Crosslinguistic Influence and Second Language Learning [1 ed.] 0367357852, 9780367357856

Citation preview

i

CROSSLINGUISTIC INFLUENCE AND SECOND LANGUAGE LEARNING

Crosslinguistic Influence and Second Language Learning provides a comprehensive overview of what is currently known about prior language knowledge and experience in second language learning. Three bodies of research are critically reviewed to achieve this goal: (i) theories of language learning that attribute critical roles to prior experience in explaining second language development, (ii) empirical studies of second language learning that have investigated roles for crosslinguistic influence, and (iii) instructional studies that have supported second language learning by addressing the negative effects of crosslinguistic influence. Using this foundation, new research directions and theorization in the field of second language acquisition are proposed. This book will serve as an excellent resource for students and scholars with interests in (instructed) second language learning, applied linguistics, cognitive psychology, psycholinguistics, and language education. Kevin McManus is the Gilbert R. Watz Early Career Professor in Language and Linguistics and an Associate Professor of Applied Linguistics at The Pennsylvania State University, USA.

ii

COGNITIVE SCIENCE AND SECOND LANGUAGE ACQUISITION SERIES Peter Robinson, Series Editor

The Cognitive Science and Second Language Acquisition series is designed to provide systematic and accessible coverage of the links between basic concepts and findings in cognitive science and second language acquisition (SLA). Titles in the series summarize issues and research in areas of cognitive science which have relevance to SLA, and when read in combination, provide a comprehensive overview of the conceptual and methodological intersects between these two fields. The series is a valuable reference for scholars who want to increase their knowledge of theoretical and operational definitions in cognitive science, and their applications to SLA. Its titles are ideal for graduate students and researchers in SLA, applied linguistics, cognitive psychology, educational psychology, and language education, and can also serve as textbooks for advanced courses in these fields. Second Language Sentence Processing Alan Juffs and Guillermo A. Rodríguez Metalinguistic Awareness and Second Language Acquisition Karen Roehr-​Brackin Connectionism and Second Language Acquisition Yasuhiro Shirai Crosslinguistic Influence and Second Language Learning Kevin McManus For more information about this series, please visit: https://​www.routledge.com/​Cognitive-​Science-​and-​Second-​Language-​ Acquisition-​Series/​book-​series/​LEACSSLAS

iii

CROSSLINGUISTIC INFLUENCE AND SECOND LANGUAGE LEARNING

Kevin McManus

iv

First published 2022 by Routledge 605 Third Avenue, New York, NY 10158 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2022 Taylor & Francis The right of Kevin McManus to be identified as author of this work has been asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-​in-​Publication Data A catalog record for this title has been requested ISBN: 978-0-367-35785-6 (hbk) ISBN: 978-0-367-35782-5 (pbk) ISBN: 978-0-429-34166-3 (ebk) DOI: 10.4324/​9780429341663 Typeset in Bembo by Newgen Publishing UK

v

For Florence Myles and Richard Waltereit

vi

vi

CONTENTS

List of Illustrations  Acknowledgments 

ix x

1 Introducing Crosslinguistic Influence 

1

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

Introduction  1 Language  2 Learning a Second Language  4 Prior Knowledge and Experience  7 Transfer  10 Cross-​language Relationships  11 Directions of Crosslinguistic Influence  15 Explicit Instruction and Crosslinguistic Influence  17 Overview of the Book  18

2 Theoretical Models of L2 Learning  2.1 2.2 2.3 2.4 2.5 2.6 2.7

Introduction  20 The Unified Competition Model  21 The Associative-​Cognitive CREED  28 Processing Determinism  35 Inhibitory Control Model  40 Synthesis of Theoretical Concepts  44 Conclusion  49

20

vi

viii Contents

3 Studies of L2 Development  3.1 3.2 3.3 3.4 3.5

4 Instructed L2 Learning  4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

50

Introduction  50 L2 Learning and Morphosyntax  50 L2 Learning and Vocabulary  62 L2 Learning and Phonology  73 Conclusion  82

84

Introduction  84 Instructional Effects in L2 Learning  84 Explicit Instruction  87 Evidence-​Based L2 Instruction: L1 Use in L2 Learning  89 L2 Instruction and Morphosyntax  90 L2 Instruction and Vocabulary  102 L2 Instruction and Phonology  109 Synthesizing Evidence about Instruction and L2 Learning  117 Conclusion  121

5 Reflections and Future Directions 

122

5.1 Introduction  122 5.2 Synthesizing Crosslinguistic Influence, L2 Learning, and Instruction  125 5.3 Theorizing Crosslinguistic Influence  131 5.4 Future Research Directions in Crosslinguistic Influence  133 5.5 Conclusion  135

Recommended Reading  References  Index 

vi

136 138 159

ix

vi

ILLUSTRATIONS

Figures 1 .1 Visualization of a Web  2.1 Simplified Visualization of Stimuli in a Stroop Task  4.1 Explicit Information Used in Benati (2005) for Teaching the Simple Past  4.2 Video Stills Depicting Concept of Ongoingness from McManus and Marsden (2017) 

16 43 88 98

Tables 1 .1 Examples of Constructions in English  2.1 Summary of Subject-​Object Information in English and German Sentences  3.1 Summary of Viewpoint Aspect Differences for Past Perfectivity and Habituality in English, French, and German  3.2 Example Stimuli in Match and Mismatch Conditions from Roberts and Liszka (2013)  4.1 Form-​meaning Mappings for Viewpoint Aspect in French with English Translations  4.2 Description of the Treatment Components Used in Session 1: Ongoingness (Present vs. Past) 

3 23 55 58 96 99

x

ACKNOWLEDGMENTS

This monograph is the culmination of many collaborations, research projects, and discussions over the years. My interest in crosslinguistic influence and second language learning really got going in 2006, shortly after I interacted with Florence Myles for the first time. We met at the Association for French Language Studies annual conference in Bristol, United Kingdom. I attended the conference as an undergraduate student to give my first research presentation. At that time, I did not know how much Florence Myles and Richard Waltereit would support and encourage me as I worked toward completing my PhD on crosslinguistic influence and second language learning. I have not looked back. I am so very grateful to the both of them because now, as a PhD advisor myself, I remind myself of the lessons I learned at that time. They taught me so much and they led by example. After graduating, I had the great fortune of working with Rosamond Mitchell at the University of Southampton and then Emma Marsden at the University of York. It is rare to have the opportunity to work with such inspirational scholars. Florence, Ros, and Emma have been the best mentors I could have ever wished to work with. I want to thank all of you for taking me under your wing and helping me along. I am very grateful for your mentorship. I am also immensely grateful to colleagues and students for sharing ideas, critiques, and inspiration over the years. Some of the people I want to mention in particular are Nicole Tracy-​Ventura, Jim Lantolf, Celeste Kinginger, Bob Schrauf, Norbert Vanek, and Leah Roberts. Nothing quite replaces just talking through projects and ideas about nothing specific and then seeing them take shape in unexpected ways. The informal ones are always the most meaningful and so I am very grateful to these people and others who have taken the time to share a coffee (or a beer!). My students have also been a great inspiration, especially Alex Magnuson, Kelly Bayas, Yulia Khoruzhaya, and the students in our SLA reading

x

xi

newgenprepdf

x

Acknowledgments  xi

group at Penn State. I feel that sometimes students think that faculty have all the answers and that we know what we are doing. We are all learning as we go along and I have learned so much from my students. There are a number of people who have made this book project possible. I am very grateful to the help and support of Routledge, especially Ze’ev Sundry and Helena Parkinson, and to Peter Robinson, all of whom worked with me along the way. I am grateful to the three reviewers who provided great suggestions on the book project. I am also very grateful to Kelly Bayas and Yulia Khoruzhaya, who helped with proofreading, formatting, and chasing down references. I would also like to acknowledge the support of the following funding agencies who supported some of the work presented in this book: the British Academy (Grant number: #PF130001) and the United States Department of Education (Grant numbers: #P017A170074, #P017A200028, #P229A180009, # P017A170072). Lastly, I want to thank Amanda Huensch for all that she does every day. She helped me so much with this project in so many different ways.

xi

1

1 INTRODUCING CROSSLINGUISTIC INFLUENCE

1.1  Introduction This book presents a comprehensive overview of crosslinguistic influence in adult second language learning. Succinctly, research about crosslinguistic influence seeks to describe and theorize the ways in which a speaker’s cumulative experience with one or more languages can influence their processing and use of other languages. This can include, for instance, the extent to which a speaker’s experience with Polish and/​or Spanish might influence their use of Mandarin Chinese. In the field of second language acquisition (SLA), researchers have studied crosslinguistic influence in variety of ways, including how first language (L1) experience shapes second language (L2) learning, how L2 learning changes L1 use, as well as how the combined experiences of L1 and L2 learning influence the learning of an additional language (for reviews, see Jarvis, 2016; Jarvis & Pavlenko, 2008; Odlin & Yu, 2016). As we will see, researching crosslinguistic influence in these different ways is necessary for informing how we think about L2 learning, including proposals about the structure and organization of a speaker’s language system. It is also one reason why our Polish–​Spanish–​Chinese example excluded the labels L1 and L2: Crosslinguistic influence is not restricted to L1 effects on L2 learning; it involves understanding all types of experience on language use. The aim of this book is to explore some of the key questions that crosslinguistic influence research seeks to address, including: • •

How do adults build knowledge of an additional language? Does learning a new language lead to broader changes in a speaker’s existing language system or do the systems of L1 and L2 knowledge appear to be unconnected? DOI: 10.4324/9780429341663-1

2

2  Introducing Crosslinguistic Influence

• •

Might connections among different languages emerge and change over time? Can instruction help L2 learners overcome some of the negative effects of crosslinguistic influence?

In terms of situating crosslinguistic influence in the broader context of what we know about L2 learning, theoretical and empirical studies have, for a long time now, attributed critical roles to prior language knowledge and experience in making sense of how adults learn and use an additional language (for reviews, see Gass et al., 2020; VanPatten et al., 2020). This evidence base includes studies of how L2 speakers process and attend to language as well as analyses of what the language input looks like. Together, the findings from these lines of research have led to clear and testable theories about the routes and rates of L2 learning (see Chapter 2) and approaches to L2 instruction that are grounded in how speakers use language (see Chapter 4). In this book, our aim is to advance new research directions and theorization in the field of SLA by critically reviewing what is known about prior language knowledge and experience in L2 learning.1 We work toward this goal through a comprehensive review of three connected bodies of research: (i) theories of language learning that attribute critical roles to prior experience in explaining L2 development, (ii) empirical studies of L2 learning that have investigated roles for crosslinguistic influence, and (iii) instructional studies that have been designed to support L2 learning by addressing the negative effects of crosslinguistic influence. We begin our review by briefly describing what we mean by “language” (Section 1.2), which is followed by a discussion of “learning a second language” (Section 1.3). Here, we note some important debates from the field of SLA about what it means to learn a new language. We then move on to discussing connected issues in crosslinguistic influence research that we will further unpack in subsequent chapters, including “prior knowledge and experience” (Section 1.4), “transfer” (Section 1.5), “cross-​language relationships” (Section 1.6), “directions of crosslinguistic influence” (Section 1.7) and “explicit instruction and crosslinguistic influence” (Section 1.8). We end this chapter with an overview of the book.

1.2  Language It is well known that theories about the nature, forms, and purposes of language are plentiful (e.g., Bybee, 2010; Chomsky, 1965; Tomasello, 2003). Some of these theories contain points of agreement (e.g., learning a particular language requires exposure to that language), but there is also considerable disagreement, including the extent to which some properties of language might be universal and/​or innate and what role general cognition might play in the development of language knowledge (for reviews, see Ambridge & Lieven, 2011; Kempe & Brooks, 2016). In this section, our aim is to outline some of the key ideas about language as discussed in this book.

2

3

2

Introducing Crosslinguistic Influence  3 TABLE 1.1  Examples of Constructions in English

Construction

Examples

Morpheme Word Complex word Idiom (filled) Idiom (partially filled) Ditransitive (double object) Passive

pre-​, -​s avocado, fish daredevil, shoo-​in go great guns jog memory he wrote her a letter the cat was chased by the wasp

This book is grounded in functionalist understandings of language, in which “the surface conventions of natural languages are created, governed, constrained, acquired, and used in the service of communicative functions” (Bates & MacWhinney, 1981, p. 192, see also Christiansen & Chater, 2016;Tomasello, 2003). Under this view, communicating ideas, functions, meanings, and intentions is what drives speakers to learn and use language. Cognitive linguistic theories propose that the forms of a language that speakers learn and use can be described and conceptualized as constructions or form-​ meaning mappings (Bybee, 2010; Goldberg, 2006, 2019; Tomasello, 2003). A construction can denote a specific meaning (when one form expresses a single meaning; e.g., avocado) or sets of meanings (when one form can express more than one meaning; e.g., -​s; see Table 1.1 for examples). The linguistic form of a construction and its use in communication is built up over time through usage and agreed upon by a community of speakers (Beckner et al., 2009; Christiansen & Chater, 2016). Constructions are therefore socially learned, a process that can lead to changes in the form of a construction (e.g., phonetic reduction, grammaticalization, see Bybee, 2010) and in the meaning of a construction (e.g., constructional change, will-​“intend” > “future”, see Traugott & Trousdale, 2013). As Table 1.1 shows, constructions can vary in terms of their functional and syntactic complexity and include morphemes, words, complex words, and more abstract syntactic frames (e.g., ditransitives and passives). Constructions can be relatively concrete, as in fish, while others can be more abstract. For example, the sentence she gave Ellie a present is made up of individual constructions (she, Ellie, a, present, gave) that when ordered in this specific way (Subject Verb Object Object) express the meaning of something being transferred (Ellis et al., 2016). Building on this understanding of constructions as the building blocks of language, the language input that speakers are exposed to and use are a critical source for building knowledge of that language (Ellis, 2006a; Goldberg & Casenhiser, 2008; MacWhinney, 2008). However, exposure to a language alone is not enough. Domain-​general learning mechanisms are needed to support the learning of a language (Bates & MacWhinney, 1989; Bybee, 2010; Christiansen & Chater, 2016).

4

4  Introducing Crosslinguistic Influence

The cognitive mechanisms used by humans to learn language are described as domain-​general because humans are understood to learn a language in the same ways that they learn any other ability or skill. Examples of the cognitive mechanisms used by humans to learn and use language include rich memory storage, analogy, categorization, and cross-​modal association (Bybee, 2010; Christiansen & Chater, 2016). When speakers encounter instances of language in the environment, for example, categorization is used to match those instances (e.g., sounds, text, images, gestures) to existing experiences with language that are already represented and stored in memory. This process allows messages to be decoded, interpreted, and also for new information to be associated with existing experiences. Our ability to categorize information cooperates with a range of other cognitive processes, including rich memory storage. Work in connectionism indicates that representations are stored and organized in self-​organizing maps (Kohonen, 1990; Li et al., 2004; Shirai, 2019), which are networks of representations stored as neurons or units. These networks are activated when input is received and compete for selection based on their specifications and how well they match to the input. Categorization and rich memory storage are just some of the domain-​general cognitive mechanisms used by humans to learn language. In sum, language learning involves the building of language knowledge through exposure to instances of language as used in communication. Language learning is supported by powerful sets of domain-​general cognitive mechanisms.

1.3  Learning a Second Language In discussions of L2 learning in adulthood, we should remind ourselves that all speakers do this with an established system of language knowledge built up from prior experience using some other language (typically L1). This is an important difference between child L1 learning and adult L2 learning. In crosslinguistic influence research, therefore, our goal is to describe and explain in what ways a speaker’s prior knowledge of language and experiences using language shape new language learning. In the field of SLA, research has repeatedly shown that adults regularly construct knowledge about a new language that is different from that built for L1 use (Alonso , 2016; Jarvis & Pavlenko, 2008; Odlin, 1989). Examples of this include instances where the same forms in L1 and L2 (e.g., articles, verb forms) express different or additional meanings. Articles in English (a, the) and languages like French (un “a”, la “the”), German (ein “a”, die “the”), and Spanish (un “a”, la “the”) are examples of this L1–​L2 difference. Even though all these languages can express definiteness or specificity with articles (e.g., a cat vs. the cat or eine Katze vs. die Katze), in languages like Spanish and German articles also express information about grammatical gender (e.g., in German die “the” is used with feminine nouns, das “the” with neuter nouns, and der “the” with masculine nouns; Durrell, 2011).

4

5

4

Introducing Crosslinguistic Influence  5

One consequence of this L1–​L2 difference is that speakers of Spanish or German, for example, can use articles to anticipate or predict upcoming nouns, but speakers of languages like English that lack grammatical gender do not (Kaan, 2014). This crosslinguistic difference is due to agreement conditions between articles and nouns that are present in languages like Spanish. This means that when Spanish speakers encounter the feminine determiner la, for example, they anticipate that the upcoming noun will be feminine as well (la-​FEM manzana-​FEM “the apple”). Since English nouns are not marked for grammatical gender, English speakers do not carry out the same type of anticipatory processing using words like the (but see DeLong et al. (2005) for anticipatory processing with singular definite articles, a vs. an, e.g., an airplane but a kite). This example illustrates an instance where the same forms (or cues) in L1 and L2 (articles) do not express the same information and are therefore not used in the same way. Learning situations like this constitute one of the more complex challenges in L2 learning because creating new form-​meaning mappings in L2 can compete with existing and established form-​meaning mappings in L1 (e.g., the same cue expresses meaning x in L1 but meaning y in L2). Even though L1–​L2 form-​meaning mapping differences can lead to persistent learnability difficulties for L2 speakers, even after considerable exposure to the target language, SLA research has shown that L2 form-​meaning mappings that are different from those used in L1 can be learned. In the English–​Spanish example, research has shown that English speakers can eventually learn that Spanish articles express grammatical gender, even though English articles do not express grammatical gender (e.g., Dussias et al., 2013). Research evidence like this has come from offline tests that require learners to produce the grammatical gender of a particular noun. An example of an offline test used in this line of research can be found in Hopp (2013), in which four images were presented on a screen and learners were instructed to say aloud each image and its color in German (e.g., das grüne Auto “the green car”, die gelbe Karte “the yellow card”). This test was used to assess whether L2 German learners could produce the correct article with nouns of different grammatical genders (i.e., did the learner use the correct article (das) with Auto?). This type of test is often described as offline because it is not time pressured, which allows speakers to draw on metalinguistic knowledge or other problem solving and reflective abilities to complete the task (R. Ellis et al., 2009). The next step in this research program has involved anticipatory processing tests that are time pressured (or online). These tests require learners to use or apply their knowledge about the grammatical gender of nouns in real time (i.e., with less preparation time and less time to reflect). In Hopp (2013), learners saw four images on a computer screen and heard a short sentence that included an article cue. Importantly, the article cue contained information about grammatical gender (e.g., wo is die gelbe Karte? “where is the yellow card?”). Learners were instructed to look at the image that matched what they heard (i.e., on hearing wo is die gelbe Karte, learners should look toward the image of the yellow card). The analysis

