# Exploring Language Education Global and Local Perspectives

Bardel et al. (Eds.)

The overarching aim of this book is to offer insight into some of the currently discussed issues at the Swedish as well as the international research frontline of Language Education in a selection of up-to-date work. The volume deals with the gap between research and practice, the importance of pronunciation in second/foreign language teaching and learning, classroom translanguaging, young learners' exposure of English outside school, and teachers' attitudes to foreign languages other than English. A further aim is to provide readers with input from research within the interconnected disciplines of Applied Linguistics, Language Education and Second Language Acquisition targeting researchers, school teachers,

teacher students, teacher educators and policy-makers.

parts of the world.

Sweden, Canada and the United States.

The content is based on a selection of talks presented at the 1st ELE Conference ('Exploring Language Education'), which was held at Stockholm University in 2018. Employing a broad thematic scope, the volume reflects the variety of perspectives on Language Education, brought together at the conference by authors working in diverse areas of the field and in different

With the first ELE conference the organizers wished to call attention to the intersection of the global and the local, in terms of linguistic and cultural diversity informing and conditioning both research questions and language education practices. The chapters represent different parts of the broad array of research directions that can be discerned under the large umbrella of Language Education, zooming in on the Western context, specifically

Exploring Language Education

LLE 2

Camilla Bardel, Christina Hedman, Katarina Rejman & Elisabeth Zetterholm (Eds.)

Language and Literature in Education 2

# **Exploring Language Education Global and Local Perspectives**

*Camilla Bardel, Christina Hedman, Katarina Rejman & Elisabeth Zetterholm (Eds.)*

Published by Stockholm University Press Stockholm University SE-106 91 Stockholm, Sweden www.stockholmuniversitypress.se

Text © The Author(s) 2022 License CC BY 4.0. Supporting Agency (funding): Områdesnämnden för humanvetenskap, Stockholms universitet

First published 2022 Cover designed by Stockholm University Press

Language and Literature in Education (Online) ISSN: 2003-8542

ISBN (Paperback): 978-91-7635-191-8 ISBN (PDF): 978-91-7635-188-8 ISBN (EPUB): 978-91-7635-189-5 ISBN (Mobi): 978-91-7635-190-1

DOI: https://doi.org/10.16993/bbz

This work is licensed under the Creative Commons Attribution 4.0 Unported License. To view a copy of this license, visit creativecommons.org/licenses /by/4.0/ or send a letter to Creative Commons, 444 Castro Street, Suite 900, Mountain View, California, 94041, USA. This license allows for copying any part of the work for personal and commercial use, providing author attribution is clearly stated.

Suggested citation: Bardel, C., Hedman, C., Rejman, K., and Zetterholm, E. (Eds.), 2022. *Exploring Language Education: Global and local perspectives*. Stockholm University Press. DOI: https://doi.org/10.16993/bbz. License: CC BY 4.0.

To read the free, open access version of this book online, visit https://doi.org/10.16993/bbz or scan this QR code with your mobile device.

# **Language and Literature in Education**

The aim of the book series *Language and Literature in Education* (ISSN 2003-8542) is to disseminate studies of language education and literature in education. Although there are a few journals publishing general educational research in the Nordic countries, there is no forum for publications in the broad field of language education, and there is definitely a gap to fill as regards educational perspectives on language and literature. Researchers operating in and outside Sweden are invited to send in proposals for monographs, thematic volumes, or collections of papers. The main language of publication is English. However, other languages are not excluded if motivated by the theme of the volume. Theoretical and empirical contributions are welcome.

This book series aims to function as reference point for researchers occupied with educational aspects of language and literature, considering the structure and content of a range of school subjects, and to embody a forum for broad perspectives of educational studies with a focus on language development, literacy and literature.

# **Editorial Board**


• Elisabeth Zetterholm, Senior Associate Professor, Department of Culture and Society (IKOS), Linköping University, Sweden

# **Titles in the series**


# **Peer review policies**

### **Guidelines for peer review**

Stockholm University Press ensures that all book publications are peer-reviewed. Each proposal submitted to the Press will be sent to a dedicated Editorial Board of experts in the subject area for evaluation. The full manuscript will be reviewed by chapter or as a whole by two external and independent experts.

A complete description of Stockholm University Press' peerreview policies can be found on the website: http://www.stockholm universitypress.se/site/peer-review-policies/

The Editorial Board of *Language and Literature in Education* applied a single-blind review during manuscript assessment. The Editors Camilla Bardel and Elisabeth Zetterholm stepped down temporarily from their roles in the Editorial Board during the review process of this book to avoid a conflict of interest. Anders Philipsson acted as chair for the Editorial Board during the evaluation process of this book.

### **Recognition for reviewers**

Stockholm University Press and the Editorial Board would like to extend a special thanks to the anonymous reviewers, who contributed to the process of editing this book, for their work and time spent on commenting about the quality of the manuscript.

# **Contents**

#### Introduction 1

*Camilla Bardel, Christina Hedman, Katarina Rejman & Elisabeth Zetterholm*

1. SLA research and L2 pedagogy: An uneasy relationship 13

*Nina Spada*

2. Teaching pronunciation: Truths and lies 39

*John M. Levis*

3. Multilingual versus monolingual classroom practices in English for academic purposes: Learning outcomes, student attitudes, and instructor observations 73

*Amanda Brown, Robert J. Lally & Laura Lisnyczyj*

4. Very young Swedish children's exposure to English outside of school 121

*Liss Kerstin Sylvén*

5. Attitudes and ambiguities – teachers' views on second foreign language education in Swedish compulsory school 157 *Gudrun Erickson, Camilla Bardel, Rakel Österberg & Monica Rosén*

Afterword 225 *Stellan Sundh*

Author biographies 229

# **Introduction**

*Camilla Bardel, Christina Hedman, Katarina Rejman & Elisabeth Zetterholm*

Acquiring the ability to use language(s) competently and in various manners is crucial both for individual development and for functional social and cultural relations, within and across regional and national borders (Byram, 2008). Successful development of what is traditionally called first, second or foreign languages (henceforth abbreviated L1, L2 and FL) is of great importance to meet that end.1 In relation to this need, as a research field, Language Education (LE) is paramount for the understanding and improvement of human social life and conditions (Council of Europe, 2001). The field of LE is vast and diversified, covering all aspects of language use in educational contexts and learning processes as well as the teaching and learning of specific languages. Being extremely broad, with several emerging subareas, the field is still in a process of self-definition, constantly seeking, as the whole educational field, its identity. Current challenges that need to be addressed are related to societal changes clashing with established research foci. From a critical perspective, language practices cannot be seen as something neutral, but embedded in a sociohistorical and political context reflecting current societal power imbalances (Blommaert, 2010; Pennycook, 2010). Although

#### **How to cite this book chapter:**

Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (2022). Introduction. In: Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (Eds.), *Exploring Language Education: Global and local perspectives*, pp. 1–11. Stockholm University Press. DOI: https://doi.org/10.16993/bbz.a. License: CC BY 4.0.

<sup>1</sup> We are aware of both the oversimplifications and the ambiguities related to the use of the labels first language (L1), second language (L2) and foreign language (FL), which may overlap and be used interchangeably in this book. Although in constant need of problematizing discussion, they are relevant in relation to many studies of the LE field, especially when considering the contexts of the studies included in this volume.

continuously approached from new points of view, fundamental matters that interest LE researchers remain the educational context, the teacher, the learner, the subject, and, not least, the processes of language teaching and learning.

*Exploring Language Education – global and local perspectives* was the theme of the first 'ELE' conference, which was arranged at Stockholm University in June 2018 and sponsored by the university's Board of Human Science and the Swedish Foundation for Humanities and Social Sciences (Riksbankens Jubileumsfond). The conference employed a broad scope, aiming to reflect a variety of perspectives on LE and brought together scholars working in different areas of the field from many parts of the world.

With this first ELE conference the organizers wished to call attention to the intersection of the global and the local, in terms of linguistic and cultural diversity, which might inform both research questions and practices within the field. Areas of interest shared by many presenters were multilingualism, Global Englishes, and specific issues related to these topics, where experienced tensions between research and practice seemed to be mutual. Yet, it also became evident how different regions and communities are contingent on local prerequisites and circumstances, leading to a number of particular challenges and assets when it comes to language education. There were unfortunately few presentations from the important area of Literature in LE (Hall, 2015), although the reach of the conference encompassed this field. We see that future conferences and volumes could have a lot to gain by widening the perspective and giving literature, verbal art and learning through fiction more visibility.

The ELE conferences will continue to offer a platform for researchers and teachers to share ideas and engage in new knowledge that develops at the different interfaces with other research fields, such as, for instance, sociolinguistics, gender studies, pedagogy and literature studies.

This book offers a selection of papers presented at the conference, focusing on some specific thematic areas of global relevance (the multilingual classroom and English for academic purposes, young learners' extramural learning of English, and the learning, teaching and assessment of languages other than English) but

clearly zooming in on the Western context, specifically Sweden and the United States. The work presented in this volume has an approach to LE that is informed by applied linguistics, SLA and pedagogy; for the sake of representativity, it is desirable that future publications emanating from the ELE conferences will be influenced by proponents also of other research traditions and paradigms.

A natural point of departure for a publication on global and local perspectives of language education is to try to capture both the general and the context specific. As for general issues, the fields of SLA and linguistics are two important foundations of LE; in the case of context specific problems, a variety of theories and methods are thinkable. This book comprises five chapters; two conceptual and three empirical studies, the latter focusing on specific contexts. Together they represent different parts of the broad array of research directions that can be discerned under the large umbrella of LE, but are linked to each other by similar linguistic and educational approaches. Two of the plenary speakers at the conference, Nina Spada and John Levis, contribute in the volume. The content of their presentations is laid out in the first chapters, which are followed by reports on work from other presenters at the conference; Amanda Brown and colleagues, Gudrun Erickson and colleagues, and Liss Kerstin Sylvén. The common ground of the five chapters is constituted by recurrent discussions of a fundamental thematic cluster: the interplay of the languages of the learner's repertoire, including the target language, and the complexity of their respective roles in language teaching and learning.

Through the two keynote speakers, the connections between L2 pedagogy, LE, SLA and linguistics are brought to the fore. The following three empirical chapters, although focusing on diverse contexts and conditions for language teaching and learning, all circle around the notions of input and use of the target language and the position of this in relation to other languages, the majority language of the community where the language is learnt or the L1s of the learners. Generally, the work presented in this volume tends to be located at the interface between LE and SLA, especially the studies by Sylvén and Brown et al., who measure aspects of linguistic knowledge. The role of attitudes in language learning is also an aspect of relevance for all three empirical studies of this book. One of the major focus points in the book is the role of the teacher, both as researcher and as participant; see for instance Spada's discussion of the relevance of LE research to the practitioner, Brown et al.'s study involving teacher-researchers, and Erickson et al.'s study based on a teachers' questionnaire.

One aim of this book is to emphasize the importance of informing the field of language teaching and language teacher training with input from research within Applied Linguistics, LE and SLA. The selection of papers aims at offering insight into some currently discussed issues in the Swedish as well as in the international context, for example the role of target language input and the attitudes toward the target language, translanguaging in the multilingual classroom, and the teaching and learning of languages other than English. The three research projects presented in empirically based chapters are concerned mainly with these issues, which should all be of relevance to language teachers, teacher trainers, and students.

The relationship between the field of SLA and L2 pedagogy is not straightforward, however. The extent of which SLA is relevant to L2 pedagogy is raised and discussed by Nina Spada in the first chapter of this volume, *SLA research and L2 pedagogy: An uneasy relationship*. While the main focus of SLA research is on the learning process and on the factors that contribute to how and to what extent second or additional languages are learnt, L2 pedagogy is concerned with theory and practice more directly related to the teaching and learning of languages. The strong focus on the learner and the interlanguage may in some cases make SLA research appear of little relevance to teachers, in the sense that implications for teaching may not be obvious.

Particularly relevant to language teaching, Spada claims, are *Instructed SLA* and *Classroom research*. Nonetheless, as argued, a gap remains between research and L2 pedagogy. Spada contemplates several possible reasons for this. From the teachers' perspective, time is often a factor. Also, teachers may find research reports too technical, and scepticism among teachers towards the relevance of research is sometimes manifested. From the researchers' perspective, one problem may be a lack of contact with the classroom and hence little understanding of teachers' questions and concerns. Another issue is related to the channels that most researchers tend to publish through. As pointed out in this chapter, professional journals or teachers' newsletters are not the kind of channels that lead to publishing credits in academia and the publication outlets that researchers aim at are often not available outside university libraries. If teachers get access to research publications, the style and the text genre make the work hard to approach for someone who is not a researcher in the field. Also, the author notices, research of interest to the researcher may be too generic and distant from the specific classroom situation of the teacher. But, as argued by the author, teachers and researchers share the common goal of making L2 learning and teaching as successful as possible and both groups have many things to learn from each other. Suggestions are therefore made about how to bridge the potential gap between them and examples of research that may reduce it are given. Researchers are invited to develop a better understanding of teachers' concerns in order to take on teacher-relevant research endeavours. The local rather than the general should then be emphasized in order to achieve significance. Teachers are encouraged to invest in professional development, to be able to take informed decisions about what research questions are relevant in their contexts, and to participate in research projects. The chapter ends with a discussion of how alternative views upon the role of the teacher in research can complement each other. The following four chapters are concerned with different aspects of second or foreign language teaching and learning, all relating to the conceptual frame laid out in the first chapter, but at the same time building on autonomous research about debated factors in non-native language learning, such as input, multilingualism, proficiency and attitudes, to mention some of the connecting points between the chapters.

In Chapter 2, *Teaching pronunciation: Truths and lies*, research about L2 pronunciation is presented and discussed by John Levis. Pronunciation and its role for communication is a linguistic area that has been neglected for years in both L2 research and teaching. It is however experiencing a revival – and a legitimate one considering its importance in language teaching, being an inevitable task for learners of a new language. One indication of the growing visibility of research on pronunciation in L2 teaching and learning is the current increase in journal publications, books and conferences on the topic, as pointed out by the author. The chapter is structured around what is referred to as four truths and four lies about pronunciation teaching and learning. These eight statements frame an overview ranging from communicative aspects of pronunciation and teaching to a final discussion of the social aspect of accent and its relation to identity. The four lies can be said to be based on opinions related to an idea that not only does pronunciation not *need* to be taught but, moreover, it *can* not be taught. On the contrary, it is claimed by Levis, pronunciation is unavoidable and essential, and teaching pronunciation works. The lies and the truths are connected to fundamental questions that should be of interest to teachers and learners as well as researchers: Is pronunciation important for oral communication, or is it enough to master vocabulary and grammar in order to communicate successfully, for example? As Levis convincingly argues, both pronunciation and listening comprehension are important components of successful communication, and an intelligible pronunciation, in combination with grammar and vocabulary, is essential. A question related to teaching and learning goals is whether an adult learner can ever achieve the pronunciation of a native speaker. As pointed out in the chapter, there is a social aspect of the varieties of pronunciation of all speakers, first- as well as second-language speakers, that needs to be considered in the teaching of pronunciation. It is through a combination of pronunciation, lexical and pragmatic competence that L2 speakers get access to (the language of) the community and are allowed to further improve their linguistic abilities.

For decades, a hot topic in applied linguistics has been the question whether language learning is best accomplished with the use of the target language only, or by using also the students' L1 (or rather, the majority language of the specific society) in the classroom. On the one hand, use of the target language only would compensate for a potentially low degree of input outside the classroom and it would give students maximal opportunity to use the language in question. On the other hand, exclusive target language use may have the effect of raising students' anxiety. Using a language the students are more proficient in may help as a means to ask questions, to translate or explain, and it may promote students' explicit linguistic knowledge. Independently of what research says about the pros and cons of comparing languages, a binary contrastive approach is problematic considering contexts where not all learners share the same L1, which probably goes for the majority of classrooms all over the world. This reality has justified a third way (see e.g., Cummins, 1998), namely multilingual language use in the classroom, an approach that is advocated by, for example, the proponents of translanguaging or other similar concepts such as heteroglossia in the classroom (Cenoz, 2013).

The third chapter, *Multilingual versus monolingual classroom practices in English for academic purposes: Learning outcomes, student attitudes, and instructor observations*, by Amanda Brown, Robert James Lally and Laura Lisnyczyj, reports from a study on the language choice in multilingual classrooms of English for Academic Purposes (EAP). There are few experimental studies on the effect of L1 vs. L2 use and there is no consensus in the frequently discussed issue of language choice in the language classroom. This study, which contrasts a monolingual and a multilingual pedagogy, empirically investigating their effects on students' learning outcomes as well as their attitudes towards classroom language use, is therefore a welcome contribution. The participants of the study were two researcher-practitioners and 50 international undergraduate students drawn from four EAP courses at a large university in the United States. However, as a whole, the classes were highly diverse linguistically and culturally, containing students with a range of different L1s; some groups were more heterogeneous than others. Two courses were instructed by the two researcher-practitioners with the standard monolingual Englishonly pedagogy. Two comparable courses were taught by the same instructors using a multilingual approach, where students were given agency in their choice of language for the learning activities. The data – student surveys, researcher-practitioner journals documenting the pedagogical content, and students' target language performance – were collected longitudinally. Qualitative and quantitative analyses were conducted on students' attitudes and their development of writing and speaking.

No statistically significant differences in learning outcomes between multilingual or monolingual conditions were found, neither were significant differences found between groups with few different L1s and groups where many different L1s were represented.

The student survey responses indicated a complex picture of attitudes toward classroom language practices, and a pedagogical reflection in relation to the multilingual classroom is conducted at the end of the chapter together with a discussion on research methodology. Students' attitudes in relation to target language only vs. multilingual classroom language practice is surely a topic worth looking into further in future research. The study is timely and represents a topic that concerns teachers in the linguistically heterogeneous classrooms of current times (cf. chapter 1 on the relevance of classroom research for teaching practice).

Another research study from the field of English as an L2 is reported in the next chapter, *Very young Swedish children's exposure to English outside of school*, by Liss Kerstin Sylvén. This is a study of young children's exposure to English in Sweden, a country where English is officially a FL and referred to as such in education policy documents. However, as the author argues, it may rather be considered an L2 due to its vast presence and extended use in society, especially among younger generations. Little is known however about how much English very young children encounter in Sweden, which is why exposure was here investigated among 7 preschool children (ages 3–5 years) and 13 primary school children in grade one (ages 7–8 years). This type of knowledge, dependent upon local language ecologies, is highly relevant to English teachers in primary school (cf. chapter 1). More specifically, the author examined both how much English the children were exposed to and through which activities this exposure occurred. This was done through interviews with the parents and language diaries kept by them. Furthermore, the level of receptive vocabulary knowledge among the children was measured with *the Peabody Picture Vocabulary Test*, PPVT-4 (Dunn & Dunn, 2007), and – for the older group – oral production of vocabulary was also examined with a picture description task. Despite the fact that the children had not yet started with any formal instruction in English, the findings indicate that none of them were complete novices in English. In fact, some of the children had an unexpectedly high amount of exposure of English and also surprisingly advanced levels of proficiency, although varying. The chapter also offers a close-up study of two children in their first school year: One child with a large amount of exposure to English and another whose exposure to English seemed to be restricted to about one hour per week. This study contributes to the body of studies indicating that the role of English in Sweden has, in practice, moved beyond that of a FL. The chapter ends with a discussion on pedagogical implications and suggestions of possible future studies.

With the final chapter of the volume, while still in the Swedish context, the focus shifts to a group of languages with very different conditions than English in Sweden. In *Attitudes and ambiguities – teachers' views on second foreign language education in Swedish compulsory school*, Gudrun Erickson, Camilla Bardel, Rakel Österberg and Monica Rosén report on a study of the situation of second foreign languages (SFLs). SFLs here refers to foreign languages encountered after English; in the Swedish compulsory school these can be studied within the subject *Modern languages*. A large number of teachers of French, German and Spanish participated in the study through an extensive questionnaire. The study focuses on three areas: teachers' professional satisfaction, teachers' use of the target language in the classroom, and the curricular status of Modern languages. These areas correspond to three levels of education, namely the individual, the pedagogical and the structural level, and they are connected to some frequently discussed issues in the current professional and policy-related debate in Sweden. As a background to the study, the authors point to the challenging conditions for SFLs, in particular in comparison to English: The attitudes, motivation and proficiency levels among the students are far from the same, a perception shared by the respondents in this study. The questionnaire – piloted and pretested – is grounded in previous research and reports, and based on conceptual considerations from the fields of teacher cognition and pedagogical content knowledge, the Common European Framework of Reference, as well as Swedish regulatory documents for language education. The chapter offers an insight to the complex field of SFLs, depicting a number of problems regarding different aspects of SFL education, for example the non-compulsory status of the subject Modern languages, or the low degree of target language use in the classroom. Despite these problems, the authors point out that the study conveys a genuine commitment among the respondents regarding their profession and the learning and development of their students. The study gives empirical insight to some of the themes Spada discusses in chapter 1.

To summarize, this book explores some current challenges of LE seen from different contexts in the Western part of the world, and from a perspective at the interface of the SLA field. The three chapters based on empirical studies investigate local problems and practices related to attitudes, proficiency, multilingualism and language use, raising questions that are applicable also in other contexts than the ones reported on. We wish to emphasize that all chapters have pedagogical implications and hope that the volume will inspire future research into matters of relevance for language teachers as well as for the research community, and for future conferences that will further explore the variety of themes relevant for LE, sketched out in the beginning of this Introduction. The editors are proud to make the volume easily accessible to teachers, researchers and students, both through its Open Access format, and considering the selection of content which will hopefully be of interest to all categories of readers, be they in one or more of these roles.

# **References**


# **1. SLA research and L2 pedagogy: An uneasy relationship2**

*Nina Spada*

# **Introduction**

Both second language acquisition (SLA) research and L2 pedagogy are subfields of Applied Linguistics. SLA is primarily concerned with the learning of second/additional languages that are learned after the acquisition of a first language (L1). The emphasis in SLA is on learning, the processes that are associated with it, and the factors that contribute to more or less successful learning. The focus of much of SLA research has been the learning of grammar, but it also includes the learning of lexical, pragmatic, and phonological features of language. L2 pedagogy is primarily concerned with theory, research and practice related to the teaching of second/additional languages. On the surface, there is an obvious connection between SLA and L2 pedagogy based on the assumption that the more we know about how languages are learned the better informed we will be about how to best teach them. The relationship between the two is more complicated than this, however, which is why the question *Should SLA research be relevant to L2 pedagogy*? has been responded to in different ways and is the source of considerable debate in the applied linguistics literature. For example, some scholars respond to this question with *Yes,* 

#### **How to cite this book chapter:**

<sup>2</sup> This paper is a slightly revised written version of a plenary presentation delivered at the Exploring Language Education (ELE) conference. Stockholm, June 2018. Some of the ideas have been expressed in other presentations including Spada & Lightbown, 2015 and Spada & Lightbown, 2019.

Spada, N. (2022). SLA research and L2 pedagogy: An uneasy relationship. In: Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (Eds.), *Exploring Language Education: Global and local perspectives*, pp. 13–37. Stockholm University Press. DOI: https://doi.org/10.16993/bbz.b. License: CC BY 4.0.

*absolutely*; others say *No, not necessarily* and still others claim that *It depends*.

Those who say *Yes, absolutely* argue that SLA is a subfield of applied linguistics which is dedicated to the study of real-world problems in which language plays a role. Therefore, SLA should have direct relevance to L2 pedagogy (Brumfit, 1997; Bygate, 2005). Those who say *No, not necessarily* argue that there are several areas of SLA research that are not relevant to L2 pedagogy, for example, naturalistic acquisition, theory development, methodological issues, as well as the nature/focus of some of the research questions investigated. This becomes evident when one takes a random look at articles published in recent mainstream SLA journals with titles such as: "*Establishing evidence of learning in experiments employing artificial linguistic systems*" (Hamrick & Sachs, 2018) and "*Investigating auditory processing of syntactic gaps with L2 speakers using pupillometry*" (Fernandez et al., 2018). In cases like these it is difficult to argue that the research is directly relevant to L2 teaching, nor is it necessary or appropriate to do so. Evelyn Hatch acknowledged the need to distinguish between SLA and L2 pedagogy over 40 years ago when the field of SLA was just developing. At that time, most researchers came to SLA as language teachers, and they were highly motivated to make connections between research and teaching. In her article entitled "*Apply with Caution*", Hatch claimed that much of SLA is not directly relevant to how languages can be best taught in L2 classrooms and argued that "The only question the researcher should answer is the one he/ she asks" (Hatch, 1979, p. 138).

Since then SLA has evolved and grown considerably; it has become a multi-faceted and interdisciplinary field (see Gass et al., 2013 & Ortega, 2009 for overviews of SLA theory and research). This includes investigations of internal/learner factors for example: cognitive (memory, aptitude), biological (learners' age) linguistic (learners' other languages), affective (learners' identity, motivation). It also includes investigations of external factors such as the linguistic environment inside the classroom (teacher talk; learner interaction); the characteristics of exposure to the L2 through study abroad programs and the use of technology inside and outside the classroom. SLA also encompasses a diverse range of theoretical perspectives. While the field has been primarily influenced by a wide range of linguistic and psychological theories, recent years have seen an increase in social (Block, 2003) and multilingual (May, 2014) perspectives, and transdisciplinary frameworks for conceptualizing the complex and multidimensional aspects of language development (Douglas Fir Group, 2016). Greater epistemological diversity is also evident as qualitative, interpretive research methodologies emerge alongside quantitative, experimental approaches that have dominated SLA since its inception.

Within the broader discipline of SLA, subfields have developed which have more direct relevance for L2 teaching. This is what has led some scholars to respond to the question *Should SLA be relevant to L2 pedagogy*? with *It depends*. Two that are considered particularly relevant to L2 pedagogy are: Instructed SLA and L2 classroom research. Instructed SLA is concerned with "Language acquisition in classrooms, language laboratories, or other settings where language is intentionally taught and/or intentionally learned" (Spada & Lightbown, 2013, p. 1). Instructed SLA "attempts to answer two questions: 1) is instruction beneficial for second language (L2) learning and 2) if so, how can the effectiveness of instruction be optimized?" (Loewen, 2015, p. 1). For many years instructed SLA has been carried out within a cognitive-linguistic perspective of L2 learning. The last two decades have seen a steady increase in instructed SLA research conducted within a socio-cultural perspective (Lantolf et al., 2018). Some specific questions that have motivated instructed SLA research include: 1) How does instructed SLA compare with naturalistic SLA? 2) Is explicit instruction more effective than implicit instruction? 3) What type of knowledge results from instruction? 4) Are some corrective feedback types more effective for L2 learning than others? 5) Is L2 instruction more effective for the development of certain language features?

L2 classroom-research examines similar questions about the relationship between the teaching and learning of languages (Chaudron, 1988; Ellis, 2012; Spada, 2019) in descriptive and quasi-experimental studies, It differs from instructed SLA in that the research takes place exclusively in classrooms. This has often resulted in more detailed descriptions of the characteristics of L2 instruction in what is referred to as process-product research (Long, 1980/1983; Spada, 2019). It is argued that there is greater "ecological validity" that comes with L2 classroom research. This refers to the "degree of similarity between a research study and the authentic context that the study is purportedly investigating" (Loewen & Plonsky, 2016, p. 56). It is believed that teachers are more likely to value research that takes place in real classrooms with real learners and teachers. This does not mean that instructed SLA research carried out in laboratories cannot be useful for teaching, but that research carried out in actual classrooms carries more weight for teachers. Of course, the application of findings from L2 classroom research does not guarantee that they will lead to improved learning outcomes because of fundamental differences between the research and educational context.

Even though both instructed SLA and L2 classroom research are more applied and investigate questions of greater relevance to L2 pedagogy, the gap remains between SLA research and L2 pedagogy. Why is this so? Below we discuss some of the reasons for this gap from both the teachers' and researchers' perspectives.

# **Why this gap? Teachers' perspectives**

One of the explanations that is frequently offered for the gap between SLA research and L2 pedagogy is that teachers do not have time to read research journals and when they do, the articles are often too technical and inaccessible as reflected in the following quotes: *In my own [school] context almost nobody reads TESOL publications—they don't have time*. (Borg, 2006); *I need somebody to interpret for me what [the researchers] are trying to say …* (Gore & Gitlin, 2004). It has also been observed that when teachers read articles in teachers' journals or newsletters little research is actually reported in them. In a review of 5 years of publications in 4 professional association journals and newsletters targeted to language teachers in the UK, Australia and the US, Marsden & Kasprowicz (2017) found that more than a third of the articles did not include references to research. It has also been reported that some teachers are skeptical about researchers' knowledge about language teaching as indicated in the following statement by an ESL teacher: *the nerve some "experts" have, to go waltzing into teachers' classrooms with their PhDs, having never been public school teachers themselves, and tell these seasoned professionals… what they should do in their classes*  (Kerekes, 2001).

Teachers' doubts about whether research is relevant to classroom teaching and learning are understandable. In a review of articles published in 2 mainstream SLA journals (*Studies in Second Language Acquisition* & *Language Learning*) from 1990 to 2010, laboratory studies were found to outnumber classroom studies by 3 to 1 (Plonsky, 2013). Again, this is not to suggest that lab studies are not relevant for teaching but that the findings from laboratory and classroom studies often differ. One important example of this emerged in research on the *negotiation of meaning*. This term refers to adjustments that are made in the speech of L2 learners (and other speakers) when breakdowns in communication occur. These linguistic adjustments include requests for repetition, comprehension checks and clarification requests (Long, 1996). Many studies that have been conducted under laboratory conditions in which L2 learners interact with each other have found a great deal of *negotiation of meaning* to occur as well as evidence for its value in SLA (Mackey, 1999; Mackey & Goo, 2007). Some research carried out in classroom settings, however, has revealed that when students interact with each other they pretend that they understand each other rather than *negotiate for meaning* (Foster, 1998). This does not mean that L2 learner-learner interaction does not play an important role in classroom L2 learning, only that it may require more preparation, guidance, and supervision than we might expect on the basis of laboratory studies.

# **Why this gap? Researchers' perspectives**

Most SLA researchers – even those who do instructed SLA research and L2 classroom research are better at communicating with researchers than with teachers, do not have a good understanding of teachers' questions and concerns, and have not taught in language classrooms for a long time. Also, most SLA researchers work at universities and in this context they are evaluated more positively when they publish in research journals because articles about pedagogy are given less weight than those focused on research. As a result, SLA researchers tend not to publish in professional journals or teachers' newsletters. This may account for the significant drop in the number of articles about L2 pedagogy in the flagship journal, *Applied Linguistics*, over the past 3 decades (Hellerman, 2015).3 Fortunately, in recent years, universities and research funding agencies have placed more weight on the dissemination of research findings, the importance of shared knowledge, and the impact of research knowledge on society, i.e., education and second/foreign language education. Over time this should help to narrow the gap.

