The globalization process has forced higher education institutions to increasingly use EMI to attract foreign students and improve their situation in the ever-more highly regarded university rankings (Dafouz & Smit Reference Dafouz and Smit2016; Wilkinson & Walsh Reference Wilkinson and Walsh2015). In fact, during the last decade EMI courses are burgeoning in contexts as diverse as Europe (Doiz, Lasagabaster & Sierra Reference Doiz, Lasagabaster and Sierra2013a; Wächter & Maiworm Reference Wächter and Maiworm2014), Asia (Kirkpatrick Reference Kirkpatrick, Murray and Scarino2014), the Middle East, South America or Africa (Dearden Reference Dearden2014). The growth of EMI is a global phenomenon, but despite its rapid spread, there is an outstanding shortage of pedagogical guidelines about how to implement effective courses. Interestingly, in most universities there is hardly any pre-service or in-service course to help practitioners cope with this new challenge, which represents a serious concern for many EMI teachers (Doiz et al. Reference Doiz, Lasagabaster and Sierra2013a; Dearden Reference Dearden2014).

One of the main characteristics of EMI programmes has to do with content teachers’ widespread feeling that language is not their concern (Aguilar Reference Aguilar2017). In a series of studies (Doiz, Lasagabaster & Sierra Reference Doiz, Lasagabaster and Sierra2011, Reference Doiz, Lasagabaster and Sierra2013a, Reference Doiz, Lasagabaster and Sierra2013b, Reference Doiz, Lasagabaster and Sierra2014), carried out by the research team I lead, we have also observed that one of the main reasons that puts teachers off from joining EMI experiences has to do with the general belief that higher education institutions do not support teachers who take on the responsibility of teaching content subjects in English. As one of the EMI teachers in Doiz & Lasagabaster's (Reference Doiz, Lasagabaster, Doiz and Lasagabaster2016a: 143‒144) study puts it:

Some participants thus feel frustrated by the lack of institutional support to work on their English, which leads them to attribute the success of the EMI programme primarily to their own commitment. Also in the Spanish context, Halbach & Lázaro (Reference Halbach and Lázaro2015) point out that most universities do not offer their lecturers any specific training related to EMI teaching. This is one of the reasons why content EMI lecturers do not consider language aspects (Aguilar Reference Aguilar2017), a widespread malaise in many European contexts (Airey Reference Airey2009) and a weakness that could be overcome by boosting collaboration between language and content teachers (Schleppegrell & O'Hallaron Reference Schleppegrell and O'Hallaron2011). This collaboration should help to raise EMI teachers’ own awareness of language aspects and it should also lead them to ponder the fact that they have a specific role to play when it comes to teaching the discipline-specific language. Language development and learning the nuances of the language used in each specialization should be an integral part of any EMI experience. Not only do students need to understand the discipline-specific vocabulary but also the conventions of the discipline (so-called English for specific purposes (ESP)). This entails students learning the specific vocabulary, syntax and punctuation characteristic of each speciality, but this approach should also encompass other language aspects and research is direly needed on how to accomplish this. Since academic skills are not always transferable from discipline to discipline, students need to adapt their discourse to the specific features of each programme (what has traditionally been tackled by English for academic purposes (EAP)). Ideally EMI courses should be underpinned by ESP and EAP courses, but unfortunately this is currently not the case in many higher education institutions. In fact, in many European universities ESP and EAP courses have been eliminated as a result of the restructuring of curricula brought about by the implementation of the Bologna process. And this despite the fact that EMI teachers need to deal with language issues in their classes in a more systematic way (Lyster Reference Lyster2007; Aguilar Reference Aguilar2017; Lo Reference Lo2015), as well as monitor their students’ linguistic output.

It is at this stage when team teaching comes to the fore, as it could help to pave the way for language and content integration, despite the fact that I need to acknowledge that team teaching is hard to establish in a systematic way for a number of reasons including workload, expense, and the fear of being observed. For the present research agenda I define team teaching as collaborative work between a content lecturer and a language lecturer in an EMI programme in which the abilities of the team members complement each other to improve the learning results so that the whole is greater than the sum of its parts (Kerridge, Kyle & Marks-Maran Reference Kerridge, Kyle and Marks-Maran2009: 93).

