EMI LECTURER TRAINING PROGRAMMES AND ACADEMIC LITERACIES: A CRITICAL INSIGHT FROM ESP

As EMI involves the integration of content and discipline-specific communication, this study focuses on intersections between ESP and EMI, especially on how ESP can contribute to improving discipline-specific communication in EMI from the perspective of content lecturers’ needs. Specifically, this study provides an overview of EMI training programmes offered by universities in Catalonia (Spain), and explores the written genres assigned by content lecturers in EMI subjects. EMI training programmes are classified according to their orientation: communication, pedagogy, and multilingualism/multiculturalism (Fortanet-Gómez, 2010; Kurtán, 2003). Our analysis pays special attention to the extent to which there is an ESP focus in such training and is complemented by a questionnaire to a group of EMI lecturers on their reported use/teaching of disciplinary genres, following Nesi and Gardner (2012). Findings shed light on practices and expectations related to discipline-specific genre pedagogy, an area that is at the crossroads between EMI and ESP. By examining lecturer perspectives on EMI alongside institutional policies and training programmes, this study can help lecturers cope with the challenges of EMI, and contribute to further developing EMI-ESP lecturer collaboration.


INTRODUCTION
English-medium instruction (EMI) is a global phenomenon, and scholarly interest in EMI is gaining momentum (Dearden, 2014;Fenton-Smith, Humphreys, & Walkinshaw, 2017).Implementation of EMI in Europe tends to be embraced as a means to integrate English and discipline-specific communication at university level, often at the expense of English for Specific Purposes (ESP) (Arnó-Macià & Mancho-Barés, 2015;Räisänen & Fortanet-Gómez, 2008).Given the burgeoning number of EMI programmes, it is important to study how ESP research and pedagogy can lead to a more informed approach to EMI.In this paper, the term EMI is used to denote a concept that has been associated with a profusion of terms.Content and Language Integrated Learning (CLIL) involves a dual focus on language and content learning, at all educational levels.The terms Integrated Vol.5(2)(2017): 266-290 has transformed ESP practice and involves multiple challenges for teachers and learners alike (Yang, 2016).
One of the challenges to interdisciplinary collaboration in an EMI context is that the implementation of EMI is left almost exclusively in the hands of content lecturers in an institutional context of compartmentalised departments and areas of knowledge (Arnó-Macià & Mancho-Barés, 2015).Set in the same scenario of EMI implementation in Catalan universities, this study adopts a complementary, 'behind the scenes' perspective.Rather than looking at EMI classes, we take a step backward and, using an ESP lens, we look at the EMI training offered to lecturers as well as the disciplinary discourse practices that lecturers reportedly use in their teaching.First, we provide an overview of the training programmes offered to (prospective) EMI lecturers across the seven public universities in Catalonia (Spain).After classifying each programme according to the dimension(s) it addresses -namely, communication, pedagogy or multilingualism (Fortanet-Gómez, 2010;Kurtán, 2003) -we look at the 'ESPness' of those programmes, i.e. whether, and to what extent, they are approached from an ESP perspective or whether they are presented as ESP for lecturers.Second, in order to find common ground for interdisciplinary collaboration, we take a closer look at EMI lecturers' self-reported use of specific disciplinary discourse practices in their teaching based on the notion of genre, which is key to the interplay between ESP and EMI.

LITERATURE REVIEW
Two important areas emerge in relation to the objectives of this study: first, the characterisation of EMI lecturer training programmes in the literature, which will allow us to analyse and classify the programmes studied in this research; and second, the teaching of specialised genres, since learning discipline-specific writing is part of disciplinary learning and can be examined from both a content and an ESP perspective.The specialised genres analysed belong to the STEM (Science, Technology, Engineering and Mathematics) disciplines and are typically found in university settings (Fortanet-Gómez, 2010;Lea & Street, 1999).

