Conceptualising Integration in CLIL and Multilingual Education 9781783096145

Citation preview

Conceptualising Integration in CLIL and Multilingual Education

BILINGUAL EDUCATION & BILINGUALISM Series Editors: Nancy H. Hornberger (University of Pennsylvania, USA) and Wayne E. Wright (Purdue University, USA) Bilingual Education and Bilingualism is an international, multidisciplinary series publishing research on the philosophy, politics, policy, provision and practice of language planning, global English, indigenous and minority language education, multilingualism, multiculturalism, biliteracy, bilingualism and bilingual education. The series aims to mirror current debates and discussions. Full details of all the books in this series and of all our other publications can be found on http://www.multilingual-matters.com, or by writing to Multilingual Matters, St Nicholas House, 31-34 High Street, Bristol BS1 2AW, UK.

BILINGUAL EDUCATION & BILINGUALISM: 101

Conceptualising Integration in CLIL and Multilingual Education

Edited by Tarja Nikula, Emma Dafouz, Pat Moore and Ute Smit

MULTILINGUAL MATTERS Bristol • Buffalo • Toronto

Library of Congress Cataloging in Publication Data A catalog record for this book is available from the Library of Congress. Names: Nikula, Tarja, editor. | Dafouz, Emma, editor. | Moore, Pat (Multilingual educator) editor. | Smit, Ute, editor. Title: Conceptualising Integration in CLIL and Multilingual Education/ Edited by Tarja Nikula, Emma Dafouz, Pat Moore and Ute Smit. Description: Bristol; Buffalo: Multilingual Matters, [2016] | Series: Bilingual Education and Bilingualism: 101 | Includes bibliographical references and index. Identifiers: LCCN 2016022309| ISBN 9781783096138 (hbk : alk. paper) | ISBN 9781783096152 (epub) | ISBN 9781783096169 (kindle) Subjects: LCSH: Language arts (Higher)--Correlation with content subjects. | Multilingual education. | Language and languages--Study and teaching (Higher) | Interdisciplinary approach in education. | Education, Bilingual. Classification: LCC P53.293 .C66 2016 | DDC 418.0071--dc23 LC record available athttps://lccn.loc.gov/2016022309 British Library Cataloguing in Publication Data A catalogue entry for this book is available from the British Library. ISBN-13: 978-1-78309-613-8 (hbk) Multilingual Matters UK: St Nicholas House, 31-34 High Street, Bristol BS1 2AW, UK. USA: UTP, 2250 Military Road, Tonawanda, NY 14150, USA. Canada: UTP, 5201 Dufferin Street, North York, Ontario M3H 5T8, Canada. Website: www.multilingual-matters.com Twitter: Multi_Ling_Mat Facebook: https://www.facebook.com/multilingualmatters Blog: www.channelviewpublications.wordpress.com Copyright © 2016 Tarja Nikula, Emma Dafouz, Pat Moore, Ute Smit and authors of individual chapters. All rights reserved. No part of this work may be reproduced in any form or by any means without permission in writing from the publisher. The policy of Multilingual Matters/Channel View Publications is to use papers that are natural, renewable and recyclable products, made from wood grown in sustainable forests. In the manufacturing process of our books, and to further support our policy, preference is given to printers that have FSC and PEFC Chain of Custody certification. The FSC and/or PEFC logos will appear on those books where full certification has been granted to the printer concerned. Typeset by Deanta Global Publishing Services Limited. Printed and bound in Great Britain by the CPI Books Group Ltd.

Contents

Contributors

vii

Acknowledgements

xi

Foreword: Integrating Content and Language in Education: Best of Both Worlds? Rick de Graaff More Than Content and Language: The Complexity of Integration in CLIL and Bilingual Education Tarja Nikula, Christiane Dalton-Puffer, Ana Llinares and Francisco Lorenzo

xiii

1

Part 1: Curriculum and Pedagogy Planning 1

Cognitive Discourse Functions: Specifying an Integrative Interdisciplinary Construct Christiane Dalton-Puffer

2

Historical Literacy in CLIL: Telling the Past in a Second Language Francisco Lorenzo and Christiane Dalton-Puffer

3

Learning Mathematics Bilingually: An Integrated Language and Mathematics Model (ILMM) of Word Problem-Solving Processes in English as a Foreign Language Angela Berger

v

29

55

73

vi

4

Contents

A Bakhtinian Perspective on Language and Content Integration: Encountering the Alien Word in Second Language Mathematics Classrooms Richard Barwell

101

Part 2: Participants 5

6

University Teachers’ Beliefs of Language and Content Integration in English-Medium Education in Multilingual University Settings Emma Dafouz, Julia Hüttner and Ute Smit CLIL Teachers’ Beliefs about Integration and about Their Professional Roles: Perspectives from a European Context Kristiina Skinnari and Eveliina Bovellan

123

145

Part 3: Practices 7

Integration of Language and Content Through Languaging in CLIL Classroom Interaction: A Conversation Analysis Perspective 171 Tom Morton and Teppo Jakonen

8

Teacher and Student Evaluative Language in CLIL Across Contexts: Integrating SFL and Pragmatic Approaches Ana Llinares and Tarja Nikula

9

Translanguaging in CLIL Classrooms Pat Moore and Tarja Nikula

189

211

Conclusion: Language Competence, Learning and Pedagogy in CLIL – Deepening and Broadening Integration Constant Leung and Tom Morton

235

References

249

Index

275

Contributors

Richard Barwell is professor of mathematics education at the Faculty of Education, University of Ottawa, Canada. In much of his research, he draws on sociolinguistic theories and techniques to examine the role of language in mathematics classrooms, particularly in contexts of language diversity. His work has been published in international journals in applied linguistics, mathematics education and general education. Angela Berger studied English and mathematics at the University of Vienna and Nottingham University, England. After having taught at different secondary schools in Vienna and London, she was an assistant at the mathematics department of the University of Vienna, obtaining a PhD in mathematics. Her dissertation focused on mathematical problem solving in English as a foreign language. She currently works as a secondary school teacher, university lecturer of didactics and teacher educator at the University College of Teacher Education in Vienna. Eveliina Bovellan works as a language teacher and a coordinator of language teaching in the city of Kuopio. In 2014, she completed her doctoral dissertation on primary teachers’ beliefs about learning and language as reflected in their views of teaching materials for CLIL. She has attended and presented in domestic and international scientific conferences and published on CLIL teacher beliefs. While working as a language teacher, Bovellan is exploring secondary CLIL teachers’ beliefs about integration and professional identities in cooperation with Kristiina Skinnari. Emma Dafouz is associate professor at the University Complutense of Madrid, where she teaches courses in applied linguistics, discourse analysis and bilingual education. Since 2001, she has focused on English-medium instruction (EMI) in higher education and currently coordinates an interdisciplinary research group, funded by the Spanish Ministry of Economy and Competitiveness, on the Internationalisation of Higher Education and EMI. The results of such research have given rise to numerous conferences, courses and publications in national and international journals (International

vii

viii

Contributors

Journal of Bilingualism and Bilingual Education, Language and Education, Modern Language Journal and Applied Linguistics). At present, she is an advisor for the Complutense Language Plan for Internationalisation. Christiane Dalton-Puffer is professor of English at the University of Vienna (English Department & Centre for Teacher Education). She has been doing research on CLIL classroom discourse and CLIL classrooms as language learning environments. Her current focus is on how cognitive discourse functions are employed by participants to construct subject knowledge. She is the author of Discourse in Content and Language Integrated Learning (2007), co-editor of several volumes and co-founder of the AILA Research Network on CLIL and Immersion Education. Rick de Graaff is professor of bilingual education and foreign language pedagogy at Utrecht University (Utrecht Institute of Linguistics and Graduate School of Teaching) in the Netherlands. His main fields of interest include processes and outcomes of content and language integrated learning, task development for communicative competence, the role of instruction in L2 acquisition and professional development of language teachers. He has published in the International Journal of Applied Linguistics, the International Journal of Bilingual Education and Bilingualism, Language Learning Journal, Studies in Second Language Acquisition, The Handbook of Language Teaching (WileyBlackwell, 2009), Technology-Mediated TBLT (Benjamins, 2014) and Building Bilingual Education Systems (CUP, 2015). Since 2011, De Graaff has been coordinator of the CLIL Research Network of AILA (International Association of Applied Linguistics). Julia Hüttner is a lecturer in applied linguistics at the Department of Modern Languages, University of Southampton. Her research focuses on the teaching and learning of English in an instructed context, with particular emphasis on disciplinary language, genre and CLIL. Her publications include the research monograph Academic Writing in a Foreign Language (Peter Lang, 2007), edited volumes and numerous journal articles (e.g. in Classroom Discourse, International Journal of Applied Linguistics, International Journal of Bilingualism and Bilingual Education and System). Teppo Jakonen works as a post doctoral researcher at the University of Jyväskylä, Finland, where he obtained his PhD in applied linguistics in 2014. His research interests include knowledge and learning in social interaction, multimodality, as well as issues related to content and language integrated (CLIL) teaching and other forms of bilingual education. Constant Leung is professor of educational linguistics in the Department of Education and Professional Studies at King’s College London. His research

Contributors

ix

interests include additional/second language development, language assessment and teacher professional knowledge. He has written and published widely on issues related to ethnic minority education, additional/second language curriculum and language assessment. His recent publications include English: A Changing Medium for Education (Multilingual Matters) and The Routledge Companion of English Studies (Routledge, co-editor: Brian Street). He is associate editor for Language Assessment Quarterly and editor of research issues for TESOL Quarterly. He is a fellow of the Academy of Social Sciences (UK). Ana Llinares is a senior lecturer in the English department at the Madrid Universidad Autónoma. Her research interests include applications of systemic functional linguistics to foreign language learning, pragmatics of foreign language learning and classroom discourse analysis, mainly applied to CLIL primary and secondary school contexts. She coordinates the UAMCLIL research group and has published widely on the topic, including coauthorship of the book The Roles of Language in CLIL, published by Cambridge University Press. She is currently co-editing the volume Applied Linguistics Perspectives on CLIL, which will be published by John Benjamins. Francisco Lorenzo is an associate professor at Universidad Pablo Olavide (Sevilla) and has held visiting research positions at Harvard University, London University and the Centre for Applied Language Studies (University of Jyväskylä, Finland). He has researched bilingual education, sociolinguistics and European language policies. His papers are published in journals such as Applied Linguistics, Language and Education, the International Journal of Bilingualism, System and the Journal of European Language Policies. He has authored or co-authored several Spanish books: Motivation and Second Languages (2007), Bilingual Education (2012) and Communicative Competence in L2 Spanish (2009). Pat Moore is a lecturer at the Universidad Pablo de Olavide in Sevilla, Spain, where she coordinates modules on bilingual education in two masters degrees. Her research interests revolve around second/foreign language learning in secondary and tertiary contexts focusing particularly on classroom discourse, learner output, materials design and teacher development. She has published related articles in Applied Linguistics, the International Journal of Bilingualism and Bilingual Education, Language Learning Journal, the European Journal of Language Policy and the International CLIL Research Journal. Tom Morton is senior lecturer in the Department of Applied Linguistics and Communication at Birkbeck, University of London. His research interests include classroom discourse and interaction in CLIL contexts, systemic functional approaches to the development of CLIL students’ academic

x

Contributors

literacy and communicative competence, and CLIL teacher education and professional development. He is co-author (with Ana Llinares and Rachel Whittaker) of The Roles of Language in CLIL (CUP). Tarja Nikula is professor in the Centre for Applied Language Studies (CALS) at the University of Jyväskylä, Finland. Her research interests include CLIL, pragmatics of foreign language learning and use, language education policies and the conceptual challenges that CLIL poses for central notions within applied linguistics. Her publications have appeared in numerous international edited volumes and journals (e.g. Applied Linguistics, Linguistics and Education, Multilingua, World Englishes). Recently, she has edited two special issues on CLIL with Christiane Dalton-Puffer for The Language Learning Journal. Kristiina Skinnari works as a postdoctoral researcher in the Centre for Applied Language Studies, University of Jyväskylä, and as a CLIL teacher and special educational needs teacher in elementary school. In 2012, she completed her doctoral dissertation on fifth and sixth graders’ language learner identities in elementary school English language learning. Since then, she has attended domestic and international scientific conferences and published articles on affordances for learner participation in CLIL teaching and on learner agency. Currently, Skinnari investigates CLIL teacher beliefs and identities in cooperation with Eveliina Bovellan. Ute Smit is associate professor at the Department of English Studies, University of Vienna. Currently, her main research interests lie at the crossroads of English-medium instruction in higher education, CLIL, English as a lingua franca, classroom discourse and language policy research. Her publications include English as a Lingua Franca in Higher Education (de Gruyter, 2010), Language Use and Language Learning in CLIL Classrooms (Benjamins, 2010) and numerous journal articles (e.g. in Applied Linguistics, International Journal of Bilingualism and Bilingual Education, Journal of Academic Writing, Language Teaching and System).

Acknowledgements

We gratefully acknowledge the support of the Academy of Finland whose research funding for the project ConCLIL – Language and Content Integration: Towards a Conceptual Framework (2011–2014, project number 250553) has made this volume possible. ConCLIL involved project partners from outside Finland having research stays in the Centre for Applied Studies (CALS) at the University of Jyväskylä. These occasions gave many opportunities for seminars, workshops and discussions and we thank CALS colleagues for the lively exchange of ideas that have helped us streamline the themes of this volume. Cooperation with the staff at Multilingual Matters has been smooth and effective. A special thank you to our editors Kim Eggleton and Sarah Williams for flexibility and for guiding us along the way, and to Tommi Grover for giving the green light to our initial idea for a book proposal. We also thank the anonymous reviewer for constructive feedback and the ConCLIL team members for the careful commenting on each other’s chapters. We are also grateful to Veronika Heuer and Ulrike Zillinger at the University of Vienna for helping us with the details of the bibliography. Apart from CLIL research, conceptual questions relating to content and language integration are also topical in research concerning other forms of education. This has been evidenced in the lively response we have received when presenting ongoing work as papers and as colloquia and symposia in national and international conferences. We thank the audiences for valuable questions and comments that helped us develop our thinking. As this volume shows, CLIL is a form of education that is practised across different geographical contexts and different educational levels. Research on CLIL would not be possible without educational institutions (schools, universities) opening their doors to researchers. We are therefore especially grateful to the CLIL schools and universities offering Englishmedium instruction and their many teachers and students in Finland, Austria, Spain, the UK and Canada who have cooperated with us over the years, welcomed us into their classrooms and lecture halls and have been

xi

xii

Acknowledgements

prepared to share their thinking on CLIL. To all of them: isot kiitokset, muchas gracias, vielen Dank, merci beaucoup, thanks so much! We hope that this volume will offer useful insights for the important and dedicated work that the schools and universities do to develop CLIL.

Foreword: Integrating Content and Language in Education: Best of Both Worlds? Rick de Graaff

In the very last section of the final chapter in Language Use and Language Learning in CLIL Classrooms, Dalton-Puffer et al. (2010b) plead for a fusion of language and content in learning, teaching and research. Integration, they argue, is more than just a combination of two elements: real fusion asks for an understanding of the characteristics and interplay of both, and of the potential aims, processes and outcomes of the fused context. This may sound as if we are creating something brand new, by integrating or fusing the two concepts or perspectives. However, it may also be the case that we are discovering something that has already been out there for a long time. We may just not have had the lens to understand (or the language to express) the full teaching, learning or research potential of an integrated perspective on content and language. Perhaps more children and students worldwide are educated in a second language (L2) than in their first language (L1). In many nations and societies, the language of schooling is not the same as the home language for a substantial group of learners. This is usually taken for granted (children will eventually learn anyway) or approached from a deficiency perspective (children need to overcome their language deficit in the educational context). Teachers, as a consequence, may feel that language development is their pupils’ own responsibility, or the responsibility of language teacher colleagues. In bilingual education, we have deliberately created challenges for content and language learning. By teaching content through an additional language, we aim to reach higher levels of L2 proficiency within the curricular programme, without lowering the aims for content learning outcomes. Research has shown that this is feasible (e.g. Admiraal et al., 2006; Verspoor et al., 2015). Many subject teachers in CLIL contexts are aware of the fact

xiii

xiv

Foreword

that they are co-responsible for the language development of their pupils (de Graaff et al., 2007). They also tend to be aware of the fact that it is through language that pupils learn their subject. This is what Coyle (2007) has called ‘language for learning’. On the other hand, it is also through content learning that pupils develop language: ‘language through learning’. Other aspects of CLIL teacher awareness are worth mentioning as well: as many CLIL teachers are L2 speakers of the ‘additional’ language they are teaching in, they also become aware of the linguistic challenges of their subject themselves, as was found in Koopman et al. (2014). Furthermore, many CLIL teachers – who may also teach their subject to mainstream classes in L1 – come to realise that language plays an important role in content teaching and learning in any language context. This is the case not only for pupils with poor language skills, but for all pupils; Llinares et al. (2012) convincingly argue that at any language level and at any cognitive level, pupils who learn content also (need to) develop their language. In this volume, the authors show that an integrated approach to teaching content and language is much more than only teaching ‘nonlanguage content […] through an additional language’ (Dalton-Puffer et al., 2010a: 3). As has been argued elsewhere (see Dalton-Puffer et al., 2014), CLIL is an umbrella term that may refer to a wide range of second or additional language contexts in education. From an integration perspective, the umbrella is even wider: an integrative approach to content and language learning is relevant for any teaching and learning, in any language (first, second or foreign/additional). Such a perspective implies that we should not only take a fusion perspective on content and additional language, but on any teaching and learning contexts: CLIL pedagogy with a real focus on integration should be an issue not only for bilingual education teachers, but for any subject teacher. Love (2009a, 2009b), in this respect, has introduced the concept of literacy pedagogical content knowledge, as an integrated part of the pedagogical knowledge that any subject teacher should develop. The present volume, therefore, may inspire and support the professional and pedagogical development of all teachers, teacher educators and curriculum developers. This is elaborated at the institutional, pedagogical and personal levels both in the introduction and in the three parts the volume consists of. At each level, the need, challenge and added value of taking an integrated and integral perspective to content and language in teaching, learning and research is addressed. From my point of view, teacher training institutes play a key role in this respect, as they prepare teachers for educational practice where pupils learn content through language and develop language through content learning. As CLIL researchers, we can help raise teacher and teacher educator awareness that language is not

Foreword

xv

only a challenge in bilingual education or for pupils from migrant or other L2 backgrounds, but for any pupil. That is, language of learning grows through learning, from preschool up to university level. This process can be supported by teachers who are aware of the integrated and integral role of language in learning. Over the past decade, the majority of CLIL research has been carried out by applied linguists, aiming to understand language use processes and language learning outcomes in bilingual contexts. However, in order to understand the role of teachers’ literacy pedagogical content knowledge for their pupils’ learning, an integrated endeavour with subject teachers, subject teacher educators and subject pedagogy researchers is needed. Only then can a shared cross-curricular sensitivity to language emerge and fully develop. Mehisto (2015: xxi), in this respect, introduces the term ‘reciprocal co-evolution’, referring to educational systems where ‘stakeholders, their understanding, actions, and the forces they are subject to and influence, […] all evolve in response to one another and in response to other external stimuli’. In terms of the present volume, these different aspects of education ‘integrate’ or ‘fuse’. It is evident, however, that this does not call for a one-size-fits-all approach: an integrated perspective on content and language is not the same in history as in physics teaching, not the same in additional language settings as in L1 or bilingual language settings, nor the same in sheltered (for weaker students) as in challenging (for stronger students) learning contexts. As is proposed in the present volume, an integrated approach must evolve at all levels: at the curriculum planning level (what), at the level of teacher and learner perceptions and beliefs (who), at the classroom pedagogy level (how), at all academic levels (when), both inside and outside classroom context (where). Such an integrated approach can only be successful if it takes the needs and challenges of the respective subjects, language contexts and participants into account. If so, the approach will also help us understand that integration in CLIL is not primarily about the subject teacher’s additional responsibility to pay attention to language, but about the inherent role of language in teaching and learning. For similar purposes, Lorenzo (2007; see also the introduction to this volume) has proposed a task-based approach to CLIL, including language-related pre-tasks, content-related main tasks and language-reflective post-tasks allowing subject teachers, subject teacher educators and subject curriculum developers to design, implement and evaluate language learning goals supporting subject learning goals. In many teacher education institutes, student teachers are educated in separate subject-specific groups. If they are taught in mixed groups, it is usually on general educational or pedagogical topics. If we take integration seriously, however, subject and language teachers must learn to collaborate in developing integrated subject- and language-related tasks. For such

xvi

Foreword

tasks, they may be aware that they need each other’s perspectives in order to develop and strengthen their own competencies and beliefs (for examples, see Dale & Tanner, 2012; de Graaff et al., 2012). Thus, future teachers can play a significant role in curriculum planning, participant perspectives and classroom practices that enhance a content and language integrated approach to learning (Hüttner et al., 2013; Kong, 2009). The integrated and integral role of language in teaching and learning is well illustrated by the statement ‘Language is a verb’: considering ‘languaging’ as a creative activity, which goes well beyond using a fixed set of language rules. This is one of the statements from the Dylan project.1 For the purpose of this volume, we may also add ‘Subject is a verb’: each subject is taught and learnt by doing. From an integrated perspective, sometimes ‘subject’ might be the main verb and ‘language’ the auxiliary, and sometimes it may be the other way around. But both are needed for successful interaction and knowledge development. As Byrnes (2005: 280) puts it, ‘the existence and stability of content separate from language’ is an illusion. We may add, vice versa, the existence and stability of language separate from content is an illusion as well. It is important to be aware of both perspectives in subject and language teacher education, as well as in subject and language pedagogy research. Following Dalton-Puffer et al. (2010b: 289), ‘an understanding of CLIL as “fusion” implies a multiperspectival view on both language and content, which, taken together, should help us understand the fusion of languageand-content. (…) We look forward to such a future for CLIL’. From my perspective, with this volume, the 2010 authors and their colleagues have kept their promise. The present book will inspire teachers, teacher educators and educational researchers alike to take a fused – or integrated – perspective.

Note (1)

The Dylan project is an integrated project funded under Framework Programme 6 of the European Union. The project embraced 20 research institutions in 12 European countries and ran for five years (2006–2011). See www.dylan-project.org.

More Than Content and Language: The Complexity of Integration in CLIL and Bilingual Education Tarja Nikula, Christiane Dalton-Puffer, Ana Llinares and Francisco Lorenzo

Introduction This volume is concerned with content and language integrated learning (CLIL). It is a form of education that has spread especially in Europe since the mid-1990s, and draws on earlier models of bilingual education such as immersion and content-based instruction (for CLIL overviews, see Coyle et al., 2010; Dalton-Puffer, 2011; Dalton-Puffer et al., 2010b). Rather than focusing on distinctions and points of convergence between different forms of bilingual education (for discussion, see Cenoz et al., 2014; Dalton-Puffer et al., 2014; Nikula & Mård-Miettinen, 2014), this volume addresses integration as a shared concern for all forms of education that have simultaneous content and language learning objectives. It aims at highlighting the complexity of integration and showing that apart from content and language matters, an array of other factors – institutional, pedagogical, personal – also need to be considered. Even though the above-mentioned duality of content and language lies at the heart of CLIL and has been acknowledged in most studies, CLIL research to date has been mainly characterised by language learning perspectives on learners’ general language skills. That is, there has been less research on how content and language integration challenges our views of language and on the importance of language in and for content learning (but see Llinares et al., 2012). Coyle’s (2007, see also Coyle et al., 2010: 41–42) well-known conceptualisation of the four Cs, i.e. content, culture, communication and cognition, as crucial considerations in CLIL has highlighted the fact that the role of language in CLIL needs to be considered from the viewpoint of language and content to be learnt, the type of language used for learning and the cultural-contextual factors 1

2

Conceptualising Integration in CLIL and Multilingual Education

involved. However, operationalising such considerations to the more concrete level of research and educational practice still remains a challenge. Accordingly, there are calls for more principled approaches to content and language integration (e.g. Cenoz et al., 2014; Dalton-Puffer et al., 2010c; Davison & Williams, 2001; Gajo, 2007; Llinares, 2015). An important step in this direction has already been taken in the volume by Llinares et al. (2012) on the roles of language in CLIL, in which they outline an approach to language that is based on systemic functional linguistics, regards form and meaning as closely intertwined and presents lexico-grammatical choices as tightly linked to the requirements of specific academic genres and registers. Also Meyer et al. (2015) highlight the importance of mapping conceptual development and language development in CLIL. The present volume continues the work on the integration of language and content in bilingual education and expands it to the levels of pedagogies, participants and practices, seeking both to contribute to the research field and to offer insights for practice. Apart from CLIL, the need for a comprehensive take on language and content integration concerns other forms of bi- and multilingual education. A more explicit orientation on the relationship between language and content in education is also increasingly important in the mainstream contexts that are becoming more and more diverse through processes of globalisation and migration (see e.g. Ahrenholz, 2010; Hélot, 2014; Little et al., 2014; Mohan et al., 2001). When discussing the content and language relationship, Davison and Williams (2001) provide an overview of different approaches to integrated language and content teaching on a continuum from curricular focus on content teaching to curricular focus on language teaching. However, they note, referring to content and language, that the different approaches still tend to be ‘imprecise about how they see these entities and how they go about integrating them effectively’ (Davison & Williams, 2001: 63). Davison and Williams (2001: 69) therefore argue for more comprehensive theory building in order to arrive at a principled basis for content and language integration and this is a driving force for this volume: we seek to complement existing CLIL research that has largely been concerned with learning outcomes or classroom practices with a research contribution that focuses on integration and how it can be conceptualised and investigated. Integration in education is broadly defined as ‘blending into a functioning or unified whole’ (Collins & O’Brien, 2011: 241). This unity is a crucial concern, yet difficult to conceptualise and explore: in CLIL, the very phrasing ‘content and language integration’ may push us towards considering language and content as separate. In fact, as will become evident throughout this volume, focusing on integration makes it necessary to explicate what is actually meant by ‘language’ and ‘content’. These concepts are easily taken for granted in ways that may disregard their inherent interdependency, which is described by Schleppegrell (2004: 1)

More Than Content and Language: The Complexity of Integration

3

as follows: ‘It is through language that school subjects are taught and through language that students’ understanding of concepts is displayed and evaluated in school contexts’. As the quote suggests, clarifying the relationship of content and language is not only useful for research and practice in CLIL and other types of content-based instruction but for any teaching and learning context, which is why we consider digging deeper into integration a worthy endeavour. A rationale for this book is thus to make a contribution both to CLIL research and to the problematics of the role of language in learning and teaching different school subjects more generally – a topical concern in many corners of the world, for example when explicating the role of language in different curricular areas, often with reference to ‘academic language’ (see e.g. Ernst-Slavit & Mason, 2011; Valdés, 2004; Zwiers, 2006). This volume is based on the international research project ConCLIL – Language and Content Integration: Towards a Conceptual Framework, funded by the Academy of Finland (2011–2014). The project, based at the University of Jyväskylä, Finland, brought together team members from Finland, Austria, Spain, the UK and Canada for the shared purpose of exploring and interrogating the notion of integration in CLIL (see http:// conclil.jyu.fi). The project made it possible to pool data from different countries. Although using data originally collected for different purposes involves making compromises in the exact similarity of background factors and variables, the availability of cross-border data has been crucial for allowing a more comprehensive handling of integration than would have been possible with narrower data sets. In addition to existing data, the project members have also been involved in collecting new interview data (see Dafouz, Hüttner & Smit and Skinnari & Bovellan, this volume). Apart from being the research focus, integration has also been an organising principle in the project. Firstly, while members, broadly speaking, share a socio-functional approach to language and language learning, they have also brought to the project different theoretical and methodological approaches ranging from socio-constructivist learning theories to systemic functional and genre approaches, as readers of this volume will find. Combining research orientations has also made it possible to foster methodological and theoretical integration, and to bring together areas that have previously rarely come together in CLIL research. Secondly, the project members represent different types of researchers: (applied) linguists, as well as general educators and subject educators who have been able to combine forces for a shared empirically based and conceptually oriented endeavour to explore integration as a core concern in CLIL. Thirdly, the different orientations have created opportunities for tackling integration as a truly multidimensional phenomenon but have at the same time helped identify core areas that keep emerging regardless of the perspective chosen. Consequently, this volume will illustrate the

4

Conceptualising Integration in CLIL and Multilingual Education

main argument resulting from the multidimensional approach of the project, namely the recognition that apart from content and language, it is necessary to extend the remit of the term ‘integration’ so that it will capture issues relating to other types of ‘integration’ that any educational institution offering CLIL or other forms of bilingual education needs to address, for example integration of different types of teacher expertise and teacher identities involved, integration as the use and merging of different languages in classroom practices, as well as the whole enterprise of integrating content and language as a meeting place for different, sometimes conflicting discourses, processes and practices.

Integration as a Broader Educational Concern The interconnectedness of language and knowledge building and, in more concrete terms, the apparent dependence of educational success on particular aspects of linguistic competence are not, of course, issues that have arisen for the first time in CLIL. In fact, ‘language and content’ has been on the agenda of sociologists, educationalists and linguists for a number of decades. In most cases, involvement in the topic has been fuelled by concerns regarding the obvious inequality among different groups of learners in meeting educational requirements, mastering the curricular content and attaining exam thresholds. An important reason identified for this inequality was that not all learners come equipped with linguistic repertoires sufficient to navigate the established practices of institutional/ formal education. In the 1970s, it was above all the parameter of social class which was employed to identify at-risk groups of learners who were seen as needing support in extending their linguistic repertoire in their first language (L1; e.g. Ammon, 1972; Bernstein, 1971, 1975; Oevermann, 1972). From the 1980s onwards, the issue of class has been overridden increasingly by the effects of mass immigration into Western industrialised (and now post-industrial) societies (e.g. Cummins, 2000; García et al., 2006). Now it is students with migrant backgrounds who seem to be getting a bad deal in the school system on the grounds that they do not have the medium of instruction as their L1. The fact that social class still operates as a factor alongside ethnic and linguistic background is frequently ignored in public discourse as well as scholarly debates (but see Block, 2014a). A range of approaches to these issues have been developed, all converging on the insight that language and content competences develop in synchrony and are, therefore, best dealt with in close combination. In the United Kingdom, the Bullock Report (1975) culminated in the recommendation that every school develop a cross-curricular language curriculum (language across the curriculum [LAC]). This initiative was, however, largely superseded by the move for standardisation from the 1980s onwards, partly because LAC was caught in a political crossfire,

More Than Content and Language: The Complexity of Integration

5

attacked by the left for highlighting explicit knowledge about language and by the right for being too political with its attention to language and class (Stubbs, 2000). In North America, content-based instruction developed out of an interest to support the academic language development not only of English as a Second Language (ESL) learners but also of those learning foreign languages (e.g. Stoller, 2004; Stryker & Leaver, 1993). The strong concern with second language learners has fuelled projects like SIOP (Sheltered Instruction Observation Protocol), a pedagogical concept designed to help teachers to plan and implement forms of teaching that support English learners’ simultaneous learning of academic knowledge and language skills (e.g. Echevarria et al., 2009), or the new Australian curriculum, with the design of Scopes and Scales outlining the language demands that need to be addressed across the curriculum in order to support ESL learners’ school success (e.g. Polias, 2003). Continental Europe entered these debates with a certain time lag owing to the historically later onset of mass immigration, but in the meantime academic and educational debates have been intense in many countries. This has motivated the Council of Europe to initiate the project Languages in Education, Languages for Education in 2005. Even though its original aim to adapt and expand the Common European Frame of Reference for Languages (Council of Europe, 2001) for the purposes of academic language has not been fulfilled as yet (see Thürmann & Vollmer, 2013), the project has produced a number of valuable documents providing curriculum writers with subject-specific as well as cross-curricular notions and categories that can be worked into national curricula (e.g. Beacco, 2010; Beacco et al., 2010; Linneweber-Lammerskitten, 2012). As regards national curricula, growing attention to the importance of language across the curriculum is reflected, for example, in the recent curriculum renewal work in Finland. The new core national curriculum was launched in Finland in 2016, and greater language emphasis is reflected, for example, in multiliteracy featuring as a core factor in the broad competence areas needed in the future, and in the acknowledgment that increasing cultural diversity needs to be accompanied with better language awareness in education (Opetushallitus, 2014). The curriculum work in Finland has no doubt been influenced by the 2006 curriculum reform in Norway, where language aims are visible in the core curriculum (The [LK06] National Curriculum for Knowledge Promotion in Primary and Secondary Education and Training) in the way each subject curriculum is expected to incorporate, in addition to its specific content requirements, five basic skills of reading, expressing oneself orally, expressing oneself in writing, developing numeracy and using digital tools (Norway, 2006). At the moment of writing, research and development also seems to be particularly intense in Germany where language and subject educators are working on a range of projects fostering cross-curricular sensitivity to

6

Conceptualising Integration in CLIL and Multilingual Education

language among policymakers as well as classroom teachers (e.g. Salem et al., 2013; Thürmann & Vollmer, 2010). In sum, it can be stated that educators and educational systems realise that learning in a linguistically challenging milieu – as is the case in multilingual education – can be facilitated with careful language intervention. A widespread notion in North America is that of sheltering (see SIOP) which invokes associations of learners ‘being out in the cold’ and images of learners receiving extra nourishment so as to be able to catch up with mainstream L1 learners. By contrast, the idea that CLIL learners are specifically needy does not seem to be widespread. Rather, they are seen as mainstream, or even a particularly capable and privileged band of said mainstream. A decisive difference between CLIL and immigrant learners is the following: even though both groups study a national curriculum in their second language (L2), the position of teachers towards said L2 is crucially different (see Martín Rojo, 2013). CLIL learners are commonly taught by content teachers who share the students’ status as non-native speakers of the language of instruction. Consequently, they see themselves as being in the ‘same boat’ as their students, having to work through the curriculum in a language that they may use with less confidence than their L1. ‘The problem’ (if there is one) is thus ‘the language’ and it is a problem shared between teachers and students. For immigrant learners this situation rarely holds: there are few countries, if any, where significant numbers of descendants from immigrant families populate the teaching profession. The normal case is frequently for linguistically unaware teachers to place the problem entirely with the learners: if only these learners knew better German/Spanish/English/Finnish (the country’s majority language) before coming to class, there would be no problem. That an adapted pedagogy might go some way to solving that problem was broadly discussed with regard to lower-class L1 learners in the 1970s but seems to have been sidelined since. The term ‘academic language’ was quite often used in earlier literature to make a distinction between everyday language and the language of schooling. The characteristics of academic language have usually been related to its higher level of abstraction, its complexity and grammatical density as well as the specialised terms and concepts within each content area (e.g. Achugar & Carpenter, 2014; Schleppegrell, 2004). The research object for work in different traditions remains the same: to explore the interface of content schemata (as in subtraction in maths, warfare in history or upwardness in PE) and language units co-occurring with these notions at the major language levels: genres, functions and lexicogrammar. The quality of multilingual education – often seen as the ultimate rationale (e.g. Brisk, 2006) – depends critically on providing content learning with enhanced language attention. Educational quality could indeed go a long way if the language in the bilingual classroom were

More Than Content and Language: The Complexity of Integration

7

predictable and identifiable. We agree with Vollmer and Thürmann (2013), however, that the large majority of research on academic language has been concerned with aspects of linguistic form rather than function (the genre approach needs to be exempted from this, though). Content and language integration needs to enhance the functional aspect in bilingual education with an elaborated theory of language and language learning and with appropriate analytical tools. Some suggestions on how to counterbalance form with function will be made in this book (cf. for instance the chapters by Dalton-Puffer and Llinares & Nikula).

The Different Perspectives on Integration Saying that language in use always entails content is trivial in that linguistic utterances always have meaning, meaning which is realised even in a routine written genre such as a shopping list. In that sense, of course, CLIL is not special at all. Moreover, the meaning potential of linguistic utterances is not realised until language users construct meaning with them and since participants may find themselves in any number of different situations or contexts, the actualised meaning or ‘content’ they will construct from a given linguistic form can vary considerably. In other words, the sense humans make of what others and they themselves say, crucially depends on how they frame what they are experiencing. The context and how it is framed is also a crucial consideration when discussing language and content integration in CLIL: we are dealing with educational, institutionalised contexts characterised by institutional logic whereby, certainly from secondary level onwards (but partly also at elementary level), schools are organised along cultural constructs called ‘subjects’, e.g. mathematics, history, foreign language, sports. These subjects are not natural kinds but the result of historical processes and rest on the social agreement that they serve a purpose in structuring knowledge and skills deemed societally relevant enough to be passed on to the younger generation in an organised fashion. The knowledge and skills themselves are laid down and talked about in documents called ‘curricula’, wherein those in authority also formulate goals that learners should have reached with regard to certain knowledge and skills areas by a certain stage in their educational careers. Consequently, these are the goals which teachers should be helping learners to achieve; in other words, this is what they should teach. However, the significance of the ‘subjects’ goes further than curriculum documents and determines the daily temporal and material reality of teachers and learners. Teachers have studied to be teachers of one or more specific subjects, they are ‘history teachers, maths teachers, French teachers’, who have been socialised into the specific discourses and

8

Conceptualising Integration in CLIL and Multilingual Education

practices of history education, mathematics education, French language education for instance. Pupils take state examinations in specific subjects and school administrators normally design timetables that are arranged into subjects following upon each other in a roughly hourly rhythm during the school day. For the learners, these hourly slots are connected with specific people (teachers), textbooks and materials, types of activities, amounts of homework to do, degrees of ‘seriousness’ and ‘importance’ and the like. Even though teachers very much have their individual styles too, there are typical things that English as a Foreign Language (EFL) teachers do that physics teachers do not and there are typical things sports teachers do that geography teachers do not. For all stakeholders involved, the shorthand for these typical constellations of people, facts and practices (which additionally vary linguaculturally, of course) is to say that ‘now it’s English/history/science’ (see Hüttner et al., 2013). This multidimensional constellation of CLIL teaching as a site where diverse institutional, educational, personal and pedagogical scripts and purposes intersect, in our view, necessitates a correspondingly multidimensional notion of integration. In this volume, we make a proposal for conceptualising this multidimensionality by identifying three perspectives on integration that will also form the basis for structuring the volume. The three perspectives have been influenced by Barwell’s (2005a) suggestion that four different dimensions can be identified in research on language and content: the policy and curriculum dimension, the institutional dimension, the classroom interactional dimension and the theoretical-methodological dimension, the latter entailing assumptions, models and theories relating to the conceptualisation of content and language in research and practice. In shorthand, the approach we suggest covers the following three perspectives on integration: the what, the who and the how. The first perspective concerns the institutional level of planning curricula and pedagogies where decisions have to be made regarding what will be integrated and how, preferably based on well-articulated ideas about the role of language in content learning. The second perspective has to do with participants, referring in particular to their perceptions of and beliefs about integration as an influential factor in any contentbased scenario. The third perspective shifts the focus to the local level of classrooms and content and language integration as a matter of in situ practices (see also Spolsky [2004] for a similar division of language policy matters into language practices, ideology and beliefs and management and planning). Figure 1 illustrates the three perspectives. As the figure suggests, the three perspectives are simultaneous and interconnected rather than discrete. That is, classroom practices are, of course, influenced by broader curriculum and pedagogy decisions and participants’ individualised beliefs and perceptions. However, it

More Than Content and Language: The Complexity of Integration

Curriculum & pedagogy planning

9

Participant perspectives

Classroom practices

Figure 1 Three perspectives on integration

is useful to explore content and language integration from these three perspectives because it makes it possible to approach integration in a more comprehensive manner than has to date been the case in CLIL research and highlights that language and content are involved in each perspective. Note that learning outcomes (an area that has received much emphasis in CLIL research to date, see e.g. Ruiz de Zarobe, 2011) is not presented in the model as a distinct perspective but rather something that may be addressed both at curriculum and pedagogy planning level (e.g. how desirable integrated content and learning objectives are articulated and defined), at participant level (how an individual perceives of and/or succeeds in CLIL) and classroom practices level (how CLIL manifests itself in language use). In the following sections, we will discuss each of these perspectives: the interface of language and content as a concern for curricular and pedagogical planning, integration as a matter of participants’ perceptions and beliefs and integration as realised in classroom practices. In sum, we focus on what is being integrated at different levels, who the participants are and what kind of perceptions they lean on or develop and how content and language integration takes place in the context of the classroom.

Integration as a concern for curriculum and pedagogy planning It is obvious that an important step in the exploration of content and language integration is how it gets conceptualised at curriculum level. As was pointed out above, the importance of curricula lies in the fact that they incorporate what education authorities define as the learning matter and learning goals within the confines of particular ‘subjects’ and

10

Conceptualising Integration in CLIL and Multilingual Education

what, consequently, is to be taught and learned in a nation’s schools at particular grade levels. Today, curricula also make recommendations on how the subject matter should be taught, especially in order to ensure that the desired competences are built. However, education systems do differ in the extent to which they actually steer and control the implementation of their curricula: some rely on standardised national exams across the board, others rely on the idea that the use of officially approved textbooks will ensure that the curriculum is implemented in classrooms but leave assessment entirely in the hands of teachers. Additionally, this may vary from subject to subject with L1 (or majority language), mathematics and first foreign language being more prone to centralised control than others. As documents, curricula contain a number of typical components: they formulate goals (what learners should know and be able to do at the end of the year/school career) and often they also say how these goals are best achieved. With regard to the specification of goals, a historical development has been taking place over a number of decades (with different onsets in different countries) that prioritises procedural over declarative knowledge: learners should not only know and remember but also be able to do. The famous taxonomy by Bloom et al. (1956) was the landmark starting point to that and a comparative-historical analysis of curricula in most Western countries (and presumably elsewhere, too) would probably show the space given to procedural knowledge increasing with successive versions of the same curriculum. Alongside the procedural knowledge element, curricula continue to cover declarative knowledge in the form of lists and tables of topic areas, concepts, facts and issues to be covered. CLIL and other forms of bilingual education involving integrated language and content teaching challenge the often taken for granted separation between ‘language subjects’ and ‘content subjects’. An integrated curriculum has been considered an innovative alternative to institutionalised conventional education, a shift from a collection curriculum, one that separates content into traditional subjects and draws on the historical canon of each discipline to delineate the borders of courses. Shifting from a collection to an integrated curriculum entails major consequences in educational organisation and praxis. Curriculum integration implies a move from mechanic to organic pedagogic practice, or put another way, from the sacred to the profane (Sadovnik, 2001: 689). Therefore, integrated classes can be seen as featuring new teaching scenarios with content teachers planning their maths or science tasks within language sensitive frameworks that include some sort of communicative practice in the L2. As a result, the linguistic nature of subjects or the subjects as represented in language are more visible in integrated practices. This, of course, requires close collaboration between content and language

More Than Content and Language: The Complexity of Integration

11

teachers and one of the important decisions for CLIL schools is how to realise such collaboration. These practices reflect classical theories of education holding that language criss-crosses all areas, and consequently that all disciplines have to take responsibilities in advanced language structure development (that is, vertical discourse, as opposed to basic streetwise and domestic language also known as horizontal discourse; Bernstein, 1999). The original formulation included a social twist: that ultimately language and social structures run parallel. Ideologically, therefore, integration must be framed in the educational tradition of progressive democratic and inclusive education. As a result, ‘curriculum integration is not simply an organizational device demanding cosmetic changes or realignments in lesson plans across various subject areas. Rather it is a way of thinking about what schools are for, about the sources of curriculum and about the uses of knowledge’ (Beane, 1997: 616). Concurrent with the idea of integrated curricula is the shift in emphasis from language competence as a set of skills to a view of literacies as social practices embedded in social and political processes and structures, a change that is carried over to curriculum planning (e.g. Wyatt-Smith & Cumming, 2000). Language education, therefore, impregnates all areas and appears in curriculum design in the description of financial literacy, health literacy and digital literacy, the expression of a new turn that demands the construction of knowledge around real-world competences. This interaction of disciplinary worlds renders a more accurate picture of knowledge construction. Classrooms are said to exist in a world of intersemiosis (O’Halloran, 2007). Maths, for instance, is often embodied in semiotic systems intertwined: language (first or second), visual imagery and mathematical symbolism co-occur and provide the informational input for knowledge construction. It is out of this blending of various inputs that knowledge takes shape. In a science class: ‘language remained the main tool of conceptualization and classification, but it was of use only when integrated with mathematical and visual representation’ (Lemke, 2002: 10). Integration, and this will be addressed in the chapters of this book, needs to explore the interfaces of content and language and sort out both content and language units in a merged template, plan or curriculum. The fact is that even when integration does appear in official curricula, language matters in content courses have often been programmed on instinct, against no linguistic backdrop and without a proper language theory supporting them, based for the most part on the impromptu language needs that teachers identify in their students. The end result has often been, as content-based instruction (CBI) experts acknowledge, that ‘while research and experience indicate the advantages of a contentdriven curriculum in foreign language classrooms [...] our educational

12

Conceptualising Integration in CLIL and Multilingual Education

bureaucracy, not interdisciplinary by nature, perpetuates the separation of language and content’ (Stryker & Leaver, 1997: 7). This suggests that truly integrated curricula are indeed a rare find, and also indicates a need for well-developed, research-based conceptualisations and models as tools for practitioners to make better sense of content and language integration and the tensions between academic and everyday language (Barwell, this volume). In this volume, chapters by Dalton-Puffer, Berger, and Lorenzo and Dalton-Puffer address these issues by suggesting ways of modelling content and language integration. For institutions and schools, the demands are many: content and language courses need to be merged in an integrated whole known by different names: integrated curriculum, school language plan or genre map (see Corson [1999] for a classical formulation and Lorenzo [2013] for a genre map for history CLIL courses). Countries like Germany, Sweden or Finland (Egelund, 2012) have implemented or drafted plans for integrated language education, normally resting on some functional language approaches in order to raise or sustain literacy levels. In some of this literature, genre-based models have been presented as a way forward in curriculum planning because genres incorporate language units and functions (terminology, structures and rhetorical moves) that content teachers know well. Language is better presented in genres, superordinate language units that mould knowledge as constructed in the discipline, than with language seen as a battery of decontextualised skills, as a content expert can better engage in language discussion when this takes the form of genres they are acquainted with: a chronicle in history, a problem statement in economics or a fieldwork observation report in biology (e.g. Lorenzo, 2013; Morton, 2010). Moving on from the curriculum level to the planning of classroom pedagogy such as lesson plans or assessment practices by teachers, integration demands cooperation between peers – teachers’ collaborative work ranks high on the list of educational enhancers in international reports like the OECD (2009) Teaching and Learning International Survey (TALIS). When language is planned across the curriculum, content and language teachers need at some point to put their heads together and produce a joint scheme that gives a sense of structure to the class. Integrated lesson plans start with content units (integers in maths or track and field sports in PE), represent them in discourse events (describing an ecosystem or comparing warfare strategies) and further split them into language units of different calibre (genres, functions/notions, vocabulary, sentence grammar units, etc.). As part of integration, every subject area plans academic tasks (lectures, field trips, observation sheets, text compositions, problems) which amount to communicative activities which may incorporate L2 practices. The end result is a hybrid of content teaching and task-based L2 teaching. Teachers select and practice language difficulties (pre-task phase) before the actual content task is done (task phase) and a final language reflection stage

More Than Content and Language: The Complexity of Integration

13

follows (post-task phase) (Lorenzo, 2007). The goal here is to map out the actual language sinews that activate content. This approach has proved useful for increasing language awareness in CLIL settings and has inspired templates for bilingual material development based on language description (Moore & Lorenzo, 2015). In the context of American sheltered instruction referred to earlier, integrated lesson plans operate in two stages: (a) under the name grammatical deconstruction, the language in the lesson is first focused on; followed by (b) text exploitation exercises such as sentence chunking, analysis of reference devices and interpretation of complex noun phrases. Research reports very positive gains in history content learning for groups taught with these literacy strategies (Schleppegrell & Oliveira, 2006). As such, this approach provides a regular systematised series of tasks that foreground language in the disciplines. The limits of this approach to date is that there is a poor mapping of the language components of the disciplines and all the emphasis is put on making input understood. Germane to sheltered instruction, another approach based on text construction and reconstruction – the reading to learn project – presents promising results in language migrant education in Sweden (Acevedo, 2010). Assessment is an integral part of the curriculum and a crucial component in every teacher’s pedagogy even if not always specified in the actual documents or explicitly defined by teachers themselves (e.g. Wewer, 2014). A practical tool for integrated assessment is suggested by Llinares et al. (2012), based on CLIL teachers’ attention to the language and literacy practices through which students will demonstrate knowledge in their subjects. The desired content goals are first identified, and then the assessment tasks through which students will demonstrate the required knowledge are chosen, by giving formative (which can also be summative) scores in a content-language integrated scale. Furthermore, these authors advocate for mediation during dynamic assessment, based on classroom interactional practices that can lead to assessment for learning (see section on integration in classroom practices below). Other proposals toward integrated content and language assessment come from Massler et al. (2014) in the form of an assessment tool based on language descriptors derived from the Common European Framework of Reference (Council of Europe, 2001) and Lenz and Studer’s (2007) lingua-level model, with the description of competences in the subject content areas and the curricula of the subjects involved in their thematic categories.