Research demonstrates that team teaching improves undergraduates’ learning and can also be a gratifying experience for the teaching staff involved, who often comment on the motivational boost it gives them, as it encourages reflection on their pedagogical knowledge and practices (Anderson & Speck Reference Anderson and Speck1998; Hanusch, Obiojofor & Volcic Reference Hanusch, Obiojofor and Volcic2009; Lester & Evans Reference Lester and Evans2009; Baeten & Simons Reference Baetens and Simons2014). In addition, it is considered to be an innovative teaching practice which does not always come at a high cost (Carpenter II, Crawford & Walden Reference Carpenter II, Crawford and Walden2007). However, it needs to be acknowledged that one of the major concerns surrounding team teaching has to do with funding (the price of having two teachers for the price of one), although there is no reason why it should become an insurmountable obstacle. Two options could be considered. Firstly, language teachers’ teaching load could be redistributed so that their team teaching becomes part of their regular teaching. If higher education institutions are willing to foster EMI as part of their internationalization process, they should be willing to embark on team teaching as it may help to pave the way to successful EMI courses. Secondly, research projects funded by the university itself or regional/national institutions could be undertaken. For example, the team-teaching experience on which this research agenda is based is funded by the Spanish Ministry of Economy and Competitiveness via a research grant.

In the context of higher education, teamwork involving language and content teachers may be one of the ways forward, as it may be an ostensibly rewarding undertaking. This collaboration would make EMI courses more effective and facilitate undergraduates’ content learning (Davison Reference Davison2006; Lyster Reference Lyster2007; Llinares, Morton & Whittaker Reference Llinares, Morton and Whittaker2012; Kong Reference Kong2014), increase their degree of satisfaction (Perry & Stuart Reference Perry and Stuart2005), and help to do away with the rigid dichotomy between language and content supported by many university teachers (Trent Reference Trent2010).

From a global perspective, the team-teaching culture is more deeply rooted in contexts such as North America than in Europe (Baeten & Simons Reference Baetens and Simons2014: 95). In fact, in the EMI programmes currently implemented throughout Europe and Asia team teaching is rarely found. As a matter of fact, only a handful of the few schemes involving language and content teachers’ working together have gone beyond the pilot project stage (Cots Reference Cots and Doiz2013). Whereas this collaboration between language and content instructors met with success at pre-university level (Pavón et al. Reference Pavón, Ávila, Gallego and Espejo2015), it is practically non-existent at tertiary level (Carpenter II et al. Reference Carpenter II, Crawford and Walden2007; Doiz et al. Reference Doiz, Lasagabaster and Sierra2013a; Aguilar Reference Aguilar2017; Wilkinson & Walsh Reference Wilkinson and Walsh2015). In fact, one of the main conclusions drawn in the ICLHE (Integrating Language and Content in Higher Education) conference that I attended in Brussels in September 2015 was that, in all European contexts where EMI experiences are currently in force, team teaching is a rarity, and the very few presentations on language and content teacher collaboration were very much welcome and unstintingly praised. Whereas there is a strong research current in favour of team teaching at pre-university levels (see Nikula et al. Reference Nikula, Dafouz, Moore and Smit2016), this is far from being the case at university level.

Team teaching can be divided, grosso modo, into two big models (Kerridge et al. Reference Kerridge, Kyle and Marks-Maran2009). In the first model the two teachers interact in class and deal with specific topics from different perspectives and later the students are invited to take part. I discard this first option because at university level it usually turns out to be rather complicated to find the necessary conditions so that two teachers can be working together in the same class at the same time. In the second model, which fits in with my proposal, the content teacher assumes the leader role, while the language lecturer sits in the class assuming a secondary role. The latter focuses their attention on the linguistic side of the class and intervenes when necessary or when their language expertise is requested. However, it is fundamental to establish the conditions under which team teaching will be carried out, which leads us to the first research task.

At university level, the higher the grade, the more complex, specialized and abstract the language needed to successfully learn the content becomes. It is also worth remembering that teaching content in a second language (L2) is especially complex in late bilingual programmes such as those implemented in higher education institutions (Johnson & Swain Reference Johnson and Swain1994; Kong Reference Kong2014; Lo & Macaro Reference Lo and Macaro2015), as in many contexts both teachers and students may have limited or no previous experience of EMI. That is why one of the key aspects of team teaching lies in the clear definition of the roles to be played by each of the members of the team (Thomas Reference Thomas1992; Anderson & Speck Reference Anderson and Speck1998; Baeten & Simons Reference Baetens and Simons2014; Cots & Clemente Reference Cots, Clemente, Knapp and Aguado2015).