Another reason for the gap between research and pedagogy is that researchers are more interested in the general than the specific and there is a tendency to assume that knowledge can be transferred from the research context to a particular classroom context. This needs to be demonstrated rather than assumed. This issue is addressed in the section below which examines some of the ways in which the gap between SLA research and language pedagogy can be narrowed. Strategies are discussed first from the researcher's perspective followed by the teacher's stance. It is also important to emphasize that there are initiatives that both teachers and researchers can take to narrow the gap one of which is to value, build on and share each other's knowledge. This means breaking down the top-down nature of the researcher/teacher relationship in which the former is considered the expert and the latter the recipient of knowledge (Freeman, 1992; Ortega, 2012). Both share the goal of making classroom teaching and learning as successful as possible and both have a great deal to learn from each other.

Researchers and teachers bring different types of knowledge and experience to the research process. Starting with their own years as students--observing teachers in what has been called

<sup>3</sup> This drop may also be related to the fact that there are more journals available today that publish articles about research on L2 pedagogy than there were in the early 1980s.

"the apprenticeship of observation" (Lortie, 1975), followed by their initial training, and continuing through their experience in the classroom, their interactions with other teachers, and the professional development activities that they engage in, teachers build up the knowledge they need to plan and carry out their daily tasks. They need to have grounding in pedagogical principles and practices including classroom management procedures and curriculum content, what the educational psychologist, Lee Shulman (1986) referred to as pedagogical and content knowledge. This includes knowledge about how languages are learned and the potential relevance of SLA research for practice. In the case of L2 teachers working in content-based approaches (e.g., immersion, CLIL), the curriculum content also includes the subject matter such as mathematics or geography. In addition, gradeschool teachers need to know about child and adolescent psychology in order to understand their students—their strengths and weaknesses, and how best to motivate and encourage them to keep learning. Importantly, through their experience in the classroom, teachers acquire the wisdom of practice (Bransford et al., 2000). Researchers acquire and develop many kinds of knowledge as well. Many classroom-based researchers started their own careers as teachers and then began graduate study to further their knowledge and skills in areas including language acquisition theories, learning theories, pedagogical theory and practice, research design and statistics, curriculum development and research findings from other studies. Both types of knowledge and experience play crucial roles in developing a better understanding of L2 teaching and learning. Acknowledging and valuing each is central to bridging the gap between researchers and teachers. Below we examine how teachers and researchers can work toward this independently.

### **What can researchers do to bridge the gap?**

### *Write for teachers*

When ones surveys the books available about L2 learning and teaching they tend to fall into two broad categories : 1) "how to" methodology texts designed for teachers with no (or very few) explicit links to research and 2) academic texts designed for researchers and post-graduate students with thorough reviews of theory and research and few links to practice. Thus, there is a need for books that present research findings and make relevant links to L2 pedagogy. One way in which researchers can accomplish this is to write about research for teachers in accessible language, in management lengths and in various forms – including articles in teachers' journals and newsletters. For many researchers this is a challenging task because while they find it easy to communicate with their peers (i.e., fellow researchers) it is not easy to do so for a teacher audience. I was recently reminded of this when I became involved as a co-editor of a book series, *The Oxford Key Concepts for the Language Classroom*, designed to present teachers with information about research in L2 classrooms in accessible and meaningful ways and to make direct connections with practice (Lightbown & Spada, 2014–2019). The books are slim volumes each focused on a specific topic (e.g., literacy, assessment, reading, grammar, vocabulary) and targeted to L2 teachers in primary and secondary schools. In order to write these books, there was a need for authors who knew the relevant research in the topic areas and could communicate about the relevance of this research to practice in non-technical and meaningful ways. This was not easy. In the end we were fortunate to have found an excellent group of writers who have completed the nine volumes of the series. All of the authors have a deep knowledge of classrooms in a variety of learning environments and they also know the research. All of them are also teacher educators who have extensive experience and a deep commitment to working with teachers to make a positive impact on teaching and learning.

Writing in more accessible formats can also include a shift from expository to narrative discourse for example, writing case studies (i.e., 'telling stories') of L2 learners and teachers (Freeman & Johnson, 1998), and presenting classroom vignettes and school profiles (Chumak-Horbatsch, 2019). Other writing initiatives that help to bridge the gap between SLA research and L2 pedagogy include the innovative OASIS (Open Accessible Summaries in Language Studies) website (Marsden & Kasprowicz, 2017). This is a database of non-technical, one-page summaries of previously published research studies available for free on the web in a searchable format (e.g., topic, focus, age, language) available for free on the internet (https://oasis-database.org).

#### *Exercise caution generalizing*

Researchers need to exercise greater caution in making generalizations about research findings because what one observes in one context does not necessarily reflect what is going on in other L2 classrooms. Indeed, research findings do not and cannot apply to all educational contexts, teachers, tasks, linguistic features, languages and learners. One example of this is the vast body of research on corrective feedback particularly recasts which are a type of indirect corrective feedback in which a learner's error is corrected while maintaining a focus on meaning and communication. In their research in French immersion classes, Lyster and Ranta (1997) observed that recasts were the preferred method of providing feedback and that students often failed to perceive them as corrective in nature. In recent reviews of a large number of studies it has been reported that the frequency, salience and effectiveness of recasts vary greatly across different teaching and learning contexts. These differences are related to age of the learners, levels of proficiency, the overall orientation of the instructional approach, teachers' language proficiency and other factors (Lyster et al., 2013). Another example of how difficult it is to generalize research findings is reported in a recent descriptive/observational study by Collins and White (2019). Their study investigating language-related episodes (LREs) between learners engaged in pair and small group interaction in intact classes showed that "even though there were many shared features among the three classes (e.g., age and proficiency of students, length and type of program, L1, L2, student familiarity with each other) there were differences in teaching approaches and classroom culture that were associated with differences in the frequency and characteristics of LREs. Thus, while the findings of their research were clearly applicable to the studied classrooms, the question remains to which other classrooms? This is further evidence that ultimately it is up to individual teachers to identify the relative importance of research findings for their own contexts.

Over the years questions have been raised about the concept of generalizability as a feasible, realistic, or appropriate goal for classroom research on L2 learning and teaching (Larsen-Freeman, 1996; Lightbown, 1985, 2000) and other terms have been proposed to replace it. Stenhouse (1975) suggests "provisional specifications" arguing that any recommendation based on research needs to be provisional and must be tested and verified and adapted by each teacher in his/her classroom. Clarke (1994) recommends the concept of "particularizability" which refers to the need for researchers to provide as much information as possible about the research that can be made usable for particular teachers (e.g., details of the conditions of the classrooms studied, characteristics of learners, materials and activities, description of teachers, description of institutional factors). The question that is typically asked with respect to generalizability is: Can inferences be made about the findings and their applicability to the larger population? The question that is typically asked with respect to particularizability is: Can the information from the research be made usable for particular teachers? Clarke also argues that teachers can better evaluate research findings in light of their own situations and determine what is relevant and meaningful in their own contexts.

Of course, it is not just up to researchers to try to bridge the gap, it is also up to teachers. Below we examine some of the ways in which teachers can do this.

# **What can teachers do to bridge the gap?**

#### *Invest in their professional development*

Teachers can invest in their professional development by attending conferences, by joining discussion and support groups with other teachers and by taking courses. In the SLA courses that I have taught to teachers over the years students sometimes come into the course with the impression that it is going to be theoretical, too heavy on research and probably not much good to them. However, students often say that it was one of the most practical courses they took in their teacher education program. What does this mean? How was their teaching helped by what they learned in the course which was not a methodology/pedagogy course? The explanation seems to be that the SLA course provided them with a better understanding of the process of learning and the different factors that contribute to more or less successful L2 learning. This gives them greater confidence in their ability to evaluate their classroom experience, to look for ways to change things that do not seem to be working and to respond to their learners' needs in concrete and practical ways.

Professional development workshops, teachers' conferences and courses are only a first step. Once teachers have heard about new ideas, they need to have someone to discuss these ideas with, in the context of their own classrooms. This kind of professional development is being done in some immersion contexts where researchers work with teachers in the implementation of researchbased pedagogy (Lyster et al., 2018). This is discussed in more detail in the section below entitled *Mediators in the researcher-teacher gap*.

#### *Participate and collaborate in research studies*

Another effort teachers can make to bridge the gap is to participate in research studies either as participants or collaborators. If they are involved as study participants, this is one way of making sure that researchers are better informed about the realities of L2 teaching and learning in classrooms. One of the ways this is best accomplished is for teachers to open their doors to researchers to spend time observing in L2 classrooms before they engage in intervention studies (Spada, 2005; Spada et al., 1996). As we know, teachers do not always trust researchers who say that they just want to "observe" classes. Even when they are assured that the purpose of the observation is to ensure that the researchers understand something about a typical day in a particular classroom, teachers may find it difficult to carry on with "business as usual" when there is an observer in the classroom. One way to mitigate the tendency for teachers to alter their classroom behavior in order to present the most positive picture of the typical day is to spend enough hours in the classroom over a period of time so that both teachers and students become comfortable. In many cases, this can be accomplished when researchers become assistant teachers or participant observers and can circulate among students and interact with them as they engage in individual or small group work. Establishing relationships like these helps to build trust between teachers and researchers and acknowledges that both bring different types of knowledge and expertise to the L2 teaching and learning process.

Collaborating with researchers to investigate questions about teaching and learning is also considered to be an effective way to bridge the research/practice gap but the challenge is obtaining time for teachers to do so. How can this be done? One model has been the establishment of research schools that are often affiliated with universities where the expectation is that research is frequently underway and teachers and researchers work together on research and teaching projects. Another model is 'field sites' in school districts. In these contexts, researchers become more familiar with the questions and concerns of teachers and teachers become more familiar with how to investigate these questions. This leads to research that is "practice-relevant by design rather than as a result of retro-fitting" (Uccelli & Snow, 2008, p. 628). Despite these advantages, it is not easy to establish research schools at universities or field sites in school districts because they require considerable resources and long-term commitments.

Another way in which teachers and researchers establish collaborative relationships occurs more organically and often emerges based on relationships established in other contexts. For example, when teachers decide to 'go back to school' and take advanced degrees they have opportunities to collaborate on research projects with their professors as well as to conduct their own research under the direction of a thesis supervisor. This collaborative work can also lead to joint publications. Some examples of research and writing collaborations within L2 classroom research that have focused on drawing learners' attention to language in meaning and content-based instruction include Kowal & Swain (1994), Doughty & Varela (1998) and Pica et al. (1998).

An example of a research project that grew out of a teacher professional development course that took place in multilingual classrooms in Canada began with a few elementary school teachers who expressed an interest in examining questions about their own practice after the course ended (Denos et al., 2009). The project included a diverse group of participants – teachers, research assistants/graduate students, teachers-in-training, and a videographer. The research questions came from the teachers and they were guided by the professor/researcher who had taught the course and worked to support, facilitate, and inform the work. Some of the factors thought to contribute to the group's success were: 1) the inclusion of diverse viewpoints as represented by the differently placed participants (with the professor and graduate students offering research, the teachers offering stories of experience in working with children); 2) the necessity of each participant to produce 'something' (usually a report of their research for publication in a teacher journal but also videos and workshop presentations to colleagues); 3) an ethos in the group that they were not looking for *answers* or 'best practices' but rather working together to identify questions and ways of gathering data to address these questions, and then documenting experimentation with various responses (Toohey, personal communication).

Another way that teachers and researchers collaborate emerges from long-term relationships that are built on trust and situated in local contexts (Spada & Lightbown, 2019). This characterizes how I engaged in L2 classroom research with my colleague Patsy Lightbown in Quebec schools over many years. Influenced by the research cycle of observation, correlation and experimentation (Rosenshine & Furst, 1973) and process-product research (Long, 1980/1983) we worked closely with teachers to investigate a wide range of questions of mutual interest about L2 teaching and learning (Lightbown & Spada, 1994; Spada & Lightbown, 1989, 2019). The nature of our sustained research in local communities gave us the opportunity to regularly return to the schools and classrooms we had visited before. This enabled us to see first-hand how the instructional practices were evolving over time in relation to the research we had carried out in the classrooms and schools.

There are many other examples of classroom researchers and teacher educators working together in different geographic


**Table 1.** Ways for researchers and teachers to bridge the gap

contexts and research paradigms making connections between research and L2 teaching. Those who have focused on primary and secondary students and their teachers include Mary Schleppegrell and colleagues working with English language learners and teachers in the US (de Oliveira & Schleppegrell, 2015) Roy Lyster and his colleagues working with French immersion learners and teachers in Canada (Lyster, 2018), and Ros Mitchell and Florence Myles with foreign language teachers in the UK (Mitchell, 2014).

Action research is often described as one of the best ways for teachers to engage in meaningful relevant research because it is contextual, small-scale and localized, aims to bring about change and improvement in practice and provides opportunities for collaborative investigation by teams of colleagues including practitioners and researchers (Burns, 1999). As with other types of collaborative research one of the challenges facing action research is getting time for teachers to do it. Institutional support is essential for teachers to be able to take time away from their full-time jobs to systematically investigate questions about their pedagogical practice.

Table 1 presents a summary of what we have discussed so far with respect to the different ways in which teachers and researchers can work independently and together to bridge the gap

between research and pedagogy. Note that the last item in the L2 teachers' column "Carry out their own research by themselves for themselves" is discussed below in the section *Alternative solutions to the researcher/teacher gap*.

# **Mediators in the researcher-teacher gap**

In discussing the researcher/teacher gap Ellis (2010) claims that both SLA researchers and L2 teachers are more likely to come into contact with classroom researchers and teacher educators than they are with each other. Therefore, he argues that it is the classroom researchers and teacher educators who function as mediators between SLA researchers and teachers. As a classroom researcher and teacher educator I agree that both play important mediator roles. Next, we examine some research that has investigated the impact that knowledge obtained in teacher education courses has on L2 teacher practice.

Several studies have explored whether teachers' beliefs about L2 learning and teaching change after participation in SLA courses. Some of this work has included a questionnaire that appears in the introduction to *How Languages are Learned*  (Lightbown & Spada, 2021). In this questionnaire teachers and teachers-in-training are asked to reflect on their views about how languages are learned and what they think this means about how languages should be taught. They are presented with a list of statements that summarize popular opinions about language learning and teaching and asked to indicate whether they strongly agree (SA), agree (A), strongly disagree (SD) or disagree (D) with them. Table 2 includes a selection of statements from the questionnaire.

Studies that have used versions of this questionnaire suggest changes in teachers' beliefs about how languages are learned after participation in SLA courses (Badger et al., 2001; Kerekes, 2001). Other studies using different questionnaires and methodologies also suggest a beneficial role in terms of changes in teachers' beliefs as a result of information about how languages are learned and taught (e.g., Borg, 2009, 2010; Freeman, 2002; Johnson, 1992).

**Table 2.** Selection of opinions about L2 teaching and learning (adapted from Lightbown & Spada, 2021)


*Note:* SA (Strongly agree); A (Agree); D (Disagree) SD (Strongly disagree)

In a study by Erlam (2008), ten general principles of effective L2 instruction were developed based on a survey of SLA research. Included among them are:


Erlam and her colleagues looked for evidence of the 10 principles in the teaching practices in Japanese and French language classrooms in secondary schools in New Zealand. This was accomplished by interviewing the teachers about the principles and asking them how they realized them in their own teaching. The results revealed that discussions about these principles improved teachers' knowledge of L2 pedagogy and gave them a shared language to talk about L2 teaching and learning. Some teachers were motivated to carry out action research based on the principles in their own classrooms. While the overall results were positive, what this study did not do was to investigate whether this new knowledge and awareness led to any changes in their teaching practices. A study by McDonough (2006) set out to do this in a graduate course by engaging graduate students who were also language teaching assistants (TA's) in action research. The TA's were asked to identify topics to investigate motivated by their own practice. Some of the topics focused on approaches to grammar instruction, techniques for encouraging class participation and syllabus design. The students were guided in the design of the research by the instructor and their classmates. The data that were collected included: TA's professional journals, TA's reflective essays, TA's action research reports and instructors' field notes. The results indicated that the TA's broadened their perspectives about research, valued the opportunity to do research rather than just reading research, gained a greater understanding of their L2 classrooms, and importantly, used that understanding to implement new practices to improve their teaching.

In a recent study of the effects of professional development courses on teachers' practice, Tedick & Zilmer (2018) investigated the experiences of 75 immersion teachers engaged in online courses that addressed how to focus on language instruction in content-based language teaching. They explored how the different types and specific features of the activities and assignments that students engaged in during the course affected their practice. The experiences that had the highest impact were ones that involved opportunities to give and receive feedback, to put into practice what they learned, as well as those that resulted in observable changes in student learning and included collaboration and opportunities for reflection. In discussing the limitations of the study, the researchers point to the fact that the data consisted of teachers' reports and perceptions in terms of what made a difference in their practice rather than direct observations and documentation of their actual classroom practice.

# **Alternative solutions to the researcher-teacher gap**

Up to this point we have discussed different ways in which the gap can be bridged between SLA research and L2 pedagogy. Of course this gap cannot be bridged unless both teachers and researchers are motivated to do so and certainly not until teachers' 'practical' knowledge is considered to be equally important as researchers' 'technical' knowledge; teachers are viewed as participants not consumers; and communication is two-way not one-way (i.e., reciprocal). Some scholars argue that this is unlikely to happen because the research- practice model, which has traditionally positioned researchers at the centre and teachers in the periphery, is not likely to change (Clarke, 1994; Freeman, 1996). Indeed, researchers working in critical pedagogy claim that advocating teacher-researcher collaboration is misguided because it simply prolongs the inequality that exists between them (Stewart, 2006). It is further argued that the gap will continue to exist because researchers and teachers inhabit different social worlds and communicate through different discourses (Kumaravadivelu, 2003; Pennycook, 1999).

In response to these concerns another path has been suggested – research done by teachers for teachers. This option moves researchers out of the equation entirely and is consistent with Allwright's (2005) exploratory practice model in which teachers develop their own ways of asking questions and reporting results using classroom activities as research tools. In this process, teachers make their own discoveries about how their own learners learn language creating and adding to their own knowledge base. The argument that the best source of knowledge about teaching comes from teachers is compatible with the Theory of Practice (Bourdieu, 1977; Kramsch, 2015) which breaks down the artificial divide between theory and practice and claims that theories about teaching and learning do not come from the outside e.g., from linguistic or cognitive theories. Instead theory emerges from practice and the building up of knowledge that comes from teachers' reflective practice which is directly relevant to teachers' social contexts. Such an approach represents a significant paradigm shift and is intellectually incompatible with the traditional theory-research-practice model. Does this mean that they are mutually exclusive? I do not think so. In my view, both efforts are needed, that is, efforts to bring researchers and teachers together to work in mutually respectful and reciprocal ways as well as efforts to position and support teachers to carry out their own research and develop their own theories of practice. I believe that a combination of these actions will contribute positively to narrowing the gap between SLA research and L2 pedagogy.

# **Conclusion**

In this paper I have discussed some of the reasons for the gap between SLA research and L2 pedagogy and suggested ways in which teachers and researchers can work separately and together to help narrow this gap. I have suggested that researchers write for teachers in accessible and summative ways, gain a better understanding of teachers' questions and concerns, and exercise caution in generalizing research findings. For teachers, I have suggested that they invest in professional development, decide what research questions and findings are relevant in their own contexts and collaborate with researchers as well as carry out research themselves. In order to make classroom teaching and learning as successful as possible there is no doubt that researchers can benefit from teachers' knowledge and experience of what is important for L2 teaching and learning and teachers can benefit from an understanding of how SLA research can contribute to L2 classroom practice.

# **References**


# **2. Teaching pronunciation: Truths and lies**

*John M. Levis*

# **Introduction**

Pronunciation has been professionally neglected in research and teaching since the advent of communicative language teaching, or CLT (Murphy & Baker, 2015). Pronunciation advocates have often described this using the metaphor of Cinderella, an image first used by Kelly (1969): "It will be obvious that pronunciation has been the Cinderella of language teaching, largely because the linguistic sciences on which its teaching rests did not achieve the sophistication of semantics, lexicology, and grammar until the 19th century" (p. 87). Kelly wrote before the advent of CLT, but the image has been kept alive by later writers. In one influential source, we read about "the Cinderella Syndrome – kept behind doors and out of sight" (Celce-Murcia et al., 1996, p. 323). In a more recent source, we see the image of disregard for pronunciation in language teaching circles. Underhill (2010) writes that "…pronunciation is the Cinderella of language teaching. It has been neglected, and disconnected from other language learning activities." As I and others have argued elsewhere (Derwing, 2019; Levis, 2019), this metaphor has more than run its course and no longer applies because of significant changes in the field of second language (L2) pronunciation, especially evident in journal publications, in conferences, and in professional books devoted to L2 pronunciation.

Journal publications are one marker of the growing visibility of L2 pronunciation. Although pronunciation-related articles have never disappeared, and indeed were always visible in journals that

**How to cite this book chapter:**

Levis, J. M. (2022). Teaching pronunciation: Truths and lies. In: Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (Eds.), *Exploring Language Education: Global and local perspectives*, pp. 39–72. Stockholm University Press. DOI: https://doi.org/10.16993/bbz.c. License: CC BY 4.0.

emphasized applied linguistics' concerns, such as *Language Learning* and *International Review of Applied Linguistics* (Levis & Sonsaat, 2017), journals dedicated to the teaching of foreign languages have had few articles on pronunciation until the past ten years. For example, the quarterly *Foreign Language Annals* started in 1968 as the flagship journal of ACTFL, the American Council of Teachers of Foreign Languages. From 1970–2007, it included just four articles related to pronunciation (less than .4% of total published articles). In a remarkable change, from 2008–2017, there were 22 articles related to L2 pronunciation. This flood of articles reflects not only a change in the journal but also in the interests of its readership. A similar shift is visible within the field at large, namely special issues in other journals in applied linguistics. *TESOL Quarterly* (2005) was the first special issue, with a focus on 'Intelligibility, Identity and World Englishes.' In 2006, *Prospect*, an Australian journal that is no longer published, released its own special issue on pronunciation with a focus on the Australian context (http://www.ameprc .mq.edu.au/resources/prospect/V21\_N1\_2006). Another Australian journal, the *Journal of Academic Language and Learning* (2015) released a special issue on 'New Directions in Pronunciation Theory and Practice.' The high-profile journal *Studies in Second Language Acquisition* (2017) dedicated a special issue to 'Task-based Language Teaching and L2 Pronunciation.' And finally, *CATESOL Journal* (2018) published a large special issue on pronunciation with 14 articles. In another important change, the first journal dedicated to L2 pronunciation launched in 2015. The *Journal of Second Language Pronunciation* (JSLP) started with two issues per year, and moved to three issues in its fifth year. This means that over 90 full-articles have been published in JSLP alone, and the number of pronunciation articles published in other journals has not been affected.

The increasing visibility and presence of dedicated pronunciation conferences is a second change in the field. The oldest conference, New Sounds, started in 1989 and is held every three years at different locations in the world. Increasingly, L2 pronunciation researchers who would previously have attended teaching-oriented conferences make up a substantial proportion of presenters. The International Conference on Native and Nonnative Accents of English, held

annually in Łódź, Poland, was the first dedicated pronunciation conference. The English Pronunciation: Issues and Practices conference began in 2009 and is a biennial conference held in Europe. Also in 2009, the annual Pronunciation in Second Language Learning and Teaching (PSLLT) conference began in North America. This conference also includes a freely-available electronic proceedings, now in its 12th year (https://apling.engl.iastate.edu/archive/). Since then, there have been other irregular conferences such as Sound to Word (Iowa City, USA in 2015), the pronunciation symposium in Wollongong, Australia (2016, 2018) and the Barcelona Pronunciation Workshop (2019) that indicate increasing interest in the field. In addition, major applied linguistics conference such as the American Association for Applied Linguistics now have dedicated strands for pronunciationrelated research.

Finally, there has been a recent explosion of professional books related to L2 pronunciation, moving beyond a focus on English to include books on a variety of other languages such as Chinese (Yang, 2016), German (O'Brien & Fagan, 2018), and Spanish (Rao, 2019), and in connecting L2 pronunciation to other applied linguistics area such as assessment (Isaacs & Trofimovich, 2016; Kang & Ginther, 2017, identity (Beinhoff, 2013; Levis & Moyer, 2014), listening (Cauldwell, 2013, 2017), ELF/EIL (Deterding, 2014; Low, 2016; Nelson, 2012), foreign accent (Moyer, 2013), and pronunciation teaching (Brown, 2014; Derwing & Munro, 2015; Grant, 2014; Murphy, 2017, etc.). I was able to identify only 7 such books for the 1990s, 12 books for the decade beginning 2000, but since 2010 there have been over 60 so far.

Despite all these changes in the field, research still reports that teachers lack confidence in their ability to teach pronunciation, in understanding pronunciation's role in the language teaching classroom, and in knowing what is true and not true about pronunciation teaching and learning (e.g., Foote et al., 2011). The purpose of this paper is to provide reliable information from current pronunciation research and teaching. Having such information may help teachers feel more confident in teaching pronunciation. I will focus on four truths about pronunciation teaching and also discuss corresponding lies that are still commonly accepted by many teachers.


**Table 1.** Truths and lies about pronunciation teaching

# **Four truths and corresponding lies**

In my teaching, I often use ice-breakers to help me get to know my students and them to know each other. A favorite ice-breaker is what I call "three truths and a lie", in which everyone has to write down four statements about themselves, and one must be a lie. All the other students need to guess which statement is a lie, a surprisingly difficult task that has the side benefit of showing learners that they can fool others in their new language. Similarly, pronunciation teaching is full of beliefs that are indefensible yet still fool many otherwise knowledgeable teachers (Thomson & Foote, 2019). The truths and lies addressed in this paper are summarized in Table 1. Each pair is explained in turn, first the lie, and then its related truth.

# *Lie 1 – Teaching speaking and listening is possible without pronunciation*

Implicit in the communicative approach to teaching oral communication skills is that it is possible to teach speaking skills and listening skills without explicitly addressing pronunciation. In pre-CLT methods such as audiolingualism, there was an assumption that if the accurate form of the language was mastered, then communicative fluency would automatically follow. This was not the case. In CLT approaches, the pendulum swung the other way, and an emphasis on fluency was presumed to result in sufficient accuracy. Pronunciation is, at a very basic level, interested in accuracy. Sounds and prosody have to be accurate enough for a listener to interpret them as expressing the intended words and meanings. This may mean that L2 pronunciation might closely match native productions, or it may mean that productions are distinctly nonnative yet interpretable as the appropriate category. Ultimately, accuracy in pronunciation means the communication of meaning, and sufficiently inaccurate forms can cause listeners to hear the wrong words, not understand the intended message, or misinterpret the meaning (Levis, 2018; Smith & Nelson, 1985). Inaccurate pronunciation can also make listeners work harder, that is, it can affect comprehensibility. Although pronunciation is only one aspect of weakened comprehensibility (Isaacs & Trofimovich, 2012), it remains an important aspect. Thus pronunciation, far from being irrelevant to L2 speaking, is always something that is critical in communicating meaning in the L2. When combined with other issues of speech, such as lexical choices or grammar, pronunciation can affect comprehensibility even more (Ruivivar & Collins, 2018).

Pronunciation also is an essential part of listening comprehension. Like speaking, listening moved from a micro-focus on accurate form in older materials (e.g., Morley, 1973) to a focus on macrolistening strategies and understanding the meaning expressed by speakers. Field (2008) describes these two aspects of listening as the bricks and mortar of speech. The bricks are the content words, those that are stressed, more easily heard, and less predictable. The mortar includes the function words, which are unstressed, harder to hear, and more predictable. Like speaking, listening instruction has suffered from myopic methodologies. Both macro-listening and micro-listening are important in L2 listening abilities, but a dominant focus on macro-listening and content word listening is almost always accompanied by a diminished focus on the role of pronunciation in successful listening, and on the function words that provide connections between content words. L2 listeners still need to negotiate the pronunciation issues that allow them to understand words in speech, that allow them to understand connected speech in careful and casual speech, and that help them to understand new voices (Cauldwell, 2018). This means they need help in noticing the role of pronunciation in identifying words in speech.


**Figure 1.** Listener-Speaker Intelligibility Matrix (adapted from Levis, 2005)

## *Truth 1 – Pronunciation is unavoidable and essential*

The first truth, that you cannot speak a language without pronunciation, is self-evident but often forgotten. A related truth is that you cannot understand others without understanding their pronunciation during both careful and casual speech. This means that for anyone who wants to speak another language, pronunciation is critical. Speaking another language is power, and pronunciation is the face of that power. It is the first thing listeners notice, and it is the most basic level of language form that allows communication to take place. As Hinofotis and Bailey (1981) describe it, there is an intelligibility threshold which, if not met, stops communication from truly taking place. Spoken intelligibility has, at the very least, both speaking and listening components. In Levis (2005), I conceptualized this in terms of a matrix with four quadrants (Figure 1) in which native and nonnative speakers play both the role of listener and speaker.

In Quadrant A, native speakers of a language speak to other native speakers. Typical assumptions for this quadrant are that communication breakdowns between native speakers and listeners will be rare, but we actually know very little about whether this is true. Dialects of the same language can be very different, and intelligibility between dialects is not guaranteed. In Chinese, for example, what most linguists would consider different languages are termed dialects, and spoken intelligibility is unlikely, but written intelligibility is likely to be certain. In English, different

dialects of the same variety such as American English are likely to be largely intelligible, but varieties of English are not always going to be easily intelligible. Still it is certain that, apart from lexical and grammatical differences, pronunciation differences are likely to play a significant role in the extent to which speakers and listeners will understand each other. Larger and more unexpected differences are likely to cause greater loss of understanding.

Quadrant B, with nonnative speakers and native listeners, is the quadrant that represents most studies of intelligibility, accentedness, and comprehensibility. This quadrant assumes that NSs are the ones who are the best decoders of spoken language, and nonnative speakers are the ones who are tasked with delivering spoken language that is decodable. Most studies of this sort make an assumption that the success of an interaction is based on the ability of nonnative speakers to make themselves understood, preferably by using a reasonable representation of a well-known native speaker accent, e.g., for English, some variant of Received Pronunciation or General American English.

Quadrant C, with native speakers and nonnative listeners, is the quadrant of language teaching and nonnative speakers encountering native speech in natural communication. Intelligibility is typically thought of as nonnative speakers having to make themselves understood to native listeners who are able to understand speech that is well delivered. But Quadrant C is an equally critical aspect of pronunciation teaching, because being able to understand the details of pronunciation in native speech is perhaps harder than making one's self understood (Rivers, 1981). Indeed, this type of listening was the core of early types of listening instruction such as found in Morley (1973), where micro-listening practice was focused upon understanding pronunciation differences between past and present tense, singular and plural forms, and many other aspects of listening that reflect pronunciation distinctions. After decades of being deemphasized, micro-listening related to pronunciation has been making a comeback because of evidence regarding the role of pronunciation in L2 listening comprehension (Cutler, 2015; Field, 2003, 2008), in regard to role of particular pronunciation features in intelligibility (Levis, 2018), and in regard to the significant role of pronunciation variations in understanding normal connected speech (Brown & Kondo-Brown, 2006; Cauldwell, 2013, 2018; Johnson, 2004; Levis, 1999; Shockey, 2003).