6

6  Introducing Crosslinguistic Influence

examined the speed and accuracy of looks toward the target image using an eye-​ tracker (for a review of eye tracking in SLA research, see Godfroid, 2019). When test items are carefully balanced for grammatical gender (e.g., only one image out of the four matches the grammatical gender of the stimulus), do learners wait until they hear the noun (Karte “card”) before looking to the target image or do they look upon hearing the article (die “the-​FEM”)? If the learner looks toward the correct image when the article cue is heard (e.g., wo is die… “where is the-​FEM …), this behavior indicates that the learner is engaging in some type of anticipatory processing. Using research designs like this, some studies have shown that English speakers can use gender-​marked articles to anticipate upcoming nouns (Kaan, 2014), even though English determiners are not used in this way. These findings indicate that L2 speakers can learn new form-​meaning mappings that are different from those used in L1. Despite such evidence, however, research has showed that not all speakers who demonstrate offline knowledge of a target feature can use this knowledge during real-​time performance (Kaan, 2014).This means that speakers might perform well in offline tests that assess knowledge about the grammatical gender of nouns, but that performance in online tests indicates that this knowledge cannot be used in real time. Findings like this have led to proposals that the knowledge L2 learners build might be qualitatively different from the knowledge built and used in L1 performance (for review, see R. Ellis et al., 2009). Here, the main claim is that even though L2 learners appear able to create explicit/​declarative and verbalizable knowledge of the L2, such as offline grammatical gender knowledge of nouns, developing L2 knowledge that can be used in more unplanned and spontaneous tasks might be more difficult (e.g., Grüter et al., 2017). Work in the anticipatory processing of Spanish grammatical gender suggests that it might not be that L2 knowledge is qualitatively different from L1 knowledge, but that other factors related to proficiency and/​or experience might play some role (see also Hopp, 2013). For example, Dussias et al. (2013) showed that low proficiency, English-​speaking learners of Spanish struggled to anticipate upcoming Spanish nouns using gender-​marked determiners in online tests despite demonstrating offline grammatical gender knowledge of Spanish nouns. This evidence suggested that knowledge of grammatical gender in Spanish might have been explicitly learned only, thus making it difficult to for learners to perform well in tasks that require rapid access to that knowledge. However, low proficiency, Italian-​speaking learners did appear to be engaging in some type of anticipatory processing in Spanish. One explanation for why the English-​and Italian-​speaking learners performed differently can be found in how English and Italian use articles with respect to Spanish. In short, Italian marks grammatical gender on articles, but English does not. As a result, Italian speakers can draw on their prior knowledge and experience of using articles to anticipate upcoming nouns, but English speakers have no such knowledge or experience to draw on. Performance from more experienced English-​speaking learners of Spanish,

6

7

6

Introducing Crosslinguistic Influence  7

however, did show evidence of anticipatory L2 processing. English speakers likely needed more time and experience using the L2 to develop anticipatory processing behaviors involving articles. This is because English speakers must build knowledge of this new form-​meaning mapping from scratch (unlike Italian speakers). In addition, practice and usage are needed to develop this new knowledge for use in real-​time processing. Indeed, research showing that usage and/​or practice can lead to the development of more automatic (or fluent) L2 abilities indicates that L2 learners can also build knowledge that is usable in real time (see DeKeyser, 1997, 2017; Segalowitz, 2010). For example, a small number of studies have tracked L2 learners’ performance longitudinally to understand the cumulative effects of usage (or practice) on performance (e.g., DeKeyser, 1997; McManus & Marsden, 2019b). This work shows that extensive and repeated opportunities for practice can lead to the development of automatic L2 abilities for target features that are crosslinguistically different. Indeed, this could be one explanation for why Dussias et al.’s (2013) advanced proficiency English-​speaking learners of Spanish were able to use their knowledge of grammatical gender in real-​time tests while less proficient English speakers showed greater difficulty. More experience processing Spanish article-​ noun combinations likely played an important role in developing more automatic L2 processing behaviors. Taken together, research has shown that adults can learn aspects of a new language that are different from L1. Research has also shown that the development of more automatic (or fluent) L2 abilities is possible with practice and experience using the L2. Training studies that track usage longitudinally, of both novel and natural languages, provide the clearest evidence in support of this as well as studies that assess different types of L2 performance (e.g., online, offline). Therefore, even though L2 knowledge can be, at least initially, explicit in nature, increased opportunities for practice and usage can lead to the development of more automatic L2 abilities.Today, these understandings are well represented in current theories of L2 learning (see VanPatten et al., 2020).

1.4  Prior Knowledge and Experience Up to this point, we have considered some of the conversations taking place in our field about what L2 learning involves, including discussions about the types of language knowledge that learners can build and the extent to which learners are able build and develop L2 knowledge that is different from that used in L1. A key factor here is that experience is critical in developing L2 abilities, both in terms of the prior experience (e.g., L1) that speakers bring to L2 learning as well as the experience that the learner has with using the L2. Both types of experience are important. In this section, we consider the prior experience that speakers bring to L2 learning. As previously noted, most adult speakers begin L2 learning with an established system of language knowledge. This description acknowledges that prior to

8

8  Introducing Crosslinguistic Influence

learning a new language most adults are accomplished and expert users of at least one language, often their L1, but sometimes other languages too (De Houwer & Ortega, 2018). This is an important difference between child and adult language learning that SLA research seeks to account for. By accomplished and expert user, we are saying that almost all adults can use their L1 (and/​or other languages) without needing too much planning, they can tell when something does not seem to make sense or sound right (even if they cannot identify why), and they can do all of this relatively quickly while often undertaking other tasks (e.g., having a conversation while walking down the street, see Segalowitz, 2010).2 At the same time, adults can use their L1 in a very deliberate and/​or reflective manner. For instance, maybe careful planning is needed because of the delicate nature of the conversation topic, perhaps somebody asked the speaker for directions to an unfamiliar location, or maybe a conversation triggered an idea or memory about something else, all of which could lead a speaker to be more reflective and/​or hesitant in how they use their language(s). This expertise in using language both very deliberately and relatively automatically is the culmination of a lifetime of daily language use, which has developed, fine-​tuned, and committed the speaker’s cognitive mechanisms for the optimal use of that language (Christiansen & Chater, 2016; Ellis, 2006a; MacWhinney, 2008). These observations about language use that adults bring to the task of learning something new are important for understanding L2 learning.This is because adults who begin learning an additional language do so after having already learned and mastered many complex ideas and concepts about the world as well as how these ideas and concepts are communicated in the language(s) they use. For example, almost all adults will have learned and mastered some understanding of the concept of time (e.g., pastness, futurity, see Comrie, 1985; Klein, 1994) as well as the ability to express and comprehend different locations in time as expressed in their language(s).This ability includes an understanding that, in English, she played tennis and she is playing tennis express different locations in time with reference to the speaker. So, we are saying that to comprehend information about time as expressed in these utterances, the English speaker needs some type of knowledge about the linguistic cues used in English to express time. In our example, this is knowledge that grammatical information expressed on English verbs can tell us something about time (e.g., -​ed to indicate past time). Not only do most adults develop expertise like this for their language(s) (i.e., knowledge of the most informative cues and the meanings they index), but this is an ability that can become well-​ rehearsed, automatic, and established with usage or practice (Anderson, 1982; DeKeyser, 2017; Segalowitz, 2010). In short, greater experience with language can optimize language use. If we apply these understandings to L2 learning, it can help us to make sense of why a large body of research has repeatedly shown that the cumulative experience of using and attending to a particular language (e.g., L1) can majorly influence the learning of an additional language, positively and negatively. The effects

8

9

8

Introducing Crosslinguistic Influence  9

of prior knowledge and experience on L2 learning can be positive (or facilitative) when the patterns of language use are similar across languages, such as when languages share sets of meanings/​concepts (e.g., time, aspect) and/​or express the same meanings/​concepts using similar means (e.g., expressing differences in time with verbal inflectional morphology, as in English and German, see Comrie, 1985). SLA research has indeed shown this to be the case: L2 learning can be easier when L1 and L2 express the same concepts using similar linguistic means (Jarvis & Pavlenko, 2008; Odlin, 1989). On the other hand, prior knowledge and experience can complicate L2 learning when languages differ in specific ways, including when shared concepts are expressed differently. For example, even though English and Mandarin Chinese share understandings of the concept of time, these languages do not express time in the same ways (Klein, 1994; Klein et al., 2000). For example, English can express temporal reference using verbal morphology (e.g., she is watching TV vs. she watched TV), but Chinese uses other means such as temporal particles (e.g., le, zhe, guo), discourse principles, and lexical means. For the Chinese-​ speaking learner of English, therefore, even though it is helpful that L1 and L2 share a conceptual understanding of time (thus not requiring new conceptual learning), the Chinese speaker’s prior L1 knowledge and experience for how to express time is less helpful. The Chinese speaker has to learn new and different ways for expressing this concept in L2 (i.e., that inflectional morphology on English verbs can provide information about time). L2 research has repeatedly shown that L2 learning can be more difficult when L1 and L2 express the same meanings/​concepts differently (Jarvis & Pavlenko, 2008; Odlin, 1989). In addition to documenting crosslinguistic differences/​similarities in usage and their impact on L2 learning, we also want to understand why prior knowledge and experience has the effect that it apparently does. For example, even though there is research evidence that the negative effects of crosslinguistic influence can be fleeting in some cases but longer lasting in others, few accounts of L2 learning agree on explanations for these findings (Lardiere, 2009; MacWhinney, 2005; O’Grady, 2015). Frequency and exposure to language input can go a long way to explaining some of these findings, but frequency-​based explanations by themselves are not sufficient (see Divjak, 2019; Ellis, 2006a). Accounts that attribute additional roles to how speakers use and process the input provide increasingly robust insights into how L2 learning unfolds (e.g., Ellis & Sagarra, 2010, 2011). Indeed, this makes good sense not only because adults bring powerful sets of cognitive learning mechanisms to the task of constructing language knowledge (categorization, chunking, rich memory storage, analogy; see Bybee, 2010; Ellis & Robinson, 2008), but also because these learning mechanisms have already been committed to the task of efficient and effective L1 use. This means that L2 speakers begin learning a new language with cognitive mechanisms that have been optimized for processing and using a different language. How speakers attend to and process new linguistic input is therefore influenced by prior experience. This

10

10  Introducing Crosslinguistic Influence

is one reason for why some of the negative effects of crosslinguistic influence can be fleeting in some cases but more persistent in others. It all depends on what experience learners bring to L2 learning. As a consequence, advancing multifactor accounts of learning that are both sensitive to the input and how prior experience biases processing are critical for understanding crosslinguistic influence and learning more generally.

1.5  Transfer In addition to the labels “prior language knowledge” and “experience”, transfer is also a commonly used term in discussions and explanations of L2 learning. Readers familiar with SLA research will be aware that transfer is often used to describe and/​or explain L2 performance that appears to be influenced by, draws on, or uses some type of prior language knowledge (e.g., L1) in the learning and use of a new language. Odlin (1989), for example, defines transfer as follows: Transfer is the influence resulting from the similarities and differences between the target language and any other language that has been previously (and perhaps imperfectively) acquired Odlin, 1989, p. 27 In this definition, Odlin sees transfer as an outcome or a result of the linguistic differences and similarities among a speaker’s languages (see also Anderson, 1983; Kellerman, 1995). Indeed, reviews of the field have indicated that this has been an influential way to think about transfer (see Gass et al., 2020), which has led to the development of methodological frameworks for identifying instances of transfer (e.g., Jarvis, 2000, 2010). At the same time, however, it has been noted that very little is known about what transfer actually is (e.g., Jarvis & Pavlenko, 2008; Odlin & Yu, 2016; Sharwood Smith & Truscott, 2006, 2014). This is a problematic state of affairs for our field for many reasons, not least because transfer is a common description and explanation for L2 learning. One consequence of this is that we can identify instances of transfer, but we do not fully understand what triggers or leads to transfer. One (common) interpretation of transfer is that it involves copying and/​or cloning one body of knowledge (e.g., L1) to create a new body of knowledge (e.g., L2). This interpretation, as Sharwood Smith and Truscott (2006, 2014) have discussed, has its roots in the “every day” interpretation of what transfer is (e.g., “transferring water” is interpreted as moving water from one container to a different container), but with some modification because transfer in L2 learning cannot mean that L1 knowledge is moved to a different place: The essential problem is that transfer, in the everyday sense of moving something from one location to another, does not make immediate and

10

1

10

Introducing Crosslinguistic Influence  11

obvious sense. The L1 elements that are supposed to be imported into an L2 do not leave the L1 and automatically impoverish it. Hence, the closest we can get to this conceptualisation is to say “transferring” something must mean “copying” or “cloning” it, leaving the original in place. Sharwood Smith & Truscott, 2006, pp. 202–​203 By interpretating transfer in terms of copying L1 knowledge, this allows the L1 system to stay intact. L2 knowledge (i.e., a copy of L1 knowledge) is then hypothesized to be modified through exposure to L2 input (see Schwartz & Sprouse, 1994, 1996, 2020). In this view, instances of negative transfer are explained as L2 knowledge that has not yet been fully modified. Even if we accept this conceptualization of transfer as a process of copying and restructuring, there are still many questions about transfer: How is transfer triggered? Do changes to transferred knowledge also affect the source (e.g., L1)? Could there be some type of exposure threshold or processing difficulty that triggers transfer? Does the copy-​ and-​restructure conceptualization of transfer entail that L1 knowledge is separate from L2 knowledge? In short, given that transfer is a common label used in descriptions and explanations of L2 learning, it is problematic that we understand so little about it. In this book, we both problematize the copy-​and-​restructure conceptualization of transfer and consider an alternative account for the effects of transfer (or crosslinguistic influence). In particular, we explore whether crosslinguistic influence can be explained more simply as the use of existing language knowledge to process new information (e.g., the use of L1 knowledge to interpret and produce L2 cues). In line with conceptualizations of transfer from cognitive psychology (Anderson, 1982; Larsen-​Freeman, 2013; Nokes, 2009), this view does not require knowledge to be copied and subsequently restructured. Instead, L2 learning involves (i) the creation of new knowledge when L1–​L2 differences exist and (ii) the creation and/​or development of selection mechanisms to manage and select among competing knowledge sources in L1 and L2 (McManus, 2021). This conceptualization of crosslinguistic influence is informed by work in emergentist accounts of L2 learning (e.g., O’Grady 2013, 2015) and language selection (Green, 1998). Explaining crosslinguistic influence in this way means that we also need to be able to explain (i) when and how new language knowledge is constructed, (ii) how L1 and L2 knowledge sources are connected, (iii) how knowledge sources in L1 and L2 are selected, and (iv) what the learning processes are that give rise to the creating of language knowledge and its selection. In this book, we focus on these questions while exploring others related to them.

1.6  Cross-​language Relationships Perhaps one of the most frequently asked questions in SLA –​but one that has been shown to be one of the difficult to investigate –​is to what extent a speaker’s

12

12  Introducing Crosslinguistic Influence

(knowledge of) different languages might be connected, and, if so, in what ways are the languages connected. Indeed, the preceding discussion highlighted that a copy-​and-​restructure view of transfer suggests that L2 learning leads to the creation of a new body of language knowledge. Whether these bodies of knowledge are connected or not is not well understood. L2 knowledge that (i) appears different from that used in L1 performance and (ii) can be used efficiently and relatively automatically likely indicates that a speaker’s L1 and L2 knowledge sources coexist (Hartsuiker & Pickering, 2008; Kroll et al., 2002; Sanoudaki & Thierry, 2014). That is, L2 learning does not appear to erode or replace L1 knowledge (although extensive L2 use has been argued to lead to convergence of L1 and L2, see Dussias & Sagarra, 2007). Indeed, as we have seen, a dominant proposal in the field is that L2 knowledge begins as a clone of L1 knowledge and that this newly cloned language knowledge can be modified through exposure to input over time to facilitate L2 processing and use (see MacWhinney, 2005; Sharwood Smith & Truscott, 2006, 2014). This account suggests that learners manage multiple language systems from the outset. Indeed, in a review of this research, Kroll et al. (2012) refer to the ability to manage multiple languages as “mental juggling”. Comparisons of L1 and L2 performance in the same speakers can provide support for claims that L1 and L2 knowledge representations coexist (Kang et al., 2018; McManus, 2021; Timmer et al., 2019). Research investigating these questions additionally suggests that a speaker’s other known languages are active to some degree when one of them is being used (for reviews, see Kroll & Bialystok, 2013; Kroll et al., 2015), meaning that L2 users cannot/​do not switch off the language not in current use (Chen et al., 2017; Dijkstra, et al., 2000). The use of event-​related potentials (ERPs) to provide an account of brain activity during language use has made an important contribution to this understanding. In Chapter 3, we discuss work by Thierry and Wu (2007) that investigated activation of L1 knowledge during L2 comprehension. Briefly, Thierry and Wu (2007) presented Chinese-​speaking learners of English with pairs of English words (e.g., post –​ mail, train –​ ham). Half of the words concealed a character repetition when translated into Chinese and half did not. The authors reported an effect of concealed character repetition on L2 performance, indicating that L2 speakers were sensitive to Chinese character repetition even though they were reading English words. These results indicate not only complex relationships between a speaker’s L1 and L2 knowledge sources, but, importantly, they suggest activation of L1 knowledge during L2 comprehension. Findings such as these are important for informing how we think about the potential connections among a speaker’s different languages. They suggest that a speaker’s knowledge of other languages is activated when only one language is being used. Given that interest in this type of research is growing and our understanding of cross-​language relationships is developing, SLA theory has sought to explain how a speaker’s language system might allow for co-​activation. In other words, how does evidence about cross-​language activation inform theorization about

12

13

12

Introducing Crosslinguistic Influence  13

the organization of a speaker’s system of language knowledge? For example, work in connectionism indicates that new language knowledge becomes integrated and connected with networks of existing language knowledge (Shirai, 2019). Knowledge representations in L1 and L2 are therefore understood to be independently represented but connected via complex networks involving different layers of linguistic form and/​or conceptual meaning (Gasser,1990; Shirai, 2019). This close network of connected language representations is one explanation for co-​activation during language use. Because L2 speakers are understood to be constantly managing (or “juggling”) multiple languages during their use of a single language and because L1 and L2 representations appear to be closely connected and tightly networked at multiple levels of representation, an additional consideration in this line of research has investigated the extent to which some type of cognitive process or mechanism might be involved in how speakers select a specific language. The question here is as follows: to what extent must L2 speakers employ or develop some type of cognitive mechanism that permits specific types of language knowledge to be selected/​used (Calabria et al., 2018; Green, 1998). In other words, can a speaker select or inhibit a particular language at a given time? The area of research investigating this question is known as language control. It is based on the understanding that, for multilinguals, the creation of L2 representations are, by themselves, not sufficient for the appropriate use of that language: “Learners of an L2 need to learn how to keep the two languages separated to avoid interference, and learn to select one language or the other in each given communicative situation” (Calabria et al., 2018, p. 221). To illustrate: For an English-​speaking learner of German, recognition of a specific ordering of letters in reading (e.g., R-​O-​C-​K) activates multiple candidate interpretations in the speaker’s languages (i.e., English, German), such as a genre of music (as in “rock music”), the naturally occurring mineral material (as in “sedimentary rock”), or a piece of clothing (as in das Rock, “skirt”). This activation of multiple candidate interpretations results in competition. The question here is how do speakers manage and select among these different candidate interpretations? In other words, how do L2 speakers manage the conflict (or competition) that comes with knowing multiple languages? One approach to addressing this question involves inhibition or inhibitory control (Abutalebi & Green, 2007; Green, 1998; Green & Abutalebi, 2013), an important cognitive mechanism understood to help manage cross-​language behavior by preventing selection of the inappropriate language during language use (Miyake et al., 2000; Stahl et al., 2014). One influential model of inhibition is Green’s (1998) Inhibitory Control Model (see Chapter 2). In a nutshell, Green (1998) proposed that language representations are tagged according to language (e.g., L1 tags for L1 representations, L2 tags for L2 representations) and that inhibitory control mechanisms facilitate use of a particular language by using language tags. Language use is facilitated by inhibiting dominant or competing responses

14

14  Introducing Crosslinguistic Influence

(e.g., inhibiting L1 knowledge representations during L2 use). Tag suppression is one way to prevent nontarget language selection. In other words, inhibitory control can help selection of L2 representations by actively suppressing L1 representations when the speaker intends to use the L2.This type of cognitive architecture is especially relevant to L2 learners because among this population L1 representations are the most dominant, at least in initial L2 learning. This is because L1 knowledge is well established, entrenched, and routinized through previous experience. In comparison, L2 representations are relatively weaker and more effortful to use. As we understand it, the language processing system will opt for the least effortful routine, which is usually the dominant L1 routine. Tag suppression facilitates L2 use by inhibiting the L1 system during L2 use. We will review Green’s (1998) Inhibitory Control Model in more detail in Chapter 2. An alternative account to understanding cross-​language relationships, competition, and how the system manages language selection is through a process of activation, which, by itself, does not rely on other cognitive mechanisms to explain language selection (see MacWhinney, 2005; MacWhinney et al., 1989; Sharwood Smith & Truscott, 2014). One influential model of activation is MacWhinney’s (2005) Unified Competition Model (UCM, see Chapter 2). The UCM draws on the combined processes of lexical activation and resonance to hypothesize how competition is resolved and how the appropriate/​desired language representations are selected. First, lexical activation is a cooperative process in which a specific language representation stimulates neighboring representations of a specific type because of connected and/​or shared features of the construction. This is possible because constructional representations are stored with information about dependencies and lexical expectations. As a result, activation of a specific representation triggers other expectations to be set up and the representations that meet those expectations will be activated (MacWhinney, 1987). For example, activation of the transitive verb pet opens up slots in the dependency structure to be filled by two arguments, such as “a petter” and “a pettee” (as in she pets the cat). In this example, activation of the transitive verb “pet” triggers expectations that two arguments are needed. Previous experience processing these structures leads to the generation of expectations. Resonance also contributes to selection processes because if the nodes of a particular language have been recently activated or are more frequently activated in general, they are understood to be in a highly resonant state, thus increasing the likelihood of their activation. Therefore, language representations that are highly resonant are understood to be selected or activated more easily than those that are of lower resonance. Appealing to any one of these processes of cooperation, activation, and inhibition alone to help explain how L2 speakers manage language knowledge sources is likely too limiting (but see MacWhinney, 2005; Sharwood Smith & Truscott, 2014). It seems possible that language processing draws on a variety of processing

14

15

14

Introducing Crosslinguistic Influence  15

mechanisms to manage different aspects of L2 behavior. We will explore these issues in greater detail in Chapter 2.

