# Multilingualism and third language acquisition

## Learning and teaching trends

Edited by

Jorge Pinto Nélia Alexandre

Contact and Multilingualism 2

### Contact and Multilingualism

Editors: Isabelle Léglise (CNRS SeDyL), Stefano Manfredi (CNRS SeDyL)

In this series:


ISSN (print): 27008541 ISSN (electronic): 2700855X

# Multilingualism and third language acquisition

Learning and teaching trends

Edited by

Jorge Pinto Nélia Alexandre

Pinto, Jorge & Nélia Alexandre (eds.). 2021. Multilingualism and third language acquisition: Learning and teaching trends (Contact and Multilingualism 2). Berlin: Language Science Press.

This title can be downloaded at: http://langsci-press.org/catalog/book/290 © 2021, the authors Published under the Creative Commons Attribution 4.0 Licence (CC BY 4.0): http://creativecommons.org/licenses/by/4.0/ ISBN: 978-3-96110-296-9 (Digital) 978-3-96110-297-6 (Hardcover)

ISSN (print): 2700-8541 ISSN (electronic): 2700-855X DOI: 10.5281/zenodo.4449726 Source code available from www.github.com/langsci/290 Collaborative reading: paperhive.org/documents/remote?type=langsci&id=290

Cover and concept of design: Ulrike Harbort Typesetting: Marten Stelling, Sebastian Nordhoff Proofreading: Amir Ghorbanpour, Bojana Bašaragin, Dora Uštulica, Eliane Lorenz, Emma Vanden Wyngaerd, Havenol Schrenk, Jean Nitzke, Jeroen van de Weijer, Leonie Twente, Lotta Aunio, Marten Stelling, Tom Bossuyt Fonts: Libertinus, Arimo, DejaVu Sans Mono Typesetting software: XƎLATEX

Language Science Press xHain Grünberger Str. 16 10243 Berlin, Germany langsci-press.org

Storage and cataloguing done by FU Berlin

## **Contents**


### Contents

### **III Language learning**


## **Acknowledgments**

We acknowledge everyone who helped with this project, especially all the contributors, who have kindly cooperated throughout the process of editing the volume. Special thanks to Sebastian Nordhoff at Language Science Press for his help with all the details of producing the final manuscript and to the Series Editors of Contact and Multilingualism, Isabelle Léglise and Stefano Manfredi, whose advice has been crucial. We are indebted for the financial support granted by the Portuguese Science Foundation (FCT), through grant UIDB/00214/2020.

Finally, we thank the following reviewers, who provided insightful suggestions to the chapters submitted to this volume:


## **Preface**

Jorge Pinto Centro de Linguística da Universidade de Lisboa

### Nélia Alexandre

Centro de Linguística da Universidade de Lisboa

In the last three decades, there have been several approaches, models and theories that have developed around the acquisition of a second language. Theories based on years of research in a wide variety of fields, including linguistics, psychology, sociology, anthropology and psycholinguistics (Freeman & Freeman 2001). However, none can cover all the needs inherent to the teaching/learning process (Cook 2001), nor has it been possible to arrive at a unified or comprehensive view of how second languages are learned (Mitchel & Myles 2004; Nunan 2001). This complexity is, among others, due to the fact that there are variations in the context where the acquisition processes occur that influence the nature of the input as well as the learning strategies used by the student, and due to biological variations of students, such as age, aptitude and intelligence, motivation, personality, and cognitive styles (Ellis 1989).

The acquisition of (several) second languages has become a subject even more complex with globalization, the growing learning of foreign languages and the increasing number of multilingual speakers. For that reason, since the beginning of this century, there has been a growth in interest in multilingualism and, consequently, a proliferation of studies on the acquisition of a third language or additional language (L3/L*n*), highlighting the differences with respect to the acquisition of an L2 and setting themselves a new area of research (Jessner 1999; Herdina & Jessner 2000; Cook 2001; Cenoz 2003). These researches emphasize the benefits of multilingual education and show how multilingual acquisition is processed. Studies in the area of L3/L*n* have largely contributed to a better understanding of the phenomenon.

Jorge Pinto & Nélia Alexandre. 2021. Preface. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, v–viii. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449763

Following this movement, the purpose of this book is to present recent inquiries in the field of multilingualism and L3, bringing together contributions from an international group of specialists from Austria, Canada, Germany, Portugal, Spain, Switzerland, Turkey, and United States. The main focuses of the articles are two: language acquisition and language learning/teaching in multilingual settings. A collection of theoretical and empirical articles from scholars of multilingualism and language acquisition makes the book a significant resource for teachers and researchers as the articles present a wide perspective from main theories to current issues and reflects new trends in the field. Since the heterogeneity and complexity that characterize multilingual acquisition and learning/teaching is a field of research in fast development and with an increasing interest rate, we believe that the texts included here will be of great relevance for the scientific community.

Part I of the volume presents different topics of L3 acquisition, such as phonology, working memory, and selective attention. Namely, in chapter one, Ghiselli presents an experimental study about working memory and selective attention of conference interpreters. The author seeks to prove the hypothesis that the time dedicated to the practice and the type of activities done during self-study would contribute to the improvement of working memory and selective attention.

In chapter two, Zhou, Freitas, and Castelo present the results of a research on the acquisition of some phonological properties that are problematic for Chinese learners of European Portuguese, namely, the developmental patterns of acquisition of the liquid consonants.

The texts included in Part II show how the research on language acquisition informs pedagogical issues. For instance, how the context of learning previous languages influences the teaching of L3 is addressed by Carvalho, in chapter three. This chapter explores the main differences between bilingual English-Spanish learners of Portuguese L3 who acquired Spanish as an L2 or as a heritage language, both in terms of their performance in the classroom and in terms of the perception of their learning process. Kucukali, in the next chapter, reports the attitudes of three Turkish multilingual teachers (speakers of English, German, and Russian) to plurilingual approaches, and the response of their L3/L*n* students to the plurilingual practices in class, showing that teachers who speak diverse foreign languages play an important role in multilingual classrooms. Paquet and Woll, in chapter five, tackle the benefits of crosslinguistic pedagogy versus classroom monolingual bias, based on an experimental study comprising forty trainees, from Canada and Mexico. The study addresses in particular their perceptions about the use of L1 or other languages in the classroom and the reasons why they use them. Part II also includes a study conducted by Barnes and

Almgren showing how to provide adequate tools for trainees to deal with multilingual classrooms in the Basque Country, starting by understanding their own multilingual skills.

Following the previous studies about acquisition and teaching, it is relevant to consider a section concerning language learning aspects. Therefore, Part III comprises texts on individual learning strategies, such as motivation and attitudes, crosslinguistic awareness, and students' perceptions about teachers' "plurilingual non-nativism". Hofer, in chapter seven, addresses the dynamic and complex nature of multilingual development as well as the need for speakers' interaction for language enhancing. In the subsequent chapter, Mayr puts forward the results of a qualitative study on the development of crosslinguistic awareness in multilingual learning settings, involving plurilingual task-based approach, carried out at a secondary school in South Tyrol, Bolzano. At the end of this part of the volume, Yanaprasart and Melo-Pfeifer compare students' perceptions of nonnative teachers' discourses and their intelligibility. The authors try to answer three main questions: (i) how students perceive a "plurilingual teacher"? (ii) how their perceptions are discursively reported? and (iii) how are these perceptions related to the profile of institutions and disciplinary fields?

Finally, we must highlight that these contributions include several different languages in contact in an acquisition/learning context: Basque, English, French, German, Italian, Ladin, Mandarin Chinese, Portuguese, Russian, Spanish, and Turkish. All the topics covered in this book are scientifically relevant, serving as support to student-teachers, teachers, as well as to all researchers whose work focuses primarily on Multilingualism and Third Language Acquisition. Particularly in a world context where schools are less and less monolingual and monocultural, and where linguistic and cultural diversity is increasing, these studies can help teachers to better cope with these situations in the classroom and provide a valuable resource for researchers.

### **References**


## **Part I**

## **Language acquisition**

## **Chapter 1**

## **Cognitive processes and interpreting expertise: Autonomous exercise of master's students**

### Serena Ghiselli

University of Bologna

The present paper describes the results of a part of a PhD project about working memory (Baddeley & Hitch 1974; Baddeley 2000; Gerver 1975; 1976; Padilla Benítez 1995) and selective attention (Cowan 2000; Moser-Mercer 2000; Seeber 2011; Timarová et al. 2014; 2015) in the training of conference interpreters.

In this experimental study, data on autonomous interpreting exercises were collected. The study group was formed of interpreting students of the master's degree in interpreting of the University of Bologna and can be divided into two subgroups, one subgroup starting the course in 2015 (27) and the other in 2016 (22). The hypothesis was that time devoted to exercise and the type of activities done during self-study would contribute to the improvement of working memory and selective attention, which were measured by a battery of psychological tests.

Before the description of empirical data, the paper includes a review of the main studies on skill acquisition (Ackerman 1988; Anderson 1995; Ericsson 2000; Von Bastian & Oberauer 2014) and on cognitive training methods in interpreter training (van Dam 1989; Dollerup & Loddegaard 1992; Benítez 2002; Gillies 2013; Andres & Behr 2015; Yenkimaleki & van Heuven 2013; 2017; Setton & Dawrant 2016a,b).

### **1 Introduction**

Interpreting expertise is not a natural ability but a hard earned result achieved by individuals with an aptitude for interpreting and thanks to targeted and constant effort. The PhD project on which this paper is based originated from the idea that

Serena Ghiselli. 2021. Cognitive processes and interpreting expertise: Autonomous exercise of master's students. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 3–23. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449765

### Serena Ghiselli

interpreters need specific training and constant practice to become professional interpreters. The project focused on working memory (WM) (Baddeley & Hitch 1974; Baddeley 2000; Gerver 1975; 1976; Padilla Benítez 1995) and selective attention (Cowan 2000; Moser-Mercer 2000; Seeber 2011; Timarová et al. 2014; 2015), which were measured through a battery of psychological tests.

In the project, one study group attended the master's degree in interpreting and the control group attended the master's degree in translation. Every group was divided into two subgroups: students who started the master's degree in 2015 and students who started the master's degree in 2016 (27 and 22 interpreting students respectively; 23 and 37 translation students respectively). For the students who started the master's degree in 2015, data were collected over two academic years, whereas for the students who started in 2016, data were collected over one academic year due to the PhD program time constraints.

For interpreting students, in addition to psychological test results, data about self-study were collected. The focus of the present paper is on self-study data and the aim is to describe autonomous exercise habits of interpreting students. These data were originally collected as part of the PhD project because they were considered to be a relevant variable that could influence psychological test results.

On the basis of bibliographical research, publications on interpreting student self-study are very scarce. Two studies included data collection about autonomous exercise (Fan 2012; Wang 2016). The picture that emerges is that targeted exercise is important to automatize interpreting practice as much as possible and that the quality of self-study is more relevant than its quantity.

The improvement of a specialised skill happens when individuals are motivated, receive feedback and can repeat training activities (Ericsson et al. 1993). Starting from this assumption, data collection focused on the frequency and on the typology of exercises done by interpreting students. Data collection aimed also at favouring participation and, for this reason, the method chosen was an email containing a brief survey that was sent to students every month. The goal was to collect a sample of data that could represent study habits and also avoid that students dropped the study because it took too much of their time.

### **2 Theoretical framework**

### **2.1 Skill acquisition and expertise**

An expert is somebody who has achieved a high level of performance and skill in a specific domain as a result of experience. Ackerman (1988: 290) described skill

### 1 Cognitive processes and interpreting expertise

acquisition as a continuous process during task practice. The process of skill acquisition involves the decline of cognitive load from novice attention-demanding processing to skilled automatic processing.

Anderson (1995) developed the concept of what happens during skill acquisition by identifying three stages in this process: the cognitive stage, the associative stage and the autonomous stage. In the first stage, novices develop declarative knowledge, by memorising a series of elements that are relevant for that ability. In the associative stage, novices gradually identify and eliminate mistakes. In the autonomous stage, the procedures novices have learnt become more and more automatic. When a novice turns verbal and declarative knowledge into procedural knowledge the learning process is almost complete.

In the case of interpreting, a crucial part of the skill acquisition process is the proficient use of WM, the short-term memory that actively decodes and stores information during complex cognitive activities (Baddeley 1997). This is a very important skill that, in the case of interpreters, needs to be trained to work at a high performance in a situation of cognitive load and stress and that has to be coordinated with an efficient attention-switching system in order to manage all the simultaneous activities involved in interpreting.

### **2.2 Expertise in conference interpreting**

One of the earliest studies about the skills that influence academic performance of interpreting students was carried out by Gerver et al. (1984). The scholars compared the results of 12 English and French tests taken by 29 students before starting an intensive course of interpreting with the results of their final exams. Tests included two recall exercises, in which students were asked to repeat two oral texts of 1000 words each without taking notes, a cloze test, that is completing missing words in English oral texts of 500 words each and an error-detection test, which involved the recognition of mistakes in an oral text. It was found that recall exercises predicted the differences in the performance of consecutive interpreting, whereas cloze tests predicted the differences in simultaneous interpreting. Generally speaking, test results were better for the students who passed final exams.

Studies on aptitude for interpreting carried out at the Advanced Schools for Interpreters and Translators of Trieste and Forlì (Pippa & Russo 2002; Russo & Pippa 2004; Russo 2014) found that on-line paraphrase is a predictor of academic success. On-line paraphrase is an exercise in which the student has to understand and report a text while listening to it and to produce another cohesive and coherent text which carries the same meaning using different words.

### Serena Ghiselli

In many domains, such as sport and music, it was demonstrated that quantity and quality of solitary activities are essential to develop skills (Ericsson 1996; 2001; 2002; Helsen et al. 1998). It is therefore reasonable that, also for interpreting, solitary activities influence performance. Only two research papers, as far as it is known, took into account the self-study habits of interpreting students: Fan (2012) and Wang (2016).

Fan (2012) analysed various factors that can influence the development of interpreting competence in a group of 30 Chinese mother tongue students at the University of Newcastle. Students answered a survey at the beginning, in the middle and at the end of the academic year. Factors taken into account included autonomous interpreting exercise. Students had to estimate the average time devoted to self-study daily, without specifying which type of exercise they did. The results were that students devoted one and a half hours to self-study daily over the first semester and one hour and 45 minutes in the second semester. No statistically significant relation between the time devoted to self-study and academic performance was found. From a regression model, a positive relation between the use of learning strategies and academic performance was found. The level of English measured through the IELTS exam also had a positive and statistically significant relation with consecutive interpreting exam results. These findings underlined that quantity counted more than the quantity of time devoted to exercise.

Wang (2016) carried out a longitudinal study with three interpreting Chinese students. Data were collected through weekly diaries and monthly interviews. The results were that students did exercises at home every day and often in groups. This paper does not give quantitative data, but from the interviews it emerged that in oral comprehension students initially understood single words and then moved to the comprehension of the global meaning. They also learnt to take less notes and memorise more. This study supports the idea that cognitive processes are essential to develop interpreting competence, which is the result of targeted practice.

In comparison with the existing studies about autonomous exercise habits of interpreting students, the study presented in this paper collects more data over a longer period of time and for a higher number of participants. The study of Fan (2012) takes only the time devoted to self-study but not the type of activity performed. The study of Wang (2016), instead, focuses only on qualitative data. The present study tries to combine both qualitative (type of activity) and quantitative data (frequency and duration) in order to provide a more detailed description of interpreting trainees' study habits.

### 1 Cognitive processes and interpreting expertise

### **2.3 Cognitive training exercises for interpreters**

In interpreting studies literature, there are some examples of exercises that can be done in class or that can be assigned for self-study. In the present paper, the main exercises have been divided into three groups according to their function and will be briefly described. The groups are exercises to improve memory, exercises to improve consecutive interpreting and exercises to improve simultaneous interpreting.

As regards the improvement of memory, Setton & Dawrant (2016b) suggested that WM limits may be put forward if information is organised in a scheme and divided into significant units. In addition, processes such as discourse analysis, note-taking and language switching have to be automatized. The authors distinguished between discourse modelling and discourse outlining (Setton & Dawrant 2016a). Discourse modelling is a generic term to indicate the process of shaping a mental model of a discourse. Discourse outlining is the act of writing a representation of this model through a list with bullet points. A discourse model is a mental model of the discourse created when listening to a text with the intention of memorising it and that helps to analyse and memorise information. The authors suggest a WM exercise called *idiomatic gist*. This exercise involves the recall of a short text (reading time: 30–45 sec.) having a sophisticated style, which forces the reader to go beyond the words and focus on the meaning behind them.

Most people remember better what they understand, what they can visualised, and what they find interesting or weird. Yenkimaleki & van Heuven (2013; 2017) think that there are three tools that interpreters can combine to improve recall: imagination, association and location. The combination of these tools implies imagining a real location in which put and divide information, to visualise an image of the discourse listened to and to create associations between elements to help recall.

Another technique that Yenkimaleki & van Heuven (2013; 2017) suggest to improve long-term memory (LTM) is storytelling that is reporting a story in the same language as the original without the help of notes. To recall a story, they highlighted the following techniques:

*Categorization:* grouping together elements that share the same characteristics;


### Serena Ghiselli

*Description:* describing the context in which there is an object, its shape or dimension.

Yenkimaleki & van Heuven (2017) carried out a study about the effects of memory training on the quality of interpreting from Farsi into English with 24 consecutive interpreting students from the University of Applied Sciences of Teheran. For a semester, they were divided into two groups: the control group received traditional training (listening and comprehension exercises in English) whereas for the study group, part of the time was dedicated to imagination and storytelling exercises. Statistical analysis pointed out a positive effect of WM exercises carried out by the study group on the quality of interpreting, especially in reducing omissions.

The technique of generalization is also at the basis of an exercise to improve WM called parataxis (Ballester & Jiminez Hurtado 1992). In this exercise the trainer reads a list of elements that are linked (e.g. eagles, hawks, kites, ospreys, buzzards etc.) and the students have to guess the general category (e.g. birds of prey). The researchers also suggested the reverse exercise for LTM, which is called synonyms: it is a brainstorming activity in which, starting from a generic term, students have to recall as many examples related to it as possible.

Among the exercises to improve consecutive interpreting skills suggested by Gillies (2013) there are paraphrasing, that is listening to speeches in a foreign language and repeating them in the same foreign language, and monolingual interpreting, which means reformulating a speech in your mother tongue using the same language. These exercises involve LTM and selective attention skills.

Gillies (2013) also mentioned that taking notes only after the speech has ended exercises the LTM or that taking notes during the speech but not using them during the translations exercises WM, LTM and selective attention. Instead, Chabasse & Dingerfelder Stone (2015) and Setton & Dawrant (2016a) suggest doing consecutive exercises without notes. In this memory training activity, students have to interpret very short texts, at the beginning (10 seconds), and then gradually interpret longer texts (up to two minutes) without taking notes.

As far as simultaneous interpreting is concerned, one of the most used preparatory exercises is sight translation, that is the oral translation of a written text, with or without previous reading (Kalina 1992; 2000; Benítez 2002; Gillies 2013; Setton & Dawrant 2016a). This exercise involves selective and divided attention, it favours chunking of the discourse and anticipation skills.

Shadowing is a preparatory exercise for simultaneous interpreting about which there are divergent opinions. This is the definition of shadowing given by Lambert (1988):

### 1 Cognitive processes and interpreting expertise

A paced, auditory tracking task which involves the immediate vocalisation of auditorily presented stimuli, i.e. word-for-word repetition in the same language, parrot-style, of a message presented through headphones. (Lambert 1988: 381)

On the basis of the differences in the time between the moment when the interpreter receives the message and the moment when the message is translated (ear-voice span, Goldman-Eisler 2002),Andres et al. (2015) distinguished between three types of shadowing:


Schewda Nicholson (1990) and Tonelli & Riccardi (1995) mentioned also another variant of this exercise, that is multiple task shadowing: Shadowing is used as an exercise to divide attention since it has to be performed together with another activity, for example recalling the content after listening or answering comprehension questions. The same principle is at the basis of the exercise *online cloze* (and error correction; Kalina 1992; Setton & Dawrant 2016b) in which, while repeating the text, the student also has to fill in missing words identified by an acoustic signal or to correct mistakes.

Another exercise to train divided attention is *Two questions at a time* (Kalina 1992; Gillies 2013). In this exercise a person reads questions about a specific topic and another person has to answer. While the answer is being given, another question is asked. Questions and answers might be in the same language or in different languages. Gillies (2013) described another variant of this exercise, in which the second person has to answer Yes/No questions and also repeat them while they listen to the next question. Gillies thinks that this exercise is more similar to simultaneous interpreting because, in contrast to shadowing, it does not only involve speaking and listening at the same time, but also understanding.

On-line paraphrase (Russo & Pippa 2004, see §2.2), also called smart shadowing in A or same-language simultaneous interpreting or within-language paraphrase (Setton & Dawrant 2016a), is another useful exercise of divided attention to prepare for simultaneous interpreting.

### Serena Ghiselli

Listening cloze is an exercise that many interpreters trainers suggested (van Dam 1989; Kalina 1992; Benítez 2002; Andres et al. 2015; Setton & Dawrant 2016b). The instructor introduces a discourse mentioning speaker, topic and context. The speech is read and gaps, identified by an acoustic signal, have to be filled in. This a comprehension exercise where the student has to show to be able to grasp the speech meaning even if some information is missing.

### **3 Data collection**

This section describes the data that were collected about autonomous exercise in the sample of interpreting students participating in this study. First of all, an overview of the characteristics of the sample will be given. Then the study plan of the master's degree in interpreting will be briefly described. To conclude, there will be a description of the monthly survey used to collect data.

### **3.1 Participants**

In Table 1.1, the characteristics of the students who took part in the monthly survey are summarised. Language A and Language B refer to the two languages of study, in which students took the entrance exam (see §3.2). Previous interpreting experience refers to any type of interpreting activity (liaison, consecutive, simultaneous or whispered interpreting) done during university courses, in a professional context or as volunteer work. Subgroup 1 and subgroup 2 refer to students recruited in two different academic years. Students from the first subgroup were recruited as volunteers in October 2015, when they were starting the first year of the master's degree in interpreting. Data collection about autonomous exercise for them started in January 2016 and ended in February 2018. The second subgroup refers to interpreting students starting the master's degree in October 2016. For this subgroup, data collection started in December 2016 and ended in February 2018. Mean age at T1 indicates the age at the first WM and selective attention test session. T1 varies from student to student because the test battery was done one person at a time by appointment. For the students in the first subgroup T1 is a day in November–December 2015, for the students in the second subgroup it is a day in October–November 2016.

### **3.2 Study plan**

The master's degree in which students were enrolled during data collection was a two-year program that included the study of two foreign languages that students could choose among English, French, German, Spanish and Russian.

### 1 Cognitive processes and interpreting expertise


Table 1.1: Study group characteristics

In order to enrol in the course, students had to pass an entrance exam in both the languages of study and in Italian. The entrance exam was changed from academic year 2015–2016 to academic year 2016–2017. The first subgroup had to take cloze tests (filling in missing words) in Language A and Language B and recall tests (repetition without notes) of two speeches (about 4 minutes long): one speech in Language A, which they had to repeat in the same language, and one speech in Language B, which had to be repeated in Italian. They also had to do an on-line paraphrase exercise (see §2.2) in which they had to reformulate a text in Italian. The second subgroup took three cloze tests and three recall tests, one in Language A, one in Language B and one in Italian for each type of exercise.

Language A involved both interpreting exams (consecutive and simultaneous) from the foreign language into Italian and from Italian into the foreign language. Language B involved exams only from the foreign language into Italian, but students could add an optional exam of 6 ECTS from Italian into Language B. 12 students (4 from the first and 12 from the second subgroup) also took the optional exam. Overall, the compulsory interpreting exams counted for 26 ECTS for Language A and 22 ECTS for Language B. After passing all the exams, students also had to pass the final exams, that is a mock conference on a specific topic, during which they had to perform four interpreting tasks, two for Language A and two for Language B.

### **3.3 Monthly survey**

Self-study is an important part of learning but it is difficult to measure because it implies the cooperation of the students, who have to give information about what they do outside the class. A balance was struck between getting reliable information and encouraging students to be constant over time in participating, by choosing to send them a monthly survey by email.

Data collection started the month after all the students had taken WM and selective attention tests the first time, that is in January 2016 for the first subgroup and December 2016 for the second subgroup. The questions were in Italian and aimed at knowing what the student had done in the day in which the email was sent. The emails were always sent in the evening and asked about that specific day.

This procedure had a double goal. On the one hand, it aimed at helping students to recall what they did since it referred to exercises they had done little time before. On the other hand, the objective was favouring the reliability of the answers, since students were not asked to calculate how much time they devoted to exercise on average but they just had to think about a specific day. When students did not answer to the email, they were sent a reminder and encouraged to choose a different day or, if they preferred, to give data about an average day of that month. Students received the emails in random working days that changed every month.

The survey included the following questions:


### 1 Cognitive processes and interpreting expertise


### **4 Results**

This section aims at displaying the findings of the data collection. In 61.41% of the answers, participants declared that the day of data collection was representative of their exercise habits. When the day did not correspond to the average (38.59%) in the majority of cases it was because students said they normally did more exercise. The average percentage of answers given over the emails sent was 85.2%. Taking this into account, the data collected can be considered a representative sample of students' study habits. The answers are based on participants' own perception of their behaviour, which could be subjective, so this needs to be considered when looking into the results.

Table 1.2 displays the data of both subgroups. The same data are shown from two different perspectives: duration and frequency of exercise. Duration is expressed with the mean of minutes devoted to exercise daily. Frequency is expressed with the percentage of exercise done out of the number of data collection requests, that is how frequently students did exercise, independently from the time they devoted to it.

Since this data collection was part of a broader PhD project, data were divided according to the dates of test sessions in order to be able to compare autonomous exercise with test results. The first subgroup took the test three times (November–December 2015, May–June 2016 and April–June 2017), whereas the second subgroup took the tests only twice (October–November 2016 and April– June 2017). For both subgroups, T1 corresponds to the beginning of the first year of the master's degree and T2 to the end of the first year, T3 for the first subgroup corresponds to the end of their second year of the master's degree. Data collection ended in February 2018 because then the last exam session of the previous academic year takes place. The type of data analysed were:

• ex: exercise, independently from the type of activity;

### Serena Ghiselli


For every data set, mean values with the corresponding standard deviations were calculated for the following periods:


The self-study profile that emerged is very diversified, with a high standard deviation, especially in the data about recall,. the majority of students did not do this exercise. On average, data about frequency are more homogenous than data about duration.

Recall exercises were performed frequently at the beginning of the master's degree and less and less as time passed. This is not surprising since recall exercises are considered as a preparatory activity for interpreting, consecutive interpreting in particular. Sight translation was done constantly over time, the same as consecutive interpreting. The exercise of simultaneous interpreting increased from the first to the second year. During the first year, simultaneous interpreting is gradually introduced in the lessons, so it is normal that this type of interpreting exercise was done more in the second year of training. Overall, the mean of minutes devoted to exercise was 63.74 (SD = 45.74) between T1 and T2, 94.62 (SD = 72.24) between T2 and T3 or after T2. As far as the frequency is concerned, students said they did at least one type of exercise in 40% of the answers. After T3, the duration of exercise for the 1st subgroup is much longer than before, 131.94 (SD = 75.41). This is probably due to the fact that in this period students had final exams, which are very stressful and demanding.

### 1 Cognitive processes and interpreting expertise


Table 1.2: Data about autonomous exercise. Column *duration* lists the mean of daily minutes. Column *frequency* lists the mean percentage of days. Values in brackets are standard deviations.

To conclude data description, the answers given to questions 6 are represented in Figure 1.1. Question 5 asked whether the student did another type of exercise and, if yes, question 6 was an open question asking which other type of exercise was done.

The mean percentage of affirmative answers to question 5 was 35.66% and 85% of the students declared at least once to have done a type of exercise different from those mentioned in the other questions. The answers given to question 6 were divided into 12 categories:

1. Shadowing (see §2.3): repetition of a text in a foreign language while listening to improve pronunciation, learn useful expressions in the foreign language and exercise on fast speeches;

Figure 1.1: Additional type of exercise


Figure 1.1 shows that the three most common typologies of exercises mentioned by students were terminology research and study (33%), listening in a foreign language (20%) and reading in a foreign language (13%).

### 1 Cognitive processes and interpreting expertise

### **5 Discussion**

As mentioned before (see §2.2), there are only two studies about autonomous exercise habits of interpreting students, Fan (2012) and Wang (2016), so the present study is one of the earliest contributions to this topic. Data are based on what students reported, so they might not be accurate. At the same time, the researcher was not their instructor and they did not get any rewards for their participation in the study, which was on a voluntary basis, so there are no apparent reasons why they should have lied.

In the study of Fan (2012), data about the autonomous exercise of interpreting students were collected over an academic year. In the first semester the mean daily time devoted to exercise was 90 minutes, whereas in the second semester it was about 105 minutes. In this study, if all types of exercises are taken into account, the mean daily time devoted to self-study was 88.43 (SD = 50.6) minutes. Time devoted to exercise increased over time, like in Fan's study. The length of self-study activities found by Fan was also confirmed, since in both studies the result is that students devote about one and a half hours a day to autonomous exercise. In this study, as the high standard deviation shows, the habits changed a lot from student to student.

Recall is a preparatory exercise and this study clearly shows that students devoted less and less time to it as they advanced in training and tended to focus only on consecutive and simultaneous interpreting. Recall can, however, be considered a valuable exercise also for more advanced students. It requires a lot of concentration, which is an essential skill both in consecutive and in simultaneous interpreting. It implies to rely only on one's mental resources without the help of notes. Notes are a valuable and necessary help in consecutive interpreting, but they cannot replace logic and critical thinking, so the interpreter can never be too dependent on them and has to make an effort to create a mind map of the message to avoid saying contradictions.

Terminology research, listening and reading in a foreign language were not included in the survey questions because the PhD project focused on cognitive aspects of interpreting. The fact that students mentioned these activities of language improvement as further self-study is in line with what trainers recommend. Reinforcing linguistic skills is in fact very important to achieve a good interpreting performance.

Students who did shadowing exercises said they did them in the foreign language to get used to fast speakers and to improve their accent and learn new expressions. This exercise is normally considered a preparatory exercise for simultaneous interpreting, but in this data set students declared to do this exercise

### Serena Ghiselli

in the first year of the master's degree (8% of the answers between T1 and T2), in the second year (4% of the answers between T2 and T3 or after T2) and, in the data collected after T3 from the students of the 1st subgroup, shadowing was also mentioned after the end of the second year classes (10% of the answers). Over the entire period, the percentage of shadowing was 6%. It emerged that, differently from what could be thought, students considered shadowing a useful exercise not only in an early stage of training but also in an advanced stage.

Eventually, the percentage of times when students declared to have done any type of exercise (40%) was quite low, since it means that in more than half of the days when they were asked, they did no exercise. This goes against what trainers would expect, but it might be due to long hours spent in class, which leave little time and energy for self-study.

### **6 Conclusion**

The data displayed here are part of a wider PhD project for which other data were collected and comparisons between different data sets were done. The present paper deals with data collected about autonomous exercise and has a descriptive approach.

From the data collected it came out that self-study habits among students were very diversified, but the mean time devoted to interpreting autonomous exercise was in line with the findings of Fan (2012), that is about one and a half hours per day.

Exercise focused more on consecutive and simultaneous interpreting activities and less on support exercises such as sight translation and recall exercises.

In the open question, where students could mention other exercises they did, most of the time they mentioned terminology research and study and listening or reading in a foreign language. This is in line with expectations, since language study and vocabulary learning are life-long learning activities for an interpreter.

Shadowing was also mentioned among other type of exercises. Shadowing presents advantages and disadvantages (Kurz 1992) but if a more difficult version of this exercise is done, like repeating a text in a foreign language having some difficult elements, such as a high speed rate or a difficult accent of the speaker, it may be useful to improve listening and speaking skills in a foreign language. Various scholars (Kalina 1992; Benítez 2002;Gillies 2013; Setton & Dawrant 2016b) suggested matching shadowing with other exercises, such as online cloze tests and comprehension questions afterwards to check whether also the message, and not only the words, was understood.

### 1 Cognitive processes and interpreting expertise

On-line paraphrase was mentioned by students as a further exercise, but only in 1% of the answers. This exercise is more difficult than shadowing and implies a thorough understanding of the message. Another exercise, that was not mentioned by students and for which both concentration and understanding are necessary, is *two questions a time* (Kalina 1992; Gillies 2013) (see §2.3).

Further developments could be carrying out an experimental study (Yenkimaleki & van Heuven 2017 is an example) to see whether more support exercises such as recall, shadowing and on-line paraphrase, that some participants mentioned, or Two questions at a time, that none of the students did, would be useful for students to improve their interpreting skills. The potential improvement of interpreting skills through targeted exercises could be verified using real interpreting tasks instead of psychological tests. In this hypothetical experimental framework, a language level assessment would be necessary to see whether the language proficiency of students is comparable.

### **References**


### Serena Ghiselli


### 1 Cognitive processes and interpreting expertise


### Serena Ghiselli

*Conference on Interpreting Studies, 9-11 November 2000*, 257–266. Amsterdam: Benjamins.


## **Chapter 2**

## **On the acquisition of European Portuguese liquid consonants by L1-Mandarin learners**

Chao Zhou University of Lisbon

Maria João Freitas University of Lisbon

### Adelina Castelo

University of Lisbon

The present study aimed to investigate the developmental patterns of acquisition of the European Portuguese liquid consonants by L1-Mandarin speakers, and to examine the prosodic effect on L2 phonological acquisition. Fourteen L1-Mandarin learners participated in a picture-naming task and results showed that the alveolar lateral was produced accurately in branching/non-branching onset, while it was frequently vocalized in coda; the palatal lateral was produced as a Mandarin palatalised lateral nearly half of the time; the tap was acquired in coda before onset and the repair strategies were context-dependent: in non-target-like realisations, it was articulated as an alveolar lateral in onset but accommodated in diverse ways in coda (epenthesis, deletion, segmental repair); the uvular rhotic was acquired fairly well due to the L1-L2 allophonic overlap. Our results suggest that the degrees of difficulty in L2 segmental learning vary as a function of the distance between L1 and L2 categories and the syllable constituency effect observed in the acquisition of the Portuguese alveolar lateral and the tap could be attributed to L2-to-L1 allophonic category mapping and L1 phonotactic restrictions, respectively.

Chao Zhou, Maria João Freitas & Adelina Castelo. 2021. On the acquisition of European Portuguese liquid consonants by L1-Mandarin learners. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 25–45. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449767

Chao Zhou, Maria João Freitas & Adelina Castelo

### **1 Introduction**

Novel liquid consonants are notoriously problematic for L2 learners. However, for an extended period of time, the acquisition of L2 liquids has been only explored within a limited number of language-pairs, such as native speakers of East Asian languages acquiring the English approximant (e.g. Aoyama et al. 2004, Brown 1998) or L1-English learners studying the French and Spanish rhotics (e.g. Colantoni & Steele 2007; 2008, Face 2006, Steele 2009, Waltmunson 2005).

The Chinese learners' struggle with the European Portuguese (EP) liquids has long been reported in the literature, while our current understanding of this difficulty is still far from being complete. Previous studies either drew their conclusions on the basis of pedagogical observations (e.g. Batalha 1995, Espadinha & Silva 2009, Martins 2008) or only studied a subset of the EP liquids (Oliveira 2016). The present study hence seeks to make both empirical and theoretical contributions by expanding on prior research in several ways. First, the acquisition of all four EP liquids (/l/, /ʎ/, /ɾ/ and /ʀ/) was elicited through an experimental task (picture-naming). Second, the L2 phonological acquisition of EP liquids was accessed across prosodic contexts (syllable and word-level positions), which has been widely shown to shape both L1 (e.g. Fikkert 1994, Freitas 1997) and L2 phonological acquisition (e.g. Waltmunson 2005, Colantoni & Steele 2008). Third, the experimental findings were discussed in the light of the current L2 speech learning models.

The paper is organized as follows: first, we review several theoretical frameworks which account for difficulties in L2 speech from different perspectives. In §1.2, the phonetic and phonological characteristics of the EP and Mandarin liquids are discussed, followed by a summary of previous studies on the acquisition of EP liquids by Chinese learners. §2 introduces the current study, including research questions and methodology. Results are presented and discussed in §3 and we offer conclusions in §4.

### **1.1 Accounts for L2 speech learning difficulties**

One of the central goals of studies on L2 phonological acquisition is to elucidate and predict difficulties in acquiring a novel sound structure, for instance, phonological features, syllabic constituents, segments, to mention a few. Numerous studies have converged on the idea that the divergence between the learners' output and the target form can be usually attributed to cross-linguistic influence (CLI), the interaction between learners' previously acquired language(s) and the target language (see Major 2008 for a general review). So it comes as no surprise

### 2 On the acquisition of Portuguese liquids by L1-Mandarin Learners

that CLI constitutes a core feature of most L2 speech models (see Colantoni et al. 2015 for a general review).

The well-known Speech Learning Model (SLM; Flege 1995) <sup>1</sup> postulates that CLI manifests itself first in perception and whether learners are able to acquire a novel segment is contingent on whether they can perceive the sound accurately, under the influence of pre-existing L1 allophonic categories. In particular, the SLM posits that the relationship between L1 and L2 sounds exists on a continuum ranging from "identical" over "similar" to "new". For simplicity, let us consider three prototypical scenarios for the moment:


In the last scenario, it is very likely that the L1-L2 category equivalence will block the formation of a new L2 category, as learners will rely on an L1-L2 composite category to process the novel sound, an example being the acquisition of English /l/ and /ɹ/ by L1-Japanese learners. According to the SLM, learning the English /l/ is expected be more demanding than /ɹ/ for Japanese learners, since [l] is perceptually closer to the Japanese category /ɾ/ (Iverson et al. 2001). This was borne out in a longitudinal study where greater improvement for English /ɹ/ than English /l/ was found both in perception and production by L1-Japanese learners (Aoyama et al. 2004).