Dimensions of EMI training programmes
Lecturer training appears to be one of the most urgent needs in the implementation of EMI.Going beyond language improvement, researchers have identified different components to be included in training programmes.In order to account for the EMI training initiatives offered by Catalan universities, we use the categorisation proposed by Kurtán (2003), and later developed by Fortanet-Gómez (2010), to identify the three main needs that training programmes should address: Vol. 5(2)(2017): 266-290 (i) communication and specific purpose language use; (ii) pedagogy and didactics; and (iii) multilingualism and multiculturalism.
Fortanet-Gómez's (2010) discussion of these three dimensions of EMI training in the context of Spanish higher education is particularly relevant to this study.Regarding the first dimension, communication and specific language use, Fortanet-Gómez (2010) refers not only to subject-specific language use, but especially to classroom discourse, in both native and non-native-speaker contexts.Thus, training may provide lecturers with resources for increasing students' attention and motivation, improving comprehension, and promoting interaction.The second dimension, pedagogy and didactics, refers to raising lecturers' awareness of the fact that translating lectures is not enough and that students' language level and needs should be considered.The third dimension addresses the role of multilingualism and multiculturalism, especially important in today's multilingual/multicultural classes and the internationalised settings in which EMI takes place.In our context, EMI classes comprise mostly local students; this is because institutional policies promote EMI as a means to develop students' language competence (Arnó-Macià & Mancho-Barés, 2015) rather than from the real need for a lingua franca.
Studies on EMI teacher training programmes in a context similar to that of our study show that the selection of their contents mostly correspond to communication (dimension 1) and pedagogy and didactics (dimension 2), for example, Ball and Lindsay (2013) in Spain, or Cloke (2017) in Italy.Applying ESP course design, the course proposed by Ellison, Aráujo, Correia, and Vieira (2017) resulted from a needs analysis and can be classified as an EAP course.The resulting scenario, according to Greere and Räsänen's (2008) continuum, would be one of non-integration, which is common in university settings, where EMI lecturers' pedagogical objectives tend to focus exclusively on content.

Teaching specialised genres
Mastering disciplinary discourses is an intrinsic part of disciplinary learning, which calls for raising students' awareness of specialised genres and their explicit teaching and assessment through the developing movement of "Writing in the Disciplines" (WID).Deane and O'Neill (2011) characterise WID as the integration of writing and disciplinary knowledge, considering writing (instruction) an integral part of disciplinary induction, with content lecturers being the most knowledgeable of the generic conventions in their own fields.And yet, EMI lecturers' self-reported low English language proficiency levels hinder them from taking on the role of disciplinary language experts (e.g.Pulcini & Campagna, 2015).It is therefore unsurprising to find that EMI lecturers do not regard themselves as disciplinary language teachers, but simply as subject specialists (Airey, 2012).
Vol. 5(2)(2017):  This situation calls for the collaboration of content and language instructors to teach discipline-specific texts.Lyster (2017) supports interdisciplinary collaboration initiatives to integrate content and language at university through ICLHE.Zhu (2004) reported on content lecturers' perspectives on the importance of teaching disciplinary writing by focusing on discipline-specific texts, with the collaboration of EAP lecturers for systematic academic writing instruction.
In the teaching of disciplinary writing, ESP plays a central role.Genre is a key notion underlying ESP (Dudley-Evans & St John, 1998), providing a principled approach to understanding how disciplinary texts are produced, so that students can gain insights into the contextual, discourse, and language features needed to produce successful texts in the relevant discourse community.Accordingly, given the importance of disciplinary writing in EMI, the integration of ESP-inspired academic writing pedagogy must be rooted in the notion of genre.
Work on academic genre analysis has developed through the compilation of corpora, including the MICASE (Michigan Corpus of Academic Spoken English), collecting data from the US (Swales, 2002), and the ELFA (English as a Lingua Franca in Academic Settings) corpus in Europe (Mauranen & Ranta, 2008), collecting academic speech from non-native speaking contexts.Regarding academic writing, the relevant corpus is the British Academic Written English corpus, BAWE (Nesi & Gardner, 2012), which contains assignments from a wide range of disciplines.Our study is based on the classification of genres from the BAWE corpus according to their social function, with genres sharing similar social purposes grouped into genre families.Table 1 shows the most common social functions in academic writing and their corresponding genre families.Nesi & Gardner, 2012: 36) This categorisation gives visibility to the variety of genres that students are expected to write, and can thus provide common ground for the collaboration between content and ESP lecturers, to inform academic writing pedagogy.