1.7  Directions of Crosslinguistic Influence This section is about the directions of crosslinguistic influence, which refers to the types of cross-​language relationships that researchers in the field of SLA have studied, including how L1 knowledge and experience can influence L2 learning, how knowledge of both L1 and L2 can influence the learning of an additional language, and how L2 learning can influence L1 use. This work has considered both unidirectional influence (e.g., L1 effects on L2 learning) as well as bidirectional influence (e.g., L1 effects on L2 use and L2 effects on L1 use, see Pavlenko & Jarvis, 2002). As previously noted, the dominant approach to researching crosslinguistic influence in our field has involved investigating the ways in which L1 knowledge and experience influences the routes and rates of L2 learning (for reviews, see Alonso Alsonso, 2016; Jarvis & Pavlenko, 2008; Odlin, 1989; see Chapter 3). More studies about L3 learning are beginning to emerge as well.These take into account the cumulative effects of L1 and L2 knowledge on L3 use (e.g., Hopp, 2019). Here, we consider how understudied directions of crosslinguistic influence, including how L2 learning can influence a speaker’s already known languages (e.g., L1), have the potential to advance knowledge and understanding about crosslinguistic influence. One view of research to date about crosslinguistic influence is that our evidence base is relatively narrow. This is because we have prioritized study designs that investigate how a speaker’s L1 background influences their learning of an additional language. Even within this particular cross-​language relationship, L2 research has predominantly focused in one direction: L1 effects on L2. In comparison, much less is known about the other directions of crosslinguistic influence (e.g., L2 effects on L1). Compared to what we know about L1 effects on L2 learning, we know relatively little about how L2 learning affects L1 use or how L3 learning affects L2 use (for reviews, see Cook, 2003; Kroll et al., 2018; Liu & Cao, 2016). This is a notable limitation, especially because the multilingual mind is increasingly understood to involve a series of multidirectional, cross-​ language connections (Grosjean, 1989), and many of the populations we study are speakers of multiple languages (Auer & Wei, 2008; De Houwer & Ortega, 2018). A focus on just one of these cross-​language relationships and in one direction prevents us from more fully understanding the complexity and multifaceted nature of language learning as well as the learning processes that potentially lead to crosslinguistic influence. To address this limitation, research is needed that takes into account other types of cross-​language relationships. If we think back to our previous discussions about the organization of the L2 speaker’s language system, we noted that this was understood to be a dynamic

16

16  Introducing Crosslinguistic Influence

FIGURE 1.1 Visualization

of a Web

and closely networked system of nodes and connections (see Shirai, 2019). MacWhinney (1987) used a metaphor of a spider’s web to conceptualize these relationships (see Figure 1.1). Using this metaphor, the claim is that if you choose a particular node in a web and pull it in one direction, the other nodes that are connected to that node move with it. Using Figure 1.1., you can imagine what this might look like. If you pull on one of the black spots, the others that are connected to that black spot also move, and so on. In short, the entire network responds to stimulation. There is broad consensus that this organization makes good sense given what we know about the co-​activation of a speaker’s languages. Based on this understanding, L2 learning would be expected to influence L1 use to some degree due to the emergence of L1–​L2 connections. This understanding is indeed borne out in the small number of studies that have investigated how L2 learning leads to changes in L1 use. In Chapter 3, we review some of this research. As we have already noted, investigating other directions of crosslinguistic influence is an important line of research if we want to understand multilingualism. Importantly for SLA research, this work suggests that additional language learning leads to changes in the nature of a learner’s existing language system(s), including changes to the ways in which L1 knowledge is represented and accessed as well as the potential convergence of L1 and L2 representations. These are important contributions to theory-​building that studies of L1 effects on L2 learning cannot/​have not sought to understand, by design. That is, investigating L1 effects on L2 learning cannot tell us about the ways in which L2 learning appears to change a speaker’s existing language system(s). Such work is critical in order to more fully understand the extent and nature of crosslinguistic influence. In sum, L2 researchers need to investigate directions of crosslinguistic influence beyond L1 effects on L2 learning in order to advance knowledge and understanding about the nature and emergence of cross-​language relationships. It

16

17

16

Introducing Crosslinguistic Influence  17

is only through a broader examination of cross-​language relationships that the field can more deeply understand the nature and effects of crosslinguistic influence.

1.8  Explicit Instruction and Crosslinguistic Influence A feature of crosslinguistic influence research is that it seeks to understand the ways in which a speaker’s experience with one or more languages influences their use of a different language. Following work in the 1970s that attempted to predict linguistic features that might be difficult to learn by describing and comparing linguistic structures (e.g., Contrastive Analysis, see Lado, 1957), it was generally concluded that comparing language structures was not a good way to approach language teaching. One reason for this is that predictions for crosslinguistic influence were not always borne out by early descriptive-​structuralist work (for discussion, see Keck & Kim, 2014; Ortega, 2009).The field’s response to this early work at understanding cross-​language relationships is now considered a knee-​jerk reaction (Divjak, 2019; MacWhinney, 1992). This is because thinking at that time was predicated on the notion that prior language knowledge and experience did not play a meaningful role in shaping the routes and rates of L2 learning (e.g., Bailey et al., 1974). Interestingly, time has shown that this conclusion was premature. Methodological and conceptual rigor in linguistic description and understanding of learning have shown that one weakness of this early descriptive-​structuralist work was that its predictions for crosslinguistic influence were largely based on structural differences between languages rather than considerations of form-​meaning mappings. Today, equipped with meaning-​based or functionalist conceptualizations of language, researchers have focused their attention on the ways in which languages are similar or different in terms of how they express similar sets of meanings (e.g., English and Mandarin Chinese both express information about time, but do so in different ways) or the extent to which some meanings appear not to be used or represented in one language but they are in another (e.g., grammatical gender in Spanish but not in English). One important contribution of this work is an awareness that prior language knowledge can have both positive and negative effects on additional language learning. In Chapter 4, we review an active but under-​studied line of research that has used evidence about crosslinguistic influence to investigate what role explicit instruction can play in reducing some of the negative effects of crosslinguistic influence. This research points to a growing body of evidence that L2 learning can be facilitated by providing explicit instruction about the nature of L1–​L2 differences. Importantly, however, this approach to instruction differs from previous approaches (e.g., Lado, 1957) in that it uses previous SLA research evidence to determine what is difficult to acquire. This contrasts with contrastive-​analysis based approaches that used descriptive-​structuralist accounts of language to predict what is difficult to learn.

18

18  Introducing Crosslinguistic Influence

1.9  Overview of the Book In the final section of this chapter, we provide an overview of the book. The main concepts and ideas that have been introduced in the previous sections of this chapter will be developed in subsequent chapters. Given that the aim of this book is to advance new avenues of research in crosslinguistic influence, each chapter focuses on what we know and what gaps still exist. One reason for this is that even though a lot is known about crosslinguistic influence, there is still considerable work to be done, as highlighted in Section 1.5 about transfer and Section 1.7 about the directions of crosslinguistic influence. In Chapter 2, readers are introduced to four theoretical models of L2 learning that assign critical roles to knowledge of other languages in understanding how speakers use and learn an additional language: The Unified Competition Model (MacWhinney, 2012), the Associate-​Cognitive CREED (Ellis, 2006a), Processing Determinism (O’Grady, 2015), and the Inhibitory Control Model (Green, 1998). These models of L2 learning all subscribe to functionalist understandings of language. The nature of their difference, therefore, is less in terms of how they conceptualize language, but more in how they understand the learning process. The theoretical foundations of each model are reviewed and the predictions each makes for L2 learning are explored. Chapter 2 ends with a synthesis of the similarities and differences among the models. Chapter 3 synthesizes and problematizes empirical L2 research about crosslinguistic influence in three areas –​morphosyntax, vocabulary, and phonology –​to provide the reader with a firm grounding and solid understanding of some important lines of crosslinguistic influence research in the field. For each area, main findings are presented and critiqued. Chapter 4 begins by contextualizing what we know about instructional effects in L2 learning, including a review of key debates about explicit knowledge in L2 learning and the extent to which instruction can benefit the routes and rates of L2 learning.These discussions lead into this chapter’s main focus: how the findings of L2 research and theories of L2 learning can be used to support L2 learning in the three areas discussed in Chapter 3 (morphosyntax, vocabulary, phonology). In this chapter, we therefore seek to highlight how instruction that addresses crosslinguistic influence can facilitate learning. Finally, in Chapter 5, the key points of Chapters 2–​4 are summarized and synthesized in order to reflect on the current state of understanding about crosslinguistic influence in L2 learning as it relates to theory, learning, and instruction. The following areas are selected for particular reflection: the organization and restructuring of knowledge, experience, cross-​language relationships, instruction, and reflections on theorizing crosslinguistic influence. Future directions in crosslinguistic L2 research are suggested. At the end of the book, suggestions for further reading are proposed, including a brief summary of the recommended texts. These readings have been selected to

18

19

18

Introducing Crosslinguistic Influence  19

allow readers to learn more about some of the topics presented in the book that connect with questions in crosslinguistic influence (e.g., connectionism, metalinguistic awareness).

Notes 1 Following convention in the field of SLA, we use the term “L2” and “additional language” to refer to non-​primary language learning (i.e., when a speaker learns a new language after already learned their first language). The labels “L2” and “additional language” are used interchangeably. 2 L1 knowledge and use among heritage language speakers and attritors can be somewhat different. For discussion, see Montrul (2010), Schmid and Köpke (2019).

20

2 THEORETICAL MODELS OF L2 LEARNING

2.1  Introduction In this chapter, we review four theoretical accounts (or models) that forefront critical roles for prior language knowledge and experience in describing and explaining L2 learning and use. The models of learning discussed include: The Unified Competition Model (MacWhinney, 2012), the Associative-​Cognitive CREED (Ellis, 2006a, 2006b), Processing Determinism (O’Grady, 2015), and the Inhibitory Control Model (Green, 1998).1 Consistent with the discussions of language and learning presented in Chapter 1, each model is grounded in conceptualizations of learning that are driven by usage. This means that the routes and rates of L2 learning are understood to be shaped by a speaker’s (i) prior knowledge and experience with a particular language and (ii) use of the L2. Usage is therefore a primary contributor to development. Indeed, an important feature of usage-​based accounts of learning is that they seek to understand and explain how speakers learn to use language in order perform a range of socio-​cognitive functions, or, following Bates and MacWhinney (1981), usage-​based accounts seek to understand how speakers go about constructing a “performance grammar”, defined as follows: a unified theory of the pragmatic, semantic, and perceptual processing strategies that adults and children use to comprehend and produce sentences, inside and outside of a discourse context. Such a grammar would focus not only on the “possession” of a rule by a language or by an individual, but on the way grammatical information is handled in real time. Bates & MacWhinney, 1981, p.190

DOI: 10.4324/9780429341663-2

20

21

20

Theoretical Models of L2 Learning  21

As a result, usage-​based accounts of L2 learning involve not only proposals about the nature of language and how it is stored and represented, but they also describe and explain how learners use language in real-​time to carry out specific communicative functions. This means that the models of L2 learning reviewed here agree that we can better understand how language is learned by focusing on how speakers use language. For each model, we begin with the “theoretical basics”. This is followed by a discussion of each theory’s predictions for L2 learning. The chapter ends with a synthesis of the four models. Attention is paid to transfer and the specific role(s) general cognitive mechanisms (e.g., attention, inhibitory control) play in learning.

2.2  The Unified Competition Model The Unified Competition Model (UCM, MacWhinney, 2008, 2012) has its origins in the Competition Model (Bates & MacWhinney, 1982; MacWhinney, 1987), which was developed to account for the nature of a speaker’s language knowledge (i.e., how and why do L1 and L2 speakers produce and interpret sentences in the ways that they do?). While the UCM shares this aim, it additionally seeks to understand “the learning process” (MacWhinney, 2017). So, the UCM seeks to explain not only the nature of a speaker’s language knowledge, but also how that system of knowledge exists in the ways that it does. The UCM proposes that language development involves the piecemeal learning of cues (or constructions, form-​meaning mappings). Accounting for how speakers go about learning new cues, therefore, lies at the heart of the UCM. Cue learnability is understood to be influenced by a range of input-​based and cognitive processing factors, including features of cues (e.g., how available and/​or reliable is a cue in the input?), interactions among different cues, and cognitive processes that may influence the production and interpretation of a cue. In short, the UCM seeks to explain the routes and rates of language learning through interactions between features of the input and cognitive learning mechanisms.

2.2.1  Theoretical Basics 2.2.1.1 Cues In the UCM, cues are the building blocks of language. Cues vary across languages in terms of their type (morphological, syntactic, prosodic, semantic, and pragmatic), availability (how frequently a cue is present), reliability (how often a cue is interpreted in the same way), and validity (the combination of availability and reliability). This means that the UCM is sensitive to what the input looks like, how the input may vary across languages, and how languages structure shared meanings in describing and explaining language learning. The UCM predicts

2

22  Theoretical Models of L2 Learning

learning difficulties when languages use the same cues in different ways. We will look at each of these cue properties in turn. First, cues can be described in terms of their type or form and include morphological cues, syntactic cues, prosodic cues, semantic cues, and pragmatic cues. An example of a syntactic cue is word order in English. In active sentences, word order is used to express subject-​object information. In the sentence the cat chased the dog, the ordering of the noun phrases (“the cat” and “the dog”) provides information about the subject and the object. In other words, the ordering of these noun phrases tells us that it was the cat who did the chasing and it was the dog who was being chased. A consequence of English using word order to express subject-​object information is that switching the order of the noun phrases additionally changes the meaning of the sentence. Compare the cat chased the dog with the dog chased the cat. When we change the order of the noun phrases, we also change the meaning of the sentence. Thinking about language in this way allows us to compare languages with reference to how they go about structuring the same meaning. In German, for example, even though word order can be used to express subject-​object information some of the time, a more reliable way to express this meaning is via case, which is a morphological cue (see Kempe, 1999). Case in German can be marked on articles, adjectives, pronouns, and nouns to indicate the subject of a sentence (nominative case), the direct object (accusative), the indirect object (dative), and the possessor of something (genitive) (see Durrell, 2011). Unlike English, word order in German is not a reliable cue for subject-​object information because changing the order of German noun phrases does not necessarily change the meaning of the sentence. For example, die Katze jagte den Hund and den Hund jagte die Katze both mean that the cat chased the dog even though the order of the noun phrases is not the same. The critical information is included in die (“the-​nominative”) and den (“the-​accusative”), which tell us information about the subject (die) and the object (den). We know that the dog was being chased because of den. If we compare English and German, we can say that English uses word order to express subject-​ object information in active sentences, but German uses case. Table 2.1 presents a summary of these differences, showing how German can change the order of the noun phrases without changing the sentence’s subject-​object meaning.The UCM predicts learning difficulties when the same cues in L1 and L2 do not express the same meaning(s). In addition to type, cues can vary in other ways that have implications for L2 learning. Cue availability refers to the extent to which a specific cue is present (or available) in the input. Importantly, however, not all cues are available in the input to the same extent. We will take as an example the cues in English for expressing past habituality, an aspectual meaning that indicates the extent to which an event or action is viewed as repeated over an extended period of time (see Comrie, 1976). Critically, multiple cues exist in English for expressing past habituality, but they are not equally as available in the input. Examples of past

2

23

2

Theoretical Models of L2 Learning  23 TABLE 2.1  Summary of Subject-​Object Information in English and German Sentences

Language

Example sentence

Subject

Object

German

die Katze jagte den Hund “the-​NOM cat chase-​PAST the-​ACC dog”

the cat

the dog

English translation

the cat chased the dog

German

den Hund jagte die Katze “the-​ACC dog chase-​PAST the-​NOM cat”

the cat

the dog

English translation

the cat chased the dog

habitual cues in English include used to (e.g., she used to play football), would (e.g., she would play football), and the Simple Past (e.g., she played football every day). Even though cues such as used to, would, and the Simple Past (e.g., played) can all be used to express a regularly repeated event in the past, they are not all used to the same extent or in the same ways. Tagliamonte and Lawrence (2000) showed in a corpus of informal conversation among British English speakers that the Simple Past (e.g., played) was used almost 70% of the time to express past habituality, whereas the cues we might associate more with past habituality, used to and would, accounted for 19% and 6% of usage, respectively.Tagliamonte and Lawrence’s analysis indicates that English cues for past habituality are not equally available in the input. Information like this is helpful for thinking about why some target features might be more difficult to learn than others. Connected to this is cue reliability, which is the extent to which the same cue leads to a consistent interpretation. In our past habituality example, cue reliability is how often the Simple Past expresses past habituality. If we compare Trish played tennis everyday with Trish played tennis yesterday, the Simple Past form played expresses past habituality in the first case (i.e., a regularly repeated event) but it expresses past perfectivity in the second case (i.e., a one-​time complete event). This comparison shows us that the Simple Past is not a very reliable cue for habituality since the Simple Past can also express other meanings. Lastly, cue validity is the joint product of availability and reliability because neither availability nor reliability on its own can explain learning well. A cue with high validity means that it is both available in the input and it reliably expresses that meaning. Cues with high validity are understood to be easier to learn than cues with low validity. By comparing the Simple Past with used to, we can use our discussion of past habituality to illustrate cue validity. We noted that the Simple Past represents a case in point for thinking about L2 learning because the Simple Past was the most available cue for past habituality in Tagliamonte and Lawrence’s corpus analysis of English. However, the Simple Past is not a very reliable cue for habituality because it also expresses past perfective meaning. In contrast, used to is a more reliable cue. This creates a complex state of

24

24  Theoretical Models of L2 Learning

affairs in terms of learning habituality in English because used to is a reliable but not a very available cue, whereas the Simple Past is an available but not a very reliable cue (see McManus & Marsden, 2019a). Taken together, the UCM’s theory of cues is a helpful way to understand how features of the input can shape and help explain L2 learning. At the same time, the UCM demonstrates that focusing on frequency alone can only provide a partial explanation for some learning difficulties (see also Gries, 2019).