On the other hand, some theories highlight that L2 speech difficulties may stem from CLI in articulation (Honikman 1964, Zimmer & Alves 2012). In production, a speaker's task is to retrieve the stored phonological representation of the indented lexical entry and decode the abstract phonological representation into articulatory gestures (phonetic implementation). As stated in articulatory setting theory (AS; Honikman 1964), each language has its own articulatory settings, which are instantiated by cross-linguistic difference in terms of tongue rest

<sup>1</sup>Other perception-based L2 models, such as the perceptual assimilation model-L2 (Best & Tyler 2007) and the automatic selection perception model (Strange 2011), are not revisited here because they do not make explicit assumptions on L2 production.

### Chao Zhou, Maria João Freitas & Adelina Castelo

position (Gick et al. 2004, Wilson & Gick 2014), and CLI is expected when L1 and L2 articulatory settings do not resemble each other. Supporting evidence can be found in Święciński (2013), where an examination using electromagnetic articulography showed that Polish beginners of English do not display a substantial difference of tongue position between L1-Polish and L2-English production, while the difference exists for advanced learners. These results indicate that learners still articulate an L2 sound within L1 articulatory settings before mastering novel gestures and/or gestural coordination.

In addition to CLI, according to the L2 speech model proposed in Colantoni & Steele (2008), universal phonetic constraints may also give rise to speech learning difficulties. Numerous studies have demonstrated that, in the development of a novel category, learners do not master it equally in all positions (e.g. Colantoni & Steele 2008, Waltmunson 2005). Cross-linguistic evidence was reported in Colantoni & Steele (2008), where L1-English learners acquired both the Spanish /ɾ/ or the French /ʁ/ faster in intervocalic onset than in word-internal coda position. The onset-coda asymmetry was argued to be effected by the universal salience of syllable onset in terms of learnability and accessibility. Moreover, Colantoni & Steele also observed that English native learners did not simultaneously acquire all the phonetic properties (voicing, place and manner of articulation) involved in the realization of the French uvular rhotic. The manner of articulation, which is considered phonetically salient, was present earlier in learners' production than the place and voicing features, revealing once again that difficulties in L2 speech learning may be attributed to phonetic constraints on production.

### **1.2 Liquid consonants in EP and in Mandarin**

EP has four segments in the class of liquids, two laterals /l/, /ʎ/ and two rhotics /ɾ/, /ʀ/ (Mateus et al. 2016).

The EP /l/ exhibits two allophonic variants, an alveolar lateral [l] in non-branching/branching onset and a velarised [ɫ] in coda (Mateus & D'Andrade 2000, Mateus et al. 2016). Both acoustic and articulatory studies show that the EP /l/ carries a certain degree of velarisation, evidenced by low F2 formant values, irrespective of syllable position and adjacent context (Andrade 1998, Marques 2010, Catarina et al. 2011). Recently, Rodrigues et al. (2019) reported that, despite the fact that the F2 values of the EP /l/ are consistently low, in support of the idea that /l/ is velarised across syllable contexts, the F3 values are relatively higher in coda than in onset, which can be regarded as an acoustic correlate of a different degree of velarisation, justifying the allophonic alternation of the EP /l/. In

### 2 On the acquisition of Portuguese liquids by L1-Mandarin Learners

terms of distribution, /l/ can occupy all syllable (singleton, complex onset and coda) and word-level (word-initial, word-medial and word-final) positions.

The other lateral, /ʎ/, is palatal and can only occur in intervocalic onset position. An MRI-based articulatory study revealed that the realization of the EP /ʎ/ requires a complete contact of the tongue blade and/or pre-dorsum with the alveolo-palatal region (Teixeira et al. 2012), in contrast to the traditional description that it is articulated at the dorso-palatal zone (Sá Nogueira 1938, apud Mateus & D'Andrade 2000).

The EP /ɾ/ is a tap, articulated with a very rapid tongue tip movement against the alveolar ridge. Although there was no register on the allophonic variation of /ɾ/ in the traditional descriptions (Mateus & D'Andrade 2000), an acoustic study revealed that its phonetic realization varies depending on the syllable position and on the adjacent segmental context (Silva 2014). In particular, in onset position, it is mostly a tap, while in coda it could be produced as a tap, an approximant or a fricative, hinging on the following consonant. Moreover, the occurrence of tongue tip closure and of the vocoid (supporting vowel) is favoured before a stop and the fricative variant is more common when the following consonant is also a fricative. In word-final position, EP /ɾ/ is often produced as a voiceless fricative (Jesus & Shadle 2005). With respect to the distribution, it can occur in all syllable and word-level positions, except word-initially.

The EP /ʀ/<sup>2</sup> is most often realized as a fricative with the place of articulation ranging from the velar to the uvular region and can be either voiced or voiceless (Rennicke & Martins 2013, Rodrigues 2015, Pereira 2020). It only occupies nonbranching onset, either word-initially or word-internally.

The Mandarin inventory includes two liquid consonants, /l/ and /ɻ/. Comparable to EP, the Mandarin /l/ is also an alveolar lateral. Regarding distribution, /l/ occurs exclusively in non-branching onset. The exact phonetic nature of the Mandarin rhotic, which is allowed both in onset and in coda (branching onset is not legitimate in Mandarin), has been an ongoing debate since it has been argued to be a retroflex approximant [ɻ] (Duanmu 2005, Lin 2007, Zhu 2007), a post-alveolar fricative [ʐ] or a retroflex vowel [əʳ] (Duanmu 2007). A recent study in which both ultrasonic image and acoustic analyses were performed on

<sup>2</sup> Some researchers analysed the EP uvular rhotic as an underlying /ɾ/ word-initially and /ɾɾ/ word-internally (e.g. Mateus & D'Andrade 2000). The debate on the nature of the EP underlying rhotic is beyond the scope of this study. However, studies on L1 phonological acquisition suggest that there seem to be two rhotics in the EP phonological inventory, since the Portuguese children processed these (the alveolar and the uvular rhotic) differently. For instance, they tended to use a lateral ([+continuant]) for the target alveolar tap but a stop ([−continuant]) for the target uvular /ʀ/ (Amorim 2014, Costa 2010).

### Chao Zhou, Maria João Freitas & Adelina Castelo

the Mandarin rhotic revealed that both retroflex approximant and post-alveolar fricative were possible allophonic variants, subject to inter-speaker difference (Xing 2019).

Additionally, it is worth noting that Mandarin has a palatalized lateral [lj] and a velar fricative /x/, which may play a role in L2 phonological acquisition of the EP liquids. [lj] is phonetically very similar to [ʎ] and [lj] was used as repair strategy for /ʎ/ by L1-English learners of Portuguese (Oliveira et al. 2016). It is thus very likely that L1-Mandarin learners will fail to discern the difference between [ʎ] and [lj] reliably. The Mandarin /x/ may be realised either with velar or glottal place of articulation (Lin 2007). The velar allophone [x] coincides with one of the possible variants of the EP /ʀ/ (Rennicke & Martins 2013, Rodrigues 2015, Pereira 2020). This overlap might facilitate the acquisition of the EP /ʀ/ (produced as [x]) and at the same time could lead to the segmental replacement with [h] in L2 production.

### **1.3 Previous studies on EP liquids by Chinese learners**

In the literature, it has long been reported that Chinese learners find Portuguese liquid consonants difficult. Among the first to register the deviant production of the EP liquids by Chinese learners in the classroom setting, Batalha (1995) found that Chinese learners tend to vocalize the word-final lateral (/l/ → [w]) and to confuse [ʎ] with [l]. In the case of /ɾ/, they may replace it with the lateral in onset position or delete it word-finally. In addition to these deviant realizations, Martins (2008) also reported the interchangeable use between [ɾ] and [ʁ] for target /ɾ/.

Oliveira (2016) tested the perception and production of the EP word-initial consonants by L1-Cantonese learners in the laboratory. Her results showed that /l/ was correctly produced 74.1% of the cases and it was most often realized as [n] when the target was not produced. The production accuracy of /ʀ/ reached 40.5% and the use of [l] was the most prevalent repair strategy.

The brief review on previous studies suggests that Chinese learners may have difficulty with all EP liquids; however, the relative degree of difficulties and the underlying reasons for such difficulties are still far from clear. In the present study, we aim to contribute to the current understanding of this L2 speech learning difficulty by testing the production of EP liquids by L1-Mandarin learners across prosodic contexts and discussing the experimental findings in the light of the L2 speech theories presented in §1.1.

### **2 The present study**

### **2.1 Research question**

The current study has three goals. Firstly, we intend to investigate the L2 acquisition of all four EP liquids, /l/, /ʎ/, /ɾ/ and /ʀ/, by L1-Mandarin learners through an experimental production task. Secondly, we aim to test the prosodic position effect on the L2 phonological acquisition of these four categories. Finally, we attempt to contribute to the ongoing discussion on the explanatory factors for L2 speech learning difficulty. To achieve these goals, the study is designed to answer the following research questions (RQs):


### **2.2 Method**

### **2.2.1 Participants**

Fourteen L1-Mandarin, L2-English learners of EP aged 19 to 20 participated in the current study. All participants were enrolled in the third year of a universitylevel Portuguese language and culture course. They come from different regions in China but all have Mandarin as their dominant language and have studied English since primary school (mean learning time = 14 years). These participants started to learn Portuguese when they entered university, which was, in all cases, after the age of 17. Before coming to Portugal, they studied EP in a formal classroom setting for two years and, at the moment of being recruited, they had been attending the Portuguese language course at the University of Lisbon for two months, where they all were assigned to B1-level<sup>3</sup> . No participant reported any hearing or speech impairment.

<sup>3</sup>According to the European framework for language assessment (Common European Framework of Reference, a definition of different language levels written by the Council of Europe), Level A corresponds to low proficiency, level B to intermediate proficiency and level C to advanced proficiency.

### Chao Zhou, Maria João Freitas & Adelina Castelo

### **2.2.2 Stimuli**

The stimuli consisted of 52 EP disyllabic or trisyllabic real words. All 42 test words were controlled for liquid type (four EP liquids), stress (all liquids occurred in stressed syllable), syllable (singleton onset, onset cluster and coda) and wordlevel position (word-initial, word-internal and word-final). Ten distractors were intermixed with the test words. All stimuli were nouns or adjectives in order to achieve a transparent relationship between the stimuli and the graphic representations. The entire stimuli list can be found in Appendix A.

### **2.2.3 Experimental task**

A picture-naming task was performed to elicit L2 production of the EP liquids across prosodic contexts. During the task, the participants were presented with pictures representing each stimulus in a random order on a computer screen via PowerPoint. The task was self-paced and took about 3 to 5 minutes. The motivation for adopting a picture-naming task rather than a word reading task, which is widely used for L2 production studies, was to avoid the orthographic influence (e.g. Hayes-Harb & Masuda 2008, Escudero & Wanrooij 2010).

Subject to the limited vocabulary size of the participants, a familiarisation task was performed a week before data collection. During the familiarisation phase, all participants were given a word list containing all stimuli with their written forms in Portuguese, in Chinese and their corresponding pictures (see some examples in Appendix B), which were used later in the picture-naming task, and they were required to memorize all the words on the list.

Participants were tested individually in a quiet room. They were told that they would see a series of pictures on the computer screen and were asked to articulate the word represented by each picture as clearly as possible. Their productions were recorded using *Praat* (Boersma & Weenink 2016) on a laptop, at an audio sampling rate of 44.1 kHz.

### **2.2.4 Data analysis**

After data collection, all sound files were imported into the program *Phon* (Rose & MacWhinney 2014), where the segmentation and the phonetic transcription were performed through an auditory analysis, combined with a visual analysis of acoustic cues present in the wave form and spectrograms by the first author, a native speaker of Mandarin and advanced L2 Portuguese speaker trained in acoustic analysis. All coding was then carefully checked by the second author,

### 2 On the acquisition of Portuguese liquids by L1-Mandarin Learners

a phonetically trained native EP linguist. The tokens on which two authors disagreed were sent to a third trained EP native phonetician. Two tokens were excluded from data analysis due to ambiguity. To determine the role of prosodic position effect on the outcome of target versus non-target production, a series of generalized logistic mixed models were run using the lme4 package (Bates et al. 2015) in R, with syllable position or word-level position as a fixed effect. Each model included random intercepts for participant and stimuli, as well as random slope for participant. All *p*-values were generated via likelihood ratio tests.

### **3 Results and discussion**

The accuracy rates were summarized with respect to the syllable (Figure 2.1) and word-level positions (Figure 2.2). The repair strategies employed by participants are presented in Table 2.1.

Figure 2.1: The accuracy rate of EP liquids produced by L1-Mandarin learners across syllable positions

Results demonstrated that in non-branching onset the EP /l/ was accurately produced in all cases (accuracy rate: 100%). This high accuracy can be attributed to the fact that the L1 Mandarin inventory likewise comprises an alveolar lateral, which seems to bear no detectable difference from the EP /l/. Consequently, as predicted by SLM (identical scenario), the reuse of the L1 lateral will lead to targetlike L2 performance. In contrast to what was reported in Oliveira (2016), the participants in the present study did not confuse the lateral with [n] in syllable onset. This difference might be explained by the distinct dialectical profiles of the participants between two studies. The participants in Oliveira (2016) were speakers of Cantonese, in which /l/ and /n/ may be freely substituted for each

Figure 2.2: The accuracy rate of EP liquids produced by L1-Mandarin learners across word positions


Table 2.1: Repair strategies used by L1-Mandarin learners for target EP liquids across prosodic positions

### 2 On the acquisition of Portuguese liquids by L1-Mandarin Learners

other at initial position (Avery & Ehrlich 1987), whereas no Cantonese speaker was recruited for the present experimental task.

A major difficulty with /l/ was detected in coda (accuracy rate: 16.7%), where learners performed substantially worse than in non-branching onset ( 2 (1) = 28.349, < 0.0001). When failing to produce the target, learners mainly vocalised it as [w] (77.3% of the time), in line with previous studies (Batalha 1995, Martins 2008). The difference between word-medial and word-final position did not reach significance ( 2 (1) = 0.043, = 0.8363). The velarised lateral [ɫ] seems to constitute a persistent challenge for L1-Mandarin learners since the [ɫ]-vocalization has also been attested in their L2 production of English (He 2014). Several explanations are plausible. The first one concerns CLI on L2 perception (SLM; Flege 1995). Guan & Kwon (2016) performed a perceptual transcription experiment testing how Russian consonants are categorized by monolingual L1-Mandarin listeners and their results showed that the velarised lateral was categorized as a /w/-like category. Accordingly, it is very likely that the EP [ɫ] is perceptually identified as /w/, presumably due to low F2 values. If this were the case, /w/ would be stored in the L2 lexicon and be retrieved consequently in production. Alternatively, the vocalisation of /l/ in coda might stem from articulatory imprecision (Honikman 1964), since the realization of [ɫ] stipulates both a coronal and a dorsal gesture, whose coordination is entirely novel to L1-Mandarin speakers. He (2014) speculated that, before mastering this novel gestural coordination, L1-Mandarin learners might only preserve the dorsal gesture, which precedes the coronal gesture in the realization of [ɫ] (Sproat & Fujimura 1993), resulting in alveolar contact loss. It is worth noting that the perceptual and articulatory explanations are not mutually exclusive as it has been shown experimentally that the articulatory and acoustic cues, whether in combination or in isolation, are sufficient for triggering lateral-vocalisation (Recasens & Espinosa 2010). This cross-modalities force may elucidate the wide distribution of [ɫ]-vocalisation, such as in synchronic variation (e.g. Recasens & Espinosa 2005; 2010), diachronic sound change (e.g. Graham 2017), L1 phonological acquisition (e.g. Freitas 1997) and L2 speech learning (He 2014).

It comes as a surprise to us that the accuracy rate of producing /l/ reached 97.6% in branching onsets, a syllable structure missing in learners' L1 Mandarin, since prior research on L2 English suggested that L1-Mandarin learners often accommodated the illicit branching onsets through epenthesis or deletion (Chen 2003, Enochson 2014). The lack of structural modifications in the acquisition of EP /l/ in branching onset might be due to the fact that learners produced a complex sound, rather than a canonical onset cluster. Complex sounds are structurally distinct from onset clusters, since two elements in a complex sound are associated to

### Chao Zhou, Maria João Freitas & Adelina Castelo

a single skeletal position, whereas each segment of a canonical branching onset has its own projection at the skeletal level (Selkirk 1982). Therefore, realising two consecutive consonants as a complex sound is structurally less demanding and it has been argued to be a strategy employed by children before the branching onset becomes available in their phonological system (Freitas 2003). This possibility can be tested in further studies that measure the duration of the EP obstruent + /l/ clusters produced by L1-Mandarin learners across different proficiency levels. If using the complex sounds is indeed an intermediate stage in the phonological acquisition of EP, a significant difference in terms of cluster duration would be attested between beginners and advanced learners. Another plausible explanation for high accuracy of /l/ in branching onset concerns learners' experience with another L2, English. All participants in the current study started to learn English around age six and reported having spoken English for 14 years on average. This long-term exposure to English, where branching onsets are common, may have led the participants to overcome the L1 structural restriction, thus facilitating the acquisition of the onset clusters of another non-native language. The positive L2 influence during the phonological acquisition of an L3 has been reported in an increasing number of studies (e.g. Tremblay 2007, Llama et al. 2010).

With respect to the palatal lateral /ʎ/, the participants produced it correctly 52.4% of the time. The use of [l] for the target /ʎ/, a repair strategy observed in the literature (Batalha 1995, Martins 2008), was rather rare in the present study (merely two tokens). Instead, our prediction in §1.2 was borne out since the participants resorted to their L1 category [lʲ], which might stem from CLI in perception or articulation. The palatal lateral [ʎ] has been shown to contain a glide-like CV (consonant to vowel) transition, a perceptual cue leading to glide-like interpretation (Colantoni 2004). The close perceptual distance between [ʎ] and [lʲ] thus might give rise to an equivalence between two categories during the construction of a novel sound category (similar scenario in SLM), resulting in the use of /lʲ/ in the L2 speech. On the other hand, the articulatory imprecision might also be responsible for the use of [lʲ] for the target /ʎ/ (Honikman 1964), since the gestural differences between [ʎ] and [lʲ] are rather subtle: [lʲ] is higher than [ʎ] at the middle of the tongue and the tongue tip during the realization of [lʲ] is more anterior than that of [ʎ] (Wong 2017). Accordingly, before acquiring target-like gestural coordination, learners might still articulate the target palatal lateral in an L1-like manner. Again, the perceptual confusability and articulatory imprecision may work in tandem, contributing to learners' difficulty with the EP /ʎ/.

The EP /ɾ/ in non-branching onset was the most problematic novel structure for L1-Mandarin learners (accuracy rate: 39%). When failing to produce the target tap, the participants uniformly turned to [l], presumably due to the perceptual

### 2 On the acquisition of Portuguese liquids by L1-Mandarin Learners

similarity between [ɾ] and [l] (similar scenario in SLM). Although the L2 perception of the EP /ɾ/ was not tested in the current study, the perceptual confusability between [ɾ] and [l] has been attested in the acquisition of Spanish by L1- Mandarin learners (Chih 2013). Accordingly, it is plausible that the EP tap might be stored as an L1-L2 composite category (/l/-/ɾ/) in the L2 lexicon, the activation of which will lead to the alternation between [l] and [ɾ] in speech production. In the case of the intervocalic /ɾ/, the articulation-based account seems to be less probable as, articulatorily speaking, [t] is also quite close to the tap (Ladefoged & Johnson 2011), but it was not used by the participants in the present study.

/ɾ/ was produced more accurately in coda (accuracy rate: 69%) than in nonbranching onset ( 2 (1) = 9.87, = 0.002). No significant difference in terms of production accuracy was found between word-internal and word-final coda ( 2 (1) = 0.01, = 0.92). The onset-coda asymmetry in L1-Mandarin learners' acquisition of the EP /ɾ/ (coda > onset) is in opposition to the one usually reported in the literature (onset > coda; e.g. Colantoni & Steele 2008, Waltmunson 2005), where the syllable onset was considered to be salient in terms of accessibility and learnability (Carlisle 1998). To our best knowledge, the preference for syllable coda over onset is novel to the L2 speech learning literature; however, it has been attested in the L1 phonological acquisition of Hebrew rhotic (Cohen 2015). The fact that the Hebrew rhotic displays less allophonic variation (more phoneme consistency) in coda than in onset leads Cohen to postulate that the phoneme consistency accelerates the development of Hebrew rhotic in coda. The consistency-based explanation does not remain plausible in the case of the acquisition of the EP tap, because it manifests more allophonic variations in coda than in onset (Silva 2014). Alternatively, the higher production accuracy of the EP /ɾ/ in coda might be due to the Mandarin phonotactic restriction. As we argued in the last paragraph, the acquisition of /ɾ/ seems to be hindered by its closest L1 category /l/; nevertheless, this L1 interference is only restricted to onset because /l/ is not licensed syllable-finally in Mandarin, which only allows nasals and a retroflex approximant in coda (Lin 2007), implying that Mandarin speakers might experience less L1 interference in coda position during the acquisition of the EP tap. The impact of L1 phonotactic constraint was evidenced by the diverse repair strategies (segmental replacement, epenthesis, deletion, metathesis) for the target /ɾ/ in coda, in comparison with onset.

The participants produced more target-like [ɾ] in branching (accuracy rate: 51.2%) than in non-branching onsets (accuracy rate: 39%), although this difference is not significant ( 2 (1) = 0.52, = 0.47). The predictor "word-level position" was found to have a significant effect ( 2 (1) = 4.55, = 0.033), indicating that the production accuracy was higher in word-medial (accuracy rate: 69%)

### Chao Zhou, Maria João Freitas & Adelina Castelo

than in word-initial branching onset (accuracy rate: 33%). Nevertheless, we attribute this word-level prosodic effect to an artefact of our experimental set-up, where the word-initial onset clusters are composed of a bilabial voiceless stop plus a tap ([pɾ]ato 'dish', [pɾ]eto 'black' and [pɾ]enda 'gift'), while two of the three word-internal onset clusters consist of a dental stop and a tap (es[tɾ]ada 'road', em[pɾ]esa 'company' and qua[dɾ]ado 'square'). In particular, we reason that the L1 interference which affects the L2 production accuracy was blocked in word-medial sequences [tɾ] and [dɾ], which were never mispronounced by the participants, due to an articulatory constraint. In particular, comparable to other syllable positions (e.g. non-branching onset), L1-Mandarin learners often realised the target tap as [l] in branching onset, due to CLI; however, such segmental repair in word-medial clusters would result in [dl] and [tl], which are consonantal sequences rarely attested cross-linguistically (Hallé & Best 2007) and the languages that currently allow these clusters are becoming less tolerant with them. For instance, Portuguese only has [tl] clusters in a few words (e.g. a[tl]ético; Mateus & D'Andrade 2000) and the tendency to avoid [tl] through rhotacism (e.g. A[tɾ]ético) was observed in Brazilian Portuguese (Cristófaro-Silva 2003) and in Angolan Portuguese (Miguel 2018). Therefore, it is likely that the relatively high accuracy of [ɾ] in word-internal onset clusters is due to the fact that the L1 interference (the use of [l]) cannot be applied to the sequences [dɾ] and [tɾ]. Future studies are suggested to take the articulatory constraint \*[dl]/[tl] into consideration when selecting test stimuli.

The EP /ʀ/ was accurately produced 80% of the time. The predictor "word-level position" did not have a significant effect ( 2 (1) = 0.65, = 0.42). The phonetic variants produced by the participants were quantified in Table 2.2, reminiscent of the production of the French /ʁ/ by L1-Mandarin learners (Steele 2002). This cross-linguistic evidence indicates that learners do not master all phonetic features simultaneously. In particular, they first target the manner feature, which is considered to be more salient, in comparison with place and voicing features (Colantoni & Steele 2008) and realize the EP /ʀ/ exclusively as a fricative. The only repair strategy attested was the production of [h], which could be explained by CLI. According to the prediction in §1.2, it seems that the EP /ʀ/ was processed by Mandarin speakers as the L1 category /x/, which alternates freely between [x] and [h] (Lin 2007). On the one hand, the velar realisation overlaps with a possible variant of the EP /ʀ/; on the other hand, the glottal realisation was regarded as a deviant production.


Table 2.2: Phonetic variants of the EP /ʀ/ produced by L1-Mandarin learners

### **4 Conclusion**

The current study contributed new experimental data to the literature on novel liquids acquisition. Results showed that not all EP liquids are equally difficult for L1 Mandarin learners, which is mediated by the relationship between L1 and L2 categories, as predicted by the SLM. Moreover, this paper has shown that the phonological development of /l/ and /ɾ/ was conditioned by syllable position, while the word-level position does not seem to play a decisive role. We reason that the syllable position effect stems from the relationship between L1 and L2 allophonic categories, in the case of /l/ and from the L1 phonotactic restriction, regarding /ɾ/. Future research should include a larger group of L1-Mandarin learners across different proficiency levels to gain a better understanding on how L2 phonological representations develop over time. Furthermore, both perceptual and production tasks are needed in order to reveal how different speech modalities interact in the L2 speech learning.

### **Abbreviations**


### **References**

Amorim, Clara. 2014. *Padrão de aquisição de contrastes do pe: A interação entre traços, segmentos e sílaba*. University of Porto, Portugal. (Doctoral dissertation). Andrade, Amália. 1998. Variação fonética do /l/ em ataque silábico em Português Europeu. In APL (ed.), *Atas do XIV Encontro Nacional da Associação Portuguesa de Linguística*, 55–76. Lisbon: Universidade de Lisboa.

Chao Zhou, Maria João Freitas & Adelina Castelo


### 2 On the acquisition of Portuguese liquids by L1-Mandarin Learners

*approaches to romance linguistics, ed. by julie auger, j. clancy clements and barbara vance. amsterdam*, 83–102. John Benjamins.


Chao Zhou, Maria João Freitas & Adelina Castelo


Chao Zhou, Maria João Freitas & Adelina Castelo


2 On the acquisition of Portuguese liquids by L1-Mandarin Learners


## **Part II**

## **Language teaching**

## **Chapter 3**

## **A close look at how context of acquisition of previous languages influences third language pedagogy: Does one model fit all?**

### Ana Carvalho

University of Arizona

A significant proportion of college students in the United States have Spanish in their linguistic repertoire. Several language programs capitalize on these students' bilingual skills to offer them the opportunity to develop proficiency in additional languages, especially cognate systems that can be acquired faster, such as French, Italian, and Portuguese. In this article, I first offer a synopsis of the research that has been developed on the acquisition of cognate languages in general, and the acquisition of Portuguese by Spanish speakers in particular. I then focus on the main premises that are considered in the curriculum designed to teach Portuguese for Spanish speakers in educational settings in the United States, pinpointing a tendency to treat Spanish speakers as a homogeneous group. Finally, I focus on incipient research that points to differences in patterns of L3 acquisition by learners who speak Spanish as their first, second, or heritage language, and problematize the assumption that they make up a homogeneous group with similar pedagogical needs.

While possible differences among L3 learners due to different contexts of acquisition of previous languages have been pointed out by previous research, only a few studies have investigated these dissimilarities. This article aims to contribute to this growing area of research by exploring important differences among English-Spanish bilingual learners of L3 Portuguese who acquired Spanish as their L2 or heritage language both in terms of their performance in class and in terms of their perception of the learning process through an examination of previous studies and incipient survey data. It concludes that while L2 Spanish speakers benefit more

Ana Carvalho. 2021. A close look at how context of acquisition of previous languages influences third language pedagogy: Does one model fit all? In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 49–67. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449770

### Ana Carvalho

from explicit teaching due to previous experience as language learners that lead to a higher level of metalinguistic awareness, heritage Spanish speakers are less fluent in metalinguistic terminology and an explicit understanding of grammar, and as such benefit from more implicit, naturalistic activities that rely on their intuition. Lastly, a number of important implications for curricular adjustments that cater for both types of Spanish speakers in L3 Portuguese classes are offered.

### **1 Introduction**

Given that an important population of English-speaking college students in the United States also speak Spanish, several language programs are capitalizing on these students' bilingual skills by offering them opportunities to develop proficiency in additional languages, especially cognate systems that can be acquired easily and quickly. In this chapter, after establishing the substantial presence of Spanish speakers in higher education in the United States, I offer a synopsis of current research on the acquisition of Portuguese as a third language (L3) by Spanish-English bilinguals, and of the main premises that are considered in the curriculum designed to teach these students. As I show, there is a generalized but inaccurate tendency to treat Spanish speakers as a homogeneous group. Incipient research points to differences in patterns of L3 acquisition depending on whether learners were born in Spanish-speaking countries (L1 Spanish speakers), learned Spanish as adults in school settings (L2 Spanish speakers), or were early bilinguals who were born and raised in the United States but learned Spanish in their homes and communities (heritage Spanish speakers). By pinpointing important differences in the way these groups acquire additional languages, I problematize the assumption that "Spanish speakers" enrolled in L3 Portuguese classes are a uniform group with similar pedagogical needs. Finally, I stress the need to create a more inclusive L3 pedagogy and suggest curricular adjustments that would cater to all types of Spanish speakers enrolled in Portuguese courses, including heritage Spanish speakers.

Although possible differences among L3 learners due to different contexts of acquisition of previous languages have been identified by previous research (Carvalho 2002; Cenoz 2011; Johnson 2004), only a few studies have investigated these dissimilarities (Carvalho & Child 2018; Carvalho & Silva 2006; Child 2014, among others). Thus, in this chapter I contribute to a growing area of research by exploring systematic differences among English-Spanish bilingual learners of L3 Portuguese who acquired Spanish as their L2 versus as a heritage language, in terms of both their performance in class and their perception of the learning process. I conclude that L2 Spanish speakers benefit from explicit teaching because

### 3 A close look at contexts of acquisition of previous languages

they have previous experience as language learners, whereas heritage Spanish speakers are less fluent in metalinguistic terminology and explicit understanding of grammar, and consequently benefit from more implicit, naturalistic activities that rely on their intuitive knowledge.

### **2 Spanish speakers in US higher education**

The United States has the fifth largest Spanish-speaking population in the world, after Mexico, Colombia, Spain, and Argentina (Escobar & Potowski 2015). This population has been growing steadily and more than 40 million inhabitants now claim Spanish as their home language (Figure 3.1), far exceeding speakers of other languages (Figure 3.2).

Figure 3.1: Spanish speakers in the United States. Original source: US Census Bureau 2011–2016

Figure 3.2: Languages other than English spoken in the US. Original source: US Census Bureau 2008–2016

### Ana Carvalho

While Spanish use is widespread throughout the United States, it is clear that the highest concentration occurs in the US south-west. In Arizona, the site of this study, more than 23% of the population speaks Spanish as their first language (Figure 3.3).

Figure 3.3: Map showing percentages of Spanish speakers in the US. Reprinted by permission of the Modern Language Association of America (www.mla.org), MLA Language Map, https://www.mla.org/ Resources/Research/MLA-Language-Map.

A fraction of these speakers attends the University of Arizona, which earned the designation of a Hispanic-serving institution (HSI) in 2018, when Hispanic enrollment in its undergraduate programs exceeded 25% (Figure 3.4).

Figure 3.4: Hispanic student percentages at the University of Arizona (Fall terms). Original source: University of Arizona's Fact Book

The University of Arizona's HSI status opens new opportunities to boost grants and research collaborations. In addition, and crucially, this new status

### 3 A close look at contexts of acquisition of previous languages

holds administrators and faculty – including language educators and applied linguists – responsible for addressing this population's needs, an important objective of the present study.

In addition to the students who acquire Spanish as their first or their heritage language, thousands of college students study Spanish as a second language. Spanish is by far the most popular language studied in schools; according to the latest Modern Language Association (MLA) 2016 report, half the students enrolled in foreign language courses are studying Spanish (Figure 3.1). Thus, a significant proportion of postsecondary students have Spanish in their linguistic repertoire, whether as a first, second, or heritage language.

These students' bilingual skills can accelerate their proficiency in additional languages, especially cognate systems such as French, Italian, and Portuguese. Spanish-English bilingual students who study cognate languages have an exceptional opportunity to become trilingual, in accordance with the Modern Language Association's recommendation that higher education should promote "speakers who have deep translingual and transcultural competence" (Ad hoc committee on foreign languages 2007: 7).

### **3 Portuguese for Spanish speakers**

In fact, the substantial presence of Spanish speakers in the US education system has prompted several initiatives to encourage Spanish-English bilinguals to learn additional languages. For example, Donato and her associates (Donato & Oliva 2015; Donato & Pasquarelli-Gascon 2015) have successfully implemented French and Italian language courses for Spanish speakers in high schools and colleges across greater Los Angeles. At the postsecondary level, the growth of Portuguese courses for Spanish speakers has driven a rapid increase in enrollment in Portuguese programs nationwide (Milleret 2012: 14). In a survey of postsecondary Portuguese programs in the United States, almost half (50 of 107 responding institutions offered a beginning-level Portuguese course specifically for Spanish speakers, Bateman & Oliveira 2014). These courses capitalize on the fact that cognate languages can be acquired rapidly and efficiently. As Wiedemann (2009) showed, Spanish speakers on average can learn Portuguese in half the time as English monolinguals because they have a high level of receptive skills from the beginning. Spanish speakers with no previous knowledge of Portuguese can understand more than 50% of what is said in standard Portuguese (Jensen 1989). Written words are even more transparent. Henriques (2000) found that monolingual Spanish speakers could comprehend up to 94% of the content

of academic texts written in Portuguese, due to the very high degree of lexical similarity between the languages.

Thus, positive transfer translates into advanced receptive skills, which lessens students' affective filter, decreases their anxiety level, increases their motivation to learn a cognate language, and serves as an effective recruiting strategy to attract bilingual students into adding another language to their repertoire relatively quickly. These advantages are, however, counterbalanced by a predisposition to non-felicitous transfer, since cross-linguistic transfer occurs at all levels of the grammar and is believed to induce early fossilization of an inter-


Table 3.1: Language (other than English) enrollments and percentage change to the previous date as reported by the MLA. Original source: MLA Report 2016

*a* Includes enrollments reported under "Arabic", "Arabic Algerian", "Arabic Classical", "Arabic Egyptian", "Arabic Gulf", "Arabic Iraqi", "Arabic Levantine", "Arabic Modern Standard", "Arabic Moroccan," "Arabic, Qur'anic," "Arabic, Sudanese," and "Arabic, Syrian."

*c* Includes enrollments reported under "Hebrew, Biblical," "Hebrew, Classical," and "Hebrew, Rabbinic." Excludes enrollments reported under "Hebrew" and "Hebrew, Biblical and Modern."

*b* Includes enrollments reported under "Greek, Ancient," "Greek, Biblical," "Greek, Koine," "Greek, New Testament," and "Greek, Old Testament." Excludes enrollments reported under "Greek," "Greek and Hebrew," and "Greek and Latin."

### 3 A close look at contexts of acquisition of previous languages

language because students are able to communicate basic meanings early in the learning process (Simões & Kelm 1991; Takeuchi 1984; Carvalho et al. 2010, among others). Even though non-facilitative and facilitative transfer processes are interconnected, and positive transfer is a crucial facilitating factor in the acquisition process, pedagogy has emphasized combating non-facilitative transfer (Carvalho 2002; Carvalho & Child 2018). In fact, the most common pedagogical techniques aim at developing metalinguistic awareness, an approach considered to support control of multilingual processing (Jessner 2006: 106). Through both contrastive analysis and focus on form, students are expected to learn to discern subtle but important differences and similarities between cognate languages, and crucially, to capitalize on the similarities while avoiding transfer of the differences. Thus, most curricula available in the United States emphasize cross-metalinguistic awareness by comparing the languages based on the learners' declarative knowledge of Spanish and their baseline knowledge of Portuguese.

Three textbooks available in the United States are designed for Spanish speakers who are learning Portuguese (Simões 1992; 2010; Bateman et al. 2016); there are also several online resources, including a podcast<sup>1</sup> . All these resources emphasize differences and similarities between the languages so that the student can use this knowledge to learn by analogy and generalization. Focusing on formal differences, these materials target classroom practice and pedagogical strategies that emphasize metalinguistic explanations and contrastive discussions that elicit declarative knowledge. The facilitation of metalinguistic awareness is believed to help learners recognize degrees of crosslinguistic relationships, capitalize on similarities, and avoid differences (cross-linguistic interference).

As Portuguese for Spanish speakers courses and instructional materials multiplied, curriculum developers and linguists began to engage in research to elucidate the particular processes involved in learning cognate languages, in order to inform teaching practices and curriculum design. Several symposia were organized, leading to the publication of selected proceedings (Simões et al. 2004; Wiedemann & Scaramucci 2008) that accompanied a call for research in a field that was heavily based on contrastive analysis (Carvalho 2002). Over the past 15 years experimental research on English-Spanish bilinguals' acquisition of L3 Portuguese has flourished (Allegro 2010; Bailey 2013; Feiden et al. 2014; Silva 2015; Trude & Tokowicz 2011, among others). In all cases, research results corroborate that as Spanish-English bilinguals learn Portuguese, transfer from Spanish is inevitable. This body of research yields to two broad generalizations. First, linguistic overlap is the strongest predictor of transfer, as predicted by Rothman's

<sup>1</sup>https://www.coerll.utexas.edu/brazilpod/tafalado/

### Ana Carvalho

(2010) typological premise model. Second, it presupposes that bilinguals are a homogenous group of learners, regardless of the different contexts in which they acquired their previous languages, a generalization that Carvalho (2002) and Cenoz (2011) have questioned. As Cenoz (2011: 80) states, it "may be a mistake not to be aware of the important difference between both types (active bilinguals and foreign language users) or to ignore the implications of dealing with one situation rather than the other." In fact, acquisitionist studies that follow the formal tradition have identified different transfer patterns which correlate with the order of acquisition of previous languages (Cabrelli Amaro & Wrembel 2016; Child 2017; Silva 2015; Giancaspro et al. 2015; Rothman 2010, among others). Meanwhile, Carvalho & Silva (2006), Carvalho & Child (2018), and Koike & Flanzer (2004) have analyzed how Spanish speakers' backgrounds influence how they learn the L3. More specifically, these authors argue, with Cenoz (2011), that it is important to consider the distinctions between L2 Spanish speakers' previous experience of learning a language in a school setting versus heritage Spanish speakers' experiences of acquiring Spanish in naturalistic environments. These differences have direct implications for teaching methods, as I discuss below.