THIS STUDY
This study aims at providing an ESP perspective on the implementation of EMI in the context of public universities in Catalonia (Spain), especially with regard to helping content lecturers address challenges involved in EMI by raising their own (and their students') awareness of the academic literacies that form part of discipline-specific communication.The analysis of EMI training programmes offered by Catalan universities and of content lecturers' teaching of specific genres can provide ESP researchers with useful information that can lead to further collaboration between both.

Research questions
In order to attain the abovementioned objectives, two research questions are posed: RQ1: To what extent are university policies promoting EMI backed by institutional support to lecturers?RQ2: What are the self-reported teaching practices of EMI lecturers and their attitudes towards the teaching of specialised genres?The answers to these two research questions can provide useful information to the stakeholders involved in EMI -policy makers, content lecturers, and ESP lecturers -to help them streamline courses and actions within a coherent approach to EMI.

Data and methodology
This study includes two types of data: (i) institutional policy documents and websites that allowed us to analyse the training and support given to (prospective) EMI lecturers; and (ii) questionnaire data from a small group of 16 experienced lecturers from the field of STEM (from one of the universities analysed).They were selected on the basis of their participation in at least one EMI training course.
The first set of data consists of university documents on internationalisation and language policy, which were analysed to identify general policies underlying the implementation of EMI.Attention was paid to actions and policies aimed at promoting EMI and supporting its implementation.The universities analysed for this research include all seven public universities in Catalonia: Autonomous University of Barcelona (UAB), University of Barcelona (UB), University of Girona (UdG), University of Lleida (UdL), Polytechnic University of Catalonia (UPC), Pompeu Fabra University (in Barcelona, UPF), and Rovira i Virgili University (in Tarragona, URV).Following this general policy analysis, a detailed analysis of the specific webpages listing EMI training courses (and other support for lecturers) 23  272  Vol. 5(2)(2017): 266-290 helped characterise such courses in terms of their scope.The titles and course descriptions published on the university websites in 2012-2013 were analysed to classify the courses according to the three dimensions discussed above: communication, pedagogy and multilingualism/multiculturalism (Fortanet-Gómez, 2010;Kurtán, 2003).After this general classification, a more detailed analysis of each course description allowed us to see the extent to which they incorporated an ESP perspective.The policy documents and websites analysed are tabled in Appendix 1.
The second strand of this research focused on the academic literacy practices of EMI lecturers from one of these universities (UdL).A group of lecturers in the STEM field completed a questionnaire (Appendix 2), which aimed at academically profiling the respondents and finding out about their reported practices and opinions on academic written genres.The questionnaire was sent to 59 lecturers (in 2015-16), who had participated in EMI training courses at the UdL.Sixteen lecturers from different degree areas responded and, for the present study, the questionnaires selected for analysis were those answered by STEM lecturers (n=9), six females and three males, averaging 3.5 years of EMI experience.Their teaching profiles are summarised in Table 2. Seven of these nine lecturers included written assignments in their responses (while the other two lecturers only mentioned either oral assignments or multiple choice questionnaires).Thus, the answers that were analysed in detail were those that involved written genres, corresponding to lecturers 2-8.For the qualitative analysis, three open-ended questions were selected to characterise the written genres dealt with in content classes and to provide an account of the lecturers' understanding and practices regarding their teaching/use of genres (with Vol.5(2)(2017): 266-290 questionnaire responses usually followed up by an email interview for further clarification).
The first question analysed (#5) elicited information on the written assignments in EMI courses: Could you briefly describe the assessed activities in which your students have to write a text in English?Participants' answers were classified according to Nesi and Gardner's (2012) genre family typology (see Table 1 above).In cases where the activity description provided by the lecturers was too broad, further information was obtained from participants using an individual email interview.
The second question analysed (#7) dealt with the assessment of those assignments (including instructions, if any, and the feedback provided): Once the students submit the activities, do you give feedback on the content of the activity and/or on how the submitted text is structured, that is whether or not it follows the formal conventions that such a text needs to have?Academic literacy is approached from the notions of genre and discourse community, which views disciplinary communicative practices as acceptable and recognised by the relevant community.In order to capture their classroom practices, lecturers were encouraged to give examples.
The third question analysed (#8) dealt with the lecturers' opinions about the importance of teaching genre writing in class: To which extent is it important to practise the formal conventions of academic/technical texts in class?

