the Great Wall, with the following prompts:

1. Long

• a. +8000 km

• b. 8 m high; 7 m wide

• c. West (west China); east (North Korea; Yellow Sea)

The task was designed to mirror oral presentation tasks the students were required to perform in their language and content classes (cf. Moore 2013), without the need to negotiate and create content ideas, given time limitations. The particular topic was chosen as it was expected that the students would be generally familiar with the content, but not with the specific data provided. In the time allocated to the task, the participants were expected to collaborate with regard to how the provided content would be presented, both linguistically and physically.

3.4 Data Analysis

Interaction data (approximately two hours in total) were initially transcribed and analysed for individual amounts of L1 and L2 use. Instances of L1 use were then coded according to the functions they performed in the dialogue (cf. Moore 2013; Storch and Aldosari 2010). Following Scott and de la Fuente (2008), relevant portions of the stimulated recall transcripts were identified to gain insights into the learners’ perspective on their own use of L1 in L2 task-based interaction. Transcription conventions, outlined below, are based on those used by van Lier (1988).

INT.:

Interviewer—the author of this chapter

…:

Interval between utterances of approximately one second

(6):

Interval between utterances if more than 5 s

e:r the:::

Lengthening of the preceding sound

-:

Abrupt cut-off

?:

Rising intonation, not necessarily a question

!:

Animated or emphatic tone

inheritance:

Speaker emphasis

˚word˚:

Utterance between symbols is noticeably quieter than surrounding talk

うん (yeah):

Japanese utterances are followed by free translation in brackets

(unint.):

Unclear or unintelligible speech

(guess):

Transcriber doubt about a word

(laugh):

laughter

((writing)):

non-verbal actions or editor’s comments

//:

turns deleted for the purpose of analysis

4 Results

After providing data on participants’ L1 use in interaction, this section presents results of the stimulated recall interviews. First, participant perspectives on their L1 use are presented, followed by their reflections on instances of L1 use during interaction. Table 2 shows the proportion of turns incorporating L1 use for each dyad.

Table 2 L1 use by dyads

Based on these data, pairs 1 and 2 can be classified as moderate L1 users, pairs 3 and 4, low L1 users, and pair 5, extensive L1 users (cf. Storch and Aldosari 2010).

4.1 Functions of L1 in Task-Based Interaction

All learners used aizuchi (backchanneling particular to Japanese) in Japanese in their interactions. It is not common for L2 pragmatics to be an explicit focus of language instruction in EMI settings, so it can be expected that learners may draw on L1 pragmatic resources when interacting with their peers (Kasper 2001). Second, learners may interact in their L1, or code-switch while interacting in L2. As shown in the example below from the study, codeswitching may occur while learners are engaging in aizuchi, or particularly Japanese-styled backchanneling devices to show the interlocutors’ engagement with each other’s utterances (LoCastro 1987; Ohta 2001; Kita and Ide 2007).

• Extract 1 Aizuchi

• 22 HARUKA: He is Chinese?

• 23 MIKA: Chinese

• 24 HARUKA: うん (yeah) (50) (writing silently)

• //

• 40 HARUKA: へ (huh) three hundred yen

• 41 MIKA: It’s so…reasonable!

• 42 HARUKA: Yes! (laughs)

• //

• 61 MIKA: how to… あの (um) … how to access

• 62 HARUKA: access access

As can be seen above, both learners employ Japanese and English to keep their English conversation going, by showing interest in their partners’ contributions (turns 24 and 41) and expressing surprise (turn 40). They also use it to manage shared thinking (turn 61), and use repetition to show engagement with their partner's contributions.

Low L1 users in the study only used the L1 for aizuchi, the use of English loan words with Japanese pronunciation and meaning, and, to a small extent, to negotiate lexical form. Ren & Yuuta used aizuchi (うん, yeah) on only three occasions. At one stage in their interaction, the following interchange occurred.

• Extract 2 Negotiating lexical forms

• 24 REN: So…it…was the most famous architecture…in the world…we were interested in…

• 25 YUUTA: It’s one of…it’s one of…world heritages

• 26 REN: yeah yeah yeah

In the stimulated recall session, Ren noted that in considering whether to use either ‘architecture’ or ‘world heritage,’ “I thought about the meaning in Japanese.”

In addition to using the L1 for aizuchi, Yuriko and Michiko used it for loan words and negotiation of form, as can be seen in extracts 3 and 4 respectively.