The final quadrant (Quadrant D) reflects an area in which pronunciation is also critically important, especially in regard to the role of English in the world. In this quadrant, we have situations in which NNSs use the target language to communicate with other NNSs, either because they do not share a common L1 or because the context requires the use of the target language (TL), such as Aviation English. Jenkins (2000) convincingly demonstrated that pronunciation is perhaps the most common reason for loss of intelligibility. Thus in all four quadrants, pronunciation is both central to communicative success and unavoidable.

If anything, this intelligibility matrix oversimplifies the extent to which pronunciation is central to successful communication. NSs and NNSs are not all the same, and languages like English also include a middle ground in which English has become nativized, as in countries like India, Singapore and Nigeria (Kachru, 1992; Nelson, 2012). In such countries, these new Englishes are closer to native varieties than to nonnative, meaning that new, stable pronunciations become part of the listener-speaker equation. In addition, proficient speakers have multiple speech styles, from the most casual to the most careful, each of which has different pronunciation patterns that will affect intelligibility.

#### *Lie 2 - Pronunciation does not need to be taught*

This lie about pronunciation is simple and influential. Going back to the earliest days of the communicative and comprehensionbased approaches in the 1970s and 1980s (Levis & Sonsaat, 2017; Murphy & Baker, 2015), there were assertions that because adult learners could not become nativelike, pronunciation should not be taught explicitly. Instead, learners should be taught to communicate in the L2. Pronunciation weaknesses would then be compensated for by successful use of the language in other ways, and pronunciation would even improve because of communicative improvement. In other words, pronunciation would take care of itself. It slowly became clear that this was not true. One of the most important areas of evidence for this was research being done with foreign teaching assistants (FTA) in the United States. These FTAs came from all over the world to study for their Master's and PhD degrees. As part of their graduate work, they taught classes in their fields to American undergraduate students, who were not experts and were studying the topic at the lowest level in general education requirements (e.g., basic chemistry or biology). Bailey (1984) reports that the dominant approach to pronunciation was to simply leave it alone, because "pronunciation patterns were resistant to change" (p. 4):

Pronunciation is a problem with high visibility. TAs often regard it as their highest priority in terms of language mediation. Students frequently cite it as the biggest obstacle to understanding. With TAs from certain countries (India is perhaps the best example), pronunciation problems mar what is otherwise highly proficient speech (Bailey, 1984, p. 34).

Hinofotis and Bailey (1981), in an earlier report, said research indicated there was "a threshold of intelligibility" (p. 124) in FTA speech, and that pronunciation was not a binary issue (native vs. nonnative) but rather a continuum along which speech could be more or less understandable.

A related thread that led to a marginalization of pronunciation teaching during this time was the idea that teaching language form was either not productive or even counterproductive to unmonitored use. This is best known from Krashen's Learning-Acquisition dichotomy (e.g., Krashen, 1977, 2013). The Learning /Acquisition distinction asserts that L2 learners have access to two different types of knowledge, that provided by instruction (Learning) and that provided by meaningful use (Acquisition). Learning is promoted by conscious improvements that are the result of explicit instruction, while Acquisition is unconscious development that comes from responding to the appropriate level of input. Learning does not become Acquisition, that is, language learned by attention to form is not available to the learner in unconscious, meaningful use of the language. This assertion was always questioned by other researchers (e.g., McLaughlin, 1978) as being too strong and based on insufficient evidence. With succeeding years, it has become clear that Focus-on-Form is critical for all kinds of second language development (e.g., Doughty & Williams, 1998), including pronunciation (Saito & Lyster, 2012).

A second problem with the view that pronunciation does not need to be taught is the question of fossilization. First put forth by Selinker (1972), fossilization has sometimes been used to describe the lack of development in pronunciation for longterm learners of a language (Derwing & Munro, 2014). Some research indicates that naturalistic pronunciation learning levels off after the first year in the L2 environment (Derwing et al., 2006), but there is also evidence that the pronunciation of longterm learners can be "defossilized" through instruction (Acton, 1984; Couper, 2006), and that changes can be evident even to naïve listeners (Derwing et al., 1998; Gordon & Darcy, 2016). In one recent study addressing fossilization, Derwing, Munro, Foote, Waugh and Fleming (2014) looked at the effects of pronunciation instruction for Southeast Asian workers in a window factory in Canada (L1: Vietnamese, Khmer). The workers had lived an average of 19 years in Canada, and they had significant pronunciation problems. For example, one worker was reported to say 'stockitt' for the word 'target', an error that caused significant intelligibility problems. The workers received 17 hours of pronunciation instruction over 3 months, with out of class assignments. The results showed intelligibility improvements both in their perception of spoken English and in their spontaneous speech. In other words, they were not really fossilized.

### *Truth 2 – Pronunciation teaching works*

The second truth about pronunciation teaching is that it is successful. Learning can occur in the absence of teaching, given the right timing, environment, motivation, etc. But this kind of improvement is limited and is most obvious within the first year being surrounded by the L2. Such a window of maximal opportunity (Derwing & Munro, 2015), once closed, does not mean L2 learners cannot improve their pronunciation. Multiple articles demonstrate that pronunciation need not fossilize, even after long periods of no improvement (e.g., Derwing et al., 2014).

Indeed, L2 learners, even those who have shown little naturalistic development, almost always improve when they practice and are instructed with some degree of regularity. For years, a critique about pronunciation teaching is that we had no certainty about its usefulness. From the Critical Period Hypothesis (Scovel, 1969) to the anti-pronunciation bias of communicative and comprehension-based approaches to teaching (Levis & Sonsaat, 2017; Murphy & Baker, 2015), there have been questions about whether pronunciation teaching was worth the time because of the evidence that adult L2 learners only rarely achieved a native-like accent (Birdsong, 2007; Bongaerts et al., 1997; Brinton, 2012). Reports of nativelike passing ability are short-lived and can be maintained for service encounters (Moyer, 2014; Piller, 2002).

This paper does not question these findings but instead argues that nativelikeness is an inappropriate goal for pronunciation teaching and learning. Although few ever achieve this level of pronunciation (Abrahamsson & Hyltenstam, 2009), becoming completely nativelike is unnecessary to intelligibility or communicative effectiveness. There is abundant evidence that even highly accented L2 speakers can be completely intelligible (Munro & Derwing, 1995).

Claims that pronunciation teaching works mean that L2 learners who are instructed in how to produce or perceive segmentals and suprasegmentals in the L2 almost always improve their production and perception of those features. Although some types of instructional interventions are likely to be more effective than others, e.g., suprasegmentals over segmentals (Derwing et al., 1998; Gordon & Darcy, 2016), the overwhelming finding about pronunciation instruction of all kinds is that it works, and often works well (Lee et al., 2015). The findings of a number of recent analyses of pronunciation studies has demonstrated this. Saito (2012) identified 15 quasi-experimental pretest/posttest studies of the effect of instruction on pronunciation improvement. Saito found that most instruction resulted in improvement, whether the instruction focused on segmentals or suprasegmentals, that improvement was more likely to be evident in controlled rather than spontaneous speech, and that focus of instruction was likely to effect the degree to which improvement was noticed in spontaneous speech. For the few studies that included control groups, meaning-oriented instruction alone did not result in pronunciation improvement. Only studies that focused explicitly on pronunciation found improvement.

A meta-analysis of 86 pronunciation studies conducted by Lee et al. (2015) found more compelling evidence for the positive effects of pronunciation instruction. The 86 studies included a wide range of instructional interventions. Overall, the metaanalysis found that there was a large effect for the results of pronunciation instruction. These effects were strongest for longer periods of instruction, interventions that provided feedback on learner language, and for more controlled language use. In other words, it appears that pronunciation instruction done over a longer period of time (up to 15 weeks) is likely to be more successful, especially if learners are given specific feedback on their pronunciation. And not surprisingly, when learners are asked to pronounce in contexts where they can focus on pronunciation form more completely, they are more likely to demonstrate improvement than when they have to juggle communicative goals and attention to pronunciation form. L2 pronunciation includes a strong element of automaticity, and controlled production is likely to improve more quickly than spontaneous production, in which attention to meaning overwhelms attention to form and leads to L1 pronunciation automaticity rather than L2.

In a narrative analysis of 75 pronunciation studies, Thomson and Derwing (2015) looked at the effect of pronunciation instruction in classrooms and in computer-assisted contexts. A narrative analysis differs from a meta-analysis in focusing more on qualitative differences between the methodology of different studies and not the quantitative analysis of success. Thomson and Derwing agreed with Lee et al. (2015) that most instruction was successful in promoting better pronunciation, but critiqued the studies based on their definitions of improvement, their targets, the types of tasks used, and the inclusion of control groups. First, most studies defined improvement in terms of whether L2 learners became more native-like in their production of the targeted pronunciation features. This focus on nativeness is different from a more appropriate goal of intelligibility, setting up a conflict between what knowledgeable pronunciation teachers and researchers regard as the ultimate goal of instruction and an outdated view of improvement. Second, most studies they examined focused on segmentals rather than suprasegmentals, despite evidence that suprasegmentals are more likely to lead to improvements in comprehensibility and intelligibility in the short run, while focusing on segmentals, even the most important segmentals, may lead to improvement that is not noticed in spontaneous speech (Neri, Cucchiarini & Strik, 2006). Third, there was a strong bias toward reading aloud. This is understandable from a perspective of experimental protocols, but reading aloud is a different type of spoken performance from spontaneous speech, and ultimately learners need to be intelligible in normal communication, not just in reading aloud. Thus, Thomson and Derwing argue for greater attention to studies that emphasize the gold standard of improvement that can be noticed by listeners (Derwing & Munro, 2009). Finally, the authors argue for the inclusion of control groups in pronunciation improvement studies because without control groups, the evidence for improvement over time will always be uncertain.

#### *Lie 3 – Teachers cannot teach pronunciation*

A more basic problem with teaching pronunciation has been reported in many articles, that is, that teachers express that they lack the training and confidence to teach pronunciation (e.g., Breitkreuz et al., 2001; Macdonald, 2002). Teacher beliefs are known to affect the ways that teachers approach pedagogy in many areas of language teaching (e.g., Farrell & Ives, 2015; Richards et al., 2001). One reason pronunciation is different is the historical state of language teacher training more generally, in which pronunciation has not been part of most teacher training programs for decades, leaving practicing teachers with the task of developing expertise through workshops and reading while they are teaching. Most do not do this. Since coursebooks also do not integrate pronunciation well, teachers often have inadequate models of practice (Levis & Sonsaat, 2016; Sonsaat, 2017). There are other reasons as well for teachers' reluctance. Pronunciation teaching often does not result in immediate changes in learners' performance. Research shows that accents of adult language learners rarely change much in naturalistic (Derwing et al., 2006) or instructed contexts (Derwing et al., 1997). If teachers take a nativeness approach to pronunciation teaching (Levis, 2005), this evidence can be discouraging. But there is little evidence that accentedness is directly related to understanding (Derwing & Munro, 2015). In other words, if teachers think pronunciation is about perfect accents, not communicative effectiveness, they are less likely to teach it.

Another reason that teachers may feel that they cannot (or need not) teach pronunciation is that they worry about how pronunciation changes may affect a learner's identity. This concern is not new (Pennington & Richards, 1986; Zuengler, 1988), but it sometimes is taken to mean that L2 learners should not have to change their pronunciation because it will change their identities (Golombek & Jordan, 2005). Such a concern seems to be overblown (Levis, 2015), and second language pronunciation work is always tied to imagined identities (LeVelle & Levis, 2014) as shown in research on passing as native speakers (Marx, 2002; Piller, 2002) and in the effects of social networks on pronunciation (Lybeck, 2002).

A more mundane reason for not teaching pronunciation is that teachers worry about being boring. In my experience, this is one of the worst reasons to avoid something important. Being boring has little to do with whether a topic is important, and learners often perceive boring and interesting quite differently from teachers. Pronunciation teaching and learning can be fun, no doubt, as is clear from many innovative materials (e.g., Hancock & McDonald, 2017; Yoshida, 2016), but this should never be the primary reason to include anything in a language classroom.

As we see from the research about pronunciation teaching's results in Truth #2, pronunciation teaching works. The many, many studies on pronunciation improvement likely include both fun and boring approaches to teaching, and the overall finding of these studies is that learners improve when they are taught. This is perhaps the most critical reason to teach pronunciation. Learners usually want help with their pronunciation. They recognize that it is central to how they are perceived, especially in the L2 environment (Gluszek & Dovidio, 2010). They often do not know what is wrong, nor what matters, nor how to participate in the L2 community of practice (Lave & Wenger, 2001). Our teaching, even when it is not perfect or extensive, can help them with their intelligibility, especially if we pay more attention to pronunciation features that make a difference such as suprasegmentals (e.g., stress, melody, rhythm), high functional load segmentals (such as /l/-/r/, /p/-/b/), and varied types of practice activities, from formbased practice to communicative practice and connections to other language skills.

### *Truth 3 – There is always a way to teach pronunciation*

A common objection to teaching pronunciation is that there is no time to do so. Teachers' class times are filled with many requirements, from the teaching of different language skills to completing testing requirements. Pronunciation may then seem like just another thing for which there is no time. This objection is specious, at least when phrased this way. Pronunciation is not taught because many teachers do not see a compelling reason to do so. Teachers always find time to teach what is important, either because a curriculum prioritizes it, an important test requires it, or because a teacher finds a topic essential to what they think learners should master. Another reason that the argument is specious has to do with an inadequate understanding of what pronunciation is and how it relates to other aspects of language that are already being taught.

As already discussed, pronunciation is an essential part of any person's communicative skills. This means that pronunciation is not simply a matter of mastering decontextualized forms and imitating a model accent, but of becoming understandable and understanding others. This means that while pronunciation may sometimes require focused attention, it does not need to be always taught apart from other aspects of language. If students are speaking and listening, they have to pronounce or understand others' pronunciation. If students are reading or learning vocabulary, they have to be able to connect the written form of words and sentences to their spoken forms. In other words, pronunciation fits naturally within other areas of language, and should mostly be taught that way. Jones (2016), in a book highlighting the ways in which pronunciation can be integrated with other aspects of language, includes chapters on reading, vocabulary, listening/speaking, presentation skills, general listening, grammar, and spelling. One of the advantages of integrating pronunciation with other aspects of language is that the relevance of pronunciation is usually more obvious when it is integrated than when it is taught alone. Another advantage is that integrating pronunciation takes little extra classroom time. It will take more time to prepare, especially when a teacher first starts trying to integrate pronunciation into established patterns of teaching. But this is true of any aspect of teaching, and is not unique to pronunciation. Following are a few ideas for how pronunciation can be integrated with the teaching of other skills.

# **Pronunciation of formulaic sequences**

Corpus studies indicate that a lot of everyday language is relatively formulaic, that is, we use similar or identical ways to communicate in routine situations. One routine situation that is important for language classes and for daily life is the use of self-introductions and repeated questions. For these routines, the pronunciation feature of prominence, or sentence stress, is essential. The example activities that follow have been successfully used in ESL classes from beginning to advanced (Levis & Muller Levis, n.d.). The task itself is important because it provides the opening for L2 learners to continue speaking. Being able to introduce oneself and have another person hear your name, even if the name is unfamiliar, is a crucial first step to interaction. I am not saying that L2 learners do not know how to say their names. Rather, they often do not know how to package the prosody of their names in a way that will help interlocutors hear it in English. Similarly, repeated questions (e.g., How ARE you? FINE. How are YOU?) are also formulaic in prosody and help L2 learners negotiate small talk routines. Both self-introductions and small talk make further conversation more possible.

These activities are built into an ice-breaker which can be done and recycled over several days, depending on the level of the learners. They are presented in a dialogue that plays out in some form or another in many daily conversations where two people meet for the first time.

### *Activity 1 – Ice breaker*

Formulaic sequences are relatively invariant routines. This icebreaker includes two sequences: self-introductions and repeated questions.


Formulaic sequences often have formulaic pronunciation. In this case, certain words have prominent syllables that are pronounced with pitch changes and extra syllable length. These pronunciation features create a melodic shape that helps listeners understand what you are saying. This means that your speech is much more likely to be intelligible.


Notice that the last word in each sentence has a prominent syllable. This happens in most English sentences, and it always happens for names.

Prominent syllables may move away from the end of the sentence depending on the structure of the conversation. For example, in questions that are said by both people (called Repeated questions), the melody of the sentence always shifts to the word "You" or "Your" when the question is repeated.


### *Activity 2 – Extending activity 1 with less control (Muller Levis & Levis, 2014)*

It is also essential for learners to be able to adjust to unexpected changes and to be able to speak more freely, so self-introductions and repeated questions should be practiced with less control and with using their own names. A dialogue in which learners use their own information allows them to do this and can be repeated with multiple interlocutors.


There are many other variations that can be used to practice repeated questions that can be integrated into later practice on question formation and interacting in formal yet formulaic situations (e.g., mixers). The point of these examples is that there are ways to give pronunciation a communicative purpose from the very beginning of instruction. This is often not done because teachers do not recognize the ways in which pronunciation features such as prominence help communicate meaning.

It should be pointed out that hearing and saying one's name, which is important in an ESL context, is likely to be just as important in other contexts, albeit with modifications about how names are pronounced. Speakers in ELF interactions, in which two or more L2 speakers of English interact, may also find names challenging but the prosodic features that allow L1 English speakers to hear names more effectively may not be operative in the same ways. Additionally, it should also be evident that L1 speakers of English may need to adjust their own presentation of their name's pronunciation in order to be effectively understood, depending on the communicative context. Ultimately, speakers in any interaction will need to converge on pronunciations that promote understanding and allow interactions to move forward.

#### *Activity 3 – Spelling your name aloud*

Self-introductions are formulaic in other ways. For example, learners may need to spell their name aloud when they are on the phone or at an appointment, especially if their name is unfamiliar to a listener. As a result, rather than repeating the name the same way, it helps to have a strategy available to be understood. This means that it is helpful to be skilled in spelling one's name aloud. Such oral spelling is often taught early in language learning, then largely ignored in more advanced classes.

In addition, because many letter names are hard to understand, even for native speakers, it helps to give a key word so listeners will know which letter is being used. A common strategy is to say the letter and then exemplify it with a common name or word clue. For example, "S as in Sad." (Of course, it is critical that the word clue be pronounced correctly!)

Other common examples are


#### *Activity 4 – Pronunciation and pragmatics in speaking*

This next activity also involves the use of prominence, in this case contrastive prominence, but it is intended for more advanced learners because of its interaction with the use of politeness while disagreeing or correcting someone else and its assumption of differing power dynamics. Correcting someone else is always face-threatening, but it is especially face-threatening if the person is of a similar or higher status. The activity is an adaptation of an exercise described by Kenworthy (1987).

Background: A major corporation is cutting back on its operations. A company spokesperson is meeting with an aide to go over the details of the information about the cutback. The spokesperson has previous information about the cutbacks, but it is has been replaced by updated information.

Directions: Each pair of students receives cards for the spokesperson and the aide. The spokesperson card starts the role play by checking the accuracy of a piece of information. The aide uses the information on the card to correct mistaken information by using appropriate prominence to highlight contrasting information. When the aide corrects information the spokesperson has, the aide should use hedging devices to be pragmatically appropriate. Some of these devices include:


### Example

Spokesperson: "So, I see that we have three factories CLOSing." Aide: "Actually, it's not quite that high; it's only **2** factories."

### **Spokesperson's card**


### **Aide's card**


# *Activity 5 – Connected speech listening*

Listening comprehension is often connected to pronunciation features, as in the following example using a cloze dialogue. The purpose of this activity is simply to hear the words that are difficult to hear in conversational speech. This type of activity can be done without extra preparation, or it can be done after asking students to predict the missing words and then listening to confirm their own expectations. Because the missing words in the dialogue are all unstressed and many have deleted [h] at the beginning of some words, such a prediction task is likely to make the task more doable, as in the example below.

Listen to the dialogue and fill in the missing words.

Jim: Did you hear about Al?

Joe: No. What \_\_\_\_\_\_\_\_\_\_ done now?

Jim: He totaled \_\_\_\_\_\_\_\_\_\_car.

He ended up in \_\_\_\_\_\_\_\_\_\_hospital.

Joe: The hospital? How bad \_\_\_\_\_\_\_\_\_\_ ?

Jim: How bad? I'm not sure.

I think he's \_\_\_\_\_\_\_\_\_\_ cast. He might \_\_\_\_\_\_\_\_\_\_ laid up \_\_\_\_\_\_\_\_\_\_ couple \_\_\_\_\_\_\_\_\_\_ months.

Teachers often say they do not teach pronunciation more because pronunciation is not connected to other things they have to teach. These few examples of how pronunciation can be connected to other language skills are only a beginning of an answer to these objections. There are many other examples of the integration of pronunciation with other language skills. Books like Jones (2016) provide extended examples of different possibilities, while books like Murphy (2017) provide models of how to approach the teaching of pronunciation in whole courses in varied contexts. What these books say is clear. If you want your students to be able to communicate, pronunciation should be an essential part of what is taught. It does not have to be the only thing taught; pronunciation teaching works, even in short interventions (Gordon & Darcy, 2016; Levis & Muller Levis, 2018). Learners can become more intelligible, even at beginning levels of spoken proficiency (Zielinski & Yates, 2014).

### *Lie 4 – Adult language learners can sound like native speakers*

Accent raises an issue that is crucial for L2 learning, teacher beliefs, and student expectations. Many learners want to sound native. They think it is possible if they just try (Levis, 2015). And teachers often think that this is an appropriate goal even if they do not believe it possible. The reasoning that underlies this is that a teacher should not try to stop a learner from aiming as high as they want, and that nativeness is a high goal. Unfortunately, adult L2 learners rarely become nativelike beyond restricted contexts of being able to pass, as in service encounters (Piller, 2002). A focus on nativeness can also lead to wrong thinking about what is possible and whether nonnativeness is a pathology that can be reduced, neutralized or modified, rather than a normal aspect of language learning (Thomson, 2014).

The mythological status of nativeness is particularly surprising when considering English pronunciation. English is a collection of accents that are very different from each other. Besides the many native accents, studies of World Englishes have shown us that there are many nativized accents such as varieties of Indian English (Pandey, 2015), Singaporean English, Nigerian English, etc. These different accents are especially appropriate models in their own contexts, as well as being appropriate though less familiar than well-known L1 accents in other areas of the world. The multiplicity of different accents point out that our most basic goal when speaking a language, whether L1, L2, or L3, is to be understood and to understand, that is, to be intelligible, not to match a particular well-known accent (Levis, 2005).

It is quite possible to be understood in a foreign language even when you do not sound native (Munro & Derwing, 1995). It is also possible to be misunderstood when you sound quite nativelike. In speaking across accents, everyone may need to converge toward variants that promote intelligibility (Jenkins, 2000; Low, 2014; Walker, 2010). For speakers whose accents are not stable or expected in a particular context, especially in immigrantreceiving contexts or in professional contexts, pronunciation issues that promote intelligibility will likely be different than in contexts that do not share the same social contexts.

#### *Truth 4 – Everyone has an accent, but not all accents are socially equal*

Accent primarily involves pronunciation. It is simply part of speaking a language. There are native accents and nonnative accents, and they can be intelligible depending on who the speaker and the listener are. This means that there are different accents, not wrong and right accents. World languages like English have many native accents. These accents may be more or less intelligible to each other, but they are all native accents. Nonnative accents vary because a language learner's L1 affects the kinds of features that are pronounced differently in the L2. Thus, we can speak of a French or Japanese or Russian accent because the typical differences in speaking the L2 are identifiable by even moderately aware listeners. Like native accents, nonnative accents are not easily changed, which is why most adult L2 learners continue to sound different from native speakers. Because nonnative accents do not change easily, those who promise to reduce an accent, fix an accent, or otherwise change a bad accent into a good accent, promise something they cannot deliver.

Despite the ubiquity of accent, there are other ways in which accents are not simply part of speaking a language. This is because pronunciation is socially significant, and accents that are communicatively equivalent are often socially judged. Even particular words may be stigmatized, such as *wash* being pronounced *warsh* in the United States. In the wider society, some accents and pronunciations are socially valued, while others are socially stigmatized. And all accents are valued and stigmatized in different social contexts, depending on the fit between the accent and the values of the community and the context in which the accent is used. The same is true of L2 accents (Gatbonton, Trofimovich, & Magid, 2005). In a context in which the speaker's L1 is not dominant, a nonnative accent may be less valued. In other words, accent includes both pronunciation patterns and indexical information. Accent is, in the words of one writer, "a set of dynamic segmental and suprasegmental habits that convey linguistic meaning along with social and situational affiliation" (Moyer, 2013, p. 11). As such, accent both serves communicative needs and signals information about social affiliation, marking in-groups and out-groups.

This social aspect of accent is not well understood in pronunciation teaching, but it may be a powerful aspect of pronunciation improvement. We know that certain aspects of pronunciation are more socially noticeable, and that such aspects carry higher sociolinguistic power. These kinds of features may be particularly important in being heard as a legitimate speaker, that is, in being recognized by becoming audible. For example, Miller (2003), in a study of immigrant high school students in Australia, found that some L2 speakers were more successful at becoming audible to others through a combination of pronunciation, lexical choices, and other elements such as humor. Audibility allowed the L2 speakers to access the language of the community and improve their language abilities, as well as to construct and express their L2 identity. Lack of audibility, in contrast, was associated with nondevelopment or even shriveling of L2 skills and identity construction. We also know that L2 speakers often get better at pronunciation when the social context encourages it. In Lybeck (2002), American women married to Norwegian men in Norway had different pronunciation success in pronouncing the Norwegian /r/, a sociolinguistically marked feature in Norwegian. The key to success seemed to be whether the women had a strong social network involving their mothers-in-law and other women. Those who had weak social networks and felt like outsiders were more likely to retain their English /ɹ/, often deliberately, as a marker of their outsiderness.

We also know that accent is both a matter of affiliation and hiding affiliation. Canadian Nicole Marx (2002), in a first-person account of her accent journey, recounts her study abroad in Germany. Because people mistook her as American, she initially took on an accent related to her second language, French, in order to avoid being characterized as American. As she continued to try to sound native-like in German, she first began to dress like other German college students so that she didn't look the part of a foreigner, then worked on aspects of her German that suggested that she was less foreign than she felt. These aspects often included sociolinguistic markers related to German dialects. Marx found that she could sometimes pass as a native speaker in limited contexts. She also found that her German accent spilled over into her English when she returned to Canada, and that it took some time for her old social context to reflect itself in her speech.

The choice of pronunciation model is also sociolinguistic. Models matter, but not in the way we think. Typically, pronunciation is taught with reference to models such as Standard Southern British or General American, but these may not be the models that are relevant to learners. In Cutler (2014), Ukrainian immigrant youth in New York City gravitated toward the speech and accent of hip-hop artists, even when they did not have a personal social connection to that community. Rather, the speech was socially important in the type of community they aspired to. When I first read this study, it struck me that the young people's choice made perfect sense as a matter of their imagined identity, but that I would struggle to be able to teach them because of my own accent and assumptions about pronunciation goals. However, it does suggest that the choice of model could be more local than most pronunciation materials can accommodate. Teachers and L2 speakers in Scotland should be able to use a Scottish variety, those in Australia an Australian variety, and those in the US South a southern variety. In other words, pronunciation should be socially appropriate to the context and social group norms.

# **Conclusion**

The teaching of pronunciation, after a long time in which it was neglected, has become much more visible again in language teaching. However, not all information about pronunciation teaching and learning is accurate. We know that pronunciation is an unavoidable part of language that does not simply improve without instruction, that intelligibility is an appropriate goal for L2 pronunciation and that a focus on nativeness is an unnecessary goal for teaching pronunciation. We know also that pronunciation teaching leads to improvement, that pronunciation can be taught effectively by integrating it with other language skills, and that pronunciation is a socially connected skill. All of these aspects suggest a growing influence for how pronunciation is approached in language teaching.

# **Acknowledgements**

Many thanks to Tim Kochem for his keen eye and thoughtful feedback on an earlier version of this paper.

# **References**


Slavic speakers. *System, 34*, 183–193. https://doi.org/10.1016 /j.system.2006.01.005


& J. Clarke (Eds.), *Supporting lifelong learning, Volume 1* (pp. 121–136). Routledge.


Zielinski, B., & Yates, L. (2014). Pronunciation instruction is not appropriate for beginning-level learners. In L. Grant, (Ed.), *Pronunciation myths: Applying second language research to classroom teaching* (pp. 56–79). University of Michigan Press.

Zuengler, J. (1988). Identity markers and L2 pronunciation. *Studies in Second Language Acquisition, 10*, 33–49. https://doi.org/10.1017 /S027226310000694X

# **3. Multilingual versus monolingual classroom practices in English for academic purposes: Learning outcomes, student attitudes, and instructor observations**

*Amanda Brown, Robert J. Lally & Laura Lisnyczyj*

# **Introduction**

There is no dispute that to learn English or any other new language effectively in both a second or foreign language context, one must receive input in, produce output in, and interact with others in the second language (L2). However, uncertainty exists around the amount of input, output, and interaction necessary and sufficient for successful second language acquisition (SLA). Perhaps to mitigate risk, most contemporary theories of language pedagogy advocate for maximal use of the L2 with minimal/no use of the first language (L1) (see discussion in e.g., Ellis & Shintani, 2014). Indeed, institutional policies and practices mandate L2-only use in both second and foreign language classrooms (Copland & Neokleous, 2011; Debreli, 2016; Jenkins, 2010; McMillan & Rivers, 2011; Rivers, 2011; Sampson, 2012; Shvidko, 2017). However, exclusive classroom use of the L2 has been labeled a "monolingual" view of pedagogy (Hall & Cook, 2012), and is drawing criticism. Recognizing this discord, Macaro (2014) has stated that the decision between using a fully 'monolingual', 'immersive', or 'English-only' pedagogy versus a 'multilingual', 'non-immersive', English plus L1/ other language pedagogy, alternatively labeled 'translanguaging'

**How to cite this book chapter:**

Brown, A., Lally, R. J. and Lisnyczyj, L. (2022). Multilingual versus monolingual classroom practices in English for academic purposes: Learning outcomes, student attitudes, and instructor observations. In: Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (Eds.), *Exploring Language Education: Global and local perspectives*, pp. 73–119. Stockholm University Press. DOI: https://doi.org/10.16993/bbz.d. License: CC BY 4.0.