The first step should consist in establishing a how-to handbook for successful classroom observation for both those observing and those being observed (Lasagabaster & Sierra Reference Lasagabaster and Sierra2011). Many teachers dislike being observed, as they find classroom observation stressful and intimidating because observers tend to exercise top-down authority (Li Reference Li2009), which is why many professionals express their anxiety and worry when it comes to classroom observation. Most university teaching staff members are unaccustomed to being observed and distrust this practice because they believe that their professional competence is going to be questioned or judged. Hence, the establishment of a congenial relationship between the content teachers and the language expert observers is a basic step as observation must be built on a foundation of trust. It seems reasonable to expect that the careful consideration of content teachers’ views regarding the prerequisites for successful classroom observation will help to improve team-teaching collaboration.

In order to carry this out an anonymous questionnaire consisting of open and closed questions can be devised. It should comprise a series of questions designed to ascertain the opinions, beliefs and attitudes of content teachers regarding classroom observation and the conditions to be met to implement a successful team-teaching experience. The questionnaire should tackle specific questions concerning how useful observation is considered to be by the participants, whether they have ever been observed in class, whether they found previous experiences useful (and if not, why), how the observation process should be done, how feelings of uneasiness and distrust could be overcome, and under what circumstances they would be willing to directly engage in activities of this nature (see Lasagabaster & Sierra Reference Lasagabaster and Sierra2011). The gathering of these data becomes essential if the following tasks are to bear the expected fruits. Moreover, to my knowledge there is no empirical account of EMI university teachers’ beliefs about the role to be played by team teaching in their professional development. If an observation scheme is to be devised, a needs analysis from the teachers’ perspective becomes an indispensable condition. Since content teachers are traditionally reluctant to team teach, they will be surveyed and language teachers will try to reach a consensus and adapt to their partners’ demands.

Teachers’ beliefs are complex and related to the micro- and macro-political context (Barcelos & Kalaja Reference Barcelos and Kalaja2011), and exert a profound influence on teaching practices (Zhang & Liu Reference Zhang and Liu2014). Therefore, researchers should uncover teachers’ beliefs because their unveiling will lead to reflection and may foster new instructional practices such as team teaching. Kubanyiova (Reference Lamb2014) reports that teachers’ practices are influenced by their prior beliefs about how languages are learned and how they should be taught, as well as by the diverse teaching methods they have internalized during their professional lives, to the extent that there is a strong correlation between teaching practices and beliefs (Kuzborska Reference Kuzborska2011). Since team teaching is fairly uncommon in EMI contexts, teachers’ beliefs with regard to this particular intervention need to be researched.

An additional useful tool to delve into teachers’ beliefs would be the organization of focus groups, as they allow the researcher to unearth opinions and beliefs that would probably not surface if the individual questionnaire only were carried out. Both instruments are complementary, but if for any reason (such as teachers’ unwillingness to fill out a questionnaire) only one is to be chosen, the focus group will probably help to collect more comprehensive and enriching data. The researcher would need to design an observation guide in which different issues should be addressed and which, apart from those aimed at the observation procedure itself, could also include questions such as the linguistic aspects EMI teachers consider to be important, which of these aspects they usually pay heed to in their classes, what their students’ main linguistic weaknesses are, which language skills they would prefer language teachers to focus on, or whether language plays any role at all in their assessment of the subject. Such data would allow the researcher to compare teachers’ beliefs and their actual practices (the latter to be observed in Research task 2). Although professional development may change beliefs, the fact is that research suggests there are significant relations between teachers’ beliefs and practices, because both are shaped by pedagogical and cultural traditions (OECD 2014). The discussion groups should focus not only on teachers’ beliefs about classroom observation but also deal with teachers’ beliefs about their roles in EMI programmes, their understanding of academic language, the language demands imposed on their students and their beliefs in the effectiveness and feasibility of team teaching.