2.2.1.2  Competition and Transfer In addition to the properties of cues, the UCM also attributes roles to the cognitive processes of competition and transfer in explaining L2 development. A key construct that links competition and transfer is cue strength, which is a direct function of cue validity. Cue strength concerns the connections (or relationships) between the different cues in a sentence. Cue strength can also be understood as the weights that speakers assign to cues when they are processing sentences in real time. Processing is easier when multiple cues agree on the same interpretation, but processing is more difficult when cues point to different interpretations. For example, in the sentence the cow chews the grass, word order and animacy cues converge on “the cow” as the subject (i.e., it is the cow that is doing the chewing). In the sentence, the grass chews the cow, however, word order and animacy cues compete because they point to different nouns as the subject: word order favors “the grass” as the subject, but animacy favors “the cow” as the agent.This means that English speakers will likely demonstrate greater difficulty processing a sentence like “the grass chews the cow” in comparison to “the cow chews the grass” because cues converge in the latter but not in the former. We can determine the strength of a cue by pitting it against a different cue to see which one wins out. In experiments with English speakers, for example, word order has been shown to be a stronger cue than animacy for assigning subject roles (MacWhinney, 2005). In terms of transfer, the UCM proposes that L2 learners transfer cue strengths from L1 for use in L2 processing. Applied to our example of subject-​object information in English, this means that word order will be transferred because word order is the stronger cue. As a result, English-​speaking L2 learners will initially use word order as a strong cue for determining subject-​object information even if the new language does not use word order to express the same meaning (such as in German where case is a strong cue for subject-​object information). In the UCM’s theory of transfer, it is proposed that “all aspects of the first language that can possibly transfer to L2 will transfer” (MacWhinney, 1997, p. 119). We will use an example from lexical processing to illustrate. Here, it is proposed that the meaning of a new L2 lexical item will be the “full conceptual structure of the most closely corresponding L1 word” (ibid.). Transfer like this is commonly found in “false friends” or cognates when L1 and L2 use similar forms (e.g.,

24

25

24

Theoretical Models of L2 Learning  25

boot in English and das Boot “boat” in German) but associate different meanings with them. Taken together, transfer in the UCM means that the L2 learner will be influenced by their knowledge and experience in producing and interpreting L1 cues: the second language learner begins learning with a parasitic lexicon, a parasitic phonology, and a parasitic set of grammatical constructs. Over time, the second language grows out of this parasitic status and becomes a full language in its own right. Ibid. As a result, cues that are transferred from L1 but that are not helpful in L2 must be reconfigured. This means that L2 cues have to be connected with new meanings in L2 (e.g., associating a new meaning to an existing cue).

2.2.2  Risk and Support Factors Lastly, we discuss a range of risk and support factors that are understood to be unique to L2 learning. The specific claim here is that the successes of L1 learning create a range of risks in L2 learning. For each risk, a corresponding support factor is proposed to counteract the negative effects of each risk. Four risk and support factors are proposed: entrenchment-​ resonance, transfer-​ decoupling, overanalysis-​chunking, and isolation-​participation. While all four risk and support factors are relevant to L2 learning, we focus on two of these here given their relevance to understanding crosslinguistic influence: entrenchment-​resonance and transfer-​decoupling. Entrenchment is a basic neurodevelopmental process in the sense that from birth humans can process general patterns, but at birth the processing of patterns is not language specific. With experience, a speaker’s processing mechanisms become committed to the patterns of their L1.That is, extensive L1 use entrenches operations that facilitate efficient L1 use, thus allowing for automatic and efficient L1 processing (Segalowitz, 2010). The support factor of resonance is one way to counteract entrenchment by allowing new experiences and memories to become consolidated and integrated within existing networks (see also Bybee, 2010). Resonance thus permits new meanings to become associated with existing cues. Resonance counteracts entrenchment by allowing the learner to create new form-​meaning pairings. This process is made possible through consolidation, which allows multiple cues to coexist (MacWhinney, 2017). When a previously encountered memory is activated, its resonant connections become consolidated and integrated with other connections in the network. Transfer is described as a risk factor because new L2 forms are aligned with similar-​looking L1 forms. When the mappings align, this is called positive transfer,

26

26  Theoretical Models of L2 Learning

but mapping mismatches lead to negative transfer (see also Chapter 1). Transfer is described as a risk factor because learners initially use L2 cues as if they are L1 cues (e.g., when English-​speaking learners use word order to interpret subject-​object information even if the L2 does not express subject-​object in this way). Initially, the L2 system is described as a “parasitic” system. The risk factor of transfer is counteracted by decoupling, which allows L2 words to be used without mediation through L1. In this case, L2 cues become cues in their own right, rather than remaining “parasitic”. MacWhinney (2017) proposes that “to achieve decoupling, the learner needs to think and operate in L2 without switching back to L1 or relying on L1 structures” (p. 300). Ways of achieving decoupling can include “assuming an L2 identity” and developing L2 “inner speech” (ibid.), which can help learners to establish ways to think and operate in L2, thus allowing meaning representations for L2 words to emerge.

2.2.3  Predictions for L2 Learning As our discussion so far indicates, the UCM makes predictions about the potential routes and rates of L2 learning and the extent to which prior language knowledge and experience play a role. In the previous section about some of the UCM’s theoretical fundamentals, we described (i) cues, (ii) competition and transfer, and (iii) risk-​support factors as three important components of the model. In this section, we consider how these components of the UCM can lead to testable predictions for L2 learning. First, cue strength, including the competition and convergence of cues, is arguably one of the clearest predictions made by the UCM, especially given the centrality of transfer and entrenchment in the model. As noted, crosslinguistic influence creates learning difficulties when L1 and L2 use different cues to express the same meaning(s). In our discussion, we used L1–​L2 differences in English and German for subject-​object information to illustrate this learning difficulty. As a reminder, subject-​object information in active sentences is cued in English with word order but with case in German. L2 learning difficulties are predicted among English-​speaking learners of German, for example, because beginning learners transfer cue strengths from L1 to L2.This entails, therefore, that, in active sentences, the English speaker would use word order for interpreting subject and object roles rather than attending to case information marked in German noun phrases. This is predicted because the strongest cue for expressing subject-​object information in English is word order. Thus, the UCM predicts that English speakers will initially use word order information for expressing and interpreting subject-​object information. Indeed, L2 research has shown that a persistent difficulty for English-​speaking learners of German is learning that word order does not reliably express subject-​ object information in German (e.g., Henry et al., 2009; Jackson, 2008). Although our example has focused on English-​speaking learners of German, this prediction

26

27

26

Theoretical Models of L2 Learning  27

can be made for any combination of languages in which L1 and L2 use different cues to index the same meaning, such as subject-​object information in English and Spanish (Fernández, 2008). Another example is the use of verb morphology to express time information in English by Mandarin Chinese speakers. This L2 learning difficulty is predicted because Mandarin Chinese expresses time using lexical cues, whereas verbal inflectional cues are important for expressing time information in English. A similar set of findings are reported in McManus (2015) for grammatical aspect in L2 French among German-​speaking learners because German expresses viewpoint aspect information lexically and pragmatically (e.g., adverbs, lexical aspect of verbs), whereas English additionally uses verb morphology (e.g., she walked vs. she was walking). Even though the field of SLA has mostly investigated crosslinguistic influence in terms of L2 learning difficulties, we must also pay attention to the ways in which crosslinguistic influence can be beneficial. In other words, how can prior knowledge and experience facilitate new language learning? In this regard, L1-​L2 similarities have been shown to lead to fewer learning difficulties. In the UCM, a crosslinguistic similarity is when L1 and L2 use the same cues to index the same meaning(s), such as the use of article cues to predict upcoming nouns among Italian-​speaking learners of Spanish (Dussias et al., 2013) and the use of verbal morphology to express habituality and ongoingness among Spanish-​ speaking learners of French (Izquierdo & Collins, 2008). Further evidence of cue-​based predictions due to transfer are reported in MacWhinney (2017), which, overall, shows a strong tendency for learners to initially transfer all cue weights from L1 for use in L2 learning. If we turn to the support factors of resonance and decoupling as examples, the UCM predicts that instruction and extensive periods of immersion /​study abroad can counteract the risk factors of entrenchment and transfer. However, we should remember that these risk factors will be persistent because adult speakers bring a lifetime of daily L1 use to the task of L2 learning. This means that the effects of instruction and immersion on counteracting entrenchment and transfer will likely vary among individuals and contexts. Furthermore, the effectiveness of instruction and immersion will also vary according to their nature, meaning that not all types of instruction (see Chapter 4) and not all types of immersion are going to be equally as effective. Nonetheless, the UCM’s predictions for resonance and decoupling on L2 learning are important. Resonance can be supported through stimulation (or activation) of previously encountered experiences or episodes. One way to achieve this is via cue focusing, in which L2 learners engage in guided explicit learning about cue strength differences in L1 and L2. Explicit information about a particular meaning (e.g., subject-​object information) followed by the ways in which L1 and L2 express this meaning is one way to address the learning difficulty, (e.g., McManus & Marsden, 2017, 2019a; Spada et al., 2005; see Chapter 4). Cue focusing can facilitate resonance by prompting the consolidation of new experiences, including linking new

28

28  Theoretical Models of L2 Learning

and established concepts into rich conceptual networks. Of particular relevance to establishing new resonant connections is practice that is systematic and deliberate as a way to activate specific types of consolidated resonant connections (Presson et al., 2014). Work in Skill Acquisition Theory (DeKeyser, 2017), for example, has shown how systematic and deliberate practice can lead to L2 development, especially in studies investigating the longitudinal signatures of practice (e.g., Cornillie et al., 2017; DeKeyser, 1997; McManus & Marsden, 2019b). Interleaving practice can be particularly effective in this regard since it intentionally spaces out opportunities for retrieval, which can prompt greater consolidation (see Rogers, 2021; Suzuki, 2018). Taken together, instruction that guides learners to attend to crosslinguistic differences while engaging in meaning-​based practice is one way to improve L2 learning outcomes. In order to counteract the risk factor of transfer so that L2 representations can be independently established rather than remaining “parasitic” on L1, extensive immersion and/​or study abroad opportunities can potentially lead to L2 development, a finding that is broadly supported in the SLA literature (McManus et al., 2021; Mitchell et al., 2017; Sanz & Morales-​Front, 2018). MacWhinney (2019, p. 16) notes that “successful second language learning depends on the decoupling of L2 from its dependence on L1”. In other words, decoupling may be possible if the learner actively and consciously tries to separate L1 from L2. Immersion and/​ or study abroad could be one way to facilitate decoupling since these contexts might encourage the learner to “think and operate in L2 without switching back to L1” (MacWhinney, 2017, p. 300) more than classroom experiences in the home country. However, spending time abroad by itself will not guarantee that speakers begin to think and operate in L2 (see Mitchell et al., 2017). In sum, the UCM predicts that meanings that are expressed (i) with different cues in L1 and L2 or (ii) with the same cues in L1 and L2 but with different cue strengths will lead to L2 learning difficulties, due to entrenchment and transfer from prior L1 use. A series of support factors are proposed as ways to counteract these L1 effects, including resonance and decoupling. Ways to implement resonance and decoupling to counteract the effects of L1 transfer and entrenchment include cue focusing (or explicit instruction) about cue differences in L1 and L2 with deliberate and systematic meaning-​based practice (for the development of new resonant connections) and immersion experiences (for decoupling). In both cases, however, the effects of instruction and immersion are likely to vary among individuals given that the effects of transfer and entrenchment are pervasive. In Chapters 3 and 4, we review in more detail SLA research that has explored these predictions.

2.3  The Associative-​Cognitive CREED The Associative-​Cognitive CREED (Ellis, 2006a, 2006b) holds that L2 learning “is governed by the same principles of associative and cognitive learning that

28

29

28

Theoretical Models of L2 Learning  29

underpin the rest of human knowledge” (Ellis, 2006b, p. 100). In this respect, learning a language is no different to learning any other skill.The major principles of this model are that L2 learning is understood to be Construction-​ based, Rational, Exemplar-​Driven, Emergent, and Dialectic (CREED). We review each of these principles as articulated in the Associative-​Cognitive CREED, paying particular attention to L2 learning as Rational, Exemplar-​Driven, and Emergent. We then consider some of the predictions made by this model for L2 learning. We should note that in contrast to some of the other models of L2 learning discussed in this chapter, Ellis considers that the Associative-​Cognitive CREED is “too broad to constitute a theory of SLA” (Ellis, 2006b, p. 112), opting instead for the label of a framework.

2.3.1  Theoretical Basics In the Associative-​Cognitive CREED, L2 learning is construction-​based, indicating that the building blocks of language are constructions, or form-​meaning pairings (Goldberg, 2006). Constructions are similar to cues (as discussed for the UCM) in that they are symbols of form-​meaning pairings that are conventionalized in communities of speakers (see also Christiansen & Chater, 2016; Chapter 1). Constructions are represented in the mind of speakers at various levels of representation with, for example, phonological, semantic, and syntactic information. The use of language by speakers is rational in that speakers’ “mental models of the way language works” are based on experience and grounded in usage (Ellis, 2006b, p. 103). That is, a speaker’s ability to use language to perform a variety of socio-​cognitive functions is influenced by their previous experiences with language. For example, the type of language used in a particular context can be influenced by a speaker’s experience using and hearing language used in similar contexts. As a result, learners may struggle to use language in a manner that more experienced speakers expect if learners have little prior experience with language in that context. Indeed, this is something educators often experience with academic writing: learning the style and conventions of academic writing takes time and experience engaging with academic writing (Hinkel, 2015; Le Ha & Baurain, 2011). Similarly, expectations about what a speaker is going to say or do is influenced by their previous experiences observing or engaging in those encounters. This is one reason why we tend to carry out practice interviews for a job, especially when we are relatively inexperienced at participating in interviews. A speaker’s use of language is thus understood to be probability-​tuned and adaptable based on encounters with language. Even though frequency and other distributional features of the input are understood to play an important role in understanding and explaining language use and development, frequency (or exposure) alone cannot explain how speakers learn and use language (see also Divjak, 2019; Gries, 2019). This is one reason why speakers can use language in new ways and can create and interpret

30

30  Theoretical Models of L2 Learning

novel uses of language without having heard or spoken those specific utterances before. Indeed, this is an ability that is grounded in exemplar-​based abstraction and attraction. Usage-​based investigations have shown that language is highly patterned and constructed out of various slot-​and-​frame patterns, some which are more open in scope than others (Dąbrowska, 2014; Ellis et al., 2016; see Chapter 1). For example, the greeting pattern “good”+ time-​of-​day can generate patterns like good afternoon and good evening. Learning language is understood to be exemplar-​based because previous experience processing and using specific slot-​and-​frame patterns can allow learners to generalize on the amount of variation within each pattern (e.g., which parts of the pattern are variable and which seem fixed, see Bybee, 2008, 2010). Therefore, experience with specific patterns is needed for generalizations to develop. In the “good” + time-​of-​day pattern, this includes knowledge about other time of day expressions that can be used as a means of creating new greetings (e.g., morning, afternoon, day, night). Other instances include patterns such as noun + s for plurals (e.g., cat -​> cats, dog -​> dogs). The key component to development is experience. Language is also emergent. This means that the properties of a construction are not static, but are dynamic and can change to meet the demands of our processing capabilities and the communicative needs of speakers (Beckner et al., 2009). One consequence of this understanding is that linguistic representations emerge, adapt, and gradually change over time in response to usage and as shaped by cognitive processing capabilities. This point was considered in Chapter 1 in our discussion of the directions of crosslinguistic influence, noting that an adult speaker’s L1 system is not static (see also Chapter 3). Representations are subject to constant change over time. When thinking about the ways in which language representations emerge, Ellis (2006b) describes the initial state of L2 learning as a “tabula repleta” that has already been tuned and committed to the task of L1 use. L2 learning is therefore processed and learned through a speaker’s prior experience, indicating that the emergence of L2 knowledge is created through L1-​tuned and L1-​committed mechanisms (see also MacWhinney, 2005). In line with the Associative-​ Cognitive CREED’s conceptualization of L2 learning as a rational and emergent process (i.e., L2 learning is grounded in a speaker’s previous experience), transfer also plays an important role. In particular, Ellis notes that: it is the very achievements of associative learning in first language acquisition that limit the input analysis of L2 and that result in the shortcomings of SLA. Associative learning theory explains these limitations too, because associative learning in animals and humans alike is affected by learned attention. Ellis, 2006b, p. 109 In this view, transfer additionally includes learned attention or the learned behaviors built up from prior experience of processing and using another language. What

30

31

30

Theoretical Models of L2 Learning  31

this means, therefore, is that learning a new language can be more or less difficult depending on how closely the new language connects with a speaker’s previous processing experience. For example, L2 learning will likely be more difficult when prior processing experience is not helpful for processing L2 data (e.g., when the same cues in L1 and L2 express different meanings). Conversely, fewer L2 learning difficulties would be predicted when speakers can process L2 data using the same processing knowledge that was built for L1 use. Two examples focusing on prior experience and information processing will be helpful to illustrate this point. We will use the relatively simple task of crossing the road as a pedestrian. The simplicity of this task does vary, of course, depending on the location. In some cities, crossing the road might be a task that an adult carries out daily over the course of their lives.When approaching the road with traffic moving in both directions, the pedestrian will probably look in one direction for oncoming traffic (e.g., left), then in the other direction (e.g., right) for traffic coming in a different direction, and then once more (e.g., left) to be sure there is no traffic before beginning to cross the road. Over time, perhaps with the guidance of an expert (maybe a parent), this is an ability that will become less labored and more effortless, thus allowing the adult to cross the road while doing some other activity (e.g., listening to the radio, having a conversation).With increasing practice, this ability to look for traffic before crossing the road would be called learned behavior. This means that we have learned to carry out a task without thinking about it too much. It is just something that we do if we want to cross the road. This is all very well until we visit a country in which there is a slight variation in how the traffic flows (e.g., vehicles move in the opposite direction, cyclists are moving down the outside lanes). For example, for somebody who was born in the UK and spent their adult life in the UK, they will likely have developed some type of learned behavior of looking right, then left, then right once again before crossing the road. However, when that person visits a country where the traffic flows in the opposite direction (e.g., USA), that learned behavior could be problematic (even life threatening!), at least temporarily, especially where fast-​moving traffic is involved or if it is a built-​up area (e.g., a large city like New York). This new situation requires the person to start by looking left, then right, then left again. That is, the person must engage in a related but different behavior for processing the same information. For those of us who have done this, the first couple of times require a lot of conscious effort (“oh, I must remember to look left this time”). Crossing the road will take us a little longer than usual because we are double checking and approaching with more caution. In fact, you will see that some large cities like London include road markings on pedestrian crossings to tell you which way to look for oncoming traffic (e.g., “Look Both Ways”, “Look Right”). This is one example where our previous experience has led to various types of automatic, learned behaviors that have to be consciously overridden every now and then.

