Abstract
My first academic interest in EMI was brought about by my doctoral student Yuen Yi Lo (Lo and Macaro, 2012, 2015), now teaching at the University of Hong Kong. So, this would have been about 12 years ago. She wanted to do a study of secondary education in Hong Kong, which was one of the early crucibles of EMI. But the interest had probably been there in the background ever since childhood because at school, at least at primary school, English was not my first language so I was learning content through a second language. Later when I began teaching in secondary schools, particularly in London, there were a lot of students with English as an Additional Language (EAL) learning their content subjects through English. So, whilst I consider EAL to be somewhat different from EMI, there are nonetheless parallels. But of course in the past two decades there has been a boom in EMI worldwide and with Content and Language Integrated Learning (CLIL), particularly in Europe, and I wanted to explore if at some point it was going to take over from Modern Foreign Languages (MFL) and English as a Foreign Language (EFL). So, I felt that it was a phenomenon that could not be ignored.
That’s a very interesting question. If we consider EMI in higher education in isolation then in some ways it hasn’t evolved very much at all. Currently the research focus is still dominated by stakeholder beliefs – that of teachers and students – and how their beliefs relate to policy either at the institutional level or the national level. Moreover, the research is typically at the single case study level – that of the single institution which makes comparisons difficult, and without comparisons it is not easy to draw conclusions about which is the best form of implementation and support. Your own study (Sahan, 2020) attempted to make comparisons within the same country of different institutions and that told us something new about policy and implementation. There are some other themes emerging but they are taking a long time to establish themselves as major research themes.
Attempts to look at what is actually going on inside the classroom is one of these: the focus on interaction is very welcome indeed. We are beginning to understand how interaction is being affected by the medium of instruction and particularly the role that proficiency of students and teachers have on shaping the interaction – but also on whether teacher expertise is there to promote quality interaction. It’s interesting that quality interaction has featured prominently in the second language acquisition (SLA) literature and also in the literature on subject learning especially in science and maths, so it is hardly surprising that it should feature in EMI settings. But a lot more work needs to be done on what goes on inside the EMI classroom and therefore what might need to change. Related to this is the theme of ‘how much English’ or ‘English-only’. There is a fair amount of research on this but one problem that I see is that there is no communication between sociolinguists and those researchers more interested in pedagogy – so, if you like, the link between the impact of EMI on the home culture and the home language with impact on successful academic outcomes. Both aspects are important but the two groups aren’t really engaging with each other – at least not sufficiently in my view.
Another theme is the impact on content learning and on English proficiency – what in previous writing I have alluded to as the components of a cost-benefit analysis of EMI implementation and expansion (Macaro, 2018; Macaro et al., 2018). One would expect that the benefit would be increased English proficiency over and above what you might acquire from EFL provision or even from EAP/ESP (English for Academic Purposes/English for Specific Purposes) provision. The corollary expectation would be that there is, at the very least, no long-term detrimental effect on content learning. There is the issue, of course, of what ‘long-term’ might be and I suppose that would have to be decided at the local level. We need some large cross-sectional but also prospective/longitudinal studies.
Well, coming back to the themes that I was talking about earlier, there is the difficulty of carrying out empirical research without first establishing a sound theoretical framework: getting to grips with the nuts and bolts of the problem. So, for example, you will find studies which have adopted terms such as L1 use, codeswitching or translanguaging but they don’t provide you with a theoretical framework which describes why that adopted terminology is the most appropriate. Similarly, you can find papers which describe EMI as CLIL but without even a minimum attempt at theorizing how the particular setting they are looking at is CLIL – in what way content and learning is being integrated.
Then there are the difficulties experienced by other fields of applied linguistics research but which are accentuated in EMI: access to classrooms where teachers might feel somewhat under threat – possibly because of their lack of confidence with their English proficiency. Also, obtaining large quantities of data for statistical analysis and, of course, the comparison between institutions and even more so between countries.
But possibly the most difficult aspect is one which I am particularly concerned about at the moment and just starting on some research. The research field is currently dominated by applied linguists. The researchers are applied linguists, the journals they publish in are essentially applied linguistics journals (even, for example,
Before teachers and policy-makers (and I stress that this should be both, as research has already demonstrated the shortcomings of a top-down approach), before these stakeholders introduce EMI, they have to ask themselves a number of questions:
Why are we introducing EMI or why are we expanding EMI provision?
Then: what sort of model of EMI is going to be the most suitable, advantageous and ultimately successful one for our context?
What is the current level of expertise that we have in our particular institution or setting? So, level of teacher proficiency; expertise of support groups, such as EAP/ESP teachers, language centres?
How much financial support and other resources can we set aside for this venture? Because the research so far seems to suggest that it needs to be really well resourced.
All these questions can be informed by institutional, national and international research. I don’t think we are at a stage where teachers and policy-makers could turn to the current research base with the confidence that they will get the answers. So if you ask me what are the key implications, then probably the most important ones to come out of the research are don’t rush into it without proper planning and make sure you are prepared to invest resources into it. I would also add that there is research evidence that EMI consolidates socio-cultural elites so the above planning questions would have to, I hope, mitigate that.
One of the key differences is the status of the home language or home languages. It is interesting that you should mention Africa because many of its countries have a number of home languages, so adopting English as the medium of instruction, particularly in Higher Education (HE), is a way of adopting a lingua franca for the various communities or ethnic groups, even though for some countries this was the language of the colonizer. It’s of note that there hasn’t been more focus on EMI in HE in Africa, which I think is something we should not be very proud of, because once again it probably comes down to money. In Latin America, I don’t know if the language is the factor – the fact that Spanish is such a widely spoken international language may be the reason why EMI expansion has been slower – that’s the impression I get from talking to colleagues in Mexico and Chile. But there seems to be less research into EMI in Latin America as our systematic review (Macaro et al., 2018) showed, and I think any update will show that the increase is not the same as in other areas.
In Europe, we are going to have to see how things develop. Clearly English CLIL and English EMI have dominated now for a couple of decades – but with Brexit it is unclear what will happen. My guess is that it will change a bit but not fundamentally. The general model in HE in Europe is to give students who embark on an EMI programme additional concurrent support (see Macaro, 2018, for models of language support) but I haven’t come across any research which takes a Europe-wide perspective on how they are being supported. If I take Italy as an example that I know a bit about (Macaro et al., 2019), then the support is very much dependent on the individual institution. Nor is there a comprehensive assessment of what differences there are in Europe with regard to access to EMI courses – whether there are proficiency exams as gateways to enrolment. I think the whole issue of enrolment methods could be looked at internationally and compared.
In China, EMI is often described as bilingual education but this seems to take many forms: from a mixture of languages in the same classroom to different modules being taught in L1 or L2. Again, it would be difficult to say exactly why this model is being proposed – even if actually it isn’t always adhered to in practice.
In terms of the different forms of implementation, the different models do seem to be geographical area specific. So the preparatory year programme (PYP) model is adopted in the Middle East and Turkey. In the latter, as you know better than me, for historical reasons; in the former my understanding is that they can afford the extra year – in terms of the resources that it entails (both students’ resources and that of the institution). But I can’t say I have come across a paper which states clearly what the reasons for adopting the PYP model in the Middle East are.
In Hong Kong, there is the long historical process and the presence of English in education that has been well documented by people like Stephen Evans (Evans, 2008, 2011; Evans and Morrison, 2011), and what you have is a continuation from secondary education EMI to EMI in HE – with no apparent need for a PYP.
But to be honest these are impressions rather than research-based comments and, again, nobody has really done a global examination of how different regions/countries or jurisdictions have implemented EMI and, most importantly, why.
Yes, I think the parallels as I mentioned earlier are quite apparent. So in SLA we have a long tradition of looking at input, interaction, student output, feedback, codeswitching. In content subjects, particularly in science, taught in the students’ L1, there has also been a strong emphasis on interaction with the general message being that the less unidirectional the teaching is, the better, and therefore they’ve looked at question-types, higher-order and lower-order, the nature of students’ responses, the Initiation-Response-Feedback sequence (and all its other manifestations) and wait-time to give students the time to formulate a quality response. Now it seems to me all these aspects of both disciplines matter in EMI and probably even more so and that you have the added complication of vocabulary knowledge at the various levels of frequency. So in an L1 medium of instruction setting, you would expect the students (EAL learners apart) to be able to deal comfortably with the everyday language, high frequency words and culturally well-known idioms. You can’t make that assumption in an EMI class. So the whole notion of what we mean by ‘proficiency’ has to once again be thought about and theorized. I recall one very interesting study (Fortanet-Gómez, 2012), which reported that EMI teachers had no problem (in terms of their own proficiency) delivering a presentation at a conference to peer-experts but much more of a problem teaching their students who were at a lower level both in terms of proficiency and confidence with the discipline’s concepts and constructs.
So there is a constant interplay between language and the concepts that are being talked about. Just think about trying to put across the concept of ‘interlanguage’ to a group of non-L1 students in SLA who already have the advantage of a strong background in language learning, then double that when you have students without a language background and even more complex concepts in science or engineering. I was struck by a further aspect the other day in a study by one of your colleagues at Oxford looking at EMI chemistry in Japan (Aizawa, forthcoming) – one of the few studies which have, to date, measured content learning. The chemistry pre- and post-tests were devised by the chemistry teachers – great. But most of the examples – at first sight – did not include a lot of language because the students in the test had to draw diagrams and construct formulae. So you might say: so where is the language problem? Well it seems to me the language problem lies in how they got to being able to draw the diagrams and constructing the formulae. Clearly a lot of receptive language had to be dealt with in order to lay the groundwork. But also productive language was underneath the surface of the action of drawing the diagram. And a further layer was whether they, at various stages in the learning process, were translating explanations of the concepts into Japanese (L1), taking notes in L1 or English and so on and so forth. And as I tried to describe in the article that you referred to: you have the label given to a concept; you then have the elegant and succinct definitions of the concept; and then the language needed to truly understand the concept and provide productive evidence that you have understood it.
I’m not sure I can predict how it
Unless we bring together content specialists and language specialists, we will find it difficult to understand the challenges faced by students and indeed by teachers (both content teachers and language support teachers). So I would like to see future research being conducted collaboratively. But I’m not holding my breath! The difficulties of collaboration, in this way when disciplines have been operating so much in silos, are not going to be surmounted easily. Collaborative research of this kind should also lead to collaborative professional development which in turn can be researched for its effectiveness – and by the way that requires long-term professional development not an intensive couple of days. Purely coincidentally, this popped up in my inbox during our interview:
So perhaps there is a chance that it will move in this direction!
We need to ascertain through research the long-term impact of EMI. The possible detrimental impacts: poor content learning; consolidation of socio-economic elites; the gradual phasing out of content material published in the L1; the psychological impact on students who are struggling. But we should also ascertain the possible long-term positive impacts: better career prospects especially for students from less wealthy countries; seeing the positive benefits of international student mobility; more diverse academic elites; and international cultural understanding.
Footnotes
Funding
ORCID iD