• Extract 3 Loan word

• 35 MICHIKO: this one is like p.. プロフィール (profile; background information)

• 36 YURIKO: yeah プロフィール … which one?

• Extract 4 Negotiating a lexical phrase

• 270 YURIKO: to…walk on the wall. (on the) wall

• 271 MICHIKO: it takes two hours 端から端 (end to end)

• 272 YURIKO: (laugh) I was too…the same same thing, but please…

• 273 MICHIKO: okay…

• 274 YURIKO: it’s oh

• 275 MICHIKO: ˚端から端˚ (˚end to end˚; low whisper)

• 276 YURIKO: end…start to end?

In the stimulated recall session, Michiko noted as follows:

253 “so, first I I was thinking in Japanese, and then change into English. So, ahh… if I change into English so is it…correct sentence or … in Japanese sentence it’s okay, but if I change in English is it good sentence?”

Dyads with moderate L1 use, used the L1 for the above functions, as well as for task management and content creation, including negotiation of (grammatical and lexical) form and meaning, involving multiple turns.

• Extract 5 Managing structure and collaborative content development

• 15 MIKA: このまえ? (before this?) I think I I want to say “This time I’d like to tell you about the Great Wall of China” but first I talk about how to how to say those those how to…explain about the (unint.) for the second and third Kanko talk about history and … souvenir?

• 16 HARUKA: Souvenir? How to…

• 17 MIKA: How to…access?

• 18 HARUKA: Yeah…

• 19 MIKA: access か (interrogative particle) (unint.)これ (this) this this one…out-outline.

• 20 HARUKA: なんと言うの (how can I say)

• 21 MIKA: part…big…uh big… much big…

Extract 5 shows Mika and Haruka drawing on the L1 minimally in managing the structure of their presentation (turn 15) and collaboratively negotiating the form of the content (turns19–21). Interestingly, when asked about which language they were thinking in during the interaction starting at turn 19, Haruka noted that she was thinking in English, while Mika noted the opposite:

• Extract 6 Language choice

• 36 MIKA: When I make sentence, I えー(hm:?)…do you…. In in my in my mind I make sentence in Japanese and translate English…and write down English sentence.

• 37 INT.: Yeah? And how about you Haruka?

• 38 HARUKA: I tried make a sentence in English, but I I have no confidence about grammar so (laughs)

Luis and Hanako’s use of L1 extended to two multi-turn exchanges—the first where they were trying to collaboratively work out the meaning of a prompt:

• Extract 7 Negotiating meaning

• 54 LUIS: 意味分かんない… (I don’t understand what this means)

• 55 HANAKO: by taxi…

• 56 LUIS: 意味分かんない … (I don’t understand what this means) two hours…

• 57 HANAKO: it takes a one hour from Beijing to there…by taxi か (interrogative particle) and it costs…by taxi…or?

• 58 LUIS: いや (no) … 何だろう (I wonder what) えっ (huh?) 有り得ない (that’s impossible)

• 59 HANAKO: うんん…(yeah, maybe…)

• 60 LUIS: 有り得ないじゃん (that’s impossible)

• 61 HANAKO: From Beijing ta- by taxi…one hour and by walk…

• 62 LUIS: hm:?…

• 63 HANAKO: two hours…

• Extract 8 Stimulated recall

• 59 INT. You’ve jumped into Japanese. (laugh) So somewhere in there you you kind of jumped into Japanese. Why do you think that was?

• 60 LUIS: I think because we didn’t know why make like walk two hours? We didn’t know the meaning? Or we know the meaning, what take what take two hours? because it was just walk two hours. But Hanako said like “From Beijing it takes you one hour”…

• 61 INT.: Yep, I mean…I know. You went back into…it was pretty clear where you kind of started talking Japanese…um…(watching video) What are you thinking there? You said, “walk two hours” right? You’re looking the paper. What were you thinking there?

• 62 LUIS: Why is it two…walk two hours?

• 63 INT.: Right, okay. And that’s about right where you….And what are you thinking?

• 64 HANAKO: えー(um:) I thought first…at first, I thought…mm…by taxi one hour and walk…by walk two hours… but he said, uh…impossible!

Their second multi-turn exchange arose when they were trying to come up with the English translation for 世界遺産 (world heritage).