(e.g., Canagarajah, 2011; Creese & Blackledge, 2010; García, 2009; García & Wei, 2014; Hornberger & Link, 2012; Lewis et al., 2012a, b), 'translingualism' (Canagarajah, 2013; Horner et al., 2011), 'codeswitching', 'dynamic bilingualism' 'fluid languaging practices', and 'plurilingualism' (e.g., García & Kano, 2014; Piccardo, 2013), is "probably the most fundamental question facing second language acquisition (SLA) researchers, language teachers, and policymakers in this second decade of the 21st century" (p. 10). In this chapter, we address this fundamental question, contrasting a monolingual versus a multilingual pedagogy and empirically investigating the effects on learning outcomes, student attitudes, and instructor observations.

# **Background**

The issue of language use in L2 classrooms has garnered considerable attention, especially in recent decades. Cook (2001) traces the belief in maximizing L2 use and minimizing or eliminating L1 use as far back as the 1880s. As a result, in contemporary research, language teachers have described a sense of guilt associated with L1 use (Macaro, 2009), especially when cautioned against overreliance on the L1 (e.g., Macaro, 2001; M. Turnbull, 2001). Learners may at least partially support an exclusive L2 policy (Shvidko, 2017), only tolerate a less than 10% use of L1 (Tang, 2002), and may even be reluctant to use the L1 at all (Storch & Wigglesworth, 2003). Indeed, use of the L1 for L2 learning may not be beneficial for learning outcomes, with negative correlations found between L1 use and text quality in L2 English writing (Weijen et al., 2009). Furthermore, institutional policies in both foreign and second language contexts may disfavor or even prohibit L1 use, often with severe penalties for non-compliance (Jenkins, 2010). Thus, teachers may fail to report or avoid L1 use in the L2 English classroom (Copland & Neokleous, 2011; Debreli, 2016; Sampson, 2012), and others may implement institutional policies inconsistently (Shvidko, 2017). As Sampson (2012) notes, professional practices around L1 use in the English language classroom have remained consistent, with change slow to nonexistent in recent decades. Precise descriptions of current institutional language policies are hard to obtain, perhaps because administrators are aware of their inaction in this area, and empirical evidence demonstrating the prevalence of L2-only policies is therefore often anecdotal. Burton and Ranjendram (2019), for example, describe the absence of an English-only policy in the context of a Canadian ESL study, but the pervasiveness of "English-only discourses circulat[ing] among students, administrators, and instructors" (p. 28). And Shvidko's (2017) "informal online survey" revealed that 14 out of 23 administrators acknowledged an active prohibition on L1 use, and 23 out of 28 lamented student L1 use in their Intensive English Programs (IEPs) in English-speaking countries. Thus, negative attitudes towards and practices around L1 use in L2 classrooms are alive and well.

However, from a theoretical perspective, a growing number of researchers have called for a multilingual approach, including use of the L1, in L2 teaching and learning (e.g., Atkinson, 1987; Cook, 2001; Crump, 2013; Cummins, 2007; Hall & Cook, 2012; Littlewood & Yu, 2011; Piccardo, 2013; Sampson, 2012; Turnbull & Dailey-O'Cain, 2009). Empirical research in this area has primarily focused on describing English as a foreign language (EFL) contexts, where, as discussed by Shvidko (2017), learners typically share an L1. Such research has documented varied frequencies of L1 use for diverse classroom functions, for example, explanations of grammar and vocabulary, implementation of task, maintenance of classroom discipline, and development of interpersonal relationships (e.g., Franklin, 1990; Inbar-Lourie, 2010; Littlewood & Yu, 2011; Ma, 2019; Macaro, 1997; Polio & Duff, 1994; Sampson, 2012; Yu & Lee, 2014). The frequency of L1 use in the L2 classroom is argued to be moderated by the type of learning task, learner age, learner proficiency level, extent of learner engagement, and teacher language background (Azkarai & May, 2015; Burton & Rajendram, 2019; DiCamilla & Anton, 2012; Lee & Macaro, 2013; Moore, 2013; Storch & Aldosari, 2010; B. Turnbull, 2018). Evidence of positive attitudes towards L1 use in L2 classrooms has been found among teachers (e.g., Debreli, 2016; Kim, 2015; Kim & Petraki, 2009) and learners (Debreli & Oyman, 2015; Kim & Petraki, 2009; Leeming, 2011; Neokleous, 2016; Shvidko, 2017), though learner and teacher attitudes may be moderated by learners' age (Macaro & Lee, 2013) and proficiency level (Burton & Rajendram, 2019; Debreli & Oyman, 2015).

In contrast to the vast majority of descriptive work, little experimental work has been conducted on use of the L1 in L2 English classrooms (Hall & Cook, 2012; Macaro, 2009), with recent calls for more intervention studies (Yu & Lee, 2014). Regarding effects on learning outcomes, in one study, use of a multilingual pedagogy improved learner perceptions of learning outcomes in EFL writing and listening; however, no control group was included for comparison (Adamson & Coulson, 2015). With inclusion of both an experimental and control group, Berning (2016) documented gains in EFL writing scores after a multilingual approach to teaching, although the treatment condition was limited to one 15-minute session. In a similar design, Arshad et al. (2015) found that use of the L1 facilitated the teaching of grammar for beginning level EFL students, but made no difference to those at higher levels. And in a series of studies, Macaro and colleagues have investigated the effects of a multilingual approach in the domain of EFL lexical development. In a study of L1 Chinese learners of EFL at various proficiency levels in their first year at university, Tian and Macaro (2012), for example, found that presentation of vocabulary using both L1 and L2 proved slightly more advantageous in tests of listening comprehension than presentation of vocabulary using L2 English only. Similar results were obtained by Zhao and Macaro (2016) in immediate and delayed post-tests of reading-based vocabulary knowledge (concrete and abstract words) among second year L1 Chinese learners of L2 English (non-majors). However, it was not clear to what extent these gains could be maintained in the long term, and Macaro et al. (2009) found no differences in learning outcomes from use of a multilingual versus monolingual approach.

While a little (quasi-)experimental research has been conducted in EFL contexts, where students generally share an L1 and exposure to the L2 outside the classroom may be limited, almost no research has been conducted in English as a Second Language (ESL) contexts in order to determine whether existing findings hold when students do not share an L1 and L2 exposure may be abundant outside the classroom. An earlier intervention study (Brown & Lally, 2019) preceding the current one involved one instructor and ESL learners at two proficiency levels, lower and upper-intermediate, with one class at each level designated as the control groups, experiencing a fully immersive English-only teaching and learning environment, and the others as the treatment groups, experiencing a non-immersive environment with use of both English and other languages. The quantitative analysis of assignment scores found no statistical differences in learning outcomes as reflected in assignment grades between the control and treatment conditions. A further analysis of course evaluations found possible student preferences in some areas for the nonimmersive classroom environment, though importantly course evaluations were standardized and did not include items specific to language policy and use.

# **The current study**

A lack of consensus exists on whether and to what extent the L1 should be used in L2 English language classrooms. On the one hand, in professional practice, English-only policies and practices remain frequently mandated (Copland & Neokleous, 2011; Debreli, 2016; Jenkins, 2010; McMillan & Rivers, 2011; Rivers, 2011; Sampson, 2012; Shvidko, 2017), yielding potentially serious negative consequences for teachers around classroom L1 use (Jenkins, 2010). Support for an English-only pedagogy is derived from the clear need for L2 input, output and interaction in TESOL and a lack of research-supported guidelines indicating how much L2 is necessary and sufficient for optimal acquisition (Ellis & Shintani, 2014), along with findings showing some positive student attitudes towards L2-only use and the possibility of negative effects of L1 use on L2 learning outcomes (Shvidko, 2017; Tang, 2002; Weijen et al., 2009). On the other hand, recent theoretical work argues against a monolingual, English-only pedagogy, instead favoring a multilingual approach or English plus L1/other language pedagogy (e.g., Atkinson, 1987; Cook, 2001; Crump, 2013; Cummins, 2007; Hall & Cook, 2012; Littlewood & Yu, 2011; Piccardo, 2013; Sampson, 2012; Turnbull & Dailey-O'Cain, 2009). Empirical work supporting this second position is chiefly descriptive (Hall & Cook, 2012; Macaro, 2009), showing the existence and functions of L1 use and positive attitudes towards L1 use among students and teachers (e.g., Azkarai & May, 2015; Burton & Rajendram, 2019; Debreli, 2016; Debreli & Oyman, 2015; DiCamilla & Anton, 2012; Kim, 2015; Kim & Petraki, 2009; Lee & Macaro, 2013; Leeming, 2011; Moore, 2013; Neokleous, 2016; Shvidko, 2017; Storch & Aldosari, 2010; B. Turnbull, 2018). Experimental work is severely limited overall (Yu & Lee, 2014) and especially so in ESL settings, generally with advantages or no effects found for a multilingual approach (Adamson & Coulson, 2015; Arshad et al., 2015; Berning, 2016; Macaro et al., 2009; Tian & Macaro, 2012; Zhao & Macaro, 2016). Thus, Macaro's (2014) question around the decision between a monolingual versus a multilingual pedagogy remains largely unresolved.

The current study constitutes part of a larger group of studies attempting to shed light on this important decision in second and foreign language teaching, across L2 proficiencies, and within and across students and teachers. This study expands on earlier research in several critical ways. First, we partially replicate a prior study (Brown & Lally, 2019) by focusing on new groups of ESL learners, implementing an experimental contrast between a monolingual versus multilingual pedagogy. Second, we generalize across instructors, testing whether effects of the intervention vary by teacher either in terms of learning outcomes or in teacher observations. Third, we focus on a mid-intermediate level of proficiency, where critical thinking is developed, to test whether the effects of intervention vary by proficiency level. And fourth, the inclusion of within and between-participant intervention-control design and a survey on classroom language practices facilitates direct examination of learner views towards the two approaches.

# **Method**

### *Participants*

Two researcher-practitioners and a total of 50 international undergraduate students drawn from four English for Academic Purposes (EAP) courses at a large university in the northeastern United States participated in this study. The researcher-practitioners rated the English proficiency level of the learners as B2 on the CEFR Global Scale. One researcher-practitioner taught two classes that were highly diverse linguistically and culturally with a total of 13 L1s represented, and the other researcher-practitioner taught two classes that were considerably less diverse, with a majority of students from one country. All classes contained students from backgrounds traditionally considered EFL, e.g., China, as well as backgrounds that could be considered 'post-colonial ESL' contexts, e.g., India.

The researcher-practitioners themselves were also multilingual; one (simultaneously) bilingual in English and Italian, with CEFR-C1 (North et al., 2018) proficiency in Spanish, and the other a native speaker of English with a CEFR-B2 level of proficiency in Spanish. Although the dual role of researcher-practitioner introduces the potential for research bias, inclusion of the instructors as researchers was essential to (1) establish classroom environments including or excluding the use of non-target languages based on instructor professional judgements (see Macaro, 2009), and (2) contribute qualitative observational data in the form of instructor journals reflecting on their experiences. Participant demographic information is summarized in Table 1 below.

### *Procedures for data collection*

Following definitions laid out in Ortega and Iberri-Shea (2005), this study was considered longitudinal, with data collected throughout a four-month period. All student participants were enrolled in CEFR-B2 level EAP classes focused on academic writing and critical thinking. The courses were taught using Communicative Language Teaching methodology with Focus-on-Form and writingbased activities that included considerable oral interaction in class. The study employed a between-within participant design with each class experiencing both the treatment and control conditions in a counterbalanced fashion. Thus, one course from each of the two instructors began with the treatment condition, while the other courses began with the control condition, and the conditions were switched midway through the study.

For the treatment condition, instructors drew on resources such as Celic and Seltzer (2013) and García et al. (2016) for instructional activities that facilitate multiple language use, which largely comprised those around pre-writing and project preparation


**Table 1.** Participant demographics, data, and analysis

NA = not applicable; NS = native speaker

such as discussion of topics, analysis of assigned readings, and background research. Students were encouraged, whenever possible, to communicate with peers in their language of choice (see Moore, 2013; Macaro, 2009), whether that was their native/ other languages or English. Statements requiring "English-only" were removed from course syllabi and supplementary materials. However, institutional polices required that grading be based on products submitted in English; thus, all final assignments were submitted in English (see Lee, 2016, for a discussion of the assessment of translingual writing). For the control condition, the standard "English-only" policy was implemented in all classes. This policy was present on course syllabi for classes that began with the control condition and was included in supplementary materials. The policy was enforced in the classroom through regular reminders to use English, although with no penalties for non-English language use.

The switch in conditions midway through the study was explicitly marked for students. Students transitioning from treatment to control conditions discussed their use of multiple languages up to that point and were challenged to continue in only English. In the reverse case, students were praised for their 'exclusive' use of English during class and challenged to flexibly switch languages henceforth. From that point, all procedures described above for the reverse condition were implemented across courses.

The between-within participant design with condition switch is desirable on ethical grounds as it does not withhold treatment from a control group, but it also facilitated a student survey on practices, policies, and attitudes towards classroom language use after students had experienced both pedagogical approaches. The survey was administered on paper during class in order to maximize participation, and thus was kept brief, comprising the following four questions:


• *How do you feel about using your native or other languages during your course?*

Consent for research participation was sought at two periods. Early in the study, volunteer students (71% of students (50/70) from all four classes) consented for work generated as part of the course to be secondarily analyzed for research purposes. Participation varied considerably among classes as shown in Table 1, with the highest participation in the two courses taught by one researcher-practitioner, which coincidentally were more culturally and linguistically diverse. Later, 61% of students (43/70) from all four classes volunteered to complete the survey on attitudes towards classroom language use. The survey was completed during class, with those choosing not to participate given an alternative activity in the same packet in order to ensure confidentiality. Levels of participation by class are indicated in Table 1, and none of the students who declined to participate in the first stage elected to participate in the survey. All consent interviews and data collection were conducted by the non-instructor researcher in order to minimize the possibility of coercion.

Finally, both researcher-practitioners completed reflective journals after each lesson throughout the four-month period. Journals documented the pedagogical content of each lesson, teacher instructions for activities as they related to language use, the languages observed in use by students for those activities, and any additional researcher-practitioner remarks, including their perceptions of student attitudes towards language use and any similarities or differences noted between classes in activity completion. These journals served as field notes demonstrating fidelity-to-condition in the procedure as well as enabling the inclusion of instructor perspectives on the intervention.

# **Analysis**

### *Quantitative analyses*

Quantitative analyses of five main student writing assignments representing a variety of rhetorical patterns were conducted. These comprised (1) *a response paper*, in which students had to demonstrate their comprehension of a video and their ability to think critically and reflectively on the video's presentation (2) *a source-based paper*, in which students employed and appropriately documented credible sources to support the main thesis of a paper, (3) *a genre analysis*, in which students explained findings drawn from analysis of an unfamiliar genre, (4) *an argument synthesis*, in which students developed an argument about the effectiveness of a visual public service announcement that presented a relevant social issue in a provocative manner, and (5) *a narrative paper*, in which students told a story in first or third person.

Scoring of assignments was based on demonstrated mastery of rhetorical modes as well as grammatical and lexical accuracy. The research-practitioners had freedom in how the standard course syllabus was implemented, and research-practitioner 1 (RP1) chose to formalize the drafting and revision process in the first half of the study, scoring each stage for some of the assignments. Thus, for RP1 a total of 12 data points comprised of ten discrete assignment scores, an interim course grade, and a final course grade for each student were available for analysis, where the final course grade was assembled from grades in final versions of each writing assignment, along with completion of non-scored blogs and class participation. Research-practitioner 2 (RP2) did not formally score drafts, but instead implemented individual oral presentations of the source-based and argument synthesis papers and written self-reflections around those oral presentations, which were scored. Thus, for RP2, a total of nine data points comprising five assignment scores, two presentation and self-reflection scores, an interim course grade, and a final course grade for each student were available for analysis, with the final course grade an assembly of grades in each writing assignment, grades in presentations and self-reflections, and class participation.

Given that scores generally reflected discrete, progressbased assignments, with each targeting a different rhetorical pattern, between-group statistical analyses focused on each score independently without inclusion of a repeated-measures analysis over time. As a result, no data points were excluded by the SPSS software even if an individual student missed one assignment, which maximized statistical power. Importantly, given the variation in assignment scoring by instructor, analyses were not conducted across instructors. Thus, one set of t-tests compared scores for the two classes across conditions taught by one researcher-practitioner, and a separate set of t-tests compared scores for the remaining two classes across conditions taught by the other researcher-practitioner.

### *Qualitative analyses*

Qualitative analyses were conducted on the student survey responses and the researcher-practitioner journals. Handwritten data initially collected on paper to maximize participation were typed to facilitate coding. For all data, a thematic analysis was conducted following applicable procedures in Nowell et al. (2017). Given the relatively open nature of the survey and practitioner journal prompts, inductive, data-driven coding was applied in an iterative fashion (Fereday & Muir-Cochrane, 2006). For the student surveys, a team-based, consensus-driven inductive analysis was conducted on responses to the latter two questions eliciting student attitudes towards English-only versus multilingual classroom language practices. For the researcher-practitioner journals, the author of an individual journal was not involved in the inductive thematic analysis in order to minimize bias.

# **Results**

### *Analysis of learning outcomes*

Several analyses of learner outcomes were conducted. Because of assignment variation across instructors, separate analyses were conducted for each of the instructor's two classes to determine differences between the pedagogical treatment and control groups. The study design was considered within-participant to the extent that all students experienced both treatment and control conditions and could comment on their experiences of and attitudes towards each. However, given that the assignments during their experience of each condition were different, meaningful statistical comparisons between participants could only be made. In the following, monolingual refers to the control condition, where instructor and students used English almost exclusively, whereas multilingual refers to the treatment condition, where the students were

**Figure 1.** Mean student scores for research-practitioner 1 by course and condition

frequently and regularly encouraged to use their native or other languages as well as English.

Figure 1 displays assignment scores for researcher-practitioner 1, who had the least cultural and linguistic diversity within courses. In this chart, the bars on the left for each assignment represent outcomes from one course, while the bars on the right represent outcomes from the second course. Note that since students in each class switched conditions midway through the study, the ordering of monolingual control and multilingual treatment changes after the interim score.

Overall, student scores were relatively high and at times close to ceiling levels. Throughout the study, the scores between classes were descriptively very comparable, with generally just a few percentage points between them. In terms of conditions, the students under the multilingual treatment condition descriptively outperformed those under the monolingual control condition at the majority of time points (8/12). However, twelve independent sample t-tests comparing scores under treatment and control conditions for each assignment revealed no statistically significant differences. The output of these analyses is shown in Table 2.


**Table 2.** Results of T-Test analysis by assignment for researcherpractitioner 1

\* Note that not all assignments were completed by all students. Revisions of papers, in particular, were only completed by a few students. One student began but did not complete the course

**Figure 2.** Mean student scores for research-practitioner 2 by course and condition

Figure 2 displays assignment scores for researcherpractitioner 2, who had the most cultural and linguistic diversity across courses. As above, the bars on the left for each assignment represent one course, while the bars on the right represent the second course for this instructor, and the ordering of monolingual control and multilingual treatment switches after the interim score when the students in each class switched conditions.

Again, student scores were generally close to ceiling levels. The students under the multilingual treatment condition descriptively outperformed those under the monolingual control condition in three cases, while the reverse pattern was seen in three cases, and in three cases the mean scores were exactly the same. Nine independent sample t-tests comparing scores under treatment and control conditions revealed no statistically significant differences, with output provided in Table 3.

In summary, results from the quantitative analyses of assignment scores revealed that scores were high overall, at times close to ceiling. This is to be expected given that assignments were progress tests, measuring mastery of discrete rhetorical patterns.


**Table 3.** Results of T-Test analysis by assignment for researcherpractitioner 2

\* Note that not all assignments were completed by all students

Descriptively, researcher-practitioner 1's students under the multilingual treatment condition generally outperformed those under the monolingual control condition, although results from researcher-practitioner 2's students were more variable. Crucially, these results cut across groups, with no individual class consistently surpassing another regardless of condition, suggesting a limited role for extraneous explanatory variables. Despite descriptive tendencies, no statistical differences between the conditions were found. Importantly, since repeated measures analyses were not conducted due to the differences in assignments throughout the study, we cannot say at this point whether individual students performed better under one pedagogical approach versus the other.

#### *Analysis of student attitudes*

Quantitative and qualitative analyses of student attitudes towards classroom language practices elicited from a brief survey were conducted. A total of 61% of the total class population and 86% of students who volunteered consent during stage one to have their coursework secondarily analyzed for research purposes also volunteered consent to complete the survey (see Table 1 for a breakdown among classes).

In response to Question 1 on language use in the classroom, 61% of students reported using languages other than English. These included Arabic, Hindi, Indonesian, Mandarin, Russian, Spanish, Ukrainian, a selection which overlapped with the known L1s of participants. However, some students did not report using languages that in particular researcher-practitioner 2 had witnessed in use during the highly linguistically diverse classes. These omitted languages included Korean and Portuguese. Furthermore, 39% of students reported using only English at some points in the course. Five of the respondents reporting English-only use were students of researcher-practitioner 1, representing 33% of the participating students from those courses. Four of the five were in the multilingual first condition such that they were experiencing the monolingual condition when the survey was administered, which may have affected their responses. Twelve of the respondents reporting English-only use came from the courses of researcher-practitioner 2, representing 43% of the participating students from those courses. However, they were generally distributed across conditions, suggesting little or no relationship between when the survey was administered (at the end of the study) and the condition applied (e.g., monolingual English-only) at that time. The difference between responses in this area from the students of researcher-practitioner 1 versus 2, albeit relatively small, might be explained by the fact that researcher-practitioner 1 saw lower participation in the study overall and thus those that did participate might have been more highly engaged with more accurate self-perceptions of language use. Overall, the observations from the researcher-practitioner journals, which documented use of languages other than English including among those individuals who reported using only English, revealed that some respondents were not entirely accurate in their self-perceptions and reporting of classroom language use, which is a weakness of self-reported data in general.

In response to Question 2 on language policy, 65% of survey respondents accurately recognized that an "English-only" policy was in effect at some point in their courses. These responses were drawn from all courses of both instructors. 28% reported that such a policy was not in effect and 7% declined to answer the question. Respondents failing to recognize the existence of an "English-only" policy at some point were distributed across multilingual and monolingual first conditions, suggesting that the presence of that policy on syllabi at the beginning of the course did not impact some. Two of the respondents failing to recall an English-only policy were students of researcher-practitioner 1, representing 13% of the participating students from those courses. Ten of the respondents recalling no English-only policy were students of researcher-practitioner 2, representing 36% of the participating students from those courses. The disparity between instructors here is larger than that in responses to questions of language use, and it is not clear what underlies the difference. It is conceivable that researcher-practitioner 1 was more explicit about the distinction between treatment and control conditions as they related to classroom policies than researcher-practitioner 2, rendering heightened student awareness of the language policy in those classes.

The remaining two survey questions were analyzed through a team-based, inductive, data-driven thematic analysis applied iteratively to the data. In response to Question 3 eliciting attitudes towards an "English-only" policy, a total of six themes were identified. These concerned the extent to which the policy was considered expected, helpful especially for facilitating conversation or thinking in the target language, necessary, not necessary or even unhelpful. These themes are illustrated below with quotes from respondents.

	- *– I do not oppose it, as this class is an English class. It makes sense to have an English only policy.*
	- *– I like English only policy, it helps me on learning better.*
	- *– It is difficult sometimes but good for us.*
	- *– This policy requires us to speak English at any time so it is an effective way to improve speaking for international students.*
	- *– It is good as it helps people of different countries to converse.*
	- *– It helps understand each other better.*
	- *– It's helpful for students to cultivate an English-based mindset meaning for students to think in English.*
	- *– I think it helps me change my channel to an English mind. I do feel that I behave differently when I am using English mind.*
	- *– It's a way to make us think about issues in English, which will help us to organize words effectively.*
	- *– Yes. It makes me think that how it has to be.*
	- *– Depends on the rigor of the [English] course. Lower [English] course could allow other languages to help expressing. Since [this class] is the highest [English] class for first year international students, English only should be in effect.*
	- *– It is good but not necessary.*
	- *– Not really helpful because sometimes we can only understand things in our native language.*

The most commonly expressed views were that an English-only policy would help English language development, followed by an equal number expressing the necessity of an English-only policy or expressing the lack of necessity and possible hindrance of an Englishonly policy. Varied other views were expressed by a minority.

In response to Question 4 eliciting attitudes towards the use of languages other than English in the English language classroom, a total of five themes were identified. These overlapped in part with those above, namely the extent to which use of languages other than the L1 was helpful for L2 development, made learners feel comfortable, was beneficial for the expression and development of ideas, was not necessary and was detrimental if the L1 of one student was not shared by other students in the class. These themes are illustrated below with quotes from respondents.

	- *– It's still useful to construct meaningful conversations that help to improve my English writing.*
	- *– If I use my native languages, I will be relax.*
	- *– It would be nice to talk in my native language as it can reduce my homesickness.*
	- *– It is a convenient way to discuss some of the tough questions.*
	- *– Don't like it. It confuses me.*
	- *– I still prefer that these parts could be completed in English.*
	- *– To learn more in English I need to talk more in English.*
	- *– When in class, students should be discouraged to speak in their own language because that divides groups into whatever language they speak. Instead, we should all be brought together during class time and encouraged to confidently interact with one another in English.*
	- *– It make me feel uncomfortable because it's* [an English class].
	- *– It could help explain some terms more easily, however, it did not impact majorly in learning for me since there is only one other* [specific language] *speaker in the class.*
	- *– Maybe unfair to some guys.*
	- *– It is a good idea if multiple people in class speak that language.*
	- *– I never used my native language in the class, as no one would understand.*

Student views on the fourth question were that the use of languages other than English can be beneficial for the expression and development of ideas, but also that such language practices were not a necessity or were even a hindrance for English language development. However, collapsing responses in the first three categories – generally helpful, promotes comfort, beneficial for expression – indicated that the majority of views expressed by participants on the issue of use of languages other than the target language were positive.

### *Analysis of research-practitioner journals*

Both researcher-practitioners kept reflective journals for the duration of the study (four months). The journals documented the pedagogical content of each lesson. Activities employed for the multilingual treatment condition included regular (daily or weekly) small-group and paired discussions as well as individual work, where students gathered and discussed information about various topics in languages other than English and subsequently produced associated writing assignments in English. The same activities were employed for the monolingual control condition, but students were asked to use English exclusively for pre-writing activities and discussion.

Given their involvement in the research, researcher-practitioner journals also focused on areas relevant to the study such as teacher instructions as they related to language use, the languages observed in use by students, and any additional remarks including instructor perceptions of student attitudes towards language use and similarities or differences noted between conditions in activity completion. The journals were examined using an inductive, data-driven, thematic analysis that was applied iteratively to the data, and the author of a given journal was not involved in the initial analysis, though was involved post hoc in the form of member checking (see e.g., Nowell et al, 2017).

#### *Research-practitioner 1*

The journal of researcher-practitioner 1 was relatively long, at 8,785 words. Excluding commentary simply recounting pedagogical content, eleven categories were initially identified in the remainder of the text. These were eventually reduced to the following five

main themes, confirmed by member checking, accompanied by representative illustrative quotes marked by the day of the observation. Note that the switch between conditions was made at Day 18.

## (1) Encouragement and amount of L1 use

The L1 was specifically encouraged by the instructor in the multilingual condition, especially early in the condition.


The introduction of the multilingual condition generated some visible reactions:


As the study progressed, no explicit encouragement was used and students in the multilingual condition still recruited multiple languages, i.e., by weeks three and four:

*– They were not told they could use their L1, but many did. (Mu1st Day 12)*

However, there were also observations about the L1 not being used, particularly noted when the conditions were switched in a class:

*– When they spoke in their L1, they were whispering. It was as if they were nervous they were doing something*  *wrong. I told them it was okay to speak loudly in their L1. (Mu2nd Day 18)*

*– Most of the pairs were not speaking in their L1 at the start; I had to remind them that they could. (Mu2nd Day 26)*

In terms of amount of L1 use, this was not recorded on a daily basis. There were a few reports of L1 use by a minority especially at the beginning of the multilingual condition:


However, there were more reports of L1 use by a majority:


(2) Mandate for and amount of L2 English use

The L2 was at times encouraged by the instructor in the monolingual condition but more often mandated as indicated by verbs such as *asked, stressed strongly, forced*. A considerable portion of the journal was devoted to comments such as the following:


In terms of how much L2 was spoken, conspicuously few journal comments documented abundant language production during the English-only policy:


Indeed, in a number of cases, the mandate to use the L2 was observed to result in markedly little oral language production:


As indicated in the final quote above, in contrast to the stated policy, use of the L1 was noted in the monolingual, Englishonly condition, with the instruction to switch to the L2:


While some L2 use was always observed, some use of L1 continued well into the monolingual condition:


(3) Presence of codeswitching / active translanguaging Language switching, labeled by the research-practitioner as codeswitching or active translanguaging, was witnessed in the multilingual condition:


A couple of these observations were around the influence of technology in this process:


### (4) Assessment of students

A number of observations were made regarding student performance. Some assessments were made with distinctions more apparent by class than by condition. Thus, the group that experienced the multilingual condition first were observed to be a strong group, with little mention of difficulties even after the condition switch to monolingual:


In contrast, the group that experienced the monolingual condition first were observed to be a weaker group, including after the condition switch to multilingual:

*– Answers were also limited. (Mo1st Day 1)*


Though some of the difficulty appeared specifically to be related to the monolingual condition:


And there were some positive general assessments of performance of the weaker group, some of which related to comfort level after condition switch.


One area where assessment was by condition to some extent was in observations of speed of activity completion. The following comments all described the same class of students, but with perceived differences in speed of performance under the different conditions:


(5) Pedagogical commentary

Pedagogical commentary described activities that characterized each condition: mono- or multilingual:


Also included were specific instances where the researcherpractitioner did not or could not implement a multilingual pedagogy


And finally, there were comments about potential off-task behaviors exhibited by the students, which could not be confirmed by the research-practitioner. These were entirely related to use of languages other than English, for the most part during the multilingual condition for both classes, but in one case during the monolingual condition when L1 was used:


Research-practitioner 2. The journal of researcher-practitioner 2 was shorter, at 3,702 words. Excluding commentary simply recounting pedagogical content, twelve categories were initially identified in the remainder of the text. These were reduced to the following four main themes, confirmed by member checking, accompanied by representative illustrative quotes marked by the day of the observation. Note that the switch between conditions for this instructor was also made on Day 18.