Once we have established the conditions under which the observation process is more likely to succeed by bringing teachers’ beliefs to light, we can proceed with the second Research task.

Before elaborating the language-related materials, language teachers need to become familiar with the content to be delivered in EMI courses. To start with, the researcher will need to familiarize him or herself with the teaching materials (books, journal articles, handouts, notes, links to web pages, etc.) to be used during the course. This knowledge is indispensable if effective language advice is to be given. In addition, classroom observation will help the language teacher analyse how language issues are tackled and consider the metalanguage students will need in their specialist domain. In their team teaching, language and content teachers will need to focus on both the macro-level curricular planning and the micro-level interactions that take place during challenging curricular tasks (Schleppegrell & O'Hallaron Reference Schleppegrell and O'Hallaron2011).

Classroom observation will allow language teachers to raise their content awareness, whereas the analysis of the classroom talk will provide them with the necessary data to figure out how content teachers’ and students’ language awareness can be heightened. A possible course of action would be to videotape a class every week over two months. The video recordings will help language teachers to gain an understanding of students’ needs, to observe classroom interactions, and to start analysing the specific genres (the different text types that students have to understand and produce) and the register (the lexical and grammatical features needed to appropriately build the genres) of the content subject. The analysis of the written materials should be combined with the transcription and examination of those video extracts that pinpoint the linguistic hurdles found in the EMI classroom.

An important role in team teaching is to guide students into using the language of the content subject and to make them aware of the different genres. After identifying the specific genres of the content subject concerned, the team teachers should focus on the lexis and grammar that is habitually linked to each specific genre. This should lead to the design of a glossary which should be the basis for content comprehension and to the creation of activities to develop students’ language awareness. The glossary is of the utmost importance because, as Basturkmen & Schakleford (Reference Basturkmen and Shackleford2015) claim, fundamental terms and concepts play a paramount role in the early stages of developing disciplinary knowledge. Nevertheless, the examination of the academic language should not be limited to specific vocabulary and grammar structures, but should also encompass genre conventions and language functions (see the seven discourse functions characterised by Dalton-Puffer Reference Dalton-Puffer and Nikula2016). As Llinares et al. (Reference Llinares, Morton and Whittaker2012: 300) put it, if EMI ‘teachers have clear statements about the genres and registers and associated language through which the content is expressed, then their feedback can be directed precisely to those features’. If content teachers help students to express their ideas in appropriate language, content teaching and learning will benefit from it. The volume by Llinares et al. (Reference Llinares, Morton and Whittaker2012) can be a very practical and illustrative resource in this respect.

What is observed in the classroom video recordings and the examination of classroom materials should become the lynchpin of a language guide that should be made available to both EMI teachers and students. The fostering of students’ attention towards features of English can be done by combining two approaches (Lyster Reference Lyster2007; Llinares et al. Reference Llinares, Morton and Whittaker2012): proactively by preselecting linguistic features to promote noticing and awareness, and reactively by providing corrective feedback to students’ production. A follow-up activity could consist of gathering data through an anonymous questionnaire to unveil whether the guide focused on language is welcomed by both teaching staff and students, whether they find it helpful when it comes to facilitating content learning in the foreign language, and whether the guide addresses all the language issues that they think need to be achieved.

To date, the vast majority of studies that tackle team teaching are based on participants’ impressions (Tan Reference Tan2011) and do not include any measurement instrument that confirms or rejects such impressions. The few studies available have been undertaken at pre-university level (Tan Reference Tan2011; Lo Reference Lo2015) and the effectiveness of team teaching has not been empirically proven. The main reason for this lack of empirical evidence lies in the widespread belief that language learning takes place in an incidental way in EMI contexts (Coyle, Hood & Marsh Reference Coyle, Hood and Marsh2010; Lo Reference Lo2015).

Since there is an overriding need for experimental research in the field of team teaching, the effectiveness of the aforementioned language-focused guide should be tested by comparing an EMI group with no team-teaching support, and a comparable EMI group that had enjoyed the purported benefits of team teaching. The study should measure English competence and content learning in both groups by using the same research tools and following a pretest/posttest design and, ideally, no differences should be found between the experimental and the control group at the initial stage of the intervention. This design would give researchers the opportunity to explore differences between solo-taught and team-taught groups of the same content subject in terms of students’ perceptions and motivational stance (see Research task 6) and achievements (language and content).