Integration as a matter for participants Above, language and content integration was discussed as a concern for curriculum design and for the planning of integrated pedagogy in the form of lesson plans and plans for assessment. The latter already brought

14

Conceptualising Integration in CLIL and Multilingual Education

in the importance of the teachers’ role but in this section the individualised character of content and language integration will be addressed in more detail from the participant – both teacher and student – perspectives: how they experience content and language integration, what their beliefs and attitudes to it are and, also, how the CLIL experiences at the interface of language and content instruction may impact upon their (professional) roles and identities. With this, we want to emphasise that the way individuals perceive content and language integration is an important mediating phase between content and language integration as articulated in curricula, pedagogical plans and guidelines, on the one hand, and, as translated into pedagogical practice, on the other hand. The role of teachers’ beliefs for their pedagogical practice is an area that has been of interest both for general educational research and for applied linguistics (for overviews, see Basturkmen, 2012; Borg, 2003) because even if there is no direct causal relationship between beliefs and action, beliefs are thought to greatly influence educational practice (e.g. Borg, 2006; Kalaja & Barcelos, 2003). In educational research, ‘teacher cognition’ is often used as an overarching term to refer to teacher beliefs, attitudes and knowledge (e.g. Borg, 2006; Woods, 1996). Similarly, student beliefs and attitudes have been widely researched in education where teacher awareness of learners’ initial state of content knowledge is a crucial component in influential models such as Pedagogical Content Knowledge (e.g. Park & Cheng, 2012; Shulman, 1986). Applied linguistics too has paid a good deal of attention to learner attitudes and cognitions relating to language learning (cf. Kalaja & Barcelos, 2003; Wesely, 2012 for an overview). In CLIL and other forms of bilingual education where language and content teaching and learning objectives merge, an important consideration is how teachers conceive of such integration, i.e. how they perceive of the relationship between content and language in their professional practice. With the focus on immersion contexts and teachers’ own reflections, Cammarata and Tedick (2012: 257) argue that one of the main influences of immersion on teachers has to do with their need to ‘revisit and reshape their teaching identity – that is envisioning themselves not only as content teachers but language teachers as well’ (which implies that teachers’ preimmersion identities are strongly shaped by the subject they teach). They also point out how balancing language and content also involves issues related to stakeholder expectations. The struggles are partly due to the fact that immersion teachers tend to mainly perceive of themselves as content teachers as they are accountable for content matter achievement (for similar observations, see Fortune et al., 2008; Tsui, 2011; Walker & Tedick, 2000). Therefore, the reshaping of teacher identity towards both content and language teaching is one often requiring considerable time, a process that could be supported, for example by teacher education.

More Than Content and Language: The Complexity of Integration

15

As far as European CLIL teachers are concerned, the same balancing of content and language is a major concern, but the equation is potentially further complicated by the fact that most teachers are non-native speakers of the instructional language themselves, which may make the identity reshaping an even more complicated journey. For example, Moate (2011) reports on an interview study with primary school CLIL teachers, which revealed how CLIL may challenge teachers’ assumed expertise and identity so that they need to re-establish their sense of professional integrity. Furthermore, the fact that CLIL teacher profiles and backgrounds may differ considerably across geographical contexts is another area that requires more attention, also from the viewpoint of teachers’ identity. For example, while in Austria and Germany secondary school teachers have dual qualifications and can be both content and language experts, in Spain and Finland they are content experts. In CLIL research, the question of teacher orientation towards content and language integration has been addressed in particular by studies examining CLIL teachers’ orientation to language. Morton’s (2012) in-depth study explores CLIL teachers’ knowledge of language or language awareness. It explicates the multidimensional nature of teachers’ orientation to knowledge, illustrated for example by the framework of ‘modes of knowing’ (Morton, 2012: 108–111), which involves personalpublic and theoretical-practical knowledge dimensions that have different emphases in different contexts. Morton (2012) investigates these different modes in relation to three teacher perspectives of language: language as a tool for learning, language as a curriculum concern and language as a matter of competence. As an overall conclusion, Morton (2012: 298) argues that CLIL teachers’ ‘language awareness in the three perspectives was largely derived from personal experience and practice, and did not generally engage with the public theories and practices’, i.e. their approach to language was not systematic and hence much L2 focus remained incidental. The conclusions by Koopman et al. (2014) are along similar lines: they explored subject teacher knowledge about language teaching in CLIL, showing that even when content teachers did display awareness of a variety of pedagogical procedures to support language teaching in their content lessons, this tended to rely on non-explicated theoretical knowledge of language learning and development. Bovellan’s study (2014) is also concerned with CLIL teacher beliefs regarding language: she interviewed Finnish primary school CLIL teachers to explore what their discussions relating to CLIL materials revealed about their views on language and learning. One of her main observations was that a language as a system view and concerns about matters of formal accuracy prevailed, rather than the integral role of language in conveying content matters. Hüttner et al. (2013) have also looked into CLIL teachers’ beliefs in a study in which they explored upper secondary level teacher and student

16

Conceptualising Integration in CLIL and Multilingual Education

beliefs about language learning, the effects and benefits of CLIL and the construction of success regarding CLIL. They found that relative absence of language management combined with strong positively oriented beliefs relating e.g. to changes in affect towards English and to increased selfconfidence as foreign language users resulted in the construction of lay theories of CLIL that emphasised its success. An important reason for identity struggles in CLIL teaching is that even if it is often argued that collaboration between content and language teachers is an important CLIL quality factor (e.g. Baetens Beardsmore, 2009; Pavón Vázquez et al., 2015), a great deal of variation remains in how much such collaboration is a part of CLIL practice, with content teachers usually having the main responsibility of CLIL teaching (e.g. Koopman et al., 2014). A pending issue so far largely neglected in CLIL teacher research is the identity of the language teacher, whose role necessarily changes in CLIL programmes probably in two main ways: (a) their subject may need to be adapted as their students may have other needs compared to those who only study the language as a subject; (b) they are expected to collaborate with content teachers in the identification of language-related areas that require attention in both content and language lessons. It is also important to take into consideration student perceptions of content and language integration as those may greatly shape their orientation to content-based teaching. So far, less research has been published on students than on teacher perceptions of content and language integration. In L2 learning research more generally, however, learner orientations have received a fair amount of attention, whether under the headings of beliefs (for overviews, see Barcelos, 2003; Wenden, 1999) or learner attitudes and motivation (for overviews, see Dörnyei & Ushioda, 2011; Gardner, 2005). Over the years, research emphases have shifted from investigating what learners believe about learning to ‘how beliefs develop, fluctuate and interact with actions, emotions, identities and affordances and how they are constructed within the micro- and macro-political contexts of learning and teaching languages’ (Barcelos & Kalaja, 2011: 282). The dynamism between fluctuating and stable aspects of learners’ mind structures as well as the importance of peers in the construction of the sense of self is also highlighted in a recent study by Skinnari (2012) on elementary school pupils’ identities as foreign language learners. Moving closer to CLIL and language learning in contexts of contentbased education, there are a number of studies conducted in immersion contexts in particular that have approached learner perspectives to language or language learning from the viewpoint of either attitudes or motivation (see e.g. Lambert, 1987; Lindholm-Leary & Ferrante, 2005; see also overview by Block, 2011). Overall, findings have suggested that immersion helps in developing positive attitudes towards both the

More Than Content and Language: The Complexity of Integration

17

instructional language(s) and their speakers. However, as Wesely (2009: 281) points out, large-scale results obtained via questionnaires and surveys may even out matters in ways that undermine the importance of ‘microcontexts’, i.e. the way in which ‘relationships with teachers, [...] peers in the immersion program, and peers outside of the immersion program, have the potential to be both positive and negative influences on student language learning motivation’. Turning to student-oriented research in CLIL, the bulk of it has concerned the impact of CLIL on their learning outcomes, both in the areas of language and content competence as well as L2 interactional skills. There are studies that have tapped into student attitudes but quite often those have explored how CLIL affects student attitudes towards languages (e.g. Lasagabaster & Sierra, 2009), towards foreign language learning (Merisuo-Storm, 2007) or how it may influence their levels of motivation more generally (Doiz et al., 2014) rather than focusing on student perceptions of CLIL as such. Recently, however, such research has started to emerge. Pihko (2010) and Coyle (2013) have both explored CLIL from the perspective of secondary-level CLIL students’ subjective learning experiences, in the contexts of Finland, and England and Scotland, respectively. Pihko’s (2010) study suggests that even in cases where students have an overall positive orientation to CLIL, performing with limited language skills is perceived as affectively demanding by some, with language anxiety leading to reduced classroom activity. Coyle (2013) focused on exploring students’ understandings of successful learning in CLIL, her results suggesting that students appreciate the spontaneity of language use and learning that extends beyond the word level towards subject-related meaning construction in CLIL. A sense of achievement was also identified as a motivating force. The role of affective factors is also brought up by Hüttner et al. (2013) whose study on CLIL students and teachers in upper secondary colleges of technology suggests that CLIL success relates to students’ sense of gaining confidence and to their self-perception as more skilful users of English than students without CLIL provision. This brief review of earlier studies points towards the importance of researching participant perceptions and beliefs about CLIL in general and towards the dearth of research focusing on participant conceptions of content and language integration in particular. These questions will be addressed in this volume in chapters by Dafouz, Hüttner and Smit as well as by Skinnari and Bovellan.

Integration as a matter of classroom practices We have sketched above how integration is, on the one hand, a matter of institutional policies and decisions as realised – whether explicitly

18

Conceptualising Integration in CLIL and Multilingual Education

or implicitly – in curricula and lesson plans, and that stakeholders’ personalised beliefs and perspectives also form a crucial consideration. This section highlights the need to also look at actual classroom practices to understand how content and language integration is realised in local and situated instances of language classroom interaction, i.e. to understand the socially and performatively situated aspect of integration. In the following, we discuss the importance of exploring how teachers integrate language and content in their discourse, how students show academic knowledge through language and how different tasks and activities lead students to approach content from different perspectives and to use varied and adequate academic and interpersonal linguistic resources. In other words, it is necessary to investigate classroom practices from all perspectives, from the moment teachers introduce new content, through to the students engaging with that content, to the type of assessment carried out and to the actual materials that scaffold these processes (Banegas, 2012). We pointed out in the section above that integration as the inherent interconnectedness of content and language and as something that gets realised as subject-specific academic literacies is not a novel idea. It is a crucial matter for any learner but it seems that contexts where L2 is used as the instructional tool have in particular brought to consciousness the need to attend to ‘language work’ also in content subjects. Schleppegrell and O’Halloran (2011: 5) provide an overview of research on teaching academic language in English-speaking (mainly US) L2 settings and point out that apart from knowledge of language in their content areas and applying this knowledge in the planning stages, teachers also ‘need strategies for engaging students in robust ways in exploring language and content in the moment-to-moment unfolding of instruction in the classroom’ (see also Cummins & Man, 2007; Lindholm-Leary & Borsato, 2006; Snow & Uccelli, 2009). In the context of bilingual education, there have been calls to carry out classroom-based research in order to understand how content and language are best learnt in integration (Cenoz et al., 2014; Leung, 2005b; Llinares et al., 2012). In immersion contexts, classroom discourse research has often been motivated by language learning concerns, hence much attention has been paid to the ways in which teachers could best make language matters salient, in order to support learner awareness of morphosyntactic matters in particular (e.g. Lyster, 2007; Lyster et al., 2013), with fewer studies orienting to the learning and using of subjectspecific aspects of language (but see Laplante, 2000; Pekarek Doehler & Ziegler, 2007). In CLIL research, language and content integration as a phenomenon of classroom practices has started to attract more attention during the last few years. With the research focus on teachers, existing CLIL

More Than Content and Language: The Complexity of Integration

19

studies with different theoretical and methodological perspectives such as discourse analysis, conversation analysis (CA) or systemic functional linguistics (e.g. Dalton-Puffer, 2007a, 2007b; Llinares et al., 2012; Morton, 2012; Nikula, 2010, 2015; Smit, 2010a), have offered important insights in this respect. For example, by focusing on aspects such as the role of different types of academic questions, the integration of content and language in teacher feedback or the way teachers differentiate between everyday and academic talk, these studies have shown that teachers, who are L2 users themselves, quite rarely make subject-relevant language use a salient focus. It thus seems that teachers need to become more aware of the scaffolding strategies that can be used to support the development of students’ linguistic resources in the expression of academic content (see Llinares et al. [2012] for suggestions on how teachers can become more aware of the different roles of language in CLIL). Another important contextual matter apart from teacher linguistic backgrounds, also worthy of research attention, is the link between the use of language in the classroom and their professional background. Some studies have shown differences between CLIL teachers with only content expertise and teachers with both EFL and content expertise, regarding for example the types of questions and feedback used and their effect on student performance (Vázquez, 2010; Whittaker & Llinares, 2009). More comparative studies across geographical contexts and educational levels are necessary in order to understand the effect of teacher expertise and training on classroom practices. With the learners in focus, the understanding of integration as an aspect of classroom practices involves paying special attention to how learners express the academic meanings required in the disciplines they are studying, how they move from oracy to literacy and how they are able to not only express factual information and ideational meanings but also to evaluate and take a stance on the academic content that they are learning. Some studies carried out in the Spanish context have looked at CLIL students’ oral and written expression of academic meaning in social science (history and geography), focusing on how this performance is related to the requirements of the curriculum and materials used (Whittaker & Llinares, 2009) and compared this performance to that of parallel groups studying the same subjects in the L1 (Llinares & Whittaker, 2010). The findings have shown that expressing meanings in a subjectappropriate manner is not an easy task for CLIL students, nor, for that matter, for students studying through L1 (for similar results, see Vollmer, 2008). In addition, longitudinal studies should be carried out that focus not only on students’ development of their foreign language competences (e.g. Ruiz de Zarobe, 2011) but also on the development of their L2 (and L1) resources to express academic meanings successfully (see Whittaker et al., 2011).

20

Conceptualising Integration in CLIL and Multilingual Education

Classroom discourse is obviously a joint accomplishment by classroom participants and it is therefore also of interest to explore content and language integration as an interactional phenomenon. Studies drawing on CA in particular have shed light on the ways in which content and language are jointly negotiated in CLIL classrooms, often paying attention to the performance of specific academic discourse functions. For example, a case study by Kupetz (2011) explores how students accomplish explaining in CLIL classrooms, drawing attention to the subtle resources students deploy to construct meaning and collaboratively negotiate subject-related content as well as linguistic form with the teacher and fellow students (see also Smit, 2010a, 2010b). Evnitskaya (2012), focusing also on the construction of explanation in eight secondary-level science lessons, takes into account both verbal and multimodal means with which explanations are co-constructed. Her findings point to great complexity and variability in how participants accomplish the action of explaining. Furthermore, an important aspect of joint meaning negotiation relates to the fact that it is often accomplished among peers. Jakonen and Morton (2015) and Jakonen (2014) have investigated interaction between peers to solve knowledge gaps concerning both language and content, and in this volume they continue it by researching ‘language’ and ‘content’ as a concern for participants in CLIL classrooms. Apart from CA approaches, research on CLIL classroom interaction as a site for negotiating language and content by both teachers and learners has also been conducted from discourse analytic and/or pragmatic and systemic functional perspectives, and with different focal interests (e.g. Dalton-Puffer, 2007a; Llinares & Morton, 2010; Nikula, 2005, 2008). Furthermore, Llinares and Morton (2012), when investigating interaction and language learning from conversation analytic and systemic functional perspectives, call for more research combining research perspectives. In this volume, a similar point is made by Llinares and Nikula who report on a study that combines discursive pragmatics and systemic functional linguistics. A special challenge in capturing and researching learners’ practices and learning processes with a combined focus on language and content is to achieve a comparable level of sophistication in theoretical and methodological approaches. After all, while school-level learners are socialised into different ‘subjects’, researchers are also socialised into different ‘disciplines’. That is why applied linguists normally have sophisticated conceptualisations and operationalisations of ‘language’ but rather impoverished or lay notions of ‘content’, whereas the reverse is true of subject educators. The solution lies, of course, in interdisciplinary research but relatively little such research has been done. This has to do both with the general obstacles to interdisciplinarity (e.g. Brewer, 1999) and the pecking order among scientific disciplines on which the study of

More Than Content and Language: The Complexity of Integration

21

language does not always rank very high. Perhaps educational realities like CLIL can serve as a catalyst in this respect: educational researchers with dual academic backgrounds might be well equipped to drive forward ‘integration’ also on the level of modelling, operationalisation and theorising. The fact that the two researchers with dual backgrounds writing in this volume (Barwell, Berger) have one foot in mathematics is probably not a coincidence: mathematics education was among the first to take a linguistic turn. Even though not the main topic in any of the contributions in this volume, it is worth bearing in mind that assessment forms an important aspect of CLIL classrooms practices and one where content and language considerations merge. Llinares et al. (2012) argue for dynamic assessment as a useful resource for assessment in CLIL. This type of assessment is related to the Vygotskian view of the teacher actively working with the learners and, rather than just focusing on what students know, the emphasis is on how they progress and the kind of mediation (scaffolding) that they need (see Leung, 2007). Integrated assessment also requires a combination of focus on form and focus on meaning. For instance, relevance should be given not only to formal recasts as a type of corrective feedback (Lyster, 2007) but also to functional recasts (Mohan & Beckett, 2003), which address language appropriateness (not correctness) to express academic content. In the area of classroom practices, the use of different tasks in enhancing integration is another research area that needs to be tackled. Kong (2009) argues that attention should be given to tasks that require complex language relations derived from curricular complex relations. In these tasks, which are genre related and, consequently, also subject related, it is necessary to identify the stages that characterise each genre in order to grasp the complex language relations that CLIL learners and teachers need to encounter (Llinares et al., 2012; Nikula, 2015). A second type of tasks that require more attention in CLIL research are those that are not subject or genre specific but form part of general teaching methodology and require students to address language and academic content in different ways: role-plays, presentations, group work sessions (see Llinares & Dalton-Puffer, 2015). With regard to how languages are integrated in the CLIL classroom, the fact that CLIL teachers are mainly non-native speakers of the target language means that their multilingual competences are more or less guaranteed. Rather than a drawback, CLIL teachers’ non-nativeness and the use of both L1 and L2 can contribute to enriching students’ content and language engagement in the classroom. For example, the positive role of translanguaging in bilingual classrooms has been demonstrated in different studies mostly focused on immigrant learners (e.g. Creese & Blackledge, 2010; Cummins, 2008a). These studies have argued for the

22

Conceptualising Integration in CLIL and Multilingual Education

interdependence of knowledge and skills across languages. In some CLIL contexts, like Spain, translanguaging in CLIL can be enhanced thanks to the collaboration between main teachers (non-native speakers) and language assistants (often native speakers) in classroom-situated activities (Dafouz & Hibler, 2013; Lorenzo et al., 2010; Méndez & Pavón, 2012). Although some of these studies have already indicated the potential advantages of using two languages in the classroom, more studies should be carried out at the micro level, focusing on concrete situated practices and activities, in order to understand the principled purpose of using the L1 and/or the L2 in the CLIL classroom, as well as the affordances that translanguaging creates for student participation and engagement in classroom activities. This is a question tackled in the chapter by Moore and Nikula in this volume.

Introducing the Chapters The volume consists of three parts, which are framed by the present introductory chapter and by a concluding overview chapter by Tom Morton and Constant Leung. Each part corresponds to the perspectives on integration introduced above. As was pointed out, we see connections between the perspectives rather than treating them as clearly demarcated entities. However, the three-way division of the volume is helpful in showing the chapters’ main entry point and contribution to the discussion on integration.

Part 1: Curriculum and Pedagogy Planning The chapters in this part relate to the ‘curriculum and pedagogical planning’ part of the model. They offer valuable tools for conceptualising and modelling the content and language relationship in ways that have genuine applicability for CLIL and other forms of bilingual education programmes, both as regards their curriculum design and pedagogical planning, and can help these programmes become clearer about their aims and objectives. In the first chapter, Christiane Dalton-Puffer approaches integration through a transdisciplinary construct of cognitive discourse functions, grounded in both educational and linguistic thinking, and links subject-specific cognitive learning goals with the linguistic representations they receive in classroom interaction. The aim of the construct is to conceptually order and reduce the multitude of academic language functions that are circulating in curricula and specialist literature alike, and to offer researchers and teacher educators a principled heuristic tool with which to access cognitive discourse functions and thus enhance their visibility and teachability. The chapter by Francisco Lorenzo and Christiane Dalton-Puffer is also concerned with modelling

More Than Content and Language: The Complexity of Integration

23

the interconnectedness of language and subject competence, with special focus on subject history. Drawing on the construct of historical literacy, the authors, firstly, outline a three-tier model of historical knowledge structures and show how these correlate with predictable units at different language levels: lexico-grammar, functions and genres. Secondly, they approach historical literacy as a developing trait, characterising the linguistic performance of learners in history courses in an L2 both in the written and spoken mode. The third chapter, by Angela Berger, is concerned with bilingual mathematics teaching and reports, based on think-aloud protocols, on learners solving arithmetic word problems to model the specific nature of the interplay between mathematics and language, including the questions of how EFL influences mathematical thinking. The chapter describes the ensuing Integrated Language and Mathematics Model (ILMM), which represents a significant step in characterising the nature of the interaction between mathematics and language during the process of solving L2 word problems and is thus an important contribution to the discussion on the nature of content and language integration in this volume. To conclude the first part, Richard Barwell examines the possibilities offered by Bakhtin’s theory of language to serve as an alternative approach to such conceptualisations of language that tend to be implicitly based on a ‘replacement’ model of learning in which learners’ informal, everyday, idiosyncratic language practices are gradually replaced by formal, standardised content area registers. Drawing on such key notions in Bakhtin’s theory as heteroglossia and centrifugal and centripetal forces, Barwell discusses how these ideas lead to a view of integration as expansion (rather than replacement) of language repertoires, highlighting the tensions that arise between these forces as the requirements of institutionally recognised forms of content language and learners’ diverse language practices meet.

Part 2: Participants This part contains the two chapters in the volume that have participant perspectives on CLIL as their starting point; both are concerned with teachers. The chapter by Emma Dafouz, Julia Hüttner and Ute Smit investigates university teachers’ beliefs on teaching and learning in English-medium education in multilingual university settings (EMEMUS) to tap into their conceptualisations of content and language integration. Drawing on the recently developed ROAD-MAPPING framework of the core dimensions in EMEMUS (Dafouz & Smit, 2014), the study analyses university teachers’ interviews conducted in Austria, Finland, Spain and the UK. Their findings suggest that teacher beliefs are shaped contextually and that integration of language and content is constructed on a continuum from ‘similarity’ to ‘difference’ when compared with

24

Conceptualising Integration in CLIL and Multilingual Education

long-standing monolingual education through the national language, these ‘similar’ or ‘different’ beliefs being contingent on perceived agency in teaching and learning, the institutions’ form of internationalisation and considerations of acculturation to university in general and the teachers’ academic disciplines more particularly. The chapter by Kristiina Skinnari and Eveliina Bovellan is concerned with CLIL teacher beliefs about integration and about its effects on their professional roles. The study is based on secondary-level CLIL teacher interviews conducted in Austria, Finland and Spain; the teachers interviewed were either science or history teachers. Their findings suggest that even if teachers acknowledge the connection between language and content and consider attempts to separate them artificial, they still often struggle in trying to understand the dual role of language and content in CLIL and its impact on their teaching. This concerns the role of language in particular, with teachers having different understandings of what language is, how it should be taught and by whom.

Part 3: Practices In this part of the book, attention is turned to integration as a matter of in situ instances of language use and interaction in CLIL classroom contexts. The chapter by Tom Morton and Teppo Jakonen explores the ways in which learners discover, and work on, their own ‘knowledge gaps’ around matters of linguistic form and meaning. Their analysis depicts in detail the linguistic, embodied and artefactual resources that learners deploy during interaction to resolve these gaps and how they achieve intersubjectivity when doing so. Their findings illustrate how a microinteractional approach can throw light on ‘cognitive’ constructs such as focus on form and languaging, and they also offer suggestions on how such an approach could be further developed in the contexts of bilingual education. In their chapter, Ana Llinares and Tarja Nikula examine the use of evaluative language in CLIL classrooms. Evaluation in classrooms is usually associated with teachers evaluating and assessing student performance. However, learning how to evaluate subject-related information and to use evaluative resources to establish social relations are important skills for students as well. Using data from different European contexts and combining systemic functional and discourse pragmatic orientations, the study explores resources of evaluation used for knowledge construction and for participation in the social context of CLIL classrooms. Their findings show that the use of evaluative language is affected by the discourse context and its interactional patterns but that cultural differences may also be at play, relating both to geographical localities and to different subject cultures. Pat Moore and Tarja Nikula close this section by approaching integration as a matter

More Than Content and Language: The Complexity of Integration

25

of language choice and merger. They use the notion translanguaging to highlight that interaction in CLIL classrooms is also about meaningful engagement in multilingual practices. The findings, based on classroom data from different European contexts, indicate two main categories of translanguaging. When content-oriented, translanguaging is explicitly oriented to and used to scaffold meaning negotiation in the teaching and learning of content. When socially oriented, translanguaging is unmarked in the unfolding talk, with participants orienting primarily to the flow of interaction. Overall, the findings suggest that multilingual practices are dynamic and functional, and contribute both to content learning and the maintenance of interactive flow but also that they are highly context dependent. To wrap up, Constant Leung and Tom Morton offer a synthesis of the volume based on each preceding chapter’s view of language competence, image or model of language learning and explicit or implicit pedagogical approach. Drawing on Bernstein’s (2000) work on pedagogic discourse, they introduce a four-part framework consisting of two intersecting axes: higher to lower disciplinary orientation in approach to language competence and higher or lower visibility in approach to language pedagogy, and argue that different positions in the framework result in tensions between what can be broadly described as centripetal or centrifugal tendencies in approaches to integration. Their contribution also shows the usefulness of ongoing dialogue on integration which will, no doubt, reveal further interfaces and convergences. We hope that the present volume can offer useful insights into this process.

Part 1 Curriculum and Pedagogy Planning

1 Cognitive Discourse Functions: Specifying an Integrative Interdisciplinary Construct Christiane Dalton-Puffer

Introduction In the introduction, we argued that content and language integrated learning (CLIL) needs to articulate substantial links between the pedagogies of different subjects like mathematics, history or economics and the pedagogy of language teaching in order to fulfil its promise of ‘dual focus’. The underlying idea of this chapter is, therefore, that integration actually lies in transdisciplinarity and that cognitive discourse functions (CDFs) constitute a conceptual and pedagogical territory where such transdisciplinarity can be achieved (Dalton-Puffer, 2013). Since learning as a cognitive event is not directly observable, the nearest we can hope to get are its observable analogues. In the case of CLIL, these analogues are to be sought in the secondary school classrooms and the discursive interaction between teachers and learners in them. Today, there is a broad consensus in education that classroom talk during lessons is the chief locus of knowledge construction and subjects are ‘talked into being’ (e.g. Edwards & Mercer, 1987; Ehlich & Rehbein, 1986; Mercer, 2000; Wells, 1999). However, it is not only the social construct of school subjects that is at issue, it is the activity of learning itself. Under a social and contextual theory of learning (implying a social and contextual theory of language), we must assume that participant verbalisations, which make the learning matter intersubjectively accessible and represent knowledge objects, thought processes and epistemological stance, are constitutive of learning itself. These verbal actions I call cognitive discourse functions (sometimes also referred to as academic language functions). CDFs thus are verbal routines that have arisen in answer to recurring demands while dealing with curricular content, knowledge items and abstract thought. The actional demands as such (e.g. classifying, hypothesising) and the requirement that students demonstrate the ability to enact them, are regular features in today’s competence oriented school-curricula. For learners in CLIL classrooms, however, operating in an imperfectly known second or foreign language, 29

30

Part 1: Curriculum and Pedagogy Planning

the linguistic resources presupposed by the enactment of these competences are often precarious, a situation that may also hold for CLIL teachers who normally share their students’ status as second language (L2) users of the medium of instruction. So subject-specific language issues would need to be addressed in the classroom, but content-subject specialist CLIL teachers view this as outside their expertise and responsibility (except vocabulary). It is my contention that CDFs and their linguistic realisation may be a pivot that can change this view and give CLIL teachers the perspective that when they are modelling/teaching how to verbalise subject-specific cognitive actions, they are not ‘doing the language teachers’ job’ but actually teaching their subject in a very substantial way. This chapter, then, approaches integration via a transdisciplinary construct of CDFs, grounded on both educational and linguistic concepts, and links subject-specific cognitive learning goals with the linguistic representations they receive in classroom interaction. The rationale of the construct lies in its aim to conceptually order and reduce the multitude of academic language functions that are circulating in curricula and specialist literature alike. Its aim is to enable researchers and teacher educators to access CDFs via a principled heuristic tool which enhances their visibility (and ultimately teachability) in naturally unfolding classroom interaction. Next, the theoretical rationale of the construct will be briefly introduced (for a full account see Dalton-Puffer, 2013). The main part of the chapter is dedicated to a description of the seven components of the CDF construct, illustrated with examples from naturalistic CLIL classroom discourse.

Theoretical Grounding and Description of the CDF Construct Multiperspectival theoretical grounding in education and linguistics The formulation of learning goals and competence models is a central concern of educational research and development and Bloom’s Taxonomy of Educational Objectives (Anderson et al., 2001; Bloom, 1956) is certainly one of the seminal texts in this respect. A cascade of publications in different contexts and educational levels (e.g. Bailey et al., 2002; Biggs & Tang, 2011) have presented further attempts at formulating coherent taxonomies and identifying verbal behaviours that can serve as indicators of learners having reached a particular learning goal (normally in the shape of can-do statements of the kind can compare X and Y, can elaborate W etc.). All of these approaches have in common that they take a curricular perspective, that is to say they set standards for, rather than examine the practice of, teaching and learning. An analogous perspective has also been adopted in the Council of Europe’s project Language(s) in Other Subjects1 which aims at

Cognitive Discourse Functions

31

systematically cataloguing the linguistic requirements arising in connection with participating in lower secondary history, science, mathematics or literature classrooms across a number of European education systems (e.g. Beacco et al., 2010), its aim being to improve support of at-risk learners. In the German context this has led to the proposal of a frame of reference for German as an L2 at lower secondary level (Thürmann & Vollmer, 2013; Vollmer & Thürmann, 2010) which I consider a milestone in making visible the ‘hidden language curriculum’ in the content subjects. In these and numerous other projects and studies, a broad range of verbs designating cognitive-verbal actions are repeatedly mentioned. A comprehensive literature review has produced an inventory of 54 such verbs in English and the extent and complexity of this lexical field clearly demands a structuring construct that makes it operationalisable for different purposes (see Dalton-Puffer, 2013; Lackner, 2012). The mapping of words onto action and action onto words is a central concern of linguistic pragmatics. Since Austin (1962) and Searle (1969), the understanding that linguistic utterances constitute actions has become universally accepted. A broad range of research has applied Searle’s typology of speech acts (representatives, directives, commissives, expressives, declaratives) in the description and comparison of numerous languages, as well as in language acquisition research (Rose & Kasper, 2001) and applied linguistics (e.g. Trimble, 1985). Trimble (1985) bases his description of English for Science and Technology on an analysis of standard communicative intentions in technical expert communication and presents their routine linguistic realisations for the benefit of L2 users of English. A similar approach was taken by 1980s East-German scholars (Hoffmann, 1988; Schmidt, 1981) who also focused on technical communication. Finally, lists of speech act verbs (e.g. Verschueren, 1980: 7) contain numerous items that also feature in the formulation of curricular learning goals and can thus be considered to represent academic language functions. My understanding of CDFs thus is one that regards them as the product of recurrent situative demands arising in the context of organised learning events, i.e. lessons. Put differently, CDFs are patterns which emerge from the needs humans have when they deal with cognitive content for the purposes of learning, representing and exchanging knowledge. They offer participants involved in knowledge-oriented communication, patterns and schemata of a discursive, lexical and grammatical nature which facilitate dealing with standard situations where knowledge is being constructed and made intersubjectively accessible.

The CDF construct As briefly sketched above, the CDF construct is based on the pragmatic postulate that it is a speaker’s communicative intentions that materialise

32

Part 1: Curriculum and Pedagogy Planning

as speech acts. In the case of CDFs, these intentions concern the desire to externalise cognitive processes. Within the logic of mainstream pragmatic theory it thus makes sense to assume that there is an underlying communicative intention of the speaker to let others know which cognitive steps they are taking in handling subject content, in sharing knowledge items and structures and in making them intersubjectively accessible. Intersubjective accessibility is the precondition for institutional learning to become possible at all and one must hence assume that such communicative intentions become relevant for all participants in the learning situation at different times. In other words, CDFs concern both learners and teachers. As one surveys the 54 verbs extracted from curricular documents (Dalton-Puffer, 2013) in terms of their underlying communicative intentions, an interesting semantic structure emerges: almost all of the verbs can be subsumed under seven basic communicative intentions, thus producing seven basic types of CDF. These are assembled in the CDF construct depicted in Table 1.1 where Column 2 contains the seven basic communicative intentions formulated in simple everyday language. The designations of the function types in the left column are a deliberate choice in order to underscore the fact that the seven elements of the construct are not entities but categories which have internal structure. The ‘speaking’ names assigned to the seven functions are indeed names and not the formally identical words. It is an attempt at establishing something like a terminology but there are problems, of course, since the semantics of natural language lexemes is fluid: different aspects of their meaning potential get activated in different contexts of use, a phenomenon which proper terminologies seek to avoid. However, since we do not have any direct evidence of the cognitive processes that these words seek to name (at least for the time being), it is impossible to attach a proper term to an unequivocally defined object. This, nevertheless, should not jeopardise the usefulness of the construct as a heuristic tool. Table 1.1 List of CDF types and their basic underlying communicative intentions Function type CDF 1 CLASSIFY CDF 2 DEFINE CDF 3 DESCRIBE CDF 4 EVALUATE CDF 5 E XPLAIN CDF 6 E XPLORE CDF 7 REPORT

Communicative intention I tell you how we can cut up the world according to certain ideas I tell you about the extension of this object of specialist knowledge I tell you details of what can be seen (also metaphorically) I tell you what my position is vis-à-vis X I give you reasons for and tell you cause/s of X I tell you something that is potential I tell you about something external to our immediate context on which I have a legitimate knowledge claim

Cognitive Discourse Functions

33

Table 1.2 CDF types and their members CLASSIFY DEFINE DESCRIBE EVALUATE E XPLAIN E XPLORE REPORT

Classify, compare, contrast, match, structure, categorise, subsume Define, identify, characterise Describe, label, identify, name, specify Evaluate, judge, argue, justify, take a stance, critique, recommend, comment, reflect, appreciate Explain, reason, express cause/effect, draw conclusions, deduce Explore, hypothesise, speculate, predict, guess, estimate, simulate, take other perspectives Report, inform, recount, narrate, present, summarise, relate

As already mentioned, the seven types are to be understood as categories which subsume a range of CDF verbs. These are represented in Table 1.2; however, a discussion of their extension and mutual relations cannot be accomplished in this chapter and must await another occasion. Some categories are more populated than others: compare EVALUATE with DEFINE. Perhaps this is because the function types themselves are different in scope. Kidd (1996) pointed out the existence of micro- and macro-functions without, however, providing criteria for distinguishing them other than their length (number of words, clauses, turns). Equally diffuse are the links to notions like logical relations, logical patterns or rhetorical patterns (see Ehlich & Rehbein, 1986; Lemke, 1990; Trimble, 1985). In other words, the internal structure of the seven function types remains unexplored for the moment. The only thing which can be said with certainty is that this structure is complex. However, not only is the internal structure of the seven function types in the CDF construct complex, but also their borders are fuzzy. The seven types are not disjunct and not necessarily exclusive. On the contrary, they frequently include each other. DEFINE always contains CLASSIFY, but not all occurrences of CLASSIFY are part of DEFINE. DESCRIBE can be part of E XPLAIN, R EPORT or DEFINE, but there are also instances of DESCRIBE which stand alone. A first exploratory study, which employs the construct in analysing a set of six CLIL physics lessons (Kröss, 2014), has shown that the seven CDF types operate on two levels, forming episodes which are themselves composed of stages that are again realised by CDFs. These moves are often the same as the episode as a whole but may be interlaced with other functions.

Description of the Seven CDF Components In this section, the seven function types will be characterised in more detail accompanied by examples from naturalistic CLIL classroom

34

Part 1: Curriculum and Pedagogy Planning

discourse. The internal structure of each function type cannot, however, be addressed in detail. All examples stem from the data pool of the ConCLIL project. Project members from Spain, Austria and Finland have contributed a total of 41 lesson transcripts from their data collections of secondary CLIL classrooms, covering the subjects history, geography, science, accounting, business studies and tourism. The data analysis for the present chapter was exploratory rather than comprehensive; its interest was to identify examples from naturalistic classroom discourse that would illustrate the CDFs that comprise the construct introduced in this chapter.

Classify Given that the corresponding cognitive activity seems to be at the heart of knowledge construction in general, CLASSIFY is a key CDF. In the words of Anderson and Krathwohl (2001: 49) ‘each subject matter has a set of categories that are used to discover new elements as well as to deal with them once they are discovered’, thus pointing out the centrality of classification for knowledge creation. Complementary to this dynamic aspect of classifying, there is the static aspect of ‘knowing about the categories and classifications of a subject area’ which is laid down in the knowledge dimension of the taxonomy of thinking skills (Anderson et al., 2001). This is a more abstract knowledge type than mere knowledge of terms or facts and it is more complex because classifications actually form links or disjunctures between specific terms and facts. Other than in the case of observable knowledge elements that can be described, weighed and measured, classifications and categories are very much a matter of agreement and convenience. In fact, each categorisation moves us away from the specific and observable towards the more abstract: whereas knowledge of specific details stems from observation and discovery, knowledge of classifications and categories has to be learned from experts, which makes knowledge of classifications an important part of developing expertise. Classifying thus involves detecting relevant features or patterns that ‘fit’ both the specific instance and the concept or principle, and the activity thus begins with the specific instance or example and requires the student finding a general concept or principle to which it can be linked or subsumed. Mohan’s influential knowledge framework (1986) accords classification a central role, making it one of three elements in his framework: Classification – Principles – Evaluation. The papers written for the Council of Europe project for history, science and mathematics (Beacco, 2010; Linneweber-Lammerskitten, 2010; Vollmer, 2010) all

Cognitive Discourse Functions

35

mention classifying but differ in the granularity they accord to it. Beacco (2010; see also Achugar & Carpenter, 2012) regards it as one of the seven central operations for the subject of history, whereas Vollmer (2010; science) and Linneweber-Lammerskitten (2010; mathematics) consider it a micro-function. Trimble’s (1985) work is important as it offers a more in-depth treatment of the range of cognitive operations involved in enacting the process of classification as well as the linguistic steps required for its verbalisation.2 He observes that novices have difficulties both in the discipline and the language (Trimble, 1985: 20) as they may be uncertain about both the direction of the classifying process (members to class or class to members), and about the nature of the base of the classification (similarity or difference?). In a biology lesson on genetics, the teacher introduces the term ‘environmental variation’ linking it to lifestyle by way of the example of muscle growth caused by exercise. After that, the teacher sets a classification activity where the students are asked to group physical features of humans into three categories: inherited, environmental, combined. Example 1 CLASSIFY (Biology, Grade 9; age 15) T:

[longish introduction of the activity] You’re supposed to fill in this table, right? With the features underneath. So […] look at this, this is what, this is exactly what you have ok? […] What are we supposed to be doing?

S:

Filling in the ((x....x )) using the corresponding

T:

Think a little bit, think a little bit about the features that we´ve got

T:

And this, these are the features, can you see the word? You can see the list below in your table, so all features that we have, [ … ] So, we´ve got the hair colour, we´ve got… the eye colour,… ..and then there are others being weight, height, [….] Ok, so this is to be filled in with features that are inherited. This is to be filled in with features that are environmental. And the last column?

S:

Features inherited and environmental

T:

So both, ok? Now, have a go. Have a go. [students work in groups to fill in the table]

In the extract, reference is made to the worksheet featuring three columns. The interaction shows how the teacher seeks to communicate that these represent the criteria by which students are to sort features of human beings.