### **4 L3 Portuguese acquisition by speakers of Spanish as L1, L2, and Heritage Language (HL)**

A brief survey at the University of Arizona, a large public university in the US Southwest, revealed that almost half the students enrolled in Fall 2018 Portuguese classes were heritage Spanish speakers who reported being "exposed to Spanish as a child in their household" (Figure 3.5, adapted from Sommer-Farias et al. 2020). The second largest group was L2 Spanish speakers, or students who learned Spanish in a school setting (28.3%), followed by L1 Spanish speakers, or students who were born in a Spanish-speaking country, in this case, typically Mexico. Less than 5% of students did not speak Spanish. Similar tendencies are evident in other universities in the US Southwest that offer Portuguese for Spanish speakers (see, for example, Milleret 2012).

The large percentage of students in Portuguese classes who speak Spanish as their heritage language has important implications for L3 curriculum development, in light of what is known about this population's language learning. Like first languages, heritage languages are primarily acquired in early childhood through an implicit, unconscious, automatic, and naturalistic process (Zyzik 2016). This process does not involve explicit knowledge, which is conscious, declarative,

### 3 A close look at contexts of acquisition of previous languages

and accessible through controlled processing (Bowles 2011). While heritage learners' previous experience with their home language puts them in clear advantages compared to L2 learners, their lack of the explicit metalinguistic knowledge that is commonly evoked in foreign language classrooms places them at a disadvantage. Studies have shown that they:


These particularities of heritage language learners have direct consequences for how third languages should be taught to students who acquired both previous languages implicitly – particularly because the teaching of cognate languages

### Ana Carvalho

focuses heavily on form and metalinguistic awareness, both of which are believed to minimize negative transfer and early fossilization.

In fact, in one of the first studies designed to distinguish different types of Spanish speakers in the classroom, Carvalho & Silva (2006) applied the think aloud protocol and retrospective interviews to explore differences between L1 and heritage Spanish speakers versus L2 Spanish speakers in learning the Portuguese subjunctive. While it is well known that the proficiency levels of heritage speakers vary substantially, the authors collected data from a pool of undergraduates enrolled in first semester Portuguese for Spanish speakers, a course whose prerequisite for all students is that they have completed two years of college-level Spanish. Their results clearly showed that L2 Spanish speakers consciously applied their explicit knowledge of grammar, whereas L1 and HL Spanish speakers tended to apply intuitive knowledge, in the forms of analogy and generalization. Examples 1 and 2 (from Carvalho & Silva 2006: 192–194) illustrate typical L2 behavior:

(1) não sei o opuesto de "ar" es "er" so "estés com raiva" que é subjuntivo 'I don't know, the opposite of "ar" is "er" so "you are (subj) mad" which is subjunctive'

In a follow-up interview, the participant explained her use of a learning strategy during the activity:

(2) [I used this verb] because sometimes we can use the present when speaking about the future or the past, I remembered a Spanish class.

Carvalho and Silva found that, in contrast, L1 Spanish speakers tended to apply intuitive reasoning. See Example 3, in which an L1 Spanish speaker explains why she picked the present subjunctive:

(3) creo que no estaba pensando solo estaba usando la intuición. 'I think I wasn't thinking but only using my intuition.' (Carvalho & Silva 2006: 192–194).

Both qualitative and quantitative data led the authors to conclude that L2 Spanish speakers in the Portuguese classroom tended to rely on explicit learning strategies, whereas L1 (and heritage) Spanish speakers favored implicit strategies. This tendency was later confirmed by Child (2014), who analyzed mood selection among three groups of participants (L1, L2, and HL Spanish speakers) based

### 3 A close look at contexts of acquisition of previous languages

on grammatical judgment and fill-in-the-blank tasks. He initially tested participants' knowledge of Spanish subjunctive, then after 10 weeks of instruction, tested Portuguese subjunctive. Even though L1 and HL speakers scored higher on the Spanish subjunctive test (their use of Spanish subjunctive was more nativelike than the L2 speakers' was), they did not score as high as their L2 Spanishspeaking classmates on the Portuguese subjunctive test, even after receiving the same amount and type of instruction. These results led Child to conclude that higher metalinguistic awareness helped L2 Spanish learners to capitalize on positive transfer of rule-based strategies.

Furthering the search for differences among bilinguals learning L3 Portuguese, Koike & Gualda (2008) analyzed how Spanish-English bilingual students acquired possessive forms when taught by implicit versus explicit methods. Pre- and posttest results revealed differences in the performance of L1, L2, and heritage Spanish speakers, depending on the type of instruction they received. The authors concluded that L2 Spanish speakers tended to do best with explicit instruction, whereas the other two groups showed less progress.

In fact, some incipient research suggests that students themselves perceive parts of the grammar that require declarative knowledge more or less difficult, depending on their linguistic background. Based on a set of questions about what is easy versus hard to learn in Portuguese, Child (2013) found the tendencies shown in Tables 3.2 and 3.3.


Table 3.2: Response percentages regarding the easiest learning aspects of Portuguese as indicated by the three groups of bilinguals ( = 108). Original source: Child 2013

In Child's sample, students who acquired Spanish as a second language found it easier to acquire declarative knowledge (grammar and conjugation), exactly opposite to heritage speakers of Spanish, who found this the most difficult aspect of learning Portuguese.

Students currently enrolled at the University of Arizona who learned Spanish as adults confirm the faciliatory role that declarative knowledge of grammar

### Ana Carvalho

Table 3.3: Response percentages regarding the most confusing learning aspects of Portuguese as indicated by the three groups of bilinguals ( = 108). Original source: Child 2013


plays in L3 Portuguese acquisition. As part of a program evaluation, students filled out surveys of their perceptions and attitudes about their experience with the Portuguese program. Crucially, one question asked them to describe how Spanish influenced their acquisition of Portuguese. The overwhelming majority of students believed that Spanish helped them to learn Portuguese. L2 Spanish speakers, in particular, often pointed to their previous experience learning Spanish as being very helpful. Examples (4–6) illustrate students' insights about how their experience learning Spanish helped them learn Portuguese.


Somewhat surprisingly, a few heritage Spanish speakers reported that their knowledge of Spanish hindered their Portuguese acquisition (7).

(7) I grew up speaking both English and Spanish so to learn another language will always be a little difficult. (Spring 2019 – Final Survey)

Other heritage speakers confirm that the emphasis on grammar interferes with their learning process (8–9).

### 3 A close look at contexts of acquisition of previous languages


These students' comments about their perception of the role of Spanish in their acquisition of Portuguese in the classroom confirm Carvalho & Child's (2018) claim that students with previous formal language training benefit from current approaches to L3 Portuguese teaching that emphasize declarative knowledge and metalinguistic awareness. Activities designed to combat negative transfer through metalinguistic discussions followed by completion tasks presuppose not only linguistic knowledge but also – and crucially – training in performing language-learning-related tasks. While these activities benefit L2 Spanish speakers learning L3 Portuguese due to what Odlin (1989: 34) called "transfer of training", their emphasis on declarative knowledge places heritage Spanish speakers at a disadvantage. Less focus on metalinguistic discussions and more emphasis on creative tasks, on the other hand, would capitalize on heritage learners' resources such as their advanced reading and listening skills, great familiarity with cognate words, and ease navigating across languages, to name only a few.

### **5 Conclusions and implications**

In sum, the evidence presented thus far led Carvalho & Child (2018) to conclude that, in general, L1 Spanish and heritage Spanish bilinguals tend to rely on intuitive, non-rule-based strategies, such as analogies, generalizations, and even explicit avoidance, whereas L2 Spanish bilinguals favor rule-based, explicit strategies. Therefore, it stands to reason that L2 Spanish bilinguals benefit from explicit instruction and feedback, whereas L1 and heritage bilinguals may benefit more from implicit instruction techniques.

Research aimed at identifying how L3 students with various language acquisition experiences may benefit from different pedagogical treatments is incipient and critically needed. However, the evidence that L2 learners prefer explicit teaching whereas heritage language learners benefit more from intuitive methods recalls Cenoz's (2013: 73) analogy that learning how to walk is comparable to

### Ana Carvalho

learning our first language, learning how to drive a car is comparable to learning a second language (since it does require declarative knowledge), and learning a third language is equivalent to learning how to drive a bus. Addressing bilinguals' unquestionable advantage in learning a third language, she claims,

The experience of driving a car, despite involving different skills and strategies, can nevertheless be extremely useful when driving another type of vehicle: the starting point is not the same as for an absolute beginner. (Cenoz 2013: 73)

Given the importance of how previous languages were acquired for L3 acquisition, a slight change to this analogy is apropos. In light of the fact that heritage speakers learned both their previous languages naturalistically during childhood, they have not received explicit instructions that aimed at declarative knowledge. In other words, they do not have the experience of "learning how to drive" before tackling their L3, since "learning how to drive" here involves a process that is equitable to "learning a second language" in a school setting. Continuing with this analogy, because heritage speakers learned both English and Spanish in naturalistic settings, they might have learned how "to walk" or simply move from one place to another in two ways, neither of which required explicit instruction.<sup>2</sup> Thus, it is not productive to teach these bilinguals how to drive a bus (speak an L3) by referencing the skills of driving a car (declarative knowledge).

These important differences provide essential information for curriculum developers who are choosing the type of instruction that best benefits all speakers of Spanish. Relatively intuitive, content-based teaching capitalizes on heritage speakers' linguistic and pragmatic repertoire and facilitates their acquisition of Portuguese as an L3. In fact, Koike & Flanzer (2004) supported the premise that heritage Spanish speakers are more likely to transfer implicit knowledge to their production of Portuguese. Analyzing the production of speech acts in Portuguese as the third language of Spanish-English bilinguals, these authors found that due to pragmatic rules shared by Spanish- and Portuguese-speaking communities but not by English-speaking communities, heritage learners outperformed L2 Spanish speakers.

Therefore, it is important for curriculum developers to consider that emphasizing explicit knowledge and metalinguistic awareness in teaching a cognate language may be appropriate to the learning needs of L2 Spanish speakers. Heritage

<sup>2</sup>As one of the reviewers rightly pointed out, the lack of metalinguistic knowledge among college students in the USA is not particular to heritage speakers, but to English monolinguals as well, given the absence of explicit grammar instruction in American K-12 education.

### 3 A close look at contexts of acquisition of previous languages

speakers, on the other hand, are more likely to benefit from activities that capitalize on their implicit linguistic knowledge. For example, Carvalho et al. (2010) presented online reading and listening activities with authentic texts that heritage speakers can easily comprehend due to a highly congruent lexical inventory and a set of grammatical structures that they have already internalized. In these activities, learners are encouraged to comprehend the meaning first, before deriving grammatical structures through paying attention to form as they answer comprehension questions. Likewise, corpus-based activities, such as those contained in the *Multilingual academic corpus of assignments: Writing and speech* (MACAWS), enable students to search for authentic examples of language use and derive patterns of use of specific features from those examples. Finally, the incorporation of multilingual and multimodal artifacts and activities, such as projects involving analysis and replication of trilingual landscapes and artistic production would not only foster multilingual reflection and awareness, but also resonate with heritage speakers' bilingual experiences. The development of empirical studies that would test the efficacy of such an approach to teach additional languages to heritage speakers is imperative and crucial for the building of a credible pedagogy that caters to different types of L3 learners.

Such an approach would take into account the consensus among scholars of heritage language pedagogy that whereas students learn an L2 through grammar, heritage language learners learn grammar through their heritage language. Torres (2013) clearly showed that heritage language learners focus primarily on the content of the task, usually being concerned with interpreting the meaning of prompts rather than analyzing them metalinguistically. Zyzik extends Torres's results, claiming that

while classroom L2 are intimately familiar with exercises that ask them to fill in the blanks, transform sentences, and replace underlined forms, HL learners will take a communicative task at face value, that is, as an opportunity to comprehend or communicate a message using any combination of grammar and vocabulary they have at their disposal. (Zyzik 2016: 121)

By incorporating approaches that capitalize on the potential for meaning making as opposed to explicit knowledge, L3 instructors, curriculum developers, and material designers could practice an inclusive L3 Portuguese pedagogy that caters to all types of Spanish speakers.

### **Abbreviations**

HSI Hispanic-serving institution

### Ana Carvalho

### **References**


3 A close look at contexts of acquisition of previous languages


3 A close look at contexts of acquisition of previous languages


## **Chapter 4**

## **Multilingual teachers, plurilingual approach and L3 acquisition: Interviews with multilingual teachers and their L3/L3+ students**

### Emel Kucukali

Marmara University

The article reports a study on plurilingual approaches (PA) of three multilingual teachers in three foreign language classrooms (English, German and Russian) at a Turkish state university. The qualitative study explored the attitudes of multilingual teachers to PA and the response of their L3/L3+ students ( = 5) to the plurilingual practices in class. The perceptions of participants were investigated from the holistic paradigm in multilingualism. Data were collected through interviews and graphic elicitation tasks. Data were analyzed through content and visual-based analysis. The results indicated three main themes based on participants' perceptions of PA: (1) implementation, (2) benefits to learning and (3) the role of multilingual teachers in L3 Acquisition. Both teachers and students displayed positive attitudes in findings (1) and (2). However, participants' perceptions were not homogeneous in finding (3), which revealed some issues. The overall findings suggest that multilingual teachers should play a crucial role in plurilingual and L3 classrooms in a Turkish context.

### **1 Introduction**

With the increase of immigration and globalization, multilingualism has attracted the attention of language researchers and educators (Cenoz & Genesee 1998; Jessner 1999; Cenoz 2013a). Research on multilingualism has created a new multilin-

Emel Kucukali. 2021. Multilingual teachers, plurilingual approach and L3 acquisition: Interviews with multilingual teachers and their L3/L3+ students. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 69–93. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449774

### Emel Kucukali

gual perspective which adopts a holistic philosophy and considers all the individual's languages as an integrated whole (Cook 1992; Grosjean 2008; Cenoz 2013a). In addition, the multilingual perspective distinguishes between how monolinguals and bi-/multilinguals learn languages or between L2 learners and L3/L3+ learners (Hufeisen 2004a). This classification resulted in new pedagogical applications some of which are called plurilingual approaches (Council of Europe 2001; Cenoz 2013a,b; Otwinowska 2014). Plurilingual approaches (PA hereafter) challenge the isolation of languages, focus on learners' plurilingual repertoires and promote the integration and diversity of languages in the classroom (Beacco et al. 2010).

As a pedagogy, PA are built on a holistic view on multilingualism (Grosjean 2008; Cenoz 2013a,b) and are suggested as a practice in multilingual education (Cenoz 2009). Therefore, research on PA focuses mostly on learners and schools in bi- and multilingual contexts of immigrant and minority communities (Cenoz 2013b; García & Li 2014). For this reason, more research is suggested on PA in different contexts with a different type of students (Cenoz 2013a,b; Cenoz & Gorter 2017). Furthermore, teachers and their language repertoire are defined as a significant variable in multilingual education (Cenoz 2009). That is why research on multilingual teachers and the influence of their languages on their teaching is also recommended (Ellis 2013). Moreover, research in the Turkish context is not enough to give a picture of PA, third language acquisition (TLA) and multilingual teachers in Turkish education. With an attempt to give light to these issues, the present study will focus on PA of three multilingual teachers in three foreign languages (FL hereafter) classrooms at a Turkish state university. The aim of the study is to explore the attitudes of multilingual teachers to PA and the response of their L3/L3+ students to plurilingual practices in class through a qualitative perspective.

### **2 Literature review**

### **2.1 Holistic view on multilingualism**

One of the generic definitions of multilingualism is "the command and/or use of two or more languages by the respective speaker" (Herdina & Jessner 2002: 52). However, apart from the generic definitions above, multilingualism could be approached from two different philosophies. They are challenging each other and are called the *holistic* (multilingual) and the *monolingual* view on multilingualism (Grosjean 2008). While the monolingual view claims that multilinguals are the sum of multiple separate monolinguals, the multilingual view proposes

### 4 Multilingual teachers, plurilingual approach and L3 acquisition

that multilinguals have a unique language system of multi-competencies. Furthermore, the multilingual view suggests that multilinguals differ with regard to language-related cognitive processing because their multi-competencies are integrated and domain-specific (Cook 1993; Grosjean 2008). Depending on the interlocutors or domain, multilinguals may shuttle between mono-, bi-, tri- or quadrilingual speech modes along a situational continuum. In the monolingual mode, bilinguals speak to monolinguals in either of their languages, while in the multilingual speech mode they speak to multilingual interlocutors by mixing multiple languages in the form of code-switching (Grosjean 2008).

### **2.2 Third language acquisition vs. second language acquisition**

The literature adopting a holistic view on multilingualism proposes that there is a difference between second language acquisition (SLA) and TLA (Jessner 1999; Herdina & Jessner 2002; Hufeisen 2004a; Marx & Hufeisen 2004; Cenoz 2013a,b).

The factor model (Hufeisen 2004a) explains the difference between SLA and TLA with the cognitive leap between the learning of the first (L2) and the second foreign language (L3). L3 learners learn differently from L2 learners because the former have the cognitive and linguistic experience of learning another foreign language during SLA. After their experience with the first foreign language (L2), L3 learners have upgraded to a significantly higher metalinguistic and metacognitive level. The following stages of learning the subsequent languages (L3+) also contribute to cognitive leaps afterwards but with little significance. In other words, while there is little cognitive distinction between L3 and L3+ learners, the gap between L2 and L3 learners is crucial due to a significant cognitive transformation between SLA and TLA (Hufeisen 2004a).

Similarly, Cenoz (2013b) suggests that TLA is different from SLA because L3 learners reactivate and relate all their languages and adapt strategies to TLA from previous learning experiences. SLA focuses on the learning of a specific language in separation, while bilingualism, multilingualism, and TLA are unified under the umbrella of involving the additional languages of multilinguals in the learning process (Cenoz & Gorter 2011).

### **2.3 Plurilingual approaches**

The holistic paradigm in multilingualism has been realized in education as pedagogical applications called plurilingual approaches (PA) (Council of Europe 2001; Beacco et al. 2010; Cenoz 2013a). PA challenge isolation of languages and promote the integration and diversity of languages in the classroom because languages

### Emel Kucukali

support and contribute to each other to build up a flexible and dynamic competence. Therefore, plurilingual approaches focus on learners' plurilingual and intercultural repertoires, raise learners' language awareness and give equal value to all languages and cultures (Beacco et al. 2010). The Council of Europe has already promoted PA in education to meet the diverse and democratic character of the European population (Council of Europe 2001; Beacco et al. 2010).

PA are defined as "making use of the learners' first and other languages to teach more effectively" (Otwinowska 2014: 102). They cover pedagogies such as alternation of languages, conscious cross-linguistic comparison and transfer, metalinguistic awareness strategies, code-switching and using cognates (Council of Europe 2001; Beacco et al. 2010; Herdina & Jessner 2002; Hufeisen 2004b; Cenoz 2013b; Jessner et al. 2016). Otwinowska (2014) also suggests the using of English as a bridge between learners' native language and their other languages by making use of similarities for better language learning.

### **2.4 Learners' metalinguistic awareness and teachers' plurilingual awareness**

Metalinguistic and cross-linguistic awareness are properties of a multilinguals' metasystem which involves "specific meta-skills" (Herdina & Jessner 2002: 129). These meta-skills contribute to the bi- and multilinguals' cognitive system and give them an advantage over monolinguals in language learning. Metalinguistic awareness "refers to the ability to focus attention on language as an object in itself or to think abstractly about language and, consequently, to play with or manipulate language" (Jessner 2006: 42). Cross-linguistic awareness is defined as the awareness of the relationships between languages (Jessner 2006: 116).

Teachers' plurilingual awareness is more compounded than that of learners and refers to "the complex ability to promote plurilingual approaches in the language classroom" (Otwinowska 2014: 103). Teachers' plurilingual awareness has three main components: cross-linguistic and metalinguistic awareness of similarities and differences between the language taught and the learners' L1, L2 and L3/L3+; knowledge about adopting a plurilingual approach in the classroom, which refers to training learners to search for similarities across the languages they know; psycholinguistic knowledge of individual learner differences that facilitate learning such as, knowing that bilingual learners differ from multilinguals in the process of language learning (Otwinowska 2014).

### 4 Multilingual teachers, plurilingual approach and L3 acquisition

### **2.5 Previous research**

Research on PA in TLA is conducted mainly in plurilingual, minority and immigrant contexts. It focuses mostly on the learners but less on the teachers and even less on both of them.

Research on learners covers the benefits of PA and prior language knowledge on L3 learning. The positive effects of PA include raising learners' metalinguistic awareness (Jessner et al. 2016) and facilitating language learning (Bono & Stratilaki 2009). These benefits are enhanced when the school languages are also used in social interactions in family and community or in school contexts which promote plurilingual practices and cross-linguistic strategies (Bono & Stratilaki 2009). Similarly, learners' prior language knowledge has a positive effect on third language learning both in plurilingual (Bono & Stratilaki 2009; Cenoz 2013b; Sánchez 2015) and formal instruction contexts (Kemp 2001; Gibson & Hufeisen 2003; De Angelis 2007). In addition, Dmitrenko (2017) found a significant correlation between the learners' degree of multilingualism and the use of multilingual strategies of language learning. The main difference between L2 and L3/L3+ learners was the preference of cross-linguistic strategies (Dmitrenko 2017).

Research on multilingual teachers and PA is little but valuable. It indicates that multilingual teachers have higher plurilingual awareness (Otwinowska 2014) and higher metalinguistic awareness (Ellis 2004; 2013) than their bi- and monolingual counterparts. On the one hand, compared to monolingual ones, multilingual teachers of English as a Second Language (ESL) in Australia are more likely to adopt cross-linguistic practices in their classrooms (Ellis 2004; 2013; Higgins & Ponte 2017) such as using knowledge of other languages at the level of phonology, lexis, grammar, syntax, discourse, and pragmatics to understand the target language and help students with their challenges (Ellis 2004; 2013). On the other hand, teachers in some European contexts have positive attitudes towards PA (Griva et al. 2016) but make little use of plurilingual strategies in L3 teaching classes (Göbel & Vieluf 2014).

In one of the few studies (Aronin & Ó Laoire 2003) in the field of TLA regarding students and their teachers, the findings indicate that trilingual students from Israel and Ireland prefer their teachers of L3 to be trilingual as well even if the teacher only teaches one language. Trilingual students prefer their trilingual teachers to use the three languages of the classroom for facilitating and practical reasons (Aronin & Ó Laoire 2003).

Research on TLA, PA and multilingual teachers in Turkish contexts is very scarce. There have been initiatives to integrate plurilingualism in education (Mirici 2008), however Turkish instructors of English reported no familiarity with or

### Emel Kucukali

training concerning plurilingualism (Çelik 2013). In terms of TLA and L3 learners, Korkmaz (2013) found that English Language Teaching (ELT) university students used their L2 (English) to make a lexical association with L3 (German or French).

### **2.6 Research questions**

For the purpose of the present paper the following research questions have been proposed.


### **3 Method**

### **3.1 Participants**

Three FL classrooms (English, German and Russian) at a Turkish state university were selected through purposeful sampling as a rich source of data (Patton 2002). Seven female participants, two of which were multilingual teachers and five of which were L3/L3+ students, were purposefully selected from these three classrooms. Teachers had at least ten years teaching experience of FL and had proficiency in three or more languages at least at B1 level according to CEFR (Council of Europe 2001). The researcher is the third multilingual teacher from the Russian class and that is why she was not interviewed. Three students are learning English and German as L3, while two students are learners of Russian as L3+. The previously learned/acquired languages of the students are at least at A2 level according to CEFR. Participants' proficiency level of their languages was self-rated. Participants' type of multilingualism was classified into "active bilinguals" and "foreign language users" (Cenoz 2013b: 82). Active bilinguals were participants who grew up, learned and experienced their languages in minority or plurilingual contexts. They were exposed to their languages on a daily basis, use them actively and demonstrate a multilingual behavior of code-switching. The foreign

### 4 Multilingual teachers, plurilingual approach and L3 acquisition

language users had acquired their FLs at school through formal instruction. They had limited exposure to their foreign languages and display monolingual practices (Cenoz 2013b: 82). To keep participants' identity confidential, codes were used instead of their names (see Table 4.1 for detailed information about the participants).

Table 4.1: Sample characteristics. *Abbreviations*. TL: Target language taught/learned in class; TM: Participants' type of multilingualism; N: Native-like; AB: Active bilingual; TeRe: Teacher/Resarcher


*<sup>a</sup>*Codes used in the study instead of participants' name

### **3.2 Data collection methods**

Data were collected in March 2016 at a Turkish state university and data collection lasted three weeks. Qualitative tools like semi-structured interviews (Fraenkle & Wallen 2009) and graphic elicitation tasks (Bagnoli 2009) were used to collect the data. Interviews were conducted in English and twice with each participant. TE, SR1 and SR2 participated in the interviews orally, while TG, SG1, SG2 and SE1 preferred to give written answers in their classrooms. The oral interviews lasted about an hour each and were audio recorded and then transcribed. Interview questions were developed by the author and aimed to elicit background data and participants' perceptions of PA and the role of multilingualism in L3/L3+ teaching/learning. Right after the first interview, TE, TG, SR1 and SR2 completed a visual task (Bagnoli 2009). Participants were asked to express visually their attitudes, feelings and the atmosphere in the classroom by freely drawing and writing on a sheet of paper (see the Appendix for the results of the elicitation tasks).

### Emel Kucukali

### **3.3 Data analysis**

Interview data were analyzed through structural, in vivo (Saldaña 2013), open and theory-driven coding (De-Cuir Gunby et al. 2011) by two independent coders. The first reading was accompanied by open and in vivo coding to elicit the key concepts from the raw data. In the second reading, the concepts from the open coding were categorized under codes by means of structural and theory-driven coding. The codes from the visual data were elicited through visual-based and text-based analysis contextual with the interviews (Bagnoli 2009). After consistency was checked, the inconsistent codes from interview and visual data were negotiated between coders and modified. Then, the final codes from verbal and visual data were compared, integrated and classified under themes to answer RQs. The final codes were classified under three themes which constructed the main findings and the answers to the RQs.

### **3.4 Validity and reliability**

The trustworthiness of the current study was supported by triangulation, member checking, and an inter-coder reliability analysis. Triangulation was used to verify the findings through multiple data sources (Miles et al. 2014) such as verbal and non-verbal data: interviews (Fraenkle & Wallen 2009) and graphic elicitation tools (Bagnoli 2009), respectively. In addition, the data form three classrooms, teachers and students were compared (Miles et al. 2014). To reinforce reliability, inter-coder reliability analysis using the Kappa statistic was performed to determine consistency between two independent coders of the data (Landis & Koch 1977). The intercoder reliability for 11 codes of interview data was found to be = 0.82 (Sig = 0.000, < 0.001), and for 10 codes of visual tasks was found to be = 0.80 (Sig = 0.000, < 0.001), which is a significant result and considered to be a substantial agreement between two coders (Viera & Garrett 2005). What is more, in order to clarify the interpretations of the data, the interview records and findings were sent to the participants for member checking (Sandelowski 1993).

### **4 Findings**

Data analysis has resulted in three main themes based on participants' perceptions of PA: (1) implementation, (2) benefits to learning and (3) the role of multilingual teachers in TLA context. The following sections will display the three main findings through both teachers' and students' perspectives obtained from the interviews and graphic elicitation tasks.

### **4.1 Implementation of PA**

The first theme and answer to RQ1 were built on two codes: "PA practices" and "intuitive implementation of PA". The codes confirm that PA are practiced by multilingual teachers in their classrooms and represent the plurilingual strategies and the systematicity they are used with, respectively.

### **4.1.1 Teachers' perspective: Interviews**

Both teachers claim to apply PA in their FL classrooms. They report to use at least three languages in the classroom and compare and contrast them to teach the target language (TL hereafter),

I use many similarities and differences when I teach…I used not only Turkish but also Italian, French, even Spanish. (TE)

In my lessons I always make use of the similarities or differences among the languages. Especially the differences, which can cause very simple transfer errors, are the main points, which I emphasize in my lessons. (TG)

Moreover, they give examples of in-class activities on grammar, vocabulary and cognates.

… during reading and also when we are focusing on grammar usages, I used many cognates when explaining English. (TE)

I make use of similar syntax, grammar or lexic between German and English. For example, comparing tempus present perfect with perfect. (TG)

When asked about the frequency and systematicity of the plurilingual activities, teachers reported unplanned and intuitive use of their strategies:

It is something I decide on the spot, whenever it is appropriate. (TE)

I did not have the opportunity to assess the results of the techniques, which I am using *instinctively* during my lessons. (TG)

### Emel Kucukali

### **4.1.2 Teachers' perspective: Visuals**

Teachers' drawings (see the Appendix, Figures 4.1–4.2) of their classrooms support the verbal findings above and also reveal that teachers adopt PA in their classrooms. TE's drawing displays group work activities with English-German-Turkish cognates. Moreover, TE has drawn multiple arrows, and has labeled multiple languages around and words like "cognates", "associations" and "multilingual examples" to express the intensive interaction via multiple languages. Similarly, TG used arrows and speech bubbles in multiple languages implying cross-linguistic communication in class. She has also used arrows labeled with the phrase "grammar+lexis" to symbolize grammatical and lexical transfer. Moreover, TG has written concrete phrases and examples related to cross-linguistic transfer such as "negative transfer", "positive transfer" and "who – wer".

### **4.1.3 Students' perspective: Interviews**

Students confirmed the plurilingual strategies used by their teachers. They said that in addition to their L1 (Turkish) and TL, their teachers use and compare other foreign languages. For example, according to the students from the German class, German and English are compared but Turkish is also integrated:

My teacher in course compares English with German. (SG1)

Apart from German, which languages does your teacher use in the classroom? (R)

English, Turkish. (SG2)

Multiple languages are integrated in English class as well:

She (teacher) speaks German, Italian, Spanish, French and uses them. She gives example for lecture's topic. (SE1)

The students from the Russian class reported that their teacher incorporates Turkish, English and Russian to teach Russian:

She were giving lessons with three languages. She uses Turkish for teaching me Russian, she gives me an English example. (SR2)

When I don't understand something in Russian, I do understand it in English or sometimes in Turkish so, we had examples in three languages. (SR1)

### **4.1.4 Students' perspective: Visuals**

Students' visuals also support the verbal findings related to RQ1. SR2's drawing implies the plurilingual practices of her teacher through speech bubbles of different students in the class, namely, "she (teacher) uses English, Russian and Turkish", "she (teacher) uses Turkish" and "she (teacher) uses English". SR1's drawing is not displaying plurilingual practices explicitly, but she used the phrase "Lots of examples" which might refer to multi-modality and variety of methods in the classroom.

### **4.2 Benefits of PA to learning**

The second finding is the answer to RQ2 and RQ3 and embraces the attitudes of teachers and the response of their students to PA practices. Participants displayed positive attitudes which are represented by two codes "facilitating effect of PA" and "autonomous learners". The codes indicate the perceived benefits of PA on students' learning.

### **4.2.1 Teachers' perspective: Interviews**

According to teachers, their students benefit emotionally and mentally from PA. Students look happier and more motivated when exposed to variety of languages because they can see the similarities between the TL and the languages they already know. They also remember and understand easier through interesting and cross-linguistic examples. Both TE and TG observed PA as beneficial and facilitating for their students:

In a positive way because I gathered many positive feedback from my students. Different examples sometimes are easier to remember. They (students) love when I use different accents. (TE)

I have the feeling through my observations that students are getting happier and more motivated when they notice that this new language is not much harder to learn than the previous one. (TG)

The other perceived benefit for students is that they become more autonomous learners by adopting and developing cross-linguistic learning strategies. According to teachers, students raised their cross-linguistic awareness by comparing languages just like their teachers. Both of the teachers have noticed that their students also started to make associations across similar languages they know and to guess words through cognates by themselves:

### Emel Kucukali

They have started to use cognates during the lesson, without my influence. They understand that it comes from a certain language looking at the root or the word and they start to make the associations themselves. (TE)

After a couple of explicit examples and comparisons between both languages English and German, students are trying to find out some other similarities by themselves. This attitude could be perhaps a sign of raising awareness. (TG)

### **4.2.2 Teachers' perspective: Visuals**

Teachers' drawings also reflect PA benefits in the classroom. TE's drawing (see the Appendix, Figure 4.1) displays an effective learning atmosphere with labels like "flipped environment", "social context" and "autonomous learning" expressed through a speech bubble "students use cognates". Similarly, TG has used a speech bubble "student cooperation in English and Turkish" in the group work with a Chinese student. TG explained how beneficial and motivating for students was to help and communicate with foreign students in multiple languages during group work activities. In addition, both teachers have drawn group works, and arrows between students and teacher implying the high energy and motivation in the classroom.

### **4.2.3 Students' perspective: Interviews**

Students' reactions to PA confirmed their teacher's claim that it was emotionally and mentally helpful for them in learning TL. Russian learners mentioned that they enjoy PA and learn or understand better thanks to this approach,

This method helped me a lot in learning vocabulary. I loved the courses, I understood everything, her system worked really good for me. (SR1)

Lessons are going enjoyable. We understand well because English and other languages for multilingual students. (SR2)

German learners expressed the cognitive effects of PA,

This create a mind mapping and I find it easy to learn. (SG1)

We can learn the lesson better … quickly. (SG2)

### 4 Multilingual teachers, plurilingual approach and L3 acquisition

Finally, the English learner described PA as "useful for our education" because her teacher uses "different ways" and "variety methods".

Similar to their teachers' perceptions, students also reported a positive effect of PA on their autonomous learning. In the same way, students claimed to have raised awareness due to plurilingual practices of their teachers. The students stated that they compare languages more or more consciously after experiencing their teacher's method. For instance, some students became more aware of languages similarities,

Teacher gave examples I hadn't noticed before. After that, I was more attentive in finding the similarities. (SR1)

Sometimes we cannot realize these similarities between English and German vocabulary. Teacher makes us realize that similarities. (SG2)

Other students developed plurilingual strategies to learn:

After my English teacher, I learned the technics. I use them more often because of the awareness I got from my teacher. (SE1)

I use this method more. These words remind me other words in other languages. (SG1)

What is more, SR2's higher cross-linguistic awareness has improved her reading skills,

Before I started lessons with my Russian teacher, I wasn't using her method ... it increases my reading comprehension skill. (SR2)

The common plurilingual strategy "picked up" by students seems to be the using of cognates from their foreign language repertoire. When asked to give examples of the similarities between languages, all of the five students listed cognates such as *Brat, sestra*–*brother, sister*, *Broun*–*brown*, *colores*–*colors*. Regardless of their TL, all students used English as a bridge between another Indo-European language such as German, Russian or Spanish. Students' L1 Turkish was not used in their examples.

### **4.2.4 Students' perspective: Visuals**

The mental and emotional benefits of PA were expressed in SR2's visual task via students' speech bubbles, namely, "we are successful", "happy classroom", "beneficial". Similarly, SR1 has placed herself in the center of the picture surrounded by arrows, which represent her feelings in the classroom. The arrows were labeled with positive emotions like "good attitude", "fun", "happy" and "attention". To express her mental challenges, SR1 labeled an arrow as "difficult subject". However, she seems to have overcome the challenges because she has also "confident", "encouragement" and "perseverance" labeled arrows to support her.

### **4.3 The attitude to multilingual teachers teaching in TLA context**

The third finding is the answer to RQ4 and covers participants' attitudes to *multilingual* teachers in TLA. Participants' perceptions were not homogeneous and were categorized into two codes "Teacher's attitudes" and "Students' attitudes" representing teachers' and students' perspective, respectively. Teachers, two L3+ students and one L3 student displayed a positive attitude to employing multilingual teachers in TLA context. However, two L3 learners revealed neutral attitudes to multilingual teachers and favored bilingual teachers. In addition, the visuals revealed some issues of mono- and bilingual students when taught by multilingual teachers in class.

### **4.3.1 Teachers' perspective: Interviews**

According to teachers, L3/L3+ should be taught by multilingual teachers who are proficient in students' L1, L2 and L3/L3+. TE believes that multilingual teachers "would be useful" for PA because they can associate the TL with learners' previously learned languages. TG's argument is that multilingual teachers can understand and analyze the L3 learning process:

Multilingualism can absolutely affect the learning process of foreign languages very positively. A multilingual teacher is capable of understanding the learning process, so that she can guess easily, where the obstacles can be in learning a new FL and can analyze and develop techniques against them. (TG)

### **4.3.2 Teachers' perspective: Visuals**

The visual tasks indirectly reveal teachers' positive attitude toward multilingualism in teaching. Their multilingualism is implied by symbols like multilingual

### 4 Multilingual teachers, plurilingual approach and L3 acquisition

speech bubbles and "transfer" and "cognate" labels. Moreover, while explaining the multilingual examples in her drawing, TE claimed to feel very "comfortable" in plurilingual teaching contexts due to her experience with plurilingual practices in her multilingual family,

I love it … because I am used to it. I've been brought up bilingual…we spoke three languages at home. (TE)

TG's drawing also implied the suitability of multilingual teachers in L3/L3+ teaching. She added a note in the corner of her picture which indicated "language problems" between her German class and a partner teacher who lacked competency in English. She explained that, when teachers' languages are fewer than or do not overlap with their students' languages, using TL only could not be enough and even lead to communication problems in TLA classrooms.

### **4.3.3 Students' perspective: Interviews**

All students appreciate their teachers' multilingualism and associate it with diverse knowledge and high competence. However, when asked to compare multilingual teachers with bi- and monolingual teachers, the perceptions of students differed. While two L3+ and one L3 student showed strong preference for multilingual teachers, two L3 learners from the German class did not differentiate between multi- and bilingual teachers. For example, L3+ learners prefer multilingual teachers to bi- and monolingual ones, because they expect multilingual teachers to use plurilingual methods and compare at least three languages in class. L3+ students do not favor mono- and bilingual teachers because there are few languages and the classes are "boring", "difficult" and "not beneficial",

I don't think I could learn Russian from a monolingual native teacher. Explaining in Russian is not easy because Russian is difficult. She (bilingual teacher) would be better than monolingual native teacher but only Russian and Turkish, it will be difficult for me. It (with multilingual teacher) is the best way for me to learn a language. We had three choices in the class. (SR1)

(with monolingual teacher) I would have been bored and scared … it's not enjoyable, I don't understand. With bilingual teacher we were OK, but my multilingual teacher gives me also Turkish, also English … This is advantages for us. (SR2)

### Emel Kucukali

The L3 student from the English class also preferred multilingual teachers to mono- and bilingual teachers because they can integrate more cultures and languages in the classroom,

If I had a monolingual native teacher, it wouldn't be beneficial for me because … cannot teach us other cultures. We can't association between other cultures. Bilingual English teacher can combine two language … but a multilingual English teacher could combine a lot of languages. (SE1)

However, the L3 students from the German class revealed a neutral attitude to multilingual teachers. For example, SG1 did not differentiate between bi- and multilingual teachers and did not consider multilingual teachers more beneficial.