(1) Encouragement, purpose, and amount of L1 use The L1 was specifically encouraged by the instructor in the multilingual condition:

*– L1 use was encouraged for those who could use it. (Mu1st Day 11)*

*– They were encouraged to use L1 in their groups and choose from movies in their L1. (Mu1st Day 16)*

The researcher-practitioner noted significant L1 use across groups during the multilingual condition for a variety of purposes:


Evaluative comments were also provided on the nature of the L1 interaction across groups during the multilingual condition:

*– This grouping produced, for the 1st time, very clear and loud use of students' L1s (Mu1st Day 4)*

Like researcher-practitioner 1, researcher-practitioner 2 also noted cases where the L1 was not being used or used minimally, which was observed throughout the multilingual condition:


In one case, upon investigation, this was because of a difference in L1 dialects:

*– I didn't hear any Arabic. .. I … learned that their* 

*dialects were very different. As a result, they preferred to communicate in English. (Mu2nd Day 30)*

In other cases, a lack of L1 use was due to L1 'singletons", i.e., those that did not share an L1 with anyone else in the class (see Table 1):

*.. the singleton pairs used English exclusively. (Mu1st Day 7)*

(2) Mandate for and amount of L2 use

Researcher-practitioner 2 also encouraged English-only use in the monolingual condition, even until late in the condition, though was perhaps less forceful than researcher-practitioner 1, at least as indicated by the verbs employed.


Researcher-practitioner 2 appeared to see some success with the English-only policy:


Success with the English-only policy appeared to some extent to be related to groupings:

*– Pairings included some Mandarin L1s together, so there was a need to remind them to use English L2 at times. (Mo2nd Day 30)*

(3) Assessment of students

A number of observations were made regarding student performance, to some extent on the basis of condition. Observations of good performance were made for the multilingual condition, and in one case in contrast to the monolingual condition, though these were generally about one specific group (i.e., that experiencing the multilingual condition first):


The superior performance of the group noted above was supported in the above:

*– I am able to get through work and can count on this group being prepared more so than the first group. (Mo2nd Day 20)* 

Overall, fewer observations of good performance were made of the monolingual condition across groups. Included here is one positive comment about students in the multilingual condition, who were unable to use their L1 as they were 'singletons':


And a couple of observations were made about comparable performance across conditions:

*– An interesting development was that 2 pairs were very quick compared to the others. Interestingly, one was an L1 pair, 1 Mandarin & 1 Mandarin/Japanese, and* 

*the other pair was a pair of singletons, Hungarian & Hindi. Mu1st Day 8*

*– Action & accuracy mirrored monolingual class (Mu1st Day 14)*

In one collection of observations, researcher-practitioner 2 noted poorer performance specifically in one group, but also in the monolingual condition:


In contrast, there was only one negative comment about the other group during their simultaneous multilingual condition:

*– Two of the* [language] *groups were rather slow and barely finished on time. (Mu1st Day 6)* 

(4) Pedagogical commentary

Much of the pedagogical commentary from researcherpractitioner 2 revolved around the efforts needed to make groupings of students that were appropriate for the condition. For the multilingual condition, the following was noted:


*– Groups of 3 & 4, so not of L1 use. Later, I mixed up groups for comparison. (Mu2nd Day 26)*

For the monolingual condition, the following was noted with comparable strategies used across classes and time and special care taken for the number of Mandarin speakers who were in the majority in both classes despite the linguistic diversity of RP 2's classes:


Finally, the researcher-practitioner noted a logistical challenge, tardiness, that affected his ability to implement effective groups, which appeared to emerge in the class under the multilingual condition towards the end of the study:

*– Lateness in class is creating a problem for in-class translanguaging activities. I'm unable to form the groups I'd like to due to late arrivals. Need to make adjustments. (Mu2nd Day 20)*

# **Discussion**

In this study, two US-based university-level ESL courses at an intermediate CEFR-B2 level were instructed by two researcherpractitioners with the standard monolingual, English-only pedagogy. Two comparable courses were taught by the same instructors using a multilingual approach, where students were given agency in their choice of language – L1 or otherwise – for in-class discussion, project preparation, and pre-writing activities. Quantitative analyses focused on learning outcomes, specifically the development of writing (paragraph and genre) and presentational speaking as well as student perceptions of classroom language practices. Qualitative analyses were conducted of responses to a survey eliciting student attitudes towards classroom language use and instructor views in the form of a researcherpractitioner's reflective journal.

Quantitative results showed generally high scores for all assignments, which was not unexpected given their nature, i.e., progress tests based on mastery of discrete aspects taught in prior lessons, e.g., rhetorical patterns. Quantitative analyses showed a few descriptive differences, to some extent favoring the multilingual treatment, but no statistically significant differences between conditions throughout the study. This is striking for a number of reasons. First, with conditions switched throughout the study on groups of different sizes, findings suggest a lack of effect by student group regardless of size. In addition, the study was conducted across research-practitioners, indicating a lack of effect by instructor. Finally, an additional unplanned variable emerged in this study of intact classes – that of cultural and linguistic diversity. Coincidentally, the classes of researcher-practitioner 1 were relatively homogenous, dominated by Chinese speakers, and similar to the classes examined in Brown and Lally (2019). However, both classes of researcher-practitioner 2 were highly diverse linguistically and culturally, which was unusual and offered the opportunity to examine the effects of pedagogical treatment by level of diversity in student demographics. Again, no statistical differences in learning outcomes were observed. Thus, like Brown and Lally (2019), this study failed to find statistically robust differences in learning outcomes associated with the form and environment of instruction as it related to student language use (target versus L1/other) in the classroom at an intermediate proficiency level. Moreover, comparable results across diverse classroom contexts (teacher, student demographics) diminished the possibility of non-relevant (to this study) differences between groups as explanatory variables.

We recognize that there was some student L1 use in the monolingual condition that was documented in both researcherpractitioner reflective journals. This may demonstrate the impracticality and even impossibility of asking students to 'switch off' their L1, especially in light of psycholinguistic evidence demonstrating the inevitability of simultaneous activation of all known languages in the mind (e.g., Runnqvist et al., 2012). However, the journals also documented strong efforts by the instructors

clearly to differentiate the conditions by maintaining an Englishonly policy during the monolingual condition and promoting student use of different languages in the multilingual condition; therefore, we have evidence of fidelity-to-condition at least as it related to instructor behavior. Overall, the findings presented here are in line with studies summarized in Macaro et al. (2009), and to some extent Zhao and Macaro (2016) and Tian and Macaro (2012), who found at least no negative outcomes and even some slight advantages of a multilingual pedagogical approach.

In terms of student perceptions of language use, notable in the current study was that some students did not accurately report their language usage and were not cognizant of the conditions employed in their classes, despite explicit instructor instructions during both conditions documented in both researcher-practitioner journals. More than one third of students reported using only English throughout, which was accurate for the few singletons in the study, but inaccurate for most of the respondents, who were specifically observed by the researcher-practitioners using languages other than English. In addition, only two thirds reported the existence of an English-only policy, which was in effect at some point for all courses. It is difficult to say what underlies these discrepancies, but their existence does have methodological implications, reflecting the weaknesses of self-report data and the need for a triangulation of data, such as the instructor journals.

Regarding student preferences for language use, some expressed positive attitudes towards an English-only policy and against multilingual classroom language practices, especially at their intermediate level of proficiency (cf. Burton & Rajendram, 2019; Debreli & Oyman, 2015; Shvidko, 2017; Tang, 2002), and raised some potentially serious issues of student equity and polarization of a multilingual policy. However, the majority expressed generally positive attitudes towards multilingual classroom language practices, emphasizing benefits such as the general helpfulness of the L1, promotion of student comfort, and assistance with development of ideas and expression. Such positive views are in line with a number of other studies (e.g., Debreli & Oyman, 2015; Kim & Petraki, 2009; Neokleous, 2016; Shvidko, 2017), especially in claims that use of the L1 can lessen cognitive load for learners (Macaro, 2005).

A significant amount of thematic overlap was observed in analyses of the researcher-practitioner journals, which signals the trustworthiness (Nowell et al., 2017) of the qualitative side of this mixed methods study. Both noted their efforts to encourage L1 use in the multilingual condition and to mandate L2 use in the monolingual condition, which as above, demonstrated fidelity-tocondition, as well as commenting on the amount of L1 and L2 use. They also provided some evaluative comments on student performance, which in some cases favored the multilingual condition. These comments revolved around the richness of discourse, accuracy and potentially speed of activity completion (see also Adamson & Coulson, 2015, for efficiency of task completion), though observations were somewhat mixed by language use, e.g., singletons using only English in the multilingual condition were also observed to perform quickly. The lack of clear distinctions in evaluations of performance by group or by condition observed by both research-practitioners, however, is supported by the lack of statistical differences found in the quantitative analysis of learning outcomes.

Although both researcher practitioners commented on various aspects of pedagogy, their focus was slightly different. Researcher-practitioner 1 described the types of activities that generated monolingual versus multilingual interactions and documented potential off-task behaviors. Researcher-practitioner 2 commented on the efforts taken to manage groupings to facilitate monolingual versus multilingual interactions (see Burton & Rajendram, 2019, for discussion of the impact on teacher attitudes of 'singletons' in the classroom) and to the logistical challenges, i.e., student tardiness, that complicated those efforts, which might have arisen due to the linguistic diversity of those groups.

Some limitations of this study offer additional opportunities for further research. Importantly, since repeated measures analyses could not be conducted, we cannot say whether individual students performed better under one pedagogical approach versus the other or the effect that cultural and linguistic background has on this issue. Further, in line with CLT, the implementation of the control and treatment conditions was primarily through oral in-class tasks, while learning outcomes were operationalized primarily through written products. Thus, an open question regards whether the impact of condition would be more visible in oral products. All above areas warrant further investigation. In addition, both researcher-practitioners commented on the amount of L1 use, which, as noted above, may have compromised to some extent the monolingual control condition. Penalties for L1 use are employed in some contexts, but this would not have been in line with the pedagogical philosophy of either of the researcher-practitioners involved here and would likely not have changed the outcome given comments above on the impossibility of deactivating the L1. Instead, a detailed description of student language use based on recordings and associated transcripts could render language use a continuous rather than a nominal variable, facilitating correlations of amount of L1/L2 use by learning outcomes. Finally, the control and treatment conditions in this study revolved around student language use; there were no descriptions of non-English use by the teacher (see Burton & Rajendram, 2019, for a discussion of teacher translanguaging), and further research could examine this issue as it relates to learning outcomes especially in an ESL context with heterogenous learner L1s.

The multilingual pedagogy employed in this study is compatible with the theoretical framework in SLA of 'multi-competence' (Cook, 1992; Cook & Wei, 2016), which considers all the known languages within an individual mind as one system, as well as with the paradigm of 'translanguaging' (e.g., Canagarajah, 2011; Creese & Blackledge, 2010; García, 2009; García & Wei, 2014; Hornberger & Link, 2012; Lewis et al., 2012a, b), alternatively labeled as 'translingualism' (Canagarajah, 2013; Horner et al. 2011), 'codeswitching', 'dynamic bilingualism' 'fluid languaging practices', and 'plurilingualism' (e.g., García & Kano, 2014; Piccardo, 2013). The study has implications for language teaching and teacher training and is in line with several decades of theoretical and descriptive work on L1 use in the L2 classroom and conclusions that an English-only mandate denies English "learners the opportunity of using an important tool" (Storch & Aldosari, 2010, p. 372).

Despite the burgeoning research literature, as Sampson (2012) has discussed, change in professional practices around L1 use in the English language classroom has been almost non-existent. In some professional contexts where institutional policies discourage or prohibit L1 use, particularly where English is taught as a foreign language, a multilingual pedagogy could be catastrophic for teachers, resulting in dismissal. In other professional contexts that have witnessed a growth in translingual practices, especially where English is taught as a second language, anecdotal and published sources still document the widespread perpetuation of English-only policies (García, 2009, Shvidko, 2017). Indeed, as seen here, at least some students still appear to prefer an English-only environment (see Storch & Wigglesworth, 2003, versus Berning, 2016, for differences in the use of L1 by learners and Shvidko, 2017, for multifarious student views).

To conclude, while understanding the potential consequences of a multilingual pedagogy in reductions of the amount of L2 input and interaction (Macaro, 2014), the results of this study support those from the larger group: four ESL classrooms examined in Brown and Lally (2019) and two French and two Arabic as a Foreign Language classrooms examined in Brown (2021). All together, these studies report either no difference or a facilitative effect of a multilingual pedagogy across 12 classrooms, in second and foreign languages, and across different learners, proficiencies, and instructors. Such parallel findings suggest at the very least no cost of a multilingual pedagogy in terms of learning outcomes as measured by assignment scores, a possible gain in terms of student attitudes, and a possible gain in terms of instructor perceptions, all findings which cast serious doubt on arguments for immersive L2 pedagogical practices. Further research is needed to determine whether and how language use in language classrooms may be adapted to optimize benefits, and all stakeholders in the learning process should feel reassured that such future research will likely not negatively impact learning.

# **References**


# **4. Very young Swedish children's exposure to English outside of school**

*Liss Kerstin Sylvén*

# **Introduction**

The spread of English as a lingua franca is rapid. In some countries, the use of English among many individuals in the younger generations is almost at a par with the use of the respective first language. Sweden, where the present study took place, is a case in point. For decades, English has been naturally occurring in Swedish society, not least through original soundtracks of TV-shows and movies originating in English-speaking countries with subtitles in Swedish rather than the soundtrack being dubbed. English words and phrases are very commonly used in advertisements of anything from perfumes to trucks. In Swedish school, English is the first foreign language (FL) to be introduced, normally in 1st–3rd grade, but often even earlier than that, already in kindergarten or preschool class. In addition, English is the only mandatory FL that has to be studied in the Swedish school system, and, at the same time, one of three core subjects (the other two being Swedish and mathematics). Other FLs are studied by approximately 87 percent of Swedish students from 6th grade and onward, but they are a voluntary choice (Skolverket, 2019a). The use of English as the medium of instruction in non-language subjects, such as history and biology, is increasingly popular. Approximately 20 percent of all upper secondary schools offer English medium instruction programs (Paulsrud, 2019). It is also found at lower levels

**How to cite this book chapter:**

Sylvén, L. K. (2022). Very young Swedish children's exposure to English outside of school. In: Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (eds.) *Exploring Language Education: Global and local perspectives*, pp. 121–155. Stockholm University Press. DOI: https://doi.org/10.16993 /bbz.e. License: CC BY 4.0.

(Berggren, 2019). In mandatory school (1st–9th grade), however, there is a limitation to the use of English as the medium of instruction set at 50 percent (Sveriges Riksdag, 2009). At tertiary level, English is omnipresent. Large amounts of literature in English are encountered at all levels, and at many universities, from master's level and upwards, English is used more or less exclusively as the language of communication in both writing and speaking (e.g., Salö, 2016). At the doctoral level, a vast majority of all theses are written in English, with an accompanying extensive summary in Swedish. Furthermore, English is commonly used as the lingua franca at companies with a global presence, such as Volvo Trucks, ABB, and Ericsson.

On an individual level, access to the Internet means that exposure to English is only a click away also for very young individuals. For instance, more than two thirds of all Swedish 2-year-olds spend time on the Internet a couple of times, or more often, per week. Further, 28 percent of 2–4 year-olds watch clips on sites such as YouTube every day, four out of ten of the four-year-olds and close to 70 percent of the eight-year-olds reported using the Internet daily (Medierådet, 2019). This is a rapid increase of both access to and use of the Internet; in 2010, the reported use of the Internet among children at the age of four and eight was 2 and 4 percent respectively.

The vast presence and use of English have sparked a debate as to whether English should rather be viewed as a second language (L2) than a FL in Sweden (Andersson, 2016; Hyltenstam, 2004; Josephson, 2004; Norén, 2006). Presently, it is still officially viewed as a FL, but many would argue that on an individual level, English can definitely be considered to be an L2 (Sundqvist & Sylvén, 2016). Even though there is no clear cut definition to distinguish an L2 from a FL, the general view seems to be that an L2 plays an important role in a country, for instance in education or government, without it necessarily being the first language of those who use it (Crystal, 2003; Fasold & Connor-Linton, 2006; Richards & Schmidt, 2002). Ringbom (1980, p. 2) offers the following distinction between acquiring an L2 and learning a FL:

In a second-language acquisition situation, the language is spoken in the immediate environment of the learner, who has good opportunities to use the language by participating in natural communication situations. In a foreign-language learning situation, the language is not spoken in the learner's immediate environment, although mass media may provide opportunities for practicing receptive skills. The learner has little or no opportunity to use the language in natural communication situations.

It appears safe to say that based on this, English is quickly moving from having been a pure FL to presently being very much of an L2 in many contexts, among them the Swedish one.

One day, I received a phone call from a woman who had read a newspaper column about the astonishing level of English proficiency in the writer's own 4 year-old son and many of his friends. The woman who called told me how this coincided perfectly with her own experience of her two granddaughters, five and seven years old at the time, and how she, a former English teacher, was at a total loss when trying to understand when, where, and how her grandchildren had come to know such a great amount of English. And with a perfectly idiomatic pronunciation on top of it all. The aim of her phone call was to inquire whether there is anybody investigating this from a scientific perspective.

That was the starting point for the study in focus in the present chapter. The study is part of a larger research project, *A Study of Young Learners of English*, STYLE, and is to be seen as a pilot study, as this is the first time exposure to, and proficiency in English among such young children, are investigated. The aim, therefore, is twofold: Firstly, to explore the amount of exposure to and proficiency in English among very young learners (VYLs), secondly, to consider various instruments to be used for assessing young children's knowledge of English. The study is guided by the following research questions:


RQ1 will be answered by the administration of language diaries and interviews among 3 to 7 year-olds. RQ2 will be answered by administering two types of test; first, the Peabody Picture Vocabulary Test, PPVT-4 (Dunn & Dunn, 2007), and second, a picture description task to these VYLs. In a larger perspective, this chapter will hopefully provide further fuel to the debate on the role of English in Sweden.

# **Theoretical perspectives and literature review**

Even though the field of second language acquisition (SLA) is wide and varied (Ellis, 2003), the need for input, output, and interaction is indisputable for L2 learning to take place (Gass & Mackey, 2006). L2 input was early seen as the (only) way to learn an L2. Later, the role of output on the part of the L2 learner was highlighted by, among others, Swain (1985). In connection with output being focused on, the significance of interaction was also highlighted as an integral part of the L2 learning process. As pointed out by Holmes and Myles (2019, p. 10), for learning to happen among young language learners "rich and plentiful input, as well as opportunities to use the language meaningfully, are necessary". Engagement in extramural English, EE, (Sundqvist & Sylvén, 2016; Sylvén, 2006) does not necessarily entail the presence of all three components of SLA theory. However, the input component is always there, most often in the form of oral L2, as in YouTube videos, and song lyrics. Written input is also very common, for instance in the form of various commands in digital games. In some forms of EE, output in the form of speaking and writing is expected of participants, not least in digital games involving many players at the same time, so called Massively Multiplayer Online Roleplaying Games (MMORPGs).

Language learning happens in a social context, thus, sociocultural theory (Lantolf, 2001; Vygotskij, 1939/1978) is a relevant theoretical lens through which to try to understand what happens to these young learners when they engage in EE activities. Traditionally, sociocultural theory talks about learning from more able peers, for instance an apprentice learning practical skills from an expert through, among other things, the use of artifacts. The concept of the Zone of Proximal Development, ZDP, is central in sociocultural theory, illustrating how novice learners, through the help of more able peers, can be pushed forward in their learning process. In the present context, there are few physical peers present. Rather, the various sources utilized during EE activities, such as YouTube, digital games, etc., seem to fulfill the role of a peer. By watching, listening, and often repeating words, phrases, and whole sentences, children become accustomed to and learn L2 English.

The importance of motivation, investment and identity for the L2 learning process in the present context cannot be underestimated (Dörnyei, 2005; Norton, 2013). As these factors have been investigated mainly among older learners than those in focus in the present study, an attempt is being made at adapting some of the underlying theories to the realities of these young individuals. Norton (2013, p. 50), for instance, argues that the notion of investment "conceives of the language learner as having a complex social history and multiple desires". While the VYLs in the present study do not necessarily have complex social histories, some of them show evidence of desires in connection with the use of L2 English. It can thus be argued that their investment in English is "an investment in [their] own identity" as L2 English users, presently and in the future.

In what follows, some of the extant literature on L2 learning extramurally and among young individuals will be accounted for. However, as there exist very few studies on EE among such young learners, studies on effects of EE among slightly older learners will also be presented.

For some time now, it has been widely acknowledged that the learning of L2 English for many individuals happens to a large degree outside of the educational sphere. Already in the 1980's, Bialystok (1981, p. 24) argued that in language learning '[t]he most functional situation would likely occur outside the classroom, in a natural setting, where conveying the message is the only essential goal of the language occasion'. Her term for use of an L2 outside the classroom is *functional practice*. Ever since, there has been an abundance of studies looking into effects on L2/FL learning of various types of EE. For instance, watching TV and movies with original soundtracks and subtitles vs. dubbed soundtracks, has attracted some scholarly interest. In their studies on Dutch adolescents learning L2 French, d'Ydewalle and Pavakanun found positive effects on vocabulary acquisition when the soundtrack was in the L2 and subtitles in the L1 (d'Ydewalle & Pavakanun, 1995; d'Ydewalle & Pavakanun, 1997; Pavakanun & d'Ydewalle, 1992). No effects were found on syntax and grammar. In the same vein, d'Ydewalle and van de Poel (1999) investigated 327 Dutch children aged between 8 and 12 who were studying Danish and French as L2s. They concluded that vocabulary gains occurred when the soundtrack was in the L2, and in particular when the L1 and L2 were fairly similar, which, in this case, was the fact for Danish and Dutch. Taking another perspective on TV as a possible source of L2 learning, Rodgers and Webb investigated effects of watching a TV-series on learners' vocabulary proficiency (Rodgers & Webb, 2011; Webb, 2007). The design with using a number of episodes from the same series entailed repetition of a large number of vocabulary items, and it was found that this indeed resulted in vocabulary gains among the informants.

Going into more detail regarding L2 input via TV, Lin (2014) concluded in a study of the occurrence of formulaic sequences in TV-genres such as drama and comedy, that it proportionally resembles such sequences found in everyday speech. Thus, watching such programs may help improve learners' proficiency in formulaic sequences that are typical of everyday oral interaction.

With the advent of the Internet, studies have looked into affordances offered online and their possible beneficial impact on L2/ FL learning. The early "hole-in-the-wall" experiment carried out by Mitra et al. (2005) provided access to a computer to young people in urban slum and rural areas in India. Results showed how the availability of a connected computer gave rise to learner autonomy (Holec, 1981) among the young individuals in the village, and how they were able to learn, not only a great deal of L2 English, but also social and academic skills.

Focusing on effects of using L2 English as the medium of instruction in non-language subjects (so called content and language integrated learning, CLIL), the unexpected result emerged that exposure to English outside of school may have been more beneficial to students' levels of receptive vocabulary than what was offered by CLIL per se (Sylvén, 2004). Also in later studies on CLIL, such exposure has been shown to be a major source for L2 progress, in particular as regards vocabulary and some aspect of writing proficiency (Sylvén, 2019). The most common types of activities conducted – fully or partly – in English were computer gaming (especially for the boys), listening to music, and browsing the Internet.

Gaming has always been an attractive activity for young individuals. With the development of digital technology, games in other languages than one own's L1 are omnipresent. The use of English as the lingua franca in online games is in many cases the default choice. Gee (2007) set up a total of 36 general learning principles in regard to what video games have to do with learning and literacy. The majority of these principles concern L2 learning (see Sylvén & Sundqvist, 2012b for a rich account of these principles and how they relate to L2 learning). Thus, the use of digital games as an effective source for language learning is highly recommended. Numerous studies confirm the L2 learning benefits of such games (for instance, Brevik, 2016; Chotipaktanasook & Reinders, 2018; Li, Peterson, & Wang, 2021; Peterson, 2012; Piirainen-Marsh & Tainio, 2009; Ranalli, 2008; Reinders, 2012; Sundqvist, 2009; Sundqvist & Wikström, 2015; Sykes, Reinhardt, Liskin-Gasparro, & Lacorte, 2012; Sylvén & Sundqvist, 2012a; Thorne, Black, & Sykes, 2009; Turgut & Irgin, 2009; Zheng, Bischoff, & Gilliland, 2015). While most studies concern vocabulary acquisition, there are also those who focus on interaction, and yet others target oral proficiency.

While the majority of studies indicating clear benefits of extramural exposure for L2 learning have focused on adolescents and young adults, few studies exist on younger learners, and possible L2 learning effects among this age-group while engaging in extramural L2 activities. Among the existing ones, several are set in the Nordic context, perhaps due to the fact that exposure to English in general is high in this area and that Nordic languages are linguistically close to English (Lindgren & Muñoz, 2013). Lefever (2010) showed how 8 year-old Icelandic children understand and can converse in basic English before formal instruction starts in school. The basic source for their learning of English was said to be various media and digital gaming. Set in Denmark, Hannibal-Jensen (2017, 2018) investigated types and amount of extramural exposure and possible correlations with L2 English proficiency among 7–11 year-olds. The general conclusion of the findings was that the Internet offers an abundance of attractive affordances for L2 English learning for young individuals, and that a large number of these individuals indeed take advantage of them. In a number of publications from the Swedish context, benefits of exposure to English outside of school in general and digital gaming in particular have been shown among 4th to 6th graders (Sundqvist & Sylvén, 2012, 2014; Sylvén & Sundqvist, 2012a). The longitudinal study by Lindgren and Enever (2017) was also set in Sweden. Three young language learners were followed during their six years in primary school. The study started when the children were in grade 1, and empirical data consisted of interviews, questionnaires, oral and written production tasks, as well as reading and listening tasks. One of the main findings was that the variation in language competency between the three informants to a large extent could be explained by the amount of extramural exposure to English. In another context, namely Flanders, De Wilde and Eyckmans (2017) focused on L2 English proficiency among 11 year-olds. This is particularly interesting as formal English instruction in Flanders starts relatively late in comparison to other European countries (Enever, 2011). The findings showed that a majority of the participating children were able to perform at the A2 level (Council of Europe, 2001), despite not having had any previous formal instruction. Here too, the main sources of L2 English learning were gaming and computer use.

Thus, given the seemingly positive influence of activities carried out in English on L2 English acquisition among both adolescents and younger learners, the present study aims at investigating even younger individuals, so called very young learners (VYLs).

### **The study**

#### *Participants*

The participants of interest in this chapter are children in two educational contexts: one in kindergarten and one in first grade


**Table 1.** Participants

of compulsory school. Both the kindergarten and the primary school are located in suburban areas, with fairly stable socioeconomic conditions. For instance, the level of education among parents is in line with that reported for Sweden in general and the percentage of recently immigrated students is three, compared to six in Sweden as a whole (https://www.skolverket.se/skolut veckling/statistik). The kindergarten group consisted of 15 individuals, of whom seven were included in the study after consent had been obtained from their parents. In the first-grade group, 18 individuals make up the entire group, and consent to participate was obtained for 13 of them. In Table 1, the participants are accounted for.

### *Material*

The empirical data consist of group discussions with the two groups of children, administration of the Peabody Picture Vocabulary Test 4, PPVT-4, (Dunn & Dunn, 2007) among all participants, language diaries from several of the participants, and a picture description task performed by 13 individuals in the group of first graders.

The PPVT-4 was originally designed for testing receptive vocabulary proficiency among L1 users of English (Dunn & Dunn, 2007). However, as vocabulary tests targeting VYLs are few, it has been used in a number of L2 contexts (Hannibal Jensen, 2017). Given its original purpose, it has its flaws when used with another aim, but there are a number of advantages with it, too. The main advantages are that it is easily administered, and that the children really seem to enjoy doing the test. In their exploratory investigation of the usefulness of the PPTV-4 for L2 English learners, Goriot et al. (2018) concluded that the test might not be trustworthy for learners completing no more than the first two sets of items. In the present study, only three of the participants did not go pass those two sets (two in kindergarten and one in first grade), whereas the majority went far beyond that point. For those who pass the two-set threshold, Goriot et al. (2018) argue that the PPTV-4 is a reliable tool for testing young learners' L2 English lexical proficiency.

The PPTV-4 consists of an easel with 228 pages, each with four colored illustrations. It is divided into 19 separate units of 12 items each, with increasing difficulty, going from high to low frequency words (Dunn & Dunn, 2007). A starting age for each unit is given, so the first set is expected to be known by children two and half to four years old, and set 12 by 14–16 year-olds. After having introduced the test and performed one or two practice items, the test administrator says the target item out loud and the child is then asked to point to the illustration depicting that word. The test should be ended when the child has pointed to an incorrect illustration four times in a row.

The language diary (see Figure 1) is an instrument that has been used in several studies in order to obtain data about learners' exposure to English outside of school. It was originally used among upper secondary school students (Sylvén, 2007), but has since been developed and used also among other age groups (Hannibal Jensen, 2017; Olsson & Sylvén, 2015; Sundqvist & Sylvén, 2014; Sylvén & Sundqvist, 2012a).

As is illustrated in Figure 1, one page of the language diary lists a number of possible activities carried out in English, such as watching YouTube-videos, listening to music, and playing digital games. The second page asks whether the activities were performed by the child alone or in the company of parents, siblings, and/or friends, as well as whether English and/or Swedish were/ was used while playing digital games and what skill/s was/were used (speaking, listening, writing, reading). There were identical two-page openings for each day of the week, and the diary was filled out during one week.

The picture illustration task was administered using three illustrations developed by the Swedish Agency for Education, for the purpose of establishing immigrant children's level of English upon


**Figure 1.** The language diary

their arrival to Sweden (Skolverket, 2019b). The material in its entirety consists of a number of tasks, the aim of which is to more accurately be able to place newly arrived children at an appropriate level in school. The illustrations used in the present study are the final step in this placement package. Each illustration is in full color, and depicts a room, a village, and a city, respectively. All three pictures are full of details (e.g., animals, people, vehicles) and various types of action (e.g., reading, listening to music, walking a dog).

# *Data collection*

After having secured the interest of participation from two teachers, one in kindergarten and one in primary school, I visited the groups respectively to present myself and the study. During the first visit, the entire group in the respective setting was involved in discussions about languages, language learning, and language use. The children were invited to share their own views on these concepts, and they offered insights into their own linguistic landscapes.

In connection with the first visit, suitable times for subsequent visits were scheduled. The kindergarten group was visited three times for the administration of the PPVT-4. The first-grade group was visited four times for the same purpose.

A copy of the language diary was handed out to each child (regardless of whether they had agreed to participate in the study or not in order not to exclude anybody). Information to the parents about how to fill it out was attached. The instructions to the children and their parents were to be as detailed as possible, and to fill in information about English activities for each day during one week. Every morning, the teacher reminded the children about the diary, and in the weekly e-mail to the parents, a reminder was also included. The diaries were then collected by the respective teacher, and I picked them up at the next visit.

The administration of the PPVT-4 was done on a one-to-one basis. One child at a time came into a separate room, where I had set up the equipment needed to perform the task. Before starting the actual test, we talked about everyday issues in a very informal manner, in order for the child to feel at ease in the situation. Each child was then informed that he or she could stop at any time, and that they should do so as soon as they felt tired or lost interest in the task. I asked each child repeatedly during the test if they wanted to continue, and made sure they did not feel pressured to continue beyond their own limits. In line with what Goriot et al. (2018) suggest, each session started at the very beginning of the PPTV-4, instead of the recommendation made in the handbook to start at an age-appropriate level (Dunn & Dunn, 2007).