36

Part 1: Curriculum and Pedagogy Planning

Define Like classifying, defining is a core activity in organised knowledge creation, as all academic disciplines require definition for the proper identification of their subject in order to determine what is and is not part of the field and also how field-specific knowledge objects are circumscribed and related to each other. Definition always expresses some kind of class membership and is thus related to classification. Accordingly, defining is mentioned as central by the authors of the history and science papers of the Council of Europe project (Beacco, 2010; Vollmer, 2010). Definitions are perhaps the best-described academic function, presumably because they have played a significant role as tools in the study of cognitive development (e.g. Snow, 1987), and have been recognised as highly relevant for academic writing (e.g. Swales, 1990; Trimble, 1985). Not least, definitions are a fairly compact affair which makes them easier to capture than other, more loosely structured CDFs.3 Research on definitions concurs in ascribing to them the following typical features: they consist of something to be defined (the definiendum) which is linked to a superordinate term or class (the definiens), followed by the specific differences that hold for the definiendum. Thus: Definiendum= Definiens+differences, or Species=Genus+Differentia. The linguistic realisation of this simple formula is not really highly complex: it requires a copula construction (an X is a Y) with the specifying features (differentia) being realised through adjectives, relative clauses or reduced relatives. The specifying features can be descriptive, comparative, functional, historical or any combination of these, as illustrated in the following two examples. (1) A spider is an eight-legged predatory arachnid with an unsegmented body that injects poison into its prey. (2) A spider is a program that visits Websites and reads their pages in order to create entries for a search engine; also called crawler. While academic writing definitions ‘come in all sizes from single words to entire books’ (Trimble, 1985: 75), spoken interaction is likely to rely on less extensive forms. Even so, a great deal of information is to be drawn from such compact, one-sentence statements like those exemplified here. In classroom interaction we are therefore likely to find less compact realisations (semi- to non-formal in Trimble’s [1985: 75–84] terms). Genus may be left out (because it is obvious from the context) or may be formulated as an additional difference (‘an arachnid is a spider’). Antonyms (‘indigenous is the opposite of foreign’) or synonyms (‘crawler’) may be used. A genus may be represented by a well-known member (‘an arachnid is a spider’, while actually it is vice versa) or by its most

Cognitive Discourse Functions

37

common characteristic (a spider makes webs). The following examples from the ConCLIL corpus give evidence of formal and semi-formal monologic definitions in CLIL classroom discourse. Example 2 Formal definitions in the CLIL corpus (2a) a kidney is an internal organ that purifies the liquids inside you,.. (2b) witnesses are people who can say... who can say i’ve seen it, i can swear that this is the truth (2c) a high involvement decision is a decision where a lot of money or a lot of time is necessary to just say yes or no Example 3 Semi-formal definitions in the CLIL corpus a structured question is a (xx) question and er there is a limited number of possibilities to answer a high involvement decision is for example if you buy a car you will look for a lot of different offers and you won’t buy the very first one It is probably no coincidence that Examples 2a–c were uttered by teachers, whereas those in Example 3 were uttered by students, thus giving insight into the differences between expert and novice users, even though the expert users show no signs of explicit awareness of what is going on. A much more frequent way of defining in classrooms is, however, to share out the task among several interlocutors, as shown in Example 4. Example 4 Interactive defining in CLIL physics (Grade 6; age 12): S1: (reading aloud): a hydraulic jack can be used to lift a heavy vehicle. The jack changes a small force into a much bigger one. It is therefore called a force multiplier. T: what’s the multiplier? …S2! S2: ein vermehrer! (E a multiplier) T: yes, how could you explain it in English? S2: ahm a thing that ahm make something stronger oder or more. T: or bigger. so this thing is used ah if you only have small force if you exert a small force to change that force into a much bigger one. An interesting and highly characteristic detail is the teacher’s use of explain in Turn 4 of the above example. What she is actually asking S2 to do is to define the notion of ‘force multiplier’ – it is not the only case in the data where the verb explain seems to be used like a hyperonym for other cognitive discourse verbs and acts like a dummy CDF as it were. This phenomenon can lead to confusion with regard to the proper extension of the CDFs explain and describe and will be discussed under those headings below.

38

Part 1: Curriculum and Pedagogy Planning

Describe Describing is an activity where a speaker/writer informs a listener/ reader about the observable features, qualities or external and sometimes also internal characteristics of something in third-person position. This something can be a given object, entity, person, situation, event or process. The Oxford English Dictionary (OED) renders this ‘ordinary current sense’ of describe as ‘to set forth in words, written or spoken, by reference to qualities, recognizable features, or characteristic marks; to give a detailed or graphic account of’ (OED, s.v. describe). In terms of what a speaker tries to ‘do’ interactionally when giving a description, describing does not seem to imply a comprehension difficulty on the part of the addressee. Rather, describing is a way of informing and thus of ‘telling you what I know’ (Gaulmyn, 1986: 125) or, more precisely, ‘telling you what I see’ (Lackner, 2012: 49). Given that attention to the not immediately obvious must be regarded as a crucial feature of specialist knowledge construction, it is not surprising that describing is invariably mentioned as a key element in academic thinking skills and academic language (see Beacco et al., 2010; Laplante, 2000; Mohan, 1986; Trimble, 1985; Vollmer & Thürmann, 2010; Widdowson, 1979, 1983; Zydatiß, 2010; among others). The fact that describing is of high relevance to a large range of disciplines from literature to chemistry to mathematics means that descriptions can take different forms in different contexts: describing birds is not the same as describing rocks or a chemical reaction (Laplante, 2000: 269). And while describing has been variously characterised as ‘distinct and clearly separable and identifiable’ and ‘intimately related to certain lexical and grammatical elements’ (Trimble, 1985: 69), it is actually mostly in science and technology contexts that a closer examination of descriptions has been conducted to identify their generic features. This has resulted in the differentiation into three types of description (Gillett et al., 2009; Trimble, 1985; Widdowson, 1979): (a) physical description: material and outward characteristics; (b) functional description: purpose of a device or institution and how their parts cohere; (c) process description: series of steps, procedures and their purpose. In general, the activity of describing involves reference to shape, size, weight, colour, texture, position in space, etc. Descriptions are typically timeless and generic, omitting reference to self and addressee and focusing attention on an object presented as universally the case (Schleppegrell, 1998: 187). Furthermore, functional and process descriptions require the author to state relations between parts and/or a ‘time order frame

Cognitive Discourse Functions

39

structure’ (Widdowson, 1983: 59). Stating such relations moves functional descriptions into the vicinity of explanations, as has been pointed out by Ehlich and Rehbein (1984: 87) among others. Thus, descriptions are frequently part of explanations, but may also be part of definitions (see above). Two exploratory studies of CDFs in history and physics lessons (Kröss, 2014; Lackner, 2012) have shown DESCRIBE to be the most frequent function in both subjects, the typical situation being that all participants orient towards an image or artefact that is physically present in the classroom and visible to everyone, verbalising what can be seen.4 Example 5 shows a typical whole-class situation. Example 5 Physical description (History, Grade 10; age 16) T:

in the picture ah did you find the sacred olive tree in the picture?

Sascha: the sacred olive tree? Daniel: hm? .. wo? S:

mm ich hab’s gleich

T:

it’s in the holy district, right in the middle, right outside the temple of Zeus

Sascha: Well Daniel: the temple of .. yeah? S:

links unten [bottom left]

T:

you find the temple of Zeus?

Daniel: Yes. in the middle of the centre T:

it’s next to this round building and just in front of it there is the sacred olive tree

What is required linguistically here, namely locative prepositional phrases, has been known to these Grade 10 students for several years, yet they hesitate, with one of them using L1 German ‘links unten’ (bottom left) and the teacher modelling the verbalisation in two of her turns. Example 6 takes us into a group-work situation where participants collaborate in describing and analysing a number of pictorial sources. Example 6 Physical description (History, Grade 9; age 15) T:

And what is he doing?

Sara:

He’s tired, yeah, he probably has a %x…x% thirteen hours or so

40

Part 1: Curriculum and Pedagogy Planning

S1:

It seems to be…un-tidy and sad

Sergio: Source B, source B, it’s a %x…x% but it’s a picture that the same factory wants that people Sara:

It shows that it is a lot easier to carry on with the work

Sergio: Yeah, and there are children… The way in which students in Example 6 do not so much verbalise actual observables depicted in the sources (‘untidy’) but move straight into interpretation (‘probably been working for 13 hours or so, sad, it is a lot easier..’) can be observed on other occasions in the ConCLIL data and might be typical of unguided description. In contrast to physical descriptions, structural and functional ones proceed much more often without visual support (especially in history) even though they are based on the notions that ‘parts form a whole’ and ‘a whole consists of parts’ (Lackner, 2012), which in principle would be amenable to visualisation. The context of Example 7 is a revision task referring to a text in the course book: Example 7 Structural description (History, Grade 10; age 16) T:

there’s some work for you. .. the government of the republic.

Ss:

(XXX)

T:

we have a senate, .. we have an assembly, .. and .. we have ... officials. ... ahm ..... the senators. the senators are .. people who ah belong to the group of patricians. two main social groups. the rich people, the nobility, the nobles, are the patricians. it says so on the right-hand side. can you find that? ... patrizier

Example 7 is also typical in another way: it shows very informal lexical choices on the part of the teacher in rendering the structure of government in the Roman Republic (‘we have…’). The following example demonstrates that process descriptions cannot only be found in the sciences but also in history; Example 8 comes from a lesson on Ancient Egypt. Example 8 Process description (History, Grade 10; age 16) S1 (I):

ahm .. they washed the body first

T:

yeah, they are preparing a mummy. they wash the body first, yeah yeah

S1 (II):

and .. they took out ah .. the brain and the insides

S1 (III): but they left the heart .. in the body

Cognitive Discourse Functions

T:

And then?

S1:

ahm they stopped .. ahm

41

In the (incomplete) three-step process description rendered in Example 8, we can observe features of informal oral language (‘they washed’) where canonical process descriptions would require the use of passive. There is also a subtle tension between the students’ rendering of the process as a kind of narrative in the past tense and the teacher using the present tense (once referring to an image on the overhead projector in the present progressive and then rephrasing it in generic present simple). The implications of these subtle linguistic choices and the genre conventions of process description, however, remain completely out of purview. Their performance is neither required nor modelled by the teacher via paraphrase or recast.

Evaluate Unlike the CDF types described up to here, EVALUATE has not been treated extensively in the literature on languages for specific purposes. Based on the definitions of the OED, the core meaning of the verb evaluate is ‘determining the value or estimating the force of something in terms of something already known’. The family of verbs thus subsumed in the EVALUATE basket includes: appraise, argue, assess, bring evidence, check, critique, contend, corroborate, debate, defend, evaluate, judge, justify, take a stance, refute, raise objections. In terms of speech-act like communicative intentions the underlying common denominator for this CDF type could be formulated as: ‘I am going to tell you my personal stance towards this. I have reasons for this position based on evidence, my previous knowledge and values’. Possibly even more so than with other function types, translingual confusions are to be reckoned with when one combines curricular analyses made in several languages into a global discussion. For example: German interpretieren, especially in its nominalisation Interpretation carries a strong evaluative element, which English interpret does not. Thus, German interpretation in literature studies is equivalent to English literary response. A similar situation prevails with German argumentieren vs. English argue as in ‘to argue a point’. German argumentieren implies the bringing forward of reasons or evidence in order to affirm one’s position and a sense of having to convince and win over an opposition, thus argumentieren in a classroom context would mean conducting something like a debate where different, usually contrary, positions have to be argued. Despite such complicating factors, the academic language functions and thinking skills literature seems to be in agreement over the importance of EVALUATE (Bailey & Butler, 2003; Beacco, 2010; Bloom, 1956;

42

Part 1: Curriculum and Pedagogy Planning

Chamot & O’Malley, 1994; Krathwohl, 2002; Mohan, 1986; Thürmann et al., 2010; Vollmer, 2011; Vollmer & Thürmann, 2010; but not Biggs & Tang, 2011). Authors who grade functions according to their complexity place evaluating near or at the complex end of the scale (e.g. Bloom, 1956; Krathwohl, 2002; also Walker, 2010a). 5 The ‘value’ of whatever is being evaluated can be quantitative (and thus expressible numerically) or qualitative (including moral judgements)6 and must be linked with some appropriate frame of reference. My working definition of EVALUATE thus also subsumes what in rational academic procedures/discourse must be part of an evaluation and is required to be explicitly expressed, namely the evidence, criteria, standards or reasons which support the evaluation that is being made. Being socialised into rational academic discourse means being socialised into taking a stance, justifying decisions and arguing one’s opinions, or, this is what education is supposed to socialise students into (see also Anderson et al., 2001: 83). Example 9 shows how a teacher demonstrates EVALUATE at the end of a sequence where the class has been working on inherited and environmental traits in humans. The teacher had illustrated this with the story of the son of Nobel prizewinner John Nash, who grew up without contact with his father, but shares his mathematical talent. The teacher calls this her ‘opinion’ but she does give arguments (‘it has a lot of support behind’). Example 9 EVALUATE (Biology, Grade 9; age 15) T:

So you think it´s both? What do you think? Do you want my opinion? Just to finish this up. Well, I do think it´s both. I do think it´s both. All right? And it´s not me that thinks it´s both, I mean, it has a lot of support behind. There´s a lot of studies that support the idea that most of the traits that we inherited, we inherited ability to do something, being good at something, sports, music, certain types of intelligence, right? (( intelligence, writing intelligence, writing skills, geometry and math skills, that kind of stuff, it´s a little bit conditioned by genes, but, however, it´s got, it´s got some environmental influence also. So, maybe the more you practice something, despite your genes, the more you practice something the better you can get at that something. Do you follow?

Example 10 is particularly interesting because it is one of the rare instances where a student contradicts the teacher’s verdict and supports her position with an argument, which is then also accepted by the teacher. The class had been discussing the social position of women in ancient Athens and a student had asked if women were allowed into theatres.

Cognitive Discourse Functions

43

Example 10 EVALUATE (History, Grade 10; age 16) T:

they could go and see plays on a special occasion. if the husband went with them, .. and so on, yeah? but they could not there were a number of jobs that they .. couldn’t do.

S1:

mechaniker?

T:

no no [SS laughing] not mechanics, they didn’t have mechanics at that time

S1:

ooh, yes of course they had they had mechanics, yes

S2:

die hatten streitwagenrennen .. und die (XXX) [they had chariot races ..and the ]

T:

die hatten .. ja streitwagen, ja, okay.[they had chariots, yes, okay] .. and they needed repair, yeah, you had mechanics, you can call them mechanics, but different mechanics because they were different cars.. all right.

Given the apparent curricular priority of EVALUATE, claims have been made about where in classroom talk EVALUATE might be expected to occur and it has been suggested that this would be the case particularly in dialogic exploratory and procedural talk (Beacco et al., 2010). Beacco et al., however, argue on a purely conceptual level, so their claims are programmatic rather than evidence based. Kröss’s (2014) study of CLIL physics lessons has shown EVALUATE to be extremely rare and also Wells’s (2009) reckoning, based on long-standing classroom research experience, provides a sobering perspective regarding the frequency and even likelihood of EVALUATE occurring in classrooms. Wells states that it is one of the paradoxes of schooling that built in inequality actually discourages stance-taking. And, once this pattern of inequality has been established as the norm in school, older children accept and even collude with it, becoming unwilling either to ask the sort of questions that might lead to a genuine instructional conversation or to go beyond giving minimal answers, even when a teacher’s question calls for an expression of their own opinions or an account of their personal experience. (Wells 2009: 4) However, educational cultures might differ in this respect as indicated by Grundler’s (2010) affirmation that argumentieren is a highly valued activity in German social science classrooms. In any case, the invitation is out for an empirical fact-finding mission across a range of subjects and educational contexts.

44

Part 1: Curriculum and Pedagogy Planning

Explain Of the function types in the construct, EXPLAIN is probably the most complex and certainly the one with the largest extension, which makes it important to contain it for our purposes. There is actually quite a bit of literature on explaining (e.g. Stein & Kucan, 2010), something which is not the case with other elements in the CDF construct. This is both helpful but also complicated because the gaps and inconsistencies in the notion are much more visible than elsewhere. In my present understanding, the word explain is used in three main ways (see OED s.v. explain v.) Explain 1: To make sth. plain or intelligible; to clear of obscurity or difficulty; to give details of or to unfold (a matter). Explain 2: To give an account of one’s intentions or motives. Explain 3: To make clear the cause, origin, or reason of. Explain 1 is a general vocabulary item and as such is closely linked to exposition. In fact, they may be synonyms where one is used more readily in formal and written contexts (exposition), the other more in oral and informal contexts (explain). ‘Instructional explanations’ as conceived by Leinhardt (2001) and ‘interactive explaining’ as introduced by Smit (2010) are linked to this understanding of explain. I exclude Explain 1 from the CDF construct, since it is too comprehensive to be of use as a specific function type. Rather, all the CDF functions in their totality could be said to contribute to ‘explaining’ in the sense of Explain 1. This said, delimitations between Explain 1, 2 and 3 are by no means categorical. This becomes evident as soon as we inspect the meaning of Explain 2: the ‘details’ of Explain 1 are now specified as intentions, reasons and motives and so implicate some kind of causality (‘X happened because Y was drunk’). Apparently, humans are strongly inclined to think of actions and events as being caused by the states and behaviours of either themselves or others (cf. Kahnemann, 2011). Ascribing the causes of events to human agents – it becomes immediately obvious – is a function that will be in demand in a subject like history. In fact, history seeks to ‘package up sequences of actions by individuals into episodes’ (Martin et al., 2010: 441) or, one might like to add, organisations, institutions and systems (e.g. the union, the monarchy, state capitalism). Such episodes are then represented as interrelated causes or consequences and because historical events usually have multiple causes, their explanations must necessarily be probabilistic and/or functional. The hard sciences, in comparison, given their different epistemology, operate on the principle that they provide exclusively deductive explanations of phenomena (i.e. Explain 3). This is something which

Cognitive Discourse Functions

45

does not necessarily stand the test at all times, but functions as a strong identity-forming principle for the sciences and is regarded as a central feature of scientific literacy (Laplante, 2000; Lemke, 1989; LinneweberLammerskitten, 2010; O’Brien et al., 1995; Vollmer, 2010). Many scientists would probably argue that such deductive (or nomothetic) explanations are the only ‘real causal explanations’. For a broader perspective on educational data (and thus within the present construct) this would be too limiting an understanding: explanations are always about answering the question ‘why?’ and are thus always causal in this sense. The understanding of EXPLAIN adopted for the CDF construct thus comprises Explain 2 and Explain 3 and is compatible with the respective definition in the taxonomy of thinking skills (Anderson et al., 2001). In sum, then, the basic communicative intention of this category is ‘I will give you reasons and tell you the cause/s of X’. Example 11 illustrates Explain 3. Example 11 EXPLAIN (Biology, Grade 10; age 16) T:

why do we have that (.) trait dark eyes if the genotype is heterozygous why

S:

ah because the dominant. the capital a is dominant but it is dark the

T:

yeah

S:

it reach the zygote

T:

exactly so it dominates or it hides hides enmascara esconde it hides I guess hm >you can say that< it hides the other one. so that’s basic genetics for a simple trait or characteristic that you can find in ↑people such as the colour of your eyes↓ (0.3) okay (.) good

Here, the learner’s attempt at an explanation is encouraged by the teacher (‘yeah’) but needs to be rephrased eventually as it is incomprehensible to anyone without a ready-made explanation in their minds to fill in from. The next example from a history lesson illustrates Explain 2. The class has been looking at an illustration of an upper-class family home in Ancient Athens; the current focus is on a room where a party is going on. Example 12 EXPLAIN (History, Grade 11; age 17) T:

…female slaves, ah there’s a boy here, the boy is allowed to take part in ah men’s parties, why?

S1:

he is male

S2:

they were good-looking

46

Part 1: Curriculum and Pedagogy Planning

T:

they were good- the boys? they were good-looking

S3:

they are learning how to be a m-

T:

yes

S3:

they learn how to be a man

T:

yes .. exactly they must learn ((Ss laughing))

S:

to be a man

Three students venture an explanation, with the teacher simply echoing their contributions, until the third and most complex explanation prompts her positive feedback ‘Yes’ and thus gets the student to repeat her insight, which is at this stage shifted linguistically from a descriptive progressive (‘are learning’) to a generalisation in present simple and then echoed once more by the teacher.

Explore EXPLORE has been chosen as the label for a family of near synonyms, which appear in each other’s definitions; these are assume, suppose, presume, conjecture; somewhat more distant semantically but still related are the verbs predict (‘to foretell on the basis of reasoning or knowledge’), guess (‘to form an approximate judgement without actual measurement’) and speculate. The more technical hypothesise with its dictionary meaning (OED) ‘to state a proposition merely as a basis for reasoning or argument, that is, without conclusive evidence’ is of course another member of the set. The semantic trait which appears to unite all these words and their uses, and which is the common denominator in the CDF construct, is that they all designate instances of ‘talking about that which is not in the here and now, and which is not firmly established past fact either’. Formulated dictionary-style, the meaning of this function type could be rendered as ‘to state a proposition without conclusive evidence (usually in the interest of laying a basis for further reasoning or argument)’. This core meaning translates into the following communicative intention which underlies all realisations of this CDF type: ‘I’m talking about something which is not in the here and now, and which is not past fact either. I do not have conclusive evidence for what I say but it can serve me/us as a basis for further reasoning’. It should be clear that in this context I am not concerned with notions of hypothesis and prediction made on the basis of a specific scientific theory as part of research processes, but with a more general, semi-expert notion of this activity (cf. Dalton-Puffer, 2007a: 138). In these non-technical terms, EXPLORE means making an assumption or prediction about what something

Cognitive Discourse Functions

47

will be like or would be like if certain conditions are met. EXPLORE therefore amounts to talking about that which is not the case now, talking about that which could be in the future or that which could have been in the past, but with a link to perceived factual reality. Hypothesising thus is a prime example of ‘talking about that which is not in the here and now’. E XPLORE in the wider sense of the function type is clearly regarded as an integral part of knowledge building in educational contexts, as shown by the fact that overviews of thinking skills, curricular goals and academic language consistently mention hypothesising and predicting among the core (e.g. Beacco et al., 2011; Kidd, 1996; Linneweber-Lammerskitten, 2010; Vollmer, 2010; Wells, 2009). If a ranking is made, E XPLORE , resp. hypothesise, is accorded a position among the highest-order functions (e.g. Anderson et al., 2001; Biggs & Tang, 2007: 91). E XPLORE is also closely connected to the notion of exploratory talk which has been given much credit in terms of furthering student understanding (e.g. Barnes, 1992; Mercer & Dawes, 2008; Moate, 2011). For its verbalisation, EXPLORE requires the use of relatively complex lexico-grammar (see Dalton-Puffer, 2007a: 138), especially modal verbs (e.g. may, will, can), modal adverbs (e.g. maybe, perhaps, possibly) and dependent clauses with conditional conjunctions (mostly if ). In the following extract, the participants use a conditional clause but transpose their explorations into present simple, as if they were treating these historical deliberations as general truths. Example 13 EXPLORE (History, Grade 10; age 16) S19: but if the husband dies? S:

gibt eh genug husbands [there are enough husbands]

T:

if the hu- if the husband dies? .. if the husband dies?

S20: she has to marry again S21: she has to marry again T:

the woman must marry again

S20: as quickly as possible S21: sag ich ja [that’s what I’m saying] T:

or the woman must go back to her father. a widow must go back to her father. because ... an Athenian woman was never allowed .. to exist without .. a man and the man was either her father or a brother or a husband

SS:

tsktsk - < G super> [brilliant]

48

Part 1: Curriculum and Pedagogy Planning

Even though it is generally difficult to find CDF episodes realised independently by students, EXPLORE seems to be among the ones where learners are particularly reluctant ‘to go at it alone’. EXPLORE is realised always either by the teacher or co-constructed between teacher and students. Example 14 illustrates the latter: Example 14 EXPLORE (Physics, Grade 11; age 17) 1 S9:

isn’t the triangle the strongest shape xxxx

2 T:

I think probably a triangle is a good shape for it. can you imagine what would happen if it actually had FOUR legs. (.) and I pushed it from the side. you know I push from the s- it’s got FOUR legs and I go (.) what happens. yeah.

3 SX-f: it would ALSO fall down. 4 T:

it would be fallin- MUCH more likely to fall down? {moves camera} I could tip it over? okay? but it’s gonna tip over here? (1) and return.

5 S3:

interesting.=

6 T:

=okay. so it MIGHT be that this idea of having a (.) triangle (.) it might be a particularly useful shape. (.) and this (.) next lesson and today’s lesson is about stability. (.) why things are stable why things stay (.) up right. yeah.

7 S10:

isn’t it maybe because the point where all the pressure is on a triangle is exactly in the middle of it? (1) because if it has four legs that doesn’t have to be? (.) but if it’s THREE (.) then it’s HAS to be in the middle.

We can see the teacher using different modal verbs but only one of the students ventures a would (Turn 3). In fact, in this particular group of learners, negative questions (‘isn’t it…’) seem to be employed as an exploring-strategy. In Turn 2, one can detect a further typical feature of EXPLORE, in the phrase ‘can you imagine what would happen if’. Thus, teachers commonly invite students to join them in the activity of EXPLORING either by questions such as in the example just cited or by exhortations, such as ‘Let’s imagine’. Given the linguistic complexity of canonical realisations of EXPLORE, it may well be that this actually prevents CLIL learners (and possibly also some teachers) from using this CDF type to a greater extent. Kröss’s (2014: 16) exploratory study of six upper secondary CLIL physics lessons at least has shown EXPLORE to be among the lesser used CDFs (undercut only by EVALUATE and CLASSIFY). This is certainly a particularly intriguing case where empirical research into expert and novice realisations of a function

Cognitive Discourse Functions

49

type is necessary in order to better understand the specific challenges faced by CLIL learners and teachers working in their L2.

Report This CDF type is about ‘what happened, when, who did it and to whom and under what circumstances’. Subsumed under this heading thus are a range of near synonyms, such as report, recount, relate, narrate, present, summarise, give an account of. As one looks at their dictionary entries (OED; Merriam Webster), they form a ‘mutual citation circle’ as they appear in each other’s definitions. We are thus justified in assuming that differences in meaning lie not in their semantic core but in conditions of use, register and/or connotation. The core that all these designations share is that they: • • •

Share the illocutionary function ‘inform’ (cf. Widdowson, 1983). Assume a reduced shared background knowledge of speaker and recipient. Foreground the referential function of language.

Their communicative intention in the CDF construct is thus ‘I tell you about something external to our immediate context on which I have a legitimate knowledge claim’. Some accounts of academic language mention summarising as a separate activity (Anderson et al., 2001; Thürmann & Vollmer, 2010), but I would claim that summarising (in the sense of a cognitive process for selecting only the most important crucial information) is part of the activities that form this particular complex of discourse functions. Schleppegrell (1998, 2001) mentions ‘summary’ as a typical element in text-types that are frequently required of undergraduate science students, but also in schooling in general. All accounts of academic discourse functions consulted featured reporting and presenting and most also included narrating. The nominal uses of some of these terms actually refer to practices which are firmly established as oral or written genres or both (e.g. historical recount, oral presentation, Erlebnisaufsatz, story) and so all of them are what Zydatiß (2010: 135) calls ‘unterrichtsbezogene diskursive Darstellungsverfahren’ [classroomoriented discursive modes of representation]. A considerable amount of research-based evidence has been assembled on certain well-established written genres, for example the lab report, which is regularly mentioned, especially in the literature on science education (Laplante, 2000; Thürmann & Vollmer, 2010). In history, the historical recount has been established as a specific genre in school history (Beacco, 2010; Heath, 1983; Schleppegrell, 2001; Veel & Coffin, 1996).

50

Part 1: Curriculum and Pedagogy Planning

Recounts are also considered a discernible feature in mathematics, where they are used for reporting on how the solution to a problem was reached (Linneweber-Lammerskitten, 2010). In (L1) literature education, R EPORT is exemplified by the typical activity of students having to render the main parts of a literary story (plot summary). With regard to oral classroom language, the evidence base is more scanty: reporting has been identified as a typical final step in a pedagogical task sequence when a student ‘reporter’ is designated to feed the results of group work back into the learning community (e.g. Bunch, 2006). The following example illustrates such a case where R EPORT is performed by a student who informs the larger learning group of the deliberations of her group concerning a certain task. Example 15 R EPORT (Business Studies, Grade 11; age 17) 01

S:

we looked if the contract of sale was valid. yes, it was valid, aahm there was no binding offer but an order from the buyer and an acknowledgement of this order from the seller. what we can do from the legal point of view: we decided that it is a fixed delivery date .. aahmm because .. i got .. a confirmation where Viking promised to deliver on thirtieth December. our legal possibilities are .. ahm cancellation of the contract without giving any other information, …[continues]

Another variant of Report in the oral mode are extended oral presentations, a monologic student genre hugely popular in Austrian classrooms (cf. also Bunch, 2009; Thürman & Vollmer, 2010). Short teacher monologues expanding on some theme are also examples of the CDF-type R EPORT. Example 16 R EPORT (Physics, Grade 10; age 16) 01

T:

so this is even three thousand years ago. (1) yah? (.) a:nd uhm even TWO thousand years ago? (1) Chinese sailors knew about magnetic (forces). (1) yah? psht. {smiles} using the uh MAGNETITE STONE. they real- the already knew about magnetic forces. (1) yeah? and why would sailors need magnetic forces? (.) or how would they USE magnetic forces?

02

S4:

to know where they are? where north is? (.) and (all that)?=

03

T:

=using what?

04

S4:

a compass?=

Cognitive Discourse Functions

05

T:

51

=using a compass. (exactly.) so we’ve been using compasses to tell where we ARE? (1) using MAGNETIC properties fo:r a REALLY long time? (.) already? (1) yah? so you use compass needles for navigation? to tell you WHERE you’re going? and which diRECtion that you wanna go to. (.) yah? which direction you wanna go IN. (2) yah? so EVEN (though) THOUSANDS AND THOUSANDS of years ago we knew about electric forces with stones like amber? {points at slide} (.) and we knew about magNETic forces with stones like magnetite? (1) we didn’t we couldn’t really EXPLAIN it. that long ago. we weren’t advanced enough to explain WHY HOW things happened (.) we just knew that they HAppened. (1) yah? that we COULD use magnetic forces to:: get ourselves from place to place. you know as sailors.

Teacher monologues like this are, however, not as frequent as is sometimes implied by commentators on teacher language. A very interesting aspect of this function-type is the relation between reporting and narrating (G berichten und erzählen) discussed in depth in a series of German articles from the 1980s (Ehlich, 1984; Rehbein, 1984).7 The upshot of the discussion is that a story is basically the reproduction of a report in a fictional space: real-life events are turned into stories with a plot and characters (usually with ‘me’ as the main character, or some third person to whom the teller of the story is close enough to render his/ her ‘inner’ view of events). This differentiation becomes rather challenging in historical narration (see Lorenzo & Dalton-Puffer, this volume), where the teller is supposed to take an external viewpoint, but experiences the pull of the personal narrative with its key role of characters. The difficulty in withstanding the pull of ‘narrative’ in oral classroom practice is also underscored by O’Brien et al. (1995: 449) who argue that chemistry learners need encouragement to give causal explanations rather than narrative accounts.

Conclusion The transdisciplinary construct of CDFs presented and illustrated in this chapter draws its conceptual foundations from both education and linguistics, and so occupies a space where the conceptual orientations of content-subjects and language education intersect. Its function is that of a heuristic and it is hoped that reference to the CDF construct will enable content-subject educators and language-subject educators to find a common language in talking about and researching student-learning in CLIL.

52

Part 1: Curriculum and Pedagogy Planning

The CDFs which make up the construct represent that element of the educational process where concepts, facts, notions, processes and cognitive operations of the subjects appear as analogues of verbalisations and vice versa. Such verbalisations naturally also encompass the written mode and the constitutive role of writing for which organised knowledge creation goes unquestioned, but, as has been evident from the examples given, my discussion and exemplification of the CDF construct concentrates on the oral mode for now. Certainly at the level of school education it is crucially the event of the lesson and face-to-face classroom interaction where knowledge construction takes place. It is the role of teachers to make the learning objects intersubjectively accessible to all members of the learning group by mediating and interpreting between expert knowledge and the knowledge states of group members. ‘Common knowledge’ (Edwards & Mercer, 1987) thus comes into being and the seven types of CDFs stand for major communicative avenues in this process. What needs to be achieved in the learning event is bidirectional: learners must be enabled to access the knowledge but they also must be enabled to then express what they have appropriated (formal exams being the most poignant occasion where this is necessary). That is to say, students need knowledge items to be accessible but also expressible. As long as these communicative avenues are given linguistic expression in everyone’s first language, the linguistic form in which these processes take shape is usually taken as unproblematic and often goes unnoticed. Even though considerable gaps exist between everyday conversation and subject-oriented conversation in levels of formality, registers, vocabularies and the like, it is taken for granted that all participants ‘know the language’ and are therefore capable of giving adequate expression to their momentary level of understanding. In other words, if someone cannot say ‘it’, they ‘don’t know it’ rather than ‘they don’t know the right words for it’. In the case of CLIL, however, this is interestingly different, as teachers and learners frequently share a status of being L2 users of the medium of instruction. Thus, being in the possession of ‘the right words’ is not taken for granted, though mostly it is only lexical items which receive conscious attention and remedy. But in order to enable learners to expressly handle knowledge items and thus truly appropriate them, lexical items alone do not suffice; expressing cognitive processes and thus CDFs are the other main pillar. Learners need to have the linguistic means to define, classify, report and evaluate etc. in their L2. These means may be picked up implicitly from their teachers (and/or by reading), but for this to happen the teachers need to be able to provide rich and repeated modelling. It may, however, also be necessary to explicitly teach aspects of CDFs, just as technical terms are taught explicitly. Matters are further complicated by the fact that an inherent tension exists in school lessons between there being oral interaction between familiars where informal language choices

Cognitive Discourse Functions

53

and high-context embeddedness prevail and there being the place where learners (ought to) get access to more formal and specialist registers. This tension is clearly exacerbated in CLIL where teachers themselves will frequently have a less secure grasp of register differences in their L2, will be less aware of necessary register switches and may thus be less reliable models for their students. The illustrative examples of CDF realisations included in this chapter show the limitations not only of students’ but also at times teachers’ linguistic means of expression in their L2; limitations which may severely limit the inclusion of crucial functions like EXPLORE in regular knowledge construction talk and thus have the potential to curb important steps in student cognitive development. So far, the evidence about how CDFs are realised by classroom participants is still illustrative, showing a good deal of implicit knowledge on the part of teachers and both limited resources and limited opportunities on the part of students. It is clear that a major effort of empirical research into CLIL classrooms in different contexts has to be undertaken to determine how participants go about constructing CDFs, and to determine how they can best be supported in doing so. This, however, also implies assembling inventories of linguistic resources that are typical of individual CDFs. A good deal of work in this direction is available from earlier work in English for Specific and Academic Purposes (ESP and EAP) but needs to be reassembled and supplemented in alignment with the CDF construct. Having made reference to ESP and EAP, a further important question needs to be addressed which is that of cultural plurality. CLIL is an international phenomenon, embedded in numerous education systems and educational cultures so that mere reference to native speaker models as has been customary in ESP and EAP must fall short of the complexity that is at issue. The CDF construct with its seven types, whose internal structure is that of family resemblance, has the potential to accommodate different cultural models tied to the basic function inscribed in each component. So, while underlying cognitive processes and communicative intentions may be very general, they are likely acted out within the local cultures of subjects and education systems. These cultural models may thus be subject-specific or lingua-cultural as school level education does not necessarily exhibit the degree of internationalisation that a label like CLIL might suggest. In sum, I hope that the CDF construct will be useful as a heuristic which is accessible to both applied linguists and subject education specialists and can thus be applied in jointly analysing (and possibly also planning) CLIL classroom interaction. So far, international CLIL research has been conducted mostly by applied linguists, but to truly do justice to CLIL, cooperation needs to be sought from subject education specialists. The CDF construct can serve as a common frame of reference when educators and researchers from each quarter work together at better understanding

54

Part 1: Curriculum and Pedagogy Planning

CLIL. This, it is hoped, will enlighten us further about how language for learning is enacted, how new ways of knowing and thinking are made accessible in classroom talk and how cognitive development is scaffolded, when classroom teaching and learning happens in an L2.

Notes (1)

Language(s) in other subjects. See http://www.coe.int/t/dg4/linguistic/langeduc/ boxd2-othersub_EN.asp (accessed 31 December 2013). (2) A table with notional language used in classification is also provided by Mohan (1986: 79, figure 4.6). (3) For an empirical study of definitions in CLIL see Dalton-Puffer (2007a, 2007b). (4) The extent to which such episodes of Describe are instances where teachers seek to instil ‘professional vision’ (Goodwin, 1994) into students raises intriguing questions but cannot be discussed in the context of this chapter. (5) Mohan (1986: 91) also formulates a staged view of evaluation over different competence levels and provides a table with notional language for doing evaluation in English (Mohan, 1986: 86). Appraisal theory (Martin & White, 2005) has since provided a comprehensive treatment of all related aspects of the English language. (6) See Llinares and Nikula (in this volume) and Llinares and Dalton-Puffer (2015) for a study on evaluative language in CLIL classrooms. (7) The English word narrative has a more general meaning than G Erzählung, which is equivalent to story (told). Narrative these days is used to designate any account which has some inner coherence (researchers for instance talk about the narrative of their articles).

2 Historical Literacy in CLIL: Telling the Past in a Second Language Francisco Lorenzo and Christiane Dalton-Puffer

Introduction The rendition of the past – history – depends critically on language but does so in a controversial manner. Narrators and tellers present events with linguistic resources which form and deform reality. Language in narratives is affected by world view, bias or sheer subjectivism and in the act of telling, events are created and recreated. For theorists in the field, then, history is not believed to exist outside language and accounts of the past presented as a grand narrative, as an objectified representation of ancient deeds and processes are simply rejected (Barthes, 1967). On the contrary, historical representation is thought to operate within the limits of natural languages, which has led to an emphasis on the role of rhetoric, the topic of narrative and the poetics of history, a research angle also dubbed the linguistic turn (Carrard, 1992; Yilmaz, 2007). Language functions of historical relevance such as causality (what brought about the events) and agentivity (the forces that caused the past and the ones that could have prevented the past from happening) have been thoroughly studied, and many definitions of the discipline have linguistic overtones. For example, history has been defined as ‘the past as a process of becoming the present’ (Hobsbawm, 1997: 21), which implies a sophisticated ability of the progressive and complex verbal tenses to render historical timescale. Others approach the discipline as the sequence of events that could have been avoided, which brings to the fore the expression of agentivity: the forces that caused the past and the ones that could have prevented the past from happening in the way it did. It is not unexpected that when history education, tasked with knowledge transmission across generations, decides to switch languages in multilingual societies, debates arise on a number of fronts: (a) the extent to which minority language students will be able to incorporate 55

56

Part 1: Curriculum and Pedagogy Planning

their own voices in the interpretation of history; (b) whether language adaptation as often required in bilingual settings preserves decisive historical functions, or rather conceals and distorts events; (c) whether the foreign structures of second language (L2)-based history may tamper with the hegemonic interpretation of national events. All of these issues have been discussed regarding American history taught in English to Hispanic students in the USA (Achugar, 2009), bilingual Arabic and Hebrew teaching of the history of the Middle East in Israel (Adwan & Bar-on, 2003) or the Franco-German teaching of European history in bilingual programmes (Breidbach et al., 2002). A proper consideration of literacy is required in connection with history in an L2, a notion that we conceptualise within the context of content and language integrated learning (CLIL) as ‘L2 historical literacy’. Within the wider context of the volume, this chapter discusses history learning and teaching in an L2. Firstly, we will outline a model to understand the dual merged ability of historical competence and language competence. Secondly, we will approach historical literacy as a developing trait and characterise the linguistic performance of learners in L2 history courses both in the written and spoken mode.

Historical Literacy and Literacies The term ‘literacy’ with the denotation ‘ability to read and write’ has been in use in English since the early 20th century in a range of disciplines (among others, in cross-cultural psychology, sociolinguistics, education studies and later also in applied linguistics [cf. Antos & Krings, 1989]). Research into the learning of reading and writing in the first language (L1) has been – and in many cases still is – dominated by a focus on the early years and by approaches that view literacy as a complex but autonomous set of skills (‘Kulturtechniken’) as exemplified by models like Bereiter and Scardamalia’s (1987) or Becker-Mrotzek’s (1997), to name only two. This position also tends to be connected with the assumption that improving the command of these technical skills leads to enhanced cognitive functioning in the learner. Taking this position as a starting point, two further developments in the educational discourse about literacy can be discerned: (a) Criticism of the ‘autonomous view’ of literacy by scholars working in non-majority or multilingual research contexts has led to the development of an alternative position, namely that literacy is very much a social practice that happens in particular social spaces (Barton, 2007 [1994]; Hornberger, 2008). New literacy studies (NLS), in particular, hold that there is more than one literacy, that literacies are social practices associated with certain domains (one such domain being,

Historical Literacy in CLIL: Telling the Past in a Second Language

57

for example, school hence school literacy), that they are related to people’s identity positions, values and conceptions of knowledge and tied to issues of power (e.g. Barton & Hamilton, 1998; Prinsloo & Baynham, 2008; Street & Leung, 2010). By and large, language users depend upon formal education at secondary level for learning and practising those literacy skills that will give them access to either higher education and/or to a job market that requires international communication. This is where the notion of biliteracy comes in: increasingly these days, young people need to possess literacy skills and literacies in more than one language. (b) The cognitive functioning view of literacy itself has experienced another ramification: powered by a broadening of the underlying concept through loosening its ties with the actual activities of reading and writing, the term ‘literacy’ has come to include general notions of skills, abilities and even knowledge. Organisation for Economic Co-operation and Development (OECD) principles of cross-curricular competences and literacy development mean that education systems are being invited to construct education as the development of various literacies (cf. OECD, 2001, 2002). We now speak of mathematical literacy, science literacy, subject-specific literacies, digital literacy, etc. in an attempt to capture the idea that whole sets of skills, practices and knowledge items in specific areas are necessary for citizens to competently negotiate their lives in modern society in ways that ensure their democratic participation. This broadening of the notion of literacy has not, however, meant that reading and writing have become less important. On the contrary, the demands of cross-curricular competence development have thrown language into relief in connection with other subjects, i.e. to work around history in relation to L2 competence (historical literacy and language literacy) or maths and L1 development (literacy and numeracy), demanding cross-curricular links between language and non-language areas. A curricular venture pointing in exactly this direction is the Council of Europe’s plan to benchmark ‘historical communication’ for school-level learners (Beacco, 2010: 10). Beacco holds that it is possible to develop a set of descriptors for the language abilities necessary in learning and teaching history by mapping and analysing in-school as well as out-of-school communication situations involving history. Plurilingual education presupposes linkage of the classroom modes of communication to the social ones involving history, so as to make transfers of proficiencies between them. At least some of the classroom modes of communication should enable learners to handle social situations of communication with historical content: either directly

58

Part 1: Curriculum and Pedagogy Planning

through the classroom use of these social forms, or indirectly, with the same proficiencies as those inherent in the social forms being developed through the classroom forms. (Beacco, 2010: 10) With this tool, Beacco (2010) has explored the links between historical knowledge and language descriptors and in-school discourse types, producing a construct called ‘historical communication’. This approach intends to formulate sets of language competence descriptors and can-do statements as found in the Common European Framework of Reference (CEFR) that map historical content and discourse and cognitive skills, hence attempting the description of integration between content and language. An example of this would be the instruction: ‘place the occurrence under discussion in a broader context (chronological, cultural)’ or ‘distinguish objectified discourse from judgement’ (Beacco, 2010: 10). Insofar as the CEFR provides a model of language description and assessment, the project could potentially show a full map of the terrain of L2 historical literacy anchored in recognisable language competence levels. Beacco, however, never proceeds to characterise the CEFR levels of historical communication into actual language description. This is where the present chapter intends to step in with regard to history.