(with monolingual teacher) I can't communicate easily because I wouldn't ask any questions to him/her… Bilingual teacher is the same (with multilingual) for me…We can learn German from the teacher who isn't multilingual teacher. (SG1)

SG2's statements about monolingual and multilingual teachers are not clear but she definitely appreciates bilingual teachers,

(bilingual teacher) can explain similarities/differences between both languages. We can understand subject better. (SG2)

### **4.3.4 Students' perspective: Visuals**

Similarly, both positive and negative attitudes to multilingual teachers have been captured through the visual tasks. In SR1's picture, multilingual teacher is appreciated with labels like "knowledge", "understanding" and "patience". In SR2's drawing, students' faces are smiling below speech bubbles saying "She is a good teacher", "I love my teacher", "and I understand all". Also, multilingual teacher is perceived as interdisciplinary and competent with a speech bubble "I would like to teach you everything". However, SR2 has also drawn one unhappy face of a student saying that her teacher "uses English but I don't understand". This means that when students' and multilingual teachers' languages do not overlap, PA may not be meaningful and even become distressful for students, which SR2 explained,

If someone doesn't know English or other languages … this is disadvantage for them because they won't understand. (SR2)

4 Multilingual teachers, plurilingual approach and L3 acquisition

### **5 Discussion**

The findings of the present study were consistent with the literature. Finding (1) that even though not trained, multilingual teachers used PA intuitively could be related to the shifting behavior of multilinguals along the situational continuum (Grosjean 2008). Multilingual teachers switch to multilingual speech mode and mix the languages when working with multilingual students. Finding (1) is also in line with previous research findings that multilingual teachers adopt crosslinguistic practices (Ellis 2004; 2013; Higgins & Ponte 2017) and display high plurilingual awareness (Otwinowska 2014). In addition, multilingual teachers from the present study made use of cross-linguistic transfer on a lexical and grammatical level which remind the strategies of multilingual teachers in Australia (Ellis 2004). Also, in the present study, participants gave mainly cognates as crosslinguistic examples and used English as a bridge between their foreign languages. These practices overlap with plurilingual strategies suggested by Otwinowska (2014) and used by Turkish ELT students to learn L3 (Korkmaz 2013). However, differently from the present study, teachers in the European context claimed to make little use of plurilingual strategies in L3 teaching classes (Göbel & Vieluf 2014). The inconsistency may derive from the difference in educational contexts and teachers' backgrounds.

Finding (2) is in line with previous research, which indicates raising learners' metalinguistic awareness (Jessner et al. 2016) and facilitating language learning (Bono & Stratilaki 2009) as benefits of PA. The difference with the present study is that teachers are also perceived as the source of students' awareness, not only the approaches used. Teachers as models of awareness are also suggested in the second component of plurilingual awareness which refers to the ability of teachers to encourage and train learners to search for similarities across the languages (Otwinowska 2014).

The data in finding (3) disclosed both benefits and some potential issues of multilingual teachers in L3/L3+ teaching. According to teachers' data, multilingual teachers are suitable for multilingual students, while mono- and bilingual teachers could have communication problems with them. On the other hand, according to students' data, mono- and bilingual students may not understand the languages of multilingual teachers. This discrepancy could be explained with the different level of metalinguistic awareness, the number and proficiency level of the languages of the participants, and therefore, different learning strategies preferred by them (Cenoz 2013b; Jessner et al. 2016; Dmitrenko 2017). The fact that students' problem was not reflected in teachers' data, means that multilingual teachers and their PA may not be appropriate for all the students in class and

### Emel Kucukali

multilingual teachers may not be aware of that. Teachers should approach their students' languages and metalinguistic level individually and adopt more flexible and embracing strategies. This issue has also been considered in the third component of teachers' plurilingual awareness as teachers' psycholinguistic knowledge of differences between bilingual and multilingual learners in the process of language learning (Otwinowska 2014). The finding that L3+ students preferred and appreciated their multilingual teachers is consistent with previous findings (Aronin & Ó Laoire 2003) and depicts the scenario when teachers' and students' linguistic repertoire overlap.

Another possible explanation of finding (3) is that students' preferences to multilingual teachers might be related to their own and their teachers' type of multilingualism, that is, being an active bilingual or foreign language user (Cenoz 2013b: 82). Both parties, the students and the teacher of the Russian classroom are active bilinguals with childhood bilingualism and they code-switch on a daily basis. The learner of English is a foreign language user but her teacher is an active bilingual, who is also with childhood bilingualism and reported to feel natural and comfortable when mixing languages. The students from Russian and English class, who are either active bilinguals themselves or have an active bilingual teacher, reported to prefer multilingual to bi- and monolingual teachers. On the other hand, both students and the teacher of the German classroom are foreign language users and those students display neutral attitudes to multilingual teachers. Being aware of the small sample of the present study, it could be assumed that participants' previous exposure to genuine multilingual behavior in their family, community and the classroom, might make them feel more comfortable with the plurilingual practices of their multilingual teachers. This is in line with the literature (Bono & Stratilaki 2009) indicating that when the school languages are also used in social interactions in family and community or in school contexts plurilingual practices and cross-linguistic strategies are promoted.

From the discussions above it can be concluded that the present study is in line with literature in terms of the holistic view on multilingualism, high plurilingual awareness of multilingual teachers, the cognitive difference between L2, L3 and L3+ learners, the influence of genuine multilingual interactions, and the benefits of PA. Unlike the previous research, the present study focused on the interaction between multilingual teachers and their students from both teachers' and students' perspective. As a result, the dual perspective and the integration of visual methods revealed some deeper and sensitive issues in classrooms with multilingual teachers.

4 Multilingual teachers, plurilingual approach and L3 acquisition

### **6 Conclusion**

The study aimed to explore the attitudes of multilingual teachers towards PA and the responds of their L3/L3+ students to plurilingual practices. The findings revealed that multilingual teachers use PA in class, which was reported to contribute to students' cross-linguistic awareness and more autonomous learning of FL. Multilingual teachers and students, and one bilingual student have positive attitudes to the employment of multilingual teachers in L3/L3+ classrooms. Teachers believe that multilingual teachers are necessary for multilingual students and, mutually, students appreciate multilingual teachers, as well. However, the discrepancies in data revealed some issues. First, mono- and bilingual teachers might have communication problems with multilingual students. Similarly, mono- and bilingual students may face language barriers because they may not understand all languages used by multilingual teachers in class. Also, two bilingual students, who have not been involved in genuine code-switching interactions in family, community and school contexts, do not find multilingual teachers more beneficial and favor bilingual teachers. Therefore, multilingual teachers are perceived as necessary and beneficial in TLA contexts but they should take their students' languages and level of metalinguistic awareness into account and adjust their approaches accordingly.

The main limitation of the present study is that it is not supported by longitudinal data and a larger sample to gain generalizable findings. In addition, qualitative data could have been triangulated with class observations, recordings and quantitative data. Also, not oral but written interviews (James 2007) were administered with four participants and visual task were not completed by all participants.

The suggested implications are giving priority to multilingual teachers in teaching and curriculum design in school contexts of third, additional, minority or immigrant languages. In addition, teacher educators should include programs like plurilingualism, multilingualism and teaching of more than one foreign language in teacher education.

For further research, investigations in different contexts with different students of different proficiency levels and degree of multilingualism are suggested. In addition, the focus on different teacher and student profiles such as mono- vs. multilingual, active multilinguals vs. foreign language uses for comparison are recommended.

### **Acknowledgements**

I would like to express my very great appreciation to Dr. Talar Kaloustian for her valuable feedback during the planning and development of this research work. I would also like to thank Marmara University for their support and enabling me to collect my data.

### **Abbreviations**

L3 Third language

PA Plurilingual approaches SE1 Student 1 learning English SG1 Student 1 learning German

FL Foreign language L2 Second language SG2 Student 2 learning German SLA Second language acquisition


### **Appendix: Graphic elicitation tasks**

L3+ All subsequent languages after L3

Figure 4.1: English teacher (TE)

### 4 Multilingual teachers, plurilingual approach and L3 acquisition

Figure 4.2: German teacher (TG)

Figure 4.3: Student 1 (SR1)

Emel Kucukali

Figure 4.4: Student 2 (SR2)

### **References**


## **Chapter 5**

## **Debunking student teachers' beliefs regarding the target-language-only rule**

### Pierre-Luc Paquet

University of Texas at Tyler

### Nina Woll

University of Quebec at Trois-Rivieres

Despite the recognized usefulness of crosslinguistic interaction, additional language (LN) teachers seem to be resistant to opening this Pandora's box to more multilingual classroom practice. While the ultimate goal of the broader research project is to promote student teachers' reflective stance concerning the potential benefits of crosslinguistic pedagogy, this study examines their beliefs regarding the TL-only rule established in their respective programs. Forty student teachers from Quebec and Mexico completed a vignette-based questionnaire where they were asked to reflect on different situations focused on the use of other languages, including L1. Results suggest that while a monolingual bias is prevalent in the Quebec context, participants from Mexico appear to be more open to resorting to other languages in the classroom. However, responses from both populations suggest that this perceived usefulness is restricted to situations where L1 is used as a vehicle rather than a resource. Suggestions for addressing these issues in teacher education are explored and avenues for further research are provided.

### **1 Introduction**

Despite the unprecedented rise of crosslinguistic pedagogy around the globe, many teacher-training programs are still constrained by a monolingual bias. Beyond any doubt, learners need maximal exposure to the target language (TL), especially when access is limited. Meanwhile, this so-called "direct method assumption" is largely responsible for an ideology of language separation that has been

Pierre-Luc Paquet & Nina Woll. 2021. Debunking student teachers' beliefs regarding the target-language-only rule. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 95–116. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449776

### Pierre-Luc Paquet & Nina Woll

prevalent in additional language (LN) pedagogy since the rise of the communicative era (Cummins 2007). However, to ban other languages from the classroom might reduce the benefits of active reflection on crosslinguistic correspondences. Throughout the last two decades, the quest for European intercomprehension has given rise to whole curricula aimed at cultivating crosslinguistic awareness (Council of Europe 2001). In North America, there have been numerous attempts to challenge these deep-rooted assumptions, namely to acknowledge the dynamics of multilingual societies in LN classrooms. For instance, Duff (2007) refers to concrete examples of classroom projects intended to foster multilingual diversity, namely to counter what she calls "assimilationist policies in Canadian schools [that] lead to subtractive bi- or multilingualism, where French or English are privileged exclusively at the expense of students' other languages" (p. 153). Accordingly, investigations into teacher beliefs regarding multilingual practices have also become more widespread. Studies conducted in different parts of the world revealed that despite the perceived usefulness of crosslinguistic classroom activities known or shown to teachers, the dominant discourse favoring exclusive TL use was not easily overridden (Arocena Egaña et al. 2015; Haukås 2016; Martínez et al. 2015). In contexts where teacher training programs that are not conceived with a view to fostering multilingual proficiency, it may thus be expected that monolingual ideology is prevalent among LN teachers, especially in contexts where the TL-rule is invoked by policy makers and educators. The goal of this study is to examine student teachers' beliefs regarding strictly monolingual vs. multilingual practice in LN classrooms in Canada and Mexico. More specifically, it addresses their perceptions of resorting to L1 or other languages and the presumed reasons for this use.

To accomplish this goal, we have taken a different approach to the narrative by weaving the trajectories of two researchers' experiences with studies in multilingualism and LN instruction. We use these personal experiences to help the readers, especially those working and learning in different teacher training programs, reflect on their own beliefs about multilingual teaching and learning, and the beneficial effects of incorporating these pedagogical practices.

### **2 The story**

To situate the theoretical foundations from which the present study has emerged, it was deemed relevant to refer to the personal and professional trajectories of two researchers whose stories might reflect that of other scholars and teachers in the field of LN pedagogy. It's the story of two young professionals who have

### 5 Debunking beliefs regarding the target-language-only rule

started academic careers in LN didactics in different parts of the world far away from their respective homes. One of the many things they had in common was that they were leading lives with multiple languages. One grew up in the Frenchspeaking province of Quebec and the other in Germany and both had learned several non-native languages at school. After years of training in higher education, both were now stranded in other countries where they used two languages at work none of which was their native language. Another common trait was that they were both trained language teachers, who had acquired experience in teaching LNs in their respective countries of birth, as well as their native languages abroad.

Throughout their teaching career, their ever-growing curiosity led them to pursue graduate studies. In the end, they both acquired PhDs in psycholinguistics, with a focus on crosslinguistic influence. More specifically, they were interested in the positive influence of previously acquired languages on the acquisition of an additional language. When they discovered the field of third language acquisition (TLA), they were both mesmerized. As they plunged into the work of scholars such as Jasone Cenoz and Britta Hufeisen, it was as if someone had explained and theorized their own trajectories as language learners.

### **2.1 Findings from third language acquisition**

Specifically, while Cenoz' work on "additive" multilingualism (Cenoz 2003) was a milestone that inspired researchers to examine the beneficial effects of learning additional languages, Hufeisen's (2000) *factor model* gives an extensive overview of the distinctive features of TLA. In particular, third language learners can draw on their knowledge of previously acquired LN, while related language learning experiences and strategies are likely to enhance cognitive factors such as metalinguistic awareness.

Although social/affective and contextual factors are also taken into consideration by this model, its main components are psycholinguistic in nature and have inspired researchers to investigate language development with a focus on the positive effects of learning and using more than two languages (e.g., Peyer et al. 2010).

What follows from these studies is that crosslinguistic interaction occurs naturally between all the languages of a multilingual. Especially when languages are typologically related, learners are led to make assumptions about underlying correspondences in function as well (Ringbom 2007). Moreover, findings reflect the premises made by the dynamic model of multilingualism (Herdina & Jessner 2002), which views multiple language learning as a dynamic process depending

### Pierre-Luc Paquet & Nina Woll

on numerous factors related to each individual learner's history. Specifically, the constant interaction between the different languages of a multilingual generates new structures that are specific to each speaker and different from monolingual systems. In this sense, the dynamic model is consistent with Cook's (1992) notion of multicompetence. What is specific to Herdina & Jessner's (2002) conceptualization of multilingual proficiency is that when lacking some relevant linguistic knowledge, the multilingual's metalinguistic abilities can make up for it. Not only were these predictions confirmed in subsequent research, but they also reflected the personal experiences of the two young scholars mentioned earlier. Namely, throughout their own trajectories as language learners, they became aware of their growing knowledge base in several languages, thus increasing their adaptability to complex communicative situations. Essentially, the conscious manipulation of the whole repertoire is likely to lead to increased levels of metalinguistic awareness, which in turn facilitates further learning (Jessner 2017). On the whole, this line of research emphasizes the benefits of learning and using multiple languages instead of focusing on the obstacles.

### **2.2 Crosslinguistic pedagogy**

When it comes to pedagogical approaches that build on these findings, the field has also been flourishing. Again, much of what the two young professionals had experienced as language teachers was reinforced by the literature on the practical implementation of crosslinguistic pedagogy in the LN classroom. Especially in Europe, changes at the policy level have generated models of language education aimed at fostering intercultural and plurilingual competence (Candelier 2007). The main focus of projects such as Eurocom (Hufeisen & Marx 2007) is to foster awareness of correspondences across languages. A number of comparative approaches such as *focus on multilingualism* (Cenoz & Gorter 2014) have been developed to tap into the repertoire of learners from various linguistic and cultural backgrounds, namely to help them discover the rich knowledge base that they can build on when learning additional languages (i.e., Leonet et al. 2020). In addition, recent developments in Canada, especially crosslinguistic awareness pedagogy (Horst et al. 2010) and plurilingual and pluricultural tasks (Galante et al. 2019; 2020), have also been a major inspiration to the two young professionals who had both been hired in teacher training programs.

In sum, years of personal and professional engagement in the learning and teaching of multiple languages, paired with the growing awareness of a community of practice in which multilingual practices are not only considered acceptable but encouraged, have grown into a feeling of empowerment for two

### 5 Debunking beliefs regarding the target-language-only rule

educators whose individual careers in language didactics and teacher training has yet to be built.

### **3 The problem**

Picture these same two young professionals, who are obsessed with language and equipped with a toolbox full of pedagogical material ready to be adapted to various linguistic and cultural contexts. Full of inspiration for practical applications that draw on the rich theoretical basis for crosslinguistic pedagogy, they are ready to share their obsession with future teachers.

And that is when they hit a wall. In fact, the student teachers were not ready. Based on our current knowledge, only a handful of studies have explored LN student teachers' and teachers' beliefs about multilingual pedagogy (Arocena Egaña et al. 2015; Creese & Blackledge 2010; De Angelis 2011; Haukås 2016; Woll 2020). It has already been established that beliefs influence teachers' pedagogical decisions, and that such beliefs are ingrained and resistant to change (Phipps & Borg 2009). Be that as it may, results of these studies are quite similar in two ways: (1) these participants think multilingualism should be promoted, but not necessarily in their own classrooms, (2) they do not feel competent enough at doing so, which according to them may be detrimental to their students' language learning (De Angelis 2011). These results indicate that experienced teachers do not feel comfortable using multilingual practices, and if that is the case, how can we expect student teachers to be open to this discussion?

Whether it is called translanguaging (García & Li 2014) or crosslinguistic interaction (Jessner 2008), multilingual practices in LN classrooms are at variance with the "monolingual principle" (Cummins 2007) that appears to be prevalent in both contexts of study. For example, anecdotal evidence from interactions with student teachers in Quebec revealed that using the TL exclusively basically meant doing their job properly. To resort to the learners' native language (L1), however, was perceived as a failure. Even if no generalizations can be drawn from such scattered statements, they reflect the "anti-L1-attitude" identified by Cook (2001) as a "mainstream element in twentieth-century teaching methodology", thus leading teachers to "feeling guilty for straying from the L2 path" (p. 405). Furthermore, Cummins pointed out that

Despite the continuing academic debate on these issues, policy and practice operate as though the "monolingual principle" had been established as axiomatic and essentially "common sense" (Cummins 2007: 224).

### Pierre-Luc Paquet & Nina Woll

A closer look at the Quebec education program provides further evidence for this monolingual orientation:


Even if there is no explicit policy of language separation in Mexico, evidence from teachers and students suggests that the monolingual principle may also reflect the dominant discourse in higher education in this part of the world. For example, Mora Pablo et al. (2011) speak of an "implicit policy" whereby teachers are "told not to use L1 and this practice continues even when the prescribed methodology is no longer in use" (p. 121).

Despite this limited amount of evidence from both educational contexts, the two professionals, who had only recently started working as teacher educators in Quebec and Mexico respectively, found themselves confronted to a similar problem: Their enthusiasm regarding crosslinguistic classroom practices and their willingness to make student teachers take a critical look at these written and unwritten norms were generally met with hostility.

Students indicated that they had trouble believing the opposite of what they had always thought to be best practice. This perceived resistance may be related to what they had been told by other professors or practicum supervisors at the university, to personal experiences during their practicum or to their own experience as language learners. In particular, research on the so-called "apprenticeship of observation" (Lortie 1975) suggests that student teachers often "fail to realize that the aspects of teaching which they perceived as students represented only a partial view of the teacher's job" (Borg 2004: 274).

### **3.1 Research questions and hypotheses**

While the ultimate goal of the broader research project is to promote student teachers' reflective stance regarding the potential benefits of crosslinguistic classroom practice, the first step to be undertaken was to examine their beliefs regarding the monolingual bias stipulated in their program of study. In particular, with a view to confronting the apprenticeship of observation, to question deeplyrooted beliefs and move beyond their limits, this study sets out to examine the student teachers' personal and professional opinions underlying those beliefs.

### 5 Debunking beliefs regarding the target-language-only rule

More specifically, to examine student teacher beliefs regarding the target-language-only rule, the present study aims at answering the following two research questions:


While the first question aims at providing an overview of factors which, according to the participants, might justify resorting to L1 or other languages, the second question aims at describing personal belief profiles. As for RQ1, certain acceptable conditions for using other languages may be prevalent within or across samples, thus pointing to possible anchor points to pursue pedagogical reflection. In other words, the most common factors could be systematically addressed and critically discussed within university programs. Regarding RQ2, a monolingual bias is expected in both student populations, yet teachers from Quebec might be more fervent promoters of the TL-only-rule given the explicit guidelines in their program of study. Overall, the anticipated results are expected to reflect concrete situations on which student teachers' beliefs are based. Finally, the reported events that appear to motivate student teachers' beliefs have the potential of informing teacher trainers on avenues for challenging the apprenticeship of observation and creating new learning opportunities.

### **4 Methodology**

### **4.1 Participants**

Twenty student teachers in Quebec aged between 22 and 52 and twenty student teachers in Mexico aged between 22 and 40 participated in this study. All forty participants were enrolled in a LN teacher-training program, more specifically in English ( = 27) and Spanish ( = 13). Subjects were recruited from four different universities, two of which were located in Quebec and two in Mexico. While the large majority of the participants were enrolled in undergraduate programs ( = 33), some participants were from different graduate programs ( = 7). All participants had at least 80 hours of classroom-based face-to-face experience teaching LNs. Moreover, all participants were bi- or multilingual learners and teachers.

Pierre-Luc Paquet & Nina Woll

### **4.2 Vignette-based technique**

To elicit participants' perceptions of the use of other languages than the TL in the classroom, this study used a vignette-based technique. Jeffries and Maeder defined vignettes as

Incomplete short stories that are written to reflect, in a less complex way, real-life situations in order to encourage discussions and potential solutions to problems where multiple solutions are possible. (Jeffries & Maeder 2005: 18).

According to Simon & Tierney (2011), using vignettes reduces defensiveness of responses which helps capture complex thought processes and stimulate critical thinking, even of sensitive information. In the present study, vignettes were used to tap into not only the participants' learning but also their teaching experience.

To select the focus of the vignettes, Mora Pablo et al. (2011) was considered, which explored the presumed reasons for resorting to L1 in an LN classroom in Mexico. This way, the extraction of linguistic situations where using other languages was qualified as acceptable for LN teachers and teacher trainers in Mexico. As for the Quebec context, the study relied on the researcher's classroom experience as teacher trainers and practicum supervisors and on the monolingual stance from the Quebec education program. Three different settings were selected, all of which were salient in both Mexico and Quebec, namely using other languages for pedagogical purposes, for establishing rapport and discipline, and for clarification. Afterwards, different real-life situations were created that are relevant to participants' experience as language learners and student teachers, and that allow them to reflect on this sensitive topic without feeling threatened or judged. Half of the vignettes presented the situations from a learner's point of view and the other half from a teacher's. More specifically, when creating the vignettes, the goal was to describe real-life situations that LN teachers and learners face during their learning and/or their professional career and in which they have to come to a decision about a potential conflict of beliefs regarding strictly monolingual vs. multilingual practice.

To design the vignettes, the general guidelines indicated by Simon & Tierney (2011) were followed. First, the vignettes were developed considering the participants' profile and experience to make them relevant and to maintain their interest. For this reason, the present study included some challenging situations, which most participants could relate to and even had previously experienced in their practice. Second, the vignettes did not exceed 200 words. As Stravakou &

### 5 Debunking beliefs regarding the target-language-only rule

Lozgka (2018: 1189) stated "[…] the provision of less information in hypothetical scenarios favors the personal elements of participants to come to the surface". That is, vignettes needed to be purposefully incomplete in order to allow for multiple solutions and to elicit participants' critical thinking. Finally, as recommended by Simon & Tierney (2011), the current study contained a reasonable number of vignettes ( = 4) and made sure that participants would take no more than 30 minutes to complete the online questionnaire (Dörnyei 2010) – see (1) and (2) for examples.

	- Sarah is a French as a second language teacher in an English-speaking high school. When it comes to presenting aspects of the French grammar, she sometimes has difficulty to make herself understood when giving explanations on those elements exclusively in the target language (French). So, when she notices that her students are not following her explanations, she instinctively changes from French to English in order to assure a better understanding. This way, she moves forward to other aspects more quickly and students seem to better understand.

The vignettes were followed by two different prompts. The first prompt requested participants to reflect on the situation described and to offer advice to the main character of the scenarios, the learner or teacher respectively. In this manner, participants adopted the role of consultants, which was central since the use of other languages is a sensitive topic to discuss with student teachers in both contexts. This way, it made the participants feel more comfortable revealing something closer to their true opinion without the feeling of being judged. The second prompt invited participants to further explain their thoughts specifying the knowledge, experiences and beliefs that may have influenced their responses.

### Pierre-Luc Paquet & Nina Woll

### **4.3 Data analysis**

To tackle the first research question, the content of all responses was analyzed thematically. Two research assistants were hired as independent raters. First, the entire data set was analyzed to identify instances in which participants referred to situations perceived as acceptable for resorting to other languages, including L1. The first coding phase constituted a clustering of recurrent patterns. Then, in a second coding phase, these were regrouped into six factors related to conditions that would justify resorting to other languages. In the last coding phase, the authors regrouped these factors into three main categories: learner factors, pedagogical strategies and practical constraints. Finally, the general trends within each category were interpreted. The frequency of each response type was listed for each context (see Table 5.1).

With a view to answering the second research question, responses were analyzed for each participant separately. First, the remaining responses, in which no justification for resorting to other languages was implied, were coded for instances where exclusive TL-only use was explicitly promoted. Depending on the recurrence of statements reflecting TL-only promotion or openness to other languages respectively, participants were classified into three belief profiles: (1) hardline-TL-only, (2) open-to-other-languages and (3) multivoicedness. The first profile corresponded with participants who showed resistance towards the use of other languages across the board, even when responding to vignettes presenting cases where using other languages seemed beneficial. Second, participants who manifested openness to using other languages in all three categories analyzed above were classified open-to-other-languages. The third classification referred to profiles where participants expressed two apparently contradicting positions within or across responses. Namely, while asserting the need for maximal TL exposure, participants in this belief profile would concurrently point to the benefit of using other languages. Instead of viewing these perspectives as plain contradictions, Bakhtin's (1981) notion of *multivoicedness* seemed more appropriate to refer to the competing views that were shown to coexist within teachers' accounts of their beliefs on various topics (Ball & Warshauer Freedman 2004). Namely, multivoiced discourse was tangible in the present study, when student teachers drew on apparently opposing ideologies to express opinions on language choices. Finally, having identified belief profiles across the sample, the frequency of profile type was also listed for each context of study (see Table 5.2).

### 5 Debunking beliefs regarding the target-language-only rule

### **5 Results and interpretation**

### **5.1 Conditions for using other languages**

The main purpose of the current study, where vignettes were implemented, was to explore student teachers' beliefs regarding strictly monolingual vs. multilingual practice in LN classrooms. This issue was particularly interesting for research since student teachers, from both contexts, seem to feel constrained by a monolingual bias. The aim of the first research question was to examine in what classroom situations the use of L1 or other languages is perceived as acceptable or even useful for student teachers. Looking at all responses, a total of 123 comments related to the present issue were obtained. All comments were then classified into one of three broad categories: Learner factors ( = 42), pedagogical strategies ( = 50) and practical constraints ( = 31). Table 5.1 provides the distribution of student teachers' responses listed in the respective categories.


Table 5.1: Conditions for using other languages

### **5.1.1 Learner factors**

As for the first category, two factors were highlighted from our participants' responses: proficiency level and anxiety. Regarding proficiency level, participants seemed to be more inclined to resorting to other languages when teaching beginner levels. This can be seen in the following excerpt:

*Considero que solo se debe de hacer uso del L1 como un recurso en niveles principiantes para garantizar la comprensión y también para optimizar tiempos.*

### Pierre-Luc Paquet & Nina Woll

'I consider that it is acceptable to use the learners' L1 as a resource only in beginner levels in order to guarantee their comprehension and to maximize time' (Participant 925 (mx), our translation)

This comment suggests that it is acceptable to resort to other languages with lower proficiency learners since they have neither the knowledge nor the strategies to understand the teacher's explanations in the TL. With reference to the learners' stress and anxiety, student teachers seem to perceive the use of other languages, mainly the use of the learners' L1, as beneficial and a powerful tool to lower their students' stress and anxiety. This can be seen in the following excerpt:

*Lorsque le ou la professeur(e) fait référence aux deux autres langues assez bien maîtrisées des élèves du groupe pour initier de la nouvelle matière en espagnol, les élèves ressentent moins de stress, font des liens et ont des références sur lesquelles se fier pour évoluer dans leur apprentissage.*

'When the teacher refers to other languages known by the students, they feel less stress, make links and have some references to rely on in order to progress in their learning' (Participant 842 (qc), our translation)

In this excerpt, the participant states that using other languages not only helps to lower anxiety, but also provides learners with the opportunity to make links between the various languages they know. L1 is described as a tool that helps with understanding language structures but more importantly as a means to reduce anxiety. Regarding this first category, learner factors, it is important to mention that there is a difference between the participants from Quebec (13 comments) and those from Mexico (29 comments). Where few comments were made by the Quebec student teachers in relation to proficiency level, participants from Mexico seem to perceive the use of other languages for beginner learners as acceptable and even necessary. According to them, beginners need more guidance in L1 in order to grasp the new system. As stated in Mora Pablo et al. (2011), this observation could suggest that in the Mexican context, the learning environment is mostly controlled by the teacher and little or no control is given to the learners.

### **5.1.2 Pedagogical strategies**

Concerning the second category, the participants' responses referred mostly to instructions and clarification, and metalinguistic description. Student teachers indicated that, when they were in the learners' shoes, they would make connections between the TL and any other languages (usually their L1). On the other

hand, from a teacher's perspective, the main concern appears to be the students understanding justifying L1 use. The following quotes illustrate the experience of two student teachers as both LN learners and teachers.

*Obviously, whenever a learner has a few languages bouncing in his head, it is favorable to use prior knowledge of given languages to make sense of the target language currently being learned*. (Participant 358 (qc))

*Cuando uso el inglés (L1 u otro idioma) para explicar cierto concepto, ellos reaccionan con expresiones de mayor entendimiento y efectivamente, entienden mejor lo explicado.*

'When I use English (L1 or other language) to explain a certain concept, they react with better understanding and indeed, understand better what was explained' (Participant 001 (mx), our translation)

The above excerpts suggest that the student teachers are aware that other languages can actually be beneficial when it comes to understanding the structure of the TL. However, as for the first category, there is an apparent difference between the participants from Quebec (17 comments) and from Mexico (33 comments). While most participants from Quebec seem to perceive using other languages as a last resort, the Mexican student teachers appear to see other languages as facilitators for the development of metalinguistic knowledge in the TL. It could be hypothesized that these observations are in line with the communicative classroom practice implemented in the Quebec education program versus the more traditional focus on forms used in Mexico.

### **5.1.3 Practical constrains**

As for the third category, participants' responses were classified in two subcategories: Time constraint and disciplinary issues. Regarding the first sub-category, participants made very few comments invoking the use of other languages for time constraint purposes. However, some of them mentioned that it all depends on the institutions' curriculum and the learners' needs. There is a certain openness to the use of other languages in the classroom due to time constraint, as the following excerpt shows:

*[…] por cuestiones de tiempo muchas veces un maestro debe de ser práctico. Considero válido el hacer uso de la lengua materna […] para lograr el objetivo y optimizar tiempo.*

### Pierre-Luc Paquet & Nina Woll

'[…] because of time constraint, in many cases the teacher needs to be practical. I consider valid the use of L1 […] in order to reach the objective and optimize time.' (Participant 925 (mx), our translation)

The above excerpt suggests that student teachers sometimes feel pressured and make decisions based on factors such as time constraint without considering their impact on the learners' language development. As for the last sub-category, student teachers seem to be inclined to resorting to the learners' L1 when it comes to disciplinary issues. This can be seen in the following excerpt:

*Speaking in the target language as much as possible, using L1 for important or meaningful events (intense need for clarification, discipline, etc.) […]* (Participant 957 (qc))

This quote is quite powerful since it actually demonstrates that when the language is used in a meaningful context, such as the one presented above, student teachers favor the use of the learners' L1. It is important to mention that most comments related to disciplinary issues were written by participants from Quebec (17 comments). In other words, it seems that the Quebec student teachers believe that this factor could be the most relevant one when it comes to resorting to other languages in the classroom.

To summarize results for the first research question, as previously mentioned, there is a definite difference between the two populations participating in the study. On the one hand, the Quebec student teachers seem to be resistant to resorting to other languages than the TL in the classroom. In fact, the only situation where it appears to be acceptable for them to use other languages is for disciplinary issues. This being said, the "anti-L1-attitude" identified by Cook (2001) regarding teaching methodologies in the twentieth century still seems to be existent nowadays. On the other hand, the Mexican student teachers demonstrated that they were open to the use of other languages in most situations. More specifically, their responses suggest that the use of the learners' L1 in beginner levels is perceived as essential for both teaching and learning purposes. Moreover, they seem to be inclined to use other languages as pedagogical strategies, for instance, for metalinguistic descriptions, instructions and clarification, which could further support the idea of a more teacher-centered approach used in this context.

### **5.2 Student teacher's belief profiles**

With reference to the second research question, this study explored to what extent student teachers in Quebec and in Mexico appear to be open to using L1 or

### 5 Debunking beliefs regarding the target-language-only rule

other languages in their personal projected practice. As previously mentioned, all participants were classified in one of the resulting belief profiles: Hardline-TL-only, open-to-other-languages and multivoicedness.

Considering the forty participants that completed the vignette-based questionnaire, a total of eight who were classified as hardline-TL-only. As for the opento-other-languages, sixteen participants showed a tendency towards this belief profile. Finally, a total of sixteen participants were identified as demonstrating multivoicedness. Before looking further at the student teachers' responses, it is important to mention that there were great discrepancies between contexts in terms of their belief profiles (see Table 5.2).

Table 5.2: Distribution of student teachers' belief profiles


As shown in Table 5.2, the student teachers from Quebec are the only ones who manifest a hardline-TL-only belief profile ( = 8). The following excerpt reflects this observation.

*Non. On doit apprendre le français en français et l'anglais en anglais.*

'No. We need to learn French in French and English in English' (Participant 816 (qc), our translation)

This excerpt suggests that the participant does not recognize the potential benefit of using other languages when it comes to learning and/or teaching a new one. In other words, participants from this belief profile do not perceive language learning as a dynamic process, as Herdina & Jessner (2002) suggested in their dynamic model of multilingualism.

With reference to the second belief profile, open-to-other-languages, most participants classified in this category were from the Mexican context ( = 13). These student teachers' responses were mentioning exclusively the use of the students' L1, excluding other languages than the TL from the equation, and tended to focus on explicit grammar instruction. This can be seen in the next quote:

*Una explicación de gramática (por ejemplo) en primera lengua ayuda a los estudiantes entender la lengua meta y las diferencias entre la lengua meta y la lengua materna mejor.*

'A grammar explanation (for example) in the L1 helps learners better understand the target language and the differences between the target language and their L1' (Participant 103 (mx), our translation)

In this quote, the participant's response is focusing on the teacher providing a metalinguistic description of a grammar concept using the learners' L1. As mentioned earlier in this section, it seems that the Mexican student teachers tend to teach in a more traditional way, which may lead them to resort to other languages (mostly the learners' L1) as they go from one grammar element to the other. Finally, the two following excerpts reflect what was identified as multivoiced discourse:

*Ella debería seguir usando la lengua meta, según lo que se me ha enseñado y según el objetivo.*

'She should keep on using the target language, according to what I was taught and in accordance with the objectives' (Participant 349 (mx), our translation)

*El uso de la lengua de los estudiantes me parece una buena estrategia cuando se puede tener dificultad para que entiendan en la lengua meta […]*

'The use of the students' L1 seems to me as a good strategy when they have difficulties to make sure they understand the target language […]' (Participant 349 (mx), our translation)

*Depending on the degree of difficulty of the grammar point she is currently explaining, it could be more efficient (time-wise) to switch to the mother tongue of the students, however it has been proven that the teacher is doing a disservice to the students if he/she uses the mother tongue extensively. It lets the learners be intellectually lazy because they know that the teacher will use their tongue, therefore they do not have to struggle to understand the target language.* (Participant 358 (qc))

As these excerpts illustrate, participants in the multivoicedness profile recognize both the importance of maximal exposure and the potential benefit of using other languages in the classroom, thus reflecting two apparently contradicting perspectives. The Mexican student teacher indicates that using exclusively the TL was enforced by teacher educators, which may not reflect the participant's professional choices. Moreover, it seems that using other languages consists in the teacher resorting to the learners' L1 to explain new TL structures. As for

the Quebec participant, using other languages appears to be a "last resort" strategy to cope with structural difficulties. In other words, when problem-solving is needed, this participant tends to resort in the learners' L1, which appears to reflect a spontaneous decision rather than a theoretically informed judgment.

To summarize results for the second research question, as previously mentioned, a line can be drawn between the two populations participating in the study. The only participants that demonstrated a hardline-TL-only position are from Quebec. More specifically, these student teachers indicated that it was essential to use the TL in the LN classroom at all time, and that resorting to other languages has a negative effect on learners' language development. As for the Mexican participants, the majority were classified in the open-to-other-languages category. The results suggest that their decisions were influenced by the presumed focus of their courses, that is to say where language is the object of study. Finally, the multivoicedness profile was distributed across contexts and will be further addressed in the discussion section.

### **6 Discussion and conclusion**

In this section, the overall findings will be reviewed and discussed with a view to debunking student teachers' beliefs regarding the TL-only rule and to making concrete suggestions to promote reflective teaching and to initiate changes in teacher training programs. As for conditions that were perceived as acceptable for resorting to other languages, the distribution of factors was different in the two student populations. Specifically, while student teachers from Mexico mostly referred to situations related to pedagogical strategies and learner factors, the main factor listed among Quebec participants was discipline.