Documents and training programmes
The analysis of institutional policies revealed a certain awareness of the need to provide EMI lecturers with support and training.This support is aimed at encouraging EMI and easing its implementation in a societal context described as having a general low English proficiency (European Commission, 2012).Support actions in policy documents reflect each institution's strategic goals.This section analyses each type of institutional support and its outreach.A general overview of university websites and language policy documents reveals three general goals and actions: Vol. 5(2)(2017): 266-290 (i) support/training: offering different types of courses, resources, and grants or support to translate or produce materials; (ii) recognition: including rewards such as increasing the credit value of EMI courses, or the consideration of foreign language skills as a merit in lecturer selection and promotion; and (iii) quality assurance: establishing requisites for lecturers to be eligible for EMI, incorporating language merits in selection/promotion, as well as those teacher support/training actions that acknowledge the challenges involved in the integration of language and content.In EMI policies, quality appears as a key word and is considered both an aim and a strategy, at the intersection between support and recognition.
These general goals and actions offer a picture of the institutional perspectives on EMI.The question that then arises is how these policy aims are reflected in actual actions undertaken by universities to provide the support laid out in policy documents.
Below is a summary of the actions undertaken (or at least announced on the websites) by universities to support and facilitate EMI implementation.First, an overview is provided of the support offered to prospective EMI lecturers (Table 3).The characteristics of these resources, together with their underlying approach, are analysed vis-à-vis the language policy of each university.
Table 3. Overview of resources offered to EMI lecturers Raising lecturers' level of English appears as the most immediate need, since all universities offer language courses at B2 level or above, corresponding to "independent user" according to the Common European Framework of Reference (CEFR) (Council of Europe, 2001).For example, UdG offers speaking and writing courses (B1-C1.2),while URV offers language courses on demand, and awards grants for external courses.Most universities also seem to envisage EMI as the integration of content and language, judging from the CLIL approach of their methodology courses.For example, UPC in its "Training Programme for English in Teaching" lays out a range of modalities, namely general language proficiency, academic English, coaching, and methodology, the latter based on CLIL.Following a similar rationale, the UB programme also combines language proficiency and Vol.5(2)(2017): 266-290 CLIL methodology (although in an integrated manner).Training courses, the most common type of support, will be analysed below.In addition, universities offer other support in the form of grants or other types of funding, especially to cover the language review of EMI materials.This type of support is envisaged in language and internationalisation policy documents as a means to attract lecturers towards EMI (in contexts where English as the medium of instruction is still scarce).It is somewhat surprising that although methodology courses for lecturers adopt a CLIL approach (apparently advocating for the integration of language and content), no provisions are made, in curricula or policy documents, for such integration either in terms of adjunct courses or full-fledged CLIL (via team teaching, for example), which thus leaves any potential integration of language in the hands of content lecturers alone.Therefore, the role of language specialists is 'behind the scenes', offering support or training.
As discussed above, quality is a key word in institutional policies.It appears in all documents analysed, especially in relation to "language quality", one of the essential institutional aims in all language plans, either because "(language) quality" is highlighted in the document (e.g. in the preamble, as a key aim/action) or because it appears frequently.At UPF the word "quality" is highly prominent, as part of the name of the unit responsible for EMI training ("Centre for Quality and Teaching Innovation").One of the "strategic lines" of the UB Language Plan is "to improve language quality", while "language quality" is a specific objective of UdG policy.Quantitatively speaking, the expression "language quality" appears 19 and 11 times, respectively, in the UB and UdG documents."Language quality" also appears as one of the objectives of the UdL -"to improve the language quality of texts produced in the working languages in all areas".Similarly, the UPC Language Plan also emphasises "language quality" in its mission (together with "knowledge, use and interaction") and is one of its "strategic areas".Specifically, one of its aims is to "ensure high standards of multilingual communication" through the promotion of "self-sufficiency" of UPC members.Some policy documents also specify quality assurance mechanisms, which include requisites for the eligibility of EMI lecturers (a PhD dissertation in English, a C1 certificate, or a stay abroad) in UB policy, with the subsequent commitment required of those lecturers, who, after training, may be asked to teach in English.The UdG language policy document specifically links quality and lecturers' aptitude (two of the stated objectives, #4 "To ensure linguistic quality in teaching", and #5 "To ensure the linguistic aptitude of academic staff").Objective 4 is further developed in terms of different types of institutional support ("language resources and specific terminology for each discipline", "review of teaching material", "grants [to support] the linguistic quality of teaching materials"), with special emphasis on training, including "English for Specific Purposes courses for lecturers".Another action stated in the plan is to "elaborate a questionnaire for all faculty to find out about their readiness and aptitude to be able to teach in a [foreign] language", which reveals certain institutional commitment with quality.What remains beyond the scope of this Vol.5(2)(2017): 266-290 study, however, is whether and to what extent institutional plans have actually been put into practice and what the outcomes are.