• Extract 9 Negotiating form

• 88 LUIS: it’s not so far…from here….and its…its ah…なんだけ?えっと.. (what is it? um…) 世界 (world) …

• 89 HANAKO: ahhh!…世界遺産 (world heritage)

• 90 LUIS: そう! (that’s right!) it’s ah:…and it’s ah: … world … world … it’s a world….it’s world …. world … (unint.) no…it’s a world … world …

• 91 HANAKO: Can I… use a dictionary?

• 92 LUIS: If you want … I think…

• //

• 95 HANAKO: Okay okay…hmm…世界遺産 (world heritage)

• 94 LUIS: it’s a world world

• 95 HANAKO: 遺産? (heritage?)

• 96 LUIS: そう、そう、そう (right, right, right) (10) world…

• 97 HANAKO: oh… I can’t find it (9) oh…okay okay…

• 98 LUIS: ahh … in … inheritance.

• 99 HANAKO: in…inheritance? hm: …

The issue here is that the term 遺産 represents both ‘inheritance’ and ‘heritage’ in English, so the correct choice only becomes clear for Luis through a clarification request in the stimulated recall session (extract 10, turn 80).

• Extract 10 Negotiating form

• 76 INT.: What’s this? What were you thinking?

• 77 LUIS: World inheritance? Why inheritance? In English…

• 78 INT.: Where where is this? Which part were you…

• 79 LUIS: At the conclusion. At the conclusion you must summarize. So, maybe you must say why you must be see this excellent place. And you have to take place because it’s not just a place…it’s a world inheritance.

• 80 INT.: a heritage?

• 81 LUIS: yeah, a heritage…

• 82 INT.: heritage, okay…

• 83 INT.: Oh did you say ‘inheritance’?

• 84 LUIS: inheritance. Yeah, it’s quite different.

• 85 INT.: Oh, it is slightly…okay, yeah, I wasn’t sure what you were talking about but…

• 86 HANAKO: I did.

• 87 INT.: The heritage…so it’s a place that’s got…

• 88 LUIS: It’s not just a place. It’s a world heritage. So, maybe if you say, ‘world heritage’…it’s like like ‘wow’!

Finally, Maki and Emi, the extensive L1 users, drew on the L1 for all the functions mentioned above. Backchanneling was performed mainly in Japanese, with うん (OK, yeah) used 43 times, various forms of ‘what?’ (何, え?) used 26 times, ね (right) used 17 times, and えっと(well, um) used 15 times. Extract 11 is typical of their interaction.

• Extract 11 Extensive L1 use

• 40 MAKI: 規模、規模って何? (scale…what’s ‘scale’ (in English))

• 41 EMI: 規模という事はイコール (scale equals)

• 42 MAKI: OK. (laugh) えっと(um)…規模 (scale).

• 43 EMI: うん、規模 (yeah, scale)

• 44 MAKI: ((checking electronic dictionary)) sc-scale…

• 45 EMI:　ううん (I see) scale

• 46 MAKI: scale…scale and … 目的だから (because it’s the purpose/aim) aim だっけ? (is it ‘aim’?) aim…and the aim…はあ (um) … the full distance なんかさ、全体はって、全体の距離は? (so, um, how do you say the whole or full… um … full distance?) so….ディスタンス (distance)… …about

• 47 EMI: 8000千キロはすごくない (Isn’t 8000 km amazing!)

• 48 MAKI: すごい (it’s amazing!) eight…eight (thousand)…thousand

Maki and Emi used their L1 to organise their interaction, negotiate form and meaning and create content. Extract 11 shows them using translation, supported by a bilingual dictionary (turns 40–46), creating content (turn 46), commenting on the content (turns 47–48) and engaging with each other’s contribution (throughout). In other words, the learners appeared to be “using the L1 to manipulate the L2 content they were creating” (Moore 2013, p. 250). In their stimulated recall session, Maki and Emi noted several instances where they were able to negotiate grammar and vocabulary via their L1 interaction. On one occasion during interaction, the following exchange occurred.

• Extract 12 Negotiating content development via the L1

• 77 EMI: ワズ (was)

• 79 MAKI: the first…emperor…ビルということ? (building, right?)

• 80 EMI: えー、ビルド(yeah, build)

• 81 MAKI: yes. build … the Great Wall of China …

In the stimulated recall, she noted “at first build … but past sentence” (turn 235), and during the presentation the correct form ‘built’ was used. On a different occasion, she noted: “I noticed it about my mistake and changed” (turn 220), though on this occasion the result was non-standard.