On the whole, Mexican student teachers seem to be more inclined to accepting the use of other languages, especially when this deviation from the TL-only rule had a pedagogical purpose, or to palliate individual learners' needs. At first sight, this tendency would seem encouraging to teacher trainers since it appears to reflect professionally motived decisions rather than spontaneous reactions to overwhelming situations. However, when answering vignettes from a teacher perspective, Mexican participants never mentioned that other languages, including the L1, could be used as a reference to better understand TL structures. Rather, the comments subsumed under "pedagogical strategies" were mostly related to explaining grammar in the L1 to ensure understanding. This apparent focus on explicit grammar instruction may stem from their own experience as learners, supposedly based on a more textbook-based, deductive approach where "the

### Pierre-Luc Paquet & Nina Woll

much-discredited presentation–practice–production procedure still prevails regardless of the pedagogical label on the coursebook" (Tomlinson & Masuhara 2018). To summarize findings regarding the Mexican participants' justifications for resorting to other languages, the latter do not reflect the kind of pedagogical reflection targeted in crosslinguistic awareness pedagogy, that is to draw on correspondences across the whole repertoire. Instead, L1 is perceived as a vehicle rather than a resource. This observation could serve as an anchor point for teacher educators.

Even though less of the Quebec participants' comments were related to "pedagogical strategies", their main focus was also on ensuring understanding when explaining grammatical structures in L1. Meanwhile, situations that would justify L1 use in this context were mostly related to discipline. Despite the fact that classroom management is not related to pedagogical strategies, the overall considerations still seem to reflect a similar concern: For really important things, teachers should use the L1. This is consistent with previous research indicating that "using the L1 for discipline signals to the students that when 'real' communication needs are at stake there is no need to use the L2" (Ellis & Shintani 2014: 234). In other words, TL should be used at all times, except when serious things are being discussed, those that require understanding. The reported experiences also reflect their apprenticeship of observation. That is, the LN instruction many Quebeckers seem to have received would essentially reflect TL-only, except for certain grammar points and discipline. While the program prescribes exclusive TL use, which would include classroom management and form-focused instruction, it may be inferred that the student teachers participating in the present study had no models available for efficient TL use in those specific situations. In sum, these findings parallel those from the Mexican context in that the comments from the Quebec population was also at variance with pedagogical reflections in which resorting to other languages is perceived as a steppingstone for crosslinguistic awareness. Despite this comparable perception of L1 as a vehicle to face serious or problematic classroom situations, the observations related to the Quebec participants' beliefs might serve as a different anchor point for teacher educators. Namely, instead of directly addressing the potential benefits of using other languages as a resource, teacher trainers might also want to debunk the idea of failure when resorting to L1. Specifically, they could examine whether exclusive TL use is favored because student teachers believe that this is what their cooperative teachers, their practicum supervisors and professors want them to do, or because they were shown how expert teachers "fail".

As for the belief profiles identified across the sample, the predictions were confirmed in that more hardline TL-only promoters were listed among the Quebec

### 5 Debunking beliefs regarding the target-language-only rule

participants, while Mexican student teachers were generally more open to using other languages. However, if openness merely translates into L1-is-ok, this does not necessarily entail pedagogical reflection. In other words, the presumed openness to using other languages may not reflect the potential for those participants to integrate crosslinguistic activities in their own classrooms. As for the profiles reflecting multivoiced discourse, which represented roughly a third of the participants in each context, they point to a compromise position, where student teachers find themselves navigating the theory-practice divide. While initially interpreted as contradictions, these multiple voices may rather be considered a reflection of student teachers' complex realities. These insights are highly valuable since they reveal apprenticeships of observation, thus providing anchor points for teacher education.

To conclude, the two researchers, who have started exploring their role as teacher educators, are hopeful. Even though barely tangible, their goal remains to implement findings from TLA research, whereby language learners make the most of their multilingual repertoire, by drawing on their linguistic resources effectively and with teacher guidance. A first step to be undertaken is to steer away from the assumption that using the TL exclusively makes for efficient language learning and teaching. To achieve this perceptual shift, there is a need to raise student-teachers, teachers as well as teacher-trainers' awareness of the importance of treating students' as multilingual learners (Arocena Egaña et al. 2015), to introduce crosslinguistic pedagogies as a regular feature of teacher-training programs (De Angelis 2011), and to create more researcher-teacher collaboration to facilitate the implementation of multilingual tasks (Galante et al. 2020). These three moves would lead people to question their own beliefs critically, to challenge those beliefs by experiencing new ways of learning and trying out new ways of teaching, and by adding a layer of pedagogical intention to the existing layers of apprenticeship of observation.

### **Abbreviations**

LN Additional language TLA Third language acquisition

### **References**

Arocena Egaña, Elizabet, Jasone Cenoz & Durk Gorter. 2015. Teachers' beliefs in multilingual education in the Basque country and in Friesland. *Journal of Immersion and Content-Based Language Education* 3(2). 169–193.

### Pierre-Luc Paquet & Nina Woll


### 5 Debunking beliefs regarding the target-language-only rule

only is not the way to go:" Teachers' perceptions of plurilingual instruction in an English program at a Canadian university. *TESOL Quarterly* In Press.


### Pierre-Luc Paquet & Nina Woll


## **Chapter 6**

## **Training teachers for the challenges of multilingual education**

### Julia Barnes

Mondragon University

### Margareta Almgren

University of the Basque Country

This chapter focuses on how trainee teachers are prepared to face the linguistic complexity they will encounter in classrooms in the Basque Autonomous Community (henceforth BAC) in Spain. Although Spanish is the home language for most children in the BAC, the overwhelming majority of parents choose Basque as language of instruction at pre-school and primary levels (Basque government 2018) independently of the language spoken at home, which means that Basque is mainly acquired through immersion programmes. The teacher training programme at Mondragon University aims at providing future pre-school and primary teachers with solid knowledge on language acquisition in bi/multilingual contexts. These future teachers are also made familiar with instruments for measuring children's language development and, furthermore, they carry out a team-work project to be presented in English, containing data on children's linguistic picture as well as reflections on their own linguistic skills and the utility of knowledge acquired for their future in multilingual education.

### **1 Introduction**

Today's globalised world is becoming increasingly aware of the fact that most of its inhabitants are no longer monolinguals. Traditionally, a much-discussed issue in childhood bilingualism has been that of mixing and separation of codes (Meisel

Julia Barnes & Margareta Almgren. 2021. Training teachers for the challenges of multilingual education. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 117–136. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449778

### Julia Barnes & Margareta Almgren

2001) or crosslinguistic influence (Lanza 1998; Müller & Hulk 2001), and much effort was given to discussing the theme of balanced or unbalanced bilingualism (De Houwer 2009b). Recently the perspective has been changing. Rather than aiming at, and evaluating, native-like competence in a bi/multilingual person's languages, the new approach takes into account each individual's total language repertoire from a holistic and dynamic perspective. This "multi-competence" perspective (Aronin 2016) proves useful when analysing the wide variety of situations in multilingual communities. In such contexts it is paramount to view "someone who knows two or more languages [to be] a different person from a monolingual, and so needs to be looked at in their own right rather than as a deficient monolingual" (Cook 2013, p.3768). As a consequence, the focus of interest is changing from monolingualism/bilingualism to multilingualism, both from the point of view of multilingual individuals and multilingual societies.

It is also important to take into account the difference made between multilingualism as a social phenomenon and multilinguality, as referred to individuals. Although in many cases multilingual individuals live in multilingual communities, this is not always the case. It may well be that in a bi/multilingual community, as the case referred to in the present article, many speakers are monolingual. On the contrary bi/multilingual speakers may also live in monolingual communities, as in the case of a German-dominant child exposed to French from one parent (Meisel 2008).

Today's multilingual speakers are also in many cases L2 speakers of some of their languages, and therefore the concept of *dominant language constellation*, DLC, (Aronin 2016) becomes extremely pertinent when analysing both social and individual use of languages. DLC refers to one person's or one community's most dominant languages and their domain of use and fits very well into the situation we will deal with. We will explore how students of the Basque Country become aware of their own multilingual skills by examining young children's language development, and how they begin to understand the different DLCs they will encounter in their professional future as infant teachers.

### **2 The sociolinguistic situation of the Basque Country**

Basque, a non-Indo-European language of unknown origin, is currently spoken by approximately 34% of the population of the Basque Autonomous Country (henceforth BAC) in northern Spain and understood by another 20% (Basque government 2017). Since Basque became co-official with Spanish in the BAC in 1979 and compulsory in education in 1982, its weight in primary education has steadily grown. It was introduced offering three educational models for parents

### 6 Training teachers for the challenges of multilingual education

to choose between: In the A model, education is provided in Spanish and Basque is a subject matter for 3 or 4 hours a week. In the B model, both Spanish and Basque are languages of instruction. In the D model, (C is not used in Basque) Basque is the language of instruction and Spanish is a subject matter for 3 or 4 hours a week.

Initially, the possibility of using Basque in education at the same level as Spanish was seen with certain incredulity, in particular by the rural Basque-speaking population by whom Spanish was perceived as the prestige language. In 1985 around 16% of primary children were schooled in Basque, and 8% in the bilingual model B. During the following decades, the positive results obtained in education in Basque, both in academic aspects and as to language skills in Basque and Spanish, slowly convinced the majority of the population of the benefits of education in Basque (Cenoz 2009). In 2018, the D model was chosen by 78.9% of the families and the B model by another 17.3% (Basque government 2018). Taking into account that Basque is L1 only for 23% of the children, it is evident that schooling in Basque means full or partial immersion for many children and is producing a generational shift in new speakers of the Basque language as a result of effective education policy in a situation of sociolinguistic complexity (Amorrortu et al. 2009; Ortega et al. 2015).

In order to respond to the increasing demand for instruction in Basque, an intense effort had to be made in order to prepare teachers for this task. On the one hand, Basque-speaking teachers in the Spanish educational system before 1982 were usually not literate in Basque. They were speakers of different oral varieties of Basque but had no instruction in the standardised written language "euskara batua", established in 1968. On the other hand, most teachers were not Basque speakers and had to start learning the language from scratch. A considerable effort has been made both by these teachers and by the department of education of the Basque government in terms of providing paid training. Due to this effort, the number of teachers able to teach (in) Basque has increased from around 5% in 1978 to over 90% in 2016 (Basque government 2017).

However, acquiring knowledge of, and even fluency in, a language is not equivalent to knowing how to teach it. And even if teachers are prepared to teach Basque as L2, or even to teach in Basque as L1, the situations they will meet in classrooms are complex; children whose home language is Basque are usually mixed with children who have had some contact with Basque outside school, or those who have had no contact whatsoever with Basque. In addition, the sociolinguistic context spans over environments where Basque is the principle language used in the community, where Basque and Spanish are used alongside each other, and environments where no Basque at all is spoken. These are factors which show that specific training is needed.

### Julia Barnes & Margareta Almgren

### **2.1 Infant and pre-primary school teacher training at Mondragon University**

In the following lines, we will focus on how trainee teachers are prepared to confront the linguistic complexity they will inevitably encounter in classrooms across the BAC. The present approach to teacher training was initiated in 2009, as part of the Bologna process, but here we will limit our discussion to data from the 2017–2018 cohort.

As pointed out, although Spanish is the home language for most children in the BAC, the overwhelming majority of parents choose Basque as language of instruction at pre-school and primary levels independently of the home language, which means that Basque is mainly acquired through immersion programmes. In addition to Basque and Spanish, English is often introduced as a third language by the early age of 4 and students are expected to reach B1 competence by the end of post-secondary education (Lasagabaster 2000). Some children are also exposed to other languages at home (Barnes 2006). All these children are schooled together in mixed groups, so it can be concluded that as a result, multilingualism is becoming the norm in the BAC (Cenoz 2009).

If a school wishes to teach 3 or 4 languages it is clearly necessary to plan the progressive introduction of each one to fulfil the linguistic objectives set down by the Basque department of education. Consequently, all future teachers from any specialism, who are themselves Basque-Spanish bilinguals (many of them L2 speakers of Basque), will need enhanced awareness of psycholinguistic and sociolinguistic requirements for educating Basque L1 children alongside Spanish L1 children in Basque immersion, together with pupils from other linguistic backgrounds, particularly in infant classrooms where language is still developing. This raises some issues from a training perpective. How aware are future teachers of young bilinguals of the linguistic complexity in their classrooms? Do they realise the variety of situations of input their pupils are exposed to? Are they familiar with instruments with which evidence of language development and sociolinguistic development can be measured? To what extent are Basque L1 and L2 trainees (who are not language teaching specialists) able to reflect upon (in writing) and act upon (in teaching) good practice in classrooms?

### **2.2 Teaching model and methodology**

To better understand this, we will analyse features such as the linguistic competence and attitudes towards language learning of Basque-Spanish bilingual trainee infant and pre-primary (0–6 years) teachers who annually complete a

### 6 Training teachers for the challenges of multilingual education

six credit module through the medium of English in years 1, 2, and 4 of their education degree, in undergraduate training in infant education at Mondragon University in the Basque Country. These students are not training to be English teachers, but we should point out that the modules are in English for two main reasons, namely to develop their personal multilingual skills (they have received compulsory English tuition for 14 years within the education system) and to give them the skills and confidence necessary to use their third language if required to do so in their future teaching activity. English is increasingly present in Basque infant education in different spaces across the curriculum, and novice teachers are often considered able to carry out activities in English.

The teaching modules within the Mondragon University curriculum which are delivered in English as a third language cover:


This paper reports on the second year module, *learning and teaching of second languages in multilingual contexts*, by the end of which students are required to have the initial theoretical knowledge to understand the principles behind the acquisition of first and second languages in monolingual, bilingual and multilingual contexts. The module includes basic concepts in relation to age factor in the acquisition of languages and of interaction and inter-language. To this effect, students are provided with literature on language acquisition (Genesee 1994; Lightbown & Spada 2013; Palenham 2004) in order to acquire theoretical knowledge to be applied in the projects they will carry out. They are also made aware of affective and emotional states that may influence language acquisition (Barnes 2006; De Houwer 2009b,a; Dewaele 2013. The module closes with a contrasting practical focus on incorporating classroom and linguistic strategies appropriate to what has been learned, through the design and peer-teaching of a short activity. This work is further developed in a module in the third year of the degree, where the topics are reviewed and developed through the medium of Basque.

The principle aim of the second year module of our interest here is to identify difficulties in communication and learning when the language of instruction is new to the learner, to show empathy to children in immersion contexts

### Julia Barnes & Margareta Almgren

and to enable students to make use of strategies to aid learning, particularly language learning, in such contexts. Additionally, students develop their own language skills, through reflection on their own language acquisition process(es) and through working through the medium of their third language, English. Academic reading skills are developed through designated chapters relating to language acquisition from Palenham (2004).

The module is divided into four units, which feed into two practical assessed Projects.

Project 1, *One child's linguistic picture*, covers the first two units of the module:


The project consists of a small-scale research study in which students are required to collect a sample of child language to examine to identify evidence of the multiple processes that take place during language acquisition and takes place over a period of two months.

Project 2 covers the second two units of the module:


Students create, peer-teach and justify an activity appropriate for a classroom context specified by the tutor. This activity should be governed by best practice and the requirements for infant language development as described in the curricular guidelines of the Basque education department. Project 2 also involves the design and preparation of a mini-conference. Student groups prepare poster presentations and accompanying handouts in which they display and explain to each other different aspects of methodology and materials appropriate for language learning in early years education. This project lasts approximately one month.

The focus of the present study is the Project 1, titled *one child's linguistic picture* (henceforth OCLP), which is central to the whole module and will be addressed in more detail in the following sections.

6 Training teachers for the challenges of multilingual education

### **3** *One child's linguistic picture* **(OCLP) project**

### **3.1 Instruments for measuring child language development**

As shown above, during their second year studies through the medium of English, teacher trainees have reflected on their own learning process as bilinguals, as learners of a foreign language (English) and their feelings when confronted with a language which is completely new to them. They have also learned about various aspects of language acquisition and bi/multilingualism. Subsequently, and in order to consolidate this learning, students take part in a multilingual smallscale research project, the OCLP, about the linguistic development of 4-year-old children in Basque and Spanish. To this effect, they are made familiar with instruments such as the Peabody picture vocabulary test (Dunn et al. 1986), the MacArthur-Bates CDI III (Dale 2007), and narrative elicitation tools. These tools will, in conjunction, provide a broad assessment of the chosen child's understanding in both languages (Peabody test), production in both languages (*Bird story*) and of parental report on the child's understanding and production in both languages (KGNZ). The goal of this data collection, rather than to undertake a scientific study on children's receptive and productive skills, is for trainee teachers to become aware of some aspects of child language acquisition.

The Peabody vocabulary test is a widely-known instrument used to measure receptive vocabulary ability, often applied with diagnostic purposes by speech therapists. First designed for testing the understanding of standard American English, its use has been extended to other languages, since it is based on linking a word the examiner says to one of four pictures. Consequently, it can be used on one and the same subject both for Basque and for Spanish for comparative purposes. It was adapted into Basque for the purposes of the OCLP project, but not validated. Since the purpose was to obtain an impression of the child's understanding, students were made aware of this and data cannot be considered scientifically reliable.

The MacArthur-Bates communicative development inventory (CDI) questionnaires are based on parental reports on their children's receptive (CDI I) and productive (CDI I, II and III) vocabulary. It was also first developed for American English, but has subsequently been adapted to nearly 100 languages throughout the world.<sup>1</sup> These questionnaires were originally developed for use by paediatricians and speech therapists, but they have since been widely used in research on child language (Fenson et al. 1994; Bates et al. 1995; Kern 2007; Ericsson et al.

<sup>1</sup>http://mb-cdi.stanford.edu/

### Julia Barnes & Margareta Almgren

2012). The CDI questionnaires have also been used to reflect language development in bilingual children in different language communities. For this purpose, monolingual questionnaires – one for each language – have been used in most cases, such as Galician-Spanish (Pérez-Pereira 2008), Basque-Spanish (Ezeizabarrena et al. 2018) and Spanish-English (Core et al. 2013). In a few cases, such as Irish-English (O'Toole & Hickey 2017) and Maltese-English (Gatt et al. 2014) specific bilingual questionnaires are available.

The CDI I for ages 8–15 months and the CDI II for ages 16–30 months have also been adapted into Basque as KGNZ I and II by Barreña et al. (2008) and into Spanish as iLG (Lopez Ornat et al. 2003). A further version for children up to 37 months of age known as CDI III was developed for American English and adapted to Basque as KGNZ III (Garcia et al. 2014). The Basque KGNZ versions have proved to be valid up to age 40 months. The CDI includes calculating mean length of utterance, based on the 3 longest sentences the child has produced, as reported by parents on the questionnaire. The MLU may be calculated on the average number of words, MLUw, or on the average number of morphemes, MLUm. Taking into account that in the Basque Country children are often exposed to both Basque and Spanish, the KGNZ also contains a section where parents are asked to estimate input in Basque, or Basque and Spanish, in relation to the number of hours a day and days a week. This is of interest because sole exposure to Basque usually takes place in the family and at early ages with incresing social exposure to Spanish over time. When comparing monolingual Basque and bilingual Basque-Spanish development the data collected with the KGNZ questionnaires reveal that if input in Basque is above 60% no significant differences appear in the productions of monolingual and bilingual subjects, prior to the age of 24 months. However, after this age amount of input is revealed to be decisive in rate of development (Ezeizabarrena et al. 2013).

So far no standardised Spanish version of the CDI III is available. For the purposes of the present OCLP project a translated version of the Basque KGNZ III was used. Evidently, the data collected should be interpreted with caution, but it should be born in mind that the main purpose was not, in this case, to collect statistically valid data on the children but to make students acquainted with different instruments and enable them to interpret their findings.

The wordless picture story *Frog, where are you?* (Mayer 1969) has been used in numerous studies on children's narrative skills (Berman & Slobin 1994; Strömquist & Verhoeven 2004). During the early years of the OCLP project it was used by the second year students at Mondragon University, but feedback suggested that children nowadays, accustomed to colourful stories in books and digital resources, seemed no longer to be attracted by the black-and-white Frog story.

### 6 Training teachers for the challenges of multilingual education

During the last two academic years of our students' training courses, the Frog story has been replaced by the *Bird story* from the COST project (Gagarina et al. 2012). The images (in their coloured version) are unfolded in sequential order to the children, who are asked to tell the story. The trainee teachers are strongly advised against allowing the test to be a question-and-answer session as the objective is to collect evidence of the child's narrative production in each language, rather than of the child's interactive skills with an interlocutor. However, the students are also encouraged to be aware of the child's anxiety, and proceed accordingly.

### **3.2 Procedure**

Teams of 3 trainees were formed and instructed to find a 3 to 4-year-old child from a family (usually from their own family and friends, local environment, or members of university staff) who were interested and agreed to let them carry out a study on the child's linguistic background and linguistic skills in Basque and Spanish. The preliminary data were first presented by each group in English on posters showing the uniqueness of each child's linguistic background with regards to sociolinguistic context in the Basque Country, family constellation and schooling.

One of the students introduced him/herself to the child as Basque-speaking, one as Spanish-speaking and one as bilingual. This was in order to apply the "one-person one-language" principle when using the Peabody test and eliciting the Bird story narration. The parents were asked to complete the questionnaires in both Basque KGNZ III and Spanish (translated KGNZ III).This ensured that at least one of the parents was a proficient Basque speaker.

Furthermore, trainees were asked to write a report reflecting in English on the language acquisition process, on the child's anxiety and strategies in encounters with a new language and on the age appropriate teacher intervention for language development. These reports were also presented orally in English in front of classmates.

### **3.3 Findings**

### **3.3.1 Children's data**

As reflected in Table 6.1, data on 20 children, aged from 2:07 to 4:01, were collected by the same number of student teams. These teams were formed of 3 students. The average age of the children was 3:07 and there were 12 girls and 8 boys. Most data were collected in the Mondragon region, which is fairly Basque-speaking,

Figure 6.1: Example of a student group poster

so young children were expected to be Basque-dominant. Based on data on the amount of input in both language as reported on the KGNZ questionnaire by parents it was possible to appreciate whether the child was Basque dominant, Spanish dominant, or had balanced input, or input from another language.


Table 6.1: Nº of subjects and linguistic characteristics

It is also worth noting that many parents who are L2 speakers of Basque tend to make an effort to speak Basque with young children, before language use becomes more complicated and they tend to speak less Basque when demands on parental language skills grow.

From Table 6.2 it can be concluded that most children had developed both Basque and Spanish to a level that made it possible to measure language skills

### 6 Training teachers for the challenges of multilingual education

with the instruments selected, except in relation to the Bird story in Spanish. Only 10 children out of 20 had attained sufficient expressive vocabulary to be able to produce coherent utterances in Spanish.


Table 6.2: Tests collected by students

### **3.3.2 Student presentations**

Having collected the data and completed a guided written report of it, each group of students presented the results of their OCLP project orally, in English, before the class. The aim was to make students aware of the diversity of linguistic microcosms that could be found in children within the age-groups they would be likely to be teaching in their future careers. The presentations were not recorded because this cohort expressed that they would not feel comfortable before the camera, but this could be done in future studies.

Through the presentations, it became evident that 3 languages constitute these students' DLC (Aronin 2016). They all switched fairly freely between Basque and Spanish when commenting among themselves on how to proceed and it is difficult to know whether they are dominant in one language or the other. However, they took care to express themselves in English when facing the audience, aware that this was the language of the classroom context. They had no special difficulties in oral expression in English, which was sufficiently fluent and only needed occasional support by written texts. Their presentations were easy to follow and well structured.

As to recurring comments in the presentations, many of them expressed their initial frustration in relation to the gap between the results of the Bird story and the Peabody test, until they had realised that they were dealing with the difference between production and comprehension. This was found to be especially striking in Spanish; all groups confessed they had thought most children did not know any Spanish until they discovered that Peabody comprehension scores were higher than expected in Spanish.

Other interesting comments were made on lexical incorporations from one language to the other and some examples were given: *de compras* ('shopping'), *morditu* ('morder' in Spanish; 'bite') in Basque by a Spanish dominant child, *eztetnai* ('I don't want to') in Spanish by a Basque dominant child. Through working with the KGNZ parental report questionnaires they discovered that there is a common, conceptual vocabulary which is reproduced in both languages – although not always: "they know the same words in different languages but some only in one language – and not necessarily the first language".

All the students pointed out that they had evidenced a high influence of family and environment on the children's linguistic development, and stressed the importance for future teachers to have knowledge of different types of sociolinguistic situations, as well as of language acquisition in a bilingual community. Their reflections on their own language learning in English and appreciation of having been able to apply theoretical knowledge acquired were very gratifying to the tutors involved.

### **3.3.3 Students' written reflections**

An analysis of the discussion and conclusion sections of the guided reports submitted by the trainees provides insights into their meta-linguistic awareness, attitudes and emotions in addition to their English language proficiency. In the following, we will present some examples of the most representative aspects from the students' written reports. The extracts are reproduced in the students' own words, to reflect the general standard of their written English production. These reports have been taken from work submitted electronically for assessment, and the students may have used dictionaries or other electronic support when preparing them. Individual hand-written reports were also made under exam conditions and are the subject of a further study.

### 3.3.3.1 Dealing with tensions

In the first place, there was a general positive feeling about working in a team, sharing knowledge and experiences. Working in teams of 3 is stressed by the students as a very positive factor that greatly contributed to the good results:

In our group we have worked very hard to carry out this project. In most times we have worked separating the work and helping each other, in the doubts we have had.

Despite we have some difficulties we have had some greatest satisfactions about our work in the project. We are proud because we have been able to carry out the project. It has been possible because we have had a really good communication and confidence in the group. So, group work has been great and we have organized what we had to do.

The students were instructed to meet the children and their families, whenever possible, on two different days, in order to separate languages and apply the "one person – one language" principle. However, it proved difficult to find family availability for more than one day of tests. In about half the cases both languages were used on the same day, as explained in the following:

At the beginning, our idea was to meet with N. two times, on the 20th of March and on the 8th of April. But one day before, N's mother called us to say that she couldn't meet with us. That way, the only day that we could meet with N, was on the 8th of April.

Most groups confess that at the start, they felt uncertain about how to approach the children and about how to handle the tests. The students describe their own feelings about the different sections of the project, the difficulties they encountered and how they gained confidence. They also interpret the child's feelings: shy, relaxed, comfortable, wanted to play, bored:

She did not tell us anything because she did not want and the level of Spanish she has was not very good so she was a little bit embarrassed.

The first day we did the exercises in Spanish and she was very shy. After we did the Peabody test and we showed her the pictures and explain her what she had to name them she was more relax...

The child respond to the activities but he is very small and he wanted to play with his toys.

The initial frustration when the children did not respond to some of the tasks was replaced by the insight that with children, flexibility is needed and the bestlaid plans may not be possible to carry out. Time must be limited and activities should not be imposed:

When we finished with the story, N. told us that she was very bored and tired. That way, we changed the plan that we had at the beginning and we had lunch with them to continue.

### Julia Barnes & Margareta Almgren

When we give her the bird story, she started to opened and closing, turning around... but she wasn't interested to told us the story. So we decided to not forced her and we let her doing what she wanted.

When discussing the results with the students it was reiterated that the aim of this study was not to produce a definitive, and scientific, picture but rather an overall impression of one child and their linguistic context. It was made clear to the second year undergraduate students that usually, in order to carry out such tests rigorously, laboratory-like conditions, specific training, and examination of validity and reliability would be required but this was not the objective for the project.

### 3.3.3.2 Knowledge gained during the process

There was a general positive feeling among the students towards the ability they gained as to the use of the instruments and they really enjoyed dealing with the data, making lists, comparing MLUm, MLUw and the Peabody profiles:

The biggest difficulty we have had with this project has been the part in which we had to count the morphemes of the longest sentences in the bird's story and those of MacArthur. We had to ask for help for this and after the explanation it has still been difficult for us, but when we finished it we felt great satisfaction.

J doesn't say any sentences in Spanish and then we analyze the morphemes of Basque sentences. In the Bird story the sentences are longer than in the MacArthur CDI but the morphemes she used are more or less the same long. We can see the similarities in the MLU since in both of them the results are almost the same.

All reports reflected that students had realised that the amount of input in the languages the child is exposed to has a decisive influence on its linguistic development. The variety among children also became significantly clear:

During the project can be seen very well that L's language is reflected in her linguistic background. She knows better Basque than Spanish as many people of Azkoitia. As Palenham (2004)says, all normal children brought up in a normal social environment acquire the language of that environment.

### 6 Training teachers for the challenges of multilingual education

Finally, their theoretical studies on language acquisition could be put into practice and all of them stressed the importance of this experience for them as future teachers:

Throughout this project we have learned many interesting things about different languages acquisition. We have read lots of articles to have enough theory before analyzing H's linguistic development. We have to say that it has been very interesting to apply all the theory in reality, analyzing the real linguistic evolution and situation of the child.

As previously discussed in relation to the oral presentations, an important discovery for students was the difference between comprehension and production. Students would insist initially that their child "only knows Basque" and took some time for the students to assimilate that their own data revealed that many of the children were able to understand some Spanish although they showed no production.

3.3.3.3 Students' meta-linguistic awareness and confidence in their own linguistic preparation

An analysis of the writing submitted by the trainees throughout the module and at its conclusion provides a wealth of insights into their meta-linguistic awareness, attitudes and emotions and their English language proficiency.

To conclude, I have to say that it has been very interesting this module because I have never thought about how I felt or how I learned the second language (Basque) or the foreign language (English).

In the other side the greatest satisfaction in carrying out this project, has been a personal satisfaction; when we have felt that we were able to make and develop a project in English. Apart from that, it has been very satisfactory understand better how a child acquired a second language, and identify the difference between the first and the second language knowledge.

### 3.3.3.4 Attitudes towards teaching

Many of the students commented on how the contact with the children through the project made them aware of important age-appropriate teacher intervention strategies, language strategies and other strategies they will need to develop in their future as teachers of young children:

### Julia Barnes & Margareta Almgren

To sumarised, to achieve good language learning and a good language development, is important to know the children every day and work with their parents coherently. What is more, the teacher has to take in to account different strategies such as:


Such features as child anxiety and the need of an appropriate approach by the teacher were also frequently included in the reports:

Another thing I have learned is that when teaching a second language you have to be very careful because of the language anxiety. For that reason, the teacher he/she has to spoke clearly, made eye contact, made feel confortable. In one future I am going to be a teacher so I think this kind of thing can be important to help to children in the learning language process.

### **4 Conclusions**

We perceive that a number of conclusions can be drawn from the tasks carried out during our project. In the first place, the trainee teachers' linguistic repertoire comprising Basque, Spanish and English permits them to make an instrumental use of all three languages for educational purposes. Although it was already clear that their level of Basque and Spanish was high enough for the studies they are currently involved in, the present work also shows that they have been able to present their oral reports in fairly fluent English (L3) in a natural way, without reading out the texts. Their written reports are complete, well structured and show a sufficient command of the language for the purpose it was designed for. This fact fits well with the "multi-competence" perspective expressed by Aronin (2016), since the aim of the task was not to evaluate native-like competence but to make students feel their L3 can serve as a tool for gaining and presenting knowledge acquired. In this sense, students' reports reflect the confidence gained in relation to their L3 skills and the realisation that English indeed forms part of their dominant language constellation.

In second place, they also reflect the satisfaction felt when realising they are able to handle the instruments used for evaluating language development, by

### 6 Training teachers for the challenges of multilingual education

interpreting the children's feelings and reactions and by overcoming their own uncertainty. In this context it should be born in mind that children's data should be interpreted with caution. On the one hand, students, although acquainted with tools for data collection, are not yet scientifically schooled enough to make these data reliable, and on the other, all the tools used had not been standardised for scientific purposes.

In third place, students seemed to appreciate the value ofthe knowledge of child language that they had acquired by carrying out the OCLP project for them as future teachers, particularly as two (or more) languages are involved.These results corroborate the importance of specific teacher training in bilingual contexts. Last but not least, students also referred to their new awareness of the complexity of family life, experienced first hand when fitting their OCLP visits around children's timetables, illnesses and parental commitments. These are valuable insights for any young teacher.

### **Abbreviations**


### **References**


Barnes, Julia. 2006. *Early trilingualism: A focus on questions*. Clevedon: Multilingual Matters.


### Julia Barnes & Margareta Almgren


### 6 Training teachers for the challenges of multilingual education

tempranos euskera-castellano. In Cecilia Rojas & Elsa Oropeza (eds.), *Diferencias individuales en la adquisición del lenguaje*, 195–218. Mexico: UNAM.


### Julia Barnes & Margareta Almgren


## **Part III**

## **Language learning**

## **Chapter 7**

## **Exploring learner attitudes in multilingual contexts: An empirical investigation at the primary school level**

### Barbara Hofer

Dyme Research Group, Innsbruck University

Recent research highlights the dynamic and complex nature and the situatedness of multilingual development. It emphasises the need for learners to have ample opportunity for interaction in order to progress on their learning trajectories and suggests that positive attitudes and emotions are key to language learning motivation and learning outcomes. Indications are that how children feel about languages and language learning can impact their learning behavior and willingness to engage with a particular L2/Ln. In this paper I investigate how sociolinguistic and educational context and amount of contact with the L2 (and/or other languages) relate to learner attitudes at the primary level. I report on work in progress carried out in variously multilingual settings in South Tyrol with the aim of establishing whether learning in multilingual contexts as opposed to monolingual surroundings has any effect on pupils' attitudes and motivations. To my knowledge no previous studies in South Tyrol have looked into attitudinal factors and/or young learners' beliefs with regards to language(s). The present research seeks to bridge this gap.

### **1 Introduction**

International research has found that the way children feel about languages and language learning can impact their learning behavior and willingness to engage. No previous studies carried out in South Tyrol that I know of, have looked into young learners' (YLs) attitudes and/or beliefs relative to language(s) and language learning. The present study therefore sets out to fill this gap by probing into primary schoolers' language-related attitudes and motivations.

Barbara Hofer. 2021. Exploring learner attitudes in multilingual contexts: An empirical investigation at the primary school level. In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 139–163. Berlin: Language Science Press. DOI: 10.5281/zenodo.4449780

### Barbara Hofer

I begin by outlining the epistemological framework within which this research is situated linking it to related theoretical perspectives, which converge on the importance of context for multilingual growth and point to the interconnectedness of cognitive activity, psychological constructs and contextual circumstances. I then provide a cursory overview of research into contextual matters and learner attitudes and examine how they may be linked to multilingual development. Next, I present the study, mapping out the aims and research question and the sociopolitical and -linguistic background in which the research is embedded. Finally, I report on selective quantitative and qualitative findings and discuss them within a dynamic systems and complexity theory frame of reference. The paper closes with implications for educational practice in South Tyrol and offers recommendations for future research.

### **2 A DMM perspective on multiple language development**

The present research is informed by the *dynamic model of multilingualism* (DMM, Herdina & Jessner 2002). DMM is widely acknowledged as a valid explanatory framework for the complex dynamics involved in multilingual development. The model has been credited for making significant contributions to current understandings of multilingualism and multilingual acquisition. Positing the total interconnectedness of the cognitive, psychological, social, cultural and physical, DMM focuses on the multilingual speaker as highly complex and adaptive systems in constant interaction with their environments.

From a DMM perspective, multilingual growth or development are seen as conditional on learner-users' perceived needs and motivations, and on GLE (General Language Effort, Herdina & Jessner 2002: 131). GLE relates to the amount of effort individuals are willing to expend in order to learn a language (Language Acquisition Effort, LAE) and to the amount of effort they invest in the maintenance of their languages (Language Maintenance Effort, LME).

DMM anticipates multilingual learner-users to benefit from a so-called multilingualism factor (M-factor) which mitigates the amount of GLE needed to learn and maintain a given L2/Ln (Ln denoting any language beyond the speaker's second language) and affords them enhanced possibilities for action. Endowing the system with special qualities, the M-factor comprises a range of skills and abilities which multilinguals develop as a function of the dynamic interactions between their multiple languages including language learning, language management and language maintenance skills, an Enhanced Multilingual Monitor (EMM), and enhanced Meta- and Crosslinguistic Abilities (MLA and XLA). Key

### 7 Exploring learner attitudes in multilingual contexts

components of the M-factor, meta- and crosslinguistic abilities are the most distinguishing features of the multilingual system (Jessner et al. 2018).

Grounded in dynamic systems theory, DMM embraces the idea of holism, which entails a global, holistic approach to multilingual phenomena. Holistic paradigms are characterised by a conception of the multilingual learner-user as an integrated and complete whole, and as constituted by their relationship to other systems (Philips 2000: 44), in other words, nested in a greater whole and continuously interacting with its surroundings.

It is against this backdrop of total interconnectedness that DMM posits the need to integrate the psycholinguistic with a sociolinguistic focus in order to take adequate account of the complexities of multilingual learning and use. Starting from the premise that "[l]inguistic aspects of individual multilingualism are shaped by the sociolinguistic setting in which the multilingual's life takes place" (Jessner 2008: 273), the implication is that the multilingual system is molded as much by social factors as by "individual cognitive factors such as motivation, anxiety, language aptitude, and self-esteem" (Jessner 2008: 274).

The situatedness of multilingual development is also highlighted in complexity theory-inspired and ecological approaches to applied linguistics (e.g. Kramsch 2012; Larsen-Freeman & Cameron 2008; O'Laoire & Aronin 2004), which point to social and cultural settings as crucial factors contributing to the individual's multilinguality. In consonance with DMM, complexity and ecological thinking endorse a view of multilingual agency and subjectivity as resulting from complex and dynamic interactions taking place on different levels and timescales, and thence as multifaceted and multi-layered, rather than self-contained and reducible to single causes.

In the following I discuss core principles underlying these paradigms and I examine how they tie in with the DMM perspective and the exploratory foci of the present study.