32

32  Theoretical Models of L2 Learning

These same mechanics apply to the processing of language just like they do for any other type of information processing. If we return to the example of the subject-​object information in English and German, the English and German sentences contain the same information about subjects and objects, but that information is encoded or located elsewhere in the sentence. Over a lifetime of prior information processing, the German speaker becomes attuned to attending to articles and their dependents to work out information about the subject.The English speaker, however, has learned that attending to the ordering of the nouns is the most informative cue. Both cues (article, word order) have high cue strength in their respective languages. The English-​speaking learner of German, therefore, has learned that word order is the cue for subject-​object information. This means that articles such as den “the” or einen “a” initially go unnoticed when processing for information about subjects and objects. This is because the English speaker’s prior processing experience has shown them that English articles are not informative cues for this type of information. This means that the English speaker must consciously override this learned behavior to process German sentences for subject-​ object information. This English-​German comparison is an example of the associative learning phenomenon of blocking (Ellis & Sagarra, 2010, 2011) because “redundant cues are overshadowed for the historical reasons that learners’ first language experience leads them to look elsewhere for the cues to interpretation” (Ellis, 2006b, p. 110). This means article cues are initially redundant to English speakers for interpreting information about subjects and objects since English articles are not informative cues for this meaning. The difficulty arises when the same cues are redundant in the L1 but informative in the L2. Taken together, the Associative-​ Cognitive CREED describes and explains the ways in which associative and cognitive learning phenomena can make L2 learning more difficult, especially when learned behaviors from the L1 have to be modified to accommodate L2 use. As in the UCM, the Associative-​Cognitive CREED makes clear that learning a language is not only about creating linguistic knowledge. An important role is also assigned to how speakers learn to use that knowledge (i.e., the development of a “performance grammar”). This is clearly something important that usage-​based models of L2 learning contribute to theory building in our field (see also Bates & MacWhinney, 1987). These models encourage us to think not only about the types of knowledge about language that speakers need to develop, but also about how speakers learn to access and assemble this knowledge for use in real-​time communication.

2.3.2  Predictions for L2 Learning A core prediction for adult L2 learning involves transfer. But, as previously noted, in addition to transfer of linguistic knowledge that has been built for L1 use, transfer also involves learning mechanisms that have been tuned and committed to

32

3

32

Theoretical Models of L2 Learning  33

processing L1 data. As a result, the Associative-​Cognitive CREED predicts transfer on at least two different levels: linguistic knowledge and learning mechanisms, both of which have been developed and fine-​tuned for comprehending the specific ways in which a speaker’s prior language experience uses language to structure information. Consequently, transfer can take multiple forms (or have multiple underlying causes), all of which are involved in L2 learning and can be difficult to separate out. For example, a particular L2 learning difficulty could be the result of (i) new cues that have not yet been learned, (ii) new processing behaviors that have not yet been learned, or (iii) both new cues and new processing behaviors have been learned but they are not yet usable during real-​time language use because of competition or other processing-​related factors. Trying to tease apart the precise cause(s) of the L2 learning difficulty is an important part of understanding L2 learning. This is one example where triangulating different task types involving different language abilities is necessary (see Chaudron, 2003; Mackey & Gass, 2016). On the one hand, L2 learning that requires the creation and development of new linguistic cues will be most difficult when L2 cues are different from those used in L1. This prediction is also made by the UCM. On the other hand, even once new cues have been learned, new processing behaviors must be created and developed so that these new cues can be used. As a result, the Associative-​ Cognitive CREED proposes that L2 learning can involve the creation of linguistic knowledge as well as the creation of new processing behaviors. We say “can involve” because what has to be learned is dependent on the nature of the learning problem. That is, then, if the processing behaviors developed from prior experience can be used in L2 learning, it seems possible (and likely) that new processing behaviors do not need to be learned (e.g., attending to verbal inflectional morphology for interpreting tense information). This, however, is a hypothesis that is yet to be more fully investigated (see Chapter 5 for further discussion). Thus, L2 learning will likely be most difficult when (i) the same cues in L1 and L2 express different meanings, and (ii) these new cues are blocked during real-​ time processing because previous L1 experience has shown these cues not to be informative (Ellis & Sagarra, 2010, 2011). As noted in Chapter 1, research in predictive L2 processing is one line of research that has contributed to this understanding, partly due to the use of offline and online tasks to determine the nature of L2 knowledge and its use in real-​time processing (see Kaan, 2014). In this line of research, even though findings show that L2 learners can acquire new cues that do not exist or are different from those in L1 (e.g., learning that articles and nouns can express grammatical gender), a contributing source to the L2 learning difficulty appears to be the extent to which these new cues can be used in real-​time processing (Dussias et al., 2013; Hopp, 2013; Kaan, 2014). One explanation for why using newly created cues in real-​time processing can be difficult is that the speaker must additionally learn new processing behaviors. This could be one reason for why learning that involves both

34

34  Theoretical Models of L2 Learning

the learning of new cues and of new processing behaviors appears to be the most difficult. Research on the use of English articles by Chinese speakers, the use of German case for interpreting subject-​object information, and Spanish articles for predicting upcoming nouns by English speakers, are examples of acquisition that involves the learning of (i) new cues and (i) new processing behaviors.This research has repeatedly shown persistent difficulties in L2 learning, but, in some cases, learners can develop the required processing behaviors. Findings suggest that the learning of new processing behaviors is probably the most difficult part of the learning puzzle, depending on the nature of what must be learned. One reason for this is that processing behaviors developed over the course of a lifetime of L1 use are automatic and optimized. In contrast, learning situations that involve the learning of new cues but not so much new processing behaviors appear to pose comparably fewer learning difficulties. One example of this is the learning of new mappings for tense and aspect (e.g., Ellis & Sagarra, 2010; McManus, 2015; Roberts & Liszka, 2013, 2020). This work has shown that although learning new form-​meaning mappings can be difficult when the specific meanings mapped to L1 and L2 forms are not the same, these new mappings are learnable. One reason for this could be that existing processing behaviors are already helpful. For example, in the case of French-​ speaking learners of English, French speakers bring already fine-​tuned processing mechanisms that attend to inflectional verbal morphology for aspectual information. In addition to learning new form-​meaning mappings, it seems, therefore, that these learners do not also need to develop new processing behaviors for this target feature.This appears to be the case for Dussias et al.’s (2013) Italian-​speaking learners of Spanish, who were able to use some articles to predict upcoming nouns. In addition, one prediction drawing on the Associative-​Cognitive CREED’s cognitive component involves consciousness in L2 learning, which includes roles for noticing, negative evidence, attending to language form, perception and attention directed by explicit instruction, and consciously guided practice (see also Spinner & Gass, 2019; Robinson, 2001; Schmidt, 1990). Learning situations that involve L1–​L2 differences in terms of (1) form-​meaning mappings and (2) processing behaviors are likely candidates that would benefit from (explicit) instruction tailored to the nature of the learning problem. There is an important emphasis on tailoring the instruction here since, as previously noted, not all types of L2 learning involve L1–​L2 differences, including at the level of form-​ meaning mappings and/​or processing behaviors. One example of this is discussed in Chapter 4, in which explicit information is used to draw learners’ attention to form-​meaning mapping differences in L1 and L2 followed by processing strategies to facilitate the development of new processing behaviors. Indeed, studies involving intensive training show that automatic processing behaviors can be developed for linguistic cues that are different in L1 and L2 (e.g., Hopp, 2014; McManus & Marsden, 2019b). Even though some research

34

35

34

Theoretical Models of L2 Learning  35

indicates that explicit instruction can lead to the creation of new form-​meaning mappings, less well understood is to what extent instruction can also help develop new processing behaviors (see Loewen, 2020), especially when blocked attention resulting from learned L1 behavior is involved. Related, it seems that the development of new processing behaviors could lead to broader changes in how speakers process information in their previously acquired languages (e.g., L1). We discuss this question in Chapter 3 where we review studies that suggest new and existing processing behaviors may converge over time. In sum, an important contribution of the Associative-​Cognitive CREED to SLA theory is its articulation about the nature of transfer in L2 learning. This includes transfer of L1 linguistic knowledge about form-​meaning mappings as well as transfer of learned processing behaviors that have been tuned and committed to facilitate automatic and efficient L1 use. An implication of this is that L2 learning difficulties resulting from crosslinguistic influence can be understood as interactions between these two effects (the creation of new cues and/​or the creation of new processing behaviors), which would likely lead to different patterns of results depending on the nature of the specific learning problem. Investigating these influences in L2 learning is also important if we are to understand questions of instructional effectiveness, including, for example, to what extent instruction should target the creation of new linguistic knowledge and/​or processing behaviors.

2.4  Processing Determinism O’Grady (2012, 2013, 2015) hypothesizes that improvement in the use of a new language is a secondary effect of the processor improving its own performance through usage. This processing-based account of L2 learning contrasts with accounts that conceptualize development as the construction and/or growth of linguistic knowledge (e.g., Rothman & Slabakova, 2018; White, 2003), which O’Grady (2013) refers to as “the illusion of language acquisition” (p. 253). In O’Grady’s view, language development conceptualized as processing improvement is emergent because the processor’s functioning is improved through experience and usage. Termed Processing Determinism (O’Grady, 2015), this proposal is shaped by two core claims about learning a new language: (1) language development is a by-​product of processing amelioration and (2) dominant processing behaviors (from L1, for instance) are transferred unless they are too costly to implement (O’Grady, 2012). These hypotheses form the backbone of our discussion of Processing Determinism.

2.4.1  Theoretical Basics In line with the other models of L2 learning reviewed in this chapter, one of the broad-​level goals of Processing Determinism (PD) is to explain how L2 speakers

36

36  Theoretical Models of L2 Learning

learn to process and produce language. In so doing, it conceptualizes a “performance grammar” (Bates & MacWhinney, 1981) with two parts: (1) some type of linguistic knowledge and (2) a processor. In PD, learners create representational knowledge of a language, in a sense similar to constructions (Ellis et al., 2016) or cues (MacWhinney, 2005). What leads to development, however, is not the construction of linguistic knowledge itself, but the ability to use this knowledge during real-​time language use. That is, then, L2 development is a consequence of processing improvement (or learning how to better use linguistic knowledge to comprehend language) and not a consequence of building linguistic representations alone. Even though the building of linguistic representations and processing improvement are related, only the latter can lead to improvement in how a speaker uses language. Already, it is possible to see connections with the Associative-​Cognitive CREED, but with a stronger emphasis on the development of processing behaviors. The driving force in PD is the Amelioration Hypothesis, which proposes that language acquisition cannot be explained as “a developmental process devoted specifically to the construction or growth of linguistic knowledge” (O’Grady, 2013, p. 254). Instead, language acquisition is conceptualized as “a side-​effect of processing amelioration” (ibid.). In PD, improvement in the comprehension and production of sentences by speakers is, in fact, the language processor getting better at its job, and not the building and storing of new linguistic representations alone. At the core of processing amelioration is an efficiency-​driven processor or processing system (see O’Grady, 2013). It is efficiency-​driven in the sense that the processor continually seeks to improve its own performance. This is because processing exerts demands on a speaker’s cognitive resources. Processing is most efficient when its operating costs can be reduced to the lowest possible minimum. The processor approaches this task in two ways: (1) freeing up working memory as efficiently as possible and (2) creating processing routines from frequently encountered and useful form-​meaning mappings and sequences of them. Working memory resources are freed up by incrementally building and processing sentences, one word at a time. When hearing a sentence like sheep eat grass, for instance, English speakers interpret the word “sheep” as soon as they hear the first word (sheep) and assign “sheep” as the agent or subject of the sentence as soon as the active verb is heard (eat). Words are understood to be dismissed from active working memory as soon as they are interpreted. Incremental processing frees up processing resources since not all words and interpretations are held in working memory together but are processed in a linear fashion and then dismissed. In addition, processing resources are freed up by creating routines from frequently encountered form-​meaning mappings. This is because the mapping procedures involved in processing are costly. For example, in the sentence she played football, the processor first assigns an interpretation to she, then accesses the meaning of the verb played, and then assigns to she the agent argument, then interprets the word to the right of played and identifies it as the patient (O’Grady,

36

37

36

Theoretical Models of L2 Learning  37

2013). The processor frees up processing resources by creating routines of regularly encountered operations. This means that instead of repeatedly carrying out the same item-​by-​item processing of words, a routine assigns an interpretation to a string of words (much in the sense of a construction, see Ellis et al., 2016). For example, this involves the creation of a processing routine from the multi-​ word structure noun-​verb-​noun in which the interpretation would be agent-​ verb-​patient (see also MacWhinney, 2008). The creation of processing routines ameliorates the functioning of the processor, “improving its speed and efficiency as [routines] are gradually strengthened and automatized” (O’Grady, 2012, p. 118). Routine strengthening occurs through repeated opportunities for processing (or practice): the processing of specific combinations of individual form-​meaning pairings becomes less demanding and more automatic as a function of usage or practice (see also DeKeyser, 2017; Segalowitz, 2010). A second component of PD for explaining L2 learning is the transfer calculus, a hypothesis that “L2 learners transfer dominant processing routines, unless the cost of implementing those routines is less favorable in the second language than in the first language” (O’Grady, 2013, p. 271). This means that automatized routines are transferred for use in L2 processing unless the linguistic properties of the L2 make using L1 routines too costly to implement. L1-​L2 differences in sentence structure is one reason why L1 processing routines might be more costly in L2. If, for example, the L1–​L2 difference requires backtracking (i.e., an initial interpretation must be revised), it is hypothesized that transfer is “blocked” (i.e., no transfer, see O’Grady, 2013). This is because backtracking is a costly procedure and is dispreferred by the processor given its design to process information incrementally. One instance of this is scopal interpretations involving negation and universal quantifiers (e.g., not + all/​every, as in Mike did not eat all the cookies) in English and Korean (see Lee, 2009; O’Grady, 2013, 2015). Findings have shown that learners do not transfer routines that require initial interpretations to be revised. Findings like these are used to support the claim that “cost blocks transfer” (O’Grady, 2015, p. 275). Another example of this learning difficulty, well attested in the L2 literature, includes interpretations of word order in English and Spanish (e.g., Fernández, 2008;VanPatten & Borst, 2012). As previously discussed, the dominant routine in English is to use word order as a cue for subject-​object information (see discussion of UCM). However, this routine would only be successful some of the time in Spanish given that Spanish uses other cues to express subject-​object information (e.g., subject-​verb agreement, animacy, direct object placement, case marking “a”). Transferring the English word order routine to Spanish would require backtracking, thus making transfer inefficient for processing. This is because the noun-​ verb-​ noun word order does not reliably map onto subject-​verb-​ object information in Spanish (similarly to our previous discussion of word order in German). In addition, variability in the input means that sometimes the English

38

38  Theoretical Models of L2 Learning

word order routine would lead to the appropriate interpretation (e.g., with subject-​verb-​object sentences, as in la niña besa al niño “the girl kisses the boy”) but sometimes the English routine would lead to the incorrect interpretation when the order of the words does not reflect subject-​object roles (e.g., with object-​verb-​ subject sentences, as in lo besa la niña “the girl kisses him”). Taken together, PD proposes that L2 learning is best understood in terms of processing amelioration, that is, in terms of the processor improving its performance by creating routines. Routines are created to make processing more efficient. Dominant and/​or automatic routines are available for transfer unless they are too costly to implement. A key interpretation of “costly” in PD is when backtracking is required (i.e., the processor revises a previous interpretation). In such cases, transfer is blocked. These principles have made an important contribution to how we understand the emergence of the speaker’s L2 system. In the next section, we consider some predictions of PD for L2 learning.

2.4.2  Predictions for L2 Learning As previously noted, two principles central to PD are the amelioration hypothesis and the transfer calculus. We consider each of these in thinking through the types of predictions PD makes for L2 learning. First, a core claim of the amelioration hypothesis is that processing resources are freed up by creating routines.This is because routines are proposed to be less costly on a speaker’s processing system than computing a series of individual operations for each item/​word encountered in a sentence string (O’Grady, 2013). Experience is a key determinant in creation of a processing routine. That is, the more often a speaker attends to a particular combination of form-​meaning mappings, the more amenable that combination of form-​meaning mappings is to routinization. Thus, one prediction of PD for L2 learning is that deliberate and systematic practice that rehearses the same processing operations would promote routinization and facilitate L2 development. Experimental intervention studies within Skill Acquisition Theory, for example, are one line of research that can inform this hypothesis (see DeKeyser, 1997, 2017). This body of research has investigated the ways in which the accuracy, speed, and stability of L2 performance can change as a function of practice (or usage). In addition, this work has subsequently interpreted signatures of L2 performance (e.g., faster and more stable processing over time) in terms of knowledge creation/​ restructuring processes (see Solovyeva & DeKeyser (2018) for lexical learning; McManus (2021) and McManus & Marsden (2019b) for grammatical learning). In these studies, learners completed systematic meaning-​oriented training over an extended period of time to practice interpreting and/​or producing a particular target feature (e.g., repeatedly attending to verbal inflections and connecting a specific inflection to a specific aspectual meaning). Taken together, PD predicts that engaging learners in systematic and deliberate meaning-​based practice involving

38

39

38

Theoretical Models of L2 Learning  39

specific target features over an extended period of time can facilitate L2 learning by supporting the formation and strengthening of processing routines. Second, as stated in the transfer calculus, PD hypothesizes that L1 processing routines are transferred only if they are not more costly to implement in L2. This is because when a word has been assigned an interpretation in the processing stream (e.g., first noun + verb = subject + verb) it is too costly to revise the initial interpretation. In cases like this, the transferred processing routine that led to this interpretation is abandoned. L1 processing routines exhibiting L1–​L2 differences would likely be “blocked regardless of the routine’s dominance in the former language” (O’Grady, 2013, p. 274). As a result, the transfer calculus predicts that L1 routines that induce revision/​ backtracking in L2 should not transfer. If we take the previously discussed English–​Spanish word order example, this would indicate that the English word-​order routine for subject-​object information would not transfer to Spanish. This is because the L1 routine would interpret the first nominal as agent on encountering the active verb (e.g., Jackie chased…). In a language with more flexible word orders like Spanish, however, the first nominal could be either subject or object. For example, in the sentences la niña besa al niño “the girl kisses the boy” and lo besa la niña “the girl kisses him”, the subject/​agent in both sentences is la niña (“the girl”), but this occurs to left of the verb in the first example and to the right of the verb in the second example. Because of this L1–​L2 difference, the L1 English routine for subject-​object interpretation should be blocked (i.e., no transfer). Instead, the learner is claimed to build a new processing routine from scratch as a function of processing L2 input, which is likely a much slower process. Under this hypothesis, English-​speaking learners of Spanish would not be expected to perform better with subject-​verb-​object word orders than object-​verb-​subject word orders, at least initially, unless some statistical bias was present in the input (e.g., that subject-​verb-​object word orders are more frequent than object-​verb-​subject word orders). This is because both routines would be built from experience processing L2 data rather than transferred from L1. Related to this prediction, it seems that some role should also be attributed to variability in the input (and Spanish word orders, in our example). This is because sometimes the L1 routine would lead to the correct interpretation and other times it would not. This variability would likely make it difficult for the processor to abandon the L1 routine. As a result, it seems likely that L1 routines might initially transfer until the processor builds up enough evidence to block its use. Such a case would lead to more accurate performance on subject-​verb-​object word order, at least initially, for example.This, however, is an empirical question that requires further exploration and empirical validation. Taken together, PD provides a series of clear and testable hypotheses for L2 learning, which, to date, have been subject to less testing than those in the UCM, for example.This is one reason why our PD-​informed predictions are a little more tentative than those discussed for the UCM.

40

40  Theoretical Models of L2 Learning

2.5  Inhibitory Control Model Green’s (1998) Inhibitory Control model seeks to explain how speakers use and manage their languages, including the ways in which competition between languages is managed and resolved and the extent to which inhibitory control might facilitate language selection (see also Abutalebi & Green, 2007; Calabria et al., 2018; Green & Abutalebi, 2013). inhibitory control is one cognitive mechanism that allows speakers to manage their languages. Although initially proposed to account for lexical processing, the Inhibitory Control Model (ICM) has been applied to other aspects of L2 use, including grammatical learning (McManus, 2021) and phonological learning (Darcy et al., 2016), as discussed in Chapters 3 and 4. In this section, we review the theoretical basics of the ICM and the ways that it can help us to better understand crosslinguistic influence in L2 learning.

