Teaching academic content in a foreign language without explicit language support can be detrimental to the learning of both the subject and the foreign tongue, researchers have said.
Practices such as English as a medium of instruction (EMI) could be ineffective compared to learning language and content separately unless students have already mastered their foreign language or receive targeted support, a new study warns.
‘It’s safe to say that EMI doesn’t work, from the data we got,’ Dr John Sweller, one of the authors of the study, told the Gazette. ‘If you try to teach adults a second language simultaneously with content, they’ll end up with very little learning in either.’
The authors explain that learning a foreign language as an adult means acquiring ‘biologically secondary knowledge’, which requires explicit instruction and ‘conscious effort’. Cognitive load theory, the study says, suggests that learning two things at the same time won’t work – and the data seems to confirm this.
Over three experiments, each with control conditions, researchers found that reading a text only in the foreign language led to the worst outcomes for both content and language learning. In all three experiments, which involved a total of 294 French university students, participants were divided into three groups. A text about a topic relevant to the students’ course was given to the first group in the foreign language, to the second group in French, and to the third in both the foreign language and French. The foreign language was German for the three groups in the first experiment and English for the three groups of the other two experiments. The students, who averaged a CEFR B1 in the foreign language, were then asked to study the texts and to perform three tests afterwards: two language and translation tests and one content test with questions in French.
Unsurprisingly, the French-only conditions yielded the best results for the content test. However, results were higher in the language and translation tests for students who read the text in both languages. The full ‘immersion’ conditions obtained the worst results in all three experiments.
Language proficiency is paramount for EMI to lead to successful learning, the authors said. Dr Sweller explained, ‘Ideally, you should not learn content in the L2 until your language levels are so high that you don’t have to consciously think about the language. Until then, according to everything we know about human cognition, you are going to struggle – both in the language and in the content.’
The idea is that we have a limited pool of resources for learning, and the brain struggles to learn two things at the same time for this reason. Co-author Dr Andre Tricot added, ‘All the resources you use to deal with the language are unavailable to learn the content.’
Universities should put in place extra support to ensure students don’t struggle with lessons delivered in English, the authors say. Co-author Dr Danielle Joulia said students would benefit from targeted language support that covers the vocabulary of the topic they are studying, ideally before each class. ‘General academic English support is not enough,’ she insisted.
Just like Clil at primary and secondary level, EMI is gaining ground in higher education around the world (see our EMI special). Although Clil has more focus on pedagogical support in the foreign language, both methodologies aim to teach content and language simultaneously. This approach doesn’t seem to tally with the findings of this study, and with cognitive load theory in general. ‘The idea that we have a limited pool of resources [see main piece] is common knowledge,’ said Dr Sweller. ‘It’s incomprehensible that there is such a divide between research and practice.’ But the divide is not only between research and practice, but also between research and policy, the authors said. ‘There is a lot of politics around internationalisation.
Much of these procedures [such as EMI] have been brought in without any prior research whatsoever,’ Dr Sweller said. And research is still lacking, with very few randomised control trials, he added.
For Dr Stephanie Roussel, the lead author of this study, there should also be a cultural perspective on the issue. ‘EMI and Clil seem to work well in countries such as Canada and Belgium, which are bilingual,’ she explained. ‘Maybe implementing such an approach in a different cultural context, such as broadly monolingual France, won’t yield the same results without strong second language support.’