This support for EMI lecturers is also reflected on institutional websites.UAB, for example, has a portal to support the integration of language and content.Placing special emphasis on multilingualism and EMI, UPF offers a "support program for multilingualism in teaching" (Multilingualism Action Plan), intended to "guarantee quality" and "level of specialization of materials" in the first years of implementation of EMI (correction, translation services).On the other hand, UPC offers, on demand, a training program linked to its internationalisation plan, which includes general language and CLIL courses, which (like other institutional plans) also pivots around the notion of "quality".Specifically, under the heading "English for teaching" (Anglès per a la docència), it states: Teaching programme linked to the UPC Internationalisation Plan, addressed to lecturers who, as indicated by their school, are planning to teach through English.General purpose of the training programme: to facilitate teaching in English with an optimal level of quality.
Like other universities, UPC also offers support for the linguistic revision of teaching materials and a website with self-access resources.The Operational Plan for Multilingualism at UdL 2013-2018 provides support to EMI lecturers through the coordination of different institutional units: UdL, through the coordination of the Institute of Languages, ICE-CFC [Centre for Professional Development] and the Department of English and Linguistics (DAL), must provide lecturers with support to facilitate English Medium Instruction.
Taking a closer look at the training offered, the programmes were analysed to find out about their interrelation with ESP, and were classified according to the dimension(s): (1) communication and specific language use; (2) pedagogy and didactics; and (3) multilingualism and interculturality.
An overview of the courses is shown in Table 4. Based on the course descriptions available on university websites, for each course the following elements were considered: approach, number of teaching hours, entry-level requirements, and contents (i.e.syllabus).The courses were then classified according to the dimension(s) addressed.
Vol. 5(2)( 2017  Training usually consists of short courses, mostly within dimensions 1 (communication) and 2 (pedagogy).For example, the course "The teaching of content in English" (UPF) integrates "classroom management" (dimension 2) and "spoken and written communication in the classroom" (dimension 1).On the other hand, the course "CLIL" (UPC) focuses on dimension 2, while the "English for Teaching Purposes" course (UAB) focuses on 1.These findings resonate with the studies on EMI training programmes reviewed in section 2.1, although the methodology courses analysed here seem to advocate for the integration of language and content (i.e. a CLIL approach).In contrast, dimension 3 is mostly absent -except for one syllabus component, "interculturality", in the UdL course.This absence may be explained by the lack of experience in multiculturalism/multilingualism in Catalan universities, whose most urgent need is language improvement and the challenges involved in the shift to teaching in a foreign language.The analysis of EMI training also identified the 'ESPness' of the courses classified as dimension 1, language and communication.In this sense, descriptions suggest that an ESP approach is taken (or, at least, included) in most courses.The title "English for Teaching Purposes" (UAB), whose syllabus includes "academic language functions", is quite revealing.Although the other programme classified as dimension 1 (the tailor-made courses at UdG) does not specify syllabus contents, it includes both general English and ESP courses on the course listing and as part of the objectives of the multilingual plan (which specifies "English courses for teaching and research", Action 3.4, page 15).
The courses combining dimensions 1 and 2 often subsume language within teaching guidelines, so that course titles reflect their pedagogical orientation (teaching content in English, CLIL, etc.), with ESP elements in the syllabi.The syllabus for "Basic skills and tools to teach content subjects in English" (UB), with a broad inventory of items ranging from classroom language to CLIL, includes one item called "English for Specific Purposes".Thus, the courses combining pedagogy and language improvement incorporate key notions in ESP, such as academic language or communication in the disciplines, which can contribute to raising EMI lecturers' awareness of academic/disciplinary communication.The inclusion of an ESP component in EMI training courses may be due to the collaboration between ESP lecturers and EMI planners, as in our context ESP lecturers are often involved in EMI training (like the authors of this study).The ESP elements found in course descriptions are identified in Table 5.However, in such short courses (presumably taken by lecturers from different disciplines), academic communication probably remains at a very general level without capturing the nuances of discipline-specific communication.These course descriptions (together with our personal experience as EMI trainers in two of the universities above) lead us to think that EMI training does not include discipline-specific communication, and in particular, the notion of academic genres.This absence calls for further insights into how those genres are actually used in EMI classes, which is a question that is partly answered in the second part of this study.
In sum, all university policies promote EMI, encouraging lecturers to take part in it.With language proficiency being one of the barriers, language courses for lecturers are considered a priority.Support is also provided through specific courses on pedagogy, usually offered to a limited number of participants, and requiring economic and staff resources (coaching, practical sessions, etc.).Documents also include other support actions, e.g.grants for the review of materials, but it remains to be seen whether they can be implemented (or sustained over time) due to the costs involved.