2.5.1  Theoretical Basics The ICM responded to early debate in the field about the ways in which humans manage tasks, including interference among tasks that compete for cognitive processing resources (Monsell, 1996). Applied to multilingualism, task management includes use of the intended language when desired. As Green notes, the ICM seeks to address the following question:“How do individuals ever manage to avoid producing a word in L1 when they wish to produce its translation equivalent in L2?” (Green, 1998, p.67). If we add some context to this question, we could also think of it like this: when an English speaker is in a bakery in Germany ordering bread, how does the speaker avoid saying the word bread when they want to say Brot? This is a relevant question for SLA theory given that a speaker’s language representations are understood to be closely networked (for review, see Shirai, 2019). It is likely that L1 and L2 representations for /​bread/​would be activated together and compete for selection, with the dominant or stronger representation winning out (see also MacWhinney, 2008). Up until relatively recently, discussion had focused on the organization of a speaker’s language system, including the nature of knowledge representations in L1 and L2 (e.g., Ervin & Osgood, 1954; Kroll & Stewart, 1994; Votaw, 1992). In comparison, little discussion at that time had focused on how speakers manage their language knowledge, including the extent to which selection of the unintended, often dominant, language can be prevented so that the desired language can be used (see Paradis, 1981). The ICM sought to address this gap. Green (1998) hypothesized that the ability to select the appropriate (or desired) language (i.e., a processing routine, cue, or construction) involved domain-​general inhibitory control processes that inhibited dominant or competing responses (Miyake et al., 2000; Stahl et al., 2014). By “inhibited”, Green proposed that certain types of language knowledge can be suppressed via cognitive mechanisms

40

41

40

Theoretical Models of L2 Learning  41

grouped under the general label of inhibitory control. They are described as “domain-​general mechanisms” because they are not dedicated to the use of language only. Inhibitory control is a core feature of information processing which can stop (suddenly and completely) or prevent a planned or ongoing thought and action (Logan, 1994). For inhibitory control to function as a means of language selection among multilinguals, language representations are hypothesized to be associated with language tags. This means that L1 representations are associated with L1 tags and L2 representations have L2 tags, for example. Tag specification is hypothesized to be part of the representation itself, indicating that the knowledge creation process involves tag specification. In addition to tag specification, Green proposes that tag specification activates connected representations with the same tags. For example, activation of L1-​tagged representations stimulates other representations with the L1 tags. Tag specification is the means and locus of language selection. While tag specification is part of language selection, tag selection alone is claimed to be insufficient for language selection. The ICM proposes that language selection is made possible through suppression of “incorrect tags”. This entails that, for example, if the speaker intends to use the L2, this is made possible by suppressing L1 tags. Inhibitory control is thought to be applied once representations have already been activated, however. For example, when representations in both L1 and L2 are activated, tag suppression prevents selection of the unintended language. Tag suppression is needed since representations in L1 and L2 are connected at multiple levels of representation. In this architecture, therefore, inhibitory control is reactive in the sense that it is applied once a specific language representation has been activated, either via external stimulation (i.e., the input) or internal stimulation (i.e., from a speaker’s conceptual/​ thought system). In sum, language selection is a reactive process that uses inhibition, a domain-​ general cognitive process, and tag specification. Language representations are associated with language tags (e.g., L1 tags, L2 tags). Upon activation, the unintended language tag is suppressed to allow the desired language to be selected. For speakers who regularly use both languages (e.g., in multilingual contexts), inhibitory control is understood to be more easily applied and released than among inexperienced L2 users. In language-​switching tasks that require learners to regularly switch between languages (see Chapters 3 and 4), inexperienced L2 learners have been shown to struggle to release previously applied inhibition of the dominant language, as evidenced by larger switch costs when switching from L2 into L1 but less so when switching from L1 into L2 (see Costa & Santesteban, 2004; Meuter & Allport, 1999). This is because L2 use requires inhibition of the dominant L1, but then switching into L1 requires this previously applied inhibition of L1 to be released or reversed. Experience with language switching is understood to be an important factor in regulating inhibitory control and therefore enhancing multilingual language use.

42

42  Theoretical Models of L2 Learning

Some L2 research has suggested that difficulty suppressing L1 tags is one explanation for crosslinguistic influence in L2 learning (Darcy et al., 2016; McManus, 2021). For example, in the preceding section about PD we considered transfer of L1 routines to interpret word order in Spanish. Selection of the inappropriate processing routine would result in negative influence, such as the English speaker interpreting the first noun phrase as the agent in Spanish object-​verb-​subject sentences. In this scenario, difficulty suppressing the dominant L1 routine leads to selection of the L1 routine over the L2 routine when both are activated, resulting in negative influence.This is why some research suggests that difficulties managing language selection can lead to negative crosslinguistic influence in L2 learning. At the same time, this performance could also indicate that there is a problem with the tag specification process (i.e., difficulties associating language knowledge with tags) as well as a problem with developing selection mechanisms that operate on language tags. Indeed, L2 studies that have traced learners’ performance over time have shown that increasing amounts of language switching can reduce the negative effects of crosslinguistic influence and therefore improve L2 performance (see Kang et al., 2018; McManus, 2021; Wu et al., 2018). One interpretation of this body of evidence is that deliberate and systematic practice that requires learners to apply and release inhibitory control might be one way to reduce the negative effects of crosslinguistic influence caused by language selection difficulties. Taken together, the ICM provides an account of language use that attributes primary importance to language selection and to the cognitive mechanisms involved in that process. It does so by proposing that language representations are tagged. Inhibition works by reactively suppressing tags associated with the undesired language. This is required since representations in L1 and L2 can be activated simultaneously. Without some means to suppress the dominant representation, the dominant representations (typically L1 among L2 learners) would be regularly selected for use over weaker representations. This is understood to be particularly relevant for unbalanced bilinguals (or L2 learners) whose L1 representations are likely to be more easily activated and selected due to prior experience using the L1 (see also Ellis, 2006a, 2006b). Tag suppression allows the more dominant language to be inhibited when necessary, thereby facilitating L2 use. Evidence of negative influence from L1 is thought to be a result of difficulties associated with tag suppression, thereby allowing the unintended language to be selected.

2.5.2  Predictions for L2 Learning In terms of using the ICM to make predictions for L2 learning and use, we should keep in mind that the model was proposed to explain the ways in which speakers select the appropriate language when needed (i.e., language control). As such, the ICM does not explain how speakers learn to select among their languages. That said, difficulties associated with language selection (i.e., selecting the

42

43

42

Theoretical Models of L2 Learning  43

inappropriate processing routine, for example) have been interpreted as one cause of crosslinguistic influence in SLA research (see Darcy et al., 2016; McManus, 2021; Tokowicz, 2014). As previously noted, the ICM makes two assumptions about the ways in which a speaker’s knowledge of multiple languages is organized. First, the ICM assumes that L2 learning leads to the development of a new/​additional body of knowledge and that different bodies of knowledge are identified with tags (e.g., “L1 tags” for L1 knowledge, “L2 tags” for L2 knowledge). Second, the ICM proposes that learners select language representations through reactive suppression of the unintended language using inhibitory control processes and tag selection. In line with other models of L2 learning reviewed in this chapter, one interpretation of the ICM’s architecture is that establishing new knowledge is a necessary condition, but establishing linguistic knowledge alone is not sufficient. In addition to the creation of L2 knowledge, inhibitory control mechanisms are required to reactively operate on language tags to allow for language selection. Therefore, the ICM predicts a relationship between inhibitory control and L2 proficiency. For example, high levels of L2 proficiency (akin to demonstrating automatic L2 knowledge, for instance) would be related with expert performance on an inhibitory control task (e.g., Simon task, Stroop task). In contrast, low levels of L2 proficiency would be related with low scores on an inhibitory control task (see also Chapter 3). These relationships are hypothesized because proficient/​expert L2 use involves reactive suppression of L1 knowledge using inhibitory control. Importantly, knowledge in L1 and L2 should be determined beforehand because absent L2 knowledge could influence the nature of the findings. Without doing so would make it difficult to determine whether the underlying issue is related to absent L2 knowledge and/​or inappropriate language selection mechanisms. Because inhibitory control has been less actively investigated in L2 research compared to other cognitive processes (e.g., working memory, see Juffs & Harrington, 2011; Wen & Li, 2019), we will briefly consider some of the types of tasks that have been used in this line of research (see also Chapter 3). In general, inhibitory control is understood to be a cognitive process that facilitates task management and interference. The Stroop task (Stroop, 1935) is one example of this type of task in which participants see a word on the computer screen written in a color and their task is to name the color of the word. A simplified visualization of this task design is presented in Figure 2.1. The Stroop task contains a series of conditions: a match condition when the color of the text and the word match (e.g., the word “BLACK” in black ink); a

BLACK FIGURE 2.1  Simplified Visualization

GLOVE

WHITE

of Stimuli in a Stroop Task

4

44  Theoretical Models of L2 Learning

mismatch condition when the color and the word do not match (e.g., the word “WHITE” in black ink); and a control condition when the color and the word are not related (e.g., the word “GLOVE” in black ink). The most taxing condition is when the color and the word do not match (the word “WHITE” in blank ink) because processing this information requires two competing responses to be managed (do you name the color or do you name the word?). As a result, speakers are understood to perform more slowly and less accurately in mismatched conditions when compared to matched conditions. Other tasks thought to tap into this same cognitive processing ability include the ABX categorization task (Gottfried, 1984), the Simon task (Simon, 1969), and the Flanker task (Eriksen & Eriksen, 1974). One additional prediction of the ICM involves language switching, which has shown to be a productive line of research in understanding inhibitory control and L2 use. Here, the prediction is that high levels of inhibitory control allow for more efficient language switching. Switching from L1 into L2 is claimed to require the suppression of L1 knowledge so that the relatively weaker L2 knowledge can be selected.Then, switching back into L1 requires the previously applied inhibition to be reversed. This process is understood to be taxing on cognitive resources, especially among inexperienced learners and/​or learners with underdeveloped language control mechanisms. As a result, L2 speakers with weaker and/​ or underdeveloped inhibitory control mechanisms are expected to show larger switch costs when switching between languages, whereas speakers with highly developed inhibitory control mechanisms will demonstrate smaller switch costs. Here, a switch cost tends to be interpreted as the amount of time it takes a speaker to switch between languages (e.g., Costa & Santesteban, 2004). Smaller switch costs are thought to indicate higher levels of inhibitory control, thus allowing the speaker to switch with fewer processing costs. Taken together, the ICM makes relevant predictions about the cognitive processes involved in L2 use and the extent to which an understanding of these can help understand crosslinguistic influence. Of most relevance is the claim that difficulties managing languages can lead to negative crosslinguistic influence, indicating that the development of effective language selection mechanisms could be one way to reduce the negative effects of crosslinguistic influence (see also Chapters 3 and 4).

2.6  Synthesis of Theoretical Concepts So far in this chapter, we have reviewed four models of L2 learning and use that attribute critical roles to prior knowledge and experience in explaining the organization, structure, and development of a speaker’s L2 abilities. In this section, our aim is to review some of the ways in which these models appear to agree as well as where they begin to diverge from one another. We focus on three topics that

4

45

4

Theoretical Models of L2 Learning  45

have featured prominently in this chapter: language knowledge and experience, inhibition, and transfer.

2.6.1  Language Knowledge and Experience The label “language knowledge” is arguably one of the most multifaceted terms in L2 research. In this chapter, we have used the term “knowledge” to refer the linguistic representations that learners build and store in memory (Bybee, 2008, 2010). L2 learners construct new representations by working with the language input they are exposed to. Importantly, however, building and storing language representations is just one part of the language learning process.We also need to be able to explain how speakers make use of these representations to carry out a range of socio-​cognitive functions in real time. Usage-​based accounts of L2 learning therefore seek to describe and explain both (i) how language is represented and stored in speakers’ minds and (ii) how these representations about language are used and handled in real-​time to comprehend and produce messages (Bates & MacWhinney, 1981). One consideration that comes to bear on this issue is that even though the biological machinery used to construct language knowledge is understood to be the same among humans (Bates & MacWhinney, 1981), not all speakers are exposed to the same types of language (Dąbrowska, 2012, 2013, 2014). As a result, an individual’s experience with language shapes their knowledge and use of language in different ways. This is one reason why we see variation in how speakers use language to carry out the same communicative functions. Indeed, these ideas are fronted in almost all the models discussed in this chapter because the learning of language is influenced by experience and driven by usage. This means that experience using a specific language influences how a new language is learned and used as well as how previously learned languages are used. We review some of the evidence to date on this issue in Chapters 3 and 4 and consider the extent to which learning a new language influences how speakers use their L1 (e.g., Bice & Kroll, 2015; Dussias & Sagarra, 2007). Together, the evidence suggests that the cumulative experience of attending to language in new ways can shape a speaker’s existing attentional and processing behaviors. These combined experiences influence how we attend to and interpret all types of information, thus developing greater consensus that crosslinguistic influence is multidirectional (Pavlenko & Jarvis, 2002). A speaker’s system of language knowledge is therefore sensitive to different types of exposure and usage and no type of knowledge is privileged. Our use of any language is therefore experience-​based and shaped by our cumulative experience using language to communicate. In addition, the ICM can help explain why the contexts in which speakers find themselves can play an important role in the development of specific language abilities, including the ease with which multilinguals can move (or switch)

46

46  Theoretical Models of L2 Learning

between languages (Abutalebi & Green, 2007; Green, 1998; Green & Abutalebi, 2013). For example, contexts that involve daily use of multiple languages have the potential to develop language selection mechanisms more efficiently than contexts in which the use of a specific language is more predictable or is biased toward a particular language (e.g., the foreign language classroom). When the contexts in which the L2 is used are predictable, language selection abilities take longer to develop. Experience can also explain why some aspects of a new language appear more difficult to learn than others. In the UCM, close attention is paid to the characteristics of cues that can influence their learnability. For example, high validity cues are easier to learn when compared with low validity cues. This is one explanation for why English speakers struggle with word order and subject-​object information in languages like German and Spanish. As previously discussed, word order can be used to interpret subject-​object information some but not all of the time in German and Spanish, indicating that there is a partial overlap in the cues used to express subject-​object information in English, German, and Spanish. Variability in the input is known to be an obstacle in L2 learning when the prior experience speakers bring to learning a new language is only partially helpful. Taken together, the experience that speakers bring to the task of learning a new language plays an important role in L2 development. This experience drives the processing and use of new languages, just as our experience with other tasks influences how we carry out new tasks (e.g., think about how the layout of your previous car influences your assumptions about where the light switch is in your new car). The types of language we are exposed to and the contexts in which we use language shape our cognition and language use in important ways as well. Our experience feeds into a multi-​component system, shaping how we use and process language, influencing L1 as much as L2 use.

2.6.2 Inhibition We have seen at least two claims about the ways in which language representations are made available and/​or are selected for use. These claims raise questions about the cognitive mechanisms that allow a speaker to utter L2 words (instead of L1 words) when they intend to do so. Accounting for this ability is particularly relevant given that learning an additional language is hypothesized to involve the creation of a new body of language knowledge. A speaker’s processing system must therefore develop mechanisms to select between L1 and L2 representations. Although both the ICM and the UCM propose that activation is an important cognitive process that facilitates language selection, the ICM additionally proposes that inhibitory control is needed given that activation occurs throughout the system. The idea that activation plays a major role in language selection is well represented in the literature (see Grosjean, 1988, 1997; Paradis, 1993). The basic driving force behind activation is that each language representation has its own

46

47

46

Theoretical Models of L2 Learning  47

level of activation (or resting level). Selection of a representation elevates its level of activation. As a result, a representation that is used/​activated more frequently can lead to a higher level of activation for that representation. One consequence of this is that items with higher levels of activation are easier to select during real-​ time processing. How an item becomes activated is understood to be a result of stimulation, either externally or internally. Given that the ICM proposes that inhibition is required for language selection, but the UCM does not, it is helpful to review why these differences exist. If we begin with the nature of language representations, representations are tagged for language in the ICM (e.g., L1 tags for L1 knowledge). Language tags can serve as a source of activation as well as inhibition. When an item with an L2 tag is activated by an external stimulus, it co-​activates other items with L2 tags in addition to items with L1 tags. This means that L1 and L2 items can be activated by the same stimulus if they share properties. Inhibition works by suppressing items with incorrect tags. So, if an external stimulus activates a specific L2 representation, inhibitory control can be applied to suppress any L1 tagged items that are simultaneously activated.This process is understood to allow selection of L2 items. In short, language processing is facilitated by suppressing representations of the unintended language (i.e., suppressing L1 tags of representations that have become co-​activated during L2 processing). In contrast, the UCM proposes that speakers manage multilingual processing through activation and resonance, rather than inhibition (MacWhinney, 2008, 2017). One reason for this is that inhibitory control is a reactive process that suppresses representations that have already been activated. In the UCM, the process of activation followed by inhibition is thought to be too costly to result in learning. Instead, language selection is facilitated by activation (and decoupling). Activation works in a similar way as discussed for the ICM, except that representations are not tagged for language. Instead, representations are hypothesized to be organized in language-​specific networks (see also Shirai, 2019), which are weighted based on prior processing experience and compete for selection based on weightings. When a specific representation is activated (via an external stimulus, for example), it activates potential candidates within the same language-​specific network based on grammatical roles (MacWhinney, 1987). Taken together, activation is an important process in understanding language selection in multilingual processing. Inhibitory control also appears to play a role in suppressing the unintended language, although the extent to which a model includes a role for inhibition depends on the architecture of the language system. When L1 and L2 networks coexist but are independent, there seems to be less need for inhibitory control, as suggested in the UCM. However, when L1 and L2 representations are hypothesized to be tightly integrated into a single network, inhibition arguably plays a larger role in preventing selection of the non-​target language, especially when stimulation co-​activates L1 and L2 items. Whether or not we argue that inhibition is needed to allow for language selection seems to be connected to how language knowledge is organized, namely in language-​specific

48

48  Theoretical Models of L2 Learning

networks or a single network. In Chapter 3, we consider some findings from L2 research in lexical processing that have investigated L1 co-​activation during L2 processing to better understand the ways in which L1 and L2 systems might be connected.

2.6.3 Transfer The models of L2 learning discussed in this chapter forefront critical roles for prior knowledge and experience in explaining L2 learning. In this section we reflect on some of these claims about transfer in order to better understand contemporary views on the role of transfer in L2 learning, paying particular attention to the UCM and PD. There are many commonalities about transfer as discussed in the UCM and PD. Both models conceptualize of transfer as a learning process that can lead to the creation of a new body of language knowledge (e.g., the L2). For example, MacWhinney proposes that cues and their weightings are transferred from L1, meaning that initial L2 cue weight settings are close to those as used in L1 processing, but “over time, these settings change in the direction of the native speakers’ settings for L2” (MacWhinney, 1997, p. 129).Transfer is therefore a process for creating new knowledge that can be modified, if required, through exposure to new input. PD also draws on transfer in the creation of new knowledge (see transfer calculus), albeit more selectively because processing routines that are more costly to implement in L2 are blocked (i.e., do not transfer).Thus, in both models, transfer of L1 knowledge plays a fundamental role in the construction of L2 knowledge. A consequence of this view of transfer is that it leads to the same knowledge existing in multiple places (or in multiple systems). In PD, for example, if L1 and L2 use the same processing routine and this is not more costly to implement in L2 than in L1 (e.g., noun-​verb-​noun = subject-​verb-​object, as in English and French), transfer leads to the same routine existing in multiple places (e.g., L1 system and L2 system). Having the same knowledge represented in multiple places is counter-​intuitive to efficiency-​driven processing. A similar architecture is proposed for the UCM given that “all aspects of the first language that can possibly transfer to L2 will transfer” (MacWhinney, 1997, p. 119). What is interesting to note, therefore, is that by attributing a key role to transfer as a way to create new language knowledge, this process also leads to the duplication of language knowledge. While this is not necessarily problematic on its own, a relevant question for SLA theory is why this type of architecture is needed. In other words, can we explain L2 learning without transfer? One reason for this architecture could be due to a tendency in the field to think about L1 and L2 as separate bodies of knowledge, thus requiring the same processing routine or cue weightings to exist in multiple places. Bernolet & Hartsuiker (2018) propose that while separate knowledge sources might exist initially, these may actually become integrated over time when the information represented in L1

48

49

48

Theoretical Models of L2 Learning  49

and L2 knowledge sources is the same and/​or similar (e.g., same processing routine in L1 and L2). This seems to be one way in which L2 processing could be seen as efficiency-​driven, by eventually reducing unnecessary duplication (see also Dussias & Sagarra, 2007). Such a view encourages us to viewing learning as a gradual move toward an integrated language system, rather than multiple systems that are connected. Alternatively, as discussed in Chapter 1, L2 learning might involve the creation of new processing routines only.This could entail that L1 knowledge is not transferred, but L2 performance that appears L1-​like is more simply the use of the L1 routine rather than transfer of L1 knowledge (McManus, 2021). Taken together, transfer has played an important role in theories of L2 learning in trying to explain and understand the ways in which L2 learners might build knowledge of a new language. A dominant approach to understanding this process has been to argue that transfer is a critical process in the creation of L2 knowledge. Under this view, the transfer process is thought to involve making a clone or copy of a speaker’s L1 knowledge, which can then be adapted based on new data, potentially leading to the newly cloned processing routine being amended or abandoned (see also Chapter 1). Over time, some proposals have considered that while L1 and L2 knowledge structures might be separate, they may eventually merge. This proposal could explain why extensive L2 exposure may affect L1 performance. It seems that L2 research should also consider if transfer is needed to explain L2 learning at all, especially in its traditional interpretation of a copy-​ and-​restructure process.