Historical Literacy in the Curriculum In many education systems, the emergence of the above-mentioned wider notion of literacy has led to a considerable reorientation in conceptualising school curricula. Many curricula now formulate learning goals in terms of competences, and competence models for school subjects have consequently been developed. This section will provide a condensed analysis of how the different readings of literacy – the general competence side and the read/write side of the notion – are intertwined and inscribed in specific curricula. Our examples come from the national curricula of Austria and Spain for upper secondary history. In both education systems, the curriculum has to be followed irrespective of whether the language of instruction is the national or regional majority language, a second or minority language or a foreign language. The Austrian curriculum for history builds on a competence model of historical literacy and defines four key ‘historical competences’: Fragekompetenz, Methodenkompetnz, Sachkompetenz and Orientierungskompetenz (questioning competence, methods competence, factual competence and orientation competence). Besides that, great emphasis is placed on the learning goal that students should develop and is summarised as a ‘personal activity repertoire enabling political engagement and formation of opinion’ (‘Entwickelneines individuellen Handlungsrepertoires für die politische Auseinandersetzung und Meinungsbildung’)

Historical Literacy in CLIL: Telling the Past in a Second Language

59

(BMUKK, 2005a). In order to get there, learners should be introduced to ‘methods and procedures’ that enable them to articulate their political ideas in speech and writing, visually and/or via modern media (BMUKK, 2005a, 2005b). The pedagogical measures and teaching techniques suggested for working towards that aim involve ‘real’ (i.e. out-of-school) and simulated (in-school) actions, as well as the fabrication of material products. A wide range of activities, text-types, genres and products is suggested for this purpose: street interviews, expert interviews, roleplays, debates, tribunals, decision games, posters, oral and written reports, newspaper articles, photo-documentation, videos, exhibitions, quizzes, tables, flyers and worksheets. Looking at this list, it is clear that, irrespective of the suggested split between ‘real-life’ and ‘simulation’, most activities mentioned implicate oral and/or written language. However, no further details as to the possible specific characteristics and purposes of individual text-types are given and it seems that the curriculum here relies on teachers’ everyday understanding of these genres. More fine-grained operators that could help to spell out the intended connection between the activities and the four historical competences aspired to occur only rarely, if at all. Some that do occur are verstehen, erklären, bewusstmachen, analysieren, fragen, interpretieren, Urteile bilden und formulieren, die eigene Position artikulieren, andere Position aufgreifen und weiterentwickeln (understand, explain, become conscious of, analyse, ask, interpret, formulate judgements, articulate own position, take up and develop other positions). The core history curriculum in Spain is also a national curriculum and operates for all languages irrespective of their status (regional, foreign or state language). History objectives are quite explicit about the language dimension of the discipline and formulate linguistic goals and language assessment indications. Much as in the Austrian curriculum, the official guidelines in Spain do not elaborate further the connection between both fields – language and history – and the level of specification is roughly as in the following extract from the history course in the first year of upper secondary school: Los estudiantes deben emplear con propiedad la terminología y el vocabulario históricos y aplicar las técnicas elementales de comentario de textos y de interpretación de mapas, gráficos y otras fuentes históricas. (BOCM, 2008) [Students must show appropriate use of terminology and historical vocabulary. Basic techniques of textual commentary and interpretation of maps, graphs and other historical sources must be practised.] At times, linguistic goals and objectives are implicit and intertwined with historical notions and knowledge structures. The following goal of the final year in secondary school demands expertise in information

60

Part 1: Curriculum and Pedagogy Planning

processing strategies and finding and sorting out alternative sources. An important number of discourse functions are involved: Buscar, seleccionar, interpretar y relacionar información procedente de fuentes diversas, -realidad, fuentes históricas, medios de comunicación o proporcionada por las tecnologías de la información-, tratarla de forma conveniente según los instrumentos propios de la Historia, obteniendo hipótesis explicativas de los procesos históricos estudiados y comunicarla con un lenguaje correcto que utilice la terminología histórica adecuada. [Students must be able to search for, select, interpret and relate information from diverse sources – real life, historical sources, communication and information media – treat them according to the appropriate historical procedures, obtain explanatory hypotheses concerning the historical processes studied and express the information in correct language using adequate historical terminology.] The two history curricula exemplified here may be diverse in goal setting, competence descriptors or pedagogical strategies, but they do share one fundamental thing: the teaching and learning of history is strongly embodied in linguistic notions and categories. Inasmuch as history teachers’ understanding of historical literacy grows from these curricular principles and/or from the official textbooks based on them, a linguistic bias, albeit largely inexplicit, seems inevitable. How this plays out with regard to learner historical literacy in their L1 is hard to fathom because empirical research into the historical literacy achievements of learners in their respective first languages (Spanish and German in our case) is not available. In the case of L2 historical literacy as required in CLIL, the language question is closer to the level of conscious awareness, however, usually only inasmuch as lexical resources are concerned. A broader concept of historical literacy is thus in order and will be introduced in the following section.

The Three Tiers of Historical Literacy History as an academic discipline and school subject has its own set of recurrent knowledge structures or content schemata. The content end of CLIL in history includes historical meanings which trigger particular mental representations. We are going to approach these three content layers, these content units, as corresponding to three categories. At the bottom level of conceptualisation there are ‘historical notions’ which express the interaction of social groups or individuals in the past (liberation, secession, union, warfare). These are articulated into ‘gestalt historical principles’, which sustain the discourse dynamics of those events. As stated above, key relations in history are cause and consequence relations, which render

Historical Literacy in CLIL: Telling the Past in a Second Language

61

the past as resulting from knock-on effects (impulses) or agentivity, the forces responsible for the historical events (Christie & Maton, 2011; Coffin, 2006a, 2006b). There are, finally, ‘historical heuristics’, which help manage the major coordinates of historical knowledge. The mere representation of events does not constitute history; the discourse of history demands cognitive academic operations. Wineburg (1998) lists sourcing, corroboration and contextualisation as the major metacognitive operations of history. Sourcing, for instance, demands the ability to express and contrast others’ accounts of the events. Corroboration requires an advanced historical competence to contrast sources in the quest for truth. Contextualisation, finally, imposes on the historian a wider lens to consider facts in an extended spatial and temporal scenario: an event reflected in isolation is a faulty narrative. The spatial and temporal coordinates need consideration too. These three types of historical knowledge structures – historical notions, gestalt historical principles and historical heuristics – have a recurring imprint in language and correlate with predictable units at different language levels: genres, functions and lexico-grammar. Genre, function and lexico-grammar language units constitute the language end in CLIL, amounting, as they do, to a linguistic switchboard activated by content. Historical knowledge structures flesh out these linguistic categories: genres (historical recounts, for example), functions (expressing causality) and lexico-grammar (including not only words and lexical chunks such as rebellion, defeat, seal, but also grammar units: past tense formation, subordinators etc.). We believe it is in this connection between content knowledge structures and language units that integration in CLIL lies. We will illustrate the three levels further, drawing on both the spoken and written mode in school history. Genres: Historical texts have particular macrostructures consisting of regular prototypical rhetorical moves (Van Dijk & Kintsch, 1983: 59). Narratives – the historical genre par excellence – revolve around an event, followed by a description of the development of a previous state of affairs, the presentation of ensuing conflicts and a closing (Adam, 2011). As stated above, integration consists of the matching of content schemata and rhetorical schemata, an interface of content and language. Of practical importance is the fact that genres are language events known and mastered by content specialists: historians are knowledgeable in the rhetorical structures of their speciality genres; both of the primary sources (chronicles, sagas, annals, cosmogonies, chorologies) and of the academic genres of history as educational endeavour (historical paper, history quiz, historical dissertation, in-class history essay). Content teachers possess an implicit knowledge of the language structure of their speciality texts and know, therefore, if those texts are well formed. It is this latent language awareness

62

Part 1: Curriculum and Pedagogy Planning

that content teachers can exploit to integrate content and language at the uppermost linguistic level, that of genres (Lorenzo, 2013). As our brief curriculum analysis above has shown, genres do play a significant organising role there, with both spoken and written genres being frequently mentioned (without normally using the word genre, however). In educational linguistics, written school history genres have received highprofile theoretical and descriptive elaboration in the context of Systemic Functional Theory (e.g. Coffin, 2006a). The extent to which subject-specific genre plays a role in oral classroom discourse is an open question for us and probably depends on one’s understanding of what extension the term ‘genre’ actually has. If one understands genre as ‘staged, goal-oriented activity’ (cf. Martin & Rose, 2003), the lesson certainly falls within this definition. As do lecture or presentation. The existence of other oral classroom genres is less well established. However, as Evnitskaya and Morton (2011), Llinares and Morton (2010) and Morton (2010) have shown, one can make out protogenres such as oral biographical accounts, in classroom interaction. These proto-genres form an interesting zone of overlap with the macro level of cognitive discourse functions (CDFs) (see the chapter by Dalton-Puffer in this volume). Functions: When talking or writing about subject matter, speakers need to express general as well as subject-specific cognitive operations such as classifying, contrasting or hypothesising, thereby making explicit the semantic relations between subject-specific terms and concepts, on the one hand, and their own subject positions vis-à-vis these knowledge objects, on the other. This intermediate level of linguistic expression has been variously referred to as academic language functions or CDFs and it is understood that learners need to master the linguistic patterns tied to them if they want to understand subject matter presented to them linguistically. They also need an active command of such functions in order to be able to demonstrate their understanding of said subject matter. The exact number and linguistic nature of these functions has not been subject to much systematic expert discussion, but Dalton-Puffer (2013) has recently proposed a concise construct consisting of seven function types where each type is a category covering more specific realisation types. The seven function types are: Classify, Define, Describe, Evaluate, Explain, Explore and Report (for a detailed account of the construct, see Dalton-Puffer ’s chapter in this volume; its theoretical foundations are discussed in Dalton-Puffer, 2013). In learning history, then, CLIL learners need to acquire the means by which they can recognise these CDFs in their L2 and ideally realise them themselves. As such activity simultaneously constitutes a large part of what we call subject-specific knowledge (as convincingly argued by, e.g. Lemke, 1990), CDFs must thus also be regarded as essential elements in historical literacy. Examples of how this plays out in CLIL history lessons will be explored below.

Historical Literacy in CLIL: Telling the Past in a Second Language

63

Lexico-grammar: Among the structural grammar patterns, some stand out for their historical salience and are more recurrent in this discipline’s texts. Examples include the structures that present an interaction of time and causes and the expression of abstraction or temporal markers (thereafter, for example). Common discourse features described in the literature include advanced structures such as: (a) a clause introduced by a temporal conjunction: ‘After the Cherokee refused to move, the Georgia militia began attacking Indian towns’; (b) non-finite dependent clauses: ‘Ignoring the Cherokee’s treaty rights, Georgia officials begin preparing for the removal’ (Achugar & Schleppegrell, 2005: 306). Grammar also operates in more tangential ways: causality can be expressed including a default resource (asyndetic coordination) consisting of the avoidance of linkers. Facts are presented as concurring and readers are prompted, sometimes misled, to find a cause and consequence relationship. This means that causes and consequences can be implicit in the lexical choices. In a historical episode, a Roman Emperor can be blamed for disaster but the reasons and causes are not expressed with morphosyntactic resources but with lexical items (in italics): ‘Fallibility lies greatly with the negligent Nero, who was so disinterested in the British provinces that laws became mere guidelines’. Asyndetic means for causal expression denote expertise in the teller and put high demands on the reader (Fitzgerald, 2011). Regarding the lexicon, the lexis of history can be extremely specific, constituting a terminology. Glossaries and wordlists are common tools for CLIL lessons and materials. However, something which may look paradoxical is that there is no direct lining up of content layers and language layers: a lexical or grammar item normally expresses an historical notion: treaty (lexical) or was faring (progressive past action). But lexical items can also realise full discourse function: a phrase like election frauds blames authorities for any ensuing events, hence expressing causality. Finally, even if anecdotally, an historical document that constitutes a genre like a placard can consist of one single utterance (Go home on a historical billboard, for example). The choice of particular words is hardly ever neutral in historical discourse and advanced functions like expressing particular views (stance taking) may be sugarcoated in lexical choices. The power of words in this regard is often highlighted. Known is the idea that ideology rests in adjectives. We may conclude so far, that the three-tier framework presented here has the potential to guide a coherent description of the complex construct that is ‘historical literacy in a second language’ and be instrumental for CLIL lesson planning and assessment, and further to understand L2 acquisition processes in immersion settings. Before we move on to apply this model to the written and oral language naturally produced in history classrooms across Europe (sixth section), two further points on historical literacy need commenting on: (a) historical

64

Part 1: Curriculum and Pedagogy Planning

literacy as a competence developing over the years; and (b) historical literacy development in an L2, i.e. when historical literacy becomes historical biliteracy.

Historical Literacy: Developmental and L2 Constraints One needs to approach literacy as an ability evolving over time: narrating the past can take a conversational or an academic form. Storytelling is a basic communicative function which demands no advanced language skills. However, when a formal recording of the past is intended, these narratives reach historical status, a discursive realisation one notch above anecdotal recounting as in casual conversation. History is articulated through advanced language structures demanding cognitive maturity and language competence. The educational literature has delineated the development of historical academic language over time, providing findings for either L1 or L2 growth. For learners, the expression of the past starts as a personal recount. Next, facts are rendered in time in a bare sequence of actors and actions (biographies and historical recount). Later in time, causation and agency take the focus of the narrative (accounts and explanations). Historical discourse reaches maturity with the rendition of multiple causes and effects, a stage where tellers gain their own voice. With this classification in mind, Martin and Rose (2003) observe that the discourse of history has two major turning points: the comprehension and production of causation, and at a later point in language growth, the comprehension and production of multiple factorial causality. Furthermore, this process has stages achieved at particular ages so that written literacy plays out over time with visible hallmarks along the way. Up to age 13, learners merely record (–13); in between the ages of 14 and 16 they can explain complex causes in the construction of the past; and finally between the ages of 16 and 18 they take a personal stance and present arguments about events with critical and personal interpretations. Recording, explaining and arguing constitute, therefore, three critical periods in the development of historical literacy (Coffin, 2006b). In an L2, the semantic and syntactic mapping of advanced historical functions (causality and appraisal) may delay historical competence. Learners of history through an L2 are put at a clear disadvantage, having to struggle with unfamiliar language categories, being deprived of fundamental knowledge to understand events in time and be able to present an opinion or develop a voice. This breach of knowledge can only be overcome through programmed attention to language through approaches such as: sheltering, adjusting, scaffolding or integrating, however they may be called in the different research strands.

Historical Literacy in CLIL: Telling the Past in a Second Language

65

L2 history students have to produce above their existing L2 capacities, a situation also dubbed ‘pushed output’ in L2 acquisition studies. As a consequence, students at early stages are tremendously faulty in their production to the extent that texts reflect a lack of competence working at a stage of performance without competence (Alison, 2002). They may engage in some discourse babbling: frame their utterances in discourse stretches, plan long stretches of discourse but then are unable to flesh them out because of insufficient competence. For all the limitations that L2 history learning may pose, a very visible overlapping occurs in L1–L2 academic language competence: L2 proficiency measures occur in parallel to an expansion of L1 cognitive academic language proficiency (CALP). During adolescence, academic language evolves and language units like sentences and noun phrases grow in length and complexity. Over time, there exists an evolution in clause-level features in complementation and adverbial subordination, which happens for both languages almost at once. The parallels between L1 and L2 academic writing competence growth are very noteworthy. Schoonen et al. (2009) note a 0.93 correlation between L1 and L2 CALP development in advanced writing, which no doubt reflects the underlying language competence that guides production in all languages. There exists a continuous crossover of L1 and L2 rhetorical structures, a factor that makes writing more effective and appropriate in both languages. Language and cognitive constraints seem to grow in parallel. Llinares and Morton (2010) have found that in the lower grades of secondary education, history is perceived and narrated as personal anecdotes of past figures disconnected from the abstract causes that constitute historical processes. Both cognitive and language constraints are believed to work in combination here.

Historical Literacy in CLIL in the Spoken and Written Mode In this section, we would like to expand on and exemplify the framework outlined in the fourth section above. It is not our intention within this chapter to offer a full-scale application of all three tiers over both modes (spoken, written) in several first and second languages. Instead, we have elaborated on specific foci, but with the three tiers (genres, functions and lexico-grammar in language and historical notions, historical gestalt categories and historical heuristics in content) providing an overall conceptual frame that makes clear the relation between what is being elaborated and what is not being discussed at this point. Given that school history genres are relatively well covered in the literature, our focus is on CDFs in the spoken mode and on lexico-grammar in the written mode.

66

Part 1: Curriculum and Pedagogy Planning

Cognitive discourse functions in CLIL history classroom talk Much like genres, CDFs appear to be constitutive parts of history curricula (as seen in the second section). They seem to be essential elements in subject learning because they serve to verbalise general and subject-specific cognitive operations and express the relations between subject-specific terms and concepts. All of this taken together constitutes a large part of what we call ‘subject-specific knowledge’ (see, for instance, Lemke, 1990). The starting point of this section, then, is the assumption that oral historical literacy rests a good deal on the ability to perform CDFs, which involves being able to do this in an L2 if the learning context is CLIL. Since in typical CLIL contexts, the classroom tends to be the main, if not the only, everyday environment where learners have to use their L2, it must be in the bilingual/CLIL classroom itself where such oral subject-specific literacy can be acquired and learned. If we take an ‘implicit learning’ or naturalistic acquisition position, oral subjectspecific literacy should be acquired via its sheer presence in classroom talk. The teacher would then be the main provider of that experience, facilitating genre models and linguistic input from which the learning can happen. From an explicit learning position, on the other hand, it follows that not only written but also oral subject-specific literacy has to be explicitly addressed by the teacher. CDFs need to be turned into pedagogical objects and students required to actively perform them. In other words, there needs to be production and there needs to be communication about how thought can be verbalised, specifically there needs to be a metalevel regarding CDFs. In order to explore the learning conditions for historical literacy in real CLIL classrooms, this section will briefly focus on the following issues: the frequency of occurrence of CDFs in history classroom discourse, who performs CDFs and in which lexico-grammatical form and finally the question of whether there is a metalevel. The examples occurring in the following discussion have been lifted from the ConCLIL project database with history lessons from Austria, Finland and Spain as well as an exploratory study of 18 Austrian upper secondary CLIL history lessons (Lackner, 2012). Expectations regarding the occurrence of CDFs in CLIL classroom talk derive from the learning goals formulated in the respective curricula (see above), as well as the literature on historical language (Coffin, 2006a; Schleppegrell & Oliveira, 2006; Zwiers, 2008) and suggest the occurrence of a broad range of functions. What is to be expected, however, is a particular emphasis on causal explanations as one of the most important aspects of history instruction (Coffin, 2006a: 265).

Historical Literacy in CLIL: Telling the Past in a Second Language

67

Lackner’s (2012) findings, however, which focus on the CDFs Define, Describe, Explain and Classify, do not show any of these to be a truly frequent phenomenon: over 18 lessons (660 minutes) a total number of 108 episodes realising one of these four functions could be observed. The distribution over the four CDFs concerned was Explain (39), Describe (38), Define (19), Classify (12) (see Dalton-Puffer, this volume, for a range of examples illustrating the realisation of all these CDFs in classroom interaction). None of the functions were found to occur more than twice in a lesson on average. As Lackner’s (2012) definition of Explain was broader than Dalton-Puffer’s (2013 and this volume) and included explanations of terms, only 17 of the 39 Explains were actual causal explanations. We agree with Lackner (2012: 69) that ‘Considering the central role causal explanations are assumed to play in the construction of historical knowledge by “giving meaning to the discipline” (Coffin, 2006a: 262), this finding is surprising’. Concerning the question of who actively performs CDFs, we note that it is overwhelmingly teachers who perform the CDFs or lead their co-construction. Teachers’ linguistic realisations of the CDFs are largely in accordance with the prototypical lexico-grammatical features stated in the literature, as exemplified in Extract 1 which features both clausal expression of cause (because) and also abstract nominalisations (the cause/s of ). Extract 1 Concept of illness in Ancient Greece (Year 11, age 17) T:

this list should include the causes of illnesses ..and the remedies. what they did to heal the illness. now, before Hippocrates, people thought they were ill because the gods punished them. the cause of the illness came from God, and the healing, the cure, also came from God

Numerous examples like this indicate that teachers do possess the procedural knowledge to perform the CDFs. Students, on the other hand, hardly ever realise a CDF on their own and also their participation in co-constructed ones frequently transpires without use of the canonical lexico-grammatical features. Extract 2 Social stratification in Sparta (Year 11, age 17) T:

slaves can be sold, helots can’t. ah ..yeah, why did the Spartans keep them oppressed? (Consequence) S1: they needed something to eat (Cause 1) S2: they were afraid of rebellions (Cause 2) T: they were afraid of rebellions, yes. They even declared war on them once every year.

68

Part 1: Curriculum and Pedagogy Planning

Turning to question four regarding the presence of a metalevel, one can observe that generally CDFs are performed in a strongly contextualised and hence implicit manner. In Lackner’s (2012) study Classify was realised implicitly, without the use of any of its signature features (such as explicit mention of the basis for a classification e.g. ‘right to vote’, ‘requirement to pay taxes’). Describe was routinely intertwined with acts of pointing to images that were present in the here and now of the classroom. There were only very few attempts to take discourse functions on a metalevel and even those few tended to remain rudimentary. The two most explicit meta-episodes found in the data were both related to talk about past or future testing situations. Here is one of them: Extract 3 How to describe a picture T:

what we did now - i ask(ed?) you what your impression of this picture is ..and you told me and you interpreted it, but if you try to imagine the situation at the Matura, if you do the same thing, this might be quite ah quite stressful and quite difficult for you to do that. so the simpler version would be if you follow ah the guidelines below: simply starting out with .. describing the picture, using these .. phrases, and once you have described the picture and you- you have talked about the obvious, .. then you could start interpreting it. ah okay, so could i ah Monika, i am not so sure, are you going to do this oral examination? okay, so, Monika, in the foreground, could you follow these rules here at the bottom.

Episodes like this seem to indicate a certain awareness of (some) teachers for the importance of CDFs, albeit not with regard to the classroom learning situation itself. Lackner (2012: 105) rightly concludes that ‘the implicit ways in which participants often realise thematic patterns is a serious obstacle to content learning’.

Lexico-grammar in history writing Historical narratives unfold in particular genre structures incorporating discourse functions which in turn are constituted by lexical and grammatical language items. Linguistic features in historical writing represent different communicative functions and different discursive dimensions. For instance, nominalisation freezes events and sets them out of time, as is key for defining. Multiple subordination facilitates multiple factorial explanations needed for higher-order functions like arguing or explaining. Finally, particular subjunctive types enable language users to present alternative histories hence hypothesising or presenting counterfactuals.

Historical Literacy in CLIL: Telling the Past in a Second Language

69

Corpus analysis of narratives – the historical proto-genre par excellence – shows that some grammatical features occur more often than others. This is not to say that there is a grammar of history, but – as stated – certain structures and lexicon correlate with historical texts and therefore, such texts share some family resemblance. Biber et al. (2006) show that particular linguistic features co-occur to form a distinguishable written register dimension known as ‘informational reports of past events’. Briefly put, this dimension includes only the preterite tense (but not imperfect), reflects a focus on past events with little background description and numerous proper names rather than third-person pronouns. This dimension also includes features of high informational prose: such as long words and premodifying attributive adjectives. The foci formulated with regard to oral discourse are relevant also for this section. Here too, we can single out the most relevant lexicogrammar and its frequency in historical discourse. To do this, we will use examples from a corpus of narratives collected in the history class in Spanish secondary schools for three consecutive years. This is a narrative of a fourth-year secondary CLIL student. Extract 4 The Industrial Revolution (8th grade, age 16) The Industrial Revolution The first Industrial Revolution started in the 18th c. in the UK and then it spreads out all over the world. One of the main changes which took place then was the invention of the train and the steamboat, as well as the construction of roads, railways and canals. The train and the steamboat were built when the steam engine was applied to transports. The first canal which was opened was the one which went from a mine to Manchester and the first railway went to Manchester too. Those new transports were used to move people and goods which became cheaper because the transport was cheap too. In addition, everything was nearer than it was before because the train or the steamboat were really quickly. Due to that, the trade grew. Those transports used to use steam power that they took from coal and it produced too much pollution. But there were terrible things that used to happen then. Children used to work in mines where tall people couldn’t enter, and people used to live in very poor conditions in cities because many people moved from the country to the cities which became dirtier and polluted. We will look at three aspects of English grammar in relation to these learner texts: the expression of time, nominalisation and adverbial subordination. See more on syntactic complexity in bilingual historical discourse in Lorenzo and Rodríguez (2014).

70

Part 1: Curriculum and Pedagogy Planning

Expression of time: Over the years, students develop a full grasp of consecutio temporum, a harmonical agreement of tenses involved in the telling of the past. Learners need to move from the expression of here and now language, to situate events in more distant places and in future and past time frames. In this, they need to produce advanced tense structures like backshifting so that what was originally spoken in the present tense is reported using the past tense as in indirect speech. Another challenging structure is the setting of two past levels of discourse, one produced earlier in time than a second past action. Backshifting and parallel past actions are necessary to express multiple factors such as causes of events and, finally, the ability to judge and take personal stances, hence acquiring a voice in the interpretation of history. The narrative above struggles with time reference: some past actions are intended to be presented as happening before another time signalled. Past perfect would have clarified this parallel timescale (‘had moved from the country’, ‘than it had been before’) but the 16-year-old L2 learner in the sample never uses this tense. The narration looks flat. Nominalisation: Christie (2002, 2012) has given a full account of the evolution of nominalisation in adolescence. Children start developing the nominal group structure at the age of 10 and, as they mature, nominal groups expand too. At around 11, they start experimenting with complex reference, a resource without which language users can barely control the functionalities of pro-forms. Later in adolescence, individuals develop complex nominals further including more defining relative clauses and prepositional phrases. As a textual resource, nominalisation is key in academic prose because it encapsulates processes into nominal abstract constructs. Complex nominalisation also increases lexical density, with texts being more tightly knit and more taxing to the reader (see Asención-Delaney & Collentine, 2011; Fang et al., 2006; Schleppegrell & Colombi, 2002). The sample shows that noun phrases are halfway through their growth process. Nouns appear premodified and also postmodified, but both preand postmodification are not in full bloom. Except for the lexical chunk ‘First Industrial revolution’, a rote-learned phrase, all noun phrases have one single premodifier, an adjective (‘the new transports’, ‘very poor conditions’). Noun postmodifications happen twice but they are performed with relative clauses only. The definition of historical terms and notions generally incorporates more long-winded expressions like the following definition of a spinning machine: ‘a simple, wooden framed machine that only cost about £6 and was used mainly by home spinners’. Adverbial subordination: In addition to nominalisation, other complex structures representing complex thoughts gradually develop over the years. One of them is adverbial subordination, key in historical discourse. Causality and the way it is expressed either through overt markers or in a tangential

Historical Literacy in CLIL: Telling the Past in a Second Language

71

manner – asyndetic causality – epitomises the core of historical reasoning: something happened because of something else. As stated, narrators can use explicit markers or express the causal meaningless overtly. The sample shows a sentence where the causal marker takes the front position, which highlights the meaning relation and makes it look very explicit: ‘Due to that, the trade grew’. Diessel (2004) has described a regular order of acquisition of adverbial clauses as part of language development: additive connectors appeared first, followed by temporal, causal and adversative ones. Three factors explain this order: the complexity of the discourse function played by connectors; the linguistic complexity of the structure (the need for utterance-initial adverbial clauses appeared later in time); and, finally, frequency in input. The acquisition of the various types of adverbial subordination seems to follow a set path in which the complexity of the meanings they transmit determines the order of acquisition. Hence, time adverbial relationships are acquired first, followed by cause, result and purpose relationships. Next, hypotheticals and counterfactuals are used, and finally, concessives, which are the last adverbial clause to be processed correctly. Concessives are, therefore, at the end point of grammaticalisation. As a major device for presenting contrast in discourse, concessives are key in academic historical discourse and other discipline languages. Its emergence is said to mark entry into adult language. In the sample, most compound sentences use coordinating conjunctions. ‘And’ is pervasive in the narrative and subordinators are mostly causal or comparative: ‘everything was nearer than it was before’. Although the text features a good command of the vocabulary related to the theme and a proper control of basic grammar structures, either age or maturation constraints preclude more sophisticated meaning relations in the presentation of the past.

Conclusion History as an academic discipline has undergone a ‘linguistic turn’, which results from the understanding that historical discourse is embodied in language. Curricula, however, provide at best a fragmented and patchy picture of what historical literacy involves in terms of what it means to be historically literate. In this chapter, we have examined the cognitive and discursive side of this connection and have suggested a three-tier model of historical biliteracy which we apply to naturalistic school history data (spoken and written). The model presents the representation of history in language units of different levels with genres including predictable discourse functions and lexico-grammar. Content and language are thus tied together in the performance of the participants in history lessons.

72

Part 1: Curriculum and Pedagogy Planning

Politically and educationally, learning history through an L2 can be a major contribution to learning otherness, a status that would enable students to rise above the boundaries of their mother tongue, to experience and consider other versions of who we were, of who we are, hence contributing to multiple identity construction. An integrated model of historical literacy, such as that presented here, should facilitate this goal and help spread the pedagogically sound implementation of CLIL.

3 Learning Mathematics Bilingually: An Integrated Language and Mathematics Model (ILMM) of Word Problem-Solving Processes in English as a Foreign Language Angela Berger

Introduction As a content and language integrated learning (CLIL) teacher in a mathematics class, what would you think about the following utterance by one of your students? The Smiths own a piece of land […] the rest of their, is their garden, the rest of their garden, for their garden […] the rest is for their garden, what fraction of their land is covered by garden? […] the rest, ummm, der Rest von ihrem Garten, ach so, der Rest ist ihr Garten, what, was für ein Bruch von ihrem Land ist ihr Garten … [translation from German: the rest of their garden, ach so (German interjection to indicate understanding), the rest is their garden, what fraction of their land is their garden (too literal translation of English word problem text resulting in unnatural, unidiomatic German)] These lines are an extract of a sample transcript of Anna, an Austrian student in a lower secondary school, attempting to decipher and resolve a mathematical word problem in English as a foreign language. What does this transcript reveal about the meaning-making process? Is the foreign language a serious obstacle to solving the problem? How does the foreign language influence the mathematical thinking of the student? And conversely, is the language affected by the mathematics in this passage?

73

74

Part 1: Curriculum and Pedagogy Planning

While the symbiosis of mathematics and language teaching in CLIL classrooms has become an integral part of curriculum planning in Europe and beyond, the specific nature of the interplay between mathematics and language has yet to be addressed theoretically and practically in mathematics education. Although the mathematics teaching literature has attempted to characterise mathematical processes involved in solving word problems (sometimes with explicit reference to linguistic activities, most notably text reception) it has, to date, failed to conceptualise interaction effects between mathematical and linguistic processes. Extensive research is therefore needed to better understand the reciprocal relationship between content and language in CLIL and other types of contentbased instruction in general and CLIL-based mathematics classrooms in particular. Singling out word problems as the epitome of the intersection between mathematical and linguistic reasoning, this chapter maps and describes in detail the interaction between subject content and foreign language from a cognitive perspective, thereby making a major contribution to conceptualising the integration of language and content at the level of individual learner mathematical activity. It reports on a study conducted to analyse the process of solving arithmetic word problems with a view to modelling the interaction between content and language during that process (Berger, 2013). The centre of interest is the specific nature of the interplay between mathematics and language, including the questions of how English as a foreign language influences mathematical thinking and learning in the process of solving word problems and how the construction of meaning and the problem-solving processes unfold. To this end, 48 think-aloud protocols of beginners in English language learning, aged between 11 and 12, were analysed. The study generated the ILMM which, for the first time, allows us to characterise the nature of the interaction between mathematics and language during the process of solving second language word problems. It represents a significant step in advancing our understanding of content and language integration in the learning process itself, thus providing valuable information for mathematics CLIL pedagogy (cf. third section in Chapter 1 of this volume).

The Relation Between Mathematics and Language It is evident that there is a close connection between mathematics and language. After all, language is the primary means through which mathematical content and knowledge is expressed. Conceptualisations of the relationship between mathematics and language, however, have changed over the years. Traditionally, mathematics was regarded as a language in its own right, often described by the metaphor mathematics is a language (Pimm, 1987). Such a representation may seem appealing as mathematics is in some ways analogous to a language as a system

Learning Mathematics Bilingually

75

of communication used by a particular community. For example, many terms – albeit homonymous with expressions used in everyday situations – have specific meanings which are applicable only in mathematical contexts. Others form part of the technical vocabulary used exclusively for mathematical purposes. In both cases, the terminology is meaningful only within the realm of mathematics. Thus, mathematics has long been considered a prime example of a jargon as an artefact to facilitate communication in a highly specialised area (Bauersfeld, 1995: 278). In this view, the language of mathematics appears to be a coherent and self-sufficient system. More recent approaches have emphasised the crucial role of natural language in the field (e.g. O’Halloran, 2005). In this alternative view, mathematics ceases to be a separate entity coexisting alongside other languages, but is subsumed and included in the more comprehensive system of natural language. The descriptive system is in fact not selfcontained, but inextricably linked to and based on the rich natural language underlying it. In other words, mathematical content or understanding is inconceivable without the flexibility of everyday language. Mathematical symbolism has evolved from natural language and requires a linguistic context to determine its function and meaning. In fact, natural language, mathematical symbols and visual representations form ‘a single unified system for meaning making’ (Lemke, 2003: 215). The close relationship between mathematics and natural language is particularly evident in teaching and learning contexts and has been consolidated by the paradigm shift in mathematics education from cognitive to social constructivist approaches. The use of natural language is a prerequisite to the learning of mathematical knowledge. Mathematical content is invariably expressed and developed through language; it is constructed and accessed by means of language (e.g. Sfard, 2008). This idea of language as a tool for mathematical thinking and construction of meaning was supported by Halliday’s (1978: 195) notion of a mathematics register as the ‘meanings that belong to the language of mathematics (the mathematical use of natural language, that is: not mathematics itself), and that a language must express if it is being used for mathematical purposes’. Accordingly, learning mathematics not only involves learning new mathematical terms but new ‘styles of meaning and modes of argument’ (Halliday, 1978: 195), including structural, textual and pragmatic knowledge of language in order to communicate mathematical ideas effectively. Since mathematical learning is firmly embedded in formal education, many writers in the field have focused on the significance of classroom discourse. Sfard and Lavie (2005), for example, point out that the classroom and its specific discourse is the primary context in which mathematical meaning-making takes place. Accordingly, it is the culture and discourse

76

Part 1: Curriculum and Pedagogy Planning

of the classroom that determine how mathematical meanings are negotiated. Ellerton’s (1989) framework for interpreting language factors in mathematics learning and Gawned’s (1990) socio-psycho-linguistic model, in particular, acknowledge the formative effect that the language of the classroom potentially has on learner mathematical understanding. This focus on classroom discourse has drawn attention to two key factors in mathematical learning. On the one hand, it has pointed to the dual function of the teacher as an agent and provider of language in the meaning construction (Ellerton & Clarkson, 1996). On the other hand, it highlights the significance of the learners’ own active languaging in the process (Bauersfeld, 1995), which requires the combination of language competence and mathematical knowledge. This active languaging is particularly relevant in the process of solving word problems, which are the epitome of mathematical tasks requiring the integration of linguistic and mathematical competence.

Learning Mathematics in a Foreign Language The relationship between mathematics and language in the context of learning is even more complex if the language of instruction differs from the learners’ first language (L1) and thus becomes an object of learning in its own right. The dual focus on mathematics and language is the key idea in CLIL programmes, which aim to provide learning in the subject matter and language in an integrated way. This combination in the mathematics classroom triggers complex processes that may not otherwise occur, which raises the question of the nature of the interplay between language, content and learning (see also Barwell, this volume). It seems intuitively reasonable to assume that the sheer complexity of mathematical content is further complicated by the foreign language. In Morris’ (1975: 52) words, ‘[t]he problems of teaching in a second language are accentuated when mathematics is the context of the dialogue. This is due essentially to the abstract nature of mathematics and the difficulties which arise in absorbing abstract concepts’. While it is likely that a foreign language poses additional challenges for learners and teachers alike, the nature of the interplay between mathematics and the foreign language has yet to be addressed more fully in theory and practice. Current research on mathematics learning in a non-native language originates from different contexts, both regionally and conceptually. One strand of research has emerged from multilingual countries like Canada, Belgium or South Africa, where immersion programmes are offered with the second language as the medium of instruction, typically with a view to fostering bilingualism. Another strand derives from CLIL-based teaching contexts, in which subjects other than languages are taught in a foreign language, usually English. Although the role of language differs

Learning Mathematics Bilingually

77

between the contexts of second or foreign language learning, bilingual or multilingual education and CLIL, and generalisations across the board may disregard the specific characteristics of each field, the various perspectives can help to shed light on the question of how an additional language interacts with mathematical content. The pertinent research can be grouped into three broad categories. Firstly, a number of studies are primarily pedagogic in orientation, focusing on the practical requirements and foci of mathematics teaching in multilingual classrooms. Adler (1998), for example, investigates possible tensions and dilemmas that teachers face in the classroom. One of these dilemmas pertains to the relationship between the accuracy of mathematical language, on the one hand, and the appropriateness of mathematical argumentation on the other. Novotná et al. (2001) discuss qualitative aspects of CLIL teacher education, identifying requisite professional competences of practising and future mathematics CLIL teachers. Secondly, research has strongly focused on the discourse and discourse practices in bilingual classrooms. Moschkovich (2007: 25), for example, analyses discourse features in bilingual mathematics classrooms and the way teachers and learners interact, acknowledging the significance of discourse in learning mathematics, which she describes as ‘a discursive activity that involves participating in a community of practice developing classroom socio-mathematical norms, and using multiple material, linguistic, and social resources’. Setati and Adler (2001) describe how students and teachers in the bi- or multilingual mathematics classroom in South Africa move between languages and discourses, where the general progression is from informal spoken language to formal written language at three different levels: from spoken to written language, from main language to English and from informal to formal mathematical language. They point out that the movement between languages in the mathematics classroom is linked to mathematical discourses. Furthermore, Clarkson (2009) proposes a model for language use in mathematics learning for multilingual students in Australia, emphasising the need to learn English not only as a communicative language for everyday conversation, but also academic English as the specialised form of the language used in subjects like mathematics. According to this model, for full management of the discourse, students and teachers generally progress from informal to more mathematical structured to academic mathematical language in the students’ first and second language as well as the teaching language. A third research focus has been on cognitive aspects of how a second or foreign language influences mathematical thinking and problem-solving processes. Investigating the participation of pupils learning English as an additional language in mainstream mathematics classrooms in the UK, Barwell (2005b) demonstrates how students’ learning comprises both mathematics and language learning and concludes that language

78

Part 1: Curriculum and Pedagogy Planning

and content are reflexively related and mediated by the social activity the learners engage in. A number of other studies have demonstrated that learning mathematics in a foreign language has a positive impact on mathematical performance (e.g. Clarkson, 1992, 2007). While some writers explain such effects by enhanced mathematical thinking through the use of a foreign language, others point to increased metacognitive awareness of bilingual learners. Bialystok (2001) concludes that multilingual learners have a higher analytical capacity, being better able to distinguish between relevant and irrelevant information. Clarkson (2007) suggests that bilingual learners perform better as they are able to reflect their own mathematical thinking more effectively and identify errors more readily. Most studies concerning mathematics learning in a second or foreign language refer directly or indirectly to one of the four basic research perspectives suggested by Dalton-Puffer and Smit (2007: 12). These perspectives unfold between two dimensions. The first one is concerned with the focus of the researchers, who can either take a macro view or a micro view of CLIL, depending on how close or distant their angle is. While the micro perspective concentrates on the influence of the CLIL environment on the people involved, the macro perspective is concerned with the implementation of CLIL programmes. The second dimension refers to the object of investigation, which can range from the processes involved in CLIL contexts to the products or results in such contexts. These two dimensions can be combined to form a grid representing four distinct research perspectives, and the existing literature can be classified according to these perspectives. Novotná and Moraová’s (2005) analysis of CLIL mathematics textbooks or Novotná et al.’s (2001) account of professional competences of future and practising mathematics CLIL teachers, for example, represent a product-oriented macro view. Clarkson (1992), on the other hand, adopts a product-oriented micro perspective in his analysis of the mathematical performance of bilingual students in Papua New Guinea. Likewise, Turnbull et al.’s (2000) investigation of the mathematical performance of French immersion students in Canada constitutes a product-oriented micro view. Although there has been some research on the processes from a micro view (e.g. Barwell, 2009; Clarkson, 2007; Gajo & Serra, 2002; Setati & Adler, 2001), this perspective is notoriously under-represented in mathematicsoriented CLIL research. While these studies have focused primarily on discourse practices in bilingual or multilingual mathematics classrooms, mathematical thinking in a second language, and cognitive, metacognitive and metalinguistic aspects of problem-solving processes, research on the precise nature of the interaction between mathematics and language is very rare. The questions of how mathematical thinking is influenced by another language, how the ‘languaging’ of mathematical content plays out and how mathematical meaning is constructed in a foreign language remain empirically unanswered. With a view to narrowing this gap, the

Learning Mathematics Bilingually

79

study reported here (Berger, 2013) sought to characterise the interaction between mathematical content and foreign language in the context of solving word problems, which represent a typical example of the interface between mathematics and language, as they elicit complex intertwined linguistic and mathematical processes (for ‘languaging‘ in the context of a history lesson see Morton & Jakonen, this volume). Learners are required to draw on mathematical and linguistic abilities in an integrated way. The main concern was how mathematical thinking and English as a foreign language interact in the process of solving word problems and how the construction of meaning on the level of individual learners’ mathematical activity unfolds.

Mathematical processes The main purpose of the research was to shed light on the interaction between mathematical content and a foreign language from a cognitive perspective, based on the analysis of how individual learners solve word problems in English. As word problems operate at the intersection of mathematical and linguistic thinking (Barwell, 2009: 64), they are eminently suitable for the purpose at hand. Word problems ‘can be defined as verbal descriptions of problem situations wherein one or more questions are raised the answer to which can be obtained by the application of mathematical operations to numerical data available in the problem statement’ (Verschaffel et al., 2000: ix). They have been investigated extensively both from a problem-solving and a linguistic point of view (e.g. De Corte & Verschaffel, 2006). Many studies have employed word problems as a methodological tool to elicit mathematically rich language in order to investigate specific aspects of the relationship between mathematics and language, highlighting the interactive and reflexive relation between language and content (e.g. Barwell, 2005c). However, none of these studies have provided a detailed account of this reciprocity. The research reported here is thus the first to enhance our understanding of how, from a cognitive perspective, mathematical activities trigger and inform linguistic processes during the problemsolving activity, and vice versa. As a first step to closing the knowledge gap, theoretical accounts of the main cognitive processes involved in solving arithmetic word problems were extracted from the relevant literature. Although the cognitive processing is characterised by individual variation, several generic processes and interaction effects have been abstracted. While some of these processes focus on mathematical activity, others are linguistically orientated. For the identification of mathematical processes, both cognitive accounts of solving mathematical word problems and modelling cycles were drawn on. Although modelling cycles are based on the cognitive processes involved in solving modelling tasks, which in comparison to

80

Part 1: Curriculum and Pedagogy Planning

word problems centre more on the extra-mathematical problem in the real world, they can help to elucidate the cognitive processing while solving word problems as well. The most relevant modelling cycles for the purpose at hand were the ones suggested by Blum and Leiß (2005, 2007) and Borromeo Ferri (2006, 2011). In contrast to modelling cycles, models of word problem-solving provide a more detailed account of text receptive aspects of the solving process. They are, therefore, more eminently suitable to characterise the linguistic dimension of mathematical activity in a foreign language. One of the most important descriptions of the processes involved in word problem-solving was provided by Kintsch and Greeno (1985). Combining insights from psycholinguistics with insights from cognitive science, they proposed a processing model of arithmetic word problems which divides the problemsolving process into a comprehension and a solution phase and integrates aspects of text comprehension into the problem-solving process. The particular significance of Kintsch and Greeno’s model lies in the importance attached to text comprehension as the basis for the problem-solving process. Reusser (1997) proposed additional levels of mental representation that mediate between the text and the mathematical structure: mental representations of the situation as episodic or factual structures, which embody the realworld structures containing the abstract mathematical problem model that is to be deduced. It is the specific story events, actions, temporal structures and intentions of the actors rather than the linguistic-semantic structures that contain information about the corresponding mathematical model. The most detailed account of the problem comprehension process and its components is Novotná’s (2004) model of the word problem-solving procedure, which is more pedagogic in orientation than the ones mentioned above. She divides the problem-solving process into three basic operations. Firstly, grasping the problem involves understanding all information that is necessary for an adequate mathematical model and the creation of that model itself. The second operation is the solution of the mathematical model and mathematical verification. Finally, the third operation involves a return into the context and contextual verification. Although these models help to understand the cognitive processes involved in mathematical problem-solving, not one of these models by itself is rich enough to capture the complexities of solving word problems in a foreign language. In order to characterise the problem-solving process more adequately, parts of the models outlined above were synthesised. The selected processes representing the mathematical content dimension defined for the purpose of this study were as follows: (1) Identification of objects. (2) Identification of the relationship between objects/identification of the episodic or factual structure of the given situation.

Learning Mathematics Bilingually

81

(3) Identification of the question/problem to be answered/resolved by mathematical means. (4) Arranging/extracting/organising/structuring of information. (5) Mathematising – creating a mathematical model. (6) Performing mathematical operations. (7) Interpreting – referring back to the question. (8) Validating. While the modelling cycles (Blum & Leiß, 2005, 2007; Borromeo Ferri, 2006, 2011) were useful for describing mathematically oriented processes, such as creating a mathematical model, working mathematically, interpreting and validating, the word problem models proved more suitable to describe the text and problem comprehension phase. Following Reusser (1985, 1997) and Novotná (2004), the comprehension phase was divided into identifying the relevant objects, identifying the relation between objects including the episodic or factual structure of the given situation or storyline and identifying the question to be answered by mathematical means. Blum and Leiß’s (2005, 2007) process of simplifying and structuring, which pertains to complex modelling tasks, was modified so as to better suit the requirements of less complex word problems. Rather than simplifying the situation or activating extra-mathematical world knowledge, less complex word problems require merely arranging and organising the relevant information.