Questionnaire to lecturers
The above overview of EMI training programmes provides an idea of what views of EMI are espoused by institutional policies, which lays the ground for making proposals from an ESP perspective.Precisely, of special interest is to delve into the genres that students are expected to write in EMI.The three groups of answers to the questionnaire below aim to find out which written genres are planned in EMI classes, and how content lecturers approach them.

Assessed written genres
The answers to Question #5, "Could you briefly describe the assessed activities in which your students have to write a text in English?", were analysed according to the social function of the genres.For the social function of demonstrating knowledge and understanding, lecturers report having planned Exercises in different contexts and formats, short-answer questions in examinations (Lecturers 3 and 4) or as part of a portfolio (Lecturers 3 and 7).For example, Lecturer 3's portfolio includes questions on multispectral remote sensing images (e.g."Why do you think that the histogram of the image band B2F_blue does not start in 0 or a value close to 0?").Developing independent reasoning is the social function of the Anthropometry essay, "Assess your own food habits" (Lecturer 2), classified as Critique rather than essay, as it involves the application of critical thinking skills.Other genres include laboratory reports (Building research skills), assigned by Lecturers 2 and 8.As for the social function of preparing for professional practice, Lecturer 5 has planned a written assignment which can be classified as Proposal, given that students are expected to respond to a professional situation -e.g dog breeders requiring clinical tests for tracking dog parentage -by submitting a report containing a literature review of research on such tests, a proposal for a clinical diagnosis protocol, cost analysis and managerial information addressed to potential clients.Other minor genres included in the report are Design Specifications (optimal testing of diagnosis protocol) and Literature Survey (accounting for published papers on similar protocols).As part of the same Vol.5(2)(2017): 266-290 assignment, but different from the Proposal, Lecturer 5 also assigns another writing activity that can be classified as Empathy Writing, specifically a promotional text of the diagnosis protocol.She gives students specific guidelines for writing that type of text (a leaflet), placing special emphasis on the notion of audience:

Lecturers' feedback/teaching of genres in the classroom
The answers to Questions #7 and #8 have to do with the lecturers' self-reported teaching practice.As for Q7, lecturers were asked: "Once the students submit the activities, do you [EMI lecturer] give feedback on the content of the activity and/or on how the submitted text is structured, that is whether or not it follows the formal conventions that such a text needs to have?".And Q8 read as follows: "To which extent is it important to practise the formal conventions of academic/technical texts in class?".In Table 7, a summary of lecturers' answers to Q7 is provided as to the presence/absence of feedback, and the type of feedback given.Other lecturers reject genre teaching with the claim that their main language objective is reading popularised science rather than writing specialized texts (Lecturer 5) -although the same lecturer reports giving students guidelines on a specific genre (see quote above).Others express their views on the limitations of genre instruction in EMI, considering it just "another tool", with the same value as other language skills, like oral interaction or reading comprehension (Lecturer 8).
In all, lecturers explicitly state the case for ESP-EMI collaboration, and even recommend the incorporation of an ESP specialist to provide disciplinary language teaching in EMI classes (i.e.fully integrating content and language instruction), with the actual content lecturer's point of not reducing content.It should be noted though, that in the engineering schools where these lecturers are based, there are no ESP courses any longer (they have been replaced by EMI courses).The only option that students have is to take extracurricular courses offered by the language centre affiliated with the university.

DISCUSSION
Regarding the first research question ("To what extent are university policies promoting EMI backed by actual support to lecturers?"),our findings reveal, first, that Catalan universities have enacted institutional policies, although there are hardly any specific guidelines on their implementation.What those policy documents acknowledge, however, is the need to support lecturers, considering the lack of tradition of English-taught programmes and especially participants' limited English proficiency.As universities prioritise language improvement and Vol.5(2)(2017): 266-290 address very general methodological EMI/CLIL issues with training courses, more specific disciplinary practices such as student disciplinary socialisation and enculturation may be overlooked (Airey, Thus, the risk is that disciplinary knowledge and induction into disciplinary practices and discourse may not receive sufficient attention (Hyland, 2000).
University policies specify three general lines of action, namely lecturer support and training (the main action undertaken), recognition, and quality assurance (the latter cutting across the other two).These broad lines are then translated into specific actions.Such detailed policies reveal an institutional concern with the promotion of EMI and the need to provide training and support.Yet, further studies need to corroborate the extent to which actions and strategies in policy documents are actually implemented.
Despite institutional commitment to the promotion of EMI, a strategic aim for most universities, its actual implementation is left to individual departments/lecturers, as most policies did not specify criteria for selection of lecturers/courses for the implementation of EMI.At this point, we should reconcile the optimistic views on the deployment of EMI that derive from the analysis of the institutional policies presented in this study with the findings from a previous study set in the same context (Arnó-Macià & Mancho-Barés, 2015), in which EMI lecturers claimed receiving no institutional support.Further studies should also monitor the quality and results of EMI programmes, so as to facilitate a principled approach to implementing EMI.
The effects of EMI training programmes fall beyond the scope of this research, but this study can pave the way for examining EMI lecturers' perspectives on academic/disciplinary literacies.Therefore, the second research question probed into the genre awareness and teaching practices of STEM lecturers.The selection of assessed written genres assigned by the EMI lecturers under analysis are Exercises, Critique, Methodology Recount, Proposal, and Empathy Writing.Some of these genres rely more on language as a meaning-making system (English, in this case) than others (Empathy Writing vs. Exercises).This identification can prepare the ground for a needs analysis on genre conventions, which eventually may feed EMI teacher training programmes.
In terms of the instruction/feedback on genres provided by EMI lecturers, we found that linguistic feedback/instruction is not an objective for content lecturers due to syllabus coverage constraints.According to lecturers, it is the task of language specialists to provide such feedback/instruction.This general unwillingness to monitor students' language production resonates with previous studies (Airey, 2012).These findings evidence not only that the main focus of EMI lecturers is on disciplinary content learning, but also the widespread lack of awareness EMI lecturers have of other formal aspects of written texts apart from terminology, overlooking the need to raise students' awareness of producing texts that meet the expectations of the disciplinary community (Bhatia, 2014).Such awareness-raising could inform the syllabus of EMI teacher training programmes, Vol.5(2)(2017): 266-290 which, in addition to dealing with general academic communication, could also delve into discipline-specific genres.Another EMI-ESP collaboration could involve the incorporation of specialists/courses in EMI programmes (as hinted by some of the lecturers in this study).Thus, a question that arises in this particular context of EMI replacing ESP is whether ESP should be reinstated, either with independent ESP courses or through tandem-taught EMI courses.Further discussion of stakeholders' views on ESP as purely language courses vs. EMI courses as discipline-specific communication is found in Arnó-Macià and Mancho-Barés (2015), which should lead ESP course designers to reappraise the approaches to ESP in this new scenario.An example could be making ESP courses subject-specific, but given current trends towards prioritising EMI, it seems clear that it is necessary to find 'survival strategies' for ESP (Williams, 2016).