4.2 Participant Reflections on Their Own L1 Use

During the stimulated recall sessions, participants were asked to reflect on their L1 use. While there was variability among individuals, low to moderate L1 users, in explaining their low L1 use, referred to: the influence of a classroom English-only policy; related ‘habits’; value judgements about the positive effects of using English and the negative effects of using Japanese; positive and negative emotions associated with language use; and the influence of translation on the final product.

In response to questions about code choice in their interactions, several students noted they were influenced by the university’s English-only policy.

• Extract 13

• 33 YURIKO: なんだろう? (I wonder what) 癖? (habit) … 習慣? (custom)

• 34 YURIKO and MICHIKO: habit! habit

• 35 INT.: habit…from?

• 36 MICHIKO: Freshman English class only…we have to only speak English

When asked how they usually prepare for oral presentations, the following exchange occurred between Luis and Hanako:

• Extract 14

• 121 LUIS: 準備の時どうする? (what do you do when you’re preparing?)

• 122 HANAKO: 準備のときは… (For prep I …)

• 123 LUIS: がんがん日本語しゃべるかな? (maybe you speak a ton of Japanese)

• 124 HANAKO: いや、授業中だけど、英語しか話せないけど。 (no, in class I can only speak English.)

• でも、本当になんか、え、これって、みたいなの時は、小声で日本で話す授業中に。 (but actually in a class situation like this right now I would probably speak Japanese in a soft voice)

• 125 LUIS: 先生には? (how about to the teacher?)

• 126 HANAKO: え? せめられないように (huh? [I’d speak English] so I won’t get criticised [for speaking Japanese]) (laugh)

• 127 INT.: So, how do you feel about speaking Japanese then?

• 128 HANAKO: ahh…it’s easy to understand…. more…than English … so, but … if I prepare in English, uh: …uh …なに、やりやすいって何と言うの (what … how do you say ‘easy to do’)

• LUIS: well?

• I can speak, um: I can speak

• INT.: well?

• HANAKO: English well umm 本番で何と言うの (how do you say ‘during the presentation’)

• 129 INT.: In the actual presentation?

• 130 HANAKO: Yes yes, in the actual presentation.

It was common for participants to make apparently contradictory comments like Hanako’s (turn 124), which contrasted what she was required to do according to the policy, which she stated that she supported, with her actual practice. Similar comments were made about learners’ ideals of using English, with practicalities based on their perceived limitations in English.

Several judgements were offered by participants in support of the English-only policy. Extracts 15 and 16, for example provide two perspectives from low L1 users.

• Extract 15

• 67 MIKA: ((it’s better to use English in class)) because I want to speak English fluently…

• 68 INT.: Yeah

• 69 MIKA: and uh…if I always use Japanese in English class…

• 70 INT.: Yes…

• 71 MIKA: it’s not good for me and uh … for for my friends

• 72 INT.: Yeah,

• 73 MIKA: and…not…I…I enjoy enjoy enjoy use…using English,

• 74 INT.: Okay.

• 75 MIKA: so I I’m okay. (laugh)

• 76 INT.: okay. Good. And how do you feel about it Haruka?

• 77 HARUKA: Using English…in my class?

• 78 INT.: Yes

• 79 HARUKA: Ah, I think it is good. Because so…if everyone speak English all the…all time?

• 80 INT.: Yes

• 81 HARUKA: So, they can enjoy speaking English…

• 82 INT.: Yes

• 83 HARUKA: And if someone speak Japanese, they are not interesting they are not interested in speaking English and maybe they are shy なんか (or something) … to speaking English.

Both learners in extract 15 make judgements about the positive and negative effects of using English (turn 67) and Japanese (turns 69–71), respectively, including references to emotional states or personality traits (turns 73 and 83). In extract 16, both learners provide reasons for their strong support for an immersion approach to language (if not content) learning.

• Extract 16

• 247 REN: Mmm…once I’d decided to use English [during today’s preparation]

• 248 INT.: Yes?

• 249 REN: I didn’t think about it. About Japanese.

• 250 INT.: OK.

• 251 YUUTA: Uh…I think we shouldn’t use Japanese when we studying English…so…yeah…

• 252 INT.: Why … why is that?