### **3 Complexity theory and ecological approaches**

Complexity theory (henceforth CT) implicates thinking in terms of connectedness, relationships and context. It is premised on the recognition that "all natural phenomena are ultimately interconnected, and that their essential properties, in fact, derive from their relationships to other things" (Capra & Luisi 2018: 2). Accordingly, CT conceives of individual learner-users as open systems who interact with and (if necessary) adapt to multiple contextual factors in time and space (cf. Herdina & Jessner 2002). Context as an intrinsic part of the system rather

### Barbara Hofer

than merely "background against which action takes place" (Larsen-Freeman & Cameron 2008: 16) is seen as playing an all-important role with the individual and her/his context subsisting in a relationship of "reciprocal causality" (Larsen-Freeman & Cameron 2008: 7). The present research dovetails with this line of thinking as it explores the complex associations between contextual factors and learner attitudes.

By the same token, ecological perspectives predicated on the understanding that systems "are bound into a functional whole by their mutual relationships" (Capra & Luisi 2018: 67) impel acknowledgement "of the interconnectedness of individuals, pairs of individuals, communities, etc." (Larsen-Freeman & Cameron 2008: 19). It is this interconnectedness and reciprocity that Kramsch & Steffensen (2008: 19) invoke when they posit that it is in contact situations (or dialogue, to use the authors' wording) that the personal, situational, and cultural merge, and interaction obtains affording learners the possibility to develop and grow. As for the current investigation, the aim is to ascertain whether different sociolinguistic ecologies and contact opportunities as provided by learner-users' everyday life realities affect their attitudes towards languages and language(s) learning.

### **4 Multilingual learning context**

Highlighting the importance of context, Blommaert et al. (2005: 203) argue that context is crucial because it "does something to people" and in so doing influences "what people can do and can become" (cf. Cenoz 2013: 79; van Geert et al. 2011: 240). Investigating learner attitudes in Sweden, Henry & Apelgren (2008: 611) confirm that there is substantial interindividual variation depending on the cultural context in which learning takes place. It then seems that if we are to understand the multilingual learner-user system better, we need to take adequate account of the larger context in which the system operates (Larsen-Freeman & Cameron 2008: 35; Paladino & Vaes 2009: 223).

In the present study, context understood as the here and now in which a system is active (Thelen & Smith 1994: 217) relates in particular to different sociolinguistic and educational settings which range from the relatively monolingual to the highly multilingual. I delineate these settings in some detail in the second part of the paper. Here, I discuss possible associations between contact with the target language community on the one hand and multilingual growth and learner attitudes on the other.

Discussing language (learning) in terms of situated practice and as requiring "participation in […] communities of practice" Blommaert et al. (2005: 206; cf.

### 7 Exploring learner attitudes in multilingual contexts

Cenoz 2013: 81) imply that participation in a given L2/Ln community of practice is conditional on the sociolinguistic context, viz., the linguistic demographics and the presence of other-language speakers in the immediate surroundings. This is supported by empirical evidence which suggests that increased contact time with L2/Ln speakers creates favourable conditions for learning because it provides important opportunities for learners to communicate (De Angelis 2012: 408; see also Kordt 2018) and thus acts as a catalyst for language(s) learning (cf. Bozzo 2014). Investigating the effects of population distribution on L1 and L2 acquisition in South Tyrol, De Angelis, for instance, found that opportunity to communicate in L2 is highly beneficial for acquisition (De Angelis 2012: 420; see also De Angelis in press). Along similar lines, Csizér & Kormos (2009: 63) argue that interaction with speakers of other languages creates important opportunities for developing L2 learners' language competence. In the DMM, acknowledgement of the centrality of contact with L2/Ln (speakers) is reflected in the postulate that regular use of the target language(s) is vital for the maintenance of both the single languages, and the overall multilingual system (Herdina & Jessner 2002: 99; see also Schmid 2011: 158 on the importance of contact and frequency of use for language maintenance).

The paper now moves on to consider the effects of affective and attitudinal factors on multilingual development.

### **5 Learner attitudes and multilingual growth**

Attitudes have been variously described as evaluative orientations to a given social object or phenomenon (Garret 2010; Cenoz 2004), and as sets of beliefs and psychological predispositions (Tódor & Dégi 2016: 124).

Research into young learners' language-related attitudes has shown that, from a very young age, children form views and hold beliefs about languages and language(s) learning (Munoz 2014; Nagy & Nikolov 2009; Nikolov 2009). As is welldocumented, learner attitudes and language-related emotions have a significant role to play in second and foreign language learning (Courtney et al. 2017: 3; Culhane 2004: 58; Portolés Falomir 2015: 77). The understanding is that affective and attitudinal factors are closely linked to learner efficiency, self-concept and learning success (Dörnyei et al. 2015; MacIntyre & Gregersen 2012: 197; MacIntyre et al. 2016; Wesely 2012). In like manner, there is general agreement that positive attitudes towards a language and/or its speakers will result in enhanced levels of motivation, increase learners' readiness to engage and contribute to higher overall attainment (MacIntyre & Gregersen 2012: 193).

### Barbara Hofer

Primary schoolers seem to be particularly influenced by what goes on in the second or foreign language classroom with teachers, methodology, classroom activities, and overall learning atmosphere all contributing to shaping their orientations towards a particular language and towards language(s) learning in general (Nikolov 1999). Research reported by Chambers (1999: 155) found that 12 to 14 year-olds' attitudes towards learning English declined as a result of classroom approaches and teaching methodology that failed to match their expectations. On a related note, Wesely (2012: 107) suggests that young learners' perception of the classroom setting and teaching methodology can have a lasting impact with early language learning experiences at the primary level potentially persisting with learners and influencing their attitudes far into adulthood. Likewise, parents act as important models (Cenoz 2004: 205; Gardner 1985 in Csizér & Kormos 2009) who exercise their influence directly (by encouraging their child or helping with homework), or indirectly through comments or (conscious and unconscious) reactions to members of the L2/Ln community (Csizér & Kormos 2009: 83; Otwinowska & De Angelis 2012: 347). In addition, as emerges from an investigation of the language practices and attitudes of minority background children in Australia (Bissoonauth 2018: 64), attitudes may be linked to religious identity and socio-cultural affiliation as well as to professional aspirations, and by implication, to perceived needs (cf. Herdina & Jessner 2002).

While there is a clear dearth of research into the relationship between language attitudes and the larger social context (Enever 2009: 28), it is understood that attitudes are context-dependent, i.e., they develop in a given socio-cultural frame or setting and are shaped by the people and events around them (Jessner & Mayer 2017: 91; Munoz 2014: 25). Csizér & Kormos (2009) note that young learners' experience of interactional encounters with speakers of other languages can influence both their disposition towards the target language and their attitudes towards the speakers of a given L2 and their culture. The tentative conclusion is then that contact with the L2 community, can "affect learners' motivated behavior", and "the energy and effort they are willing to put into L2 learning" (Csizér & Kormos 2009: 63). Crucially, this must be taken to apply in particular to multilingual contexts where the complexity of the socio-cultural and socio-political fabric is confounded by the presence of several languages and speech communities. In settings where historical liabilities and social tensions weigh heavily, intergroup relations and SLA (and multilingual acquisition) are additionally complicated by prejudice and ideological differences with language-related beliefs and attitudes reflective of political viewpoints potentially affecting individuals' motivation to learn a given language and/or engage with the L2/Ln community. While little is known about how young learners are affected by the complex

### 7 Exploring learner attitudes in multilingual contexts

spillover effects of politico-ideological narratives, the beliefs and values held by parents and significant others must be taken to have some (significant) impact on their attitudes and language learning motivation. This said, an interesting phenomenon has been observed amongst Spanish youth. Woolard & Frekko (2013 in Lasagabaster 2017: 586) found that young people in Spain are increasingly distancing themselves from the prevailing nationalist rhetoric and are instead embracing a more cosmopolitan attitude which, it is to be expected, will translate into greater respect for linguistic and cultural diversity.

Investigating emergent multilingual learners' preferences in Ireland, Harris et al. (2009: 4) evidenced positive overall attitudes towards languages among primary schoolers participating in a pilot project which saw the introduction of a new L3 in the final two years of elementary school. 84% of the children in the study stated that they were glad to learn a foreign language in addition to L2. Positive attitudes towards L2 and L3 were also found by Henry & Apelgren (2008). However, they report more favourable attitudes towards the more recently introduced L3 compared to L2 in 10 to 12-year-olds in Sweden and interpret this as a sign that pupils perceive the newly introduced language as more exciting and fun than the by now familiar L2 (p. 618). In addition, Henry & Apelgren observed attitudinal changes over time with (girls' and boys') attitudes to both, L2 and L3, declining between grades 4 and 6 (p. 613). Dynamic changes of a similar nature are also reported in Cenoz (2004: 214).

More recent research into young learners' pragmatic awareness and attitudes in Valencia (Portolés Falomir 2015: 172) points to age as an important factor. The younger children in Portolés Falomir's study displayed more favourable attitudes towards both the minority language Catalan and L3 English, while older children showed a preference for the majority language Spanish. The author attributes these findings to (1) younger children's less biased and prejudiced stance towards (minority) languages (cf. Cenoz 2004: 213 for similar results), and (2) to an agglomerate of political, social and psychological factors (Portolés Falomir 2015: 172). An alternative explanation for the observed attitudinal discrepancy between differently aged children is advanced by Cenoz (2004: 214). She suggests that older learners may be dissatisfied with the more academic and grammarfocused instructional approaches typically provided to their age groups.

Comparing the attitudes of young learners in type A, B and D educational models in the Basque country, Lasagabaster (2005), evidenced important differences in learner attitudes amongst students in different linguistic models and with different home languages (Lasagabaster 2017: 585; see also Portolés Falomir 2015: 172). Relatedly, Latino-background students in Spain have been found to show a preference for Spanish and English as the dominant prestigious codes and less

### Barbara Hofer

positive attitudes towards Basque and Catalan as the more peripheral languages (Lasagabaster 2017: 589). Interview data have revealed a tendency amongst Latino youth to be critical of mandatory minority language instruction for all immigrants (Lapresta Rey et al. 2010 in Lasagabaster 2017: 589). Empirical evidence suggests that alternative multilingualism-oriented teaching approaches (such as CLIL) may have the potential to act as motivation booster effectuating important attitudinal shifts in learner-users of different ages. A study by Lasagabaster & Sierra (2009) in Lasagabaster 2017: 590), for instance, found that students in CLIL programmes exhibit more positive attitudes towards English, Spanish and Basque compared to students who do not receive CLIL instruction (see also §6.5 for a similar finding).

In summary, it is fair to say that the findings yielded by research into young multilingual learners' attitudes are far from conclusive. Moreover, as emerges from the above, differential research foci and settings render comparability of results extremely difficult and generalisation to other contexts almost impossible. In some way, making sense of findings can be thought of as resembling the task of combining puzzle pieces into a coherent whole whereby a major difficulty consists in filling in the (many) missing bits. The conclusion to be drawn at this point then is that much more research in the field is needed if we are to make progress in lifting the veil on the processes driving young multilinguals' willingness to engage with languages.

It is the aim of the present study to work towards this ambitious target. More specifically, the study looks to illuminate the complex associations between young emergent multilinguals' attitudes towards languages (learning) and the larger sociolinguistic and educational environment in which they are nested.

The current study forms part of a large-scale research project into young learner-users' multilingual competences, strategy use and motivations. Data evaluation for the study is still in progress. For obvious reasons, the focus of the present paper is more restricted in scope lying as it does on possible correlations between sociolinguistic/educational context and learner attitudes. In the following I give a brief overview of the socio-historical backdrop against which this research is set. Thereafter, I outline the study design and procedures.

### **6 The study**

### **6.1 The South Tyrol context**

An autonomous and officially trilingual province, South Tyrol is Italy's northernmost region bordering on Austria (of which it formed part prior to WWI).

### 7 Exploring learner attitudes in multilingual contexts

South Tyrol is home to three linguistic groups including a German-, Italian- and Ladin-speaking community. According to the 2011 census, the German-speaking community constitutes the numerically strongest group, making up 64% of the local population, followed by the Italian-speaking community with 24% and a small community of Ladin speakers at 4%.

De Angelis (2012) refers to the South Tyrol as an area of language conflict. The conflict she identifies can be traced to two incisive historical events: the Treaty of Versailles, which in the aftermath of WWI stipulated that South Tyrol be ceded to Italy, and the Fascist period during which the German minority suffered oppression and hardship at the hands of a Fascist regime who, in an attempt to italianise the region, decreed that the German language be banned from all public places and institutions, including schools.

The region's troubled past has resulted in the widespread fear that too much contact with the language of "the other" might, in the long run, lead to loss of the German language and identity (De Angelis 2012). This is very much reflected in a German-only ideology in German-language schools which have historically sought to keep classrooms as Germanophone as possible in the conviction that this would enhance students' competency in L1 German and, by extension, safeguard and reinforce the status of the German language in South Tyrol (Egger 1977). The instantiation of bilingual or multilingual instructional models, as have long become reality in the Ladin valleys and, more recently also, in Italian elementary schools in the region, have been successfully forestalled by the local authorities. In German-language schools, multilingual programmes are, as of yet, few and far between and tolerated rather than welcomed.

It is important to note that this monolingual bias extends primarily to the L2 Italian. Augmenting teaching hours for L3 English, for instance, is not contested in the same offensive manner as increasing teaching time for L2 Italian. This is partly due to the fact that some sections of the public (and the authorities) still regard Italian as the language of the enemy and warn against watering down article 19 of the Autonomy Statute, which guarantees the right of German-speakers to school education in their L1. For fear to lose voters, the political majority party (whose founding fathers had fought hard for the rights of the German minority group in the decades following WWII) tend to exercise cautious reserve when it comes to positioning themselves on this contentious issue. The demeanor of the German school board is similarly reticent. As for the parents, they are often in two minds about what to believe. However, for many, the concern is not so much a political one but is rather related to their offspring's academic achievement and job prospects in an increasingly global marketplace. Parents' support for more (and/or improved) second and additional language teaching and for multilingual educational programmes has grown appreciably over the years. This de-

### Barbara Hofer

velopment is important for two reasons, firstly because political decision-makers will not be able to ignore voters' wishes indefinitely (so, change on this front is bound to come sooner or later) and secondly because parents' positive stance on languages learning is bound to have a positive impact on their children's motivation and learning behavior (cf. Gardner 1985).

### **6.2 Aims and research interests**

Following from the above, the focus of the present paper is on language (learning) attitudes in young learner-users in varyingly multilingual life realities in South Tyrol. The aim is to ascertain whether children in diverse sociolinguistic and learning contexts and with different levels of exposure to L2 hold different attitudes towards languages and language learning, in particular towards L2 Italian. The following research question has been formulated for this purpose:

RQ: Does sociolinguistic and/or learning context, and level of exposure to L2 Italian, affect young learner-users' attitudes and motivations with regards to languages and language learning?

Based on previous research, it is anticipated that pupils' overall language attitudes will be positive. However, it is hypothesised that children in more multilingual surroundings (such as provided by the bigger and linguistically more diverse towns in South Tyrol) come to adopt more favourable attitudes towards the L2 and L2 community because they use the language more frequently and interact more with speakers of the L2 (and/or other languages for that matter), thus experiencing the functional and personal value of bi- or multilingual ability first hand. Identification with the L2 community (referred to as integrativeness in the pertinent literature) is known to be a strong predictor of learner motivation (Gardner 1985). As for attitudes towards L3 English, it is important to note that English as a foreign language does not carry the historico-political onus of L2 Italian, which is why attitudes to English are expected to be positive regardless of the context in which it is studied.

### **6.3 Participants**

209 children in their 5th and final year of primary school took part in the study. The children were on average 10 years old and were drawn from 10 Germanlanguage schools located in various parts of South Tyrol. Settings differ considerably in terms of their sociolinguistic and educational realities. While some children live and attend school in remote villages where they hardly ever encounter

### 7 Exploring learner attitudes in multilingual contexts

L2 Italian (or other languages) outside of the school context, others come from larger towns with a high concentration of L2 Italian and/or other-language speakers. Classroom composition also differs substantially from school to school with some classrooms being relatively homogeneous (German-speaking) and others clearly heterogenous (i.e. multilingual). To boot, some children study in mainstream educational models, while others receive multilingual instruction. Mainstream educational programmes typically provide subject matter teaching in L1 German, with Italian being taught as L2 for 3–4 hours a week (NB: in first grade only 1 hour of L2 Italian is provided). L3 English is introduced in 4th grade and is taught for a total of 2 hours per week. Conversely, in the multilingual programmes both German and Italian are (though with varying intensity) used as vehicular languages from first grade onwards and L3 English is taught as a foreign language (for 2 hours a week). In addition, there is a strong focus on crosslinguistic comparison and learning as a means of fostering language awareness and cross-language skills, and, by implication, multilingual competency.

On the basis of participants' sociolinguistic and educational backgrounds (and their resulting differential multilingual experiences), learners were grouped into 5 cohorts and positioned on a continuum ranging from the relatively monolingual to the decidedly multilingual.

Figure 7.1 shows how degrees of mono- or multilingualism have been operationalised for the purposes of this study. This figure provides an overview of the various learning/educational contexts and illustrates how the 5 learner groups are arranged on the continuum. The graph is meant to be read from top to bottom with the least multilingual setting at the top and the most multilingual context at the bottom.

As shown in Figure 7.1, the majority of schools are collocated at the less multilingual end of the continuum. Group 1, for example, comprises 5 schools, all providing subject teaching in L1 German and offering L2 Italian for 3–4 hours (from grade 2) and L3 English for 2 hours per week (from grade 4). In addition, pupils in group 1 live and study in relatively monolingual contexts with both the language constellation in their classrooms and the wider sociolinguistic surroundings being very homogeneous, i.e., Germanophone.

Further down the continuum, contexts get more and more multilingual with some schools (groups 2 and 3) being located in areas where there is a high concentration of L2 (and/or other-language) speakers in the neighbourhood (NB: groups 2 and 3 differ from each other in terms of classroom composition and the wider sociolinguistic ecology).

The school constituting group 4 is situated in a comparatively monolingual environment (see Figure 7.2) but purposely compensates for the demographics-

### Barbara Hofer

 

*Least multilingual learning/educational contexts* 

Group 1: Learners in traditional educational programmes and in relatively monolingual classes and sociolinguistic contexts: (schools 1, 2, 3, 4, 5)

Group 2: Learners in traditional educational programmes in a relatively monolingual classroom but a strong presence of the L2 outside of school (school 8)

Group 3: Learners in traditional educational programmes and with high levels of contact with the L2 Italian in and outside of the classroom (school 7, 9)

Group 4: Learners in a multilingual programme with relatively monolingual classroom composition but otherwise relatively monolingual (i.e. German) sociolinguistic context (school 6)

Group 5: Learners in a multilingual programme with multilingual classroom composition and multilingual sociolinguistic contexts (school 10)

y *Most multilingual learning/educational contexts*

Figure 7.1: Learning/educational context

related lack of contact with L2/n by providing for multilingual (cross-language and awareness-focused) learning within the school walls.

Group 5 represents the most multilingual context with pupils studying in a distinctly multilingual classroom (both, in terms of the linguistic diversity of its student population and in terms of the teaching approach) and living in surroundings where L2 Italian and/or other languages are heard and spoken on a daily basis.

It is important to note that the two schools providing multilingual instruction, i.e. school 6 (= group 4) and school 10 (= group 5; see Figure 7.2) have a policy of fostering multilingualism and multiliteracy, which clearly qualifies them as multilingual schools (Cenoz 2009: 32). School 10 (i.e., group 5) additionally provides for bilingual (German-Italian) literacy instruction from grade 1. In contrast to the distinct multilingual policy promoted by these schools, traditional or mainstream educational models do not (generally or overtly) aim at multilingualism and typically contend themselves with offering some (limited) L2 and L3 instruction. In these schools, promoting learners' competency in L1 German tends to take priority over second and foreign language learning.

Figure 7.2 further zooms in on the sociolinguistic and sociocultural contexts in which the schools are embedded. It shows that while some schools are located in areas (mostly small villages) with a mere 2% of Italian speakers in the immediate *Relatively monolingual sociolinguistic context (low concentration of L2 speakers). Source: 2011 census.*

ݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݓݔ ݒ School 1: 2.18% Italian speakers; 97.82% German speakers; 0% Ladin speakers School 2: 3.37% Italian speakers; 96.51% German speakers; 0.12% Ladin speakers School 3: 4.58% Italian speakers; 95.42% German speakers; 0% Ladin speakers School 4–5: 7.91% Italian speakers; 91.80% German speakers; 0.29% Ladin speakers School 6: 14.91% Italian speakers; 83.14% German speakers; 1.95% Ladin speakers School 7: 70.42% Italian speakers; 29.07% German speakers; 0.51% Ladin speakers Schools 8–10: 73.00% Italian speakers; 26.29% German speakers; 0.71% Ladin speakers *Highly multilingual sociolinguistic context (high concentration of L2 speakers)*

Figure 7.2: Sociolinguistic context

surroundings, others (in larger towns) are situated in areas with a percentage of well over 70% of Italian speakers. The percentages provided refer to the presence of Italian, German and Ladin speakers in the villages or towns in which the given schools are located. Again, the graph is intended to be read from top to bottom with the (relatively) monolingual contexts at the top end and the multilingual settings at the bottom end.

### **6.4 Instruments and procedure**

To gain insight into young emergent multilinguals' language-related attitudes and motivations, a questionnaire comprising a total of 26 items was administered (due to space restrictions, only 4 of these are discussed here). The questionnaire is divided into three sections, one focusing on language learning in general, one on learning L2 Italian and one on learning L3 English. The questionnaire was designed by the present author for the purposes of this research and probes into participants' attitudes towards their L2 and L3, their perceptions relative to their own language learning experiences (in and outside of school), and relative to themselves as learners. In two open questions, the questionnaire further elicits children's beliefs about the importance of learning Italian (L2) and English (L3), and about what they like best about learning languages.

### Barbara Hofer

The questionnaire was administered during school time under the supervision of the class teacher. Prior to administration, parents' consent was obtained, and pupils and parents were informed about the aims of the study. Pupils were given 30 minutes to complete the questionnaire. Completion of the multi-item questionnaire required them to indicate (on a 4-point Likert-like scale) whether and to what extent they agree or disagree with statements such as:

	- b. Learning languages comes easy to me
	- c. I am afraid of making mistakes
	- d. I like learning about other languages and cultures
	- e. When I am older, I want to learn more languages
	- f. I like learning Italian
	- g. I like learning English
	- h. I would like to be able to speak Italian really well one day
	- i. I would like to be able to speak English really well one day
	- j. I like my Italian class
	- k. I like my English class
	- l. I think I am good at learning Italian
	- m. I think I am good at learning English
	- n. My parents say that it is important to learn Italian
	- o. My parents say that it is important to learn English

In addition to completing the questionnaire, 42 pupils (recruited from 10 schools) took part in a short ten-minute interview in which they were asked questions about their use of strategies and how they learn languages. As part of the interview, children indicated with a smiling or stern face whether and how much they liked or disliked learning languages. In a second step, pupils then explained how their drawings were to be interpreted.

Interviews were semi-structured and were conducted during lesson time in a quiet part of the respective schools. Children were interviewed individually. Participation was voluntary. The working language was German. All interviews were digitally recorded, transcribed and coded according to thematic themes.

For the qualitative evaluation, interviewees' utterances were categorised as positive (+), negative (−), or positive with some reservation (~). The overall score for comments was built cumulatively and is presented in Table 7.5 (page 157).

In the following, I present first selective results relative to children's attitudes to L2 Italian as obtained from the analysis of the questionnaires and interviews.

### 7 Exploring learner attitudes in multilingual contexts

### **6.5 Results**

### **6.5.1 Quantitative findings**

The statistical analysis was performed utilising SPSS. Since the data are not normally distributed, the Kruskal-Wallis test as the non-parametric alternative to the one-way between-groups analysis of variance was used. To calculate the strength of the relationships between the variables, Kendall-Tau-b/Spearman correlations were applied. The significance level is 0.05.

While overall, pupils' attitudes to L2 Italian are very positive (with 45.6% of children stating that they like Italian very much, 49.3% reporting they are very glad to be learning Italian, 61.6% indicating they are highly motivated to study Italian, and 45.6% declaring they very much like learning L2 in school), there are significant between-group differences as revealed by the Kruskal-Wallis test. The groups at the more monolingual end (i.e., groups 1 and 2) score significantly lower compared to the groups at the more multilingual end (groups 3, 4 and 5). Statistical differences obtained for pupils' answers relative to their attitudes to learning Italian ("I like the Italian language", = 0.038; "I am glad to learn Italian", = 0.034; "I like learning Italian at school", = 0.015) and relative to their (LL) motivation ("I would like to know Italian really well, that's why I study a lot", = 0.036). The crosstabulations in the following detail the pertinent results according to groups. Results are organised around 4 items drawn from the questionnaire.

Table 7.1 presents the results for participants' response to the questionnaire item "I like the Italian language". The Kruskal-Wallis test yielded a significant between-groups effect for this item at = 0.016, = 0.168 (very weak follow-


Table 7.1: Crosstabulation attitudes: "I like the Italian language"

### Barbara Hofer

Answers (%) Group "No" "A bit" "A lot" "Very much" 1 98/101 1.0 13.3 40.8 44.9 2 19/19 5.3 0.0 63.2 31.6 3 54/55 0.0 11.1 38.9 50.0 4 17/17 0.0 0.0 23.5 76.5 5 15/15 0.0 0.0 33.3 66.7 Total 203/209 1.0 9.4 40.4 49.3 = 0.015, = 0.170 (very weak effect) Cramér's V test: = 0.070, = 0.181 (very weak effect)

Table 7.2: Crosstabulation attitudes: "I am glad to learn Italian"

ing Spearman). As can be gleaned from the percentages reported, the children in groups 3, 4 and 5 (positioned at the more multilingual end of the educational and/or sociolinguistic continuum) attained the highest scores at 57.4%, 64.7%, and 53.3% respectively. No child in groups 4 and 5 reported not liking L2 Italian. Groups 1 (38.4%) and 2 (26.3%) score significantly lower than the groups at the multilingual end.

Table 7.2 reports the results for pupils' answers to the item "I am glad to learn Italian". The Kruskal-Wallis test yielded a significant between-groups effect for this item at = 0.015, = 0.170 (very weak following Spearman). An additionally run Cramér's V test revealed a sign. level of = 0.070, = 0.181 (very weak effect). Pupils in group 4 reach the highest score with 76.5% indicating they are very glad to be studying Italian L2. At 66.7%, the score for group 5 is the second highest. In group 3 the percentage is 50%. No pupils in either of the three groups reports not being happy to learn Italian. Scores for groups 1 and 2 are significantly lower (44.9% and 31.6% respectively) than for groups 3, 4 and 5.

Table 7.3 provides the results of children's answer to the questionnaire item "I would like to know Italian really well, that's why I study…". A significant effect was evinced at = 0.018, = 0.165 (very weak effect following Spearman). The additionally run Cramér's V test revealed a sign. level of = 0.003, = 0.220 (weak effect). 82.4% of pupils in group 4 reported a high level of motivation to improve their current level of L2. In group 5, the percentage is 66.7%. With 72.2% of pupils indicating they are working very hard to improve their L2, group 3 surpasses group 5 on this item. No child in groups 4 and 5 reports not wanting to work hard in order to improve their L2 proficiency. Scores for groups 1 and 2 are visibly lower (56.1% and 36.8%).

### 7 Exploring learner attitudes in multilingual contexts

Table 7.3: Crosstabulation attitudes: "I would like to know Italian really well, that's why I study a lot"


Table 7.4 shows the results for respondents' reply to the item "I like learning Italian at school". A significant group effect was evidenced at = 0.002, = 0.219 (weak). Cramér's V test indicates significance at = 0.115, = 0.172 (very weak effect). 76.5% of children in group 4 report liking to learn L2 at school. The percentage for group 5 is 53.3%, that for group 3 is 51.9%. No pupil in any of the three groups reports not liking their L2 lessons. Groups 1 (36.4%) and 2 (42.1%) score much lower than the groups at the multilingual end.

The ensuing section enlarges on the qualitative findings which serve as a complement to the statistical data.

Table 7.4: Crosstabulation attitudes: "I like learning Italian at school"


### **6.5.2 Qualitative findings**

As emerged from the interview data (and more specifically from children's explanations of their drawings), pupils' responses are overwhelmingly positive confirming the results yielded by the statistical evaluation of the questionnaires. Only a handful of children reported not enjoying language(s) learning (5 out of 42). It is indicative that it is only children in group 1 (the group at the more monolingual end) who express negative views about language(s) learning. Reasons therefor seem to be linked to the efforts required to build and maintain adequate or required proficiency levels in L2/L3 (see examples below). One is inclined to conjecture that participants in group 1 may be somewhat less conscious of the functional utility and personal enrichment of additional languages.

*Sprachen zu lernen gefällt mir nicht soo.* (NOESch) 'I don't really like learning languages.'

*Es ist halt schwierig Sprachen zu lernen, ja und in der Schule da ist das Sprachenlernen auch das anstrengendste.* (FELGr) 'Well, it's difficult to learn languages, and in school it's really hard work.'

*Es ist vielleicht viel Arbeit […] Es gefällt mir nicht so gut.* (MIRISch) 'is perhaps hard work, don't like it that much.'

*Aber Sprachen lernen ist nicht so mein Ding.* (LUCIKna) 'learning languages is not really my cup of tea.'

Most children, however, report enjoying languages and language(s) learning. As can be gleaned from the comments below, primary schoolers show a tendency to perceive languages as cool, fun, and exciting:

*Weil ich die Sprachen toll finde.* (MOMBr) 'Because I think languages are great.'

*Ich finde Sprachen sehr interessant, weil man auch etwas dazu lernt.* (STEFLei) 'I think languages are very interesting because you always learn something new.'

*Sprachen sind cool.* (DAVSch) 'Languages are cool.'

*Sprachen gefallen mir eigentlich gut, es sind viele neue tolle Wörter, besonders English, da muss man Wörter auch anders aussprechen, das gefällt mir.* (LAURIZo)

### 7 Exploring learner attitudes in multilingual contexts

'I quite like languages, there are many new and interesting words, especially in English, there you also have to pronounce words differently. I like that.'

*… also Sprachenlernen find ich toll* (ROSTau) 'Language learning is cool'

Pupils' L(L) attitudes (as proffered in the course of the interviews) are summarised in Table 7.1. As can be seen from the figures adduced, the children in group 5 are the most positive about language(s) learning with all children in this group stating that they enjoy languages and language(s) learning. As for groups 2, 3, and 4, their attitudes are also predominantly positive with one child in each group, however, voicing reservations. Attitudinal patterns differ even more for group 1 where 5 children (out of 18) declare that they do not like language(s) learning and 5 more children express some degree of reservation (of the sort "it's hard work", "it requires a lot of effort", etc.).

Table 7.5: Language (learning) attitudes as reported in the individual interviews. +: positive; −: negative; ~: positive but with minor reservation


### **6.6 Discussion**

The present study examined the associations between sociolinguistic/learning context and young multilingual learner-users' attitude to L2 Italian in South Tyrol. As a consequence of the specific demographics and spatial distribution of the linguistic groups, primary schoolers enrolled in German-language schools in South Tyrol may have different levels of exposure to Italian depending on whether they live in peripheral or more central locations. Accordingly, contact times differ substantially for children in remote rural villages vs the larger towns

or capital. In addition, (language) education in mainstream versus multilinguallyoriented classrooms can differ considerably. Following from these markedly differential initial (macro-level) conditions, it was hypothesised that pupils' language learning experience reflects in their L(L) attitudes and motivation to learn Italian. This hypothesis is borne out by the quantitative and qualitative analysis.

While overall, participants have been found to hold favourable attitudes towards L2 and L2 learning (thus confirming extant research in the field), a number of interesting findings have emerged from which equally interesting conclusions can be drawn. First, it is noteworthy that group 4 scored highest on all 4 items, i.e. they exhibit the highest level of positive attitudes towards Italian and the highest level of motivation to study it, surpassing even group 5 as the most multilingual cohort. On the learning continuum, group 4 is positioned at the more multilingual end. The children in this group benefit from a multilingual learning environment and a holistic integrative teaching approach, but the larger sociolinguistic ecology is relatively monolingual. This finding is somewhat unexpected. However, what we might infer therefrom is that even (comparatively) limited contact time within the framework of an integrated multilingual education programme can have significantly positive effects on young learner-users' LL attitudes and motivation.

Second, it is particularly worthy of consideration that group 3 overtakes group 5 (but not group 4) in the score for item 1 ("I would like to speak Italian really well, so I study Italian very much"). This may be linked to a perception among the children in group 3 that their level of proficiency does not match that of more competent (or native) Italian speakers in their immediate surroundings. Living in a highly multilingual environment (in and around the capital), the children in group 3 may have developed a certain awareness of the knowledge gap that separates them from more competent others, together with high expectations of themselves as learner-users of Italian. In addition, they may, owing to their distinctly multilingual life experience, be astutely aware of the necessity and benefits of mastering the language of their neighbours. Embedded as they are in a relatively monolingual lifeworld, pupils in group 4 may feel similar about the need to study hard in order to make headway in Italian (which may be why they scored highest on this item). On a purely speculative note, group 5 may be more confident about their Italian-language competency, which could be linked to their school's holistic teaching approach, and acknowledgment of multilingual competences as by nature asymmetrical and domain-specific. Alternatively, it may also be the case that cohort 5 has come to adopt a more multilingual mindset which leads them to view their L2 competence from a resource-oriented rather than a deficit-oriented perspective.

### 7 Exploring learner attitudes in multilingual contexts

Concluding, the main findings can be summarised as follows. As revealed by the quantitative analysis, pupils' overall attitudes towards (learning) Italian are predominantly positive. The same holds for their motivated behaviour and their willingness to invest time and effort to improve their level of proficiency in Italian. The statistical results are in line with the findings yielded by the interview data. Overall, the outcomes of the current study then substantiate the findings reported by previous research which found a high appreciation for language(s) among young learners (Nikolov 1999, Shameem 2004). The results also reinforce the notion that important attitudinal benefits can accrue from early multilingual learning (Hélot 2008).

### **7 Concluding remarks**

This paper has reported on work in progress. A selective review of the data set however allows for first important conclusions to be drawn. Overall language (learning) attitudes and motivation amongst the primary schoolers participating in the study turned out to be highly positive. Sociolinguistic and educational context, as it has been established, make a substantive difference in so far as children in more linguistically diverse settings with more exposure to L2 Italian hold more positive attitudes and are more motivated to expend time and effort to improve their L2 language skills. Based on the findings yielded, it can be concluded that type of early multilingual experience and amount of contact time with L2 (and/or other languages) do affect children's LL attitudes and motivation. An important implication of this research is that formal multilingual learning programmes can act as attitudinal and/or motivational booster and thus compensate for lack of interactional opportunity in the immediate surroundings. Consonant with recent complexity- and dynamic systems-informed theorising and on the understanding that positive attitudinal patterns can be effected through multilingual instructional programmes, the present author calls for the wider implementation of multilingual models in South Tyrol (and beyond), with a vision to fostering not only learner-user attitudes and motivation (and ultimately learning outcomes), but also social cohesion and intercultural respect.

As for the larger research project underway, the next step will be to look into the intricate relations between contextual factors, learner attitudes and multilingual ability. Working within a dynamic systems/complexity theory framework, we do not anticipate simple straightforward cause-effect relationships but complex interdependencies and interactions between the multiple factors involved.

### **Abbreviations**


### **References**


## **Chapter 8**

## **Building bridges between languages: How students develop crosslinguistic awareness in multilingual learning settings**

### Gisela Mayr

Free University of Bolzano

The present qualitative case study investigates the acquisition of increased crosslinguistic awareness in a multilingual learning setting, and the different development attributable to the linguistic backgrounds of emergent multilingual students. The study was carried out in a secondary school in South Tyrol, belonging to the German-speaking school system, where plurilingual task-based modules were inserted in regular language lessons. The languages involved were: German, Italian, English, French, Latin and Ladin. Students were involved in complex plurilingual problem solving processes during the elaboration of the language production. Thanks to this, crosslinguistic awareness could be trained and fostered, as plurilingual negotiating processes arose from the plurilingual and multimodal input provided by the teacher.

### **1 Introduction**

### **1.1 Crosslinguistic awareness: A language learning competence**

Crosslinguistic awareness, according to Jessner, is the ability of multilinguals to make implicit or explicit use of the connections and overlappings that exist between the different language systems in the human brain during language production and use (Jessner 2006: 116). It is a conglomerate of competences in the performance domain, including, on the one hand, the selective analysis of linguistic

### Gisela Mayr

structures accompanied by a repertoire of abilities that allow the speaker to successfully handle deficiencies in language use and the ability to anticipate arising problems in communication by identifying and selecting appropriate strategies to overcome them. On the other hand, this presupposes a more general monitoring of the language-processing process and production, aimed at error analysis and correction as well as the optimization of communication (Luo et al. 2010; De Angelis & Dewaele 2011; Herdina & Jessner 2002). It also implies the capacity of multilinguals to apply metalinguistic abilities to a certain context by activating specific linguistic resources across their languages.

Recognizing, and availing oneself of crosslinguistic interaction (CLIN), entails the ability to make use of transfer and inference as well as code-switching, codemixing, translanguaging, and crosslinguistic borrowing in oral communication. This means all interlingual correspondences, regularities and contrasts are employed to optimize communication. Multilinguals enact multilingual compensatory strategies (Jessner 2006: 87), and are therefore able to switch to a metamode, where language production is constantly surveyed. (ibid. 87). The resulting metacognitive translingual transfer is, to a large extent, unconscious (De Angelis & Dewaele 2011; Gibson & Hufeisen 2011; Vidgren 2013), but if made conscious can become a competency, and a strategy to master complex multilingual communicative situations. It is also assumed that multilingual learners, in contrast to monolingual learners, not only rely on L1 for transfer, but in many cases prefer L2. The reason is that L2, unlike L1, is a consciously acquired language and can more easily provide comparable structures and words (Herdina & Jessner 2002: 79; House 2004: 64; Müller-Lancé 2006: 178f). This so-called L2-factor leads to the assumption that the L3-acquisition process differs substantially from the L2-acquisition process (Hufeisen 2011; Cenoz 2013; Herdina & Jessner 2002). Comparative studies have shown that early bilingualism has a positive influence on any further language acquisition (Cenoz & Valencia 1994; Lasagabaster 1997; Pilar Safont 2003; Ringbom 1987). It is demonstrated that test- persons achieve higher proficiency in L3 English if they have achieved a high proficiency level in L1 as well as L2 (De Angelis & Jessner 2012). The latter study also shows that the L2 factor, and its influence on subsequent language acquisition is, to a large degree, dependent upon proficiency and the psychotypological perception of L2. If L2 is perceived as distant and unfamiliar, this will inhibit transfer from L2 to subsequent languages.