Language-related processes Problem-solving in a foreign language also triggers additional cognitive processes with a focus on language, both receptive and productive. In order to characterise problem-solving from a linguistic perspective, psycholinguistic accounts of language comprehension and production were investigated. Common text reception models characterise reading either as a bottom-up process (Bernhardt, 1991), in which linguistic resources are drawn on to combine smaller units into larger ones, or as a top-down process, in which larger units such as context or world knowledge affect the perception of smaller units (Birch, 2007). Weir and Khalifa (2008) have proposed the most comprehensive model of cognitive processing in reading to date. They postulate a central processing core as a sequence of cognitive processes with different levels of complexity. At the most basic levels, the processes include word recognition, lexical access, syntactic parsing and establishing the propositional meaning at the clause and sentence level. From then onwards, the reader may have to go beyond what is explicitly given in the text so as to understand the implied meaning relations among different parts of the text. In order to build a mental model of the text, this creative process of ‘inferencing’ integrates new information into what is already known in

82

Part 1: Curriculum and Pedagogy Planning

such a way that a consistent, meaningful and relevant understanding of the complete text can be accomplished. Finally, the reader creates a text-level structure, which implies an understanding of the hierarchy of meanings and all information relevant for the meaning of the whole text. The most significant text production models for the purpose at hand were Levelt’s (1989) blueprint of the speaker and its adaptations for bilingual processing (De Bot, 1992; Poulisse & Bongaerts, 1994). According to these models, the speech production system comprises several specialised, autonomous and automatic components, including conceptualisation, formulation, articulation and self-monitoring. The conceptualiser initiates the production process by generating a preverbal message. The preverbal message is subsequently taken up by the formulator and converted into a linguistic message through both grammatical and phonological encoding. While the former involves retrieving appropriate lexical entries and syntactic forms so as to generate a surface structure, the latter involves determining the phonetic form of the utterance. The resulting phonetic plan is then realised by the articulator, which transforms the articulatory plan into overt speech. The speech comprehension system is a mediating component, which monitors both internal and overt speech, a process also referred to as ‘self-monitoring’. In order to describe the linguistically oriented processes involved in the solving of mathematical word problems in a foreign language, the most relevant parts of the models outlined above were synthesised. The processes representing the language dimension defined in relation to the data at hand were as follows: Text reception: (1) (2) (3) (4)

Lexical access. Syntactic parsing. Establishing propositional meaning. Inferencing/building a mental model. Text production:

(5) Generating a communicative intention/conceptualising a preverbal message. (6) Activating lexical resources/building a syntactic structure. (7) Articulation/phonological expression. Weir and Khalifa’s model seems to capture effectively the text receptive processes, including lexical access, syntactic parsing, constructing propositional meaning at the clause and sentence level and inferencing and building a mental model. Unlike in Weir and Khalifa’s model, however,

Learning Mathematics Bilingually

83

word recognition and lexical access were not separated into two different processes but were considered as one. Moreover, inferencing and building a mental model of the text were combined into one processing category. As regards language production, the salient processes based on Levelt’s model and subsequent adaptations included generating a communicative intention and conceptualising a preverbal message, activating lexical resources and building a syntactic structure and articulation and phonological expression. The review of the relevant models so far allowed a more detailed understanding of the types of cognitive processes involved in solving mathematical word problems in English as a foreign language. It was assumed that the solving procedure involves both mathematically and linguistically oriented activities, and although content and language are strongly connected, the primary cognitive focus can be on one of the two. The language dimension represents linguistic activities which may require deliberate effort. For example, comprehending the word problem in a foreign language involves conscious text reception, and interpreting the results and writing an answer sentence requires explicit text production. The content dimension represents those activities with a mathematical focus, such as creating a mathematical model or performing mathematical operations. Moreover, the procedure also demonstrated that processes in one dimension interact closely with processes in the other dimension. Mathematical processes can trigger language-oriented activities and vice versa. For instance, the cognitive focus can shift from mathematical processes back to text reception when the results prove erroneous during the interpretation phase. Combining mathematically and linguistically oriented processes into one integrated problem-solving model helps to explain and predict this interplay. While the cognitive models to date have focused on one of the two dimensions, they fail to take adequate account of possible interaction effects. The study at hand was designed expressly to characterise the relations between the mathematical content dimension and the language dimension.

An Integrated Language and Mathematics Model On the basis of the empirical results of this study, the interaction between mathematically and linguistically oriented processes can be presented in an ILMM given in Figure 3.1. For the first time, it is possible to model the interaction between content and language-related processing while solving mathematical word problems in a foreign language in an integrative way. This unprecedented representation of how language and mathematical content interact makes a noteworthy contribution to conceptualising content and language integration in mathematics education. While this section provides a summary description of the

84

Part 1: Curriculum and Pedagogy Planning

Figure 3.1 An integrated language and mathematics model (ILMM)

ILMM, the following sections give a more detailed specification of the model and illustrate its application to learner activity. In this model, the cognitive processes listed on the left form the language dimension, both in a receptive and a productive respect; the cognitive processes listed on the right represent the mathematics dimension. The most significant part of the model is the centre, representing the interaction between the two dimensions. A number of brackets and arrows in the centre of the model indicate the nature of this interplay. The ILMM also shows the close connection between some specific levels of processing. Dotted lines connecting the two dimensions signify clusters of processes which are characterised by extended phases of dual focus on both language and mathematics or frequent switching between the two dimensions. The top section of the model indicates that successful performance on arithmetic word problems requires successful text reception. The identification of the mathematically relevant objects and the relationship between these objects is possible only once the propositional meaning has been established. In practice, the first two mathematically oriented processes often concur with the construction of a mental model of the text. The identification of the problem to be solved by mathematical

Learning Mathematics Bilingually

85

means, the arrangement and organisation of the mathematically relevant information and the creation of a mathematical model depend on accurate propositional meaning and an adequate representation of the text. The given link between text reception on the one hand and the arrangement of mathematical information and mathematising on the other hand takes account of the fact that intense processing of the foreign language text, particularly the increased focus on establishing propositional meaning and building a mental model, entails extended phases of organising mathematical information and creating a mathematical model. This implies that learners using a foreign language tend to use the text more profoundly for stepwise deduction of a mathematical model. The link between text reception and mathematically oriented processes also takes into account that translations of the problem text into the learners’ L1, triggered by the high linguistic demands during the text reception processes, i.e. establishing propositional meaning and building a mental model, encourage additional reflection of both linguistic and mathematical meaning. The bottom section of the model shows that mathematically oriented processes initiate text production. The link indicates that, owing to linguistic demands when trying to express mathematical content adequately, bilingual learners often place an increased focus on text production, both in terms of frequency and extent, particularly while creating a mathematical model and performing mathematical operations. This extended focus, in turn, entails additional linguistic and mathematical reflection. The link between mathematics and language also takes account of the fact that the process of interpreting, i.e. referring back to the question or problem and formulating an answer statement, requires particular attention to text production. Finally, the model shows the connection between the interpretation of the problem and the validation of the provisional results on the one hand and receptive language processes on the other. Both interpreting and validating often require the learners to build a mental model of the problem text once again and thus extend the cognitive focus to text reception. As can be seen in the ILMM, the interaction effects are reciprocal in the sense that language and mathematics influence one another. Some of these effects will be exemplified in the application section below.

Specification and Operationalisation of the ILMM The interaction between mathematics and the foreign language presented in the ILMM above is based on the analysis of think-aloud protocols (Ericsson & Simon, 1993; Knoblich & Öllinger, 2006) of Austrian learners aged 11 or 12 while solving arithmetic and word problems. That is, the participants were asked to verbalise as adequately

86

Part 1: Curriculum and Pedagogy Planning

as possible their thoughts while engaging in mathematical activity. The resulting protocols provided information on both the sequence and content of mental processes, allowing the researcher to make inferences about the underlying reasoning. The data collection took place at an Austrian lower secondary school. Two groups of subjects were recruited for the study. While the first consisted of 13 learners attending special international classes in which English had been used as the language of instruction in several subjects for one and a half years, the control group consisted of 16 students who had been taught in their L1 only. All participants spoke German as their native language. Both groups were asked to complete a total of 17 tasks, which were chosen in cooperation with the class teacher and two other mathematics teaching experts with a view to obtaining a varied sample of representative tasks (for more information on the selection of the word problems see Berger, 2013: 183). For the control group, the word problems were translated into German. After an extensive practice phase to familiarise themselves with the technique, the participants were videotaped while solving the given tasks and verbalising their thoughts in the language of their choice. Any observations of situational and behavioural facets deemed relevant for a detailed understanding of the thinking process were noted down. Once the tasks had been completed, the participants were interviewed about their think-aloud experiences and attitudes towards learning in a foreign language, which allowed conclusions about the reliability and validity of the video data. Finally, a random sample of 48 videotaped problemsolving performances was selected for transcription. For the data analysis, the transcripts were enriched by descriptions of any external activity observed in the video performances. Annotations like ‘the student is reading the task prompt’, ‘the student is pointing to relevant information in the text’ or ‘the student is writing an answer sentence’ helped to make more reliable inferences about the focus of attention. Then, the transcripts were divided into the smallest possible meaning units. Table 3.1 gives an example of how the transcript was divided into units. The unit boundaries were established in a multimodal way by considering various expression modalities, including linguistic (e.g. codeswitching, self-correction, false starts), prosodic (e.g. intonation, pauses), paralinguistic (e.g. facial expressions, body language, eye direction, laughs, sighs), content-related (e.g. onset of new information) and action-oriented features (e.g. change of activity, repetition of activity, termination of activity). Such features did not usually occur singly but in clusters. A speech pause, for example, appeared in combination with a change of activity, a different eye direction and new information. The co-occurrence of several features helped to demarcate the unit boundaries more clearly.

Learning Mathematics Bilingually

87

Table 3.1 Illustration of transcript units Units 10.

Transcript Susie is twelve years (1s)

11.

and she has two brothers -.-

12.

13.

one of them is three years OLDer (2s) plus THREE -.-

14.

u:nd zwölf -.-

15.

und zwölf plus drei i:st -.fünfzehn.

16.

Criteria for unit boundaries pause, change of eye direction, change of activity, onset of new information unit pause, change of eye direction, onset of new information unit pause, change of eye direction, change of activity pause, change of eye direction, change of activity, code switching pause, fresh start

External activity writes 12; then adds S =; student script: S = 12

reads word problem text again; runs a pencil down the text while reading; looks up briefly reads word problem text again; looks at written notes writes +3 = ; student script: S = 12+3 =

pause, change of activity

looks at notes; points with pencil to notes looks up

falling intonation, change of activity, change of eye direction

writes 15; student script: S = 12+3 = 15

As a next step, the units were analysed. The main aim of the analysis was to deduce the cognitive processes underlying each unit. To this end, a tabled matrix was created, with the cognitive processes for both dimensions listed horizontally and individual units of the transcripts listed vertically. This matrix enabled the researcher to identify both the sequence and types of processes that the learners engaged in while solving the word problems. At the same time, the matrix was instrumental in revealing parallel foci on mathematics and language and interaction effects between the two dimensions. Table 3.2 provides an extract of the data analysis. In addition to identifying the processes, the quality of these processes was evaluated. Three levels of quality were distinguished: complete (C) for successful execution, partial (P) for partly successful execution or unsuccessful (U) for failed execution of the process. If it was impossible to judge whether or not the process had been completed successfully, it was coded as questionable (Q). A total of 1297 units were analysed in this way. Every decision in this analysis was an interpretive inference based on the holistic consideration of a number of multimodal factors, including the verbalisations, external activities, surface-level phenomena,

5.

4.

3.

2.

1.

Transcript

Susie is twelve years O:LD -.and she has two brothers.One of them is three years old and -.the other is double -.her age.-

C

C

P

P

P

C

C

b C

C

a C

C

P

U

C

c C

U

d

e

Language f

g

Q

A Q

B

C

D

E

Mathematics F

Learner age 11, female, international class with English as the language of instruction

G

H

reads given word problem text

reads given word problem text

reads given word problem text for the first time; reads passage again later on reads given word problem text; points with pen to text reads given word problem text; reading error: ‘old’ instead of ‘older’

External activity and comments

Susie is 12 years old and she has two brothers. One of them is three years older and the other is double her age. How old are her brothers? Write an answer!

Word problem text:

Word problem 1

Table 3.2 Matrix for transcript analysis

88 Part 1: Curriculum and Pedagogy Planning

Learning Mathematics Bilingually

89

contextual cues and the learners’ written notes. To increase the reliability of the procedure, the coding process was conducted twice, with a random sample of the transcripts being coded independently by another expert and equivocal cases being reviewed in discussion groups. More detailed information on the research methodology and the development of the ILMM can be found in Berger (2013: 132ff). The following section illustrates some of the key findings with extracts from the think-aloud protocols, focusing primarily on the upper half of the ILMM, i.e. the interaction between text reception and mathematical processes leading up to the creation of a mathematical model.

Application of the ILMM One of the key findings of the analysis was that working in a foreign language leads to extended phases of text reception and prolongs learner engagement with mathematical content, thereby providing additional opportunities for both linguistic and conceptual reflection. This increased activity manifests itself in a number of ways. Firstly, and quite expectedly, solving mathematical word problems in a foreign language requires additional cognitive resources in relation to text reception. It is linguistic difficulties, in particular, in the reading comprehension phase that lead to longer periods of text-based meaning construction, particularly while establishing propositional meaning and building a mental model. Such difficulties, in turn, result in a larger number of repetitions and selfcorrections in the reading process, and thus require a greater focus on text receptive activity. Secondly, extended phases of text reception owing to linguistic difficulties can retard, impede or even preclude mathematical activity. The analysis has shown that obstacles posed by the foreign language can obstruct the shift of focus to mathematically oriented processes. Sometimes, insurmountable difficulties in the text reception phase lead to a premature termination of the problem-solving process. Thirdly, solving mathematical word problems in a foreign language requires additional cognitive resources in relation to mathematical activities, particularly during the preparatory phase of arranging the relevant mathematical information for further processing, the stepwise deduction of a mathematical model and the interpretation phase. Fourthly, working in a foreign language leads to extended phases of parallel focus on linguistically and mathematically oriented processes or recursive switching between both dimensions. One insight that follows from these observations is that effective mathematical activity in a foreign language requires successful text reception. On the one hand, proper construction of propositional meaning in the language dimension is a precondition for the identification of objects and the identification of the relationship between these objects in the

90

Part 1: Curriculum and Pedagogy Planning

content dimension. That is, learners can identify the relevant objects and the relationships among them in the course of establishing propositional meaning at the earliest. On the other hand, effective inferencing and building a mental model is a precondition for the identification of the problem to be resolved by mathematical means, the arranging or structuring of the information and the creation of a mathematical model. That is, the identification of the pertinent question, the structuring of the relevant information and the creation of a mathematical model in the content dimension cannot be accomplished unless a mental text model has been created in the language dimension. While these findings may not be entirely surprising, the fact that learners working in their L1, by comparison, may well be able to work mathematically without intensive text reception or even full comprehension was less predictable. The data show that L1 users sometimes pick keywords selectively, shifting the focus more readily to mathematical processes without ensuring the adequacy of their text comprehension. They seem to be able to switch to mathematical activities at an earlier stage in their text reception phase than learners working in the foreign language. At the same time, it is interesting to note that this early shift of focus to the content dimension can be a source of error. Expeditious rather than detailed processing of the word problem text may result in a premature switch to mathematical activity and incorrect outcomes. Other major insights from the analysis pertain to the specific nature of the interaction between mathematics and the foreign language. The findings enable a more precise description of interaction effects between the two dimensions. Firstly, the learners’ active engagement with the word problem text when establishing propositional meaning and building a mental model leads to more frequent, systematic phases of arranging, extracting, organising and structuring of the essential information. In contrast to their monolingual peers, learners working in the foreign language display increased mathematical activity concerning the organisation of the relevant details before they are in a position to create a mathematical model. This kind of structuring of the information is not only more frequent, but typically occurs in combination with text receptive processes. Table 3.3 illustrates how the cognitive focus extends from text reception to strategically relevant mathematical activities. In this excerpt, the learner starts the text reception process over owing to difficulties in the comprehension phase. After the first attempt to build a mental text model, she starts rereading the word problem text in Unit 14. In Unit 15, she extends her focus to the mathematical process of organising the data as she notes down a relevant fraction given in the text. In the following unit, she returns to the word problem. It becomes clear that she has successfully established propositional meaning when she defines the fraction as housing area and notes down the extracted

Learning Mathematics Bilingually

91

information. Subsequently, she alternates between establishing propositional meaning and writing down the information several times. This example illustrates that the relevant mathematical objects are extracted and arranged for further use in the course of intensive text reception. Another interaction effect pertains to the creation of a mathematical model. A stronger focus on establishing propositional meaning and building a mental model in the foreign language often results in more extended and frequent processes of mathematising, i.e. creating an explicit mathematical model. Not only do such phases occur more frequently in the problem-solving process, they are also more often accompanied by intensive text reception. That is, the learners display a dual focus on both the text model and the mathematical model. In contrast, learners working in the L1 seem to process the word problem text less intensively and create mathematical models less explicitly. A systematic deduction of the underlying mathematical structure in the course of intensive text reception was less frequently observed with monolingual learners. Furthermore, the analysis has shown that difficulties in the word problem-solving process often trigger translations into the L1, which in turn initiate additional reflections, both linguistically and conceptually. Such translations were generally prompted by two types of difficulties. Firstly, it is problems in the reading comprehension phase that initiate translations, particularly while establishing propositional meaning and building a mental model of the text. Unsuccessful attempts to establish meaning are often resolved by means of translations into the L1. The learners seem to switch to their L1 to establish a text model or test their provisional interpretation of the text. This autonomous languaging process and the resulting alternative expressions in the L1 open up additional opportunities for linguistic as well as conceptual reflection. Secondly, problems arising when the focus is on both text reception and mathematical processes at the same time can also trigger translations into the L1. Learners often resort to their L1 when their cognitive resources are directed to the construction of a mental text model and any of the processes leading to the construction of a mathematical model simultaneously. Again, such translations provide additional opportunities for linguistic and conceptual reflection (cf. the functionality of translanguaging in CLIL classroom discourse as discussed in Moore & Nikula, this volume). Table 3.4 illustrates how the learner, Anna, who we met at the beginning of this chapter for that matter, uses translations of the word problem text into the L1 to overcome difficulties in the meaning construction process. This transcript shows the recursive nature of the text comprehension process, which is characterised by a number of repetitions, corrections

1

1

1

Transcript

OF IT is covered by their house.

Ein Drittel ist gleich.-

HOUSE. Ein Sechstel.-

17.

18. 19.

The Smiths OWN a piece of LA:ND.EIN DRITtel.-

16.

15.

14.

C

C

C

C

a

C

C

b

C

C

c

d

e

f

g

P

C

P

C

A

P

C

P

B

C

C

P

P

D

E

Learner age 12, female, international class with English as the language of instruction Language Mathematics

What fraction of their land is covered by garden? Write down your answer!

The rest is for their garden.

F

G

H

The Smiths own a piece of land. 3 of it is covered by their house. 6 is wood and 4 is covered by grass.

Word problem text:

Word problem 2

Table 3.3 Extract 1

External activity and comments

text; parallel focus of attention on language and mathematics rereads word problem text; parallel focus of attention on language and mathematics writes: = ; focus of attention on mathematics writes: house rereads word problem text; parallel focus of attention on language and mathematics (continued)

writes: 13 ; eyes on word problem

rereads word problem text

92 Part 1: Curriculum and Pedagogy Planning

is WOOD.ist gleich.-

einSECHStel ist gleich.wood -.und ein Viertel.-

is covered by grass.

Ein Viertel ist gleich.-

grass.-

20. 21.

22. 23. 24.

25.

26.

27.

Transcript

Table 3.3 (Continued)

C

C

C

a

C

C

b

C

C

c

d

e

Language f

g

C

P

C

A

C

P

C

B

C

C

P

P C

P

D

E

Mathematics F

G

H

looks at note pad; writes: grass

looks at note pad writes: wood looks at word problem text; rereads word problem text; parallel focus of attention on language and mathematics rereads word problem text; parallel focus of attention on language and mathematics looks at note pad; focus of attention on mathematics; writes: 14 =

rereads word problem text looks at note pad; focus of attention on mathematics; writes: 16 =

External activity and comments

Learning Mathematics Bilingually 93

is.their garden. The rest of their garden.-

28.

29. 30. 31.

13. 14.

The rest of the garden. What fraction of their land is covered by garden? Write down your a:nswer. The Smiths OWN a piece of LA:ND.... The rest of their-

11. 12.

10.

1.

Transcript The Smiths O:WN a PIE:CE of land -.… The rest of the garden -.-

C C P

P

C C

P C

P

a C

P U

U

C C

U C

U

b C

U U

U

C C

U C

d

e

Language

U

c C

f

g

U

C

A

U

B

C

D

E

Mathematics

Learner age 12, female, international class with English as the language of instruction

Word problem 2

Table 3.4 Extract 2

F

G

H

reading error

(continued)

reading error; focus of attention on language; monitor is activated correction

reads given word problem text rereads word problem text

reading error: ‘The rest of the garden’ instead of ‘The rest is for their garden’; monitor is activated erroneous reading is repeated reads given word problem text

External activity and comments reads given word problem text for the first time; reads passage again later on

94 Part 1: Curriculum and Pedagogy Planning

von ihrem Land I:ST.ahm.covered.-

mit GRA.GARten.

57. 58. 59.

60. 61.

55. 56.

What fraction of their land is covered by garden? Write down your answer. Ahm. Welcher BRUCH

… The rest is for their garden.

What fraction of their land is covered by garden?

Transcript FOR THEIR garden.

54.

53.

33.

32.

Table 3.4 (Continued)

U C

C

C

C

C

a C

C

C

C

C

b P

U Qi

Pi

C Pi

C

Q

C

c Q

Qi

d

U C

P Q

P

e

Language

U C

P U

P

f

U C

P

P

g

Q

A Q

Q

B Q

U Q

P

P

P

U

U

C

D

E

Mathematics F

G

H

(continued)

because of the previous reading error it is unclear, in how far word problem text has been understood

lexical resource in L1 not available reads expression from given word problem text

reads word problem text again; because of the previous reading error it is unclear, in how far word problem text has been understood tries to grasp the question to be answered; attempts to translate the question later on reads word problem text again translates the question to be answered; conscious language production; focus of attention on language and mathematics

External activity and comments correction; unclear in how far word problem text has been understood because of the previous reading error it is unclear, in how far word problem text has been understood

Learning Mathematics Bilingually 95

… Die Familie.- Smith.-

… The rest.- ahm.-

Der Rest von ihrem.-

97.

98.

… What FRACtion is covered by their ga:rden?

Transcript ahm.- acht.nein. The Smiths own land. … the rest.- for.the rest is for their garden. What fraction of LA:ND is covered by garden?

87.

77.

74.

72. 73.

62. 63. 64.

Table 3.4 (Continued)

C

P

P

P C

a

C

P

P

U C

b

C

Q

Q

U Q

Ci

c

d

P

P

e

Language

P

P

f

P

P

g

U

P

P

U U

C

A

U

P

U U

B

Q

U

C

D

E

Mathematics F

G P U

H

erroneous translation

reads word problem text again;

(continued)

translates word problem text; conscious language production; focus of attention on language and mathematics

reads question again; because of the previous reading error it is unclear, in how far word problem text and question to be answered have been understood

because of the previous reading error it is unclear, in how far question to be answered is understood

reads word problem text again

External activity and comments tries to make a connection with the question

96 Part 1: Curriculum and Pedagogy Planning

GARten.ach so.der Rest ist ihr Garten. What.Was für ein Bruch.von ihrem Land ist .ihr Ga:rten.

… Wie viel ist der Garten?

99. 100. 101. 102. 103. 104. 105.

123.

Transcript

Table 3.4 (Continued)

C

a

b

Ci

d

P P C

C

e C

Language

Ci

c

P P C

C

f C

P P C

C

g C

A U Q Q

B U Q Q

C

P P P Q

C

D

E

Mathematics F

G

H

rephrase

because of the previous reading error it is unclear, in how far question to be answered is understood

revised translation

External activity and comments monitor is activated

Learning Mathematics Bilingually 97

98

Part 1: Curriculum and Pedagogy Planning

and fresh starts. As the learner reads the word problem text for the first time, it can be assumed that the focus of attention is on text reception. Units 1–55 clearly show the learner’s strenuous and unsuccessful efforts to understand the word problem. It is the difficulties in establishing propositional meaning that lead to repetitions and iterations in her attempt to establish an adequate mental model of the text. In Unit 56, the student uses her L1 for the first time with a view to resolving the problems. She translates parts of the main question from English into German, thus generating an alternative linguistic representation of the intended conceptual meaning. What is interesting about this passage is that difficulties in this translation process itself trigger additional reflections. Another instance of translation occurs in Unit 98. As the student is unable to establish propositional meaning and thus an adequate model of the text, she fails to identify the relevant objects or the factual structure of the problem situation. Her new attempt to do so involves translation of the English problem text into German, which requires her to process the text again. The languaging and translation process triggers reflections that help her recognise the inadequacy of her provisional text model. This epiphanic moment is emphasised by the German interjection ach so indicating understanding in Unit 100 and followed by the correct rendering of the passage into German, which helps her establish the propositional meaning. Subsequently, the learner also translates the main question into German, aiming to get a clearer text model and to identify the mathematical problem. The German paraphrase in Unit 123 seems to be an indication that the question has now been fully understood. This example illustrates that the translations from the foreign into the first language offer additional opportunities for reflection both at the linguistic and conceptual level.

Conclusion The study reported in this chapter has provided some answers to the questions of how the dimension of a foreign language influences mathematical thinking and working, how mathematical content and processes are verbalised and how meaning is constructed during the problem-solving process in a foreign language. Whereas previous approaches to mathematical problem-solving have merely pointed to the relationship between language and mathematics, this one has made a substantial contribution to characterising the interaction between language and mathematics on the level of individual learners’ mathematical activity. Think-aloud protocols were analysed to map and describe the specific nature of the interplay between content and language while learners were resolving mathematical word problems in English as a foreign language. The resulting ILMM shows how the two dimensions

Learning Mathematics Bilingually

99

representing cognitive linguistic and mathematical processes work in concert during the problem-solving activity. This model is innovative as it goes beyond describing the two dimensions; instead, it characterises the specific nature of the interplay between them, revealing that working in a foreign language typically leads to extended phases of text reception and perpetuates learner focus on mathematical content, thereby providing additional opportunities for both linguistic and mathematical reflection. The model has great impact potential, both theoretically and practically. It calls for a solid integration between language and mathematics in the mathematics classroom, whether CLIL based or otherwise. This integration entails more deliberate attention to both receptive and productive language activities on the teachers’ part. The think-aloud protocols have shown, for example, that learners resolving word problems in a foreign language spend considerably more time building a mental model of the text compared to learners who are working in their L1, not least because the foreign language demands a great deal of cognitive effort from the learners. While such extra demands may impede mathematical activity if they exceed the learners’ current level of ability, they provide opportunities for additional engagement with and reflection of mathematical content if they correspond to the learners’ level of potential development and if teachers are able to capitalise on them. The notion of the foreign language being an obstacle to mathematical learning can thus be reinterpreted as the foreign language being a valuable asset to mathematical learning. Accordingly, Anna’s repetitions, self-corrections, fresh starts, pauses and translations quoted at the beginning of this chapter appear to be signs of a highly reflective learning process involving two languages. In this view, mathematics and language do not merely coexist side by side as two separate dimensions but form an integrated whole.

4 A Bakhtinian Perspective on Language and Content Integration: Encountering the Alien Word in Second Language Mathematics Classrooms Richard Barwell

Introduction A basic premise of this book is that research on language and content integration has not paid sufficient attention to integration. Language and content integration can be examined in relation to a number of different aspects of the educational process, including at a curriculum or policy level, at an institutional level or at the level of classroom interaction (Barwell, 2005a; Nikula et al., this volume). In this chapter, I focus on integration at the level of classroom interaction. This focus is distinct from research that looks at integration in curriculum and policy (e.g. Leung, 2005b) or research that looks at integration at an institutional level (e.g. Arkoudis, 2005). While there has been a good deal of research on classroom interaction in content and language integration contexts, some of which I will discuss below, this work has not yet developed a well worked out theoretical account of integration itself. While there is agreement that the separation of language and content is problematic, there is a need for perspectives that ‘transcend such an understanding that conceptualizes language and curricular content as separate reified entities and instead think of them as one process’ (Dalton-Puffer, 2011: 196). In this chapter, I argue that an approach based on Bakhtin’s (1981) dialogic theory of language has the potential to form the basis for such a conceptualisation. A Bakhtinian perspective views language in relational terms, highlights a form of tension inherent in language and understands any instance of

101

102

Part 1: Curriculum and Pedagogy Planning

language use as an encounter with otherness. This approach also makes it possible to link specific moments of classroom interaction to broader social and historical context. To illustrate these ideas, I will draw on examples from an ethnographic study of interaction in elementary school second language mathematics classrooms conducted in Canada.

Language and Content Research of Relevance to Mathematics Classroom Contexts Existing theoretical approaches to the classroom interaction level of language and content integration, for the most part, fall into three groups, although these three groups have much in common. Research from a cognitive perspective looks at content learning in terms of the acquisition of curriculum content, and looks at language learning in terms of the acquisition of correct linguistic forms. The unit of analysis in both cases is the individual learner. A good example of this perspective is Lyster’s (2007) work in immersion settings, in which he sets out a ‘counterbalanced’ approach to teaching languages through content. He highlights research showing that content classrooms, even where specialist language teaching support is available, frequently rely on rather implicit methods of language teaching, such as unmarked recasts. Lyster analyses the use of recasts and repetition, pointing out that in the immersion setting, they serve two different functions: confirmation of meaning and confirmation of linguistic form. This dual function results in several different combinations of feedback: (a) Recasts that simultaneously confirm meaning but disconfirm form. (b) Repetitions that simultaneously confirm form but disconfirm meaning. (c) Repetitions that confirm form and meaning. (see Lyster, 2007: 104) Lyster argues that in this kind of feedback, the focus on form (i.e. language) is too implicit to be productive for student language learning. He goes on to discuss research in which drawing student attention explicitly to incorrect forms, for example, appears to be more productive. Based on this work, Lyster (2007: 126) advocates a counterbalanced approach, which ‘promotes continued second language growth by inciting learners to shift their attentional focus in a way that balances their awareness of getting two for one, that is, learning both language and content together ’. It is apparent that the relationship between languagerelated and content-related feedback is complex. Specifically, Lyster ’s argument concerns explicit attention to either language or content. While this approach does not deny the integrated nature of language and content learning, it conceptualises integration discretely: attention is

A Bakhtinian Perspective on Language and Content Integration

103

either on language or on content. This model, then, can be likened to an interweaving of two distinct strands, rather than a fusion in which any utterance is understood to be both language and content. Research from a functional perspective is generally based on Halliday’s systemic functional linguistics. From this perspective, content is examined in terms of language structures typical of the content discipline (e.g. mathematics), while language is understood in terms of Halliday’s three metafunctions. This approach is typified by Mohan’s (1986) theory of knowledge structures, developed in the context of English as a Second Language (ESL) education. For example, in one study using this approach, Mohan and Slater (2005) analyse interaction in an elementary school second language science classroom. Through their analysis, Mohan and Slater show how scientific theory is principally organised in interrelated taxonomies, a key knowledge structure in the scientific register. The students’ emerging understanding of a taxonomy relating to magnetism is built up through discussion of their experiences in class with magnets. In learning content, in this case science, students need to be able to interpret subtle shifts in the choice of (grammatical) process. In their own learning of English, if they are to be successful in science, they need to master the appropriate use of these different processes and the shifts between them. Since systemic functional linguistics is not a general theory of learning, research from a functional perspective is often combined with Vygotskian theory, including the notions of mediation, scaffolding and the zone of proximal development (e.g. Gibbons, 1998, 2003). These theories are compatible in their attention to social and semiotic aspects of language and learning. Vygotskian perspectives on learning highlight the importance, for example, of more knowledgeable others in guiding learners to develop ‘higher mental functions’ through the internalisation of the kind of language structures investigated by Mohan. Recent research in content and language integrated learning (CLIL) classrooms has also drawn on systemic functional linguistics, including work that highlights genre as a key ‘knowledge structure’ on which to focus in order to develop an integrated approach to teaching and learning (e.g. Lorenzo & Dalton-Puffer, this volume; Morton, 2010). These approaches offer a view of integration in which language and content learning are goal directed, with content represented by particular scientific linguistic structures, such as taxonomies. Students learn from more knowledgeable others how to use these structures and so learn to organise (and hence think about) content in more advanced ways. Finally, research from a discursive perspective draws on approaches like pragmatics and conversation analysis to examine classroom discourse processes. Research from this perspective tends to highlight interactional variation across languages and contexts.

104

Part 1: Curriculum and Pedagogy Planning

Nikula’s (2005, 2007, 2010) studies of interaction in Finnish CLIL and non-CLIL classes provide a good example of this approach. Nikula examined pragmatic features of interaction in CLIL classes in chemistry, physics and history, as well as in English as a Foreign Language (EFL) classes in secondary schools in the same town in Finland. Her analysis highlights differences between the CLIL and EFL classes in how English and Finnish are used, as well as in pragmatic features such as politeness. In a related study, Nikula (2010) compares the interactional features of biology lessons taught by the same teacher in Finnish and in English. In Finnish, the teacher used a lecture-based style; lessons in English were more interactive. As a result, student participation is rather different depending on which language is used. My own research, conducted in the UK, provides a second example. This work investigated the participation of learners of English as an Additional Language (EAL) in mainstream primary school mathematics (Barwell, 2005b, 2005c, 2005d). My analysis was based on discursive psychology (e.g. Edwards, 1997), an approach that explicitly looks at the discursive construction of cognitive processes, such as learning, through interaction (see Morton & Jakonen, this volume, for a similar approach). In one analysis, I examined the reflexive relationship between language learning and mathematics learning during an activity in which two Year 5 girls (ages 9–10) wrote and solved mathematical word problems. My analysis traced in detail the discursive practices through which the girls accomplished language and content learning. For example, they engaged in a lengthy discussion of verb tense that was closely related to discussion of the mathematical structure of the problem they were writing. Hence, at a detailed interactional level, mathematics learning reflexively depends on language learning and vice versa. In research from a discursive perspective more generally, then, learning is seen as an interactional phenomenon, accomplished jointly by participants. Learning content therefore entails learning to participate in content area discourse. While I have proposed these three distinct perspectives as a way to organise research on language and content integration, they clearly have much in common. Indeed, Dalton-Puffer (2007a) draws on all three in her research in CLIL classrooms in Austria. She presents analyses drawing on the cognitive, functional and discursive perspectives to highlight different strengths and challenges of CLIL. One commonality across these approaches is the assumption that learning content must be accompanied by a parallel learning of specific forms of language. While this assumption is not, in general terms, controversial – learning mathematics clearly has some relationship with learning the language of mathematics – the nature of this relationship is at the heart of any conceptualisation of integration (see Berger, this volume). In much of the research in the three strands I have sketched above, this relationship is, often implicitly, a directional one. That is, learners progress

A Bakhtinian Perspective on Language and Content Integration

105

from using informal, everyday and often erroneous forms of the target language towards mastery of formal, correctly used content registers in the target language (Renshaw & Brown, 2007). Thus, Lyster’s (2007) different kinds of feedback are mechanisms for moving learners from grammatical idiosyncrasy to standard forms in the target language. Similarly, Mohan and Slater’s (2005) analysis of knowledge structures is designed to show the kinds of content language learners need to master. The directionality implicit in this model arguably also entails a sense of progression or replacement, with the informal and inaccurate being replaced over time by formal and accurate language use. This kind of model can be traced, at least in part, to a broadly Vygotskian view of learning in school as a dialectic process of developing ‘scientific’ concepts, which gradually replace naïve, everyday concepts. In Vygotskian theory, this process occurs through internalisation, based on another dualism in which external forms replace internal ones. In terms of pedagogy, the teacher’s task is to guide learners towards formal content language in the target language, whether through explicit exercises or implicit feedback (this perspective reflects broader trends in educational research: see, for example, Daniels, 2002; Forman et al., 1993; Mercer, 2000). Indeed, the required language forms in many cases are specified in curriculum documents (for a discussion related to mathematics, see Barwell, in press). There is an institutional dimension to this process, in which learners must grapple with new forms of language in order to be successful in content learning (or, at least, in content performance). In some respects, however, this approach is problematic. In a study of mathematicians’ use of language to explain advanced concepts to a general audience, I showed how experts use formal mathematical language and everyday language (Barwell, 2013). This finding suggests that proficiency in content language involves an expansion of a repertoire of language practices, rather than a replacement of everyday language with content language. To develop an ‘expansion’ perspective on integration, I draw in particular on Bakhtin’s (1981) work.

A Bakhtinian Perspective on Learning Language and Content Bakhtin’s ideas have informed much research on classroom interaction in mathematics education (e.g. Renshaw & Brown, 2007; Roth & Radford, 2011; Sfard, 2008). The majority of this work, however, combines Bakhtin’s ideas with Vygotsky’s, with the latter’s providing the organising conceptualisation of learning and language. Matusov (2011), however, argues that such an approach is problematic (see also Wegerif, 2008). Moreover, these combinations of Vygotsky and Bakhtin’s

106

Part 1: Curriculum and Pedagogy Planning

ideas tend to underplay Bakhtin’s strong theory of language, in favour of more psychological readings of his work. In this chapter, I develop a more thoroughgoing Bakhtinian approach to thinking about language and content learning. I will examine four key concepts derived from Bakhtin’s (1981) work and show how they can be used to think about integration as an expansion of repertoires of language practices, as well as some of the associated consequences of such an approach. First, from a Bakhtinian perspective, meaning is understood in terms of response, rather than a choice made within a semiotic system. We are always engaged in dialogue and this dialogue requires us to respond. In responding, we highlight some things, ignore others and participate in a chain of utterances. Our response, which must be seen in relation to that which preceded it, produces meaning. As Holquist (2002: 48) has it, ‘lack of water means nothing without the response of thirst’. Thus, meaning, always provisional, arises from the dialogic relations between utterances, words, social languages, genres and so on. The meaning of any particular utterance is never unique; multiple meanings can always arise, since many different responses are always possible. Moreover, subsequent utterances can recast earlier ones, so that meaning shifts as interaction unfolds. A dialogic perspective is distinct from a dialectic approach (see Wegerif, 2008), so that Bakhtin’s theory of language stands in contrast to the thinking of Vygotsky’s (1978) work or, for that matter, that of Voloshinov (1986). Second, Bakhtin (1981) understands language in terms of the diversity of relations it reflects and for which he (or his translators) uses the term heteroglossia. Heteroglossia is ever-present: At any given moment of its evolution, language is stratified not only into linguistic dialects in the strict sense of the word […] but also […] into languages that are socio-ideological: languages of social groups, ‘professional’ and ‘generic’ languages, languages of generations and so forth. (Bakhtin, 1981: 271–272) It is important to note here the link between language and sociohistorical concerns. Language is not simply a collection of neutral meanings waiting to be used; it always arises in particular situations to accomplish particular social ends. Indeed, a Bakhtinian perspective foregrounds the social meanings produced by language in use (and ‘backgrounds’ the referential ‘content’) (Bailey, 2012). The notion of heteroglossia has been widely taken up, particularly in the sociolinguistics of multilingualism, in recent years (e.g. Blackledge & Creese, 2010, 2014; Blommaert, 2010; Makoni & Pennycook, 2007). In this work, language use is understood in terms of repertoires of language practices on which speakers draw to suit the requirements of a particular situation. These

A Bakhtinian Perspective on Language and Content Integration

107

requirements include those related to the work of social relations, of identity construction and so on. Third, for Bakhtin (1981), language is tension filled. Specifically, Bakhtin identifies a tension between heteroglossia and what he calls unitary language, which refers to the idea of fixed forms of language. Bakhtin characterises this tension using the metaphor of centripetal and centrifugal forces. Centripetal forces push towards a centre and so represent the pressure towards uniformity in language. Of course, some degree of uniformity is necessary. Centripetal forces may include institutional and political forces, such as school language policies or the official languages of nation states (Duranti, 1998: 76). But they also include more local, everyday forces, such as the tendency of speakers to notice or correct deviations from a presumed standard form (which they often call ‘errors’). Centrifugal forces push away from the centre, representing the production of heteroglossia. Centrifugal forces tend to be much more local, arising from the diverse social languages within any speaker’s repertoire and the relations between different speakers’ repertoires, and hence are more likely to include the languages of marginalised speakers (Duranti, 1998: 76). The key point for Bakhtin, in the metaphor of centripetal and centrifugal forces, is that the two forces are in tension. Both forces are present in and shape every utterance, representing in some sense societal tensions between institutional norms and individual agency. Finally, for Bakhtin (1981), language always entails an encounter with otherness, for which he uses the expression ‘the alien word’. This otherness is derived from the sociohistorical dimension of language. Any utterance is not simply an expression of an individual’s idea; it expresses a host of ‘other’ ideas that derive from preceding usage and must be understood in the light of preceding utterances: Language is not a neutral medium that passes freely and easily into the private property of the speaker’s intentions; it is populated— overpopulated—with the intentions of others. Expropriating it, forcing it to submit to one’s own intentions and accents, is a difficult and complicated process. (Bakhtin, 1981: 294) Moreover, since we can never escape from being in relation to the language that precedes us, this language in some sense defines who we are. It is in this sense that Wegerif (2008: 353) proposes, ‘for each participant in a dialogue, the voice of the other is an outside perspective that contains them within it’. There are several implications of these ideas for the notion of language and content integration. First, content is not simply constructed by individuals through participation in classroom processes. If meaning is understood in terms of socially situated response, content must be

108

Part 1: Curriculum and Pedagogy Planning

understood as irreducibly social in nature, jointly produced by participants. Heteroglossia is ever-present. In particular, it is present in content-area discourses. Heteroglossia is present, for example, in mathematical language: there is no one mathematical language; there are many mathematical languages – those of different domains of mathematics, of different communities of mathematicians, of school mathematics, of textbooks, of mathematics teachers and so on. Moreover, from a dialogic, relational perspective, learning is not a process of replacement (Wegerif, 2008) in which mathematical language gradually replaces everyday forms of expression; both will be present and recognised through their relation to each other (Barwell, in press). Interaction in CLIL, immersion or other forms of second language classroom (or indeed any classroom) is shaped by the tension between centripetal and centrifugal language forces. Hence, on a general level, the institutional requirements to learn a particular language and to learn the language of a particular content area are in tension with students’ and teachers’ various diverse ways of participating. Students may bring to their participation in classroom interaction diverse, overlapping, but non-identical, repertoires of spoken and written language. Their task is to expand these repertoires to include content-related practices from the target language. In content and language classrooms, then, students must expropriate the language of content to make meaning in acceptable form. Finally, both the target language and the language of the content area are fundamentally forms of ‘alien word’ for the students. This otherness precedes them. Students must engage in dialogue with this otherness and find ways to make the words their own. Indeed students do this all the time. The centripetal aim of mathematics teaching, however, is that students learn to do this in institutionally sanctioned ways. The relationship between the students’ intentions and those alien intentions that already populate the language available to them is necessarily dialogic. In it is this sense that, in Wegerif’s terms (2008), as students struggle to expropriate content language, they are ‘contained’ within this language; in some sense, it defines who they are and what they may be able to say. At the same time, the teacher is contained within the students’ language. The integration of language learning and content learning, then, unfolds in the dialogue between these outside perspectives that contain the students and the teacher.

Two Examples The two examples discussed in the rest of this chapter come from an ethnographic study of the learning of mathematics in second language classrooms in Canada. Canada has two official languages, English and French, as well as First Nation, Métis, Inuit and immigrant community

A Bakhtinian Perspective on Language and Content Integration

109

languages. Data were collected in four different classroom settings representing four different second language education contexts. One of the goals of the study is to investigate the relationship between forms of interaction in second language mathematics classrooms and the particular second language context of the classroom. For this reason, sites were chosen to represent a variety of forms in second language classrooms. The data referred to in this chapter come from two of these settings: • •

A sheltered ESL mathematics Year 5–6 (ages 10–12 years) class in an Anglophone elementary school. The class mainly consisted of Creespeaking students. A Year 5–6 (ages 10–12 years) classe d’accueil in a Francophone school in an urban area in the province of Quebec. The classe d’accueil is a sheltered class for new immigrants who do not speak French on arrival.

Although the official classroom language is different in the two examples, the institutional and policy contexts are of greater significance for what happens in each case. In the first setting, the students’ language, Cree, has little official status or value (despite being a Canadian language that predates English and French in North America). In the second setting, the purpose of the class is to ‘françicise’ the children: that is, to teach them the official language of the province of Quebec. In each case, I conducted classroom observations, audio recordings of classroom interaction, as well as interviews with some of the students and each teacher. I also collected examples of student work and other relevant artefacts from the classroom. I spent a school year visiting the first setting and three months visiting the second. To illustrate how a Bakhtinian perspective is useful in understanding integration, I present two episodes.