The picture description task was administered in much the same way as the PPVT-4, with the exception of a recorder being placed on the table. Each child was made aware of the presence of the recorder, and was asked if they felt comfortable with being recorded (which all of them did). As explained above, there were three different illustrations they could choose from, and after having made their choice, they were asked to tell me, in English, as much as they could about what they saw in the picture. Afterwards, each recording was moved from the recording device to a computer, and transcribed.

The transcriptions are verbatim, but do not contain any other information (such as pauses, etc.). The transcripts were cleared of everything except the child's own utterances, and subsequently,

each picture description was run through LexTutor (https://www .lextutor.ca/) in order to analyze the total number of words, and the distribution of words into frequency groups. The specific VocabProfiler used was VP-Kids. VP Kids is based on various studies into children's oral production and matches children's texts against 10 modified 250-word lists generalized from several empirical studies of children's oral productions (Murphy, 1957; Johnson, 1971; Hopkins, 1979; Moe et al., 1982). This version of the VP is to be preferred when analyzing young children's lexical growth, as the Classic or BNC versions of VP are too weak for that purpose (Horst & Collins, 2006), especially in the crucial K-2 phase (https://www.lextutor.ca/).

The group discussions in both groups focused on languages and language use. Among the first graders, a number of children had other L1s than Swedish. Spanish, Bosnian, and Arabic were among the languages represented. The kindergarten group was more homogeneous as regards L1s, only one of the participants had another L1 than Swedish, namely Arabic. During the discussions, it was evident that many of the children encountered and used English in a variety of contexts, an observation that was subsequently reinforced in the language diaries. The children in both groups seemed very much to appreciate the opportunity to talk about their own languages and language use.

### *Ethical considerations*

Conducting research among such young individuals as was the case in the present study, entails careful ethical consideration (Larsson, Williams, & Zetterquist, 2019). Ethical guidelines were adhered to (e.g., Mackey & Gass, 2011; https://codex.uu.se/). Information about the study was sent out to all parents, outlining the aims and methods of the study. In addition, the respective teacher offered details at information meetings. Parental consent was obtained for all participating children. Each child was informed about the voluntary nature about participation and that he or she could withdraw from the study at any time. In addition, as explained above, during the administration of the PPTV-4, every participant was told that they should only continue as long as they thought it was fun. Everybody, including those whose parents had not given their consent to participation was invited to take part in all the activities involved in the study in order to make sure no one felt left out. In subsequent analyses of the material, only those for whom consent had been obtained were taken into consideration.

In presentations and publications involving this study, all names have been anonymized, and the real name and exact locations of the school and preschool are not revealed.

### *Validity, reliability, and generalizability*

The validity of the present study is secured through the triangular methodology employed: talks with the respective group, L2 English proficiency tests, and interviews with both children and their teachers.

Face validity, defined as "the degree to which test respondents view the content of a test and its items as relevant to the context in which the test is being administered" (Holden, 2010), of the tests was ensured by using tests developed for this particular age group. All participants were informed about the purpose of the tests, both at group level in both groups and individually, and that their participation was voluntary. Those who decided to participate seemed to enjoy the tasks they were asked to do.

The broader concept of validity is defined by Messick (1989, p. 13) as "an integrated evaluative judgement of the degree to which empirical evidence and theoretical rationales support the adequacy and appropriateness of inferences and actions based on test scores or other modes of assessment". In the present study, validity is deemed as high as appropriate and relevant test instruments were used to investigate L2 English proficiency among VYLs. Inferences based on the results, however, need to be made with great caution due to the low number of participants.

The reliability of the results obtained in the study is likewise deemed as high. Great efforts were made to ensure reliability. First, by asking the children orally to give an account of their exposure to English, what was reported in the language diaries was controlled for. In the vast majority of cases, both sources coincided. Second, to make sure that an authentic depiction of the level of proficiency was shown, each child was given every possibility to display their abilities in English in the two tests administered. And, third, measures were taken to ensure that parents and teachers were in agreement with the aims and methods of the study.

As the study only included two groups of children, one in kindergarten and one in first grade, the results are in no way generalizable to a larger population. However, the findings are an indication of the influence of L2 English on the young generation and how it can have effects on an individual as well as a societal level.

# **Results**

In this section, results from the language diary and the two tests of language proficiency are accounted for. First, the results of the language diary are presented, followed by those obtained on the PPVT-4 and the picture description task. Finally, figures illustrating the language diary in correlation with the results on the PPVT-4 and picture description task respectively are offered.

### *The language diary*

The language diary was filled out by a total of 15 individuals, five in the kindergarten group and ten in the group of first graders. In Table 2, the results obtained are illustrated, with time of exposure per activity, total time of exposure, gender and group.

As is evident from the numbers reported in Table 2, the most popular activity overall is watching YouTube, followed by computer gaming. Watching movies and listening to music show roughly the same figures and come in third place, while reading and speaking are activities that are fairly uncommon. The spread of engagement in various activities should be noted; while the average time for exposure to English was 430 minutes, the standard deviation was 361.

In Figure 2, the results are shown for the language diary among the kindergarten group.

Mehmet stands out in this group with his 1230 minutes of exposure to English during the week reported, compared to Moya



**Figure 2.** Total activity (in minutes during the week investigated), kindergarten

**Figure 3.** Total activity (in minutes during the week investigated), first-grade

who reported 160. The average for these children was 657, with a standard deviation of 450 minutes (max: 1230, min: 160).

Figure 3 illustrates the results for the first graders.

The mean for the first-graders is 317 minutes, with a standard deviation of 265 (max: 983, min: 75). As is illustrated in Figure 3,

**Figure 4.** Results on PPVT-4, kindergarten

the spread is vast between Ava, who was exposed to English for a total of 983 minutes during the week of the language diary, and Louis whose corresponding number of minutes was 75.

### *The PPVT-4*

In the pre-school group, a total of seven children completed the PPTV-4 test. Figure 4 illustrates the results.

The black line in Figure 4 illustrates the total number of items completed, and the grey the number of correct items. On average, the preschoolers completed 62 items, with a standard deviation (SD) of 14.6. The average score of correct answers was 43 (SD = 18.7). The spread is seen between Sue at the low end of the continuum, who did 36 items, and got 22 out of them correct, and Ethel at the opposite end, who did 72 items with no errors.

In the group of first graders, a total of 13 children completed the PPVT-4, and in Figure 5, the results are illustrated.

In Figure 5, we see that the first graders, in general, aimed higher in the sense that they made attempts at a larger number of total items on average, namely 89 (SD = 36.5). Their average score of correct items was 59 (SD = 30.4). Here, the spread is between Eve, who aimed for 24 items, and scored 13 of them correct, and Ava whose corresponding figures were 144 and 132.

**Figure 5.** Results on the PPVT-4, first grade

**Figure 6.** Number of words produced, first grade

# *The picture description task*

The results of the picture description task are presented in the form of total number of words produced by each child. As accounted for above, the picture description task was only administered among the first graders, as it was deemed too demanding for the children in kindergarten. Figure 6 illustrates the results.

As seen in Figure 6, the spread is enormous, from Ava's 824 words to Vincent's 17. The mean in this group is 121.92 words, with a standard deviation of 221.78. While all of the children use words from levels 1, 2, and 3, Ava and Dave are the only ones uttering words belonging to level 10 (examples from Ava: *age*, *search*, *zebra*; examples from Dave: *lamb*, *zebra*). Among the off-list words, that is words that are not included in any of the frequency levels, Ava uses the largest number with words such as *completely*, *definitely*, *normally*, *includes*, *otherwise*, and *toddler*.

### *Correlations between exposure and test results*

In order to check for possible correlations between amount of exposure to English and performance on the PPTV-4 test and the picture description task, correlation analyses were done. Among the first graders who completed the PPTV-4 test, only nine also filled out the language diary. Due to the low number of data points, it is questionable to carry out any statistical analyses and results should definitely be interpreted with great caution. However, a bivariate correlation analysis was run which showed that there is a statistically significant correlation between amount of exposure and results on the PPTV-4 test at the 0.01 level (see Table 3).

In Figure 7, the correlations between the exposure to English outside of school and the results on the PPTV-4 among the first graders are illustrated.

Among the participants who did the picture description task, only nine also filled out the language diary. Again, the results from the bivariate correlation analysis should therefore be interpreted with caution. The analysis showed that there is a statistically significant correlation between amount of exposure to L2 English and number of tokens produced in the picture description task at the 0.01 level (see further Table 4).

In Figure 8, these correlations are illustrated, showing how the two lines are fairly parallel, with some notable exceptions.

In the following section, these results will be discussed.


**Table 3.** Correlation between amount of English activities and results on the PPTV-4 test

\*\* Correlation is significant at the 0.01 level (2-tailed)

**Figure 7.** Correlation between amount of English exposure and results on the PPTV-4


**Table 4.** Correlation between amount of English activities and results on the picture description task

\*\* Correlation is significant at the 0.01 level (2-tailed)

**Figure 8.** Correlation between amount of English exposure and results on the picture description task

# **Discussion**

Before dwelling on the results themselves, some methodological issues merit attention. One of the aims of the present study was to test various instruments in order to investigate VYLs exposure to and proficiency in English. For this purpose, group discussions, language diaries, and the administration of language test were used. In the following, these instruments will be evaluated and discussed.

Group discussions were used in order for the children in both groups to become acquainted with me and the topic under investigation, as well as for me to get to know the children a little bit. These discussions served their purpose very well. The children were allowed to talk about their own language use, which one/s they used, when, where and with whom. This was an apparent topic of interest to the majority of children. In the end, we had to cut these sessions short, as they could have talked for much longer than was planned for. In future studies involving these age groups, therefore, more time and effort should be put into this kind of activity. One idea may be to arrange for smaller groups to discuss, and to tape-record such discussions.

The language diary (see Figure 1) was filled out by the children's parents. This in itself is problematic, as the parents cannot be aware of everything their child does, or in what language. Therefore, the results of the diaries should be viewed only as crude approximations of reality. However, in subsequent talks with the children, what had been reported in the diaries seemed to be in agreement with what the child him/herself told me verbally. Time is also an issue. It may be very difficult for parents to such young children to find the time and sit down to fill the diary out. A final consideration is the fact that only written instructions were offered to the parents. Ideally, a presentation of and introduction to the diary should have been done in a real-life meeting, where questions could be asked. Unfortunately, this was not possible in the present study but will definitely be part of any subsequent studies using the same instrument.

As regards the use of the PPVT-4, it has been pointed out above that it was originally intended to be used to establish vocabulary knowledge among L1 speakers of English. Something similar targeting L2/FL users of English would, of course, have been preferred, but in the absence of such tests this served its purpose well in the present study (cf., Goriot et al., 2018). The fact that the children seemed to appreciate the format is the most important one, and gives face validity to the test.

The picture description task also served its purpose very well in the present study. The children liked the illustrations, and every child could at least say something in English about the chosen picture. When VYLs are in focus, the opportunity for them to really show what they know is important, and using this type of task gives everybody a chance to do so.

Moving on to the results themselves, three main points, based in the research questions outlined above, will now be discussed. First, the time and amount of extramural exposure to L2 English among VYLs; second the apparent influence of such exposure; and third, the differences in proficiency in L2 English. The pedagogical implications will be discussed in a separate section.

Exposure to L2 English among these VYLs, as evidenced in the language diaries, varies to a very large degree. While some children have virtually no such exposure at all, others are exposed to English more or less throughout the day. Ethel, five years old in kindergarten, is one example of the latter group. She uses English to communicate with her peer Mohammad, a recent immigrant to Sweden, to help him when he does not understand the spoken Swedish, and to play with him. When she goes home, she watches YouTube videos and plays digital games, such as the Sims and Minecraft, all in English. In addition, she and her older sister, Ava, use English when they play together. All of these situations exemplify what Bialystok (1981) refers to as functional practice and provide ample opportunities for some of the important building blocks of L2 acquisition; input, output and interaction (Gass & Mackey, 2006).

At the other extreme, there is Louis in first grade, who only comes in contact with English through singing some songs together with his mother and sister, and a little bit of playing the computer game NHL together with a friend, during a total time of 75 minutes during the week of the diary. In other words, there are vast differences in type and amount of exposure to English among these children. This means that there are various types of individuals represented in the sample: those who are exposed several hours per day and those whose total exposure barely reaches one hour per week. Even though the spread is large in this sample, it correlates with previous findings among young individuals. Hannibal Jensen (2017, p. 10), for instance, found that 8 and 10 year-olds in Denmark on average were exposed to L2 English in their spare time for 366 minutes, with a standard deviation of 316. Likewise, the Swedish 11–12 year-olds in Sylvén and Sundqvist (2012a, p. 311) reported an average of 564 minutes of English exposure, with a standard deviation of 474. Thus, it may be concluded that the VYLs in the present study adhere to similar trends previously reported among slightly older individuals.

As regards the influence of exposure to English among VYLs, it seems as though there are some indications of correlations between type and amount of exposure on the one hand, and L2 English proficiency on the other. Ava in first grade is an outstanding example of someone with large amounts of exposure to English. Mainly, she watches YouTube videos and plays digital games of her own choice. Her level of proficiency is the highest measured in this study. The results on the PPVT-4 indicate a level of a native 14–16 year-old. In the picture description task, she produced the largest number of words, totaling 824. Further investigation into her unusually strong L2 English skills are warranted, but one explanation may be that all she does in English in her free time is at her own will. She decides what YouTube clips to watch and what digital games to play. Thus, the activities are driven by integrative, rather than instrumental, motivation, or what Dörnyei (2005, pp. 65–119) refers to as self-motivation. This, in turn, contributes to the large investment in these activities (Norton, 2013). Another explanation most likely is found in the fact that Ava and her little sister Ethel use English as their mode of communication while playing. The fact that they together, in a non-threatening environment, can play with the language, imitate what they have heard somebody say in, for instance, a YouTube clip, and practice on their own terms is key to the development of proficiency level (cf., Holmes & Myles, 2019). A positive spiral is created between input, output and interaction, and language development takes place and progresses (Gass & Mackey, 2006). It should be pointed out that, according to her teacher and parents, Ava's proficiency in L1 Swedish is normal, and she does not excel in any other school subjects.

In contrast to Ava, there is Louis, who barely had any exposure to English. On the PPVT-4, he scored 46, which was the second lowest result among the first-graders. On the picture description task, he produced a total of 41 tokens. Ava and Louis are examples of the interrelatedness at both ends of the continuum of investment and language learning (Norton, 2013). Ava invests a great deal of her time and interests in activities involving English, contributing to high proficiency levels. She often talks about her knowledge of English, how she loves using English as much as possible, and how she counts on using it in the future, thus displaying clear signs of identifying herself with her high proficiency in English. In addition, many of her peers refer to Ava as the best one in the class as regards knowledge of English, thus contributing with an external identification of someone who knows English well. At the other end of the continuum we find Louis who prefers to engage in activities not involving English which most likely impacted on his low results on the current measures. Instead of engaging in, for instance, digital games or other computer related activities, Louis is intensely involved in sports and outdoor activities.

There are also those whose extramural exposure does not correlate with their results in a straight forward manner. One of them is Mary who was exposed for a total of 360 minutes, which places her slightly above the mean of her age group. Her result on the PPVT-4 of 59 matches the group mean of 61 fairly well, but on the picture-description task her scores were among the lowest measured with only 23 tokens. There are several plausible explanations to these findings. One is that there were signs of her not being perfectly at ease in performing the picture description task. For instance, she had to be probed and scaffolded to a larger extent than was the case for the other children in order to produce output. Another is that the picture description task involves language production, which is much more challenging than what is required in the PPVT-4 test, where a finger pointing to the relevant image is all that is needed. It is also possible that the activities she engaged in were not conducive to L2 learning.

The causal link between exposure to and interest in L2 English, however, has yet to be established. In other words, we do not know if those who spend time on activities in English do so because of an initial interest in language in general, and English in particular, or if the availability of English through various activities leads to an interest in taking part in them. However, the fact that the children involved in the present study are as young as 3 to 8 years old, indicate that the mere existence of such a multitude of possibilities of engaging in activities involving English is what sparks the interest to increase the exposure. Needless to say, further studies involving this age group are decisive in order to test this hypothesis.

A vital issue in need of discussion is the enormous variation of L2 English proficiency seen in this study. First, it should be pointed out that the sample is very limited, and only receptive vocabulary knowledge and, to a limited extent, oral production have been tested. Nevertheless, the results indicate levels of proficiency ranging from those expected from 2–4 year-old natives as defined in the PPVT-4 test books (Dunn & Dunn, 2007), all the way up to the level of 14–16 year-old natives. It is intriguing that despite the fact that the study was conducted among children in preschool and in first grade, who had not yet started any formal education in English, none of the participants showed evidence of being complete novices to the language. It is an indication that the presence of English in everyday society, as exemplified above is the case in Sweden, indeed leads to some basic knowledge even without any formal education, and that in some cases it results in English becoming more like another L1. This, in turn, has some pedagogical implications which are discussed next.

# **Pedagogical implications**

The results of this study illustrate the vast heterogeneity in level of proficiency in English found among very young children in today's Sweden, and surely other countries with a similar context as regards availability of English (see, e.g., De Wilde, Brysbaert, & Eyckmans, 2020; Hannibal Jensen, 2018). Such heterogeneity puts strains on the educational system, where, thus far, the teaching has been streamlined for students expected to be novices when they start school. The findings here rather suggest that some school-starters are more or less fluent in English, alongside others who only have a very basic knowledge. Further research needs to establish the generalizability of this study.

Should the findings be corroborated in larger studies, English as a school subject needs to be revised. In order for VYLs who already are well acquainted with the language not to lose interest in the subject of English in school, other types of teaching than those implemented presently are necessary. This, in turn, necessitates measures to be taken to collect information about each individual's previous exposure to English (see, e.g., Sundqvist & Sylvén, 2016) as well as to perform some kind of early placement test in order to identify the various levels of proficiency in English among VYLs. In many cases, it may be relevant to teach English as a second, rather than a foreign, language. And, as touched upon above, in some cases, it may even be relevant to teach it as another first language.

# **Future work**

As has repeatedly been pointed out, the study presented here is very limited in scope. Therefore, a natural continuation of this line of research would be to include larger numbers of informants representing all parts of the country. Are there, for instance, differences between children growing up in urban vs. rural environments as regards the use of English in connection with various activities and its effects? In addition, in-depth, qualitative case-studies of individuals exposed to large amounts and different types of English and their respective levels of English proficiency, are necessary for a better understanding of any causal relationships. Effects of extramural and pre-school exposure to English on the teaching of English as a subject in school also need to be investigated. The findings reported on in this chapter indicate that profound changes to the English subject are to be expected in the future.

# **References**


Lindgren, E., & Enever, J. (2017). Employing mixed methods for the construction of thick descriptions of early language learning. In

J. Enever & E. Lindgren (Eds.), *Early Language Learning. Complexity and mixed methods* (pp. 201–221). Multilingual Matters.


analysis. *ReCALL, 24*(3), 3561–3381. DOI: https://doi.org/10.1017 /S0958344012000195


*Digital games in language learning and teaching* (pp. 189–208). Palgrave Macmillan.


*ReCALL, 24*(3), 302–321. DOI: https://doi.org/10.1017/S09583 4401200016X


# **5. Attitudes and ambiguities – teachers' views on second foreign language education in Swedish compulsory school**

*Gudrun Erickson, Camilla Bardel, Rakel Österberg & Monica Rosén*

# **Introduction**

Sweden is a multilingual country with around 20% of the population born outside Sweden4 and up to 200 different languages spoken in society (Institutet för språk och folkminnen, 2020). This is reflected in the national curricula for compulsory and upper secondary school, with separate syllabuses for Swedish, Swedish as a second language, Mother tongue tuition5 , Sami, and Swedish Sign language for the Hearing. In addition, three syllabuses are provided for so-called foreign languages, namely English, Modern languages and Chinese.

The overarching aim of the questionnaire underlying the current study is to give voice to a large group of teachers of Second foreign languages (SFLs), by mapping, describing and reflecting on their reported practices and perceptions regarding a substantial

#### **How to cite this book chapter:**

Erickson, G., Bardel, C., Österberg, R. and Rosén, M. (2022). Attitudes and ambiguities – teachers' views on second foreign language education in Swedish compulsory school. In: Bardel, C., Hedman, C., Rejman, K. and Zetterholm, E. (Eds.), *Exploring Language Education: Global and local perspectives*, pp. 157–201. Stockholm University Press. DOI: https://doi.org /10.16993/bbz.f. License: CC BY 4.0.

<sup>4</sup> https://www.scb.se/hitta-statistik/sverige-i-siffror/manniskorna-i-sverige /utrikes-fodda/

<sup>5</sup> Sweden has five official national minority languages: Finnish, Meänkieli, Romani chib, Sami, and Yiddish. Sami has its own syllabus, whereas the other four are included separately in the syllabus for Mother tongue tuition.

number of issues within the broad domain of language education, including aspects of learning, teaching, assessment, and framework factors. The extensive teacher questionnaire (translated English version in Appendix 1) was part of a research project on Modern languages in compulsory school (see below). Some of the most salient results were reported at an early stage to the category of stakeholders enabling the survey, namely teachers (Erickson et al., 2018).

In this chapter, three areas often discussed in the Swedish context are focused upon, namely *professional satisfaction, teachers' target language use* and *the curricular status of Modern languages*. These issues are also discussed in reports and studies on Modern languages (e.g., Lärarnas Riksförbund, 2016; Skolinspektionen, 2010; Tholin, 2019). Furthermore, the three areas can be seen as representing three fundamental levels of education, namely the individual, pedagogical and structural levels (cf. Erickson et al., 2015). Some background information is first given about the TAL project and about discussions concerning the school subject Modern languages in Sweden. In addition, the conceptual basis for the study is outlined and the iterative development of the questionnaire is described. Before reporting the results, the data collection, the sample achieved, and the analytical approach taken are presented. The results section that follows is organised according to the three above mentioned issues. In the last part of the chapter, the results are discussed and some possible implications of the findings are outlined.

# **Background**

In the following, the research project within which the current study was conducted is presented, as are foreign languages in Swedish compulsory school, including current discussions regarding the subject Modern languages, the latter underlying the choice of the three issues focused upon in the present chapter.

# *The TAL project*

The purpose of the project *Learning, Teaching and Assessment of Second Foreign Languages – an Alignment Study on Oral*  *Language Proficiency in the Swedish School Context*, funded by the Swedish Research Council between 2016 and 2018, was to achieve a better understanding of SFL in Swedish compulsory school, with special attention given to oral proficiency (Granfeldt et al., 2016).6 The target languages in focus were French, German and Spanish.

The project adopted a holistic perspective on language education and examined factors at the individual level, the school level and the societal level. A major part of the project consisted of surveys with school-leaders and teachers, combined with field studies at 15 schools in the country, drawn from the initial sample of schools. The school leaders' questionnaire was sent to a stratified random sample of 416 schools (with sampling assistance from Statistics Sweden/SCB) in September 2016. The response rate was 34% (n=143), and the questionnaire targeted school leaders' educational and professional background and attitudes to second foreign languages, as well as school frame factors, including resources (Granfeldt et al., 2019).

The questionnaire examining teachers' perspectives, which is the focal point of the current chapter, was developed in a collaborative process involving language teachers as well as researchers and included piloting at different stages. It was administered in 2017 and sent to teachers of French, German and Spanish in the schools that had been sampled for the project, as mentioned above (see further below, under Methodology).

### *Second foreign languages in Sweden*

The tradition to study modern languages in Sweden goes back to the 19th century, when, initially, French was the first foreign language. It was also possible to study German and English in secondary education (Hyltenstam & Österberg, 2010). In 1859, German became the first foreign language (Psykologisk Pedagogisk Uppslagsbok, 1956) and remained so until 1946, when after the end of World War II English took over this role (Hyltenstam & Österberg, 2010). In 1952/53, English was made a compulsory subject for all students as from school year five. This starting

<sup>6</sup> Project website at https://www.tal.lu.se

point has since then been gradually lowered, today being year three at the latest (Johansson, 2004; Malmberg, 2000; Tholin & Lindqvist, 2009).

With English as the first foreign language, studies of French or German became possible from school year 7, and from 1962 to 1969 it was compulsory to take one of these languages in order to be admitted to upper secondary school (Tholin, 2019).

There have been important policy changes regarding SFLs since 1969, two of which deserve to be highlighted here. First, in 1994, Spanish was introduced as a possible alternative to French and German and is today by far the most frequent choice among beginners (Bardel et al., 2019; Tholin, 2019). Second, the latest starting point has recently been lowered from year 7 to year 6 (Persson, 2018). Both these changes, together with others, for example an increase of teaching time (Tholin, 2019), aimed at raising motivation among students to study a second foreign language (Bardel et al., 2019). It also needs to be pointed out that Modern languages is part of the so called 'Language choice' (Swe: *Språkvalet*), a group of language electives in Swedish compulsory school, from which students have to choose one. This group currently (2020) comprises Modern languages, additional studies in Swedish and/or English, Mother tongue tuition, and Swedish Sign language for the Hearing (Skolförordningen, 2011:185; Tholin, 2019).

However, in comparison with English, the SFLs face very different conditions. The strong position of English concerning attitudes, motivation and proficiency level is far from the same when it comes to French, German and Spanish (Bardel et al., 2019; European Commission, 2012a; European Commission, 2012b). Furthermore, in the same way as school children today learn English in out-of-school activities, younger children also have many chances to acquire some English already in preschool age, not least via digital media (Sylvén, 2022).

It is important to recall that the possibility to study an SFL, introduced after English, is offered to all students.7 In compulsory

<sup>7</sup> There is also a possibility to take a third foreign language, starting in school year 8, within the so-called 'Student's choice' (Swe: *Elevens val*). Very few students make this choice (in 2019/20, 0,8%; n=966)

school, at least two of the languages French, German and Spanish must be offered by the school organiser within the Language choice, and a large majority of children – 88.7% of the cohort in the autumn of 20198 – start with one of these languages, in spite of its optional status. Of the total cohort in 2019, 17.5% of the students chose French, 19.8% German, and 51.5% Spanish (0.2% other modern languages).9

While most students start with a second foreign language, a number of them drop out during the years up to year 9 (Krigh, 2019; Tholin, 2019). According to statistics from the National Agency for Education, between 2015 and 2019 final grades of Modern languages were awarded to an average of 69% of the students.10

Tholin's (2019) and Krigh's (2019) studies offer an important background to the current study, but compared to the large and developing body of educational research on the learning, teaching and assessment of English in Sweden, there is a lack of studies on the second foreign languages. The TAL project, from which the data for the current text emanate, represents an exception.

#### *Current discussions about Modern languages*

Certain issues regarding SFLs in Sweden are the subject of recurrent discussions in the media and among different categories of stakeholders, and have been so for quite some time. One of these issues concerns Modern language teachers' satisfaction with their work, which has been shown to be alarmingly low. An example of this is a survey conducted by one of the large teacher unions showing that, during the past few years, more than 60% of

8 https://www.skolverket.se/skolutveckling/statistik/sok-statistik-omforskola-skola-och-vuxenutbildning?sok=SokC&omrade=Skolor%20och %20elever&lasar=2019/20&run=1 Table 7A

https://www.skolverket.se/skolutveckling/statistik/sok-statistik-om-forskola -skola-och-vuxenutbildning?sok=SokC&omrade=Skolor%20och%20elever &lasar=2019/20&run=1 Table 7B

<sup>9</sup> See footnote 5.

<sup>10</sup> https://www.skolverket.se/skolutveckling/statistik/sok-statistik-om -forskola-skola-och-vuxenutbildning?sok=SokC&omrade=Betyg%20 %C3%A5rskurs%209&lasar=2018/19

Modern language teachers have considered leaving the profession (Lärarnas Riksförbund, 2016). Reasons vary, with low salaries, heavy work load and lack of in-service education being the most frequently mentioned negative aspects. Another issue, which has been a topic for discussion at least since the mid 1900s, was brought forward in 2010 by the Swedish Schools Inspectorate in a critical report concerning SFL education in compulsory school (Skolinspektionen, 2010), highlighting the question of target language use, which was found to be low both among teachers and students. The Inspectorate saw this as evidence of weak compliance with the national curriculum and syllabus that are characterised by a clearly functional and competence-based view of language, with active language use in focus. A third example of a discussion that has been going on for a long time and still evokes strong opinions is the status of SFLs, or more precisely whether it should be mandatory for all students in compulsory school or remain an elective subject. Here, a certain change can be noted, for example when comparing a study conducted by the National Board of Education in the late 20th century (Skolöverstyrelsen, 1991) and the survey conducted by a teachers' union some twenty years later (Lärarnas Riksförbund, 2016). In the older study, only about 12% of the responding SFL teachers expressed a positive attitude to a mandatory second foreign language, whereas the corresponding response in the more recent survey was roughly 50%.

### *Focus of the study*

The three issues mentioned – professional satisfaction, teachers' target language use11 and the curricular status of Modern languages – constitute the focal point of the present text, chosen because of their importance in the professional and policy-related debate referred to above. They represent different levels of language education, with professional satisfaction at the individual level and teachers' target language use at the pedagogical

<sup>11</sup> For reasons of focus and space, students' target language use, albeit of obvious interest and connected to teachers' use, is not focused upon in the current chapter.

level, whereas the curricular status of Modern languages in Swedish compulsory school represents the structural level. Reviewing the somewhat disparate sources mentioned in the previous paragraph, one can conclude that they try to deal with problems of attitudes and motivation in relation to other foreign languages than English. The relevance of bringing these three issues together and trying to find out what they mean to teachers who work in the field becomes obvious. As shown below, the different parts of the teacher questionnaire cover the three levels of language education, the individual, the pedagogical and the structural, by asking questions about the respondents' background; learning and teaching, assessment and grading; frame factors and attitudes.

The three different issues focused upon, representing three levels of language education, taken together require a broad conceptual basis at the individual, teacher level, the pedagogical content level, and the structural level (see further *Conceptual considerations*, p. 167, and *The Questionnaire*, p. 169). Furthermore, they have all been the subject of previous research, although not in connection with each other and not with the methodology used here which further explains the rationale behind this study. Conceptual considerations of significance to the content of the questionnaire and the analysis of the responses will briefly be further discussed in the Methodology section of the text. First, however, some mentioning of previous studies related to the three issues seems relevant.