However, in the ideal case, crosslinguistic awareness can lead to increased metalinguistic awareness, which in turn allows for cross-lingual lexical consultation. In this case, procedural knowledge, the knowledge about how something is done, as well as declarative knowledge, the basic knowledge about something, is drawn

### 8 Building bridges between languages

mainly on L2, and L1 loses its predominant role for transfer. As a result, grammatical error recognition and analysis become much more effective and productive (Bialystok 2004; Gibson & Hufeisen 2011). This induces multilinguals to be more risk-taking during language production, since they can avail themselves of increased cognitive control. Cenoz postulates the existence of different levels of metalinguistic awareness, which have an effect on the plurilingual lexicon and support multilingual speakers (Cenoz 2013). They can rely on a metasystem: the interlanguage formed during L3 acquisition (Herdina & Jessner 2002: 131–161).

All these competences, however, are not to be considered as given, but rather develop when language systems interact with each other. Therefore, crosslinguistic awareness can be exercised by adopting specific plurilingual learning settings that promote multilingual strategies in the classroom. Thanks to this, the consciousness of the different interrelations existing between diverse language systems could be trained also in monolinguals, and their ability to draw on implicit and explicit declarative as well as procedural knowledge to determine similarities, and differences raised. This initiates a process in which all languages may take on the role of bridge languages, and assume different functions, according to the specific needs of the speaker. Due to their increased ability to handle multilingual discourse, multilingual speakers in such a context often assume model role function in conversation and positively influence the communicative competence, and language learning process of less proficient students. Thus, multilinguals can practice their abilities in a learning setting, where there normally is no space for transfer and CLIN, and at the same time take over the role of mediators between languages, cultures, and worldviews. This way, they can initiate individual learning processes that best comply with their multilingual biography, and in the meantime enhance the learning process of monolingual learners. (Jessner 2006).

Relatedly, crosslinguistic awareness is not only relevant on the linguistic level, but can initiate transcultural learning as well, by acknowledging that language acquisition processes strongly depend on the emotional dispositions of the learners towards the individual languages and cultures (Burwitz-Melzer 2012: 29). This implies that languages are socially and politically charged, and their development associated with the historical and cultural development of a certain community at a given time. Crosslinguistic awareness, then, as intended in its political and social dimension, implies the critical questioning of power structures behind common language use by different subjects, social classes and cultures, as well as the consciousness of the presence of gender issues in language use (ibid. 29; Morkötter 2005: 28f). This ultimately leads to the acceptance of differences in all their

### Gisela Mayr

forms, and fosters an inclusive attitude with regard to the different languages of instruction in school, and in the society in general.

In addition to these cognitive and socio-cultural aspects of crosslinguistic awareness, the present study wants to analyze the effect of plurilingual interaction in linguistically heterogeneous groups, with regard to the possible change of attitudes and dispositions of the individual speakers. This is only made possible by reflecting critically upon each person's language-learning history in comparison with that of others. In particular, monolingual learners can recognize attitudes, and dispositions of multilingual learners, and adopt them, with the result that these monolinguals can then more actively and autonomously participate in the multilingual problem-solving processes.

In order to implement the above-mentioned forms of learning, plurilingual learning settings are required, which put a focus on the identity of the learner and their language biography. It is necessary to take account of the order, and mode of acquisition of each individual language, which altogether forms a complex system of languages and emotions. This means that each newly acquired language gives access to new experience and a new perception of the world. The acquisition of a second or third language is associated by the awareness that each person's own identity is not unalterable but can change and expand, and that this expansion grows in complexity with each further language (Reich & Krumm 2013: 88). Such forms of crosslinguistic awareness imply the consciousness that each language biography is related to attitudes and emotions, and that these influence the way we communicate and perceive the world.

### **1.2 Plurilingual task-based learning**

The tradition of task-based language learning (TBLL) can be seen as a further development of the communicative approach to foreign-language teaching (FLT), and has established itself as one of the most successful innovations in FLT over the last few decades. In task-based language teaching (TBLT), learners face meaningful and relevant tasks, and use the target language to solve real-world problems in a functional way. It is no longer the aim of teaching to impart grammatical structures, which in TBLT are acquired indirectly while students deal with diverse contents and tasks that involve them emotionally and cognitively (Hallet 2012). Learners cooperate autonomously in groups, this way initiating problem-solving processes, where solutions are found in an act of collaborative learning, and where meaning is continuously re-negotiated among peers. Thus, a change in perspective takes place, and learning is no longer seen as the transfer of knowledge mainly provided by the teacher, but rather as the transformation

### 8 Building bridges between languages

of knowledge, performed by the learners themselves in autonomy (Ellis 2003). This form of learning raises intrinsic motivation, and in the meantime focuses on the acquisition of learning strategies. It leads to the forms of self-reflection, which give the learner the opportunity to see themselves at the center of action, and make autonomous decisions with regard to their personal learning process. Using foreign languages as a realistic means of communication in the classroom gives the learner the impression of being a competent speaker by boosting selfconfidence. This way the ideal conditions are created for the development of the communicative competence along with the forms of social learning (Dewaele 2010: 84f).

However, until now there have been no attempts to adapt TBLT to the needs of plurilingual teaching and learning at school, although various researchers have explicitly requested a change in perspective from monolingual to plurilingual forms of learning (Kramsch 2009; Hallet 2015; Martinez & Schröder-Sura 2003). The present research study is a first attempt to link TBLT to forms of plurilingual learning, thus fostering the ability of the individual to enact cultural and linguistic inclusion in a society characterized by pluralistic discourse. Plurilingual TBLT is an approach, where the input for the students is not only multimodal but also provided in more languages (5 languages in this case: German, Italian, English, French, Latin).

While solving the task, students work with documents in more languages about the same or similar topics. They compare and analyze these documents in the course of the task solving process, complete the task and finally elaborate a plurilingual output (Mayr 2020). This promotes a form of learning, in which the learners develop the ability to communicate not only in different languages simultaneously, but also to mediate between these languages, and cultural reference systems in a process of continuous comparison (Meißner & Morkötter 2009: 88). Being provided with plurilingual and multimodal input, the learners are given the opportunity to work with more than one language, and adapt their language use to different communicative needs or purposes. In this process of continuous mediation, and translation from one language to the other, the learners can develop and activate their multilingual repertoire, while in the meantime also increasing their crosslinguistic awareness. Thanks to the plurilingual TBLTapproach, it is therefore possible to promote crosslinguistic awareness, which also according to Allgäuer-Hackl & Jessner (2013) can be acquired in a process of learning, under the condition that all languages are incorporated, and critical reflection is promoted. This is also the main objective of the multilingual taskbased learning project introduced in this study.

### **2 Research project and questions**

This case study was carried out at a secondary school in Bolzano, South Tyrol. The region is situated in Italy, and is characterized by a minority language situation, as a minority population of German speaking people live there. This population has been granted self-determination to a large degree, and can avail itself of an autonomous German-speaking school system, where Italian is L2 and English is taught as L3. However, due to a troubled history of coexistence between the two linguistic groups, there is still a negative influence on the psychotypological perception of Italian L2 as will be shown in the data analysis. Due to the L2-factor and its importance for language learning, this affects subsequent language acquisition. The students in the class subject of the case study, follow a particular language curriculum described below:


The students, thus, on average should have at least a B2 level (CEFR) in Italian, although in many cases the actual proficiency level is lower, a B1–B2 level in English and an A2 level in French.

For the data collection, a group of four students was observed during plurilingual TBLT classes over a period of eight months, when five plurilingual modules of the duration of 10 hours each were inserted in regular language classes.

Before the beginning of the project, all 22 students were administered a questionnaire on their language biography, and studies. In order to comply with the principle of the maximum possible diversification, and to be able to observe how the learning process develops under different conditions, it was necessary to find students with divergent social, cultural and linguistic backgrounds, as well as the proficiency levels and character. Four students were chosen on the basis of the outcomes of the questionnaires. The following criteria were adopted:


### 8 Building bridges between languages


### Additional criteria:


During the implementation phase, three stimulated recalls (SR) were carried out with each of the four students at a two-month interval. In addition, audio and video recordings of the peer interaction among students during the negotiation processes, and the outputs were made and analyzed. The discourse analytical approach adopted was the "documentary method" (Bohnsack et al. 2013). At the end of the project, a retrospective interview was carried out with each student. The data analysis focused on the following research questions:


In the following outline of the data analysis, the statements of the students regarding crosslinguistic awareness, and the final interview are summarized, and integrated by the analysis of the audio and video recordings of the discursive processes. The students were anonymized, and given fictitious names.

### **3 Data analysis**

### **3.1 Student 1: Amelie**

### **3.1.1 Background**

This student's social background is characterized by diglossia, which means that she is used to switching from a standard German variety to the South Tyrolean dialect, commonly used in lifeworld discourse but also at school. This implies that she has the ability to switch from one variety of German to the other according to specific communicative needs. Amelie speaks the dialect at home, with her

### Gisela Mayr

friends, and at school with her peers. A standard variety of German is usually used within the classroom, both during the lessons themselves, and when talking to the teachers. Amelie comes from a rural region, and therefore rarely has the occasion to speak Italian (the second most widely spoken language in South Tyrol). She has learned Italian only at school. Even though she has been learning Italian since her first year in primary school, she still considers her proficiency as only sufficient to get by, and perceives the language as distant and unfamiliar.

### **3.1.2 The learning processes**

The student claimed that it was very difficult for her to learn to switch from one language to the other. At the beginning of the study, the language activation mechanism was still in a monolingual mode (Grosjean 2007). In fact, when claiming that "when a text was in German it was not possible for me to speak about it in English for example", (translation from German by the author), she referred explicitly to this difficulty. Only during the project did Amelie shift from a monolingual mode to a multilingual one. Thus, later she said that "the more you speak the better you get used to switching ... and if you don't how to continue, another language may be helpful". Amelie here found that the languages in her repertoire complemented and supported each other. This also implied that during the language production process, code-switching assumed a scaffolding function in order to help her cope with difficult situations. She also realized that code-switching could be used in a multilingual discourse for different strategic purposes.

She stated that she had learnt to use previous knowledge and experience as well as new crosslinguistic knowledge to help accelerate her learning process in all languages. Amelie learnt that it was helpful to expose oneself to challenging, multilingual situations, and develop compensatory strategies to handle them.

The data show that the student activated self-regulatory forms of learning, which meant that she activated strategies and metacognition to identify areas of learning, and sought out the most suitable strategies to enhance her learning process. As a result of this, the student used predominantly French in the negotiation processes at the beginning of the project. When she said that "the French pronunciation is the most difficult one, and thanks to the practice during these modules it became better and better" she demonstrated that she chose a language for practice she didn't often have the opportunity to speak. Her language choice and progressive improvement showed that multilingual learning increased the student's consciousness about her different levels of proficiency in the different languages of her repertoire. Thanks to the learning setting, she was

### 8 Building bridges between languages

capable of taking action and found new ways of learning. Amelie learnt to reflect on her own language production and to observe it from an outside perspective so as to critically analyze and correct it wherever necessary (in fact she continuously looked up the correct pronunciation for French words). Furthermore, she understood that she could draw on her functional multilingualism to handle demanding situations and mastered also sub-areas of technical languages.

The audio recordings show that, during the learning process, she would first assess her interlocutors' language proficiency, and then adapt her language use to their needs. This allowed her to reflect on her own attitudes and habitual language use. She showed that she had acquired an inclusive attitude towards students who were new in the class and didn't speak any Italian when she said that "always when we spoke with him we avoided Italian, because we didn't know whether he understood us".

Amelie herself had quite a negative psychotypological perception of Italian before the start of the project, as the findings of her questionnaire showed. As a result, her motivation to activate and use Italian only manifested itself in the last module. When she said "I started switching to Italian too, which I didn't do at all at the beginning", she signaled a turning point in her language-learning process, because, as she began to include Italian L2 in her active multilingual repertoire. The way she perceived the language had changed, and consequently the further language acquisition process had, too. The multilingual learning setting allowed forms of social learning and imitation that gave the student the opportunity to change her perception of Italian, thereby rendering it accessible to her as a source of transfer and CLIN for L3/Lx. The disposition of the student changed and she opened up to plurilingualism. This allowed her to resort more and more to crosslinguistic lexical consultation and to understand that there was occasionally no one-to-one correspondence in the meaning of words in different languages, and that sometimes it was simply not possible to translate a word from one language to another. She stated that especially when analyzing literary texts "when you switch from one language to the other, and you want to use the same word, you realize that you can't translate, but you need to find an appropriate word." In this context while she was trying to find an appropriate translation for the dialect word *tratzen* (the dialect word for tease) and she wasn't able to, Amelie also recognized that the South Tyrolean dialect was her language of emotional socialization (Pavlenko 2011), and that it allowed her to best express particular emotions. Plurilingual literary learning helped the student to tackl the problem of polysemy and ambiguity in the plurilingual discourse, while also developing a sense of transcultural awareness that contributed to her overall awareness of the plurality and heterogeneity as well as hybridity of cultures (Hufeisen 2010:

201). She also learned to influence the course of the conversation by strategically code-switching when she wanted to achieve a certain effect. For example, when her peers were losing concentration and started using the German dialect, she brought them back on track by switching repeatedly to Italian using the expression *ehm iniziamo?* ('can we start?').

### **3.2 Student 2: Sarah**

### **3.2.1 Background**

This student comes from the Gardena valley, where Ladin is still spoken, mainly as a heritage language. According to the assertions in her questionnaire, Sarah speaks Ladin and South Tyrolean dialect in her family, and also claims that Italian is spoken in a wider familiar context with relatives. She states that within her family she assumes different language roles, depending on her interlocutor. She never uses Ladin at school or in class. It is an unspoken rule for her to restrict the use of this language only to her family, and to the sphere of her strictly Ladinspeaking friendships. Because of this she also feels that part of her personality is excluded from her scholastic career.

### **3.2.2 The learning processes**

Sarah was a very active student and it could be observed that during the study she developed the ability to observe the conversation from a meta perspective, and monitored it. Sarah then intervened with regulating strategies when she realized that the conversation drifted off-topic and her peers were losing concentration, by repeatedly switching to Italian and using the imperative *concentriamoci* ('let's concentrate'). Sarah had become the reference person for all the questions and doubts concerning Italian, since this was the language she knew better than anyone in the group. She became a role model and used the deriving authority repeatedly to bring her peers back to concentration. Through her language behavior in the group, she uncovered linguistic hierarchies. Thanks to her repeated switches to Italian and French she induced her peers to reduce the use of English, the predominant language in this group, and use other languages more frequently. Sarah boosted the use of Italian, which was the language least often chosen at the beginning of the project. Sarah like Amelie often resorted to the languages she wanted to practice, and this way learnt to judge her linguistic knowledge in the different languages using translingual criteria. She compared her linguistic competences in the different languages, and made good use of the gained knowledge for her further learning process. When translating from one language into another, she resorted to her complete multilingual repertoire as a source. She

included all the languages of instruction in her learning process, so that in a continuous activity of mediation these could support and complete each other. This way language monitoring was activated, and supported her language acquisition. In the interview she stated: "now when I read a book in a foreign language, and I don't understand something, I simply deduce it from other languages, instead of using a dictionary. I even use Latin, it helps to be able to switch from one language to another, to solve problems".

Sarah learned to use previously acquired knowledge, strategies, linguistic and content-specific know-how, as well as procedural knowledge on different levels to cope with difficult situations, and to speed up her language-learning process. This student too used codeswitching as a scaffolding strategy in linguistically demanding situations during language production. This required her to evaluate her own linguistic abilities in the different languages correctly as well as those of her interlocutors, in order to monitor her language production, and intervene with corrections wherever necessary. Meaning was identified through a process of transcultural and crosslinguistic analysis and comparison. Sarah perceived her multilingual repertoire as a network of interrelations and intersections, which formed a whole, and continuously edited her language production. The statement "the project helped me to find different ways to express myself without having to switch to German all the time" implies that during the learning process, a generic multilingual repertoire could develop that made available different patterns and strategies to acquire the knowledge of the world, and thus form a critical personal opinion. Sarah gained more and more the ability to deduce unknown words from her knowledge of other languages. This meant that her ability to transfer knowledge and strategies from one language to the other had accelerated, which implied increased levels of language monitoring and crosslinguistic awareness.

Her heritage language, Ladin, however, remained excluded from the languagelearning process and found no place within the school. But she became conscious of this, and the fact that her language use depended on the situation and the social setting. Sarah acknowledged that she often used Ladin for transfer, and that it was always present but never actually used at school. Multilingual literacy training helped her realize that different literary texts belonged to different cultural reference systems and that, thanks to these plurilingual modules, different discourse worlds and genres were brought into contact with each other. Sarah learnt that different languages were associated with different emotions for her, and that these emotions had a biographical origin. For instance when using Italian she says that "it is the language I associate with my father, the memories I have of him speaking in front of an audience".

### **3.3 Student 3: Vera**

### **3.3.1 Background**

This student lives in an urban area and has a bilingual German/Italian background. However, the student does not define herself as bilingual in the questionnaire, and her assertions about the two languages are contradictory. Vera gives German a prominent position, even though her biography is clearly bilingual with a predominance of Italian in certain spheres. Her attitude reflects a different emotional perception of the two languages as well as a lack of awareness about the actual nature of her multilingual background.

### **3.3.2 The learning processes**

Vera was a very active and interested student, and in contrast to the other three students, already felt at ease in multilingual situations at an early stage of the project. This allowed her to become a role model for the other students in the group. She often employed codeswitching and translanguaging as discourse-strategic instruments, which was an indicator that her multilingual communicative competences were quite highly developed right from the beginning.

There were many examples where the student showed that CLIN was a means for her to monitor her spoken production and correct it when necessary. Right from the beginning she would make decisions about grammatical and lexical correctness, using crosslinguistic consultations in different areas. This demonstrated a high level of crosslinguistic awareness and an efficient language monitor. There were several examples in the recordings that showed that her multilingual repertoire was constantly present and that translanguaging was naturally part of her and her language production. She could apply her language monitor according to the specific linguistic needs that emerged during the communication. In the sentence "he who dares der wagt to crave means to want something" for example she uses translanguaging as well as paraphrasing in one sentence to explain a verse from a poem to her peers.

In the process of language mediation, Vera realized that it was not always possible to translate words from one language to another, and that meaning and connotation varied according to the historical development of the different languages. When she states that "when dealing with Christmas in different cultures, it becomes evident that it has many different meanings" she shows that language mediation, for her, meant playing with different meanings in the light of different cultural reference systems. This ability also allowed her to create new meaning by combining various connotative aspects, so for instance she created the term

### 8 Building bridges between languages

"traffic decoration" for "Christmas decorations," combining the German "Straßendekorationen" with the English word.

Vera was, to a large extent, indifferent towards language hierarchies. The multilingual setting helped her live out her plurilingual identity, which supported her in further language acquisition. The multilingual learning context provided greater clarity for her with regard to her own multilingual background. She could experience herself consciously as a multilingual subject, and identify herself as such in front of the others. In the meantime, she realized that the process of multilingual language production accelerated with time and that this reflected the degree of activation of her multilingual repertoire, she claimed that all the languages of her repertoire "seem as if they were one language only". On a semantic level, the student discovered that polysemy was linked to multilingualism and that meaning was often so multifaceted that it reflected reality "like a kaleidoscope".

Transfer, in this case, occurred on many different levels such as: grammatical, lexical, textual but also social and emotional level. On the lexical level, for instance, there was evidence that instead of just single words or expressions, Vera developed the ability to transfer chunks that could be combined and recombined within one language or between languages, according to her needs. This induced her to use multilingual metaphors allowing her to express meaning by combining and overlapping different metaphorical and semantic fields, thereby reaching a high degree of expressive complexity. The student could infer unknown meaning in forms of intercomprehension-learning by using her multilingual repertoire (Hufeisen 2004; Meißner 2005; Meißner & Morkötter 2009). According to her, multilingual learning allowed her to create new bridges between languages and these bridges helped her to reach *das Ganze* ('the whole'). A statement that showed increased awareness about the fact that the multilingual repertoire is a transient and ever-changing unity, made up of different languages intricately linked to each other. Thanks to working with multiple meanings, Vera could develop a refined perception of slight semantic nuances provided by different languages.

This student, like the two previous ones, developed the ability to better judge her interlocutors' linguistic proficiency and dispositions and consequently regulated her behavior so as to allow successful communication. Vera could identify herself as a multilingual speaker in front of the others, and use her ability to mediate between the different worlds at their disposal. This not only changed her but also the others' attitude towards the languages used in the course of the project. English was no longer the dominant foreign language, as was commonly the case due to its social prestige. Instead, other languages were perceived as equally important and interesting, and therefore became more familiar.

### **3.4 Student 4: Andrea**

### **3.4.1 Background**

Andrea grew up in a German-speaking family in an urban area. Within her family, she speaks exclusively the local German dialect. Her statements are in many cases contradictory but they show that she has a rather negative view of her language competencies in the other languages. Although she lives in an urban area, where German as well as Italian are commonly used, she claims that she almost never has the opportunity to use Italian in her free time, and that her acquisition of the language has been restricted mainly to school, and the interactions with the teacher in class. Andrea therefore perceives her learning of Italian to have plateaued.

### **3.4.2 The learning processes**

The student claimed that at the beginning of the project it was very challenging for her to switch between languages and that she always had to think about how to proceed. Over the course of the project, however, thanks to a habituation process, the use of multiple languages and codeswitching became easier and easier for her. Andrea stated that the activation of her multilingual repertoire did not only imply a quantitative enhancement in her spoken production, but also a qualitative enhancement because she learnt to use different languages as a communicative strategy, which meant that she used codeswitching to express specific meaning. Thanks to these strategies she was able to acquire new words in more languages, in fact she asserted that "I am able to express myself better, because I have learnt more complex words"

Over time, thanks to the forms of social learning, the students became aware of the fact that the error correction could take place in autonomy, and in a crosslinguistic mode, using more than one language to facilitate the process. Her statement that "sometimes if you translate a word you need to find a synonym, or to paraphrase it, or make a description, but this is not important the important thing is to make yourself understood", showed that she not only made use of compensatory strategies when needed, but thanks to the forms of imitation learning, she stored knowledge that she could then activate at a later point in time. Eventually, she tried to use Italian more and more, but her language production was characterized by frequent mistakes, which in most cases she was not able to correct. However, Andrea became aware of the fact that her Italian repertoire lacked colloquial forms, due to the exclusive use of this language in class. The student

therefore sought to increase her competency in this field as well, by using Italian in colloquial situations. This meant that, thanks to the multilingual learning setting, this student too acquired a more detailed awareness of her competences in the different languages, and used forms of self-regulated learning to adjust to her evolving needs. Andrea, like the other students, applied her functional multilingualism to enhance her communicative effectiveness, and was thereby able to adjust her communicative behavior to the requirements of her interlocutors.

Her psychotypological perception of Italian L2 changed towards the end of the project, since she perceived herself as a more competent speaker of this language, in fact she asserted that working with different languages at the same time "reduces prejudices". This allowed Andrea to make better use of L2 for transfer, which in turn gave her the possibility to play her part, when she was the only one in the group to find the Italian translation for *scar* (*ciccatrice*). At the same time the simultaneous use of more languages helped Andrea to overcome preconceptions, and to open up new access points to the different languages that "are all of the same importance". Thanks to her newly acquired plurilingual reading skills, she was able to transfer knowledge from one text to the other, and thereby realized that the construction of meaning in a reading process based on plurilingual inputs was much more complex than in monolingual reading. Contrastive multilingual reading exercises enhanced her critical view of historical and social phenomena, and she realized that the construction of meaning in the reading process in many cases relied on complex crosslinguistic and transcultural word knowledge. So for instance she realizes that the word *patriotic*/*patriotisch*/*patriotico*/ *patriotique* had profoundly different connotations in the different languages, and that this fact was attributable historical reasons. Andrea also acquired items of subject-specific vocabulary in in this context, and used them adequately.

At the end of the project, the student claimed that she had begun to perceive all the different languages as one single language, and that in addition to critically analyzing different languages and cultures, multilingualism in itself had become a new culture to her, and a new way of being in the world.

### **4 Conclusions**

To pick up on key aspects it can be said that plurilingual TBLT initiates different learning processes in the learners, depending on their language biography. This creates a plurilingual learning setting that helps learners activate their plurilingual resources, and make use of them in communication. This way, their linguistic repertoire is expanded, and both receptive as well as productive skills in all

### Gisela Mayr

languages are developed. This form of learning forces the learner to constantly compare both linguistic and cultural content, and consequently to develop and expanded awareness of the similarities and differences between languages and cultures. Students undergo a consciousness-raising process, that leads them to better judge their own linguistic competences in the different languages. They learn to use their multilingual repertoire as well as the interrelated resources at their disposal to optimize their language production. Therefore, with regard to the first research question, it can be stated that plurilingual learning with TBLT methodology increases crosslinguistic awareness in students, and that the degree and the extent to which it is increased is strongly dependent upon biographical aspects. All students in fact developed an increased crosslinguistic awareness, but the degree to which this awareness developed was strongly influenced by their linguistic background and language biography as well as prior knowledge.

With regard to the second and third research questions, the data clearly show that there is a focus on the activation of transfer strategies, which are adopted at different levels to overcome linguistically demanding situations. Most notably, lexical transfer and cross-lexical consultation between L2 and L1 can be mentioned here. While plurilingual students already have access to transfer strategies, thanks to the practice in the classroom, they become aware of how these strategies can be used for communication as well as for their own languagelearning process. This leads to the ability to activate the use of transfer for every language in their repertoire, where the choice is determined not so much on how long a language has been learnt, or the language family, but rather by the needs, and whether in a specific communicative situation one language can be more useful than another to provide solutions to a certain problem. Students, who come from a predominant monolingual social background, and have experienced mainly consecutive language learning on, can approach transfer from L2 and L3 thanks to imitation learning and forms of social learning, and this way also change both their psychotypological perception and attitude towards their L2 (Italian). The increased attention, and the simultaneous use of more languages enables all the learners to activate the forms of intercomprehensive learning by making use of their linguistic and translingual knowledge to deduce the meaning of unknown words, and expressions. All students claimed that multilingual learning accelerated their language-learning process, as they experienced new ways of language acquisition.

At the same time, students also tend to transfer meaning from one language to the other, thus changing the composition of their own system of cultural and linguistic reference. Thanks to contrastive reading exercises, they learn that the construction of meaning is culture-specific, and that multilingual discourse is characterized by ambiguity, fluidity and polysemy. Due to this, the meaningmaking process becomes transcultural and more complex, since it is based on more than one cultural reference system. This gives the learners the opportunity to play with multiple meanings and metaphors by recomposing and repositioning them in new and unforeseen ways.

Communicative strategies such as codeswitching and translanguaging on the one hand are implemented by students with a less plurilingual background (students 2 and 4) to overcome difficult linguistic situations, thus allowing them to regulate their learning process in a way that helps them identify problem areas, and search for new ways of learning. On the other hand, students with a more plurilingual background (students 1 and 3) learn to use codeswitching and translanguaging also strategically to express their multilingual personae and to regulate discourse. All students develop an increased awareness for the needs of the interlocutors in plurilingual settings and try to adapt their language production to them.

New learning paths are discovered along this way. Compensatory strategies such as codeswitching are used in difficult linguistic situations with a scaffolding function, thus supporting the learners in their attempt to approach their own ZPD (zone of proximal development Vygozky). The contrastive use of languages, promoted by the plurilingual inputs, induces the learner to reflect on their own personal language production, and to critically monitor and correct it where necessary. This way, their proficiency in different languages is perceived more clearly, and their specific needs can be identified. Thanks to the self-regulated forms of learning, the necessary steps are taken by the individual learner to comply with the identified shortcomings.

The activation of the multilingual repertoire through the simultaneous use of more languages accelerates with time in all students. This means that the languages are more easily retrievable, and that new ways of learning based on the interaction between languages can be found. Linguistic hierarchies are thereby laid bare and recognized as such, and the students develop a consciousness for their own emotional approach to the languages in question. They realize that each language is associated with particular emotions and that these emotions tend to depend, to a large extent, on each student's own language background, influencing the way each language is used.

### Gisela Mayr

### **Abbreviations**


### **References**


## **Chapter 9**

## **Students' perceptions of plurilingual nonnative teachers in higher education: An added or a mudded value?**

### Patchareerat Yanaprasart

University of Geneva

### Sílvia Melo-Pfeifer

University of Hamburg

In this study we compare students' perceptions of expatriate nonnative teachers in two higher education institutions, one in Geneva, Switzerland, and the other in Hamburg, Germany. Relying on a theoretical framework that crisscrosses aspects of internationalization of higher education and students' perceptions of nonnative discourse and its intelligibility, the current study compares how students in both universities perceive nonnative teachers' performances in the classroom and the impact that these perceived performances may have on their academic achievements. Results point out that, in both institutions, despite their different sociolinguistic profiles, the interviewees tend to positively value multilingualism and plurilingual repertoires. However, it emerges that Swiss students express willingness to position themselves absolutely positive, whereas German students are more neutral regarding the added value of the "plurilingual nonnativism".

### **1 Introduction**

Higher education scenarios have been dealing with an increase in issues such as the internationalization of staff, students and teachers. In this context, educational institutions are not only expected to attract international students, but also an increasing number of teachers who can teach and socialize in other languages

Patchareerat Yanaprasart & Sílvia Melo-Pfeifer. 2021. Students' perceptions of plurilingual nonnative teachers in higher education: An added or a mudded value? In Jorge Pinto & Nélia Alexandre (eds.), *Multilingualism and third language acquisition: Learning and teaching trends*, 185–206. Berlin: Language Science Press. DOI: 10.5281/ zenodo.4449784

### Patchareerat Yanaprasart & Sílvia Melo-Pfeifer

than their first language. This "other" language can either be the international language (Mueller 2018, for English), the local language of the institution (Melo-Pfeifer 2017, for German) or the language of the discipline (Yanaprasart 2019). The common point of these teachers is that they are all "nonnative teachers" (Dervin & Badrinathan 2011) of the language of instruction.

Furthermore, this linguistic situation underscores the language competences of this group of teachers, who are, de facto, bi-plurilingual (Mueller 2018). The question is to know whether such a plurilingual profile is recognized by students, in what way and under what conditions. Besides, is there any difference, from a student's perspective, between a native and a plurilingual nonnative teacher when they teach in a foreign language for some and a first language for others (Taillefer 2004)? If yes, in what way is such a difference described?

This article analyses and discusses the perceptions of "native and nonnative" teachers held by students. In the case of nonnatives, although plurilingual, they are still nonnative speakers of the teaching language. How is a "plurilingual teacher" perceived (Llurda 2005; Varghese et al. 2005) by students? In what way are their perceptions discursively reported (Miller 2010) and how do these perceptions relate to the profile of the institutions and of the disciplinary fields?

Unlike the work of Medgyes (1992; 1994), which focuses on teachers' views, our research is similar to that of Li-Chua Chen & Van Tien Nguyen (2011) on student views of teachers with varied linguistic origins and to that of Lasagabaster & Sierra (2002), whose focus is also based on students' perceptions of native and nonnative English speakers. More precisely, we will deal with the problematics of if, whether, why and how the "monolingual habitus" (Gogolin 2008) and the "plurilingual mind" (Menghini 2017) are visible in the students' discourse and how far one habitus or another influences the ways teachers' linguistic and pedagogic competences are perceived, recognized, legitimized, and (de)valorized (Kramsch 1997; Clark & Paran 2007).

After the theoretical framework, we will present the empirical study. The methodological design section will provide information about: (i) institutional context of data collection; (ii) data collection instruments; (iii) data collection procedures and target audience; and (iv) data analysis procedures. The results section will outline the definitions of being a native speaker and the students' perceptions of nonnative teaching practices. The final section will close with discussion, concluding remarks and perspectives.

### 9 Students' perceptions of plurilingual nonnative teachers in higher education

### **1.1 Native and nonnative: A dichotomy worth visiting to understand students' perceptions of expatriate teachers in higher education?**

As stated by Kang et al. (2015: 683), "campuses are becoming increasingly diverse", a specific development being "the increasing number of nonnative instructors" (idem: 684). With the effort to maintain high standards so as to be recognized as an international institution, each university faces challenges in managing language diversity that expatriate teachers and international students bring with them (Yanaprasart 2018). Inevitably, this phenomenon has prompted the question of whether the language competences that these expatriate nonnative teachers bring should or should not be valorized pedagogically and institutionally; if yes, how do we do that in a multilingual academic context (Blommaert & Verschueren 1998) where diversity and tensions are present?

Studies undertaken by Subtirelu (2015) and by Kang et al. (2015) concede that students' attitudes towards international teachers tend to be guided by a monolingual bias and that they therefore tend to evaluate nonnative teachers as less competent or less comprehensible. Kang et al. (2015) state further that native students tend to perceive nonnative lecturers as linguistically inadequate or lacking in linguistic accuracy, despite the fact that intelligibility in teaching has to be negotiated and co-constructed, as in any communicative situation, for the sake of a mutual understanding requiring both "interpretability" and "intelligibility skills" (Candlin 1982). According to Kang et al. (2015: 684), "undergraduates often perceive – whether rightly or wrongly – deficiencies in the intelligibility of international instructors".

According to Rajagopalan (2005: 284), in a study on nonnative teachers of English, "native speakers" are considered as "the true custodians of the language, the only ones authorized to serve". An ideal teacher is portrayed as to have "near native" qualities (Coppieters 1987), or to be a "pseudo native speaker" (Medgyes 1994), producing "a native-like pronunciation", possessing "high-level language" abilities (especially with regard to idiomatic language) and showing "confident language use" (Ofra 2005). This "nativespeakerism"<sup>1</sup> model (Gnutzmann 1999; Holliday 2006) still reflects the "ideal monolingual native speaker". This traditional dichotomy, "native" versus "nonnative speaker", claimed by Derivry (2006) as a theoretical linguistic abstraction, has to be questioned in the era of globalization, where people speak more and more languages. Therefore, individuals acquire multiple foreign languages and become multicompetent users of multiple languages (Cook 2002).

<sup>1</sup>This is a theory suggesting that a foreign language learner becomes and behaves in general as a native speaker in his/her mastery of the language.

### Patchareerat Yanaprasart & Sílvia Melo-Pfeifer

Notwithstanding the outdated conceptions attached to the dichotomous labels "native" and "nonnative", however emphasized by Derivry (2006) as important, it turns out that these concepts carry a potential explanatory adequacy and thus a heuristic validity. As pointed out again by the research of Kang and her colleagues, "failures in communication between native speakers and nonnative speakers are typically attributed to problems with nonnative speakers' proficiency" (Kang et al. 2015: 681). Following these lines, while lack of proficiency and accent may be perceived as scapegoats regarding the (negative) evaluation of teachers' performances (at both a linguistic and a scientific level), it should be acknowledged that arguments are usually related to the duality of native vs. nonnative skills. The second is usually caracterized in terms of "broken" linguistic skills (see Lindemann & Moran 2017 for further explanations about the ideologies attached to the adjective "broken"). Both Lindemann & Moran (2017), in the United States, and Melo-Pfeifer (2017), in Germany, conclude that nonnative discourse is perceived as being related to "having an accent", making mistakes and sometimes lacking in comprehensibility. As reported by Kang et al. (2015: 682), "in one particularly difficult and sensitive situation – the U.S. undergraduate classroom taught by an international teacher assistant – students' complaints are frequently more a function of their own stereotyped expectations than of ITAs' [international teacher assistants] objective language performance". Nevertheless, the authors acknowledge that "although some ITAs' lack of English proficiency can indeed hinder undergraduates' ability to comprehend subject material (…), students' linguistic stereotyping plays a powerful role adversely affecting their comprehension of ITAs over and above legitimate issues of ITA oral proficiency" (idem: 684).

While we can agree that studies in national contexts constructed as monolingual (such as Hamburg) may be irrelevant or even inadequate to analyze how students' perceptions work in multilingual ones (such as Geneva), it is strikingly important to note that local and expatriate teachers are evaluated differently (Subtirelu 2015). Whereas in some contexts, as reported by Lindemann & Moran (2017: 650), this evaluation may be related to general "negative attitudes toward nonnative speech in the US", in other contexts this may be related to a monolingual mindset in academic institutions, which tend to value multilingualism and plurilingual competences only when they are perceived as profitable or relevant in some scientific areas (Berthoud et al. 2013; Gajo 2013; Melo-Pfeifer 2017; Yanaprasart & Lüdi 2018).

Teaching requires more than language competences. Teachers also perform the functions of transmitter, vector and negotiator of knowledge, evaluator, speech stimulator and mediator (Gajo 2005). Teachers play the roles of facilitators and

### 9 Students' perceptions of plurilingual nonnative teachers in higher education

coaches in framing new perspectives (Roussi & Cherkaoui 2011). However, when it comes to being evaluated, at least in the U.S., "ratings are not simply a neutral measurement of the speakers' language [and teaching, we would add] ability but instead reflect listener factors as well" (Kang et al. 2015: 700; see also Lindemann & Subtirelu 2013).

### **2 Empirical study: Methodological design**

Our study is conducted jointly at the Universities of Geneva and of Hamburg, in the scope of an exploratory project entitled *Students' social representations of plurilingual nonnative teachers: Between a monolingual and multilingual perception of multilingualism* (Yanaprasart & Melo-Pfeifer 2017; 2019). Its aim is to analyze the perceptions of students towards their teachers' plurilingual competences when teaching in a foreign language. In what way and under what conditions does a monolingual or multilingual environment shape the perceptions of the concerned social actors?

### **2.1 Institutional context of data collection**

Because perceptions of students may be influenced by the context, either national or local, we will briefly present the institutions where data were collected.