The research should be longitudinal because studies indicate that short-term EMI courses may not be sufficient to yield any significant language gain. In fact, Aguilar & Muñoz (Reference Aguilar and Muñoz2014) report that the gains in English proficiency by engineering students who followed a 60-hour course in English were small, especially in the case of grammar, a language skill in which the improvement between the pretest and the posttest did not reach statistical significance. These authors also warn about the need to gauge the effect of students’ proficiency on language gains after completing the EMI course, in the sense that students with an advanced level of proficiency (Oxford Placement Test Band 6) and those with an upper intermediate level (Oxford Placement Test Band 5) did not make any gains in the skills under scrutiny (listening and grammar), whereas only those with the initial lowest English proficiency (elementary level or Oxford Placement Test Band 3) obtained significant gains in both language skills. Therefore, and since less proficient students are the only ones that reaped any language gain in such a short period, the conclusion to be drawn is that a longitudinal design will help researchers analyse in more detail the impact of team teaching by considering students’ different levels of proficiency. A longitudinal study of two to three years will more than likely help to examine language gains in depth and will provide researchers with a much more detailed picture of the effect of team teaching on students. Since synchronic studies are unable to capture the complex L2 + content learning process, several authors (Davison Reference Davison2006; Evans & Morrison Reference Evan and Morrison2011) have called for more longitudinal research, especially from English-medium universities that are not rooted in English-dominant societal and academic contexts. Although different variables (especially from outside the classroom, such as private English classes or English exposure outside the formal learning context) may affect the language proficiency in longitudinal studies, they can also be controlled by administering questionnaires to gather data regarding such variables and to ensure that the control and the experimental groups are comparable. The same applies to content learning (e.g. to control whether students are attending private content classes).

As far as content learning is concerned, the researcher should ensure that the team-taught EMI groups and the solo-taught EMI groups are comparable in the pre-intervention stage, for example by gathering university entrance results or by including any diagnostic assessment at the beginning of the academic term to gauge students’ previous knowledge on a given content subject. After the implementation of team teaching, both groups’ content learning should be compared at different points in time: for example, by comparing final grades in EMI subjects at the end of each semester for two academic years.

In the linguistic sphere, the procedure could be as follows. In the first semester the language teacher researcher will focus on analysing all the materials and the video recordings, whereas in the second semester he or she will start to implement the team- teaching intervention. At the beginning of the first EMI course the researcher will administer a general English proficiency test (e.g. Oxford Placement Test) and will assess students’ subject-specific language proficiency based on their written and/or oral production (the latter assessment will rest on the usual tasks completed in the EMI classes and therefore will not require any additional time, that is, class activities will be used to measure students’ subject-specific language proficiency); the same type of general English proficiency test will be administered and similar activities will be performed at the end of the course once the linguistic advice presented in the guide has been given to the participants. Due to the longitudinal nature of the intervention proposed, the language tests should be carried out at the end of every term and should span over two to three years, so that the impact of team teaching can be scrutinized in the long run. This will allow the comparison of results at the pre- and post-team- teaching intervention and gauge whether the cross-curricular collaboration has helped to reap the expected linguistic benefits. Researchers can also focus on what interaction patterns take place in EMI settings and which are more likely to foster content and language learning.

Improving pronunciation is one of the language aspects that many EMI teachers deem essential (Ball & Lindsay Reference Ball, Lindsay and Doiz2013), which is why it could become another research avenue. Based on the weaknesses detected in the video recordings, short pronunciation training sessions could be delivered at the end of the class for only ten minutes (the content teacher may be likely to allow this every 15 days if it only requires a short span of time), and the researcher could analyse how this short training affects both students’ pronunciation and their confidence as EMI users by the end of the course. An oral presentation in the content classroom could also be video recorded and, combined with an interview, serve as a basis to check how students’ pronunciation bears on their EMI ideal self (see Research task 6 for further information on the ideal self; pronunciation could be analysed as part of students’ ideal self).

A mostly untapped research avenue that deserves further attention is concerned with the impact of team teaching on subject-specific language development, as research studies have focused on general English proficiency since research has mainly been undertaken by applied linguists who are not content teachers. It is my contention that, among the different language aspects to be considered, the language guide should also focus on the use of the first language (L1) in EMI classes, an issue that deserves further attention (see Research task 5).