# **Previous research**

As for *Professional satisfaction*, a number of national and international studies have been conducted. In Sweden, broad national evaluations were conducted between 1989 and 2003, all of them comprising questions to teachers (e.g., Skolverket, 2004). Internationally, teachers' attitudes to their profession have been studied more recently in the TALIS surveys (OECD, 2019; Skolverket, 2020a). TALIS, *The Teaching and Learning International Survey*, is organised by the OECD (Organisation for Economic Co-operation and Development) and focuses on school leaders' and teachers' work. In the latest survey (2018), 48 countries around the world took part. Results show that, on the whole, nine out of ten Swedish teachers of a broad range of subjects declared themselves satisfied with their work, with only ten per cent regretting their choice of profession. Regarding the perceived status of their work, Swedish teachers are considerably more hesitant than their international colleagues, the OECD mean for high status being 26% and the Swedish corresponding proportion 11%, to be compared with 18% in Denmark, 35% in Norway and 58% in Finland (Skolverket, 2020a). Finally, it can be noted that surveys among teachers of SFLs regarding professional satisfaction are scarce, the teacher union survey from 2016 being an exception. However, studies of English as a foreign language (e.g., Bonnet, 2004; Skolverket/Erickson, 2004), highlight a number of opinions expressed by teachers, concerning for example the lack of in-service education, and a perceived low degree of interest in the language teaching profession.

*Teachers' target language use* (TLU) has quite a prominent role in the current national syllabuses for foreign languages in Sweden (Skolverket, 2011). In the syllabuses for English and Modern languages for upper secondary school, it is explicitly stated that "Instruction should in all essence be conducted in English/ the target language".12 However, this sentence is not included in the syllabus for lower secondary school, although there are apparent similarities in the documents concerning the description of other aspects of language learning and teaching. This can be assumed to be related to the age of the students rather than the level of competence, this since Modern languages in upper secondary school range all the way from beginners to the highest level described in the syllabus.

TLU in foreign language classrooms has been the object of a large number of studies, mostly focusing on teachers (for studies before 2002, see Turnbull & Arnett, 2002; for more recent research on the issue, see Shin et al., 2019). In addition, a consistent monolingual, target language approach for teachers as well

<sup>12</sup> Swe: *Undervisningen ska i allt väsentligt bedrivas på engelska/ målspråket*.

as for students has been – and still is – strongly promoted in discussions about learner autonomy and the implementation of the action-oriented language competence described in the CEFR (Little, 2009; Little et al., 2017). Although a certain change is noticeable regarding the perception of TLU as self-evident, or even indispensable (e.g., Krulatz et al., 2016; Littlewood & Yu, 2011), monolingual approaches still have a very strong position in language education. In the Nordic context, however, studies of TLU are relatively scarce, especially regarding languages other than English. An exception is Stoltz (2011), who studied the use of French among teachers and students in two French upper secondary language classrooms in Sweden, finding that Swedish was used to a considerable extent, most often related to specific instructional activities and with the aim of facilitating students' understanding. In a recent study, also of French, set in a Norwegian lower secondary school context, Norwegian L1 was found to be the language of instruction in most classes (Thue Vold & Brkan, 2020) and that students' use of the target language was clearly limited. Further, the frequency and effects of teachers' target language use were studied in the European Survey on Language Competences (European Commission, 2012b). Results, albeit not all of them significant, indicate that "the more teachers speak the target language during lessons, the higher the score on the language test" (p. 64).

*The curricular status of Modern languages* has been discussed for a very long time in Sweden, but the subject has never been mandatory for all students. As from the 1960s, there were two levels of courses: one 'general' and one 'special'13, where the general course was intended to offer a less demanding alternative. These 'alternative courses' were first introduced for English and Mathematics and later also for French and German (Malmberg, 2000). Both courses granted access to upper secondary school (Giota & Emanuelsson, 2015). The existence of the alternative SFL courses became fairly short, and they were abolished in 1980. The two courses for English and Mathematics were kept but were increasingly criticised, not least for

<sup>13</sup> Swe: *Allmän kurs* and *särskild kurs*.

contributing to students' tactical choices and for increasing segregation based on class and gender (Giota & Emanuelsson, 2015; Lindblad & Eriksson, 1987). The courses were abolished when the 1994 curricula were launched (Malmberg, 2000; Marklund, 1985).

As mentioned in the beginning of the current chapter, considerable structural changes regarding Modern languages were introduced in the 1994 national curricula, with the ambition to strengthen students' motivation to learn more languages. Periodically, the issue of a mandatory SFL for all students has been debated in the general media and also in teachers' journals, often with strong opinions expressed both for keeping the optional status and for introducing obligatory SFL. However, political discussions have not been very loud, and it seems clear that opinions about this issue are often divided, also within political parties. The National Agency for Education has approached the Ministry for Education on several occasions since the early 2000s, most recently in 2018 (Skolverket, 2018), suggesting reforms to make more students study a second foreign language, not least by abolishing the option to choose additional Swedish or English, but so far no changes have been made.

# **Methodology**

The methodology of the study will be described from three points of view, namely the development of the questionnaire; the collection of data, including the achieved teacher sample, and the analytical approach.

# *The development of the TAL teacher questionnaire*

The development of the TAL teacher questionnaire (henceforward, the TTQ) was preceded by a number of discussions within the project group and with different educational researchers and experienced teachers, in particular regarding recipient related issues, ranging from sampling to delivery mode and format. Furthermore, conceptual and empirical considerations and analyses had an essential role in the actual development of the instrument.

#### *Conceptual considerations*

The overarching purpose of the TTQ was to focus on a broad spectrum of issues related to teachers' experiences and perceptions of their profession and to offer possibilities to view these aspects from different angles. Consequently, a broad conceptual basis was needed for the development of the TTQ, but also for the analyses and interpretations. In this study, research on teacher cognition (Borg, 2003; van Driel & Berry, 2012) was essential, as were theories – general and subject specific – on Pedagogical content knowledge, PCK (Shulman, 1986; Watzke, 2007). A conceptual basis in theories about teacher cognition proved highly relevant in the analyses of all three issues focused upon in this chapter. PCK, together with the Common European Framework of Reference for Languages, CEFR (Council of Europe, 2001), on language learning, teaching and assessment, were indispensable in approaching and analysing the whole language educational area focused upon in the questionnaire, in particular the question of teachers' target language use. Theories about communicative competence and its role in action oriented language learning, teaching and assessment (Hymes, 1972; Canale & Swain, 1980; Larsen-Freeman, 2000) were of obvious importance, separately within different domains, but also as expressed and operationalised, for example, in the CEFR and its accompanying Companion Volume (Council of Europe, 2020), and in the Swedish national curriculum and language syllabuses (Skolverket, 2011), the latter affecting, in an obvious way, language teachers' daily practices.

Furthermore, to better capture aspects of systemic, school and collegial environment, Frame factor theory was relevant, that is, theories focusing on external factors affecting the lives of schools and teachers (Dahllöf, 1967; Lundgren, 1999), as well as work on Practice architectures (Kemmis et al., 2014), including so called sayings and doings (Schatzki, 2010), which further strengthen the social aspects by adding relations between different agents to the landscape of factors forming conditions and practices at schools. Frame factor theory – both regarding material and relational aspects (Lundgren, 1999 and Kemmis et al., 2014) – proved very useful in interpreting and analysing professional satisfaction as well as the curricular status of Modern languages.

### *Empirical considerations*

Three empirical paths were followed to optimise the development of the questionnaire. One of these was to form a reference group, including people with different competences related to language teaching, learning and assessment, most of them also teachers of languages. A number of the members had substantial experience of language test development at the national level, which meant that they were well acquainted with the national curricula and syllabuses and that they worked with teacher questionnaires on a regular basis. In the group, some of the members also had experiences from different research projects related to language education at large to bring into the work. This group met throughout the development process and provided valuable comments and advice to the project group on the different steps undertaken and on the questionnaire as a whole.

A second way to develop the TTQ was to study previous examples of national and international questionnaires for language teachers. Nationally, this meant examples from National evaluation rounds of English (1989–2003) and a research project focusing on language teachers' role in assessment (Skolverket/Oscarson & Apelgren, 2005; Oscarson & Apelgren, 2011). In addition, the previously mentioned survey targeting the subject Modern languages in Swedish schools, labelled 'Languages – so much more than English'14 served as a useful source (Lärarnas Riksförbund, 2016). Internationally, questionnaires from two European surveys were studied: 'The Assessment of pupils' skills in English in eight European Countries' (Bonnet, 2004, ed.) and 'The European Survey on Language Competences' (European Commission, 2012c). Although most of these questionnaires focus on assessment, they all contain a substantial number of relevant questions related to language education in a broad sense.

Finally, in the empirical phase of developing the TTQ, different rounds of questions and sets of questions were piloted, analyzed and discussed in the reference and project groups. Preliminary versions of the full questionnaire were pretested in three different rounds with language teachers from different parts of the country

<sup>14</sup> Swe: *Språk – så mycket mer än engelska*.

and adjustments were made based on their comments. All in all, this meant that approximately 40 people were in touch with the TTQ during its developmental stage and were able to actively influence its final design.

With these actions undertaken in the development of the TTQ, we feel confident that the content and item formats have been firmly grounded in a combination of conceptual considerations, current practice and previous research.

### *The questionnaire*

A digital questionnaire was considered most feasible for respondents as well as researchers, and the aspiration was also to, thereby, optimise the response rate. As for format, it was deemed important to combine selected response and open-ended questions, thus creating possibilities for initial overviews and comparisons of the data as well as deeper understanding of the responses. The final questionnaire (Appendix 1)15 that was answered anonymously online comprises 50 questions, divided into four thematic sections focusing on Respondent background information (Q 1–13); Learning and teaching (Q 14–24), and Assessment and grading (Q 25–34) [cf., for example, Council of Europe (2001); Shulman (1986)]; Frame factors and attitudes (Q 35–50), [cf. Borg (2003); Kemmis et al. (2014); Lundgren (1999)]. In particular, the sections focusing on learning, teaching and assessment were related to the current national syllabus for Modern languages, but a number of questions also focused on issues beyond the scope of the national regulatory documents, for example concerning teaching methods and practices (cf. Watzke, 2007), which in accordance with the national curriculum are to be decided locally, by individual teachers and schools. The questions were of different length and complexity, with eight of multiple-choice (MC) type only, 11 requiring open responses only and 31 using a mix of closed and open formats, with MC items or Likert scales plus space where respondents were prompted to comment on the issue in focus. The last two questions differed from the others in asking teachers

<sup>15</sup> In the appended questionnaire, the 23 questions (46% of all questions) actively used in the analyses are italicized.

to express themselves freely on the contribution of Modern languages to students' general education and development, and the respondents' teaching practices regarding SFL oral competence (see Erickson et al., 2018 for an overview of the general results).

### *Data collection and achieved teacher sample*

To ensure a representative sample of schools, and also to enable possible analyses between the different studies within the TAL project, it was decided to distribute the TTQ to the 416 schools initially sampled for the project by Statistics Sweden. As already mentioned, the 416 schools were extracted using a stratified random method to ensure a representative group of schools across the country regarding socio-economic as well as geographical inclusion parameters (for further information, see Granfeldt et al., 2019). All in all, 315 responses were received, representing 186 schools, which gives a response rate at the school level of c. 45%. This may seem very low, but according to communication with Statistics Sweden16, this is a slightly higher number than is usually expected in surveys of the current kind. The underlying reasons are obviously multifaceted, but what may be seen as an essential aspect of the relative reluctance to respond, for example, to questionnaires, is the rapidly growing interest in educational research that has brought about a large number of studies requiring active collaboration between teachers/head teachers and researchers. Fortunately, however, the responses received can be compared to population data for teachers of Modern languages, provided by the National Agency for Education17, which gives useful information regarding the representativity of the data.

Table 1 gives an overview of some relevant background statistics of the 315 participating teachers, as reported in the questionnaire.

<sup>16</sup> Statistics Sweden/SCB is a Swedish government administrative authority, which reports to the Ministry of Finance and is responsible for official statistics and other government statistics. Statistics Sweden shall, on behalf of the "Riksdag" (Parliament), provide customers with good quality statistics.

<sup>17</sup> https://www.skolverket.se/skolutveckling/statistik/sok-statistik-om -forskola-skola-och-vuxenutbildning?sok=SokB&omr=Personal&run=1


**Table 1.** Background information TTQ


18 Independent schools in Sweden teach according to the national curricula and are free of charge.

As shown in the table, the teachers responding were fairly evenly distributed across the three target languages in focus, French, German and Spanish. As compared to national statistics of Modern languages for school year 2017/2018, i.e., the year when the questionnaire was administered, this indicates a certain overrepresentation of teachers of French and German, and a corresponding underrepresentation of teachers of Spanish. However, differences are quite small.19 Regarding number of female and male teachers, the proportions in the TTQ group were similar to national statistics for the year in focus, 2017/18.20

85% of the respondents were teaching at a municipal school, which is similar to national statistics for school year 2017/2018.21 As for formal qualifications, however, the group of respondents in the TTQ study reported a considerably higher degree of national teacher certification and teacher education including the target language than the whole group of Modern language teachers, with 65% in the whole group (TTQ: 84%), and 70, 73 and 58% for French, German and Spanish, respectively (TTQ: 89, 85, 78%).22

Analyses of the responses to the questionnaire, based on the different selection criteria used by Statistics Sweden, show reasonable representativity of schools in relation to the 416 schools contacted initially. This means that the TTQ schools do not differ in any systematic way from the full sample regarding geographic, demographic and socio-economic variables. However, it is impor-

<sup>19</sup> https://www.skolverket.se/skolutveckling/statistik/sok-statistik-om -forskola-skola-och-vuxenutbildning?sok=SokC&omrade=Personal&lasar =2017%2F18&run=1;Table 5.A

<sup>20</sup> https://www.skolverket.se/skolutveckling/statistik/sok-statistik-om -forskola-skola-och-vuxenutbildning?sok=SokC&omrade=Personal&lasar =2017%2F18

<sup>21</sup> https://www.ekonomifakta.se/fakta/valfarden-i-privat-regi/skolan-i-privat -regi/elever-i-friskola/

<sup>22</sup> https://siris.skolverket.se/reports/rwservlet?cmdkey=common&geo

<sup>=1&</sup>amp;report=personal\_amne2&p\_flik=G&p\_verksform=11&p\_hman

<sup>=&</sup>amp;p\_niva=S&p\_amne=&P\_VERKSAMHETSAR=2017&P\_KOMMUNKOD

<sup>=&</sup>amp;P\_LANKOD=&p\_skolkod=&p\_hmankod=

tant to bear in mind, that this does not mean that the individual teachers are necessarily representative, for example in terms of education or opinions. Taking part in the survey was optional, and it is highly likely that those who chose to do so were interested in the topic and thus reflected a certain degree of positive selection, which needs to be taken into account when interpreting the responses.

### *Analytical approach*

In this study we present a selection of primary analyses of the TTQ data. The variables described were chosen to illuminate the three issues focused upon in the current chapter: teachers' professional satisfaction, teachers' target language use, and the curricular status of second foreign languages. Thus, this is not a complete account and analysis of the entire survey.

We have chosen to present only standard descriptive statistics, such as frequencies distributions on key questions, and comparisons of means on selected Likert-scale questions. However, our discussion of these response patterns is also supported by analyses of relations between variables expressing perceptions or attitudes and background variables using correlations and p-values.

Selected open responses have been subject to iterative reading by the researchers, targeting explicitly and implicitly expressed perceptions regarding the three issues in focus. These perceptions have also been categorised and validated through independent coding by three researchers, discussed and subsequently agreed upon. In the text, these categories are used in relation to the perceptions expressed on the scales of the attitude items.

A separation has been made between the different target languages, to enable analyses of possible similarities and differences between answers from respondents teaching French, German and Spanish. However, since comparing the three languages was not the main aim of the study, this will be accounted for and commented on only in cases where the results show a noticeable difference relevant to the issue in focus.

# **Results**

In the following, results from the questionnaire regarding the three issues focused upon in the current text will be presented, one at a time, with the numbers of questions actively used in the analyses given within brackets. Special attention will be given to instances where certain ambiguities emerge. Finally, the different results are discussed in the concluding section of the text.

# *Professional satisfaction*

The fourth thematic section in the TTQ focused on frame factors and attitudes and comprised 16 questions altogether. Eight of these questions, together with one from the section on assessment, aimed to capture what may be referred to as professional satisfaction, which to a considerable, albeit individually varying, extent can be connected conceptually as well as practically to different conditions in the working environment, with material as well as relational characteristics (cf. Borg, 2003; Kemmis et al., 2014; Lundgren, 1999).

The questions chosen to capture levels of professional satisfaction among the respondents focused on three aspects that commonly emerge as having an influential role, namely colleagues and collegial cooperation, in-service education, and perceived attitudes to the language teaching profession.

Two open-ended questions focused on number of colleagues (the numbers of the questions in the questionnaire given within square brackets):

*'How many teachers of Modern languages are there at your school?'* [38],

and

*'How many teachers teach the target language at your school?'*  [39]

The responses to these questions, in particular the first one, highlighted a problem regarding terminology, namely the concept of 'Modern languages', which has been used in national curricula since the year 2000 (instead of naming the individual languages) but still causes certain problems of interpretation and use. The respondents' definition was sometimes too wide, including both English and different mother tongues, sometimes too narrow, presumably referring to the target language only. This made some answers difficult to interpret and results should be treated with caution. However, the most common number of Modern language teachers reported, in order of frequency, were 3 *(n=76; 25%)*, 4 *(n=70; 23%)* and 5 teachers of Modern languages *(n=51; 17%)*, followed by 2 *(n=35; 12%)*.

The question about target language, that is, French, German or Spanish, was apparently easier to interpret, with only very few answers indicating a possible misunderstanding of the terminology. As shown in Table 2, the following responses were given.


**Table 2.** Number of teachers teaching the target language at the respondent's school

As can be seen, the largest group of respondents, about 45%, report on being the only teacher of the target language at her/his school, followed by 25% having one colleague teaching the same language. Comments show that lack, or shortage, of colleagues is often considered as negative.

A separation of French, German and Spanish shows clear differences, with 60% of the teachers of French reporting to be the only target language teacher at the school, as compared to 52% for German and 26% for Spanish. Consequently, teachers of Spanish much more often have colleagues teaching the same language. This can obviously be related to the higher number of students/groups for Spanish as compared to the other two languages, but it may also reflect conditions related to organisation and forms of employment in different municipalities and schools.

In a subsequent question [40], the respondents were asked to describe in what ways teachers of Modern languages collaborate at their schools. Here, the problem of terminology mentioned previously was further emphasized: It was often not possible to determine how the question had been interpreted, and a considerable number of teachers apparently used Modern languages for their own target language. English was mentioned in four comments out of 281, and Swedish once; other languages not at all. As for the answers to the question, the responses were quite scattered, both concerning frequency and content. It was quite clear, however, that the most common example of collaboration given concerned assessment and/or grading. This may be compared to the outcome of question 28, where teachers were asked to indicate on a 5-point scale how often they cooperated with colleagues on assessment issues, and where the mean was 3.08, hence in between the alternatives often and seldom.

Another question in the TTQ [41] focused on language related in-service education during the past five years. As shown in Table 3, quite a negative view emerges.

As can be seen, less than a third of the respondents report on recent in-service education related to the target language, and only



one out of ten on such education in a target language speaking country. Furthermore, only 19% feel that they have been able to influence the education themselves. The following two quotations may serve as illustrations to what is often reported23: *'Only simple things in my own town, without/or with very low costs'* and *'Paid for my own five-week course last autumn. Didn't receive any economic support at all and have never had any during my nearly 20 years teaching languages'*. However, there are also positive examples given, for example activities offered by different Swedish universities ('language teacher days'), the Swedish language teachers' union, and language and culture institutes (*Institut Français*, *Goethe Institut* and *Instituto Cervantes*). Some local initiatives are also mentioned: *'Once every academic year, there is usually an opportunity to meet the other teachers of X in the municipality, which is very valuable'*. Regarding language in-service education, the results in the current study coincide to a large extent with similar studies, which we will return to in the concluding discussion.

Three questions in the TTQ focused on the respondents' perceptions regarding external views on *the value of knowing Modern languages*. Here, the views in society [44], at the respondent's own school [45] and among students [46] were asked for in multiple choice questions with five alternative answers in a Likert scale. The responses are summarised in Table 4.

<sup>23</sup> Quotations translated into English by the researchers.


**Table 4.** How do you think the following agents value target language competence?

As can be seen, the respondents did not consider TL competence very highly valued, especially not by society in general or by students. As shown by the mean value, the attitudes at their own schools were deemed a bit more positive, but only to some extent, with a mean value slightly above the 'neutral' middle value 3. Some teacher comments summarize attitudes often expressed by the teachers:


and

– *The subject has low priority but it has become better. Our school management now think that most students should study a modern language instead of the Swedish/English option ....*

What the respondents convey regarding number of colleagues, in-service education and perceived attitudes to their profession as teachers of Modern languages is quite negative. Therefore, it is somewhat surprising, albeit a bit ambiguous, that the answers to the question *'Would you choose to become a teacher of Modern languages today as well'* [43] is answered with Yes by 84% of the respondents, with [only] 16% saying No. However, some of the comments following this dichotomous question shed some light on the complex issues of professional satisfaction. Positive features frequently mentioned were the value of languages, personally as well as globally; creativity; contact with young people; learning by teaching and, very frequently, joy from 'opening doors' and seeing people grow. Negative comments were of course also made, albeit not very frequently. Factors mentioned here included workload, working environment, stress, and lacking motivation among students. Some also emphasised that as long as Modern languages is an optional subject, it will never have the same status as other subjects. We will return to this structural aspect later in the text, after having had a closer look at the pedagogical issue of target language use.

### *Teachers' target language use*

Oral language proficiency was emphasised in the TAL project, and consequently the teacher questionnaire comprised a number of questions about this competence, not least focusing on the frequency and nature of target language use in the classroom, both regarding teachers and students. In the following, the analyses presented focus on teachers' target language use.

A question at the beginning of the questionnaire, targeting the respondents' current language confidence, may serve as an interesting, albeit not wholly compatible, baseline for further questions about target language use. The question was phrased in the following way: *'As compared to when you were a novice teacher of the target language, how confident do you feel in your language use today?'* [10]. Responses were given both on a five-point Likert scale ranging from 'Much more confident' to 'Much less confident' and in (optional) individual comments. Table 5 shows the results that emerged.


**Table 5.** Current target language confidence as compared to being a novice teacher

As can be seen, the large majority of the respondents declared that they feel more confident to use the target language today as compared to when they were novice teachers, the largest group – 45% – even much more confident. It is also worth noticing that nobody chose the alternative 'Much less confident' and only nine individuals the second lowest alternative. Analyses of the 94 individual comments (30% of the respondents) show that close to twenty of those commenting characterise the target language as either their first or second language, or a language used daily in their family. It is also clear that the respondents emphasise the importance of frequent and authentic contact with their target language, usually through visits to countries where the language is spoken or via in-service training in general. However, these options are not depicted as something common. Interestingly, some teachers also point out that they are (much) more confident today. However, this does not necessarily have to do with language but with pedagogy; they know how to teach. Confidence may obviously also be applicable to other aspects of using the language in focus, not only regarding oral competence. Finally, when comparing responses based on the different target languages, the teachers of Spanish report a somewhat higher increase of confidence as compared to being new in the profession, with 80% choosing the two most positive alternatives, as compared to around 70% for French and German.

To capture the respondents' degree of active target language use in the classroom, one question focused on the average amount of time for all school years (usually four, but sometimes three).24 The question was phrased in the following way: *'During how large a proportion of your lesson time do you speak in the target language? Think of all school years together and estimate an average percentage*.*'* [14].25 As shown in Table 6, the question generated the following responses.

The table shows a distribution of intervals of percentages where 41% of the respondents report an average use of the target language of between 26 and 50% of the time, with 31% below and 28% above that range. Following the Likert scale question, the respondents were asked to comment on the interval chosen. Here some examples of situations were given, based on pre-testing experiences: *'Please describe the situation more thoroughly (for example, if it varies for school year, content etc.)*. A large number of comments were given (253 = 80%), generating a fairly clear picture of the TLU issue, and of the opposite, namely use of Swedish, i.e., the national majority language and L1 for most students.


**Table 6.** Average target language use across school years 6/7–9

<sup>24</sup> At the time of the study, most students started their SFL in school year 6, but a substantial number (c. 37%) also in year 7. As from the autumn of 2018, starting in year 6 is mandatory.

<sup>25</sup> The reason for this general wording was to avoid taking it explicitly for granted that there were differences, for example between different school years.

By far the most common of these comments concerned *differences between instruction for students at different levels of competence*. A large number of respondents (n = 110; 43%) described a situation when the first instruction, in school years 6 and 7, is mostly delivered in or via Swedish, and that teachers' use of the target language then increases gradually up to the end of compulsory school. The most common explanation given to this situation concerns students' limited comprehension, often used synonymously with learning, expressed in this way in one of the comments: *'It varies between years. The more [X] they have learnt, the more [X] I speak'*. Another frequent teacher comment (n = 56; 22%) focused on *explanations of grammar*, which were commonly delivered in Swedish. Here as well, Swedish, or rather L1, was described as a prerequisite for understanding and learning. A third, fairly frequent comment (n = 28; 11%) concerned language teaching methodology and pointed to procedures including translation, for example of instructions, between the target language and Swedish, either done by the teacher him/herself or students. It is worth noticing that none of the 253 responses mentioned students with another L1 than Swedish.

Following the question on the 'longitudinal' use of the target language, there was one focusing on situational and content related aspects, the first one being *'How often do you use the target language in the following situations?'* [15] The two situations focused upon were talking to one student or more and talking to the whole group. As shown in Tables 7 and 8, the following responses were given.


**Table 7.** Target language use 'when talking to one or more students'


**Table 8.** Target language use 'when talking to the whole group'

As can be seen, there is a certain difference in frequency between the two situations, not very large, however significant and worth considering. The difference may be easier to detect in an analysis of mean values, which shows 3.37 for TLU with one student or more and 3.82 for TLU with the whole group. Some possible explanations to the discrepancy may be found in the comments following the Likert scales, where the respondents were asked to describe *'concrete situations when you most often use the target language (for example to greet, tell, explain or instruct)'*. The responses here were very similar to those in the preceding question concerning average TL use over time, and reflect a situation where the target language is very often used for purposes of classroom management (greeting, informing, instructing, planning etc.) but also for social small talk and questions about texts. A number of teachers also take the opportunity to repeat and emphasise when the TL is not used: *'I ALWAYS speak in Swedish when it comes to pure grammar'*.

Albeit not the focal point of teachers' target language use, a question focusing on the type of oral proficiency assessed by the teachers may serve as an indication of an attitude to oral SFL language proficiency, where spontaneous language use is not in focus, at least not when evaluating the relatively low level of proficiency that is expected at the end of compulsory school (roughly equivalent to CEFR level A2.1). The question [30], was three-dimensional in its focus, namely asking about teachers' practices regarding oral *situations* assessed, the *constellations* of students, and the students' *partners/audience*. Here, the four highest ranked options all emanated from prepared topics, which means that spontaneous, unprepared topics were not used at all as often, in spite of the fact that oral interaction has a prominent role in the national syllabus for Modern languages.

Finally, it should be mentioned that differences between the three different languages in focus were quite modest regarding reported teachers' target language use, especially concerning situational use, where only very small differences were seen. In the question concerning estimated TLU time across years, there was a slight tendency of a wider distribution of percentages in German and Spanish, with more low and high values, and a certain central tendency for French – 52% choosing TLU for 26–50 per cent of the time, as compared to 33 and 39% for German and Spanish, respectively. With numbers of respondents quite low, 94 for French, 105 for German and 102 for Spanish, interpretations and possible conclusions obviously need to be handled with caution.

### *The curricular status of Modern languages*

One question at the end of the TTQ focuses on a structural issue that has been discussed for a long time in Sweden, namely the curricular status of Modern languages in compulsory school – whether it should be mandatory or, as today, an elective among other languages.

The question focusing on the curricular status of Modern languages in the TTQ was a Yes/No question followed by space for comments. The question was phrased in the following way: *'Do you think Modern languages should be a mandatory subject in compulsory school?* [47], with responses summarised in Table 9.


**Table 9.** Do you think SFL should be a mandatory subject in compulsory school?

As shown, the large majority, 68%, of the respondents answered Yes to the question about mandatory SFL, 28% said No and 4% chose not to answer. As compared to results from earlier surveys, this is a clear increase of those in favour of a second obligatory foreign language in compulsory school. To find out about teachers' comments on their answers, an analysis of the 98 open comments available (31% of the respondents) was undertaken. Three of the researchers independently categorised all comments as positive or negative, or both, without knowing whether they were preceded by a Yes or a No to the main question. Consensus in this part of the analysis was very high. In the next step, different categories of comments within each group were identified, compared, discussed, and eventually agreed upon.

Four arguments were typically found in the positive comments. Learning/knowing an additional foreign language was characterised as:


In this category, the first and the last group were the largest. Examples of comments, one per group, are the following:

– *If you are going to have even the smallest chance of competing about jobs in today's world, you must master more languages than English.*


Aspects identified in the No-answers were the following, often expressing concerns about the students. An additional foreign language was considered negative for


Overlaps between the aspects were common, which complicated the grouping of comments to some extent. However, the largest group of comments saying No to mandatory SFL was the one focusing on the demands of Pass grades in English and Swedish, followed by a mixed category encompassing issues of learning difficulties coupled with lack of motivation.

Examples of No-comments, one per category, are the following:


Some types of comments occurred in different contexts, sometimes following both a Yes and a No to the initial question, which made them difficult to fit into the identified categories. This is also the reason why a detailed account of numbers for the different aspects would be neither quite possible to define, nor meaningful in the interpretation of the results. However, this does not mean that these comments are less relevant; on the contrary, they are well worth mentioning to make the picture of the situation clearer and more complete. In particular four different aspects were mentioned, namely individual student features related to *maturity* rather than to aptitude or motivation; the need for *adaptation*  and *individualisation* of instruction as well as resources for *special support* for individual students ('as in other subjects', some respondents pointed out). In addition, a widening of electives was mentioned, both regarding languages and other subjects, for example, both Japanese and more practical subjects, crafts in particular. One teacher also mentions the issue of students' different mother tongues that are sometimes studied and graded as Modern languages (an option mentioned in the curriculum).

Finally, it needs to be pointed out that the analyses of comments revealed a certain degree of ambiguity in the Yes responses, namely that, strictly speaking, about a fifth of them were not totally positive to mandatory SFLs, but conditional in their message, that is naming the exceptions that should be made to the obligatory status. These exceptions often concerned students with learning difficulties, general and/or regarding English and Swedish, as well as newly arrived migrant students. The following comments may serve as examples:

*– To achieve the goals of the EU with two languages besides one's mother tongue. An exception, though, for those students who have a special reason not to be able to study the language, certain difficulties for example.*

*– Not for those who have only just arrived or have enormous problems with English.*

As for the three language groups, that is the teachers of French, German and Spanish, there were no large differences in responses. However, teachers of German were slightly less positive to mandatory SFL than their colleagues in French and Spanish, with 75% Yes responses for French, 66% for German, and 73% for Spanish.