2.7  Conclusion This chapter has focused on L2 models about the learning and use an additional language. Four models were reviewed: The Unified Competition Model, the Associative-​ Cognitive CREED, Processing Determinism, and the Inhibitory Control Model. In addition to reviewing the theoretical basics of each model, predictions for L2 learning were also considered. Following our review, three topics that connected with these theories were selected for further discussion and reflection (language knowledge and experience, inhibition, and transfer). Each discussion section topicalized these areas to prepare the reader for Chapters 3 and 4.The information discussed in this chapter will be developed and carried forward in the subsequent chapters, especially in understanding how L2 research can explain learning by paying attention to the roles played by experience, inhibition, and transfer.

Note 1 In this chapter and throughout the book we use the label “model” (in place of “theory” or “framework”) when referring to these theoretical accounts (for discussion of these labels in the field of SLA research, see Gass et al., 2020; Jordan, 2004; VanPatten et al., 2020).

50

3 STUDIES OF L2 DEVELOPMENT

3.1  Introduction We reviewed in Chapter 2 four models of L2 learning that make important claims about the ways in which prior linguistic knowledge and experience can help us to describe and explain the routes and rates of L2 development. In so doing, we reviewed some of the main ideas and concepts used in theoretical accounts of crosslinguistic influence and L2 learning. In this chapter and in Chapter 4 we add to these theoretical accounts by focusing on empirical studies of L2 learning that have investigated roles for prior linguistic knowledge and experience in L2 development. This empirical evidence is included in two chapters to address two different but connected questions: (i) what do we know about crosslinguistic effects in L2 development (this chapter)? and (ii) can instruction informed by this evidence support L2 learning by reducing the negative effects of crosslinguistic influence (Chapter 4)? In terms of the organization of these chapters, we have selected representative studies from three areas of language: morphosyntax, vocabulary, and phonology. Cross-​referencing is provided throughout the chapter to connected lines of research.

3.2  L2 Learning and Morphosyntax In this section, we review studies that have investigated the acquisition of morphosyntax. This review includes studies involving complete beginners (Ellis & Sagarra, 2010, 2011) as well as more advanced learners (McManus, 2015). In addition, the review includes study designs that compare learners from different L1 backgrounds (e.g., Roberts & Liszka, 2013) as well as studies that have examined how L1 performance changes as a function of L2 learning (Dussias & Sagarra, DOI: 10.4324/9780429341663-3

50

51

50

Studies of L2 Development  51

2007).Together, these studies contribute different perspectives that advance knowledge and understanding about the effects of crosslinguistic influence in language learning.

3.2.1  Learned Attention and Blocking In a series of studies designed to investigate the effects of learned attention and blocking in initial L2 learning (see Chapter 2), Ellis and Sagarra (2010, 2011; Ellis et al., 2014) investigated the extent to which prior knowledge of one cue (e.g., adverbs) blocked acquisition of other cues (e.g., verbs) when both expressed the same temporal meaning. These studies examined, therefore, whether some aspects of L2 grammar learning that have been well-​documented to be more difficult to learn than others (e.g., verbal morphology) can be explained by blocking and learned attention. The findings provided empirical support for many of the ideas discussed in Chapter 2 with regard to prior knowledge and experience in L2 learning (Ellis, 2006a, 2006b). While this line of research initially examined blocking effects among English-​speaking learners of Latin, subsequent studies involving Chinese, Russian, and Spanish speakers provided further support for blocking effects in L2 learning (see Ellis & Sagarra, 2011). We review this line of research given its importance for understanding crosslinguistic influence in L2 learning. An important point of interest in this research was understanding the extent to which a lifetime of prior language use influenced L2 learning. For example: L1-derived knowledge that there are reliable lexical cues for temporal reference (words like yesterday, gestern, hier, ayer) might block the acquisition of verb tense morphology from analysis of utterances such as yesterday I walked or hier nous sommes allés au cinéma “yesterday we went to the movies”. (Ellis & Sagarra, 2010, p. 556) In other words, if a speaker had associated a specific cue with a specific meaning, in what ways might this experience and knowledge influence new language learning? This question was investigated by providing English speakers with short-​ term training of Latin followed by receptive and production testing. Critically, speakers received one of the following types of training: verb training only, adverb training only, no training.The question of interest was whether the initial learning of verb cues blocked attention to adverb cues and vice versa. That is, then, does associating one sets of cues with one outcome block the learning of subsequent cues expressing the same outcome? In Ellis and Sagarra (2010), English speakers with no prior knowledge of Latin were assigned to one of the three groups (verb pretraining, adverb pretraining, no pretraining). In the adverb pretraining condition, learners saw on screen either hodie (“today”) or heri (“yesterday”) and had to select the meaning of the adverb

52

52  Studies of L2 Development

from two options, “today” or “yesterday”. In the verb pretraining condition, learners saw on screen either cogito “I think” or cogitavi “I thought” and had to select the meaning of the verb from two options,“I think” or “I thought”. Learners received feedback in both conditions.When the response was correct, learners saw “correct” on screen.When the response was incorrect, learners received corrective feedback, such “wrong –​the meaning of hodie is today” (ibid., p. 558).The purpose of this training phase was to develop knowledge of a specific type of cue (e.g., hodie/​cogito) and a specific meaning (e.g., today/​I think). No training was received by learners in the no pretraining condition. Phase two involved sentence decoding. Learners read two-​word sentences that combined an adverb and a verb (e.g., hodie cogito “today I think”; heri cogitavi “yesterday I thought”). Participants had to select whether the sentence referred to “the present, the past, or the future”. Previously unseen future time adverbs (cras “tomorrow”) and verbs (cogitabo “I will think”) were included as a control to see how learners performed on cues when no type of pretraining was given at all. Feedback was provided for incorrect answers. Phases three and four included receptive and production testing, respectively. The reception training provided participants with two-​ word sentences (e.g., cogitavi heri) and asked them to rate the temporal reference of the sentence on a five-​point scale: 1 = past, 2 = between past and present, 3 = present, 4 = between present and future, 5 = future. No feedback was provided. The production testing phase provided participants with the English translations of the Latin phrases (e.g., “yesterday”, “I thought”) and asked them to type in the Latin equivalents. No feedback was provided. The results showed that type of pretraining was important. Learners who were pretrained on adverbs paid more attention to adverbs than verbs and vice versa in the verb pretraining condition. Learners in the no pretraining condition attended to adverbs and verbs to an equal extent.This patterning of results also played out for the future verb and adverb conditions even though no group received pretraining on cues expressing future time. While some of these results were confirmed in a subsequent close replication, the blocking effect in the verb pretraining condition was not replicated (Ellis et al., 2014). That is, pretraining on verbs did not block subsequent attention to adverbs. Although the replication was not able to explain why this between-​study difference occurred for the verb pretraining condition, the authors speculated that maybe the backgrounds of the learners recruited in the two studies were different. In a connected line of research, Ellis and Sagarra (2010, 2011) included learners from three other L1 backgrounds, Chinese, Russian, and Spanish, all of whom completed the no pretraining condition (i.e., no pretraining on verbs or adverbs). Adding in participants from these language backgrounds allowed the researchers to understand how speakers with potentially different attentional biases due to prior knowledge and experience performed when no pretraining was included. For example, would the speakers of these language all attend to the same cues

52

53

52

Studies of L2 Development  53

irrespective of language background, or would their prior knowledge and experience with particular languages lead them to perform differently? Without any type of pretraining, Chinese speakers showed a strong tendency to attend to adverb cues for interpreting temporal reference of two-​sentence strings (Ellis & Sagarra, 2010). Although a similar tendency was found for Russian and Spanish speakers (Ellis & Sagarra, 2011), this finding is difficult to interpret because the 2011 study included a larger number of verb cues than used in the 2010 study with Chinese speakers. This means that the Russian and Spanish speakers had to learn a larger number of verb cues than the Chinese speakers (for discussion, see McManus, 2020). The greater number of verb cues relative to adverb cues could be one explanation for why the Russian and Spanish speakers performed better on the adverb cues (i.e., there was less to learn). Nonetheless, putting this between-​study difference aside, we can see that all groups attended to adverb cues more than verb cues. Perhaps most interestingly, this effect was most pronounced among Chinese speakers. In addition, Russian and Spanish speakers appeared to show more sensitivity to verb cues than the Chinese and English speakers. In terms of why all groups appeared to attend to adverb cues more than verb cues without pretraining, Ellis and Sagarra (2011) suggest that this may be a consequence of the cue dimensions of adverbs (see also Ellis et al., 2014), because of “the relative salience, simplicity, and reliability of adverbial cues compared to verbal inflections and the fact that adult language learners have prior knowledge of the use of adverbial temporal references from their L1” (Ellis & Sagarra, 2011, pp. 615–​16). In terms of why Russian and Spanish speakers were more sensitive to verbal cues, this is likely because they are experienced users of L1s that have rich inflectional morphology, thus making them potentially more sensitive to verbal morphology. Taken together, this line of research indicates that prior experience can play an important role in determining the routes of initial L2 learning. Even though the initial effects of blocking following verb pretraining did not appear to be borne out in subsequent analyses (Ellis et al., 2014), the effects for adverb pretraining were strong. This line of work has also indicated that learners from a range of different L1 backgrounds appear to attend to adverb cues over verb cues. However, further replication work is needed to balance the number of adverb and verb cues. This is needed so that we can be clear that adverb effects among the Russian and Spanish speakers was not due to a small number of adverb cues relative to verb cues.

3.2.2  Remapping Meaning to Form As noted in Chapter 1, SLA research has investigated the extent to which learners can acquire L2 forms that express different meanings (or are used differently) to those used in L1 (e.g., articles in English and Spanish). For example, McManus (2013, 2015) investigated how English-​and German-​ speakers expressed and

54

54  Studies of L2 Development

interpreted aspectual information in L2 French and whether L2 performance varied as a function of L1 background and/​or L2 proficiency. These languages were studied because each one expresses viewpoint aspect (or grammatical aspect) in different ways (see Comrie, 1976; Smith, 1997). Other lines of research investigating aspectual morphology to address similar questions have included studies of verb morphology in L2 Japanese by speakers of English, German, and Russian (e.g., Sugaya & Shirai, 2007) as well as L2 French by speakers of Spanish and English (e.g., Izquierdo & Collins, 2008). Taken together, this body of work has documented how similarities and differences in how L1 and L2 structure the same meaning(s) can influence the routes and rates of L2 learning. Before looking in more detail at this research, we will take a moment to describe some of the form-​meaning mapping differences between English, French, and German for viewpoint aspect (Smith, 1997). Viewpoint aspect is a semantic category that expresses how speakers present or view events in time. A past perfective viewpoint presents an event as complete (e.g., she played squash), an ongoing viewpoint presents an event as in progress (e.g., she was playing squash), and a habitual viewpoint presents an event as regularly repeated (e.g., she used to play squash every day or she played squash every day). A viewpoint aspect meaning such as ongoingness, for instance, can be expressed with verb morphology in English and French (elle mangeait “she was eating”), but German expresses this meaning using lexical (e.g., verb telicity and other lexical cues) and pragmatic means (see Bohnemeyer & Swift, 2004). Thus, as the previous English examples illustrate, English can use verbal morphology to distinguish among some (but not all) viewpoint aspect meanings, such as between past ongoingness and past perfectivity (e.g., I was eating vs. I ate). French can similarly use a grammatical contrast on the verb to contrast past ongoingness from past perfectivity (e.g., je mangeais “I was eating” vs. j’ai mangé “I ate”). In German, however, the verbal morphology differences between ich habe Fußball gespielt (“I played football”) and ich spielte Fußball (“I played football”) with the Perfekt and the Preterit, respectively, do not express different viewpoint aspect meanings (see Bohnemeyer & Swift, 2004). Furthermore, even though English and French use verbal morphology to express viewpoint aspect, the individual meanings they map to specific verb forms is critically different (Ayoun, 2013; Kihlstedt, 1998; Labeau, 2005). French maps the past imperfective meanings of ongoingness and habituality to Imparfait verb morphology (e.g., elle jouait au foot “she was playing /​used to play football”), but English splits these meanings up by mapping them to different forms: ongoingness to the past progressive (e.g., yesterday I was playing football after school) and habituality to the Simple Past (e.g., every morning I played football after school) and other forms (e.g., every morning I used to play football after school, every morning I would play football after school). Therefore, a relevant distinction between English and French is that French distinguishes between perfective and habitual meaning using verb morphology, whereas English can use the same form to express both meanings (McManus, 2015). Although used to and would can be used to express past habituality, these

54

5

54

Studies of L2 Development  55 TABLE 3.1  Summary of Viewpoint Aspect Differences for Past Perfectivity and

Habituality in English, French, and German Viewpoint

Language

Example sentence

Aspect form

Perfective

French

Jack a joué au foot

Passé Composé

English

Jack played football

Simple Past

German

Jack spielte Fußball Jack hat Fußball gespielt

Preterit Perfekt

French

Jack jouait au foot (chaque jeudi)

Imparfait

English

Jack played football (every Thursday) Jack used to play football (every Thursday) Jack would play football (every Thursday)

Simple Past Used to Would

German

Jack spielte Fußball (jeden Donnerstag) Jack hat Fußball gespielt (jeden Donnerstag)

Preterit Perfekt

Habitual

verb forms have been shown to be less frequently used than the Simple Past (Tagliamonte & Lawrence, 2000). These differences are summarized in Table 3.1. Given these differences between English, French, and German for expressing viewpoint aspect, McManus (2013, 2015) investigated the extent to which new and different form-​meaning mappings could be learned and the extent to which L2 learning was influenced by the nature of the L1–​L2 similarity/​difference (see also Izquierdo & Collins, 2008; Sugaya & Shirai, 2007). Given the form-​meaning mapping differences between these languages, the learning challenge for English-​ and German-​speaking learners of French is not the same. English-​speaking learners of French must modify an existing system (reconfigure the meanings mapped to specific aspect forms), whereas German-​speaking learners of French must learn a new system in which viewpoint aspect is expressed using grammatical rather than lexico-​pragmatic means. It was predicted that modifying an existing system would be more difficult than learning a new way to express viewpoint aspect. The study design compared English-​and German-​speaking learners of French at two different levels of L2 proficiency. Performance on a French c-​test was used as an independent measure of proficiency. A low group included learners in Year 1 of a university degree program in French and a high group was made up of learners in Year 4 of a university program in French. In addition to the c-​test, learners completed a sentence interpretation task and two picture-​based spoken narratives. The interpretation task assessed learners’ comprehension of L2 sentences, while the spoken narratives assessed learners’ oral production abilities.The picture-​based narrative described in McManus (2015), for instance, used images of two sisters talking about a vacation from 2006. In the sequence of images, learners saw the trip the sisters took as well as some reflections made by the sisters about what their lives used to be like when they were younger.This sequencing contrasted past perfective events (the vacation in 2006) with past habitual events (reflections on their

56

56  Studies of L2 Development

childhood) that required use of different tense forms in French, Passé Composé for past perfective events and Imparfait for past habitual events. Analyses of these data showed important differences between the groups in terms of L1 background and L2 proficiency. We focus our attention here on McManus (2015) but see also McManus (2013) for a similar patterning of results that additionally included comprehension data. First, in terms of the use of aspect morphology to refer to past perfective events, the findings indicated more target-​like performance among learners in the high group than learners in the low group, as evidenced by accurate and appropriate use of the Passé Composé among high group learners to describe past perfective events. No meaningful differences were found among English and German speakers in the high groups. In the low groups, German-​speaking learners’ performance was more target-​like than the English-​speaking learners. These results suggested that L1 background influenced performance among low proficiency learners but not high proficiency learners. Second, for past habitual events, the results showed meaningful differences between all groups, indicating that L2 performance appeared to be influenced by both L2 proficiency and L1 background.Although performance in the high groups was the most target-​like, German speakers’ performance was more accurate than the English speakers’ performance. In the low groups, however, English speakers performed more accurately than the German speakers. Further analyses showed that performance in the German low group was quite similar in perfective and habitual contexts because these learners tended to use the same verb form irrespective of viewpoint aspect context, whereas learners in the English low group tended to use a different tense form in habitual and perfective contexts. Taken together, these results indicated important differences among learners that appeared explainable by L1 background. Importantly, German speakers in the low group were not using verbal morphology to contrast past perfective from past habitual viewpoint aspect meaning, but high proficiency German speakers did. In contrast, English speakers at both proficiency levels did use verbal morphology to contrast the different viewpoint aspect meanings. Thus, these results suggest an influence of L1 background in two ways. First, a positive effect of crosslinguistic influence was found among English speakers in the low proficiency group because they used verbal morphology to contrast viewpoint aspect meanings in L2 French, leading to more target-​like use of verbal morphology. Second, a negative effect of crosslinguistic influence was found among German speakers in the low group because this group of speakers did not use verbal morphology to contrast viewpoint aspect meanings in L2 French, leading to patterns of verbal morphology use that were more L1-​like than target-​like. One explanation for this behavior is that German does not use verbal morphology to distinguish between perfective and habitual meanings. The observation that low proficiency German speakers performed in a way that was L1-​like suggests use of L1 form-​meaning mappings in L2 learning.

56

57

56

Studies of L2 Development  57

In terms of explaining differences among high-​ g roup learners, McManus (2015) suggests that German speakers outperformed English speakers potentially because learning a new way to express viewpoint aspect was potentially easier than modifying an existing system. Given that English also uses verbal morphology to express viewpoint aspect, it may have been more difficult for these speakers to notice that the specific viewpoint meanings expressed by specific forms was not the same in L1 and L2.What this learning trajectory of acquiring target-​like form-​ meaning mappings looks like, however, is unavailable from this study since only cross-​sectional data were collected. It therefore remains unknown to what extent the routes or trajectories of L2 learning were similar for all learners.