Episode 1: Cat and mouse The first episode comes from the classe d’accueil. French was the language used for the official business of the class, including during mathematics, as well as for interaction involving the teacher. For the teacher, the purpose here was not simply that the students learn French, but that they learn to think in French (‘le réflexe de réfléchir en français’). Learning mathematics is secondary to, and shaped by, this first goal. Hence, although mathematics was certainly a regular part of what the students studied, less time was devoted to it than is the case in mainstream classes. Moreover, even during mathematics lessons, much attention was given to the language of mathematics, including vocabulary (e.g. polygon, angle, convex) and pronunciation (e.g. polygon, angle). One strand of work focused on aspects of geometry, including the concept of convex and non-convex shapes. In a convex shape, all points

110

Part 1: Curriculum and Pedagogy Planning

inside the shape can be joined by a straight line to any other point inside the shape, without the straight line passing outside the shape at any point. A rectangle is convex; an L-shape is not. To explain this idea to the students, the teacher asked them to imagine a cat chasing a mouse inside the shape. If the mouse has nowhere to hide, because the cat can always see the mouse, then the shape is convex. This strand of work included an activity in which the teacher sorted different geometric forms, first collectively, organised around the blackboard, and then in small groups. For the latter activity, the teacher distributed bags of prepared shapes and asked students to classify them into two groups and explain their rule. She asked a student to explain the classification of one group who had in fact separated their shapes into convex and non-convex piles. At this point, the word convex had not been used (S refers to specific students; see Appendix for transcription conventions): Teacher:

S17: Teacher: S17: Teacher: S17:

ok les au:tres (.) regarde comment ils les ont classés et essaye de deviner comment ils ont fait pour les classer comme ça (…) ok est-ce qu’il y a quelqu’un qui a trouvé comment ils ont fait pour les classer? S17? er il a mis la boula comme ça il a mis aussi les choses comme ça les les choses comme ça rec-tangle le rectangle (.) c’est comme la même la même (.) la boule il est comme la même (…) la même grosseur? oui non pas la même grosseur la même c’est comme c’est comme co:mme je sais pas comment le dire en français (.) c’est comme la même chose

Translation: Teacher: S17: Teacher: S17: Teacher: S17:

okay everyone (.) look how they sorted them and try to guess what they did to sort them like that (…) okay is there anyone who has found how what they did to sort them? eer he put the ball-a like that he put also the things like that the the things like that rec-tangle the rectangle (.) it’s like the same the same (.) the ball it is like the same (…) the same size? yes no not the same size the same it’s like it’s like li:ke I don’t know how to say it in French (.) it’s like the same thing

A Bakhtinian Perspective on Language and Content Integration

111

Later work included students being asked to explain whether a given shape was convex or non-convex, as in the following exchange: Teacher: Student: Teacher: Student: Teacher: Student: Teacher: Student: Teacher: Student: Teacher:

donc c’est: (2) c’est convexe? (6) quand le chat et la souris ça peut arriver qu’ils ne se voient pas est-ce que c’est convexe ou non-convexe converse c’est convexe? garde ici si le chat et la souris leuh (.) ils se voient partout converse non: (.) convexe ok redis-le (.) non-convexe non (.) con (.) non-converse? con-vexe non-convexe? con (.) v-v-vexe non convexe ouais c’est pas un B ah tu comprends?

Translation: Teacher: Student: Teacher: Student: Teacher: Student: Teacher: Student: Teacher: Student: Teacher:

so it’s: (2) it’s convex? (6) when the cat and the mouse it’s possible that they can’t see each other is that convex or non-convex? converse it’s convex? look here if the cat and the mouse leuh (.) they can see each other everywhere converse non: (.) convex okay say it again (.) non-convex non (.) con (.) non-converse? con-vex non-convex? con (.) v-v-vex non-convex yeah it’s not a B ah you understand?

These extracts illustrate the heteroglossia apparent in the class. In the first extract, S17 struggles to explain her classification in French, mainly relying on words like ‘the same’, ‘like that’ and ‘thing’. This highly informal mathematical language is difficult for anyone else to follow, although presumably it makes sense for the student. This exchange illustrates the tension of centripetal and centrifugal forces, in that the student is required to speak in French, and so draw on certain parts of her language repertoire, while avoiding using other parts (her first language [L1], for example).

112

Part 1: Curriculum and Pedagogy Planning

Meanwhile in the second extract a student is, again, caught in the tension between his own way of expressing himself and the more standard mathematical French expected by the teacher. In particular, one student, who is from a Spanish-speaking background, has his pronunciation corrected by the teacher: her response to the student’s first enunciation shown in the transcript as ‘converse’ suggests that she heard the pronunciation ‘conberse’, and was expecting to hear ‘convexe’. She therefore moves to reiterating first the ‘x’ sound and then the ‘v’ sound in ‘convex’. The teacher ’s rehearsal of this word is an example of what Bakhtin (1981) calls ‘double-voicedness’. The teacher revoices the student’s utterance, reworking his pronunciation, shaping the student’s intentions with her own. There was a sense in this class, therefore, that heteroglossia was acknowledged, even as it was marked as undesirable. The teacher seems to be aware that students speak their home languages or other languages they have in common with their peers when she is not present or when she is elsewhere in the classroom. During an interview, she seemed to approach this fact light-heartedly while maintaining that the students should not be breaking the language policy: en classe je le remarque moins parce que ils le font plus quand ils savent que j’entends moins hehehehuen je sais que dehors euh ils parlent beaucoup (.) parce que les autres me le disent mais ils savent qu’ils ont pas le droit (.) [Teacher interview] Translation: In class I notice it less because they do it more when they know that I can’t hear them so well hehehehuen I know that outside euh they speak [their home languages] a lot (.) because the others tell me but they know that they aren’t allowed (.) [Teacher interview] For the teacher, students will take more care to speak French if they are aware that she is listening. Students understand the use of French somewhat differently. During a group interview, one student said that they sometimes find it difficult to communicate their thoughts to the teacher in just one language: quand on est ici (.) il oui on dirait que quand on n’sais pas comment expliquer à la madame on dit euuh on (.) un peu en français (.) un peu espagnol, on dit comme ohh non mais c’est en espagnol et pas français […] oui, on doit faire que la madame comprend [Student small group interview] Translation: when we’re here (.) it yes we’d say that when we don’t know how to explain to the teacher we say euuh we (.) a bit in French (.) a bit Spanish, we say like ohh no but it’s in Spanish and not French […]

A Bakhtinian Perspective on Language and Content Integration

113

yes, we have to make sure that the teacher understands. [Student small group interview] For the students, they need to speak French to their teacher so that she will understand. It is therefore clear that the interaction in this classroom is shaped by the tension between institutional and, indeed, political requirements that the students speak in French (they must attend the class because of Quebec education law) and the various language repertoires that they bring with them. These institutional and political requirements are reflected in the ‘policing’ of language use apparent in the comments of both the teacher and the students (see Amir & Musk, 2013). These brief exchanges illustrate how this tension shapes student learning. The students are encountering an alien word, not simply in the specific word ‘convex’, nor even in the distinction between convex and non-convex, but in the way of talking about and thinking about shapes in mathematics. The requirements of this alien word also relate to its location within French more broadly. Hence, in the exchange, one student’s non-standard pronunciation (‘converse’ or possibly ‘conberse’) leads to a repair sequence in which the teacher displays the interpretation that that the student has not understood the concept. When the student repeats ‘converse’, the teacher displays a new interpretation: the student does not have the right pronunciation and asks him to rehearse the desired sounds. The idea of ‘right’ pronunciation, of course, is a centripetal force. So pronunciation and a display of mathematical understanding are intimately interrelated. A dialogic view of language highlights two things about integration in this sequence. First, the relation between the students’ language forms and those used or introduced by the teacher is dialogic. The meaning of these forms is not inherent in the words; it arises in the relations between them. The distinction between converse and convex, for example, is only meaningful when both possible pronunciations are present. Similarly, the sequential occurrence of S17’s utterances about a set of convex shapes followed by the teacher’s use of the term ‘convex’ sets up a dialogic relation. The students’ words encounter the mathematical otherness of ‘convex’ and vice versa. Second, the language of convex and non-convex and the associated ways of talking and thinking do not replace the students’ informal forms of expressing these ideas. The latter are necessary and inextricably linked with the former. Thus, in developing this mathematical idea, both convex and ‘like that’ are necessary and both are present. This allows the students to expand their repertoire of language practices (learning language) to talk about and think about shapes in new ways (learning content).

114

Part 1: Curriculum and Pedagogy Planning

Episode 2: Odd and even The second episode was observed in the sheltered ESL class. As in the classe d’accueil, the teacher emphasised the importance of learning English as a goal for the class, making particular mention of mathematical vocabulary. She also mentioned the need to develop student mathematical skills. The activity I refer to as the mystery number game stood out as one in which both language and mathematics appeared to be particularly relevant. The activity consisted of the teacher choosing a number and then either giving a series of clues or inviting questions from the students. The students’ task was to identify the number, either based on the teacher’s clues or on her responses to their questions. In later versions, the students took the role of the teacher. The activity occurred on several different occasions, including the example discussed below, which was the first time the game was introduced. The extract that follows represents part of the interaction relating to the first mystery number. Teacher:

sh this is a (.) two: digit number (.) the first thing you should do is write down (.) a line cause you know it’s two: digits (2.0) ^okay^ (2.0) this numbe:r i::s and don’t scre:am it out if you know it just write it down (.) this number i::s an o::dd number which means that it’s (.) o:ne (.) three: (.) fi:ve (.) what comes next? [Teacher writes on the blackboard: “_ _” and then “ 1 3 5 ”] S1: don’t know Students: ^(laugh)^ Students: [seven (.) nine Teacher: [seve:n (.) ni:ne [(.) so it’s an o:dd number not even number o::dd number you don’t have to write that down just remember it o:dd number [Teacher writes on the blackboard “ 7 9 ”] Student: [odd number S1: so if you if you don’t get the right one what happens if you write any of those numbers Teacher: no you don’t need to write any of these numbers don’t (.) remember if you’re between five (.) five this way or five that way (.) okay? (.) you know it is o:dd so it’s one three five bla::::h it’s not even it’s not two four six eight ten (.) okay Various forms of this activity recurred and evolved, including the following extract from seven weeks later: Teacher:

o:k I have [a number written down ask me a question so you can figure it out for a prize

A Bakhtinian Perspective on Language and Content Integration

S1: S2: S3: Teacher: S1: S2: S3: Teacher: S: Teacher: S1: S2: S3: Teacher: S1: S2: Teacher:

115

[(...) four five is it a one digit number? no it is [no:t a one digit number [eight ten five yes ^(laughs)^ is it a three digit number? no it is no:t a three [digit number [I don’t know how to play this game [is it a two digit number? [is it a two digit number it is a two: digit number so there’s two numbers (.) ok two (.) [numbers [a::h is it less than fifty? it is less than fifty (.) less [than fifty

The mystery number activity was conducted almost entirely in English, underlining the centripetal forces for the use of English in the school. English is the only permitted language in the class. Indeed, a printed notice displayed in the classroom said ‘remember to speak in English’. As with the previous example, the singular use of English is as a consequence of institutional and political centripetal forces (for an extended analysis of the centripetal and centrifugal forces influencing this class, see Barwell, 2014). Centrifugal forces are represented by the students’ informal ways of talking about mathematics. I have included the two extracts to show how the students gradually take on the more formal language used by the teacher. In the first, it is the teacher who talks about the number of digits, for example. In the second, a student asks ‘is it a 1-digit number?’ The voice of the teacher is apparent in this second utterance. The process of socialisation can be seen at work within the first extract. The teacher glosses odd number with examples, explicitly asking the students to complete the sequence and then contrasting it with even number. The students participate in an expansion of odd number by completing the sequence of odd numbers: the teacher begins ‘one, three, five’ before explicitly inviting the students to continue the sequence, ‘what comes next’. The contrast pair odd number/even number relates odd number with its antonym. These kinds of patterns socialise the students into formal mathematical language and ways of thinking, not simply vocabulary or word meanings. Completing the sequence of odd numbers socialises the students into a commonly accepted meaning of odd number and a way of

116

Part 1: Curriculum and Pedagogy Planning

talking about odd numbers. It is important to note that odd number has a variety of meanings, depending on the context. It could refer to a strangely written number (odd=strange); any number ending in a 1, 3, 5, 7 or 9; or, more formally, to any number of the form 2k+1, where k is a natural number. Meanings for odd number emerge from the interaction: • • • •

one of the numbers one, three, five, seven or nine; the symbols 1, 3, 5, 7 and 9 (written on the blackboard); ‘not even number’; ‘not two, four, six, eight, ten’.

These meanings arise as responses by the teacher, both to immediately preceding utterances (many of them her own), and to the extended interactional history of her work with these students over the preceding several months. As such, they construct the students in particular ways. For example, they construct the students as not necessarily knowing what an odd number is, or as not necessarily recognising the term ‘odd number’; indeed, in a second language classroom, these two forms of not knowing (not knowing a concept and not knowing the term used for a concept) may well be conflated (see Barwell, 2012; Morton & Jakonen, this volume; Reis & Barwell, 2013, for more on the discursive construction of ‘not knowing’). The students’ task is to expropriate the language of odd numbers, including ways of talking about and discussing odd numbers, but this language comes already populated by the voice of their teacher, a voice, moreover, that inevitably constructs them as knowing some things (e.g. numbers) and not knowing others (e.g. odd number). There is no separation here between learning language and learning mathematics. The meaning of odd number for the students starts with completing the sequence by saying ‘seven nine’ in unison with the teacher and is thus jointly produced. Looking at the episode from a slightly broader perspective, it is apparent that the interaction offers the students ways to expand their repertoire of problem-solving language, not in terms of vocabulary but in terms of how one can talk about a problem in mathematics. These possibilities include writing down information, expanding a word (odd) into a set of specific possibilities (1, 3, 5, 7 or 9) and eliminating other specific possibilities (‘it’s not even it’s not two four six eight ten’). The second extract illustrates how their repertoire for participation in this activity has expanded. The students are able to pose appropriate questions as well as to logically sequence questions, such as when they eliminate one digit numbers and then three digit numbers and confirm that the answer is a two digit number. The meaning of these utterances derive from their place in the sequence of questions, but also from their place in the sequence over several weeks of playing different versions of the game. The students are starting to expropriate the words for their own ends.

A Bakhtinian Perspective on Language and Content Integration

117

On the other hand, this kind of mathematical thinking is representative of a particular form of mathematics that for these students may be seen as an alien word. They have little choice but to respond. As with the cat and mouse example, the dialogic view of language highlights two things about integration in this class. Again, the meaning of the language forms expropriated by the students is not inherent in the words; it arises in the relations between them. The meaning of ‘odd’ or of a question of the form ‘Is it an x-digit number?’ emerges through the exchange of responses between students and teacher. When students use these words, they carry to some extent the voice of the teacher and vice versa. Second, the new mathematical language used by the students does not simply replace their previous ways of talking. Rather, the students expand their repertoire of language practices in the context of the mystery number game and (learning language) to talk about and reason about number in new ways (learning content).

Discussion and Concluding Remarks Broadly speaking, language, for Bakhtin, exists in and depends on relations between sounds, words, speakers, genres and so on. From this perspective, language is always situated: each utterance is a response to previous utterances (and hence always historical) and is designed and shaped for an addressee, real or imagined. Content-area meaning, therefore, emerges as responses in dialogue. A couple of key issues arise from adopting this approach to examine integration. First, as the examples I have presented illustrate, the tension between centripetal and centrifugal forces is present in second language classrooms. The examples illustrate how this tension shapes the classroom interaction and hence the particular nature of content and language integration. In the first extract from the cat and mouse episode, the learning of mathematics and language is shaped by a concern for pronunciation. This attention to pronunciation may take attention from other aspects of the task. Furthermore, these language forces are not simply local concerns; they derive from a broader sociohistorical aspect of language. The expectation that students will learn a particular language or talk about content in a particular way relates to political choices and the history of the language in the content area. More specifically, learning and teaching content domains like mathematics are always circumscribed by both institutional (school, curriculum, policy) and disciplinary (mathematics, science, history, etc.) expectations. As a result, there is often a sense, at least for teachers, of driving learners towards particular content language, while eliminating or leaving behind institutionally undesirable forms. Moreover, to be successful (in school terms), learners are expected to develop proficiency in the required forms. The mystery number game, for example, must be

118

Part 1: Curriculum and Pedagogy Planning

understood in the broader context of a curriculum that values questioning, problem-solving and learning mathematics through discussion. There is evidence in the extracts that the learners develop suitable language practices related to these broader curriculum expectations. At the same time, however, for these learners, such language practices are in some respects ‘alien’. The assumptions of replacement can then be seen to be politically loaded. Pedagogic orientations towards replacing some forms of language with other, institutionally sanctioned forms, could amount to a silencing of learners’ voices and the imposition of preferred alternatives. Bakhtin, however, suggests that language is always shaped by a tension between these different positions – between, in this case, learners’ voices and institutional requirements. Indeed, the two examples illustrate how students do not replace old ways of talking with new ones. Both are necessary. Student understanding of content language derives from its relation to their own ways of talking about the same phenomena. For example, in the mystery number game, the students’ use of mathematical questions in the later extract is meaningful only in relation to their participation in earlier versions of the activity. No instance of language is entirely academic or entirely informal; in some sense it is always a bit of both. The immediate implication of this perspective on integration at the level of classroom interaction is the need to develop suitable approaches to pedagogy that work with language tensions and, in particular, with students’ voices and ways of using languages. Such approaches would not forego the idea that students should learn accepted forms of language in any given content area. Content language is necessary. A consequence of the Bakhtinian perspective I have proposed in this chapter is that students’ informal or non-standard language is also necessary, not least because notions like ‘formal language’ or ‘content language’ cannot exist without contrasting forms. Such approaches to pedagogy would, therefore, support learners to expand their content-related language repertoires and to develop awarenesses about which forms of language are used in different situations. This approach is inherently ‘integrated’, since learning language and learning content amount to the same thing. Distinctions between ‘language focused’ and ‘content focused’ are not necessary. Learning is understood as the process of student expropriation of content (or other) language, a process that requires students to grapple with the intentions that already populate this language. As such, the students are defined by the interaction in their classrooms, such as when in the mystery number game, the students are defined as not knowing what an odd number is, or when in the convex extract, the student is defined as not having correct pronunciation. These ways of defining students do not exist outside of language; they are a consequence of the sociohistorical and indeed the political dimension of

A Bakhtinian Perspective on Language and Content Integration

119

language in the mathematics classroom. It is in this way that the students are contained by their teachers’ mathematical language, just as the teachers are contained by the language of their students.

Acknowledgements The research project referred to in this chapter was funded by the Social Science and Humanities Research Council of Canada, grant number 410-2008-0544, and the ideas were developed in part through participation in the ConCLIL project. I am grateful to the students and teachers who participated in the research, and to Élysée-Robert Cadet, Maya Shrestha, Maha Sinno and Marc-Alexandre Prudhomme for their assistance.

120

Part 1: Curriculum and Pedagogy Planning

Appendix: Transcription Conventions: Bold

emphasis

(.)

pause < second

(2.0)

pause > 2 seconds

(...)

untranscribable

( )

encloses uncertain transcription

?

rising intonation

^^

whispered speech

::

elongated vowels

[

overlaps

Part 2 Participants

5 University Teachers’ Beliefs of Language and Content Integration in English-Medium Education in Multilingual University Settings Emma Dafouz, Julia Hüttner and Ute Smit

Introduction The Englishisation of university education has attracted a great deal of research attention over recent years, in part motivated by the growing numbers of programmes offered in English. Since the early 2000s, the number of English-medium programmes has increased tenfold, from about 700 to 7000 (Brenn-White & van Rest, 2012; Wächter & Maiworm, 2014). Besides this impressive growth, the use of English for university teaching also merits analysis from a qualitative perspective, as it is an indicator of the recent political endeavours of Higher Education Institutions (HEIs) to join the internationalisation movement and, at the same time, reflects the diversified communicative roles of English as a leading language not only of academia and international business, but also of the social media and the internet more widely. Correspondingly, the reasons for Englishising tertiary educational programmes are varied and combine enabling student and staff mobility, opening up access to international expert communities (academic and business) and entering into collaboration or competition with other HEIs (e.g. Gajo et al., 2013; Wilkinson, 2013). In other words, English is primarily used in the tertiary setting because of its function as academic lingua franca so that the rapidly changing groups of students and teachers can communicate despite their otherwise diverse linguistic repertoires (Björkman, 2013; Mauranen, 2012; Smit & Dafouz, 2012). At the same time, some tertiary institutions also introduce English as medium of instruction in programmes attended largely by home students 123

124

Part 2: Participants

with an upper intermediate level of English (e.g. Barnard & McLellan, 2013; Dafouz et al., 2014), resulting in bilingual educational scenarios with an explicit language learning endeavour. In lower levels of education, in contrast, English (or any other target language) complements the educational language shared by the participants. Using English is usually not a communicative necessity, but a didactic decision giving it a much clearer pedagogical purpose. Having said this, we fully acknowledge that there is no clear divide between tertiary and compulsory education and that, especially at the upper secondary level, school-based learners are also expected to come with certain levels of language proficiency, to interact with international guests or visiting students and to gain knowledge and expertise that might open doors to international expert communities (e.g. Hüttner et al., 2013). However, the times when secondary education requires the use of English as a lingua franca are surely much scarcer than at the tertiary level and form the exception, while they are the rule in most English-medium tertiary programmes. The specificity of university education is also reflected in the many labels that researchers have suggested over the years (Coleman, 2006; Hellekjær, 2010; Wilkinson & Zegers, 2006), of which probably Englishmedium instruction (EMI) and integrating content and language in higher education (ICLHE) are most widely used. These two acronyms, however, as argued in Smit and Dafouz (2012: 3–4), represent different orientations, with EMI adopting a discursive perspective and ICLHE a more pedagogically oriented one. In an attempt to take a wider approach and avoid former connotations or associations, this chapter will use a more recent coinage, English-medium education in multilingual university settings (or EMEMUS). This term focuses specifically on English as the medium of instruction in tertiary education, reflecting the indisputable international and communicational roles of English in academia, the internet, business and research while, concurrently, not ‘specify[ing] any particular pedagogical approach or research agenda’ (Dafouz & Smit, 2014). Furthermore, in our view, ‘[m]ultilingual universities are seen as sites where bilingual or multilingual education, whether official or unofficial, partial or comprehensive, pedagogically explicit or implicit, may be represented’ (Dafouz & Smit, 2014). Against this multilingual background, our study aims to examine lecturer conceptualisations of content and language integration and provide a better understanding of this complex and frequently implicit process. While teaching language and content concurrently arguably lies at the heart of EMEMUS programmes, this view is not shared by all university teachers. We would argue that this situation provides relevant insights into stakeholder conceptualisations and therefore gives due attention to perspectives, attaching little overt importance to the integration of

University Teachers’ Beliefs of Language and Content Integration

125

language and content for the overall process of academic learning. By choosing teacher beliefs as our analytical focus, we aim to provide an emic perspective into EMEMUS from one of the main stakeholder groups, and so concurrently acknowledge the importance of beliefs in shaping teacher practice and understanding of their practice. While interest in teacher beliefs is not new (see section below), extant research has largely taken the form of studies in one specific location. In the case of tertiary education, so far the majority of studies on perceptions of Englishisation focus on individual institutions (Dafouz et al., 2007; Fortanet-Gómez, 2013), countries (Serra, 2011) or specific disciplines, such as physics (Airey, 2012) or economics (Unterberger, 2014). While these studies provide an in-depth understanding of specific educational cases, we argue that there is value in studying multiple locations across Europe in order to establish to what extent diverse conceptualisations seem primarily context-bound (and thus views are shared among participants in one context) or participant-specific (in which case similar views might be found across contexts). More precisely, this contribution allows for a deeper understanding of stakeholder beliefs of language and content integration in two ways. Firstly, by using data from a number of European sites, we are able to gain insights into the similarities and differences conceptualised throughout a range of EMEMUS sites and go beyond local perspectives. The database employed consists of 18 interviews with teachers from four different HEIs, who, having been involved in teaching international student groups and/or their subjects through the medium of English, shared their ideas on, among other topics, the integration of content and language. Secondly, we will analyse the data with the help of the recently developed ROADMAPPING framework (Dafouz & Smit, 2014), which offers insights into the dynamic and complex co-construction of beliefs within and across university settings.

Theoretical Considerations The ROAD-MAPPING framework Given the diversity and multiplicity of HEIs across the European scenario, the implementation of EMEMUS is inevitably multifaceted and, thus, its conceptualisation presents a challenging task. Within this context, the ROAD-MAPPING framework was recently developed to offer a holistic and dynamic theoretical account of the core dimensions that seem to operate in EMEMUS (Dafouz & Smit, 2014). Drawing mainly on sociolinguistic, ecolinguistic and language policy considerations (e.g. Blommaert, 2010; Hult, 2010; Scollon & Scollon, 2004), this framework identifies six core dimensions, namely, Roles of English in relation to other

126

Part 2: Participants

languages (RO), Academic Disciplines (AD), (language) Management (M), Agents (A), Practices and Processes (PP) and Internationalisation and Glocalisation (ING). Without going into detailed descriptions, which fall beyond the scope of this chapter, working definitions for each of these dimensions are provided next. Following the order in the acronym ROAD-MAPPING, first, RO focuses on the multiple factors that interact when considering the functions of English within HEIs. Among these perspectives we may include, for example, language policy considerations in multilingual and multidisciplinary settings, institutional decisions linked to national or supranational strategies or educational reasons connected to classroom practices. The second dimension, AD, refers to the way in which academia in general and the different disciplines more particularly follow specific teaching and learning principles, develop certain academic and discursive conventions and implement different assessment methods, according to their respective epistemologies (Becher, 1989; cf. also Llinares & Nikula, this volume). Thirdly, M is concerned with language policy statements and declarations, which aim to influence language practices in society. In university settings, while these language policies may be issued by relevant institutions (such as ministries of education or university management), in actual practice these policies may have little impact on everyday classroom teaching. In other words, language managerial decisions are closely linked to agent beliefs and behaviour and thus need to be viewed in relation to one another. This idea takes us to the next dimension, A, which encompasses the social players engaged in EMEMUS. These participants may behave differently or support distinct strategies, depending on whether they act as individuals (i.e. teachers, students, curriculum developers, administrative staff, etc.) or as members of an institution (i.e. university management, faculty, etc.) or a specific division (i.e. departments, research projects, etc.). In any case, agent beliefs and their subsequent actions will vary depending on the roles they adopt, their professional concerns or their disciplinary backgrounds. Fifthly, PP refer to the actual teaching and learning activities that agents construct and that are constructed by specific EMEMUS realities. Finally, ING explain the different drives operating both at the macro and the micro levels of HEIs, their connection with international and local factors and how these drives often come into conflict and need to be taken into account. True to the dynamic and intersecting nature of the ROAD-MAPPING framework, there are also multiple sites of potential access to any data aimed at shedding further light on the conceptualisations of EMEMUS. For this project, we decided to choose A as a point of entry and, more specifically, content teachers (university lecturers) involved in Englishmedium education. By choosing this access point, we are focusing on one of the three overarching ‘dimensions of integration’ covered in this

University Teachers’ Beliefs of Language and Content Integration

127

volume, namely participant perspectives or ‘the who’ (see Introduction). In view of our research interests and data set, our analytical lens will complement A with ING and AD. Somewhat surprisingly, RO turned out to be a less central dimension and will thus not be used in the following. Most likely reflecting their content expertise, our interviewees used arguments linked to the diversified content of teaching rather than the diversity of roles languages can play. Finally, the investigative spotlight on teachers and their beliefs means that (classroom) PP and M decisions are not directly accessible and remain outside our analytical lens.

Teacher beliefs The study of beliefs in educational research has had a chequered history with initial research dating to the early 20th century, declining with the rise of behaviourism mid-century, only to regain interest from cognitive, sociological and psychological perspectives from the 1970s onwards (Thompson, 1992) and increasingly from a discursive perspective since the 1990s (Barcelos, 2003). The precise definition and delimitation of teacher beliefs is a contested matter, and we subscribe here to the very inclusive view of beliefs as constituting the complex cluster of intuitive, subjective knowledge about the nature of language, language use and language learning, taking into account both cognitive and social dimensions, as well as cultural assumptions (cf. Barcelos, 2003: 8ff). While the boundaries between such subjective knowledge or belief, and more objective, shared knowledge are far from clear-cut, we share Thompson’s (1992) understanding that the former are characterised by disputability and the latter by allowing for an agreed means of evidence-based procedures of assessing their validity. For the current investigation, the feature of disputability is encapsulated primarily in gaining stakeholders’ individualised perspectives on EMEMUS and it is precisely this disputability that allows for a better understanding of the elements that are fore- or backgrounded in participant views of EMEMUS. A number of research perspectives have been employed, but in this project we follow the contextual approach, which focuses on capturing beliefs as linked to specific contexts and as inherently dynamic and complex (Mercer, 2012; Mori, 1999) and so allows for an investigation of belief constructs without establishing a priori links to actions. This perspective also highlights the role of discourse as the locus of (co-) construction of these beliefs and not merely as a way of making these beliefs visible (cf. Kalaja, 2003; Kramsch, 2003), which aids in differentiating between disputable beliefs and (overtly) shared knowledge. The rationale for studying stakeholder beliefs, especially in educational endeavours, has its origin in (largely psychological) research highlighting

128

Part 2: Participants

the importance of teacher beliefs as interacting – in a complex manner – with teaching practice (e.g. Borg, 2006; Farrell & Kun, 2008; Li & Walsh, 2011; Phipps & Borg, 2009), and importantly in their role in supporting or hindering change in teacher learning and development (e.g. Basturkmen, 2012; Johnson, 1994; Mansour, 2009). The positioning of beliefs as an integral part of ‘extended language policy’ (Shohamy, 2006; see also Spolsky, 2004) provides further support for using such beliefs as an access point in this research (see also Gajo et al., 2013). Within content and language integrated learning (CLIL) and other bilingual forms of primary and secondary education, the role of beliefs in (re)shaping teachers’ professional identities to include both language and content emerges. At a general level, this research highlights changes in teacher self-perceptions on a journey towards a reconceptualisation as content and language teachers (Cammarata & Tedick, 2012; Moate, 2011). On a more specific level, Morton (2012) addressed the actual language knowledge and awareness of teachers, and Koopman et al. (2014) and Bovellan (2014) focused on participant views of the language teaching elements in their own classrooms. Hüttner et al. (2013) explored the relevance of teacher and student beliefs in the implementation of CLIL and found similarly that success and failure of the CLIL endeavour was constructed in stakeholder beliefs in the first instance. Our primary interest in this study lies not in the relationship between these beliefs and actual teaching practice; we acknowledge that there might well be a mismatch between these two. Rather, we aim to gain insights into the conceptualisations of the integration (or lack thereof) of language and content in EMEMUS in the minds of the university teachers involved. Although not a major focus of extant research in general, there has been recent interest in understanding stakeholder perceptions in tertiary-level English-medium education. Generally, such studies focus on the prevalent beliefs held by teachers of one HEI (e.g. Doiz et al., 2012; Fortanet-Goméz, 2012; Hu, 2014), of teachers working at HEIs in one country (Serra, 2011) or of one academic discipline (e.g. Airey, 2012; Unterberger, 2014). Notable for its broader focus is Jenkins (2014), who used a questionnaire study to tap into university teacher beliefs across 24 countries and various disciplines, addressing issues we would mainly classify under RO in the ROAD-MAPPING framework. Her findings indicate that respondents largely positioned themselves midway on a University–English–Ideology continuum, ranging from an expectation of total conformity to Anglo norms on campus to a total acceptance of diversity on campus (Jenkins, 2014: 162). The overall implications appear to be that for the majority of university lecturers in EMEMUS (native, standard) English is the natural choice of medium of instruction and any tolerance expressed towards deviations from native speaker English norms is one of practical

University Teachers’ Beliefs of Language and Content Integration

129

acceptance, rather than of paradigm shift. While Jenkins (2014: 132) did not address issues of language and content integration specifically, she does highlight issues related to support for students in EMEMUS, which most respondents felt to be less than adequate.

Methodology In addition to its theoretical value to this study, the ROAD-MAPPING framework and its main characteristics – the six dimensions, their complex interplay and EMEMUS as discursive construction – also form the basis of our research methodology. On a surface level, the point mentioned last translates into discourse-focused research methods, such as the qualitative content analysis of interviews. More fundamentally, however, the discursive view foregrounds and topicalises ‘the fluid and dynamic nature of relationships among discourse processes across dimensions of social organisation’ (Hult, 2010: 40). Put more concretely, this means that participants’ discursive constructions are understood as constituted by, and constitutive of, the respective EMEMUS scenario or context. When investigating such discursive constructions, it is thus paramount to widen the analytical focus by including the ‘context’ and treating it on a par with the construction itself. An aid to this end is the ROAD-MAPPING dimensions and their complex interplay, which, as will be shown below, allow for a multilayered, dynamic and holistic analysis. In view of our aim to go beyond the local nature of most previous studies on (tertiary) teacher beliefs (but see Jenkins, 2014) and complement their findings by offering insights into teacher beliefs on EMEMUS across sites, our corpus of teacher interviews consists of 18 interviews, which were carried out at four different HEIs: •

•

•

At the Finnish university (interviews 1–3), three multilingual Finnish lecturers in education and physics with first language (L1) Finnish were interviewed in relation to their teaching to postgraduate local and international students. At the UK university (interviews 4–8), five education and engineering lecturers, all but one monolingual English speakers, shared their thoughts on teaching mixed postgraduate groups of home and international students. At the Austrian tourism college (interviews 9–13), five multilingual teachers, all of whom had professional careers in international hospitality, related their experiences of teaching diverse hospitality subjects to largely international student groups in an undergraduate diploma carrying programme. Three teachers were L1 speakers of German, while English was the L1 of the other two.

130

•

Part 2: Participants

At the Spanish university (interviews 14–18), five mono- or bilingual Spanish lecturers with managerial duties revealed their views on introducing English-medium teaching to groups of predominantly home students in economics and business studies.

All interviews were conducted at the respective HEIs and in the default language of communication between interviewer and interviewee, i.e. either the shared L1 or, in case of different L1s, English as lingua franca. Consequently, the corpus is multilingual as it contains interviews carried out in English, Spanish and German (for more information on the interviews see Table A.1, Appendix). The interviews themselves were semi-structured in design (e.g. Dörnyei, 2007), which means that they were based on interview guides identifying topic areas and possible questions. Additionally, the interviewees were encouraged to introduce additional topics and ideas they considered relevant. A further specificity of the data pool is that the Austrian and Spanish interviews were originally collected in connection with earlier projects (Dafouz et al., 2007; Smit, 2010a). As the foci of these projects overlap largely with the present study, these interviews offer rich data on teacher beliefs in EMEMUS and are therefore used for the present study. In fact, the interview guides of the Spanish and Austrian studies were used as bases for developing the interview guide for the interviews carried out in Finland and in the UK. Motivated by the research interest in ‘tapping into lecturer beliefs about EMEMUS across various institutions in different European settings’, the interviews focused on the ideas, opinions, experience(s) and expectations of teachers who had been involved in teaching international student groups and/or their subjects through the medium of English. More precisely, the teachers were invited to share their thoughts and views on (a) internationalisation and language management; (b) teaching and learning in EMEMUS; and (c) the integration of content and language, which is the focus of this chapter. All interviews were subjected to the same analytical procedure, which followed established steps of qualitative content analysis (e.g. Cohen et al., 2011; Lichtman, 2013) and made use of the online qualitative research software Dedoose (www.dedoose.com). The set of codes was established on the basis of the interview guide topics, in combination with the ROADMAPPING framework, in particular the dimensions A, AD, ING and RO (Dafouz & Smit, 2014).

Findings and Discussion A clear finding emerging from the interview analysis is that the integration of language and content learning is conceptualised quite

University Teachers’ Beliefs of Language and Content Integration

131

diversely across sites (and participants). This finding is conceived along a continuum of perceived ‘similarity’ or ‘difference’ of this integration in EMEMUS and the teaching and learning of academic content through the discipline-specific L1. The labels ‘similarity’ and ‘difference’ should be interpreted in relation to how English-medium education is constructed as similar or different from traditional or long-standing monolingual education. In other words, whether the lecturers interviewed regard this experience as comparable to their regular teaching practices (i.e. teaching through their and their students’ L1, usually the national language) or whether differences emerge as a result of a change in the teaching and learning conditions. What is also worth noting is that some participants overtly or implicitly challenge the relevance of an integration of language and content. This can take the form of negating any differences between EMEMUS and other settings, or of conceptualising specific disciplines as numerate only (and hence downplaying the role of any language in it), or finally, of acknowledging integration only implicitly through a focus on related peripheral issues, such as the added effort if integration is considered as inherently difficult or the related cultural issues when dealing with an international group of students. We consider all of these positions in the analysis as they provide further insights into the perceptions of language and content integration, even if ‘by absence’ rather than ‘by presence’. In the following discussion, we will outline the positions expressed along this continuum with regard to three dimensions of the ROADMAPPING framework, i.e. A, ING and AD.

The continuum of similarity to difference in Agents (A) At the cluster of the continuum highlighting similarities between EMEMUS and L1 practices in university settings, we find an expression of beliefs that focus on an engagement in what is viewed as appropriate academic practices, which are considered as shared by all stakeholders, regardless of their status as L1 or second language (L2) speakers of the language of education. At one extreme, this can involve the denial of any significant differences at all between the two educational approaches. With regard to the language and content interface, a relevant variant of this position highlights the similarity of learning academic language practices, in conjunction with relevant academic content, among both L1 and L2 speakers of English. This position is expressed solely by participants at the UK site, who appear to envisage a type of ‘community of practice’ (Lave & Wenger, 1991) into which students (of all backgrounds) are socialised. The focus of these comments lies on the outcome (achieved or desired) of student learning in terms of

132

Part 2: Participants

discipline-specific academic language use. In general, most teachers agree that the process of acculturation requires time (and at least some support) for students to develop the adequate skills for university acculturation. Importantly, however, this is not considered to be distinct between L1 and L2 English speakers. What is noticeable is that this view of academic acculturation implies little to no agency on the part of the teachers (and to some extent also of the students) with regard to any integration of language and content learning; arguably, this limits teacher responsibility for facilitating the process of learning both L2 and academic content. A clear example of this acculturation concept of language and content integration can be seen from a UK respondent in Extract 1. Extract 1 they get initiated to a certain type of speak in a way and master students through attending the sessions, through reading research papers and books they are initiated in a particular kind of language, it’s the academic language of education and in a way they have to use that language in order to communicate meaning (UK 5) As regards the differences between EMEMUS and L1-based practices, the most commonly expressed beliefs cluster around three main themes. Firstly, the increased preparation time for teaching is mentioned, including the need to rely on extra support in preparation, often in the form of consulting perceived language experts among colleagues. In contrast to the beliefs clustering at the ‘similarity’ end of the A dimension, here we can see how agency is foregrounded in the actions that teachers need to take in order to make the content and language integration of their students’ learning effective. A large group of these beliefs do not highlight the integration of language and content as such, but focus on the prerequisites of such an integration, which encompass providing more materials to students and preparing classes in greater detail, including teachers’ work on language items they predict being needed and/or difficult. Thus, integration itself is not conceptualised in any detailed manner, but merely viewed as ‘hard’ and/or effortful. The second theme clusters around (perceived) difficulties of student learning and is intrinsically linked to the third theme of changes in teaching practice aimed at reducing these student difficulties. Most of the differences perceived are linked to a general view of, especially languagerelated, difficulties of students in coping with EMEMUS. These might relate to general proficiency or also to more specific elements of disciplinary language use, such as ‘subtlety and nuance’ as stated by one respondent (UK 6). However, the link between such language-related issues and other types of influences causing these difficulties is made explicitly

University Teachers’ Beliefs of Language and Content Integration

133

by several participants. Thus, when talking about academic writing, several respondents highlight different student cultural and academic backgrounds as potential, additional reasons for perceived difficulties, as can be seen in the following excerpt from a Finnish teacher: Extract 2 if a student comes from an African country, from a school system where you don’t write a lot, you more speak and perhaps the examinations are spoken situations, so you are not at all used to write - [...] so is it due to the language or is it due to the writing skills (FIN 3) Mostly, however, a description of these differences in the teaching situation is inherently linked to the changes made (or required) to alleviate them. Thus, the integration of content and (foreign) language is conceptualised as inherently difficult and lecturers construct themselves very clearly as having both the agency and the responsibility of making these adaptations in order to facilitate student learning in EMEMUS. Such changes can involve offering extra language support, reducing the speed at which lectures are delivered, provision of additional visual support, including complementing verbal explanation with diagrammatic representations, offering explanations of unknown words, of which the technical words are seen as easier to identify than the non-technical ones and encouraging (or requiring) students to read a lot of disciplinary texts. An additional perspective in this cluster of beliefs is to downplay the role of (any) language in a specific discipline, by focusing on numerical information, as outlined in the following quotation by an Austrian respondent: Extract 3 well, I can [get] round the problem because if I use numbers and (x) on the board that mean things I don’t have to use very many words (AUT 9) Some of the specific language adaptations that lecturers mentioned focus on providing greater clarity and avoiding any ambiguities, often in instructions or in exam situations, for example in: Extract 4 when I’m writing exam questions, I have to be very careful not to use phrases which might be um, I don’t know, double negatives or colloquial or ask questions in a negative way and I think really that- (.) that again that’s to the benefit of all of everyone particularly [...] it’s particularly

134

Part 2: Participants

important in exams because they can’t put up their hand and ask a question but I think, yeah I think that’s something that I should be doing even if I didn’t have any international students (UK 8) The extracts cited above indicate that, firstly, language and content integration is conceptualised by A in a complex manner and, secondly, language and content themselves are not viewed as monolithic blocks but as diversified entities. As these findings show, A conceptualisations of EMEMUS as similar to or different from L1-medium education are inherently linked to their practices; thus, if EMEMUS is conceptualised as different, most lecturers appear to introduce changes to their practice in order to alleviate these differences and create what is perceived as an ‘equal playing field’ for all students. Underlying such a conceptualisation is the view that integrating content and language is inherently difficult, while often the nature of these difficulties will not be further specified. In those cases where A highlight the similarities, practices that do not differentiate between students or between educational strands are suggested as self-evident.

The continuum of similarity to difference in Internationalisation and Glocalisation (ING) Within the cluster of beliefs highlighting similarity, participants express the view that English should be used as the exclusive means of communication at all times in a rather monolingual fashion. This is mostly observed in the case of the Spanish university where multilingual or multicultural practices are not usually implemented, most likely as a result of the institution’s shorter experience with international programmes and students. Extract 5 illustrates this situation: Extract 5 I don’t allow my students to talk to me in Spanish in class, other teachers do. Mine talk to me in English in the corridor, during office hours. If it’s a group in English, it is in English. If I was the only one who spoke English, how would they improve? (ESP 14) As regards the difference pole of the continuum, English is seen as the lingua franca, with other languages playing a role depending on factors such as: the number of competent speakers in the learner groups, the teacher’s ability to use other languages or the wider roles of the languages in question. This view is frequently expressed in settings with a longer tradition in multicultural and multilingual programmes. In Extract 6, set in the UK context, the use of other languages was regarded as positive when the respective languages were used among proficient speakers, thus helping students with their work, especially at the beginning of the programme.

University Teachers’ Beliefs of Language and Content Integration

135

Extract 6 often you’ll hear translation going on in the lecture or during something, you know, if they’re trying to work out something, there’ll be a lot of Mandarin and a lot of note-taking and their own notes are in Chinese (UK 6) The potential relevance of the students’ L1 in the learning process was also mentioned by other teachers when either observing students making active use of their L1 or when offering occasional translations in languages they are competent in themselves (see Extract 7). In the case of one of the Finnish lecturers interviewed there was explicit rethinking or reflection along this line (see Extract 8): Extract 7 Given that there are a lot of German speakers I will occasionally if I get a question somebody say what is that in German I will answer. Seems unkind not to and I will also do that for French, but that’s the end of it, because I can’t do any more (AUT 9) Extract 8 So also thinking that what needs to be in English and what can be in Finnish [..] so when because somehow appreciating their native language and their mother language that you must be able to learn and think some of the very basic and core things perhaps first in your own language (FIN 3) In addition, we find that the largely international profile of the students and teachers is seen by the A to have an impact on the academic content taught. Along this line, access to more international content and the inclusion or exclusion of different topics in the teaching curriculum are often mentioned: Extract 9 In accountancy for example or statistics, supply and demand are the same, mean, deviation... it’s international […] accountancy is absolutely inserted in the international context […] without lowering the level of demand and the content, there might be topics that could be reduced or reorganised (ESP 18) This curriculum adaptation seems to be a relevant factor across most disciplines and garners explicit mention by the social scientists among the interviewees. In our data, all of the lecturers in education, for instance, argue

136

Part 2: Participants

that their teaching of educational concepts requires in-class discussion, explicit contextualisation and terminological clarifications. Although the interviewees sometimes identify such student diversity as potentially problematic, they also see it as an enabling feature that allows educators to think beyond their own acculturation. This can even go so far as to rethink, as mentioned before, which content is most relevant for international learner groups, as in this case of educational management: Extract 10 it demands us to carefully think for example [..] of our curriculum and what we teach because our - for example when our students, and many of our students come from very poor countries and the basic mission and vision of our programme is that the students are then able to develop education in their countries so, so kind of how we teach the students whose country is in in a totally different situation as Finland kind of - we can’t be a country or an institute who says that this is the way how you should do it (FIN 3) Overall, then, the beliefs concerning the integration of languages other than English in the international classroom are clearly separate depending on the type of internationalisation experienced by the interviewees. Thus, while teachers with practice in multilingual groups usually understand that English functions as a lingua franca and recognise the relevance of other languages for the teaching and learning process (Airey & Linder, 2008; Smit, 2010a), the traditionally monolingual settings see this multilingualism quite differently and render the use of the students’ L1 as reflecting a number of difficulties. Among these perceived ‘deficiencies’, a lack of teacher or student language proficiency, inadequate materials or interferences with the L1 are mentioned. Especially in less internationally oriented contexts, even the use of the (shared) L1 or any other language than English can be conceptualised as a ‘failure’ on the part of the speakers or indeed the programme itself. ING drives also foreground the importance of curriculum adaptation as a result of a change in the language of instruction and the growing diversity of university classrooms. In this context, teachers coincide in the need to seriously rethink the integration of content and language for more inclusive teaching and learning practices.