Founded in 1559, the University of Geneva is a public research university located in Geneva, Switzerland. Today, this French-speaking university is the third largest university in Switzerland by number of students (16,935 in 2017). Thirtyseven percent of them are international whereas 20% come from other parts of Switzerland, all together representing 151 countries. Sixty percent of teachers and scientists are of foreign origin (2,854).<sup>2</sup>

The University of Hamburg was founded in 1919 and is the largest research and educational institution in northern Germany. In 2017 the university had 43,326 appointed students, with 5,433 (13%) being classified as "international".<sup>3</sup> In terms of teachers' profiles, 15% (from among 4,640 teachers and scientists) are called *Ausländer/innen* ('foreigners'). As reported by Mueller, resorting to an on-line questionnaire, "in total 279 languages, including dialects, varieties, creole languages, pidgin languages, sign languages, one deaf-blind manual alphabet, programming languages and artificial languages were reported to be understood,

<sup>2</sup> Source: https://www.unige.ch/stat/fr/statistiques/, https://www.unige.ch/stat/fr/statistiques/ chiffresetudiants/, https://www.unige.ch/stat/fr/statistiques/personnel/.

<sup>3</sup> Source: https://www.uni-hamburg.de/uhh/profil/fakten.html.

### Patchareerat Yanaprasart & Sílvia Melo-Pfeifer

and/or spoken and/or written by students and instructors" (Mueller 2018: 366) and 93 self-reported mother tongues were identified. Despite this internationalization level, most of the German students we interviewed for this study agreed that the linguistic environment is mostly monolingual. Even though the university is considered multilingual, as students and teachers have plurilingual repertoires, Mueller states that "at the university of Hamburg, we can observe a (…) situation, as the local standardized language is German, and English serves as a complementary or additional language" (2018: 361).

This brief presentation allows us to see that being integrated in a multilingual country (with three official languages), the University of Geneva also has a more significant percentage of both international students and teachers when compared to Hamburg. Another interesting difference is that Germany is perceived in the linguistic imaginary as being monolingual. This information, which may have an impact on the visual and acoustic linguistic landscapes of the institutions and, thus, on students' perceptions of "linguistic normality" in academia (either monolingual or multilingual), might help to explain some of the differences in the collected data. An underlying hypothesis of this comparative work is that the linguistic environment of the universities could influence how students perceive plurilingual nonnative teachers in both contexts.

### **2.2 Data collection instruments**

Methodologically, several types of data were collected, namely the analysis of official documents such as the institutional language policy and the bachelor and master programs of a Swiss-French university (University of Geneva, Yanaprasart 2020) and the analysis of questionnaires and interviews with students in foreign language education programs (University of Geneva and of Hamburg).

In terms of the questionnaires and interviews, these were collaboratively constructed by the researchers in the two contexts and calibrated in order to be understood in both of them. In terms of collected data, we gathered information on:


### 9 Students' perceptions of plurilingual nonnative teachers in higher education


The themes mentioned in the theoretical part were questioned in the most neutral way possible, the goal being to discover the spontaneous and current position of students on these issues. All these questions were treated according to content analysis (identification of thematic strands). Table 9.1 provides an overview of the data collected in both institutions.


Table 9.1: Overview of the data collected (Year 2016/2017)

### **2.3 Data collection procedures and target audience**

At the University of Geneva, in the first place, a semi-open questionnaire was submitted to 31 students in the field of didactics of French as a foreign language. Twelve are students enrolled in French as a foreign language diploma (DEFLE), compared to two in bachelor's and 17 in master's (MAFLE, MS Management, MA in Theology, MA in English, MA in History and French, MA in German Studies). Twenty-two point five (22.5) percent are male. The ages of the respondents vary from 20 to 42 years.

At the University of Hamburg, respondents were mainly prospective Spanish and French teachers. In both contexts, the study was designed as a small sample to encourage respondents to write as much as possible about the questions, allowing for in-depth discourse and content analysis. Another issue was the design of a comparative study that would allow the authors to compare the more or less equal numbers of students in both fields. Table 9.2 provides a glimpse of the profile of the respondents in the two higher education institutions.

### Patchareerat Yanaprasart & Sílvia Melo-Pfeifer


Table 9.2: The respondents' profiles

### **2.4 Data analysis procedures**

For the questionnaire's open questions and for the interviews, we followed a content analysis (Bardin 1993), and for the closed questions a quantitative analysis, without statistical aims, but just to identify tendencies. The content analysis of the questions makes it possible to characterize the judgments towards the aforementioned themes. It is a question of the participants' representations and not of any truth they would express except "their lived truth". The results present what students think of their teachers. The concept of representation is used by Durkheim (1960) to explain that between various social groups circulate representations of others. Boyer (1995) puts this notion in relation to the "ethnosociocultural" aspects of each group. We find the adjective "social" in Jodelet's (2003: 53) definition, for whom social representation is "a form of knowledge, socially

elaborated and shared, with a practical aim and contributing to the construction of a reality common to a social whole". This concept is therefore of particular importance in social life, notably in the field of assimilation of knowledge, where it plays a constitutive role. This notion is also central to our surveys, which aim to confront the main advantages and constraints of multilingualism in education.

In this paper, we will combine both quantitative and qualitative data analysis, from interviews and questionnaires, regarding the definitions of "native speaker" and the perception of classroom practices of both native and nonnative teachers.

### **3 Presentation of results**

### **3.1 Definitions of being a native speaker**

According to most respondents, the identity of a native speaker is perceived primarily by his/her accent (or the perceived lack of). Furthermore, a native speaker is someone who speaks his or her first language.

The distinction between a native speaker and a nonnative speaker will be for me at the level of the mother tongue. The mother tongue is the native language of that person, the first language that this person speaks. (UGE\_8)

More specifically, the notion of "native" is most of the time associated with "origin", "maternal", "first", even "natural". For some, a native refers particularly to someone who has an "intuitive" and spontaneous mastery of the language "since forever", who learned the language "in a natural way", not at school, whereas a nonnative is someone who learned the language later on.

A nonnative has learned the language, while for a native, it is more intuitive. A native is someone who has mastered the language since forever; it is his mother tongue, while a nonnative is someone who learned later. (UGE \_5)

The natural aspect stays with the native.

From my point of view, a native is one who grew up in one language and learns it very little. We do not learn it at school, but in a natural way. This natural aspect always remains with the native. (UGE\_7)

The following quotation from a German student also brings this representation through the repetition of the adverb "naturally" to the point:

### Patchareerat Yanaprasart & Sílvia Melo-Pfeifer

The native speaker of a language has (.) naturally no accent and has no – well – naturally, he also makes errors, everybody makes errors, I also make errors in German, but they do not make such a high frequency of errors as nonnative speakers. And also, regarding vocabulary, the lexicon (..) everything is for him much easier and yes (…) and on the contrary he also naturally - well (..), right. (UH\_1)

The responses highlight that a native speaker is someone who has not only grown up in that language, but who has had years of study in that language. A native is someone who has grown up in one language and has been learning it since childhood. Others argue that there is not much difference in terms of language competence, because there are nonnatives who master the language better than natives. It is also claimed that a native is a person who belongs to a given culture, who has grown up in that culture, or a person who has spent many years in that "natural" cultural immersion and feels at home.

As natives, some people who are part of a culture, who grew up in this culture or who spent many years in this immersion, for which this culture is already natural. I claim a concept a little wider. I look at myself as a nomad and I need to feel good wherever I am. (UGE\_4)

The sense of "feeling good" "in a natural way", as described above, provides a broader meaning than the belief of "nativism" linked to "national culture" that "is incarnated in certain groups of individuals born in a given country [...] and that whoever is born outside of this country of parents speaking another language is unable to achieve the status of native speaker" (Amin 2004: 78 quoted in Annous 2011: 213).

Over and above that, a native, according to respondents, will use more idiomatic expressions than a nonnative. That being said, what seems to distinguish a native from a nonnative is a certain delicacy of language, notably pictorial expressions and cultural references in daily language use. It is this particular way of speaking, or the idiomatic usage that our respondents prefer to emphasize, which is a far cry from the "idealistic" ideology of the native speaker who is generally associated with the "perfect mastery of the language" (Annous 2011). A German student states that native speakers can be very inspirational, because they symbolize the level of perfection they want to achieve – or should achieve – as teachers of a foreign language:

I: And what for an influence has your teachers' "native-speakerism" on your scientific or academic education?

### 9 Students' perceptions of plurilingual nonnative teachers in higher education

B: (..) It naturally has an influence, because /ahh/ well, we really try to express ourselves as a native speaker. It is ya somehow the norm or what we should attain and... (UH\_1)

All in all, we find some of these answers in the characteristics of a native speaker provided by Davies (2003: 210–211): (1) the native acquired his L1 during his childhood; (2) intuitively, he knows what is acceptable and correct in grammatical terms; (3) he knows how to differentiate the grammatical aspects of his L1 from those of another language; (4) he is able to produce spontaneous and fluid speech, his communicative competence (production and comprehension) is varied; (5) he knows how to be creative in writing (literature, metaphors ...); and (6) he has the unique ability to interpret and translate in his own language. However, this can be a comparative fallacy (Bley-Vromon 1983), since only the first characteristic of the native (the language has not been learned during childhood) cannot apply to nonnatives. Everything is possible for the rest, according to the motivation and the possibilities offered to the speaker to practice the language, in an educational context, of international mobility or simply outside a "native" context, as also perceived by the interviewed students.

### **3.2 Perceptions of practices in class**

In this subsection, we will outline the perception of classroom practices of the two groups of teachers. We resort to quantitative data analysis of the questionnaires, following a comparative perspective between both contexts. We also introduce qualitative data from our interviews in order to explain the quantitative results.

Figure 9.1 shows the results of quantitative data related to the evaluation of perceived classroom practices. The analysis of it makes clear that Swiss and Germans tend to position themselves differently regarding the way the sentences are formulated (see highlights in yellow). Thus, Swiss students tend to answer either positively or negatively at the extreme end of the scale (positions 1 or 6), while German students seem to be more cautious and tend to position themselves on neutral values of the scale (positions 3 and 4). These different positions may be explained by the linguistic diversity of the class population. On the Geneva side, 25.8% of the respondents have two first languages, 36.6% declare trilingual; 20% speak four languages, 10% are pentalingual, 13.3% sextalingual, 3.3% septalingual and the same percentage octalingual. On the German side, students declare themselves mostly as German (41 out of 60) and having German as a mother tongue (55 out of 60).


Figure 9.1: Observations on teachers' language and teaching practices (1: I fully agree; 6: I don't agree at all)

### 9 Students' perceptions of plurilingual nonnative teachers in higher education

The choice to have only native teachers was rejected by a large percentage of respondents: 49.9% (UH) and 82% (UGE), respectively. Among the justifications given was that it would interrupt interculturality; it would be a shame to miss out on very good teachers, especially since "a nonnative can have the same abilities to teach as a native" (UGE). Some students are of the opinion that "diversity is always a linguistic and cultural richness" (UGE). At this point, according to these students, nonnative teachers are more comfortable with cultural differences, more likely to help students to deal with them, and to share their own culture and to step out of their comfort zone (Pratt 1991; Yanaprasart 2017).

With a nonnative teacher, I think there are also cultural aspects. S/he also learned the cultural aspects of the teaching language. (UGE\_9)

Well, good, also there, that (s)he has perhaps other perspectives, isn't it? As someone who is representative of the culture him/herself. /Ähm/ perhaps is (s)he more objective, perhaps not, I don't really know. Perhaps there is not such a thing as objectivity. But well (..) another perspective and (s)he has also perhaps more experience regarding intercultural encounters. (UH\_3)

The interesting thing underlined by the interviewed students is about diversity: a diversity that comes to "us", which can open the mind to different points of view and perspectives. If it is very good to have cultural references from local teachers, having expatriate teachers should be encouraged so as to have contacts with different accents and intercultural experiences. Diversity is a synonym for richness, and mixing means force and dynamic. A mixed team should be created, and a balanced collaboration between local and international teachers should be encouraged and optimized: "I like the mixing of teachers. Diversity of teachers signifies richness. For me, it's something positive." (UGE\_4)

Furthermore, while Swiss students do not tend to perceive native teachers as more interesting, motivating or better organized, German students tend to perceive these characteristics as not being necessarily attached to the linguistic skills, but more skeptically: regarding the need to perfectly master the language, a German student states that correctness is "not necessary for transmitting content" and that "we learn from errors" (from anonymous questionnaire). In both cases, however, organization is not perceived as being particularly attached to the native teacher. Regarding the question of the organization of courses, the vast majority of respondents believe that good organization "has nothing to do with being native". On the other hand, it seems to them that "nonnative teachers make more efforts" to organize their courses. Additionally, just a small number

of respondents agree with the finding that native teachers are more demanding in class, in both contexts. But again, some divergences emerge: while, for some interviewees, the requirements are the same, for others, it is quite the opposite: it is nonnative teachers who are more demanding. As for a third group of remarks, the requirement does not depend on the origin of the teacher, it depends on character.

Also different is the positioning of German and Swiss students regarding their perception of degree of requirement and the ability to evaluate students. While Swiss students do not envisage such a relationship, German students, again, are quite undecided and avoid a clear positioning. When asked about the assessment skills, all respondents think that both categories of teachers have the same abilities. Indeed, "the ability to evaluate does not depend on the language of instruction", wrote one respondent. "It's not an ability related to language proficiency", argues another from the same institution. Nevertheless, many more Swiss than German students are willing to attest that their grades do not depend on teachers' linguistic skills. Another interesting feature concerns the different assessment of rapidity of learning with native speakers: while the majority of the Swiss students (61%) answer that they do not learn easier and more rapidly with native speakers, German students position themselves again in the middle values, which leaves some place for thoughts on the impact of comprehensiveness of teachers' input into their cognitive work in the classroom. So, having classes with nonnative speakers may be perceived as delaying the acquisition of content, at least from the perspective of the inquired German students.

The majority of the surveyed students share the opinion about the importance of the "perfect mastery" of the language of instruction. It is important to note that, in both universities, in order to deconstruct stereotypes about being bilingual or plurilingual, students stress the impact of explicit instruction acquired in the classroom about bilingualism and "partial" linguistic proficiency:

I have always thought that being bilingual means speaking perfectly two languages at the same level, but it is not, in fact not necessarily. Anyway, we always learn a language, even our own mother tongue; we will continue to learn until we die. (UGE\_5)

I would really like to speak as a native speaker, I would really love it a lot, I guess, that it is very / very difficult to attain, I don't know. /ahm/ but, yes, the time spent in my university had had an influence on me – and it is perhaps good so – the time at the university had shown me that is not necessary [to speak like a native speaker], and I had this impression previously (laughs). (UH\_3)

### 9 Students' perceptions of plurilingual nonnative teachers in higher education

To the question of whether it is more important to speak the language of teaching perfectly than to speak several languages (sentence 8), certain diversity in the answers can be observed. If no one answered "absolutely", 21.6% said "yes" and "rather yes". It is the "absolutely", "yes" and "rather yes" answers that win in the following two observations in both contexts: knowing several languages is more and more important for students and speaking several languages is more and more important for teachers (sentences 9 and 10). On the learners' side, it has been shown that "it is an undeniable wealth; an asset for life; an advantage in our present society; for certain mobility; for studies; to read articles, books written in a foreign language" (UGE). With regard to teachers, the fact that they must "work with foreign colleagues, or even with foreign universities, they need to speak several languages" (UGE). More precisely, plurilingual resources are "vital for acquiring and transmitting knowledge" (UGE). But, even if plurilingual skills are perceived as very positive from both sides, German students are less effusive regarding this optimistic evaluation. This cautious position may be correlated with the answers to question 8: more German than Swiss students are skeptical regarding the sentence "It is more important to perfectly speak the teaching language than to speak multiple languages", meaning that the plurilingual competence is less positively evaluated than mastering perfectly the language of instruction.

For the last question of whether the English language should replace the local language, it is the answer "rather not" that wins most of the German answers, with 35%, while on the Swiss side, the answers are "rather not" at 12.5%, "no" at 17% and "not at all" at 58%. Note that at the University of Geneva, the French language is the main language of instruction for all subjects within the bachelor degree. A passive or good knowledge – in a second language, in this case in English – is recommended or even necessary to study. That is to say, most disciplines recommend or require knowledge in several languages: 78.57% for university baccalaureates and 69% for master's degrees. This requirement reflects a desire to maintain the local language while allowing the integration of an international dimension in curricula.<sup>4</sup> We may say that this university language policy has a certain impact on the students' perceptions towards the role and place of English in the study program.

<sup>4</sup>At the University of Geneva, half of 71 disciplines for the master degree (52.11%) require knowledge of two languages: French-English (81.08%), English-French (18.91%); 14.08% (French); 11.26% (English); 19.71% (Trilingual): French-English-German (6 masters), German-French-English (5 masters), French-English-combined languages (3 masters). No such requirements exist in Hamburg, except for Bachelor and Master studies, related, for example, to foreign language teaching.

Patchareerat Yanaprasart & Sílvia Melo-Pfeifer

### **4 Discussion and concluding remarks**

In light of the foregoing, students do not report any significant superiority of "nativism" between the two groups of teachers. In addition, most respondents are of the opinion that there is no relationship between the way of teaching and the language skills of the teacher. Furthermore, no difference is perceived in evaluation practices. What students expect from their teachers in both categories is that they are able to make the learning process relevant and motivating, that they are sensitive and responsive to student needs, and finally are able to respect their learners as individuals with their own aspirations.

If there is a difference, it concerns more a cultural than a linguistic aspect. It is about a question of personality: "it is more related to the person and to his pedagogical sensitivity than to the mother tongue". Unlike Medgyes' conclusion (1992; 1994), according to which the difference between native and nonnative teachers from the point of view of the teachers interviewed comes mainly from language competence, at least in the field of foreign language teaching, our sample is of the opinion that no one is better than anyone because of his/her linguistic competence, but by reason of his/her professional, didactic and pedagogical skills. A difference would be mainly in the way of transmitting knowledge, as one student said: "It's not enough to be a native speaker but you have to know how to teach" (UGE). For Bento (2011: 104), both linguistic skills (the knowledge of "describing" language-culture) and teaching skills (the knowledge to "transmit" in a given context) are necessary. While the results of Medgyes' research (1992; 1994) in the field of foreign language teaching reveal that language competence is the main cause of the different ways of teaching native and nonnative teachers, none of our responses shows a correlation between way of teaching and linguistic competence, and this across different disciplinary fields. According to Medgyes, this difference in his study does not necessarily imply that a nonnative does not teach as well as a native, because the former finds effective teaching strategies to compensate for his/her possible linguistic weaknesses.

To this point, the students' perception of the ability of L1 teachers to use the language "naturally" may explain why they are portrayed as having shown more confidence in their skills, particularly in grammar, conversation and pronunciation, and were specifically appreciated for their knowledge in the culture of the language taught. Nonnative teachers have been described as experienced and understandable as having lived the same trajectory, which seems to be positive and encouraging for learners who aspire to follow in their footsteps, portrayed as an accessible and feasible model from a professional perspective. Castellotti (2011: 46) talks about an enhancement of approximation, linking, circumvention

### 9 Students' perceptions of plurilingual nonnative teachers in higher education

and transfer capabilities (our translation). By learning with teachers whose conditions are close to their own, these students hope to acquire the same skills. By means of their plurilingual resources acquired and by plurilingual pedagogical practices, teachers learn to build up translinguistic and intercultural strategies while promoting a dynamic vision of language competence, as well as intercultural awareness when teaching in multicultural and multilingual classes. So, learners and teachers develop reflexive and critical skills side by side.

The main comparative differences across the two contexts are located at the perception of the "plurilingual nonnativism". The plurilingual teacher is observed with more enthusiasm and acceptance on the Swiss side than on the German side, which views with more caution the advantages of having nonnative teachers and the understanding of the plurilingual repertoires as partial linguistic skills. As a matter of fact, in regard to the Hamburg side, the findings suggest two main points:


Terms such as *native* and *nonnative*, as well as *norm* and *correctness*, consistently emerge as heuristic categories in the German interviews. Concerning the Geneva students, their answers suggest a firm and favorable position to plurilingualism. Teachers' competences are not evaluated in terms of being a native or nonnative teacher, but instead of as being able to speak one, two or more languages (Bento 2011). Partial skills in multiple languages are positively portrayed. As such, the diversity of teachers' linguistic repertoires represents cognitive, pedagogical, communicative and didactic resources and strategies.

Taking the above into account, in the eyes of the surveyed students, teachers' plurilingual skills potentially represent, in terms of knowledge acquisition and transmission, an "added value", an undeniable strength and an inescapable advantage to impart knowledge in a multilingual and multicultural classroom, if only they feel prepared to do it. Otherwise, a "partial" linguistic proficiency is just a "mudded" value since most of the time students and teachers abide by the monolingual and monoglossic habitus of the educative institutions (mainly in the German context). So, if as mentioned by Cook (1991; 2008), the knowledge of two or more languages in one mind contributes considerably to the quality of teaching/learning in terms of motivation and cognitive development, these

### Patchareerat Yanaprasart & Sílvia Melo-Pfeifer

advantages should be more thematized, developed and discussed, particularly in language learning classrooms and in teaching education practices and supervision. The challenge for institutions is to change the environment by changing the perception of "language-as-problem" or "language-as-right" to "language-asresource" (Ruiz 1984).

In light of these reflections, it is especially in a transformative (Savin-Baden 2008) "troublesome space" (Montgomery 2011) that the "monolingual habitus" (Gogolin 2008) leaves its place to the "plurilingual mind" (Menghini 2017), where the use of a language can occur not as a simple, fixed and rigid code, but as a complex, yet flexible, dynamic and evolving tool so as to better explore knowledge in the service of learning quality (Berthoud 2016). A diversification of knowledge will be optimal in an environment where different cultures and languages have the opportunity to interact in a hybrid and multilateral perspective (Yanaprasart 2018), thus constituting a rewarding model of inspiration to open up to global knowledge.

### **References**


Bardin, Laurence. 1993. *L'analyse de contenu*. Paris: PUF.


9 Students' perceptions of plurilingual nonnative teachers in higher education


Patchareerat Yanaprasart & Sílvia Melo-Pfeifer


### 9 Students' perceptions of plurilingual nonnative teachers in higher education

*ucation and the Chinese learner*, 129–148. Hong Kong: Hong Kong University Press.


Patchareerat Yanaprasart & Sílvia Melo-Pfeifer


Ackerman, Philip L., 3, 4 Allegro, Fernando, 55 Allgäuer-Hackl, Elisabeth, 169 Alves, Ubiritã, 27 Amin, Nutzhat, 194 Amorim, Clara, 29 Amorrortu, Estibaliz, 119 Anderson, John R., 3, 5 Andrade, Amália, 28 Andres, Dörte, 3, 9, 10 Annous, Samer, 194 Aoyama, Katsura, 26, 27 Apelgren, Brit M., 142, 145 Arocena Egaña, Elizabet, 96, 99, 113 Aronin, Larissa, 73, 86, 118, 127, 132, 141 Avery, Peter, 35 Baddeley, Alan, 3–5 Badrinathan, Vasumathi, 186 Bagnoli, Anna, 75, 76 Bailey, Ann, 55 Bakhtin, Mikhail, 104 Ball, Arnetha, 104 Ballester, Ana, 8 Bardin, Laurence, 192 Barnes, Julia, 120, 121 Barreña, Andoni, 124 Batalha, Graciete Nogueira, 26, 30, 35, 36 Bateman, Blair, 53, 55 Bates, Douglas, 33

Bates, Elizabeth, 123 Beacco, Jean-Claude, 70–72 Beaudrie, Sara M., 57 Behr, Martina, 3 Benítez, Padilla, 3, 8–10, 18 Bento, Margaret, 200, 201 Berman, Ruth, 124 Berthoud, Anne-Claude, 188, 202 Best, Catherine, 27, 38 Bialystok, Ellen, 167 Bissoonauth, Anu, 144 Blackledge, Adrian, 99 Bley-Vromon, Robert, 195 Blommaert, Jan, 142, 187 Boersma, Paul, 32 Bohnsack, Ralf, 171 Bono, Mariana, 73, 85, 86 Borg, Michaela, 100 Borg, Simon, 99 Bowles, Melissa, 57 Boyer, Henri, 192 Bozzo, Luisa, 143 Brown, Cynthia, 26 Burwitz-Melzer, Eva, 167 Cabrelli Amaro, Jennifer, 56 Cameron, Lynn, 141, 142 Candelier, Michel, 98 Candlin, Christopher, 187 Capra, Fritjof, 141, 142 Carlisle, Robert, 37 Carreira, Maria M, 57

Carvalho, Ana M., 50, 55, 56, 58, 61, 63 Castellotti, Véronique, 200 Catarina, Oliveira, 28 Çelik, Servet, 74 Cenoz, Jasone, v, 50, 56, 61, 62, 69–75, 85, 86, 97, 98, 119, 120, 142– 145, 150, 166, 167 Chabasse, Catherine, 8 Chambers, Gary N, 144 Chen, Szu-wei, 35 Cherkaoui, Kenza, 189 Chih, Martín Te-chi, 37 Child, Michael, 50, 55, 56, 58–61 Clark, Elizabeth, 186 Cohen, Evan-Gary, 37 Colantoni, Laura, 26–28, 36–38 Cook, Vivian, v, 71, 118 Cook, Vivian J., 70, 98, 99, 108, 187, 201 Coppieters, René, 187 Core, Cynthia, 124 Correa, Maite, 57 Costa, Teresa, 29 Council of Europe, 70–72, 74, 96 Courtney, Louise, 143 Cowan, Nelson, 3, 4 Creese, Angela, 99 Cristófaro-Silva, Thaïs, 38 Csizér, Kata, 143, 144 Culhane, Stephen F., 143 Cummins, Jim, 96, 99 D'Andrade, Ernesto, 28, 29, 38 Dale, Philip, 123 Davies, Alan, 195 Dawrant, Andrew, 3, 7–10, 18

De Angelis, Gessica, 73, 99, 113, 143, 144, 147, 166

De Houwer, Annick, 118, 121 De-Cuir Gunby, Jessica T., 76 Dégi, Zsuzsanna, 143 Derivry, Martine, 187, 188 Dervin, Fred, 186 Dewaele, Jean-Marc, 121, 166, 169 Dingerfelder Stone, Maren, 8 Dmitrenko, Violetta, 73, 85 Dollerup, Cay, 3 Donato, Clorinda, 53 Dörnyei, Zoltan, 103, 143 Duanmu, San, 29 Duff, Patricia, 96 Dunn, Lloyd, 123 Durkheim, Emile, 192 Egger, Kurt, 147 Ehrlich, Susan, 35 Ellis, Elizabeth M., 70, 73, 85 Ellis, Rod, v, 112, 169 Enever, Janet, 144 Enochson, Kelly, 35 Ericsson, Karl, 3 Ericsson, Karl Anders, 4, 6 Ericsson, Mårten, 123 Escobar, Anna, 51 Escudero, Paola, 32 Espadinha, Maria. Antónia, 26 Espinosa, Aina, 35 Ezeizabarrena, Maria Jose, 124 Face, Timothy L, 26 Falomir, Laura Portolés, 143, 145 Fan, Dinghong, 4, 6, 17, 18 Feiden, Juliana, 55 Fenson, Larry, 123 Fikkert, Paula, 26 Flanzer, Vivian, 56, 62 Flege, James, 27, 35

Fraenkle, Jack R., 75, 76 Freeman, David, v Freeman, Yvonne, v Freitas, Maria João, 26, 35, 36 Frekko, Susan, 145 Fujimura, Osamu, 35 Gagarina, Natalia, 125 Gajo, Laurent, 188 Galante, Angelica, 98, 113 Garcia, Iñaki, 124 García, Ofelia, 70, 99 Gardner, Robert C., 144, 148 Garret, Peter, 143 Garrett, Joanne M., 76 Gatt, Daniela, 124 Genesee, Fred, 69, 121 Gerver, David, 3–5 Giancaspro, David, 56 Gibson, Martha, 73, 166, 167 Gick, Bryan, 28 Gillies, Andrew, 3, 8, 9, 18, 19 Gnutzmann, Claus, 187 Göbel, Kerstin, 73, 85 Gogolin, Ingrid, 186, 202 Goldman-Eisler, Frieda, 9 Gorter, Durk, 70, 71, 98 Graham, Lamar, 35 Gregersen, Tammy, 143 Griva, Eleni, 73 Grosjean, Francis, 172 Grosjean, François, 70, 71, 85 Gualda, Ricardo, 59 Guan, Qianwen, 35 Hallé, Pierre, 38 Hallet, Wolfgang, 168, 169

Harris, John, 145 Haukås, Åsta, 96, 99 Hayes-Harb, Rachel, 32 He, Yunjuan, 35 Hélot, Christine, 159 Helsen, Werner F., 6 Henriques, Eunice R., 53 Henry, Alastair, 142, 145 Herdina, Philip, v, 70–72, 97, 98, 109, 140, 141, 143, 144, 166, 167 Hickey, Tina M., 124 Higgins, Christina, 73, 85 Hitch, Graham, 3, 4 Holliday, Adiran, 187 Honikman, Beatrice, 27, 35, 36 Horst, Marlise, 98 House, Juliane, 166 Hufeisen, Britta, 70–73, 97, 98, 166, 167, 173, 177 Hulk, Aafke, 118 Iverson, Paul, 27 James, Nalita, 87 Jeffries, Carolyn, 102 Jensen, John, 53 Jessner, Ulrike, v, 55, 69–73, 85, 97– 99, 109, 140, 141, 143, 144, 165–167, 169 Jesus, Luís, 29 Jiminez Hurtado, Catalina, 8 Jodelet, Denise, 192 Johnson, Keith, 37, 50 Kalina, Sylvia, 8–10, 18, 19 Kang, Okim, 187–189 Kelm, Orlando, 55 Kemp, Charlotte, 73 Kern, Sophie, 123 Koch, Gary G., 76 Koike, Dale, 56, 59, 62

Kordt, Birgit, 143 Korkmaz, Şule Çelik, 74, 85 Kormos, Judith, 143, 144 Kramsch, Claire, 141, 142, 169, 186 Krumm, Hans-Jürgen, 168 Kurz, Ingrid, 18 Kwon, Harim, 35 Ladefoged, Peter, 37 Lambert, Sylvie, 8, 9 Landis, J. Richard, 76 Lanza, Elizabeth, 118 Lapresta Rey, Cecilio, 146 Larsen-Freeman, Diane, 141, 142 Lasagabaster, David, 120, 145, 146, 166, 186 Leonet, Oihana, 98 Li, Wei, 70, 99 Li-Chua Chen, Megali, 186 Lightbown, Patsy, 121 Lin, Yen-Hwei, 29, 30, 37, 38 Lindemann, Stephanie, 188, 189 Llama, Raquel, 36 Llurda, Enric, 186 Loddegaard, Anne, 3 Lopez Ornat, Susana, 124 Lortie, Dan, 100 Lozgka, Evangelia, 102 Lüdi, Georges, 188 Luisi, Pier L., 141, 142 Luo, Lin, 166

MacIntyre, Peter, 143 MacWhinney, Brain, 32 Maeder, Dale, 102 Major, Roy, 26 Marques, Isabelle, 28 Martinez, Hélène, 169 Martínez, Ramón Antonio, 96 Martins, Marlene Moreira, 26, 30, 35, 36 Martins, Pedro, 29, 30 Marx, Nicole, 71, 98 Masuda, Kyoko, 32 Masuhara, Hitomi, 112 Mateus, Maria Helena, 28, 29, 38 Mayer, Kerstin, 144 Mayer, Mercer, 124 Mayr, Gisela, 169 Medgyes, Péter, 186, 187 Meisel, Jürgen, 117, 118 Meißner, Franz-Josef, 169, 177 Melo-Pfeifer, Sílvia, 186, 188, 189 Menghini, Michela, 186, 202 MEQ, 100 Miguel, Afonso João, 38 Miles, Matthew B., 76 Miller, Jennifer, 186 Milleret, Margot, 53, 56 Mirici, Ismail Hakki, 73 Mitchel, Rosamond, v Montgomery, Catherine, 202 Montrul, Silvina, 57 Mora Pablo, Irasema, 100, 102, 106 Moran, Katherine, 188 Morkötter, Steffi, 167, 169, 177 Moser-Mercer, Barbara, 3, 4 Mueller, Jessica, 186, 190 Müller, Natascha, 118 Müller-Lancé, Johannes, 166 Munoz, Carmen, 143, 144 Myles, Florence, v Nagy, Kriztina, 143 Nikolov, Marianne, 143, 144, 159

Nunan, David, v

Ó Laoire, Muiris, 73, 86

O'Laoire, Muriel, 141 Oberauer, Klaus, 3 Odlin, Terence, 61 Ofra, Inbar-Lourie, 187 Oliva, Cedric, 53 Oliveira, Daniela Santos, 30 Oliveira, Desirée, 53 Oliveira, Diana, 26, 30, 33 Ortega, Ane, 119 Otwinowska, Agnieszka, 70, 72, 73, 85, 86, 144 O'Toole, Ciara, 124 Padilla Benítez, Presentación, 3, 4 Paladino, Maria P., 142 Palenham, Kenneth, 121, 122, 130 Paran, Amos, 186 Pasquarelli-Gascon, Violet, 53 Patton, Michael Quinn, 74 Pavlenko, Aneta, 173 Pereira, Rodrigo, 29, 30 Pérez-Pereira, Miguel, 124 Peyer, Elisabeth, 97 Philips, Denis, 141 Phipps, Simon, 99 Pilar Safont, Jordà, 166 Pippa, Salvador, 5, 9 Ponte, Eva, 73, 85 Potowski, Kim, 51, 57 Pratt, Mary Louise, 197 Rajagopalan, Kanavilil, 187 Recasens, Daniel, 35 Reich, Hans H., 168 Rennicke, Iris, 29, 30 Riccardi, Alessandra, 9 Ringbom, Håkan, 97, 166 Rodrigues, Susana, 28–30

Rose, Yvan, 32

Ruiz, Richard, 202 Russo, Mariachiara, 5, 9 Sá Nogueira, Rodrigo, 29 Saldaña, Johnny, 76 Sánchez, Laura, 73 Sandelowski, Margaret, 76 Savin-Baden, Maggi, 202 Scaramucci, Matilde, 55 Schewda Nicholson, Nancy, 9 Schmid, Monika S., 143 Schröder-Sura, Anna, 169 Seeber, Kilian, 3, 4 Selkirk, Elisabeth, 36 Setton, Robin, 3, 7–10, 18 Shadle, Christine, 29 Shameem, Nikhat, 159 Shintani, Natsuko, 112 Sierra, Juan Manuel, 146, 186 Silva, Ana, 29, 37 Silva, Antonio J. B., 50, 56, 58 Silva, Gláucia V., 55, 56 Silva, Roberval, 26 Simões, Antonio, 55 Simon, Marielle, 102, 103 Slobin, Dan, 124 Smith, Linda B., 142 Sommer-Farias, Bruna, 56, 57 Spada, Nina, 121 Sproat, Richard, 35 Steele, Jeffrey, 26, 28, 37, 38 Steffensen, Sune V., 142 Strange, Winifred, 27 Stratilaki, Sofia, 73, 85, 86 Stravakou, Pelagia, 102 Strömquist, Sven, 124 Subtirelu, Nicholas, 187–189

Rothman, Janson, 55, 56 Roussi, Maria, 189

Święciński, Radek, 28 Taillefer, Garcia, 186 Takeuchi, Nadoca, 55 Teixeira, António, 29 Thelen, Esther, 142 Tierney, Robin, 102, 103 Timarová, Šárka, 3, 4 Tódor, Erika-Maria, 143 Tokowicz, Natasha, 55 Tomlinson, Brian, 112 Tonelli, Livia, 9 Torres, Julio, 63 Tremblay, Marie Claude, 36 Trude, Alison, 55 Tyler, Michael, 27 Vaes, Jeroen, 142 Valencia, José F., 166 van Dam, Ine M., 3, 10 van Geert, Paul, 142 van Heuven, Vincent, 3, 7, 8, 19 Van Tien Nguyen, Etienne, 186 Varghese, Manka M., 186 Verhoeven, Ludo, 124 Verschueren, Jef, 187 Vidgren, Noora, 166 Vieluf, Svenja, 73, 85 Viera, Anthony J., 76 Von Bastian, Claudia C, 3 Wallen, Norman E., 75, 76 Waltmunson, Jeremy, 26, 28, 37 Wang, Binhua, 4, 6, 17 Wanrooij, Karin, 32 Warshauer Freedman, Sarah, 104 Weenink, David, 32 Wesely, Pamela M., 143, 144

Wilson, Ian, 28 Woll, Nina, 99 Wong, Nicole, 36 Woolard, Kathryn, 145 Wrembel, Magdalena, 56 Xing, Kaiyue, 30 Yanaprasart, Patchareerat, 186–190, 197, 202 Yenkimaleki, Mahmood, 3, 7, 8, 19 Zhu, Xiaonong, 29 Zimmer, Márcia, 27 Zyzik, Eve, 56, 63

Wiedemann, Lyris, 53, 55

## Multilingualism and third language acquisition

The purpose of this book is to present recent studies in the field of multilingualism and L3, bringing together contributions from an international group of specialists from Austria, Canada, Germany, Portugal, Spain, Switzerland, Turkey, and United States. The main focuses of the articles are three: language acquisition, language learning and teaching.

A collection of theoretical and empirical articles from scholars of multilingualism and language acquisition makes the book a significant resource as the papers present a wide perspective from main theories to current issues, reflecting new trends in the field.

The authors focus on the heterogeneity and complexity that characterize third language acquisition, multilingual learning and teaching. As the issues addressed in this book intersect, it represents an asset and therefore the texts will be of great relevance for the scientific community.

Part I presents different topics of L3 acquisition, such as syntax, phonology, working memory and selective attention, and lexicon. Part II comprises texts that show how the research on language acquisition informs pedagogical issues. For instance, the role of the knowledge of previous languages in the teaching of L3, the attitudes of multilingual teachers to plurilingual approaches, and the benefits of crosslinguistic pedagogy versus classroom monolingual bias. In sequence, Part III consists of texts on individual learning strategies, such as motivation and attitudes, crosslinguistic awareness, and students' perceptions about teachers' "plurilingual nonnativism".

All these chapters include several different languages in contact in an acquisition/learning context: Basque, English, French, German, Italian, Ladin, Portuguese, Russian, Spanish, and Turkish.