CONCLUSIONS AND QUESTIONS FOR FURTHER RESEARCH
This paper has probed into institutional EMI policies and lecturers' reported teaching practices, to take EMI and ESP collaboration a step forward towards improving students' discipline-specific academic literacy.Although this study is restricted to the analysis of policy documents, without looking at their implementation, such a broad overview of institutional plans across Catalan universities offers valuable findings on EMI policies.EMI is promoted as part of internationalisation processes, and more specifically the development of students' English language proficiency.However, such general top-down policies are hardly connected with actual practices, implemented at a bottom-up level, with actions left in the hands of individual lecturers (Dafouz-Milne, Camacho-Miñano, & Urquía-Grande, 2014).This situation may lead to a lack of coherence between top-down policies and bottom-up implementation.A follow-up study should now examine the implementation of such programmes, especially the training offered to lecturers.Gathering EMI lecturers' reactions would help reappraise training courses and pinpoint the most relevant areas.
The analysis of EMI training reveals that genre pedagogy is absent (let alone, critical approaches, e.g.Benesch, 2009), although most of the lecturers surveyed assign a variety of disciplinary writing tasks.However, they are reluctant to address linguistic/communication issues in their content classes mainly because the syllabus would need to be reduced, and because language instruction is the task of language specialists.Thus, EMI lecturers do not approach genre instruction, revealing instruction models based on 'a pedagogy of osmosis' (Turner, 2011), whereby academic literacy is expected to result from long-term exposition rather than focussed instruction.This situation can be overcome by raising EMI lecturers' genre awareness, especially by making such conventions visible and teachable to students with the help of ESP lecturers.
Vol. 5(2)(2017):  Due to their tradition in genre studies, and the knowledge gained by teaching disciplinary communication skills, ESP instructors can collaborate with EMI lecturers in the design of genre-based EMI courses as well as training programmes that make disciplinary literacy visible.Thus, EMI students will benefit not only from EMI instructors' comments on the contents of the texts produced, but also on how to use language to produce the desired communicative/social effects in disciplinary communication.This study is a first step in this direction, and data from EMI training courses and lecturers' reported practices will have to be complemented with evaluations of the effects of training on EMI lecturers and closer examination of the genres used in EMI teaching.

Table 1 .
Classification of written academic genres (adapted from

Table 2 .
Lecturers' EMI experience in Bachelor and Master programmes

Table 6 .
Table 6 summarises the findings, listing the main genres used.Classification of written genres assigned by lecturers

Table 7 .
Distribution of feedback depending on genresData from Table7bring to light that lecturers' feedback mostly focuses on content, while some lecturers also provide feedback on terminology and genre conventions.Special mention can be made of Lecturer 5, who gives specific guidelines on genre, register, and text structure: