This thesis is a study of language use in English language classrooms in primary schools in Hungary. The focus of the study is on the use of the target language (English) and the mother tongue (Hungarian) by the teachers and the learners. The teachers are all Hungarian native speakers, with varying levels of competence and previous experience in communicative language teaching, and this presents a challenge to the adoption of a communicative approach to the teaching of English. The National Core Curriculum endorses the communicative approach, with the expectation that the target language will be used as much as possible. However, in practice, the mother tongue is widely used in these classrooms, both by the teachers and by the students. There is therefore a conflict between policy and practice: the policy is that the target language should be used wherever possible, whereas the practice is that the use of the target language is limited to predictable and routine contexts. It is this conflict which constitutes the central question which is addressed in this thesis: how do teachers resolve the conflict between what they are expected to do, and what they feel capable of doing. Data from classrooms and interviews were collected and analysed, using both quantitative and qualitative techniques. The focus of the analysis was on the amount and function of the use of the mother tongue by the teachers. Comparisons were drawn between teachers of Grade 4 pupils who started to learn English in Grade 1 and those who started in Grade 4. This analysis is complemented by evidence concerning the teachers‘ beliefs and understandings about the pressures and constraints which affect their teaching of English to young learners. The results suggest that the possibility of communicative language teaching in these classrooms is constrained by various factors, including the limitations in the children‘s cognitive capabilities and the proficiency level of the children, and the teachers‘ preference for using their previous methods which included grammar, translation and memorisation; also by curriculum requirements such as the use of the textbook, and the necessity to prepare the children for examinations. The implications of these findings for curriculum development in foreign language teaching in other comparable contexts are discussed.

Previous studies have shown that teachers and students using English as the medium of instruction (EMI) in science classrooms encounter many language challenges with teaching and learning processes. Problems include the limited English communication skills of science teachers, the lack of EMI training for science teachers, the students' different language abilities and science teachers' beliefs that they are not responsible for addressing students' language needs in science. Teachers' lack of language awareness has led to poor teaching practices and limited interactions in the classrooms. This lack of language awareness, in turn, suggests that there are limited opportunities for students to learn English as a second language in the science classroom. This study extends the research on EMI classroom interactions in Hong Kong (Lo and Macaro, 2012) to the previously unexamined context of senior secondary science classrooms. A total of 19 teachers and 545 students from grades 10 and 11 EMI science class were recruited in Hong Kong from 'early-full EMI' schools (full EMI instruction from grades 7 to 12) and 'late-partial EMI' schools (Chinese medium from grades 7 to 9 and partial EMI instruction from grades 10 to 12). The project used multiple sources of qualitative data (i.e. semi-structured interviews and 33 videotaped classroom observations) to explore the similarities and differences in classroom interactions during the first and second years of the senior science curriculum (grades 10 and 11) in the two types of EMI schools. This project also investigated these science teachers' and students' perceptions of EMI teaching and learning processes, their preference of instructional language and their beliefs about teaching and learning in the EMI environment. Interviews also probed teachers' language awareness, teachers' and students' belief about EMI, students' self-concepts in science (students' perceptions or beliefs about their ability to do well in science, see Wilkins, 2004)) and their perceptions of language challenges and coping strategies in EMI classrooms. The results from the observational data show similar interactional patterns in both early-full and late-partial EMI science classrooms when measured as percentages of interaction time, distribution of time between teacher and student talk and frequency of pedagogical functions. However, the nature of the interactions is different. In late-partial EMI schools, overall, there are more (but shorter) student initiations and responses, more use of higher-order questions from the teachers but less direct feedback to students. Both teachers and students tend to use their L1 more. In both types of schools, there was less interaction time and a lower maximum length of student turns and more L1 use in grade 11 than in grade 10. The discourse analysis of the four biology lesson transcripts also shows that both early-full and late-partial EMI students predominantly produced incomplete sentences consisting of short, technical nouns or noun phrases referring to scientific items. Science teachers rarely made any attempts to correct their students' language mistakes, nor did they encourage students to produce a complete sentence. This lack of teacher feedback on students' L2 language production perhaps reflects the fact that EMI science teachers rarely provide comprehensible input to facilitate students' L2 language learning. These findings suggest the important role of the teacher's modified input in teacher-student interaction in developing students' content knowledge and language skills. The adoption of EMI appears to lead to the development of students' comprehension of content knowledge more than development of their language production skills. As a result of their language shortfalls, the students' L2 productive skills remain under-developed despite English instruction. This lack of language support by teachers appears to indicate a gap between the aims of the EMI policy and its implementation. The interview and questionnaire data show that the science teachers from both the early-full and late-partial EMI schools held many of the same views about their EMI teaching experiences, but they differed in their attitudes towards the value of English language skills and their language awareness. The early-full EMI teachers believed English language skills were important and these early-full teachers have a higher language awareness than the late-partial EMI science teachers. Students from both types of schools also held similar views about their EMI learning, indicating that they welcome the adoption of EMI instruction. However, while the late-partial EMI students see EMI as an opportunity to improve their English, those in the early-full EMI schools believe that EMI discouraged them from learning. By providing an evidence-based, pedagogically focused analysis of teacher and student classroom interactions and their perceptions, this research sheds light on ways to improve the quality of instructional practices in different EMI classrooms in Hong Kong and in similar contexts around the world.