# **Discussion**

The 315 responses from teachers across the country generate a rich and diverse picture of practices and perceptions characterizing second foreign language education, and they also clearly illustrate the contextuality and complexity of the teaching profession at different levels (cf. Borg, 2003; Shulman, 1986). As shown in the analyses, what seems to be certain interesting ambiguities emerge in the responses to and comments on the three issues focused upon. In the following, these issues will be briefly discussed under the headings of the three levels that they represent: the individual, pedagogical and structural levels.

### *The individual level*

At the individual level, the respondents' professional satisfaction was looked into based on questions focusing on three aspects of well-being and self-esteem in relation to being a SFL teacher: having colleagues, access to in-service education, and external appreciation of the value of learning, teaching and knowing a SFL, indirectly, that is, the professional choice of the respondents. Taken together, the picture conveyed is not a very positive one, with reported lack, or shortage, of colleagues, especially those teaching the same target language, weak provision of in-service education, and what is felt to be a generally lukewarm interest in SFLs as such (cf. contextual and relational factors described, for example, by Borg, 2003, Lundgren, 1999, and Kemmis et al., 2014). In spite of this, however, the large majority of the respondents (84%) answer yes to a question if they would choose to become teachers again. Whether this reflects a profound satisfaction with the profession, or a feeling that there are not many alternatives for people with an education in languages is of course not possible to establish. It is noteworthy, though, that teachers in the open comments to several questions speak very positively about the teaching profession, in particular focusing on the aspect of working with young people in the process of learning and development. Based on the ambiguity that lies within the discrepancy between the reported examples of negative experiences of the profession, and the strong expression of liking of the same profession, we will now briefly discuss the different ambiguities that emerge in the data.

It is sometimes claimed that the teaching profession can be very lonely with lack of regular peer communication and support. This was one of the reasons to include questions on colleagues in Modern languages and the specific target language. Furthermore, adequate professional development as well as the feeling of respect for languages and language education are essential. It seems clear that a majority of the respondents were the only teachers of the specific TL at their schools, which was sometimes mentioned as a reason for weak or no collaboration between colleagues. Here, a widening of the concept of 'colleague' seems essential, especially given the fact that all Modern languages in the Swedish school system (except Chinese) have the same national syllabus and the English syllabus is almost identical, albeit operating at different proficiency levels for students of the same age. Also, professional collaboration with teachers of Swedish L1 and L2, as well as Mother tongue, would enable the chance of strengthening the language education context with mutual benefits for the different teacher groups and, in the long run, also for students. Regarding the very weak provision of in-service education, the results from the current study coincide with similar studies, at both national and international levels (e.g., Bonnet, 2004; Lärarnas Riksförbund, 2016). To interpret this, several aspects have to be considered, for example the decentralised organisation of schooling in Sweden, which makes decisions regarding teachers' professional development quite variable. It is worth noticing, however, that efforts to facilitate and promote collegial learning are increasing, with the aim of enhancing professional development

by combining learning with collegial collaboration (Timperley, 2011). Work of this kind has grown considerably, in Sweden supported by major national initiatives and investments, for example the program "Språksprånget"26 ('the Language leap') offering online, in-service materials and modules for Modern languages.

### *The pedagogical level*

At the pedagogical level, the responses to the questions focusing on teachers' target language use is another example of potential ambiguity, most of all with the discrepancy between teachers' reported increase of confidence in their target language proficiency and their not too frequent use of the language in the classroom. The tradition of teaching a foreign language through, or with the consistent support of, the L1, or rather the majority language of the country, goes back in history, with its roots in the teaching of classical languages (Littlewood & Yu, 2011), and is sometimes characterised as teaching *about* the target language rather than *in* the language. What seems to be very firm beliefs regarding this practice is expressed in the TTQ responses, especially concerning the importance of adapting the amount of input to the age and proficiency level of the students, that is speaking much more Swedish in the early years, and, even more clearly, to adapt one's SFL use to content, especially regarding grammar, which is reported to be taught through Swedish almost exclusively. Certain priorities made in the assessment of students' oral language competence also strengthen the impression that the functional and action-oriented approach to language proficiency (cf. Council of Europe, 2001; Skolverket, 2011) may not be altogether embraced.

TLU is an issue that has been, and is, being focused upon in research as well as in public debates about language teaching, with strong opinions on both sides. According to Thue Vold and Brkan (2020), the current status in research is one of "judicious use of L1" (Shin et al., 2019), with contextual adaptations needed, however with maximised exposure to the TL "because it is essential for the development of communicative language abilities"

<sup>26</sup> https://www.skolverket.se/skolutveckling/kurser-och-utbildningar /sprakspranget---kompetensutveckling-for--larare-i-moderna-sprak

(p. 2), especially in FL and SFL contexts, where out-of-school exposure to the language is limited.

Two reflections can be made in relation to the questionnaire responses to the issue of Target language use, the first being that Swedish is most often seen as a necessary means to understand both in general and, in particular, regarding formal aspects of the language. None of the 253 responses to the TLU question include any reference to the fact that well above 20 percent of the students in Swedish compulsory school do not have Swedish as their L127 (Skolverket, 2020b; no data available on the proportion in SFL groups). Consequently, it cannot be taken for granted that these students are helped by references to Swedish. If references to individual students' L1s are considered necessary, it requires that teachers have a solid 'multicontrastive' knowledge and awareness as pointed out by Ohlander (1988; see also Erickson, 1990, and Tornberg, 2020); otherwise some students are favoured in a way that may be considered problematic in an inclusive 'school for all', a concept introduced in connection with the launching of a national curriculum for Swedish compulsory school in 1980, Lgr 8028 (Skolöverstyrelsen, 1980).

Secondly, the attitudes expressed regarding TLU often reflect a traditional view of language learning and teaching that can probably, at least to some extent, be related to the poor provision of in-service education. Consequently, there seems to be plenty of room for improvement, both regarding more theoretical and conceptual aspects of language and language education, and practical examples of teaching, implementing an action-oriented approach and promoting communicative language competence. In this, collegial collaboration seems to be one of the positive and constructive ways to move forward. Returning to the issue of professional satisfaction and the widening of the concept of colleagues discussed, collaboration between (S)FL teachers and teachers of Swedish as a second language may be an interesting path to explore, for example regarding the use of the target language. With groups that are usually very linguistically mixed, teachers of Swedish as a second language normally have to stick to the target language as the only

<sup>27</sup> https://siris.skolverket.se/reports/rwservlet?cmdkey=common&geo

<sup>=1&</sup>amp;report=gr\_elever&p\_sub=1&p\_ar=2019&p\_lankod=&p\_kommunkod =&p\_skolkod=&p\_hmantyp=&p\_hmankod=&p\_flik=G

<sup>28</sup> Lgr = "Läroplan för grundskolan" (Curriculum for compulsory school)

common denominator, and this may be of interest to discuss in relation also to other language subjects.

### *The structural level*

One question, in particular, focused on the structural level, namely national regulations regarding the curricular status of second foreign languages (Dahllöf, 1967; Tholin & Lindqvist, 2009). Here, a gradual change of attitudes has taken place over time, from a firm SFL teachers' No at the end of the last century to the inclusion of a SFL in the group of mandatory subjects in compulsory school (Skolöverstyrelsen, 1991), to a clear Yes expressed in the TTQ. However, analyses of the comments to the question reveal a certain degree of ambiguity in this case as well, showing that in around a fifth of the responses saying Yes to mandatory SFL, there were exceptions defined, often using the same arguments and expressing the same concerns that were put forward by those negative to a change: newly arrived migrant students were mentioned as were learners struggling with Swedish and English, and those with 'different difficulties'. Adjusting the numbers for these conditional Yes responses only changes the outcome of the question marginally, but the ambiguity as such seems of interest, since it reflects a view of certain subjects, in this case SFLs, being less possible to master for all students, whereas others, for example Physics, Geography and Arts, are not questioned, at least not explicitly. Putting it differently, the question could be asked, whether a second foreign language is seen as a subject for all or just for some of the students in the Swedish, distinctly inclusive compulsory school. A considerable number of respondents also highlight this aspect in their comments and problematise the message given by the fact that Modern languages are optional, wondering whether this contributes to the perceived low status of the subject in society, at their schools and among students.

Another aspect of clear interest is that the respondents were not equally eager to comment on their answers regarding the curricular status of Modern languages; 51% of the comments submitted were made by respondents having chosen the No alternative to mandatory SFL, although this group represented less than a third of the total number of respondents. There may be several reasons for this, maybe one being that this group felt the need to explain a negative answer, and perhaps an opinion perceived as less in line with contemporary discussions regarding inclusion and 'a school for all' (Skolöverstyrelsen, 1980).

An aspect not commented on in the responses is the fact that Modern languages are part of the 'Language choice' in the curriculum. This means that choosing one of the options in this group is mandatory but that none of the alternative subjects are individually compulsory. This is a question that has been discussed for a long time (see *Second foreign languages in Sweden* at the beginning of the chapter) and where different modifications have been suggested, the most frequent one being to abolish additional English and Swedish. However, what is seldom talked about, even less discussed, is the fact that Mother tongue tuition is included in this group. This means that a large number of students with another L1 than Swedish may have to choose between a SFL and their first language. This is an unfortunate situation that will need serious discussions at the structural level.

# **Concluding remarks**

In conclusion, the responses to the TAL teacher questionnaire generated a large amount of interesting information, including a number of ambiguities that may serve as food for thought and inspiration for continued discussion and development of (second) foreign language education and teacher training. The material has obviously only been partially possible to report on and discuss in the current chapter. However, a few very general points can be made at this stage, first that the study is an example of successful teacher-researcher collaboration, both regarding the development of the questionnaire and in teachers taking time to respond in ways that often go far beyond expectation regarding willingness to share experiences and reflections. Perhaps the most essential comment to make, though, is that, in spite of a number of problems made clear regarding different aspects of SFL education at the individual, pedagogical and structural levels, the responses convey genuine commitment among the respondents regarding their profession and, not least, the learning and development of their students. This may be illustrated by the following quotation from one of the teachers in the study in response to the question why s/he chose to become a teacher of SFL:

*'Because language is such a fantastic tool, unlocking communication and relations between people. If I can help a single student to meet a new situation, person, or a new context thanks to her/ his language, then I have succeeded!"*<sup>29</sup>

# **References**


<sup>29</sup> Swe: "För att språk är ett sådant fantastiskt verktyg, som låser upp kommunikation och relationer mellan människor. Om jag kan hjälpa en enda elev att möta en ny situation, människa eller ett nytt sammanhang tack vare språket så har jag lyckats!"


*English in lower secondary school – A close examination]*. Högskolan i Borås.


# **Appendix 1**

## ITALICIZED QUESTIONS = QUESTIONS USED IN THE TEXT *(2, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 28, 30, 38, 39, 40, 41, 43, 44, 45, 46, 47)*

Teacher survey on learning, teaching and assessment of Modern Languages

A Background information. Here are some questions about your school and your educational background.

**1. The name of the school where you teach, and the municipality:**

### *2. Which of the following languages are you teaching at present?*

French Spanish German

We would like you to respond to this questionnaire based on the language that you are most experienced in teaching. In the following, it will be referred to as the "target language". Please, fill in the language that you choose:

**3. Do you teach Modern languages in more than one school?**

Yes No

If yes in question 3, please answer according to the school where you do your main teaching.

# *4. Fill in the kind of school where you work:*

Municipal Independent

Other

If other, please note which:

**5. Which years/ school forms are there at your school? (Several marks may be needed.)**

Preschool-Year 5

Year 6

Years 7–9

Upper secondary preparatory class

Special school

*6. Do you have a national teacher certification?*


No

If yes, please note the combination of school subjects and type of school /years:

# *7. Describe your educational background. Mark the alternative that is most applicable:*


Type of teaching degree (also foreign teaching degree) and year of graduation:

Please fill in the combination of school subjects and type of school/ years part of your teaching degree:

If ongoing teaching education, describe how far you have come:

Indicate the combination of school subjects and the type of school years you aim for:

Have you completed your degree with studies in an additional language? (e.g., the target language or another modern language)?


No

Note which language, number of credits (new or old credits):

Indicate the total number of credits that you have obtained in the target language and if it refers to new or old credits.

Are you planning to study in order to acquire certification in the target language?

Yes No

How are you planning to complete your education (indicate school form and how many credits you are going to study/have left to study):

Are you planning to acquire teacher certification through supplementary teacher training (KPU)?

Yes

No

Indicate school form you are planning for, and when you have planned to study:

Other educational background, please indicate below:

*8. Have you spent a longer period of time in the area where the target language is spoken?*

Yes No

If yes, indicate how long (in months):

*9. Is the target language your first language?*

*10. As compared to when you were a novice teacher of the target language, how confident do you feel in your language use today?*


*11. How long have you been teaching the target language?*

12. Year of birth:

	- Female
		- Male
		- Other gender id
		- Don't want to say

## B Learning and teaching

Here are some questions regarding learning and teaching in the target language.

*14. During how large a proportion of your lesson time do you speak in the target language? Think of all school years together and estimate an average percentage.*


Give a more detailed description of the situation (e.g., if it varies with school year, content or something else)

*15. How often do you use the target language in the following situations? (Mark one of the boxes in a scale from "Very often" to "Very seldom". Proceed in the same way with the questions with similar scales).*

Indicate specific situations when you mostly use the target language (e.g., to greet, tell, explain or give instructions):

### **16. How often do the students do the following in the target language (working methods)?**


**17. How often do your students use the target language in the following situations?**


### **18. Is the national syllabus for Modern languages a living document to you in your teaching?**

# **19. How often do you use the syllabus for the following?**


**20. Is the Common European Framework of Reference for Languages (CEFR) a living document to you?**

### **21. How often do you and your students use the following resources (please note that there are some overlaps)?**


Comments about the above resources or else (e.g., paper or digital versions of the resources mentioned above):

### **22. How easy do you consider it to find materials which work well for development of the following in the target language?**


### **23. How easy do you think it is for your students to develop the following in the target language?**


**24. How often do you focus on the following in your target language teaching? Take all school years into consideration***.*



# C Assessment

Here are some questions about assessment and grading in the target language.

### **25. How do you find assessment in the target language?**

5 Uncomplicated


1 Very complicated

### Comments:

### **26. How do you find grading in the target language?**

Comments:

**27. What degree of support do you feel that you have of the following when assessing and grading students' knowledge in the target language?**



*28. How often do you collaborate with colleagues in assessing in the target language?*


Examples of and comments on collaboration:

**29. How often do you use the following for assessment in the target language?**



*30. How often do you base your assessment of students' oral skills in the target language on the following situations?*

Comments on situations, content, recordings etc.:

**31. How easy do you think it is for your students in school year 9 to achieve the course requirements for oral proficiency regarding the following?**



**32. How often do you give feedback on the following aspects of the students' language skills?**


**33. How important do you consider the following when you award final grades in the target language in year 9?**


**34. Do you use the support materials for assessment (tests) from the National Agency of Education (NAE) when awarding grades in year 9?**


Comment:

Any other comments on assessment and grading in the target language, or in general:

D Frame factors

Finally, here are some questions on the conditions for teaching the target language.

**35. Approximately, how many students are there normally in a teaching group in the target language? As a minimum? As a maximum?**

Is there any difference between the school years or the modern languages regarding the number of students in the teaching groups?

**36. Are there mixed groups regarding school year in the teaching of Modern languages?**

Yes No

Comment:

**37. Is the school where you teach a so-called 1–1 school (meaning that every student has a laptop, computer or tablet)?**

Comment:

*38. How many teachers of Modern languages are there at your school?*

*39. How many teachers teach the target language at your school?*

*40. In what way do the teachers of Modern languages collaborate at your school?*

*41. During the past five years, have you taken part in in-service training related to the target language?*


If yes, describe in detail (content of the education and country where the course was held):

**42. Why did you choose to become a teacher of Modern languages?**

*43. Would you choose to become a TL teacher today as well? Please give your reasons.*

Yes No

Motive(s):

*44. How do you think society values TL competence?* 


Comments:

*45. How do you think your school values TL competence??*

5 Very highly

4 3

2 1 Very low

Comments:

*46. How do you think your pupils value TL competence?*

5 Very highly 4 3 2 1 Very low

Comments:

*47. Do you think Modern languages should be a mandatory subject in compulsory school?*

Yes No

Comments:

**48. In what ways do you think that Modern languages contribute to students' general education and development?**

**49. We would be very grateful if you could describe and further comment on how you work with oral proficiency in the target language, e.g., when it comes to planning, materials, work inside and outside the classroom, assessment etc. By answering, you will**  **contribute importantly to an area that still lacks systemized knowledge.** 

**50. Please give your opinion on the content and design of the questionnaire:**

**Many thanks for your participation!**

# **Afterword**

There is an intense debate in Sweden as well as in many other countries in the world about the significance and role of foreign language learning and teaching. This discussion takes place in the current context of societies where international communication is a natural part of many citizens' everyday life. Virtual encounters in foreign languages are frequent both at people's work and in their free time. A virtual reality is what people are used to in meetings with the help of digital tools. This phenomenon requires skills in foreign languages. For efficient communication across the countries in Europe, citizens need to be able to communicate in at least one foreign language. The English language often takes this role and in communication between people it is to a great extent used as a lingua franca. This puts new demands on material and methods used in the foreign language classroom. We need to think about how to teach foreign languages in order to prepare young people for this lingua franca communication. The focus in that context is on effective communication and interaction. At the same time, learners' have expectations on more student-active approaches in the language learning classrooms, as well as the fact that the learners' own experiences of the language outside school, so called extra-mural English, is considered, and the language teacher is to meet these demands.

The contributing authors in this book all provide useful and relevant insights into various aspects of research on the teaching and learning of foreign languages as well as findings related to what teachers need in the new language learning and teaching situation described above. This research field is vast and comprises for instance both studies of specific languages in use in the classroom, and interdisciplinary approaches and investigations. Consequently, this volume with its contributions by participants from the Exploring Language Education conference at Stockholm University in 2018 reflects this diversity and creativity found among language teachers and researchers all over the world. For language teachers at schools, the book provides highly relevant areas of interest and the challenge is to find the concrete application of the findings in classroom settings.

What is inspiring, after having read the contributions, is that even though the focus is on Sweden and the United States, the teaching of foreign languages is international in itself so that local practitioners and researchers all over the world can find something of interest in the conceptual or the empirical presentations, although some are positioned in a local context.

The relationships between broad concepts in the field of the didactics of foreign languages are relevant to discuss and possibly to clarify as done by Nina Spada and John Levis. The fact that widely spread and used concepts in language education can be differently used in various cultural settings makes it important to discuss and define them in order for teachers all over the world to fully understand each other in discussions and comparisons of methods, material and curricula. Since teachers are to be updated and well acquainted not only with documented experiences and conclusions by fellow teachers from classrooms but also with findings by researchers, this book is extremely useful and valuable for foreign language teachers. It is a strength of the present volume that it can bring together researchers and practitioners in fruitful discussions within the field of second or foreign language learning and teaching.

Furthermore, there is no doubt about the fact that this book has parts that are relevant in the field of language teacher training. Naturally, both the introducing chapters which treat language education, second language acquisition and linguistics, and the three chapters presenting findings from empirical studies contribute in a sophisticated manner to the understanding of the complexity of language learning and teaching.

What is presently emphasized more and more is that the foreign language teacher is regarded as both a practitioner in the classroom and a researcher. Therefore, it is a pleasure to see that the five chapters in the book address issues that are really in focus in the discussions among foreign language teachers, language teacher students and teacher educators. One of these issues is related to oral production and working with our attitudes to foreign accents and our perceptions of native versus non-native production in the classroom and whether to apply a more

normative or descriptive approach. Another issue is translanguaging which currently is a phenomenon discussed both from a descriptive point of view and a teaching strategy in the classroom for successful communication. The last issue to mention is on the one hand the dominance and special role of the English language in the field of language learning, and, on the other hand, the emergence of the awareness of the numerous other languages that are used in a multilingual and multicultural society which definitely have significant roles to play in communication today, and in language teaching and learning in the future. In this context Nina Spada's discussion of second language acquisition and L2 pedagogy is extremely interesting.

The fact that a chapter is dedicated to oral proficiency and the communicative aspects on teaching and learning pronunciation is really relevant for the current situation in societies across the world. Today, the research field of English as a lingua franca makes more and more impressions on practitioners and as a consequence, dialects and accents in relation to identity are concepts that are given more attention and prominence in the learning of a foreign language. The language teacher now faces new challenges in the learning of oral production and interaction. It is interesting to see that in the learning of oral proficiency we do not only consider pronunciation but see several more aspects, such as pragmatic and situational competence. At the same time, the traditional question of whether using the target language only or not in the classroom is treated. In the context of the multilingual classroom where there are several mother tongues represented among the learners, there is an interest in turning to a more authentic-like situation when all the linguistic resources available are expected to be in use for effective communication. In this context it is therefore valuable that the chapter by Amanda Brown, Robert James Lally and Laura Lisnyczyj deals with classes with students with different L1, differences in cultural backgrounds and with multilingual backgrounds. It goes without saying that students' attitudes to classroom language practices is a highly complex area and that we need more studies in this dynamic and innovative field of research.

Young learners' learning of English is in the same way a slightly new research area, at least in Sweden. It is consequently important that this area is represented in the text by Liss Kerstin Sylvén. This chapter deals with the significance of the exposure of English outside school, that is, extra-mural English. We may expect an increasing interest in this area of research during the coming years, especially when it comes to the acquisition of vocabulary and listening comprehension among young learners. For primary school teachers, these concrete findings are of great interest for understanding when, where and how the learning of English takes place for the very young ones.

English has a very strong position as a foreign language in Sweden, as is the case in many other countries in the world. In Sweden it is sometimes even questioned if the English language should be regarded as a foreign or a second language. Other foreign languages such as French, German and Spanish currently have a totally different situation. The last chapter clearly shows these different, and sometimes even problematic and challenging, conditions for these three languages at Swedish schools. The authors highlight central concepts in language learning such as the learners' motivation and attitudes as important factors to consider in order to come to terms with the current situation. This is definitely an area worth looking into if we are going to reach the goal in the European Union that all citizens are to master two more languages besides their mother tongue.

With this first volume from the first ELE conference, an important step is taken by providing a selection of current thematic areas in research where fundamental questions are raised. We all look forward to future ELE conferences when hopefully, new perspectives from all parts of the world together with both global and local perspectives on areas which are not discussed in this volume will be covered. These areas are, for instance, the didactics of literature in the foreign language classroom and young learners' acquisition of language skills in extra-mural English.

> Stellan Sundh Senior Lecturer/Associate Professor Department of Education Uppsala University

# **Author biographies**

**Camilla Bardel** is Professor of Modern languages and Language Education at Stockholm University, where she has previously served as Professor of Italian. Her main research interests are multilingualism, second and third language teaching and learning, cross-linguistic influence, and language policy. She has published articles, books and book chapters in the field of language education, advanced second language development and third language learning, and she currently coordinates the graduate school *Learning, teaching and assessment of Swedish and English. Multilingualism as an asset and a challenge – PhD program for teacher educators*.

ORCID: https://orcid.org/0000-0001-5615-3688

**Amanda Brown** is an associate professor of Linguistics and the Director of the Linguistic Studies Program at Syracuse University. She investigates the psycholinguistic effects of multicompetence and associated implications for language pedagogy, assessment, and teacher preparation. She teaches courses in second language teaching/learning and holds additional administrative positions in graduate programs in linguistics and language teaching. ORCID: https://orcid.org/0000-0001-7199-9323

**Gudrun Erickson** is Professor of Education in Language and Assessment at the University of Gothenburg. Her background is in language teaching and teacher education, with long experience of national and international projects focussing on assessment of language competences. This includes work on language curricula and quality frameworks for different educational levels, and scientific responsibility for the development of national tests and assessment materials for foreign languages in Sweden. Erickson was President of EALTA 2013-2016 and received the British Council International Assessment Award for 2020. She has published widely, nationally and internationally, on issues of language pedagogy. ORCID: https://orcid.org/0000-0002-1145-6164

**Christina Hedman** got her PhD from the Centre for Research on Bilingualism, Stockholm University in 2009, and she holds a position as full Professor in Swedish as a second language at the Department of Teaching and Learning, Stockholm University, Sweden. Her research encompasses multilingualism in an educational perspective, as well as in ideological and critical policy perspectives. In recent research, a focus has been on the teaching of *Swedish as a second language* in various educational settings, including a collaborative project on *Critical Multilingual Language Awareness* among migrant adolescents. ORCID: http://orcid.org/0000-0001-8869-6687

**Robert J. Lally** is an assistant teaching professor of English to speakers of other languages at Syracuse University. He has taught general, business, and academic English in Colombia, Ecuador, Venezuela, China, and Japan. He currently teaches academic writing and discipline-specific and professional English language courses at the undergraduate and graduate levels. He is interested in multicompetence and its implications for language pedagogy. ORCID: https://orcid.org/0000-0002-4687-7858

**John Levis** (PhD, University of Illinois, 1996) is Professor of Applied Linguistics and TESL at Iowa State University. He has published articles in many top journals, is the author of *Intelligibility, Oral Communication, and the Teaching of Pronunciation*  (Cambridge University Press), and is co-author of *Second Language Pronunciation: Bridging the Gap Between Research and Teaching* (Wiley Blackwell)*.* He initiated the Pronunciation in Second Language Learning and Teaching Conference in 2009, has been the editor of most of its conference proceedings, and is founding and current editor of the *Journal of Second Language Pronunciation*. His primary interests are pronunciation teaching, suprasegmentals, and speech intelligibility.

ORCID: https://orcid.org/0000-0001-7405-5969

**Laura Lisnyczyj** is an assistant teaching professor of English to speakers of other languages at Syracuse University. She has taught numerous courses at the university level since 2002 including research and writing skills, and cross-disciplinary professional communication skills for teaching assistants, graduate students, and undergraduate students. Her research interests include heritage language acquisition and multicompetence in ESOL instruction.

ORCID: https://orcid.org/0000-0003-3454-2441

**Katarina Rejman**, PhD, is a senior lecturer at the Department of Teaching and Learning at Stockholm University. Her main research interests are Swedish as a first language and literature in education but also include literature, reading and writing in Swedish as a second language. Rejman teaches courses in the undergraduate teacher program as well as in the subject teacher program and she supervises teacher students' degree projects in various areas. ORCID: https://orcid.org/0000-0002-0778-6777

**Monica Rosén** is Professor of Education at the University of Gothenburg. Her fields of expertise encompass educational assessment, measurement and quantitative research methods. She has been involved in both national and international projects on assessment of students' educational outcome, in which her own research, over and above the methodological parts, has covered reading achievement from pedagogical and equity viewpoints. She is currently running several research schools focusing on educational assessment and quantitative research methodology and since 2019 she is the Link convenor for Network 9 Assessment, Evaluation, Testing and Measurement within the 'European Educational Research Association' (EERA). ORCID: https://orcid.org/0000-0003-3128-7895

**Nina Spada** is Professor Emerita in the Language and Literacies Education program at OISE, University of Toronto. Her research has investigated the effects of different types of instruction on the L2 development of children and adults in second and foreign language classrooms. Dr. Spada is committed to making research findings accessible and relevant to teachers. She is co-author of *How Languages are Learned* published by Oxford University Press and co-editor of the *Language Learning and Language Teaching* book series published by John Benjamins. Dr. Spada is Past President of the American Association for Applied Linguistics. https://www.oise.utoronto.ca/ctl/Faculty\_Profiles/1475/Nina \_Spada.html

**Stellan Sundh** is an Associate Professor of Language Didactics and Director of Studies at the Department of Education, Uppsala University, Sweden. His main duties are within teacher education and in courses at advanced level. He is engaged in several projects with international cooperation and with a special interest in non-native use of the English language and intercultural communication. His research areas cover didactic issues such as student-interactive video-conferences and linguistics with a focus on young learners' uses of English as a foreign language. ORCID: https://orcid.org/0000-0003-1940-3865

**Liss Kerstin Sylvén** is Professor of Language Education at the University of Gothenburg. She obtained her PhD in English linguistics, and her research interests include computer assisted language learning, content and language integrated learning (CLIL), second language vocabulary acquisition, motivation, individual differences and extramural English. Sylvén has co-authored a book about extramural English (Palgrave MacMillan, 2016), and edited a volume about CLIL (Multilingual Matters, 2019). She has published in, among others, *CALICO Journal*, *the International Journal of Bilingual Education and Bilingualism*, *Journal of Immersion and Content-Based Education*, *Studies in Second Language Learning and Teaching*, and *ReCALL*. ORCID: https://orcid.org/0000-0002-1655-5639

**Elisabeth Zetterholm**, PhD in Phonetics, is a Senior Associate Professor in Swedish as a Second Language at Linköping University, Sweden. Her main research interests are the teaching of pronunciation in Swedish as a second language, multilingual young and adult students writing and literacy development as

well as mother tongue teachers' education in Sweden. Zetterholm has published her research in journals, book chapters and textbooks in Swedish and English. She is also one of the initiators of the Nordic Network for L2 Pronunciation.

ORCID: https://orcid.org/0000-0002-7688-4696

**Rakel Österberg** holds a PhD in Spanish and is specialized in second language acquisition. She is a senior lecturer at the Department of Romance Studies and Classics at Stockholm University, where she teaches grammar, mediation and linguistics and supervises and teaches in the teacher training program. Her research interests cover second language acquisition, discourse-syntax, Spanish as a FL in the school context, academic writing, Spanish as a heritage language and motivation.

ORCID: https://orcid.org/0000-0002-6125-4755

The overarching aim of this book is to offer insight into some of the currently discussed issues at the Swedish as well as the international research frontline of Language Education in a selection of up-to-date work. The volume deals with the gap between research and practice, the importance of pronunciation in second/foreign language teaching and learning, classroom translanguaging, young learners' exposure of English outside school, and teachers' attitudes to foreign languages other than English. A further aim is to provide readers with input from research within the interconnected disciplines of Applied Linguistics, Language Education and Second Language Acquisition targeting researchers, school teachers, teacher students, teacher educators and policy-makers.

The content is based on a selection of talks presented at the 1st ELE Conference ('Exploring Language Education'), which was held at Stockholm University in 2018. Employing a broad thematic scope, the volume reflects the variety of perspectives on Language Education, brought together at the conference by authors working in diverse areas of the field and in different parts of the world.

With the first ELE conference the organizers wished to call attention to the intersection of the global and the local, in terms of linguistic and cultural diversity informing and conditioning both research questions and language education practices. The chapters represent different parts of the broad array of research directions that can be discerned under the large umbrella of Language Education, zooming in on the Western context, specifically Sweden, Canada and the United States.

Bardel et al. (Eds.)

Exploring Language Education

LLE 2

Exploring Language Education

Katarina Rejman & Elisabeth Zetterholm (Eds.)

Language and Literature in Education 2

Global and Local Perspectives

Camilla Bardel, Christina Hedman,