Researchers should entertain the two research questions put forward by Dafouz & Camacho-Miñano (Reference Dafouz and Camacho-Miñano2016), namely how the use of different assessment formats impact on solo-taught and team-taught students’ academic results, and whether the student distribution and results by level of academic performance (low achievers, medium achievers and high achievers) are the same in both groups.

In the case of assessment formats, certain types of assessment may be more demanding for EMI students, such as oral presentations, written essays and active participation in seminars, in contrast to others less linguistically demanding, such as mathematical problem-solving exercises. The focus on different types will help to avoid the negative effects that assessment may provoke, especially among lower-achieving students. In the literature, two main types of assessment are often distinguished: summative assessment (also known as assessment of learning), which is administered at the end of the learning process to judge students’ mastery of the material, usually in the form of an exam; and formative assessment (assessment for learning), that is, an ongoing process that is administered through various means during the term or academic year with a view to continuously assessing students to identify learning gaps and work to fill them. The latter requires a student-centred approach in which teachers act as mediators (McCallum, Hargreaves & Gipps Reference McCallum, Hargreaves and Gipps2000). Stiggins (Reference Stiggins2007: 30) claims that ‘Assessment for learning (as opposed to assessment of learning) has a profoundly positive impact on achievement, especially for struggling learners, as has been verified through rigorous scientific research conducted around the world’.

In this vein, Dafouz & Camacho-Miñano (Reference Dafouz and Camacho-Miñano2016: 65) underscore the need ‘to establish whether there could be a relationship between students’ English language proficiency and the language demands of the aforementioned assessment formats used’. Since students’ English proficiency is one of the main concerns of EMI teachers (Dearden Reference Dearden2014; Doiz, Lasagabaster & Sierra Reference Doiz, Lasagabaster and Sierra2014), this is an issue well worth considering. It goes without saying that the content and assessment formats should be the same for both the experimental and the control groups.

As for the different levels of academic performance, the misgivings about the selective nature of EMI students harboured by some authors (see Pérez-Cañado Reference Pérez-Cañado2016) do not apply, because our objective is not to compare EMI and non-EMI university students, but rather solo-taught and team-taught EMI students to observe whether team teaching is equally beneficial for low, medium and high achievers. This will also open up the possibility of studying how assessment types and students’ academic performance interact (Basse Reference Basse2016).

The use of stakeholders’ L1 or codeswitching has been under-researched in EMI settings. The spread of English-only language policies in many contexts (Swain, Kirkpatrick & Cummins Reference Swain, Kirkpatrick and Cummins2011; Mazak & Carroll Reference Mazak and Carroll2017) has exerted pressure on content teachers to avoid the use of their other languages in EMI classes. Consequently, content teachers have felt some sense of guilt about using their own and their students’ linguistic resources (Swain et al. Reference Swain, Kirkpatrick and Cummins2011; Macaro Reference Macaro, Barnard and McLellan2014) because the use of students’ first languages ‘has tended to be disparaged by textbook writers, methodologists and educational policy makers in many countries’ (Barnard & McLellan Reference Barnard and McLellan2014: 2). On some other occasions content teachers do not have any indication from their institutions on the use of the L1, which leads to various courses of action.

Studies have revealed the disparity in content teachers’ beliefs about the use of the L1 in both English as a foreign language (Macaro Reference Macaro, Turnbull and Dailey-O'Cain2009) and EMI classes (Doiz & Lasagabaster Reference Doiz, Lasagabaster, Mazak and Carroll2016b). This disparity ranges from non-acceptance to acceptance of L1 use, although most EMI teachers tend to exclude students’ mother tongue. Nevertheless, content teachers admit to resorting to the L1 in some circumstances, which reveals a mismatch between beliefs and practices (Doiz & Lasagabaster Reference Doiz, Lasagabaster, Mazak and Carroll2016b). Costa (Reference Costa2012) claims that university EMI teachers and students resort to the L1 when an explanation or a translation of a word or whole expressions is required. Conversely, Doiz & Lasagabaster (Reference Doiz, Lasagabaster, Mazak and Carroll2016b) observed through focus groups that the use of the L1 is not typically accepted in classroom interactions and assessment tasks, although, curiously enough, it is quite common outside the class context (i.e. office hours and one-on-one interactions). The divide between inside and outside class seems thus to be worth investigating.