• //

• 262 REN: translation from Japanese: (Well, right we can talk in Japanese better so… it’s better right? But even if you speak in Japanese you can’t (say it) in English … you can’t, so… You don’t learn anything from it. … Science and other subjects, for example … When you learn other subjects it’s difficult (to use English) so we should use Japanese, but in the case of English….um…language… … using words/language is the best, right?) … to use is the best way do you put it, I think.

• 267 YUUTA: The more touch uh…the more we touch English the…

• 268 REN: more?

• 269 YUUTA: the more our English skills better.

Finally, the influence of translation on language production was seen as a negative by the following dyad.

• Extract 17

• 307 MICHIKO: 英語でなんか、もの、文法とかを考えないと(ああ)なんか、日本語で考えちゃうから英語が日本語の文章みたいになっちゃう。(so, if we don’t think of things and grammar in English, we’ll um think of things in Japanese and our English sentences will sound like Japanese.)

• 308 YURIKO: あ、なるほど。 If we think of…past thing Japanese sentence….

• 309 INT.: yes

• 310 YURIKO: and translate…なんだけ? (what was it?)

• 311 MICHIKO: in?

• 312 YURIKO: translate English sentence is not good, we think.

• 313 INT.: right.

• 314 YURIKO: and [because] so, we have to think about only English sentence if I sp-speak English…

• 315 MICHIKO: so because um…if I…think, if I’m thinking in Japanese so English sentence like Japanese style so, it’s not good for…us to learn study English, so…

Learners who were moderate to extensive L1 users, in contrast, explained their code choice with reference to their ability to explain complex phenomena in Japanese, and frustration with their own or others’ perceived inability to do the same in English.

• Extract 18

• 133 LUIS: of course … we write we write in English, but we talk our ideas we share our ideas in Eng. … Japanese //

  so, it’s not that we cannot speak English together. It’s just that if we speak English we have to explain. It’s more complicate…it’s complicate to explain that. So, it’s quite easy to speak Japanese.

• 135 LUIS: … in my class, almost nobody can talk English and if you’re talking English you have to explain a lot of things stuffs … because if you talk like this you they cannot understand too much so maybe you have to // like use easier words // it is quite complicated I mean it’s quite boring explaining all the guys all the stuff

In contrast to Hanako’s use of English discussed in extract 14 above, Luis notes the convenience of using the L1 with a classmate of similar L2 proficiency (turn 133), as well as the frustration of using the L2 in interaction with classmates of lower L2 proficiency (turn 135).

Finally, in contrast with Ren’s comments in extract 16, Maki and Emi explain their challenges as learners of lower proficiency, and as extensive L1 users.

• Extract 19

• 74 EMI: translation from Japanese: (um … in English…I can get a bit mixed up, but I feel I can get across what I want to say more in Japanese)

• 75 INT.: OK. How about you, Maki?

• 76 MAKI: (um, even trying in English I can’t express what I truly think well…)

• 77 INT.: Right…

• 78 MAKI: (um, I don’t know if the other person thinks…or really understands (what I’m saying in English), but in Japanese I have confidence (they understand)).

5 Discussion and Conclusion

Like previous research into the use of available linguistic resources in task-based interaction (Antón and DiCamilla 1998; DiCamilla and Antón 2012; Moore 2013; Storch and Aldosari 2010; Swain and Lapkin 2000; Villamil and de Guerrero 1996), this study, conducted in an EMI context where English is used as a foreign language, has found that learners naturally and variably draw on their L1 linguistic resources to manage, negotiate and construct content in preparation for an oral presentation task performed entirely in their L2, English. The data show participants drawing on the L1 for social and cognitive purposes. Specifically, all learners used aizuchi for a range of purposes, including signalling, inter alia: dis/agreement (extracts 1, 7, 9, 11); surprise or amazement (extracts 1, 11); wondering (extract 1); or confusion (extracts 1, 5, 7). Low L1 users drew on the L1 briefly, in the use of loan words and the negotiation of lexical form (extracts 2–5). In addition to these functions, moderate L1 users drew on the L1 for a range of functions, including managing and structuring the task (extract 5), negotiating meaning (extracts 7–8) and negotiating form (extracts 9–10). In contrast, extensive L1 users appeared to manipulate the L2 content development via the L1, drawing on all the functions outlined above (extracts 11–12).