The continuum of similarity to difference in Academic Disciplines (AD) As regards teacher beliefs on the integration of content and language in AD, some interviewees initially argued that English-medium teaching

University Teachers’ Beliefs of Language and Content Integration

137

does not change what they are teaching or how they are doing it. This similarity is advocated either because lecturers identify their teaching focus as remaining on content exclusively (see Extract 11) or because they regard the content of the disciplines as equally new – and difficult – for all of their students, whether L1 or L2 speakers of English (see Extracts 12 and 13). Extract 11 I think it’s really difficult to see those situations as kind of language teaching situations, because it’s more or less I see that it’s a content thing (FIN 3) Extract 12 I’m teaching, so that’s just a general principle I suppose I try to apply for whoever, knowing that even for a lot of, you know, our home-grown British students, a lot of the language and terminology is hard for them. So I just try to be not one of those academics who just talks in riddles (UK 6) Extract 13 I don’t know if it’s an education-specific thing but there’s a lot of specific terminology which is likely to be outside of their normal conversation of English, you know, so they might [be] at a very proficient level in terms of conversational English, but academic English is an entirely different thing which I think is where we come up against a lot of the problems: there’s loads of terminology and that can be really overwhelming. So it’s not just understanding how the terminology fits in but of course how that relates to the set of new ideas and understandings including different, very different philosophical positions on knowledge, you know, just fundamental stuff philosophy and all that. So yes there’s an awful lot to learn on that level I think (UK 6) Additionally, in the non-UK sites, this conceptualisation of similarity is often linked to views of English as the language of disciplinary knowledge, and hence the language in which the most specialised and updated content knowledge is transmitted. In this sense, some lecturers express that, since the subject matter knowledge they read and teach is already coded in English textbooks and materials, their teaching practices in the target language are often ‘the same’ as those in the L1. There are, however, also some positions expressed at the point of difference. The most extreme point of view taken in this respect, albeit rarely, relates to the choice of disciplines that can – or cannot – be taught

138

Part 2: Participants

effectively in English, or for that matter, in another language. In the case of Spanish tax law, for example, participants view this subject as impossible or not useful to teach through English, as it displays a specific lexicon in Spanish that does not travel well in another language and is tightly tied to national models that should not be assessed in another language. While assessment was not a central topic of the interview guide and was only explicitly addressed in one sub-question, the number of comments found in reaction to this question reflects its centrality. By and large, the responses reveal that, except for the tourism college education set in Austria, all institutions stand on the similarity end of the continuum and insist on using identical assessment criteria and practices for all students, irrespective of the language of instruction. Extract 14 I correct their spelling, given that if I correct accents and grammar in Spanish and don’t allow spelling ‘haber’ without the ‘h’, then I won’t permit spelling mistakes in English (ESP 14) On the difference side, teachers of the tourism college in Austria describe a flexible assessment system, with teachers being free in adapting their assessment to the respective student group. The only other adaptation reported by a teacher comes from the UK institution, where the testing procedure was changed for all learners in order to accommodate international students. Otherwise, the UK interviewees agree with all other non-Austrian teachers that assessment is not variable but unchangeable and consistent for everyone. Regarding assessing student work, an integrative understanding of content and language comes to the fore in preparatory practices offered, such as to encourage students to read intensively on the exam topic, to change the wording of exam questions to aid comprehension (see Extract 4) and to give workshops and specific sets of instructions on pre-assessment and post-assessment activities, such as providing feedback. Feedback practices seem to feature more in the UK interviews, arguably relating to the focus on providing supportive and formative feedback in that institution. In the Finnish context, preparatory help is offered by language experts: a split in teacher roles that is also visible in the assessment practices of content teachers (Extract 15). This finding connects with the complex profile of the A engaged in EMEMUS. Extract 15 when we are as the supervisors for the theses for example I’m not able to-to correct the language in the theses (FIN 1)

University Teachers’ Beliefs of Language and Content Integration

139

In sum, interviewee comments on preparatory as well as assessment practices reveal that they are generally aware of the integration of language and content when it comes to the high-stakes relevance of assessing their international students. At the same time, differences in their beliefs draw on, and thus perpetuate, their institutional assessment ideologies as well as practical circumstances and facts of language ability and professional responsibility. Going beyond the geographical sites, these differences in assessment beliefs might also reflect a more fundamental divergence in assessment ideologies between institutions. Thus, it may not only be a question of different national settings and their assessment cultures, but also of the contrast between professionally oriented vs. academically oriented sites. The tourism education college, for instance, educates directly for careers in international tourism management (see Extract 16), while other university institutions have the double function of catering for the professional or business world and also for the academic one. On a speculative note, it might be argued that the university teachers act on their understanding of more rigid requirements of the imagined academic communities, while the college teachers respond to the perceived flexible requirements of the world of work. Extract 16 I am not someone who says ‘there is an absolute fail and an absolute distinction’. Rather, I really feel that assessing is done in relation to the group and in relation to the learners’ abilities. This is so because I’ve experienced quite some surprises, you know. We’re educating for the world of work and at [student] reunions I could really see that some of the weakest [students] had made the most impressive careers; it simply took them some time to reach their full potential. That’s why I think that assessment is nothing absolute, but it has to remain relative, contingent on the student group, and there is then an average. (AUT 13) To summarise, the similarity of learning academic content is expressed by lecturers as a necessity for all learners, irrespective of their origin or L1. This learning of content is inextricably integrated with the development of their language competence and academic expertise, which will ultimately enable students to become familiar with the subject matter content and language. More specific differences emerge, however, when lecturers refer to particular academic disciplines both from the point of view of their distinct epistemologies (Winberg et al., 2013) and/or the disciplinary language used in each of these (e.g. the language of accountancy, of tourism management or of education).

140

Part 2: Participants

In these cases, the integration of language and content within such specificities needs to be seen not as a unified phenomenon but as one that deserves contextualised responses.

Conclusions In an endeavour to complement the proliferating studies focusing on the Englishisation and internationalisation efforts of individual HEIs, this contribution has argued for applying a wider analytical lens (Hult, 2010) by exploring EMEMUS. In view of the complex contextual specificities of any HEI, such a research undertaking needs to place centre stage the dynamic and complex nature of EME practices, made possible by applying the recently developed ROAD-MAPPING framework (Dafouz & Smit, 2014). With its dynamically conceived dimensions metaphorically pivoting around the discourses by which EMEMUS is co-constructed, the framework provides a conceptual basis as well as a methodological frame for analysis. Most importantly, the framework allows for research across different HEIs, which in this study are taken from different countries and academic disciplines, and represent different forms of Englishisation or internationalisation. The study focused on university teacher beliefs regarding Englishmedium education or, more precisely, on ‘language and content integration’ as the interviewees perceived it based on their own teaching experience. Reflecting the dynamic, complex and contextualised nature of beliefs, the data set, consisting of 18 interviews conducted in four different European countries, turned out to be a rich source of discursive constructions of the contingencies and affordances of content teaching and learning in English in international student groups. The findings became tangible in relation to the ROAD-MAPPING framework and in particular the three dimensions of ING, AD and A. To start with the central outcome, the educational practices in EMEMUS were compared with those in idealised monolingual education, i.e. when teacher and students share the same linguacultural background and use their default language of education, and constructed along a continuum of perceived ‘similarity’ or ‘difference’ between EMEMUS and monolingual practices. With regard to the kind and degree of ING experienced, the ‘similarity’ end point of the continuum can be equated with a strongly felt ‘monolingual habitus’ (Gogolin, 1994), i.e. the long-standing ideology of European universities to use a single language for education. Such a belief could be found at the Spanish HEI, which was also the site at the beginning of internationalisation, thus dealing with mainly local students. All the other HEIs had truly international student groups, correlating it seems with a growing ‘multilingual habitus’ in the sense that, generally, the students’ multilingual repertoires were acknowledged as relevant for their learning and academic success.

University Teachers’ Beliefs of Language and Content Integration

141

From the point of view of AD, the ‘similarity’ end point represents the belief that all students go through a process of acculturation (and contingent language learning) into academia in general and, more particularly, into one discipline such as engineering or physics. Towards the ‘difference’ end point, acculturation is singled out as a specific necessity for the international students, either for their own good (some UK voices) or to everybody’s advantage (other HEIs). At the same time and in relation to social sciences, especially education, Englishmedium education is seen as bringing along different subject topics or ways of dealing with them in class. This finding already shifts the focus to the A involved, and by that, given the data set, mainly the teachers. On the ‘similarity’ end, assessment was mentioned as the same for all students in all universities (but not in the profession-oriented tourism college), thus possibly hinting at differences between institutional, maybe also national values. More importantly, however, interviewees across all sites pointed out how they see EMEMUS requiring or affording different practices as regards, for instance, preparing classes, marking, reading, speaking, explaining or participating in class. While, from an applied linguistic perspective, it is surely reassuring to find such a level of awareness among content subject experts, a word of caution is due: this is a purely qualitative study and the interviewees, who volunteered to participate, are, presumably, more interested in the topic than many of their colleagues. Besides such a possible bias in teacher beliefs elicited, this study is also limited to teacher perspectives and we want to acknowledge that teacher beliefs should not be mistaken as representing those of other stakeholders as well. On the contrary, complementary research into student beliefs can reveal very different evaluations of EMEMUS practices (e.g. Baker & Hüttner, 2014). But even if none of these findings are (intended to be) generalisable, they are valuable for our concern in that the range of such ‘similar’ to ‘different’ beliefs is not only interesting in itself, but is also enlightening when interpreted with regard to conceptualising content and language integration. First of all, it needs to be acknowledged that the views foregrounding ‘similarities’ between English-medium and traditional teaching tend to downplay the relevance of the interplay of language and content, either by claiming that ‘language’ plays an unimportant role in the teaching and learning practices, or by arguing that it is equally challenging for all learners and thus a normal or natural part of university learning. While such a belief taken on broadly and unquestioningly would result in the denial of integrating content and language as relevant to the learning process, this is not the outcome of the present study. Instead of representing monolithic views on English-medium education, all interviewees revealed a range of beliefs that illustrate their general realisation that teaching and learning draw on content and language in intricate combinations. Each of the respondents constructed their teaching practices as contingent on content

142

Part 2: Participants

topics and language use, but also pedagogical needs, aims and methods. ‘Content’ comes across as composite, combining particular disciplinary topics, general academic cultural aspects and wider sociocultural knowledge; and also ‘language’ is not viewed monolithically. Possibly also triggered by the interview questions, a structuralist view of language is more prominent and materialises in teachers juxtaposing English, the respective national language and student L1s. At the same time, the post-structuralist understanding of ‘languaging’, i.e. ‘emphasizing the agency of speakers in an ongoing process of interactive meaning-making’ (García & Li, 2014: 9), can also be detected as, for instance, in the acknowledgement of students profiting from drawing on their plurilingual repertoires or in English functioning as academic language and as lingua franca for international communities. Overall, then, ‘language and content integration’ comes to the fore as a central, complex and dynamic concern, shared by all 18 university teachers interviewed, across their national, institutional and disciplinary contexts of EMEMUS. Finally, this study has been the first to apply the ROAD-MAPPING framework across EMEMUS sites. Much more needs to be done, such as the extension to the study participants hinted at above by also including students and their voices. We hope that the potential of the framework will be tested – and contested – in such and similar ways in further investigations.

Note (1)

Data coding was realised iteratively and in close collaboration among the three authors and Tarja Nikula, whose input at the beginning of the coding process is gratefully acknowledged. The jointly established ‘code-book’, including the codes, their definitions and example extracts was used for the excerpt-based interpretation.

University Teachers’ Beliefs of Language and Content Integration

143

Appendix Table A1 Information on interviews included in database Academic discipline

Gender

EMI teaching experience (in years=yrs)

Linguistic repertoire (self-reported)*

No

Country

1

FIN

Education

M

5–9

Multi (E=L2)

2

FIN

Physics

M

Unknown

Multi (E=L2)

3

FIN

Education

F

up to 4 yrs

Multi (E=L2)

4

UK

Education

M

10 yrs plus

Mono (E=L1)

5

UK

Education

F

5–9 yrs

Multi (E=L2)

6

UK

Education

F

up to 4 yrs

Mono (E=L1)

7

UK

Education

F

unknown

Mono (E=L1)

8

UK

Engineering

M

5–9 yrs

Mono (E=L1)

9

AUT

Tourism

M

up to 4 yrs

Multi (E=L1)

10

AUT

Tourism

F

up to 4 yrs

Bilin (E=L1)

11

AUT

Tourism

M

10 yrs plus

Bilin (E=L2)

12

AUT

Tourism

F

up to 4 yrs

Multi (E=L2)

13

AUT

Tourism

M

10 yrs plus

Multi (E=L2)

14

ESP

Economics

F

up to 3 yrs

Bilin (E=L2)

15

ESP

Economics

F

No experience

Mono (no E)

16

ESP

Economics

M

up to 3yrs

Bilin (E=L2)

17

ESP

Economics

F

No experience

Bilin (E=L2)

18

ESP

Economics

F

No experience

Bilin (E=L2)

* The interviewees’ self-reported language proficiencies are characterised by the following criteria: (a) ‘multi’: multilingual; ‘bilin’: bilingual; ‘mono’: monolingual; and (b) role of English as first language (L1) or as a second or foreign language (L2).

6 CLIL Teachers’ Beliefs about Integration and about Their Professional Roles: Perspectives from a European Context Kristiina Skinnari and Eveliina Bovellan

Introduction Integrating language and content is at the very heart of content and language integrated learning (CLIL). However, defining integration is problematic for both practitioners and researchers in the field because of the varying understandings and diversity of the practical realisations of CLIL in different contexts (Eurydice, 2006: 55; Nikula et al., 2013: 72). Therefore, it is important to research the very notion of integration. Our research intends to give an insight into what integration means to the practitioners of CLIL, more specifically to teachers. Studying integration and professional roles through teacher beliefs yields emic practitioner perspectives, still needed in CLIL research (Moate, 2013: 16). Teacher views of integration affect the ways they enact integration in the classroom and the understandings of their roles, choices and responsibilities. According to Hüttner et al. (2013: 275), beliefs are important contributors to how CLIL is defined and manifested. Teacher perspectives on CLIL have already been investigated separately in different countries (e.g. Escobar Urmeneta [2013] in Spain; Wegner [2012] in Germany; Moate [2011] in Finland), but we aim to find a fresh approach to the multifaceted phenomenon of integration through relating three European CLIL contexts: Austria, Finland and Spain. This study is socioculturally oriented in its basic assumption of human action as culturally mediated and always situated in a specific context, which means taking into account the teachers’ local and wider cultural backgrounds. We expect the three contexts to offer an interesting possibility

145

146

Part 2: Participants

for setting the data side by side to identify both shared and specific features of the different sites (see Nikula et al., 2013: 92). Although the flexibility of CLIL poses problems for CLIL research if findings from one context are uncritically transferred to another (Cenoz et al., 2014), we agree with Dalton-Puffer et al. (2014: 3) that as long as contextual variables are taken into account, this is a risk worth taking.

Teacher Beliefs and Professional Roles Beliefs can be defined as a complex set of variables based on attitudes, experiences and expectations. Thus, they constitute a basis for teacher action, reflect the nature of their instruction and guide their decisionmaking (Thompson, 1992: 138). When exploring CLIL teacher beliefs about integration, we focus on how they see the roles and relationships of language and content in CLIL. According to studies on secondaryand tertiary-level CLIL teachers, teachers view themselves mainly as content experts (Banegas, 2012; Hüttner et al., 2013). In contrast, teacher beliefs about language learning and teaching in CLIL vary considerably (Bovellan, 2014). Successful CLIL seems to imply that the content teacher involved is committed to language and communication (Marsh et al., 2001: 139). Since teacher beliefs affect classroom activities and strongly affect teacher behaviour and choices (Pajares, 1992: 326), investigating them is expected to help in explaining why CLIL is understood and implemented in so many different ways. Previous research has shown (e.g. Tan [2011] in a content-based instruction [CBI] context) that teacher beliefs form a crucial factor guiding teacher pedagogical practices when teaching through a foreign language (FL). Tan (2011: 328) suspects that this is due to the absence of formal teacher training on integrating content and language teaching. This may well be an important factor contributing to the strong role of teacher beliefs in CLIL. Furthermore, teacher beliefs are intricate in nature and their relationship with classroom practices is complicated (Basturkmen, 2012: 1; NegueruelaAzarola, 2011: 360). Barcelos and Kalaja (2003: 232–233) aptly describe beliefs as dynamic and emergent, socially constructed and contextually situated, experiential, mediated, paradoxical and contradictory. Finally, teacher beliefs are mostly organised; in other words, individual beliefs seem consistent with one another so that one idea about teaching cannot be changed without affecting another (Correa et al., 2008: 151; Kagan, 1992: 76). Sasajima (2013: 64) investigated Japanese English as a Foreign Language (EFL) teacher beliefs and found that CLIL has ‘a potential to influence teacher mindsets’. The complex nature of teacher beliefs as reflected in this chapter is illustrated in Figure 6.1 which is adapted from Bovellan (2014: 55).

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

1.Culture

147

2. educational background

experience of life

social background

experience as a learner

Teacher beliefs 5. attitudes assumptions expectations

3. prior knowledge

4. teaching experience

Figure 6.1 Factors influencing teacher beliefs about integration and their professional roles in CLIL (adapted from Bovellan, 2014: 55)

Figure 6.1 highlights the five perspectives on beliefs that are relevant for investigating integration and teacher roles in CLIL in this study: culture, educational background, prior knowledge, teaching experience and implicit assumptions, attitudes and expectations. The other areas mentioned in Bovellan (2014), i.e. experience of learning, experience of life and social background, remain background factors since this study is not biographically or narratively oriented but focuses more straightforwardly on teachers’ professional roles. Culture (1) is the major background premise for this study, not only because it is based on three different European contexts. Teaching is cultural practice and consequently all teacher beliefs are influenced by contextually varying cultural factors. Furthermore, teachers’ educational backgrounds (2) are shaped by culturally formed practices, such as teacher education in different countries and the contextspecific requirements for a CLIL teacher. Typically, CLIL teachers in secondary education are content teachers with varying knowledge of the CLIL language and FL teaching pedagogy. In his study on Spanish CLIL teacher language awareness, Morton (2012: 298) discovered that teacher knowledge combined representations of the role of language with practical knowledge of teaching, including knowledge of learners, subject and self. Teachers’ prior knowledge (3) also includes their knowledge of specific CLIL pedagogy. At the beginning of their CLIL career, in particular, many teachers confront challenges in teaching through a FL and have difficulties balancing between two foci: content and language (Mehisto, 2008: 113; Moate, 2011). According to research from different countries, CLIL teachers share similar experiences (4) about the challenges in their work, such as the feeling of inadequate FL competence,

148

Part 2: Participants

lack of time, suitable materials and collegial or administrative support, insufficient training and various linguistic or motivational struggles of the students. On the other hand, the length of teaching experience in CLIL does not necessarily explain how teachers view CLIL (Bovellan, 2014). Although teacher attitudes, assumptions and expectations (5) vary individually, they are situationally constructed and reflect the values of the communities they belong to. Besides beliefs related to teaching and learning, teachers also hold beliefs about themselves and their socially constructed positions (Cross & Hong, 2012: 958). These professional identities consist of internalised meanings and expectations linked to questions about who one is and who one wants to be in association with teacher roles (Korthagen, 2004: 81; Stryker & Burke, 2000: 298). Roles are positions in social structure, externally defined in relation to others by what one does or should do (Burke & Stets, 2009; Stryker & Burke, 2000: 289). Institutionally demarcated roles provide teachers with a predefined status which comprises rights, duties and obligations (Wright, 1987: 12). For example, there can be a split in the teacher roles of language experts and content teachers in CLIL (see Dafouz, Hüttner & Smit this volume). This role division may be productive in secondary education if tandem teaching is made possible (Baetens Beardsmore, 2009: 210). CLIL teacher roles may be unclear if there is a lack of knowledge concerning the aims of teaching and if teachers do not know what is expected from them (Banegas, 2012: 47). Despite their institutionally determined and contextually constrained professional roles, teachers are active individual agents who have varying possibilities to make choices concerning their work (see Eurydice, 2008). Embedded in personal understandings and definitions of their professional roles are ‘the root beliefs about what is meaningful and important in teaching’ (Mahlios, 2012: 20). Thus, when exploring teacher beliefs about their roles, we also approach their personal meanings and related professional identities to some extent. Our foci of interest in this study are teacher beliefs about CLIL and about their professional roles. These are more specifically illustrated in the following research questions: • •

What do teachers’ accounts of how they understand CLIL reveal about their beliefs about integration? How are teachers’ professional roles reflected in the interviews?

CLIL in Three Contexts The ConCLIL project has enabled data collection from three countries with diverse CLIL policies and repertoires. Our study focuses on CLIL through English which is clearly the dominant language used in CLIL in

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

149

Austria, Finland and Spain (British Council, 2012; Dalton-Puffer et al., 2011; Kangasvieri et al., 2011). In Austria and Finland, CLIL has been an expanding trend since the 1990s. In both countries, it is a bottom-up, local enterprise with no or little official support but a lot of freedom for schools and teachers to implement it (Dalton-Puffer et al., 2011; Kangasvieri et al., 2011). In the Austrian CLIL schools in this study, approximately 50% of teaching was conducted through an FL and 50% through the first language (L1), German. As a rule, applicants for the CLIL stream undergo an informal FL competence test. Some teachers are doubly qualified in both the FL and in the content subject but there are no formal requirements for the teacher language competence. Native speaker (NS) assistants are frequently present in the classroom as linguistic support. In Finland, CLIL has developed as a bottom-up system in big cities since the 1990s, initialised by the municipality or individual schools. Educational or administrative authorities neither support nor restrict the implementation of CLIL to a significant degree. CLIL in Finland varies from very light, ‘language-shower’ type versions to international classes where almost all teaching is conducted through English. The term ‘immersion’ is only used in Finland for Finnish/Swedish context. The students for CLIL mostly have to apply and pass a test to be accepted, but paradoxically, equality is emphasised in Finnish education. At the secondary stage, teachers are often qualified only in the content subject and requirements for their FL competence are not always fulfilled. The Spanish teachers participating in this study came from Andalusia where CLIL is implemented top-down (Lorenzo, 2010). It enjoys strong governmental support but also tight control. Unlike in the two other contexts, Spanish secondary school teachers are required to show their second language (L2) competence by passing official language tests for advanced learners (Eurydice, 2006: 43). In the ‘Language Rich Europe’ study, where 24 European countries were surveyed for their language education policies, Spain was the only country to report widespread CLIL at the primary stage (British Council, 2012: 40). Despite its frequency, CLIL in the secondary stage still seems to be under transformation in Spain and testing for CLIL is rarely done (British Council, 2012: 197). Andalusia has invested a great deal in preparing English language materials that fit the national curriculum. Additionally, not only NS assistants but also NS content experts are typical in Andalusian CLIL classrooms.

Methods and Data Previous studies on teacher beliefs have shown that interviews are especially suitable for studying beliefs (Kalaja, 2003), also in the CLIL context (Dalton-Puffer & Smit, 2013: 550). With underlying assumptions

150

Part 2: Participants

and experiences, beliefs are not a straightforward research object. Since beliefs are constructions of reality, ways of seeing the world, interviews are a natural form of exploring them. Also, teacher experiences and beliefs about their roles are displayed in the interviews. The data for this study consist of semi-structured interviews with 12 secondary school teachers from Austria, Finland and Spain who all teach their subject – science or history – at least partly through English. The designed foci of the study, teacher beliefs about what integration in CLIL is and about their roles and identities, were informed by Morton’s (2012: 298) finding that CLIL teacher knowledge about the role of language in CLIL is in relation to their practical knowledge of learners, subject and self, for example. As we first conceptualised integration as the roles and relation of content and language (Dalton-Puffer et al., 2013: 216), a decision was made to ask teachers about their experiences and views about CLIL in general and in their own content subject, about the role of language in CLIL and about their own role in teaching through English. In each country, the interviewers covered the same questions or topic areas: background information concerning the teacher and the school, teacher biographies, teachers’ primary and secondary identities and their understandings of learners and learning. The interviews were all conducted in English since it functioned as the lingua franca of the interview partners and allowed for a data set that both authors could work with. After audiotaping and transcribing the interviews, all data were analysed by content analytical methods: close reading, thematic analysis and comparison (cf. Dafouz et al., this volume).

Participants The 12 participants are introduced in Table 6.1. Since this is not an in-depth study of individuals and for reasons of protecting participant anonymity, we did not focus on participant background information but have listed their gender (female/male), the number of native (NS) and non-native (NNS) English speakers and the years of CLIL experience.

Table 6.1 The participants Austria (A)

Finland (F)

Andalusia, Spain (S)

5

4

3

(F/M)

3/2

3/1

1/2

(NS/NNS)

2/3

1/3

1/2

4–13 years

2–11 years

6–10 years

N=

CLIL experience

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

151

The abbreviation after each excerpt shows three facts about the respondent: (a) the country of origin, (b) whether the respondent is an NS or an NNS of English and (c) which content subject the respondent teaches through English. The abbreviation Hi stands for history which also includes civics. The abbreviation for science subjects (physics, chemistry and biology) is Sc. All the NS teachers also taught EFL. One of them had worked as an EFL teacher at a secondary school but at the time of the interview was teaching various content subjects at the primary stage.

Findings Five major themes indicating teacher beliefs about integration and about their professional roles emerged in the teacher interviews. The themes will be discussed in the following sections, beginning with teacher content orientation. The second theme deals with teacher orientation to language and language teaching. Closely linked to this is the third theme on subject-specific language. Fourth, teacher understanding of language as a tool will be explored and the fifth section is intended to formulate a bigger picture of integration, referring to it as more than simply combining language and content.

Teacher content orientation Generally, content goals were prioritised by the secondary school teachers, which reflects the predominant teacher belief that learning content is more important than learning FL (cf. also Morton, 2012: 214). This is not surprising since the teachers were mainly secondary school subject experts and content learning is often stated in the curriculum as the main goal (Davison & Williams, 2001: 62), as for example in the Finnish context in Extract 1. Extract 1 it’s very much like sort of content-driven, it’s like that they have to have the same knowledge at the end as the ones who are doing it in Finnish. (F2NS/Hi) What is considered to be ‘the same’ knowledge, however, may be open to cultural adaptation when teaching through an FL, as was declared by a Spanish history teacher: Extract 2 I tell you the truth, I don’t teach exactly the same (.) mm, syllabus when I am teaching in the Spanish or in English. I focus much more in the Anglo-Saxon culture and in Anglo-Saxon history. (S2NNS/Hi)

152

Part 2: Participants

This adaptation works in both ways so that culturally and linguistically foreign content may also be ‘domesticated’ to fit the local context and curriculum. A Finnish science teacher explained this as follows: Extract 3 I’m not sure if it’s even necessary for me to (.) to express the phenomenon or the (.) or the subject (.) in the same way that a: native speaker would, or how would these things be taught in UK, for example. I think it’s enough when (.) I (.) I do it (.) from my point of view, from my culture, from my language. (F1NNS/Sc) Unlike Extract 2, in Extract 3, foreign culture does not seem to be a natural part of CLIL since the teacher emphasises his ‘point of view’, his ‘culture’ and his ‘language’. Some teachers were not afraid of adjusting content goals if their personal beliefs about the goals of CLIL diverged from the official ones. In Extract 4, an Austrian teacher describes her autonomous decisions when choosing content. Similarly, the Spanish respondent below explains that ‘other considerations’ than content may ‘become more important’. Extract 4 I always ask myself, is this really important, do I have to teach it or do I just do it because it’s in the book or it’s in the curriculum […] it’s a loss of contents (.) but I don’t think that I lose important contents. (A4NNS/Hi) Extract 5 other considerations about the science, about the person, about the future, when you use or spend some of your time to work with communicative skill you have to take the time from other activity that you fully used to advance in the contents of, to get (.) mm, deeper in the, but […] the difference of contents, they are not so important, and become more important other considerations. (S3NNS/Sc) These ‘other considerations’ seemed to be at the heart of integration and CLIL in general, manifested as new goals and a new kind of educational thinking. We will return to these issues later in the chapter. Most often the teachers appeared to be fairly confident with fulfilling the content requirements. However, a few expressed their concern about not meeting the subject goals in CLIL. The following respondent from Austria, for instance, described this as a ‘small fear inside’ of her, caused by the extra time needed for ‘explaining some language issues’:

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

153

Extract 6 there’s always this small fear inside of me that because of the English they don’t have as much physics as they would usually have, I know this is just not right because I’ve seen it and I do (.) I do more or less the same in the CLIL classes than the non-CLIL classes so there’s not a reason but it’s this fear that probably content teacher sometimes have because they need to take some time for explaining some language issues. (A1NNS/Sc) This fear was mitigated by the teacher’s belief that the students gain something more valuable if they miss part of the content. Although teachers sometimes had to make compromises with content goals, all the teachers agreed their content expertise was not questioned. Being a content expert was reconfirmed as the main role legitimated by teachers’ institutional position, official training and the expectations of others concerning their work. However, as explicated in Extract 7, after making a clear statement about being a history or a science teacher, many of the teachers continued with additional or even contradictory explanations defining their personal goals and positioning of self. Extract 7 I’m teaching history, I am a historian (.) I’m not an English teacher, I have never studied to teach hist, um English […] I am fully convinced that if my teaching is useful it’s not because of history but because of English […] that sounds a bit (.) weird from a historian. (S2NNS/Hi)

Orientation to language and language teaching Content teachers were not always aware of the role of language in teaching and learning their subject. Some teachers described language as ‘a side-effect’, ‘by-product’ or ‘a spice’ that was transparent and problemfree and needed very little attention. Thus, they seemed to believe that language is acquired naturally through FL content learning. Consequently, the role of language was considered minor, the focus being on content. However, the ‘extra’ the language brought seemed to be problematic to define, as Extract 8 illustrates. Extract 8 -- not really teaching English, partly teaching English as well but it comes kind of […] a side effect somehow (F4NNS/Hi) This belief about ‘natural’ language learning was to be detected among the majority of teachers (cf. also Dalton-Puffer, 2011: 193). They

154

Part 2: Participants

believed language already exists in all content teaching and does not need special attention (see Extract 9). The everyday use of the FL was thought to naturalise its use and NS assistants, where available, were believed to support natural language learning. Extract 9 I think this sort of like two for the price of one type thing, the fact that they’re actually using the language that they’re learning actually in a sort of real way. (F2NS/Hi) According to the teachers, natural language learning was connected with easy-going and relaxed ways of learning, often in out-of-classroom contexts. This seems to resemble the ‘Net Gen’ natural way of learning: young people of the internet generation, who in their online networks have grown up collaborating, sharing and creating together (Tapscott, 2009: 137), learning English from the media, internet games and popular culture, for example (Luukka et al., 2008). As for teacher roles, all teachers positioned themselves as content experts and authorities. There were also teachers who felt that teaching through English did not affect their teaching at all. However, many of them shared the idea of ongoing negotiation of meaning and construction of knowledge in the classroom. Extract 10 English teaching is much more interactive than L1 teaching. I have a much closer contact, verbally at least, with my students […] because I have to help them and stop, I have to go slower, and in L1 I am not forced to do so that much […] they are asking more often, I help them more often, the way I teach is completely different. (S2/Hi) Most teachers had at some point thought about the role of FL in their lessons though they did not often talk explicitly about language awareness. When asked about the role of language in CLIL, a Finnish history teacher referred to learning the academic language of history (see Extract 11) with concepts such as ‘causes and consequences’, ‘reasoning’ and ‘explaining the reasons’. Extract 11 [History requires] thinking about causes and consequences and kind of, that kind of reasoning why did something happen, that is probably the most important question, explaining the reasons and so that can be done in any language. (F4NNS/Hi)

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

155

An Austrian CLIL teacher concluded that EFL teacher language awareness is higher than that of content teachers, even in their L1: Extract 12 They [physics teachers] just talk and talk and they don’t even know that this is a new word. If you teach in L2 you are more aware of that. (A1NNS/Sc) Although Extract 12 suggests greater language awareness of CLIL teachers, most of the data do not support this. On the other hand, many teachers said that they had chosen CLIL because they were initially interested in language issues and especially in the English language and culture. Some NNS teachers and all of the NS teachers made comparisons between the use of the student L1 and English, showing awareness of the different languages in the classroom. The teachers’ personal relationships with the English language and their experiences and confidence in using English may have affected their beliefs about language teaching and learning and the role of language in CLIL. For example, an NS teacher said that his own children had learned languages naturally in their bilingual family and was convinced that this would also work for other children in educational settings. On the other hand, a teacher not confident with English did not believe her students would learn English in CLIL lessons and was afraid she might even do some harm to their EFL development (see Extract 13). Extract 13 I’m not the language teacher so I have no idea of how to teach language, I maybe, maybe I do some harm. (F3NNS/Hi) This reflects a belief that FL teaching should be restricted to the FL classroom, which is in contrast to the idea of using language for communication in CLIL, pointed out by the majority of teachers. Their view is in line with Hüttner et al.’s (2013) recent research where both teachers and learners characterised language learning in CLIL as ‘doing’, a repeated practice with abundant FL use. A teacher’s idea of learners ‘communicating about science’ (A1 NNS/Ph) illustrates combining language and content in action (see also Extract 11). Many teachers place enhanced communication skills as central goals in CLIL. Some teachers were very concerned about their students’ FL learning and highlighted their responsibility to teach language issues. Teaching through an FL challenges content teachers and their conventional roles in the classroom. Marsh et al. (2001: 138) remark that CLIL will necessarily involve change that has to be accepted by the teachers. Although increasing

156

Part 2: Participants

awareness of language pedagogy is useful for CLIL subject teachers to support their students’ language learning (Koopman et al., 2014), knowledge in language pedagogy was not important to all teachers. An Austrian CLIL teacher described her role in combining language and content as follows: Extract 14 for me it was obvious combination of both [English and subject] I, I really, what I really appreciate are the, the means of, of language teaching which I can use in teaching biology […] I still teach biology and I, I, that’s my subject and the language is a means to teach […] I would never be an English teacher, not a language teacher. (A3NNS/Sc) From the interview, it is evident that this teacher had both interest and experience in teaching English; yet, she drew a clear line and excluded language teaching from her teacher role. This is obviously due to the teacher’s subject background, having a degree in biology but not in English. She said she would not consider herself a language teacher out of sheer respect for the EFL teachers in her school community whom she regarded as the real language experts. One NS teacher defined herself as ‘language sensitive’ and said that language issues were constantly at the back of her mind when teaching history. When she was asked whether her students might think differently of her because she also teaches English, she first said that her primary role, history teacher, is clear in the classroom but then hesitated: Extract 15 I think that they, they see me as both but I think one of the things because, I think that one of the things because I teach them English as well […] but that’s just because of that, it isn’t (.) because I’m teaching them English during the history lessons, it’s because they (.) or I don’t really know how they, oh yeah. (F2NS/Hi) This hesitation shows how difficult it is to generalise teachers’ shifting positions and roles. Teachers’ institutional roles seem to be very fixed at schools but in CLIL they might become confused and contradictory. Extract 16 I am probably more of an English teacher than I’m trying to be seen, I would like to be seen but I think I’m more of […] somewhere in between there. (A5NS/Hi)

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

157

Being an English teacher in this excerpt might refer to a traditional EFL teacher position and a way of teaching where the ‘teachers are tied to the books’ which was criticised by this teacher. The same teacher gave the impression that he was very much aware of the role of the English language and student language development by claiming that the students ‘should be aware of the language’ but ‘not think about the fact that they are doing something in a foreign language’. Language teachers and subject teachers were real or potential partners who, however, had very traditional institutional roles and tasks. This was evident, for example, in the ways teachers described who was allowed to assess the students’ content knowledge or language skills. A CLIL teacher may work with language in the classroom but in most cases is not involved in testing or evaluating students’ English skills. In sum, there were several reasons why the teachers saw themselves as responsible for student language learning. These included, for example, their language awareness and their confidence in using English. The teachers saw their primary role as that of content specialists in their respective subjects, whereas their orientation towards teaching language and adopting the role of an English teacher varied a lot.

Subject-specific language From previous research, we know that the academic languages of science and history differ from each other in subject-specific idiosyncrasies, which is also an aspect that teachers are at least partially aware of (Dafouz et al., this volume; Llinares et al., 2012: 109). All the teachers found common ground for their subject and language in vocabularies, terms and concepts (Extract 17), but only some teachers spoke about there being a language of history or a language of science (Extract 18). Extract 17 My responsibility is to give them words for a scientific report […] content-specific vocabulary. (A1NNS/Sc) Extract 18 There’s a certain, like language of history which probably you have to get like used to. (F2NS/Hi) Both history and science teachers talked about teaching key vocabulary and explaining concepts and terms when they described language use in their CLIL lessons. They also shared the difficulty of explaining abstract terms in L1 and L2, as demonstrated in Extract 19.

158

Part 2: Participants

Extract 19 it’s difficult even in Finnish, the language, the language, the terms, the language question, even so, it’s so theoretical (.) very often because you have to teach, like if I ask (.) people in the high street that, you know, gallup that what you, what the nationalism, liberalism, a: socialism means, I think that it would be very difficult for them, yeah, difficult terms, difficult concepts […] it’s more and more so abstract things. (F3NNS/Hi) However, as Dalton-Puffer maintains in this volume, it is the teachers’ task to make the content accessible to the learners by mediating between the expert knowledge of the teacher and the knowledge states of the learners. English is generally seen as the language of science (Huber, 1998). An Austrian physics teacher stated: ‘If you study science you need to learn English’. Also, the similarity of L1 and L2 science terms (German and English in Extract 20) supported student understanding of content. Extract 20 With the scientific language I never had a lot of problems, especially because it’s more or less the same […] if you know a term in German it’s the same in English when you talk about, like scientific things […] Latin and Greek origin words that was actually no problem […] in science, no matter if you take an English textbook or German textbook, it’s gonna be the same content […] global community just agreed on what kind of units you use […] the communication gets easier for us […] if we all use meters and litres [vs. local measuring units]. (A1NNS/Sc) This Austrian physics teacher further suggested that using English might even have advantages in teaching and learning new physics concepts because the students ‘know the word belongs to the physics content and they do not confuse it with the everyday term’. The exact nature of the scientific definitions and conceptual difficulty was emphasised by a physics teacher: Extract 21 I often talk about that if there are some definitions for or some phenomenons [sic], or quantities (.) the definition might, might be very long and hard to understand and I’m trying to explain that that is the way it has to be said because if it’s said in a shorter way (.) or in a somehow easier way it, it might change the (.) the point or the meaning. (F1NNS/Sc)

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

159

On the other hand, although concepts and definitions belong to the realm of language, for this respondent language did not seem to play a role in learning physics and chemistry: Extract 22 Maybe of course there are some areas in physics and chemistry that might be more challenging to teach, but I don’t think it has anything to do with the language. (F1NNS/Sc) Here, the teacher might also have entertained the thought that there is content beyond language, unlike the contemporary idea implying that language and content are mutually constitutive (Breidbach, 2006; Luukka et al., 2008). While history teachers agreed with their science colleagues that abstract terms and phenomena were challenging to teach and learn, both in L1 and L2, they expressed different opinions on the choice of language more generally: using English was sometimes seen as problematic and artificial, especially if they had to teach local or national history and civics, which are very contextually situated. Extract 23 especially in civics, there is so many new kind of things every year, like a: what means ‘uusioperhe’ [reconstituted family in English] or that kind of things and the new kind of words are usually, there is very often the other words which are not (.) even (.) there is no English words for them because they are, like Finnish social security questions. (F3NNS/Hi) The Finnish history teachers had to teach Finnish history in English despite their own preferences. Nevertheless, an Austrian colleague taught some parts of history and geography through L1, considering it more natural to talk about sensitive issues such as concentration camps through the mother tongue. Extract 24 I think we are an Austrian school and we have to know Austrian history in German, too […] when I talk about feelings, for example when I talk about concentration camps during the Second World War, I don’t do it in English […] German is kind of the feeling language. (A4NNS/Hi) In this respect, this study shares Wiesemes’ (2007) findings on teaching the holocaust through an FL in Germany.

160

Part 2: Participants

On the surface, the most obvious difference between disciplines is that of vocabulary which is often technical and abstract. As Llinares et al. (2012: 109, 111) argue, it is fundamental that CLIL learners understand the particular language of the subject which constructs the store of knowledge that makes up the discipline. Lorenzo and Dalton-Puffer (this volume) note that teacher knowledge of subject-specific language is implicit, in other words, their language awareness is latent.

Language as a tool Language as a tool is a common metaphor for integration in the CLIL literature. Language as a vehicle or tool for content learning, and not as an object as such, is also a curriculum issue (Finnish National Board of Education, 2004: 272). Interestingly, this belief is present in the teacher accounts, but its meanings seemed to vary. One perspective was that language is a teacher’s ‘tool box’, consisting of techniques applied for content and language teaching purposes. The idea of tools seemed to refer to two opposite interpretations: enhanced awareness of the role of language in CLIL, on the one hand, and language as a mere tool with content learning as the main goal, on the other. ‘Tool experts’ were believed to be the EFL teachers or NS assistants because they were considered to have the best English skills. They were called on for help when explicit language teaching methods were required: Extract 25 he [a native speaker colleague] always checks some vocabulary and then he had very good ideas about (.) teaching in general, not only CLIL, CLIL, in CLIL groups (F1NNS/Sc) The idea of language as a tool for teaching and learning became apparent when the teachers reported about the teaching and learning practices in the classroom in very concrete ways. For example, language was used as a tool to improve content-targeted communication, when the teacher corrected the language mistakes that hampered communication and meaning-making, and not only corrected language for language’s sake. The teacher, though not necessarily an FL expert, was expected to respond to emergent language issues in the classroom, explain content through language, open up new or challenging content matters by means of language or answer student questions as shown in Extract 26. Extract 26 if somebody needs help (.) for example have a question, I didn’t understand this, then I’m, I just (.) it depends on the person who I’m

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

161

talking to, I might continue, continue using English, explain it in different way, in different words but still using English. But sometimes, sometimes I just might change the language back to Finnish to make it (.) of course more easier […] if I feel that this person needs to be (.) a: spoken at this moment in Finnish rather than in English. (F1NNS/Sc) Many practical language-related tools that teachers used for content learning in CLIL were detected in the data, including the use of L1 and L2 in teaching, scaffolding the content with the help of, e.g. mind maps or summaries, taking teaching out of the classroom and teaching through interaction and communication. Extract 27 illustrates how a Finnish CLIL teacher uses notes, key points, a mind map, a summary, talk and discussion as language tools in her classroom: Extract 27 I (.) maybe give them some like notes and key points and maybe you know a mind map, all that type of thing which I suppose perhaps is maybe a summary of what they’ve got in their history textbook […] it’s not like a direct, direct translation as such […] we talk a lot and discuss a lot (F2NS/Hi) Being aware of or sensitive to language is a teacher tool that requires practice and knowledge of language. Yet, content remains the primary thing, as the following extract shows. Extract 28 although I’m aware all the time of course all the, to make it sort of language-sensitive and all that I’m aware of that all the time but it’s still the history, it’s like the prime thing. (F2NS/Hi) Alternatively, language in CLIL was seen predominantly as the learners’ tool. The teachers with this belief emphasised that language and content are learned in interaction, requiring abundant FL use from the students and elaboration of the content with each other. A Spanish teacher describes his focus on learner communication in CLIL as follows: Extract 29 It’s something I did in my English classes as well, just using English as a language, just communicating in English (S1NS/En+primary subjects) Many teachers believed that learning content through an FL had not only cognitive but also affective consequences. This finding is supported

162

Part 2: Participants

by Lorenzo and Dalton-Puffer (this volume), who argue that in history, CLIL students can learn otherness and rise above the boundaries of their L1. Our study shows that teachers believed CLIL enhances student respect towards different peoples and cultures and boosts their courage and willingness to communicate in the FL, and their motivation to learn content matters. The FL can also enhance student motivation to learn content that would not be as interesting if studied in L1, as seen in Extract 30. Extract 30 Some of them have said ‘oh, we don’t really like history but we like English’, so […] learning through English […] makes it more [sic] (F2NS/Hi) In sum, teachers interpret language as a tool in CLIL in both abstract and concrete ways. According to Mohan (2001: 107), ‘language as a medium of learning requires a functional theory of language and discourse’. In this study, Mohan’s idea, although not theoretically formulated, was represented in the CLIL teacher understanding of language as communication rather than a system. In a broad sense, the metaphor of language as a tool might be considered a synonym for integration. After all, integration cannot come about on its own, but requires support, or ‘tools’, such as scaffolding, communication in L2 or correcting language mistakes.