In an overview article on the direction classroom codeswitching research should take, Macaro (Reference Macaro, Barnard and McLellan2014) proposes some research tasks that need to be undertaken in order to shed light on some pressing questions that still remain answered. The tasks are the following (p. 21):

1. 1. To investigate how EMI teachers codeswitch when it comes to explaining complex vocabulary and concepts and whether their explanations are more effective if some L1 information is used. To carry out this research task, the researcher first selects classroom extracts in which the teacher codeswitches to explain some new vocabulary/concepts. Then students are invited to view the extracts and judge the efficiency of the use of the L1 on a three-point scale: very efficient, efficient or not efficient. Afterwards, they are asked to justify their choice. Once this is done, the researcher analyses the students’ responses and tries to reach some conclusions as to when and how the use of the L1 renders more effective.

2. 2. To study whether codeswitching is equally beneficial to all students irrespective of their proficiency level. That is, do lower proficiency EMI students benefit more from codeswitching than high proficiency students? The information gathered in the task in the previous paragraph can be divided according to students’ language proficiency, which will allow the researcher to scrutinize whether their judgements about L1-use efficiency vary depending on their proficiency level.

3. 3. To examine EMI teachers’ and students’ attitudes towards codeswitching, since studies specifically focused on EMI university stakeholders’ beliefs are rather scant. The stakeholders could be asked about their attitudes towards codeswitching by means of questionnaires and/or focus groups.

Team teaching could play a definitive role with regard to these research objectives, as shared reflection could help team teachers draw conclusions about the more adequate use of the L1 and could help formulate guidelines (that is, specify when and under what circumstances content teachers should resort to the L1). These guidelines will help to settle some content teachers’ inner conflict in their attempt to find a balance between some universities’ official English-only language policy and the possibility of using students’ whole linguistic repertoire to help them better grasp the lecture content (Tien Reference Tien, Barnard and McLellan2014). The data gathered through classroom observation could serve as the basis to develop content teachers’ language awareness and to challenge their more-often-than-not monolingual view of language codes (Cook Reference Cook and Chapelle2012; Lin Reference Lin2015). Reflection will help team teachers to establish some guidance, since the use of students’ linguistic repertoire can foster students’ engagement in the classroom (Lewis, Jones & Baker Reference Lewis, Jones and Baker2012).

According to Swain & Lapkin (Reference Swain and Lapkin2013), the literature indicates that the L1 is used for languaging both cognition and affect, but researchers need to look into the influence of this L1 languaging on L2 development: ‘We need both qualitative and quantitative studies, descriptive and experimental studies that focus on this issue’ (p. 123). Lin (Reference Lin2015: 80) is more specific and underscores that research on the impact of L1 use should try to fill the gap regarding the lack of focus on the planned use of L1 in the curriculum. This author asserts that the literature analyses the teacher's unplanned, spontaneous spoken use of L1, but systematic L1 use should also be considered, as it can help to reach the curricular goals.

Motivation is a keyword in any discussion about the learning process, as it is ‘a highly important determiner of success in all forms of instructed knowledge acquisition’ (Henry Reference Henry2012: 24). Studies (Anderson & Speck Reference Anderson and Speck1998; Carpenter II et al. Reference Carpenter II, Crawford and Walden2007; Lester & Evans Reference Lester and Evans2009) indicate that, as each teacher involved in team teaching concentrates on different aspects, their collaboration promotes students’ motivation, since students receive increased support and even individual attention. Researchers will have to assess whether greater attention to the language component affects students’ motivation in any significant manner. The impact of team teaching on both teachers’ and students’ motivation thus offers many interesting options for further research.

Team teaching can improve feedback of students’ oral and written performance, as language teachers may discuss students’ written work and oral presentations in order to provide solutions to their linguistic problems. This support should contribute to students’ interest, motivation and learning. As for content teachers, team teaching can create a sense of support, community and collaboration (Baeten & Simons Reference Baetens and Simons2014), which should also enhance their motivation, but this also needs to be researched. Even if the potentially negative aspects of team teaching (such as struggles across professional domains and the uncertainties involved) come to light, they will undoubtedly be worth examining, because these experiences will help to improve future team-teaching enterprises. The research task would vary depending on the number of teachers/students involved. Were the researcher able to gather information from a large number of participants, a questionnaire should be designed in order to analyse what particular aspects of motivation have been affected by team teaching.