In explaining their L1 use, low to moderate users noted that they were influenced by the institution’s English-only policy (extracts 13–15), and several participants spoke positively about the use of English and negatively about the use of Japanese (extracts 14, 15, 17). Positive affect-related comments linked English use to enjoyment and Japanese use to lack of interest or shyness (extract 15). Other comments related to the negative effects of Japanese-English translation on the finished product (extract 17). In contrast, moderate to extensive L1 users offered two complementary perspectives related to language proficiency: the frustration of higher proficiency learners not being understood by those of lower proficiency in English (extract 18), and the confusion, lack of confidence and difficulty experienced by lower proficiency learners in using the L2 (extract 19).

In light of these findings, Swain and Lapkin’s (2013; cf. also DiCamilla and Antón 2012) first guiding principle for L1 use in classroom interaction (i.e., use L1 for languaging about complex concepts in the production of L2 texts) can be seen as a recognition of what many learners actually do, regardless of the existence or otherwise of an institutional language policy (Swain et al. 2011). Especially for learners of lower proficiency, and in tasks where the task performance is in the target language, Swain and Lapkin argue “[p]ermitting the students to use their L1 to language (at times when the complexity of the task makes it necessary to do so) still allows for the target language to play a key role in the activity” (p. 114). This signifies a move away from a generic English-only policy, to one which is task-centric (Moore 2013), or based on the expected demands of the task given to learners. The use of the term ‘permitting’ indicates both the influence of the teacher on learners’ in-class behaviour (see extract 14, where Hanako notes that she would “probably speak Japanese in a soft voice” in class), and a recommendation for agreement between teachers and learners that there are valid uses for the L1 in L2 learning.

Swain and Lapkin’s second principle (for teachers to make their expectations clear and create a supportive environment) relates to the perspective that there is an inextricable link between cognition and emotion in language learning. In the data for the current study, there were several instances where social/emotional negotiations were intertwined with the development of the task at hand. The extensive use of aizuchi was the most obvious of these, but all learners drew on both the L1 and L2 to achieve intersubjectivity (Antón and DiCamilla 1998; Moore 2013), from expressing surprise (extract 1) or confusion (extract 7) at information provided in the prompts, to the use of laughter or repetition to engage with each other’s contributions and keep the cognitive work of the task on track. Their third principle (for L1 use to be “purposeful, not random” [p. 123, see above]), is more difficult to achieve in learner-learner interaction than in teacher-fronted classrooms. The participant reflection data reported above suggest that individual learners develop their own principles for language use during interaction with their peers, and these appear to be influenced by a range of factors (noted above), including institutional policy, prior experience with similar tasks, their own and their peers’ language proficiency, and perceived task complexity. Further research into raising learners’ awareness of links between code choice in interaction and language performance is needed to determine whether code choice practices of extensive L1 users can be changed. It is interesting to note that several participants in the study expressed support for the ideal of the English-only policy of their institution, while noting that the practicalities of different situations led them to draw on their L1 in interaction. While this could be interpreted as learners simply adopting the policy as presented by their institution, the principled reflections provided by participants in this study suggest that the learners themselves may be a valuable source for discussions surrounding language policy and practice.

While the study does provide support for Swain and Lapkin’s (2013) and others’ principles of L1 use, it also supports the contention that L1 use emerges naturally in learner interaction, and that contextual factors need to be taken into account when planning language use or policy for the classroom (Moore 2013; Swain et al. 2011). In addition, it has been seen that learners interact based on their own working principles and language ideologies, informed by institutional policy, but they also diverge from these in practice in logical and potentially predictable ways, in response to their own or others’ actual or perceived limitations. For the reason that code choice practice emerges and is influenced by the cognitive and emotional pressure of the task-in-process, teachers’ task-based policies may be based both on the task as designed or implemented by teachers, as well as what is known about how learners adapt and respond to this in the task-in-process (Breen 1989).

The study was limited in terms of generalizability by the facts that it was conducted on a small number of self-selected dyads and that it was not conducted in the participants’ classes. As such the study cannot be claimed to represent normal classroom interaction for these participants, or how participants who are unfamiliar with each other might perform the task. Nonetheless, participants commented on the similarity of the research task to tasks they had recently performed in class, and made reference to the influence of the institution’s ‘English only’ policy on their interaction, indicating that findings are of value in understanding classroom interaction. The study provides valuable insights, from learners themselves, into how learners in the specific context of EMI in a Japanese university draw on their existing L1 and L2 linguistic resources to complete an oral L2 task which is representative of oral tasks in many international universities.