Additional benefits of integration CLIL teachers often found benefits and set goals beyond the obvious targets of learning content and language. There were more general pedagogical aims such as enhancing communication, widening perspectives, cultivating respect and understanding or employing new learning and teaching methods. CLIL was linked to a new pedagogy and learning culture. Some teachers maintained that CLIL might enhance student cognitive skills, for example, by offering them new schemes for thinking. Extract 31 I work in a skill that […] can produce the student the change of their mind and the connection their neurons of different problem that they have not, they have not to know to connect […] can produce a learning by chance […] better is the thing more difficult (S3NNS/Sc) It was not always clear how this development actually happened but many teachers had experienced better learning in CLIL than in mainstream classes. CLIL learner mental skills were understood to improve or transform because of the use of the FL in content teaching.

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

163

Extract 32 When the students learn in L2 it’s a bit more challenging thinking structure for them […] working in different languages can help your thinking skills […] somehow the patterns and ideas behind the language, you don’t get too stuck with the language (F4NNS/Hi) This teacher introduced two paradoxical ideas of the role of FL in learning: it both challenges and liberates learner thinking. Some teachers reasoned that the challenge of using an FL enhances understanding because it slows down the learning process and leads to a better focus and further elaboration of the topics to be learnt. Often, teachers also emphasised learner participation in classroom activity. Students’ greater language awareness may also affect their L1 academic language development. Many teachers reported that learning in English made the learners and teachers more aware of the relationship of learning and language. Extract 33 CLIL students explain better and are better in linguistic things in their L1 as well […] They are used to present their work, communication skills in L1 are also enhanced. (A1NNS/Sc) As seen in this example, teachers furthermore believed that CLIL and student communicativeness correlate to a degree. This was understood to help the students to prepare for the future and their potential international working careers with high demands for communicative competence with people from different cultures and backgrounds. The teacher ’s role in the internationalisation of the students and taking responsibility for the learners’ future skills was taken very seriously, particularly by the Spanish teachers. Teachers also said that in CLIL lessons they introduced new collaborative working methods and related skills such as research and presentation skills. An Austrian teacher claimed that CLIL students were thus ‘communicating science better ’ (A1) and explained integration as communication: Extract 34 this whole combining content and language is about having them talk or write or listen or whatever about science. (A1NNS/Sc) As regards more social ways of learning, teacher and learner roles were in transition. Some teachers considered themselves as much learners as they considered their students to be. They expressed solidarity towards their

164

Part 2: Participants

students but sometimes also their lack of confidence as users of English. New positions for teachers and learners were constructed: as co-learners, especially in learning English. Extract 35 at least at times I have a feeling that […] we are all kind of learning, I’m helping them to learn and we are learning together and finding out things together […] when it comes to L1, I’m a specialist, I know my subject-specific language […] but in English I would kind of […] we are studying together. (F4NNS/Hi) An important finding in this study was that not only the content and language but also people and cultures ‘integrate and get integrated’ in CLIL. By this, teachers referred to cooperation, support and affiliation within and outside of their school communities. Especially important were the NS classroom assistants and colleagues from English-speaking countries. Extract 36 The native teachers, I would not be as motivated as I am now if they weren’t there […] it brings you into talking English […] they bring in a lot of new culture, other aspects of teaching […] looking at different countries and see how they teach science helps you a lot to broaden your horizon […] they bring a whole new pedagogical culture. (A1NNS/Sc) A native English-speaking teacher working in Austria said that he brought a new, more communicative and cooperative culture to a school where teachers had a strong tradition of working alone. Teachers also talked about integrating with international or English NS communities outside school which resulted in the development of attitudes and the internationalisation of their students, and about cooperation with their colleagues, especially EFL teachers. They often complained about not having enough resources for cooperation and networking. All these perspectives convey the idea that CLIL teaching works best in communities with cooperation and support from others. For example, team teaching experiences were found to be very positive by the interviewed teachers and many of them would like that to increase.

Conclusions Teacher beliefs detected in this study imply that language and content integration and teacher roles in CLIL were understood in varying ways, yet there were similarities across the contexts as well. Often, teachers

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

165

explained their views not only by emphasising their national or local contexts but also by referring to personal and professional experiences, histories, goals and interests. In this context, looking at Figure 6.1, it becomes clear that the factors included in it (i.e. culture, educational background, prior knowledge, teaching experience and implicit assumptions, attitudes and expectations) are relevant for understanding the breadth of teacher beliefs. In general, the interviews imply that the teachers believed the integration of content and language to happen in the classrooms in a natural and dynamic way and therefore separating them from each other would be artificial. This is in line with the understanding of integration in CLIL as a process where language and content are entwined and cannot fruitfully be distinguished from each other (e.g. Barwell, this volume; Dalton-Puffer, 2011: 196). Although the primacy of content over language and the teachers’ content specialist roles were strong and evident, many teachers were willing to compromise content for other goals, often not directly related to content or language learning, such as increasing cultural awareness and respect. In comparison to their content teacher roles, the roles in relation to language teaching were much more varying, dynamic and even contradictory. The teachers also had different understandings of what language is, how it should be taught and by whom. To clear up the confusion of tasks and roles when managing the dual teaching requirements of teaching content and language, we maintain that teachers would arguably profit from in-service training on combining content and language (Llinares et al., 2012: 5; Morton, 2012: 116) and more tools and resources for cooperation, teamwork, curriculum work and materials design. Although new ideas and methods accompany CLIL, it is difficult to implement them because of inflexible institutional traditions and constraints. After all, teaching in CLIL, like all teaching, consists predominantly of teacher-fronted episodes (Llinares & Nikula, this volume). Teachers’ separate roles in different subjects are persistent and as one of the participants stated, ‘a CLIL teacher does not officially exist’ (A4NNS/Hi). Assessment, in particular, seems to engender problems with role boundaries. New understandings of language as subject-specific genres and a vehicle for communication expand the role of a subject teacher and, consequently, the role of the language teachers might be in transformation, as well. Understanding the academic language of a subject requires awareness of the fact that language instruction is an inseparable part of every lesson and every subject specialist is responsible for language teaching (Breidbach, 2006: 14). However, this does not seem to be evident for all of the teachers participating in this study and is undoubtedly worth further investigation in future research. Like Dalton-Puffer (this volume) argues, more cooperation with subject education specialists is

166

Part 2: Participants

also necessary to raise teacher awareness of the role of language in CLIL. Language awareness is increasingly highlighted in the international debate about education and attempts are being made to implement the concept on a practical level, e.g. in the new core curriculum in Finland, taking effect in 2016. Our study shows that teachers often struggle in trying to understand their dual role in CLIL. Teachers who had thought deeply about the essence of CLIL described it as a qualitative educational transformation which was a new professional challenge for them, including more than adding content and language elements together. Consequently, CLIL can break the cycle of fossilised routines and lead to new perspectives and positions (Breidbach & Viebrock, 2012: 12). For instance, many teachers reported having introduced new teaching methods which resulted in more active student involvement. Furthermore, a new teacher role as a co-learner with the students was welcomed by teachers in all three countries. Teachers seem to be aware of the various demands coming from diverse quarters (e.g. educational authorities, parents) and have their individual opinions and ways of responding to those demands. In Andalusia, in particular, the official language proficiency requirements set considerable demands for CLIL teachers. Although teachers considered their FL competence adequate for CLIL teaching, they had to prove it by attending a formal language test. Austrian and Finnish teachers, on the other hand, preferred courses on subject-pedagogical topics and communicative classroom language to training where only general knowledge about CLIL was offered. Our results indicate that CLIL teachers are familiar with at least some public theories and CLIL research (cf. Morton, 2012: 267). This knowledge enhances their understanding but also causes pressure if they compare their ‘real life’ situations to idealistic goals. CLIL teachers’ work at the secondary stage is often lonely. Local school-level decisions about what and how to teach in CLIL might facilitate their work, but would require resources for cooperation with colleagues and NS assistants and support from the school heads. CLIL teachers also expect more visible recognition of their work, which often requires extra time and effort. CLIL research has been conducted for quite some time and in many places worldwide, but the cross-fertilisation of findings is still lacking. Like Dalton-Puffer and Smit (2013: 556) argue, more dialogue between CLIL researchers and local practitioners from different international contexts is called for to enhance mutual gain. This study has shown that, despite the diverse contexts, secondary school teachers in different countries share many understandings of integration in CLIL and of their own role in it. This is confirmed by Dafouz et al. (this volume) who argue that the orientation of the educational institution (whether it is more or less academic) largely affects how teachers understand integration in CLIL.

CLIL Teachers’ Beliefs about Integration and about Their Professional Roles

167

The data of this study are rich and illuminating and the descriptions of the settings allow careful comparisons but, like in qualitative research in general, they do not allow simplistic interpretations or generalisations. In addition, exploring contextual factors in a more nuanced way could yield more systematic comparisons about the country-specific differences in teacher beliefs and roles. These research aims might be met by a longitudinal, ethnographic study from a community perspective that takes into account the voices of different stakeholders and explores the dynamics between them to combine the different focal points in a more holistic and complex picture of CLIL.

Part 3 Practices

7 Integration of Language and Content Through Languaging in CLIL Classroom Interaction: A Conversation Analysis Perspective Tom Morton and Teppo Jakonen

Introduction This chapter focuses on an aspect of integration which has not until now received a great deal of empirical attention, that of the interactional negotiation of language knowledge among peers as they carry out content tasks in content and language integrated learning (CLIL) classrooms. Generally, integration in CLIL is seen as primarily a teacher’s matter, in which aspects of language can be either proactively singled out for attention or dealt with reactively in response to student performance (Lyster, 2007). In both cases, it is the teacher who decides which aspects of language to focus on, and may do so according to a wide range of possible criteria, whether to do with the content-relatedness of the language item or with factors relating to second language (L2) development, such as aiding students to notice less salient and/or more recalcitrant language forms. The present study complements this work by adding another dimension, the usually much less visible work carried out by learners as they deal with their own knowledge about language issues during peer interaction. The chapter is thus positioned within the third dimension of the framework in this volume, that of the local level of classrooms, in which content and language integration is seen as a matter of in situ practices. Its focus is the interactive work done by students in CLIL classrooms as they discover, work on and try to resolve issues of language knowledge that come up as they work individually on content tasks. The tasks themselves are purely content-oriented, that is, they are not pedagogically engineered to provide for an explicit focus on language. This means that language 171

172

Part 3: Practices

issues that come up are those that present some kind of obstacle for individual learners as they work to complete the task. However, by asking their peers for help, they turn the individual problem into a collective one in which they not only attempt to resolve the knowledge gap but, in order to do so, have to negotiate precisely which aspects of knowledge are indeed problematic. Thus, we examine instances where it is a participant (in this case student) matter if, when, where and for what purposes language appears as an issue, and when it does, which aspects of language are focused on, and which resources are used in dealing with it, in the context of individual seatwork on a content task in a CLIL setting. The main theoretical and methodological orientation we use to throw light on these phenomena is that of conversation analysis (CA) as it is used to investigate issues of L2 learning, particularly in the work of researchers such as Lee (2010) and Sahlström (2011). However, in the spirit of integration which characterises this volume, we use such work to build upon and illuminate other concepts from the wider field of studies of L2 pedagogy: focus on form (FonF) and Swain’s (e.g. 2006) work on ‘languaging’. Both of these constructs have a central concern with knowledge, understanding and learning in relation to language, which are central to the pedagogic goals in (language) classrooms. We argue that recent CA work on epistemics (e.g. Heritage, 2012; Stivers et al., 2011), which takes an interactional perspective on issues of knowing and understanding, has much to offer in fleshing out these concepts by showing the actual work that gets done, turn by turn, as learners attempt to resolve issues of language knowledge in peer interaction in a CLIL setting. The chapter is organised as follows. The next section provides an overview of the issues around integrating a FonF in CLIL instruction, drawing mainly on concepts relating to communicative language teaching. This is followed by a section which looks at work on peer interaction on resolving language issues in L2 classrooms, notably Swain and colleagues’ work on language-related episodes (LREs), collaborative dialogue and languaging. This leads into a section in which we review work which takes a CA approach to L2 learning, particularly recent work on epistemics in interaction. We argue that such a perspective can enrich our understanding of collaborative dialogue around language-related problems by providing fine-grained analyses of the sequences of actions, and the multimodal (linguistic, textual, artefactual, embodied) resources used as learners work to resolve knowledge gaps. After briefly describing the research context and participants, the main middle section of the chapter contains a detailed analysis of two extracts from one episode in a history lesson in which the word ‘bruises’ became the focus of attention. The analyses highlight the interactional work that was necessary not only to resolve issues of form and meaning around this lexical item, but

Integration of Language and Content Through Languaging

173

also to establish mutual understanding about which of these aspects were actually the source of the problem. We conclude the chapter by identifying implications for ‘integration’ in CLIL research and practice, suggesting that while there is a need to cultivate affordances for the more contingent forms of content and language integration analysed in the chapter, it would also be possible to integrate these fine-grained multimodal analyses in research designs that address learner appropriation of pre-specified areas of language chosen for their content relevance.

Focus on Form and Languaging in CLIL Most advocates of CLIL and other content-based approaches such as immersion would claim as one of its major advantages that it provides ideal opportunities for language learners to engage in meaning-focused tasks in which the main purpose is communication around cognitively engaging content. In this sense, these claims mirror those which advocate ‘strong’ communicative forms of L2 pedagogy, such as task-based learning. However, in both types of L2 pedagogy, early enthusiasm that purely meaning-focused approaches would on their own lead to positive results in terms of attainment was replaced by a growing consensus that it was necessary to incorporate ways of directing learner attention to L2 forms, particularly those which appeared to be most recalcitrant even after relatively long periods of immersion. Within the broad field of contentbased approaches to L2 pedagogy, Lyster’s (2007) ‘counterbalanced approach’ is the most comprehensive framework to date for ensuring that learners are afforded opportunities to pay attention to L2 forms as they take part in meaning-focused content instruction. However, within the general field of L2 pedagogy, there is not a consensus as to how form and meaning can be best integrated. Leaving aside the possibility of ‘language bath’ approaches in which there is no attention to language forms, most other options can be characterised under the broad heading of form-focused instruction (FFI), which, according to Loewen (2011: 577), is ‘any instructional activity attempting to draw learners’ attention to specific linguistic items’. This can include a focus on formS approach, in which items from a grammatical syllabus are preselected for instructional treatment, or a FonF approach, in which learners’ attention is drawn incidentally to language items as they emerge during meaning-focused activity (Long, 1991). Loewen’s (2011: 579) more recent characterisation of FonF as ‘brief attention, either planned or incidental, to problematic language items within a larger communicative context’ expands it beyond purely incidental attention, allowing for preplanned items to be included. In terms of what ‘form’ can mean, there is quite a wide possible scope. Nassaji and Fotos (2011: 13), for example, give a broad definition

174

Part 3: Practices

encompassing ‘grammatical, phonological, lexical and pragmatic forms’. Despite this wide range of linguistic phenomena that can potentially be the object of attention, there is evidence from recent research (e.g. Costa, 2012; Morton, 2015) that in CLIL classrooms where FonF occurs it is mainly lexical in orientation, including attention to subjectrelated terminology. In the broader field of L2 pedagogy, lexical FonF is attracting growing research attention, as studies by Laufer and Girsai (2008) and Tian and Macaro (2012) attest. Indeed, Laufer and Girsai’s (2008: 695) definition of lexical FonF is relevant to CLIL classroom communication, as vocabulary items become the focus of learner attention if they are ‘necessary for the completion of a communicative, or an authentic language task’. In the data analysed in this chapter, it is indeed a vocabulary item, the word ‘bruises’, that is necessary for one CLIL history student to complete his authentic language task (writing a short essay about life in Tudor times), and which becomes the focus of attention of the learners seated at his table. It is interesting from a CLIL point of view that L2 pedagogy researchers such as Nassaji and Fotos see FonF as a matter of integration. For them, FonF is ‘an instructional option that calls for an integration of grammar and communication in L2 teaching’ (Nassaji & Fotos, 2011: 1). If we replace ‘grammar’ with ‘language’ and ‘communication’ with ‘content’, we can see that CLIL research on integration is really grappling with a wider problem that affects L2 education generally. CLIL research and practice need to address the same issues around integrating form and communication (content). Nassaji and Fotos point out that FonF can be implemented explicitly or implicitly, with or without prior planning, integrated with communicative tasks or carried out in separate formfocused activities. CLIL faces these questions but within a context of pressure to attain content-learning outcomes as well, which is not the case in L2 teaching. Within CLIL contexts, it would seem reasonable to go for ‘integrated’ FonF in which learner attention is directed to aspects of language as they participate in meaning-focused activity. Such FonF can be engineered into the task, perhaps by using ‘input-enhancement’ (Sharwood Smith & Truscott, 2014) methods such as underlining or putting words in bold font. However, it is not always possible to predict which language items may become problematic or salient in some way for learners, and thus there are strong arguments in favour of allowing learners, at least some of the time, to find and work on their own problematic linguistic issues, and collaborate to find solutions to them. These arguments can be based on individual factors, such as learners’ individual states of interlanguage development, or from a sociocultural perspective, in terms of their zones of proximal development (ZPDs) (Vygotsky, 1978) where they may reach new understandings of language systems through interaction with peers.

Integration of Language and Content Through Languaging

175

Alternatively, arguments can centre around the interactive dialogue which emerges as learners collectively engage in problem-solving around language items which are necessary for the completion of the content tasks they are involved in. Of course, from a sociocultural perspective, both are connected, as it is through participation in activity mediated by others, and by appropriate cultural tools, that learners progress to more advanced cognitive functioning. Such collaborative forms of dialogue around understandings of aspects of language have been a focus of research in L2 education by Swain and colleagues for over two decades. The earlier work focused on what they called ‘language-related episodes’, defined as ‘any part of a dialogue where the students talk about the language they are producing, question their language use, or correct themselves or others’ (Swain & Lapkin, 1998: 326). Later, this work took a more explicitly sociocultural turn, and the term used to describe these interactional practices was ‘collaborative dialogue’ (Swain, 2000). More recently, Swain and colleagues have been using the term ‘languaging’ to refer to verbalisations in which learners work on their individual or joint understanding of conceptual linguistic material, such as the meanings and use of passive constructions in French (Swain et al., 2009). Swain (2006) provides a detailed definition of languaging as The process of making meaning and shaping knowledge and experience through language. It is part of what constitutes learning. Languaging about language is one of the ways we learn language. This means that the languaging (the dialogue or private speech) about language that learners engage in takes on new significance. In it, we can observe learners operating on linguistic data and coming to an understanding of previously less well understood material. In languaging, we see learning taking place. (Swain, 2006: 98) While Swain and colleagues’ recent work has focused on languaging in contexts in which there is a very clear pre-specification of which aspects of language are to be focused on and this is done through tightly engineered tasks (e.g. the 2009 work on passives in French), the definition above allows room for less tightly pre-specified and designed contexts for languaging to occur. These can be the type of FonF episodes described above, which emerge contingently in the midst of content-related tasks. In fact, CLIL is a context in which there is potential for a wide range of languaging behaviour. As Swain et al. (2009) point out, languaging is an important part of the learning process, whether what is being learned is history, mathematics or French. In CLIL settings, we might expect learners, if instructional activities are organised in such a way to encourage it, to engage in languaging around both subject-specific content

176

Part 3: Practices

and language. However, the current study, as it is concerned with the role of FonF within CLIL, limits its scope to episodes in which language is the object of attention. In sociocultural terms, languaging can be seen as both an external process of overt social speech or ‘communicated thinking’, and as an internal process, speech for oneself or ‘dialogical thinking’ (Haenen, 2001). Haenen (2001: 163) points out that the benefits of communicated thinking ‘point to the importance of verbal interaction, small-group work and cooperative learning in the classroom’. Given these strong sociocultural arguments, a further step is to throw more light on how such work is done, by examining the interactive practices and resources used by learners as they engage in ‘communicated thinking’ or overt languaging. Doing so is not to negate whichever internal cognitive processes may be activated by engaging in such activity, but to bracket them at least temporarily by taking an agnostic epistemological stance towards them and seeing such ‘cognitive’ matters as learning, knowing and understanding as publicly displayed and played out by participants as they engage in interaction. To do so, we need to turn to conversation analytic approaches to issues of learning, knowing and understanding in classrooms.

CA Perspective on Learning, Embodiment and Understanding in Classroom Interaction Since around the turn of the millennium, the use of a conversation analytic methodology has rapidly increased in applied linguistic investigations of the connections between language learning and social interaction. A major inspiration for this line of work has been the seminal critique of mainstream second language acquisition (SLA) research by Firth and Wagner (1997), who used a CA lens to question what they considered as a too individualistic and mechanistic treatment of discourse in SLA. Instead, they argued for more sensitivity towards local context and recognition of participants’ own perspectives, something which CA with its ethnomethodological origins is well equipped to examine (for an introduction, see Hutchby & Wooffitt, 2008; Psathas, 1995; ten Have, 2007). Firth and Wagner’s (1997) critique has since then inspired a large number of CA-based studies addressing language learning, as part of a research area generally known as CA-for-SLA or simply CA-SLA (Kasper & Wagner, 2011; Pekarek Doehler, 2010). Taking into account participants’ own perspectives as they manifest in observable conduct helps to better understand the social base of theoretical constructs such as ‘languaging’. As it is, the kinds of conversational activities involved in ‘languaging’ are in fact left underspecified in

Integration of Language and Content Through Languaging

177

Swain’s (2006) definition of the concept. There is therefore a need for interactional research such as CA to investigate the contexts in which ‘languaging’ occurs (or does not occur). We suggest that useful questions guiding such an enquiry are, firstly, what students do when they are seen to be ‘languaging’ in instances of social interaction. Investigating the way specific actions, produced for the practical purposes of a particular, locally relevant activity, work as building blocks for languaging may help to identify social practices of a fairly general order that might afford learning. On the other hand, doing so may help us to understand the ways in which learning itself is social in nature. One example of the kinds of activities ‘languaging’ may sometimes involve is the routine task faced by teachers and students in language classrooms of identifying and working on distinct learning objects. As described by Lee (2010), an important part of pedagogic activities in an English as a Second Language (ESL) classroom is that participants examine each other’s talk as an analytic resource to determine which linguistic objects are in need of explicit teaching. In the empirical examples he provides, learning objects are found in turns-at-talk in which students orient to something as somehow unfamiliar, for example by asking for the meaning or a synonym of a specific word, or turns in which the teacher interprets items as possibly unknown to the students by providing clarification or checking their meaning. Markee (2008) has shown how such ‘learning objects’ may emerge and be addressed in interactions over a period of time. Markee advocates a learning behaviour tracking (LBT) methodology which allows such objects to be both tracked along trajectories of occurrence and analysed for how participants orient to, and possibly incorporate, learning objects into their communicative repertoires. A second question regarding ‘languaging’ to which a CA can contribute is a detailed investigation of the kinds of resources that go into the accomplishment of such interactional events, including the processes of knowledge formation and understanding put forward by Swain (2006). When interacting, we do not only rely on speech to create and convey meaning, but habitually also use our body and coordinate physical objects to get our business done. This has been amply demonstrated by what may be termed as ‘multimodal CA’ (see e.g. Deppermann, 2013). Even the most mundane classroom activities are produced by assembling together various resources, such as talk, gesture, textbooks and task sheets to name but a few. Accordingly, activities that we often take as produced by an individual student, such as explanations of scientific principles in a CLIL geography classroom (Kupetz, 2011), require the collaboration of several participants and the coordination of semiotic resources, such as gaze, gesture and facial expressions as well as classroom tools such as an overhead projector. Moreover, the way we use our body to communicate is not limited to relaying and receiving visual information, but also involves

178

Part 3: Practices

tactile (touch) and kinesthetic (ability to sense one’s own movement) senses and practices, which may be used to manage knowledge. Streeck (2013) has used the term ‘lived body’ to acknowledge the fact that bodies accumulate experiences with the material world and are used to relay those experiences to others. He suggests that this kind of ‘muscle memory’ of how to handle objects is involved in situations where word searches are accompanied with gestures that identify the searched-for referent, for example ‘pinching’ to invoke a roll of adhesive tape. What all this points to is a need to reconsider the formation of interactional meaning and perhaps even individual experiences as not confined to ‘language’ only, and take this into account in researching CLIL classroom interaction (for multimodal studies in CLIL contexts, see also Evnitskaya, 2012; Jakonen & Morton, 2015; Kääntä & Piirainen-Marsh, 2013). Language learning involves knowledge and understanding, two concepts which also feature prominently in Swain’s (2006) definition of ‘languaging’. Recent conversation analytic work on epistemics (Heritage, 2012; Stivers et al., 2011) has shown that knowledge can be seen as not only an intramental domain but also as embodied in and realised through interactional practices. It has also demonstrated that speakers’ understandings of what they know and take others to know, as well as the relative ranking of speakers’ knowledgeability are at stake in many different conversational practices (Heritage, 2012; Heritage & Raymond, 2005). The classroom is perhaps a prime example of an institutional context in which such interactional management of knowledge and knowledge states is pervasive, as activities are deliberately designed for the purpose of teaching and learning. For example, Koole (2010) describes the way in which teachers provide one-on-one instruction in mathematics classrooms and argues that they regularly call for the instructed student to produce two different types of displays of epistemic access regarding the object of the instruction, those of understanding and knowing. According to Koole, these displays represent different interactional objects, which systematically occur in different sequential environments. Thus, when teachers take a longer turn to explain how to proceed with a specific problem, they frequently end their turn by requesting the student to claim understanding of the just-prior advice (e.g. ‘Do you understand?’ followed simply by ‘Yes’). Alternatively, when the taskexplanation sequence has what Koole refers to as ‘dialogical organisation’ whereby the teacher guides the student through multiple question-andanswer sequences, teachers frequently request the student to demonstrate evidence that he/she knows the instructed object or can proceed with the task. Elsewhere, Koole (2012) argues that when teachers manage understanding problems, they tend to assume the existence of a specific problem and invite the student to align with it instead of inviting the student to formulate a problem. He also suggests that the potentially harmful effects of this practice may be exacerbated in multilingual

Integration of Language and Content Through Languaging

179

maths classrooms, because the student’s difficulty may not be directly related to the mathematics content, but may be a language issue, such as not understanding the language of the task rubric. Koole labels such language issues as ‘textual’ problems, distinguishing them from what he describes as ‘conceptual’ and ‘procedural’ problems. This is interesting for our CLIL context, as language issues may not simply be seen as impediments to ongoing content task work, but may be the instructional focus themselves, thus blurring the distinction between conceptual and textual epistemic issues.

Data and Participants The two data extracts analysed in this chapter are drawn from a corpus of video-recordings of CLIL history lessons taught to 14- to 15-year-old students in a secondary school in Finland. The lessons come from an optional course which students could elect to take within the school’s bilingual programme. The teacher was a native Finnish speaker who is fluent in English. As will be seen in the data analyses, although the teacher had established an official ‘English-only’ policy, the students switched between the two codes, Finnish and English, in their peer interaction. In order to pick up group interaction, three cameras were set up in three corners of the room and digital voice recorders were placed at each table. The data were transcribed using CA conventions (see the Appendix), and not only linguistic actions but actions using other modes of communication, such as gaze and gesture, were transcribed. Turns in which Finnish was used were given idiomatic translations (in italics). The analysis followed the CA methodological procedure (see Sidnell [2010] for a description) of building ‘collections’ of examples of interactional practices in which phenomena of interest to the study were found, in this case, interactional sequences where students worked collaboratively to resolve knowledge gaps. While some of these ‘knowledge gaps’ related to conceptual or procedural matters (see Jakonen & Morton, 2015), a considerable number concerned language issues, and these were the focus of our analyses. For space reasons, we analyse a single case here. This case serves well to highlight the range of phenomena that were found across the examples in the collection. It also shows the layers of complexity involved in reaching a shared focus of attention before a problem can be resolved.

Analysis We have divided the analysed stretch of interaction into two extracts to illustrate how CLIL students’ focus on linguistic form emerges from, and is embedded in, wider classroom activities. We also intend to illuminate how reaching such a (joint) focus can sometimes be a matter

180

Part 3: Practices

of interactional struggle that involves the management of epistemic and social relations. In the extracts, three secondary school students are independently writing their essays on schooling in Tudor times as part of an English-medium CLIL history course when one of them, Sakari, initiates talk and asks for the English spelling of a Finnish word ‘mustelmia’ (‘bruises’). As the extracts take place during independent seatwork, they do not comprise continuous conversation, but can be more aptly characterised as instances of incipient talk (see e.g. Szymanski, 1999), i.e. situations where talk may re-emerge and lapse as individual students solicit assistance from each other. Extract 1 ‘How do you write ‘bruises’?

01 Sakari

-> miten kirjotetaa (.) mustelmia how do you write bruises 02 Susanna ↑hmh {SHIFTS GAZE FROM HER TASK TO SAKARI

03 Sakari

mustelmia bruises

04 Susanna -> RAISES HAND, PALM UPWARDS, THEN PUTS IT ON THE TABLE 05 (2.0) 06 SMILES AND LEANS FOREHEAD AGAINST THE BACK OF HAND

07 Sakari ->

englanniks in English

08 Susanna -> RAISES HEAD AND SHIFTS GAZE TO SAKARI, SMILING 09 -> hh £mää en tiiä oota£ hh £I dunno wait£ 10 RESTS CHIN AGAINST HER PALM, GAZE UPWARDS 11 (1.5) 12 Sakari -> joku /bru:ses/ something like /bru:ses/ 13 (1.7) 14 Susanna-> joo yeah {SHIFTS GAZE TO SAKARI

15 Sakari

nii mut mite= yeah but how 16 Susanna-> =ei (.) ne o arpia no they are ‘arpia’ ((scars))

Integration of Language and Content Through Languaging 17 18 Sakari -> 19 20 Susanna 21 Sakari 22 23 Susanna 24 25

181

(1.0) °hhjoo° °hh yeah° (0.6) no onha yes they are @no hhjoo@ well yeah / yeah right 2.0 / AT 1.2, SU MOVES EYEBALLS TO LEFT TOP CORNER

(eiku) (no I mean) SHIFTS GAZE TO INKA; INKA’S GAZE ON SOFIA

->

↑mikä o mustelma englanniks ↑what’s ‘mustelma’ in English

26 (1.0) / INKA YAWNS AND SHIFTS GAZE TO SAKARI 27 Susanna -> black (lälläl) ((SMILES)) 28 Inka -> black £hole£ 29 Susanna [hhh hehe {GAZE TO SAKARI

30 Sakari 31 32 T

-> [eiku- (.) black dots no I mean (.) black dots (0.9)

{WALKS BEHIND SAKARI

33 Sakari -> 34 T

PLACES BOTH HAND ON THE DESK AND LEANS FORWARD

35 36

/bɹu:si:s/ (.)

->

what is (.) like (.) mustelma

37

>I’ll give you-
£sss(h)hh£ -> {GIVES ‘THE FINGER’ TO SAKARI {INKA SHIFTS GAZE FROM TASK TO SAKARI

41 42 Sakari

(2.2) / SUSANNA SHIFTS GAZE BACK TO TASK

/pruses/

The sequence begins as Sakari interrupts his independent task work to request, in Finnish, how the word ‘mustelmia’ (‘bruises’) is spelt at line 1. Note that line 1 does not detail the language of the referent word’s orthography, which in fact later turns out to need further clarification. The immediate response provided by Susanna is an ‘open’ repair initiator (Drew, 1997) which identifies trouble with the turn but does not indicate its nature. However, we can see that at line 3, Sakari orients to the problem being one of non-hearing as he simply repeats the target item. By omitting the reference to spelling, he, however, makes this second

182

Part 3: Practices

occasion of the request ambiguous as regards what he invites Susanna to do to the word; that is, line 3 on its own could be heard as a confirmation that Susanna is to provide an English translation for ‘mustelmia’. Instead of a display of knowledge, which would be the sequentially relevant action at line 4, Susanna makes an intricate series of hand gestures (lines 4–6), smiling as she does so (see transcript images). She begins by withdrawing from mutual gaze and shifting her gaze to the open palm she holds in front of her for some time (first image). However, she quickly places her hand on the desk and cracks into a smile (second image), only to subsequently lean her forehead against the back of the same hand (third image). The withdrawal of gaze from co-present parties, in this case by shifting it to the open palm, can indicate that a word search is to be conducted alone (cf. Goodwin & Goodwin, 1986). Moreover, an empty palm displays exactly that it is empty, i.e. that something is missing. Seen this way, Susanna’s subsequent smile and leaning of her forehead against the back of her hand have a somewhat self-sanctioning quality: she is treating the answer as ‘elementary’, something that she should know and be able to produce. However, by investigating Sakari’s reaction to clarify the language of the word (spelling) he is asking for, we can notice that it treats Susanna’s smile and gesturing as having conveyed an understanding that Sakari may be after the spelling of the Finnish language word, in which case the smile is open to being interpreted as ridiculing Sakari for presenting such a simple question. As the language of the knowledge object has now been clarified, Susanna’s claim of not knowing the answer at line 9 works to confirm to Sakari that ridicule was not ‘meant’ at lines 4–6. Immediately following, she also shows her willingness to help out with the problem by asking for more time and assuming a ‘thinking face’ (Goodwin & Goodwin, 1986; Hellermann & Pekarek Doehler, 2010), gazing up (lines 8–10, see image). As she is ‘thinking’ a possible response, Sakari demonstrates at line 12 that he in fact knows the English language word, ‘bruises’, whose correct spelling he is seeking. However, the way Susanna, following a fairly lengthy silence, receives the candidate word with an acceptance token at line 14, and does not address the spelling issue, indicates that for her the pursued knowledge object is still the English language translation of ‘mustelmia’ rather than its spelling. This is in fact what Sakari’s subsequent turn at line 15 also attends to by treating line 14 as a contribution that has failed to resolve his knowledge gap. Thus, at this stage, a misalignment regarding the nature of the knowledge gap has been made evident. At this point, Susanna changes her position and contests Sakari’s earlier suggestion of ‘bruises’ as the English language equivalent of ‘mustelmia’. She does this by presenting another Finnish word, ‘arpia’ (‘scars’) as the correct translation of ‘bruises’ at line 16. This action halts the search for the

Integration of Language and Content Through Languaging

183

correct spelling, as the speakers will first have to agree upon an appropriate English word, the spelling of which they can then determine. Sakari does not back down but uses soft voice and flat intonation to construct the acceptance token ‘joo’ (‘yeah’) as ironic at lines 18 and 21. With no side initially qualifying their claims to knowing what ‘bruises’ means, the apparent stalemate becomes a matter to be solved by means of external resources. Susanna’s facial micro-gesture of moving her eyeballs to top corner at line 22 during a 2.0 second silence already projects that she might be backing down, and at line 23 she uses the particle ‘eiku’ (~ ‘no I mean’) to project repair (cf. Sorjonen & Laakso, 2005: 251), which attends to the degree of knowledgeability she has previously claimed for herself. Following this, at line 24, she turns to the third group member, Inka, to request the correct English translation of ‘mustelma’ (singular of ‘mustelmia’), thereby asking her to be an arbiter for the dispute. The three students’ suggested translations at lines 28–30 are marked as instances of language play by means of smiles and laughter particles. It is only when the teacher approaches the group and Sakari grabs the opportunity to address her at line 33 that the more serious quest continues. The teacher’s response (line 35) ratifies the correctness of Sakari’s original display of knowledge, that ‘mustelmia’ is indeed ‘bruises’ in English. By bringing him a dictionary, she orients to the search as a spelling-related one – after all, the students are supposed to be independently writing essays. After the teacher has left the group, Susanna jokingly reprimands Sakari for his just-ratified epistemic primacy within the group by giving him the finger, accompanied with chuckled laughter particles. Extract 1 illustrates how students need to conduct interactional work to accomplish a shared understanding of what the targeted language form is before it can be addressed, and how this work can sometimes become a matter of struggle. It also shows how shifts can be established in the joint focus of attention, so that what began as a quest for the correct spelling of a candidate word made relevant a question of whether or not that word is itself a correct translation of a first language (L1) word. While working on such different aspects of ‘bruises’, students also need to manage epistemic relations in (peer) interaction in which knowing or not knowing can be a consequential matter for social relations, vividly expressed by Susanna’s reaction following the teacher’s ratification of Sakari’s stance as correct. By now, the students have established that ‘bruises’ is an appropriate translation for ‘mustelmia’ and that it does not mean the same as ‘arpia’ (‘scars’). In Extract 2, we see how the original focus on spelling is resumed after Sakari has spent some time browsing the dictionary which was brought to him by the teacher before he again solicits the help of the group at line 66.

184

Part 3: Practices

Extract 2 ‘How do you use this?’ 66 Sakari -> 67 68 69 70 71 72 73 74 75

Inka -> Sakari-> -> Inka Susanna

↑mite- ↑mistä (.) ↑miten tää käytetää ↑how- ↑where- (.) ↑how do you use this (2.5) en tiiä (.) mää kirjota jotenki ((CLOSES DICTIONARY)) I dunno (.) I’ll write something what word you are for= =I know now. (1.2) but I don’t know how to write it. REACHES OUT FOR THE DICTIONARY LEANS TOWARDS SAKARI

mikä what [/bruses/

76 Sakari -> 77 Susanna -> BEGINS TO WRITE ON A PIECE OF PAPER 78 (ei oo) it’s not 79 ERASES HER WRITING 80 ässällä with an S 81 SHIFTS GAZE TO INKA WHO IS HANDLING THE DICTIONARY 82 Sakari ämmä [ämmä ämmä M M M 83 Susanna-> [/bru:si:s/ 84 ((33 LINES REMOVED – STUDENTS SAY ALOUD WORDS IN THE DICTIONARY)) 117 Inka -> °° 118 Susanna TAPS THE DICTIONARY 119 Sakari näy[tä sho [w me 120 Susanna [(on /bruise/) (is bruise) 121 Inka [°black and blue° 122 (.) 123 -> °se o bee är /uise/° °it is B R /uise/° (2.0) / SUSANNA WRITES ON PAPER 124 125 Inka 126 Susanna

POINTS AT SUSANNA’S PAPER

°tää on sitte äs° °and this is an S°

127 (2.1) 128 Sakari -> /bruise/ hehe 129 Susanna-> niinku bruises like bruises

Sakari’s query at line 66 is procedural, targeting how the dictionary is ‘used’ to find the correct spelling. Given that there is no audible or visible uptake during the long silence at line 67, he seems to give up on this project before Inka pitches in with an offer to help out at line 68. Notice how Inka’s turn changes the language of the epistemic work to English, so that Sakari’s response (lines 70–72) quite elegantly claims a shift in epistemic

Integration of Language and Content Through Languaging

185

state and redefines the nature of the problem assumed in the previous turn: that is, Sakari ‘now knows’ the word but is unable to spell it. As it turns out, the correct spelling becomes a matter to be resolved by all three students, who participate in different ways. Inka begins to search the dictionary for a translation for ‘mustelmia’ in the Finnish–English section (line 73), guided by Sakari (line 82); Susanna confirms what the word is with Sakari (lines 74–76) and begins to scribble it on a piece of paper, coordinating her physical activity with self-talk that focuses on the letters in the spelling (lines 77–80). After a spate of language play in which random words are read aloud from the dictionary (not shown in the extract), Inka finds the word ‘mustelmia’ and promptly reports the possible translations to the others (lines 117, 121). The spelling problem is finally resolved as Susanna writes the correct spelling on the paper slip and it is shown to Sakari (see image). To finish off the instruction by peers, Susanna corrects the missing s in Sakari’s turn to indicate plural spelling at lines 128–129. The analyses of these two extracts show that students can deploy a broad range of resources to resolve contingent problems during individual task work in the CLIL classroom. They not only draw on L1, L2, the dictionary and writing paper, which they coordinate with and by their bodies, but they also tap into the knowledge resources of each other and the teacher. Moreover, on occasions, it may be difficult to distinguish between these resources, such as when the students use pronunciation to convey the spelling of ‘bruise(s)’. For example, Inka’s pronunciation of the word at line 117 as she announces the spelling in the dictionary is different from that of Susanna’s at line 83 as she follows Inka’s search. At line 117, Inka pronounces the English word ‘bruise’ as if it followed the conventions of Finnish orthography, which relatively consistently reflects the phonemes of the language (see also Jakonen & Morton, 2015). This kind of translanguaging (García & Li, 2014) is therefore designed to provide sufficient information to Sakari of the correct word spelling, even if it turns out to need partial letter-by-letter spelling at line 123. In terms of the outcome of the problem-solving sequence, evidence in the form of Sakari’s written text shows that the word ‘bruises’ was successfully incorporated into the content task he had been doing (Figure 7.1). In this sense, the analyses of the ‘languaging’ practices can be seen as an example of Markee’s (2008) LBT, as they show how the learning object ‘bruises’ occurs over the period of the interactional episode in which it was the focus of attention, how it was oriented to in different ways by the participants concerned and how it was eventually incorporated into Sakari’s (written) repertoire. Returning to Koole’s (2012) ideas about the ownership of problems in a multilingual classroom, it is clear here that this problem was not a product of the teacher’s understanding of what Sakari needed, although the teacher was used as a resource in eventually resolving the

186

Part 3: Practices

Figure 7.1 Sakari’s essay

issue (both through her own oriented-to epistemic primacy on language matters and as a purveyor of an additional resource, the dictionary). As Swain (2000: 12) has pointed out in her earlier work on collaborative dialogue around language problems, ‘it is impossible to predict what pairs of students will talk about’. The task that Sakari was doing did not oblige him to write about school life in Tudor times in such a colourful way, and it was left to him to find, with the help of other mediational means (in sociocultural terms), the resources he needed to complete the task in the way he wanted to. ‘Bruises’, seen in this way, is not just a ‘language’ or ‘textual’ problem, but is an example of FonF integrated with content.

Discussion and Conclusion In this chapter, we have addressed the issue of language and content integration from the perspective of participants by using a multimodal CA methodology to illuminate through a detailed sequential analysis how students use a wide range of linguistic (L1 and L2), artefactual and embodied resources to identify and resolve a language problem arising in the midst of individual work on a content-based writing task. Our approach has been integrative at a theoretical level, as we have shown how two key constructs in L2 pedagogy, FonF and languaging, can be respecified using the analytic resources of CA, particularly recent work on epistemics in interaction. Doing so does not attempt to deny the importance of these cognitive constructs, whether they are derived from cognitivist SLA research or sociocultural theory, but to provide a much more fine-grained picture of how FonF and languaging are actually accomplished in a CLIL classroom. We have seen that in the wider field of L2 pedagogy, the integration of ‘grammar and communication’ is a pervasive problem, and various solutions ranging from no focus at all to the various options within FFI have been proposed. These issues are no less pressing in CLIL, especially in the context of claims that it is a form

Integration of Language and Content Through Languaging

187

of communicative language teaching par excellence. However, as Tian and Macaro (2012) point out, even within language teaching, FonF raises serious issues about how much time can be dedicated to it. As they put it, ‘How much lesson time should be devoted to intentional vocabulary learning, when one considers all the other demands made on teachers and learners alike, is an important question’ (Tian & Macaro, 2012: 382). If this is the case in classes where language learning is the only aim, then it becomes all the more problematic in CLIL, where the demands of content learning are usually paramount. The construct of languaging can be a rich resource for further research in CLIL. As Swain et al. (2009) point out, the use of language to shape experience, build on and share knowledge and effect conceptual change is common to the teaching of all subjects. CLIL offers a privileged setting to encourage and observe languaging across language and other subject domains, especially multilingual languaging practices, which could be a valuable addition to recent work investigating the construct of translanguaging (see the chapter in this volume by Moore & Nikula; also Creese & Blackledge, 2010; Gallagher & Colohan, 2014; García & Li, 2014). Bringing the analytic power of multimodal CA can illuminate languaging and translanguaging practices by showing the already complex interactional work that is the foundation that underwrites all work on learning objects, whether they be in ‘content’ or ‘language’ domains. As Macbeth (2011) shows, there is an ‘understanding’ that needs to be established interactionally among participants before any understanding of curricular content can take place. The ‘bruises’ episode analysed in this chapter was an example of languaging practices around a language item which popped up incidentally because one student decided to use it in his essay. It is thus an example of incidental, unplanned, student-led FonF. CLIL as an environment may be especially rich in affordances for such practices to occur, but it may not be enough to leave integration to chance in this way. The kind of work on languaging carried out by Swain and colleagues looks at how learners develop understanding of pre-specified areas of language which are important in the L2 curriculum, such as the passive voice. In CLIL, it would be of interest to develop opportunities for languaging around language areas chosen for their functional importance in relation to content. In this way, the kind of highly interactionally sensitive work shown here could be incorporated into more tightly engineered research designs involving pre and post-tests, ideally based on content-based tasks. Such an approach would match the integration of content and language at the level of classroom practice with the integration of research perspectives and methodologies drawn from different traditions, in the pragmatic applied linguistics spirit of finding solutions to, and deepening understanding of, real-world problems.

188

Part 3: Practices

Appendix: Transcription Symbols (based on Jefferson, 2004) wo::rd

prolonged sound

(.)

silence less than 0.5 seconds

(2.0)

duration of a silence

(word)

uncertain transcription

word

emphasised talk

£word£

talk produced with smiley voice

slower pace than in surrounding talk

>word