The three components of Dörnyei's (Reference Dörnyei2005, Reference Dörnyei, Dörnyei and Ushioda2009) L2 Motivational Self System could help to structure the motivation questionnaire and divide it into three sections:

1. 1. Ideal EMI teacher/student self: Lack of motivation is likely to be caused by an underdeveloped ideal self. The ideal L2 self is an effective motivator if the teacher/student has a desired future self-image. Thus, the more positive the individual's disposition towards the L2 and its speakers, as well as towards L2 use is, the more attractive their idealized L2 self will be. It is assumed that the EMI teacher/student will be motivated to endeavour to reduce the discrepancy between their actual and ideal teaching selves. The individual's motivation will stem from their self-image. Teachers/students are expected to have a strong image of themselves as speakers of English, which allegedly will inspire them to strive to become highly proficient speakers of this language. In this vein, self-competence beliefs should also be investigated, as they are conceptualizations of estimates of ability in particular activities (in our case in EMI settings) and expectations of future performance.

2. 2. Ought-to EMI teacher/student self: The ought-to self concerns the attributes that one believes one ought to possess to meet others’ expectations and to avoid possible negative outcomes. It therefore embodies the EMI teacher's/students’ representation of their obligations and responsibilities which are dependent on extrinsic incentives which may bring about negative consequences if these are not satisfied. This would encompass colleague- or institution-induced views about EMI that may be in conflict with the individual's ideal self. For example, a teacher/student who may not be very interested in taking part in EMI experiences but who eventually joins the programme to align with their colleagues and live up to the institution's expectations, e.g. a higher education institution that regards EMI as a fundamental element of its internationalization process and expects the teaching staff/students to take the plunge.

3. 3. EMI experience: The third component relates to how the teacher's/student's experience in EMI teaching contexts, the language policy implemented by the institution, or the teachers’/students’ reactions to EMI courses affect the individual's motivation. Studies undertaken in countries such as Hungary (Csizér & Lukács Reference Csizér and Lukács2010), Indonesia (Lamb Reference Lamb2007), China, Iran and Japan (Taguchi, Magid & Papi Reference Taguchi, Magid, Papi, Dörnyei and Ushioda2009) indicate that the ideal L2 self significantly affects a learner's L2 motivation. However, the impact of EMI on students’ possible selves (representing the individuals’ ideas of what they might become, what they would like to become, and what they are afraid of becoming) has hitherto hardly been touched upon (Lasagabaster Reference Lasagabaster2016). Since EMI is mushrooming in education systems all over the world, this seems to be a question well worth studying, namely, whether EMI affects both the ideal L2 self and the ought-to self, or whether one prevails over the other.

If the EMI group is small, or simply if qualitative data are also to be collected, then Ushioda's (Reference Ushioda, Dörnyei and Ushioda2009) approach to motivation should be considered. This author claims that a focus on the person rather than on abstract learners (and individual differences in a theoretical sense that provide uniform accounts, for example, by referring to the statistical mean) is also needed, because teachers and students are necessarily located in particular cultural and educational contexts in which their experiences and self-states may facilitate or hinder their learning engagement. Ideally, a mixed methods approach will help to enhance the quality of the data. Therefore, the researcher could rely on discussion groups to complement the quantitative data. The discussion groups add a human dimension to impersonal data while the participants talk freely in interactive groups and their discussion sheds light on shared beliefs; the method also allows individual differences to be voiced. This tool also helps to uncover contradictions that may not be detected in the questionnaires; however, it does enable the researchers to delve into insights that could otherwise remain hidden. Dörnyei's aforementioned tripartite motivational system can also serve as the basis for the questions to be tackled in the discussion groups (see Doiz & Lasagabaster Reference Doiz, Lasagabaster, Doiz and Lasagabaster2016a). The analysis of these three components before and after the team-teaching experience from both a quantitative and a qualitative perspective will provide the researcher with a very vivid picture of the motivational impact of the intervention based on the currently most influential L2 motivational model.