3.2.3  Comparing L2 Performance Online and Offline Understanding the types of knowledge created during L2 learning and its availability for use in real-​time performance constitutes a major source of theorization in the field (see Chapters 1 and 2). This is because early debates in SLA suggested that declarative (or explicit) knowledge of language was useful during reflective and untimed activities only, thus potentially making declarative knowledge of language less useful for more spontaneous and/​or unplanned language use (for discussion, see Roehr-​Brackin, 2018). One approach to understanding this area of research has examined L2 speakers’ performance in tasks designed to assess different types of language knowledge, such as tasks with and without time pressure and with a primary focus on meaning (see R. Ellis et al., 2009; Rebuschat, 2013). Two studies conducted by Roberts and Liszka (2013, 2020) represent one example of this research. In these studies, the authors compared L2 performance in an online task (a self-​paced reading task) and in an offline task (an acceptability task) among learners of different L1 backgrounds (for examples of other L2 research studies incorporating online and offline tasks, see Grüter et al., 2012; Ionin et al., 2021; Roberts et al., 2008). A central question driving this research is as follows: to what extent can learners develop automatic L2 abilities for target features that express different meanings in L1 and L2? Roberts and Liszka (2013) examined the extent and nature of crosslinguistic influence among French-​and German-​speaking learners of English. In that study, learners were tested to understand how sensitive they were to temporal-​aspectual violations between an adverb and the main verb in sentences. This was achieved using an experimental design that included “match” and “mismatch” conditions. The temporal features of the adverbs (e.g., last week, since last week) and verbs (e.g., went, has gone) were compatible in the matched conditions and incompatible in the mismatched conditions. Example stimuli in the match and mismatch conditions are shown in Table 3.2 (the critical items are indicated with bold type face for adverbs and underlining for verbs). Using this experimental design, Roberts and Liszka (2013) tested the following hypothesis: if learners had “fully acquired the semantics underlying the

58

58  Studies of L2 Development TABLE 3.2  Example Stimuli in Match and Mismatch Conditions from Roberts and

Liszka (2013) Condition

Example stimulus

Match

Last week, James went swimming every day. Now he’s getting bored of it. Since last week, James has gone swimming every day. Now he’s getting bored of it.

Mismatch

Since last week, James went swimming every day. Now he’s getting bored of it. Last year, James has gone swimming every day. Now he’s getting bored of it.

morphological marking of tense and aspect, then they should be sensitive to the mismatch between a fronted temporal adverbial and the tensed clause that follows in both off-​and on-​line comprehension” (p. 420).This question is of interest given the L1–​L2 pairings under investigation (as discussed in the preceding section). As such, the learning task facing French-​and German-​speaking learners of English is different. French speakers must reconfigure existing form-​meaning mappings for viewpoint aspect and German speakers must learn to use verbal morphology to express this meaning. In addition, Roberts and Liszka (2013) investigated whether learners who demonstrated offline knowledge of tense-​aspect in English could also use this knowledge in online tasks. Previous research has tended to show that offline L2 knowledge of a target feature may not always be available for use in online tasks (for review, see R. Ellis et al., 2009). To test this claim, two data collection tools were used, an offline acceptability judgement task and an online self-​paced reading task. In the offline judgement task, learners “were asked to read each sentence and then to assess its acceptability on a scale from 1 (least acceptable) to 6 (most acceptable)” (Roberts & Liszka, 2013, p. 421; for a review of judgment tasks in L2 research, see Spinner & Gass, 2019). In the self-​paced reading task, each sentence was presented one word at a time on a computer screen. A push-​button box was used to bring up the subsequent word in each sentence. Each new word replaced the previous word on the computer screen (for review and discussion of self-​paced reading tasks in SLA research, see Marsden et al., 2018). The difference between the tasks is that the acceptability judgement task was used to determine if learners were aware of or could detect any problems or issues with the sentences, whereas the self-​paced reading task was used to determine whether learners’ reading times would slow down due to any problems or issues in the sentences. Because the self-​paced reading task presented one word at a time, it is possible for us to better understand where in the sentence learners may have perceived a problem or issue. For example, if learners had perceived a mismatch between the adverb and the verb (e.g., last week James has gone swimming…), a reaction time slowdown may

58

59

58

Studies of L2 Development  59

be detectable on or shortly following the processing of the verb phrase (Jegerski, 2014). In contrast, since learners assign an overall score for a sentence in the acceptability task, it can be difficult to known why or what aspect of the sentence contributed to the acceptability rating.Therefore, combining evidence from online and offline tests can provide different types of information about L2 processing. Findings showed that performance in the offline judgement task was similar among the groups. Participants rated as more acceptable sentences that contained an adverb-​verb match (e.g., last week James went swimming every day), while they rated as less acceptable sentences that contained an adverb-​verb mismatch (e.g., since last week James went swimming every day).This evidence suggested that participants were able to detect conflict in the tense-​aspect meanings of adverbs, verbs, and their congruence in English (see also Roberts & Liszka, 2020). In contrast, results from the self-​paced reading test showed differences between the French and German speakers. Reading times among the French speakers were slower in mismatch conditions relative to match conditions. No comparable evidence of a processing slowdown was found among the German speakers. This means that the French speakers slowed down after reading a verb that was not congruent in tense-​aspect meaning with the previously seen adverb (e.g., as in since last week James went swimming every day), but the German speakers did not slow down. Slowing down at or shortly after a mismatch like this is understood to indicate sensitivity to some type of violation, whereas no change in reading time suggests a lack of sensitivity to the violation (Jegerski, 2014). This general pattern of results was also found in Roberts and Liszka (2020). Overall, Roberts and Liszka’s findings show that incorporating different types of tasks can help us to better understand potential relationships between type of knowledge and L2 performance, especially when the tasks assess real-​time language use. If the study had included the offline judgement task only, the study’s conclusions would have suggested no meaningful differences among the German and French speakers. While one interpretation for this patterning of results is that tense-​ aspect knowledge of adverb-​ verb combinations was represented declaratively only among the German speakers, thus potentially explaining why no reading slowdowns were found in the mismatched conditions in the self-​paced reading task, the prior experience that these learners bring to the task of L2 learning is also relevant. Prior L1 experience processing adverb and verb combinations for aspectual congruency would have sensitized French speakers to potential violations between an adverb and the verb for tense and aspect, whereas German speakers bring no such experience to the L2 learning task.This is because German verb morphology does not express viewpoint aspect (see Bohnemeyer & Swift, 2004). The previous experience that the French speakers had in attending to adverb-​verb matches likely played an important role in shaping their processing of adverb-​verb combinations in L2 English. These results also reinforce the earlier conclusion that the nature of the L2 learning task can depend on the specific knowledge and experience that speakers

60

60  Studies of L2 Development

have accumulated. We can make sense of these different types of experience in terms of positive and negative transfer. Positive transfer can facilitate L2 learning when the same property is expressed similarly in L1 and L2, but negative transfer (i.e., when the same property is expressed differently in L1 and L2) can lead to L2 difficulties. Indeed, the observation that aspect is grammaticalized in English and French but not in German appears to be an additional explanation for these findings.This means that crosslinguistic influence would have increased L2 learning difficulty among the German-​speaking learners, but it would have facilitated L2 learning among the French-​speaking learners.

3.2.4  Converging Representations in L1 and L2 In Chapters 1 and 2, we noted that a dominant approach to understanding crosslinguistic influence in the field has primarily involved assessing how L1 knowledge and experience influences L2 learning. Studying this issue has examined how differences and similarities between L1 and L2 for specific form-​meaning mappings (e.g., viewpoint aspect) can influence L2 learning. Research exploring other less investigated directions of crosslinguistic influence has added to what we know about this topic by examining the extent to which L2 learning might also influence L1 performance (Kroll et al., 2018; Liu & Cao, 2016). This body of research is driven by conceptualizations of the multilingual mind as a network of multi-​directional, cross-​language connections, rather than “two monolinguals in one person” (Grosjean, 1989, p. 3). In this view, often referred to as “the wholistic view”, the multilingual mind is not a collection of independent language systems, but a highly connected network of knowledge representations. Research investigating changes to L1 during L2 learning thus seeks to better understand the ways in which L1–​L2 connections emerge and are maintained (Hernandez et al., 2005). Although these questions are less often studied in SLA research, they are relevant for theories that seek to explain the ways in which a speaker’s knowledge and use of a new language emerges over time and with usage. One reason why L2 learning might influence how a speaker uses their L1 is that a speaker’s system of language knowledge most likely adapts as new language knowledge is created and stored (see Chapter 2). One study that has contributed to this line of research is Dussias and Sagarra (2007), who compared Spanish-​ speaking learners of English with different amounts of L2 exposure/​experience. A limited English exposure group included Spanish-​speaking learners of English who were studying English while living in Spain, and an extensive exposure group included Spanish-​speaking learners of English who were living in the United States. Overall, the learners in the United States had spent longer in an English environment than the learners in Spain (7.1 years compared with 8.5 months, respectively). The main difference, therefore, is that the two groups differed with respect to the type and amount of English exposure. In addition to the type of English immersion experience, learners were asked to self-​report their proficiency

60

61

60

Studies of L2 Development  61

in Spanish and English, which indicated very high levels of Spanish language ability in both groups (≥ 9.2 on a 10-​point scale for reading, listening, speaking, and writing). In terms of English language ability, self-​ratings for proficiency were also relatively high (≥ 7.4 on a 10-​point scale), but scores were higher in the extensive group than in the limited exposure group. Spanish speakers living in the United States self-​reported higher English language abilities than the Spanish speakers living in Spain. In order to investigate the extent to which L2 experience influenced learners’ L1 performance, participants read Spanish (L1) sentences that included a complex noun phrase followed by a relative clause, such as: Un ladrón armado le disparó a la hermana del actor que estaba en el balcón (“An armed robber shot the sister of the actor who was on the balcony”, from Dussias & Sagarra, 2007). This target feature was selected because it is temporarily ambiguous in both languages (i.e., was the sister or the actor on the balcony?). However, previous research has shown that English speakers show preference for attachment to the NP2 (the actor was on the balcony), while Spanish speakers show preference for attachment to the NP1 (the sister was on the balcony). Nouns and modifiers in Spanish, but not in English, must agree in gender and number, which can be used as a cue to favor one type of interpretation over another. For example, in the sentence el policía arrestó al hermano de la niñera que estaba enferma desde hacía tiempo (“the police arrested the brother of the (female) babysitter who had been ill (fem) for a while”, from Dussias & Sagarra, 2007), NP2 attachment is favored. This is because the complex NP includes two nouns that are each marked for a different grammatical gender, brother-​MASCULINE and baby-​sitter-​FEMININE, but the relative clause attaches to the feminine noun because it too is marked for feminine gender (enferma). If the complex NP included two feminine nouns (e.g., the mother of the girl), then Spanish speakers would likely opt for NP1 (mother) over NP2 (girl) (Cuetos & Mitchell, 1988; Dussias, 2003). The aim of this study was to investigate whether increasing amounts of L2 exposure also led to changes in L1 attachment preferences. To assess L2 learners’ attachment preferences, Dussias and Sagarra (2007) used eye-​tracking to document the processing strategies of the different groups (for a review of eye-​tracking methodology in L2 research, see Godfroid, 2019). Analyses focused on the critical region, defined as the adjective within the relative clause given its role in helping to facilitate attachment toward NP1 or NP2 in the complex NP. Overall, the results showed that learners with limited exposure to English favored NP1 attachment (like Spanish L1 speakers), whereas learners with extensive exposure to English favored NP2 attachment (like English L1 speakers). In other words, Spanish speakers’ attachment preferences changed with greater English immersion experience (see also Dussias, 2001, 2003; Dussias et al., 2017). Given that all participants were born in Spain and had learned English in adulthood, the participants in this study likely had well-​established and efficient L1 processing behaviors, as also indicated by their high self-​ratings of Spanish

62

62  Studies of L2 Development

proficiency (although proficiency data other than self-​reports would be helpful for determining this claim). Despite this experience, however, exposure to a new language appeared to have influenced speakers’ well-​tuned L1 system, suggesting “permeability of the first language system” (Dussias & Sagarra, 2007, p. 112). The authors suggest that this extensive immersion experience led to the “convergence” of L1 and L2 knowledge sources: “NP1 attachment […] was replaced by NP2 attachment” (ibid., p. 212). This interpretation suggests that in initial L2 learning separate language systems may emerge, but, over time and with increasing L2 use, a speaker’s different L1 and L2 systems may converge into a single language system (see also Bernolet et al., 2007; Hwang et al., 2018), even when L1–​L2 differences exist. Taken together, this research suggests that L2 learning can influence a speaker’s L1 system, including convergence of L1 and L2 knowledge sources and changes to the representations and/​or the ways in which language representations are accessed and/​or selected. Because these claims are based on cross-​sectional data collected at a single point in time, future research should investigate what the developmental trajectories and learning processes are that led to these types of behaviors. One way to explore these trajectories would be to use longitudinal designs involving assessments of L1 and L2 processing at regular intervals over an extended period of time. Given the relevance of findings like this for how we think about the emergence and development of L2 knowledge as well as the nature of connections among L1 and L2 representations, we further explore these questions in subsequent chapter sections about lexical and phonological learning.

3.3  L2 Learning and Vocabulary Just like research in L2 grammar learning, studies of vocabulary learning have made important contributions to how we think about and theorize L2 learning (e.g., Nation & Webb, 2011; Schmitt, 2010; Schmitt & Schmitt, 2020). In this section we review studies that have examined crosslinguistic influence in lexical L2 learning. A well-​known contribution to understanding in this area has involved studies of path and motion because not all languages express the same meanings in the same ways. For example, languages like English express the semantics of path and motion using combinations of verbs and particles (e.g., go into the bank), while languages like French fuse motion and path together in verbs (e.g., entrer dans la banque “enter the bank”, see Cadierno, 2004, 2012). Other work in this section reviews studies that have examined the extent to which L1 representations might be activated during L2 processing (Jiang et al., 2020) and how L2 learning can influence L1 use (Kroll et al., 2002), as well as the extent to which inhibitory control and L2 lexical learning are related (Linck et al., 2012). Taken together, the studies reviewed in this section provide important insights into the nature of crosslinguistic influence in L2 lexical learning.

62

63

62

Studies of L2 Development  63

3.3.1  Remapping Meaning in L2 As discussed in Chapters 1 and 2, an important consideration in SLA theories is how to account for L2 development. In Chapter 2, we saw a range of models of L2 learning that forefront key roles for different cognitive mechanisms in explaining development (e.g., transfer, competition). One unifying feature of these models is that input is needed for learning to take place, which is indeed a major component of almost all theories of L2 learning (for review, see Gass, 2017). In the UCM and the Associative-​Cognitive CREED, for example, features of the input can make certain aspects of the L2 easier to learn than others (i.e., the type, availability, and reliability of cues). Larrañaga et al. (2012) addressed this question by studying motion events in L2 Spanish. In this way, they investigated how English-​ speaking learners’ expression of motion was influenced by features of the input. This question is particularly relevant for understanding crosslinguistic influence since motion events in English and Spanish are expressed differently (Slobin, 2004; Talmy, 1985, 2000a, 2000b). This means that even though languages can express the same types of information about motion, differences exist in how specific languages package the same semantic information. Indeed, this is not so different from our previous discussion of viewpoint aspect in English, French, and German. This is because English and Spanish differ in how they package information about motion into combinations of verbs and prepositions (see Cadierno, 2004, 2012). Briefly, a motion event is understood to include four elements: the Figure, the Ground, path (the direction of movement), and motion (the type of motion). For instance: • •

Iñaki-​FIGURE walked-​MOTION+MANNER to-​PATH the shop-​GROUND Iñaki-​FIGURE subió-​MOTION+PATH corriendo-​MANNER a la azotea-​GROUND “ Iñaki ran up to the roof ”

These English and Spanish examples (from Larrañaga et al., 2012) show how English and Spanish can differ in how they express motion events. Languages like English are called Satellite-​framed languages because they map path onto prepositions (or satellites) that co-​occur with the verb (e.g., go + into or go + down). In contrast, languages like Spanish are called Verb-​framed languages because they use verbs to express path (e.g., entrar “go in” or salir “come out”). In short, Satellite-​and Verb-​framed languages differ in how they map the semantics of path to forms: path is mapped to verbs in Verb-​framed languages (e.g., Spanish) but to satellites in Satellite-​framed languages (e.g., English). Understanding whether L2 speakers can repackage how path is expressed in a new language constitutes a major domain of research in lexical semantics (for reviews, see Cadierno, 2010, 2017). In Larrañaga et al. (2012), this question was addressed by investigating how English-​speaking learners with different amounts of L2 Spanish learning experience expressed information about path (i.e., a

64

64  Studies of L2 Development

boundary crossing). It was hypothesized that greater amounts of experience and exposure to Spanish would lead English speakers to repackage how they express the semantics of path in L2 Spanish, thus demonstrating remapping of lexical semantic information from satellites to verbs. Evidence for this would include the use of verbs to express path information. It was also hypothesized that the existence of Latinate path verbs in English (e.g., enter, exit) may facilitate the L2 learning of expressing path information in Spanish. Thus, features of the L2 input and prior experience with similar verbs in their L1 were proposed as potential explanations for L2 development. The study included English-​speaking learners of L2 Spanish at three different levels of university-​level instruction, levels one, two, and three. Level three students had additionally completed a short stay abroad in a Spanish-​speaking country. Participants completed a c-​test as an index of general L2 proficiency. The main data collection instrument was a picture-​based narrative of a bank robber. Of particular interest was one scene from the picture-​based narrative in which the robber crosses a boundary by running into a bank. Prior to narrating the story in Spanish, learners completed a “warming up session” in which they introduced themselves and counted up to 30 in Spanish. The results showed large amounts of similarity among all levels in how they expressed path information in L2 Spanish, indicating no meaningful between-​ group differences, despite differences in the amount of Spanish learning experience. Even though all groups used some path verbs (e.g., entrar “enter”, venir “come”), these were produced the least frequently among the most proficient/​ experienced group of L2 learners. As a result, no evidence was found that showed more Spanish-​like production of motion events as a function of language exposure and/​or proficiency. Overall, all groups showed frequent use of verbs to express manner and motion, as in their L1. Taken together, the authors concluded that these results showed that learners “transfer the lexical information from the English verb onto the Spanish equivalent” (ibid., p. 133). If we extrapolate beyond the specifics of this study, one question that remains is in what ways can this line of research inform understanding about crosslinguistic influence? Larrañaga et al.’s (2012) results indicated persistent L1 influence, even for learners who had spent some time abroad in a Spanish-​speaking country. The authors report analyses from a general corpus of Spanish (Davies, 2002) showing path verbs (e.g., entrar “enter”) to be relatively common in Spanish, likely suggesting that learners would have been exposed to instances of path expressed with verbs in both classroom and study abroad contexts. This hypothesis, however, is an empirical question that future research should address. Even though general corpora are useful for understanding general patterns in usage, corpus-​based studies of the classroom input are needed to better understand the input that learners are exposed to in these specific contexts so that claims about exposure and L2 learning can be substantiated (see Collins et al., 2012; Huensch, 2019).

64

65

64

Studies of L2 Development  65

The authors additionally suggest that differences between English and Spanish in how path is expressed may be quite subtle (or even too subtle) to be easily noticed, thus making path a feature of Spanish that might take longer for English-​ speakers to learn. Of course, the ease or difficulty of learning path in L2 will be dependent on the experience that learners bring to the task of L2 learning. Thus, learning to express path in L2 would not be predicted to be universally difficult. Given the subtlety of this English-​Spanish difference, a lack of sensitivity or awareness on the part of the learners for how English and Spanish express the semantics of path could explain these findings. Exploring ways to draw learners’ attention to this L1–​L2 difference would be a useful area of exploration in future research, especially given that noticing/​awareness can be a critical learning support in instructed contexts (see Leow, 2015). In addition, both Spanish and English can use verb and preposition combinations to express these meanings (e.g., correr a la tienda “run to the store”), and English can express the semantics of path using both verb-​framed and satellite-​framed patterns (e.g., enter the store, walk into the store). Factors like this suggest that some degree of optionality may prevent learners from quickly noticing that path tends to be encoded verbally in Spanish (see also Chapter 2). As a result, exposure to instances of path verbs alone might be insufficient for learning. It is possible that encouraging learners to notice these English-​Spanish differences could be helpful to better support L2 development (Vanek, 2020; see also Chapter 4). In the next chapter, we will review examples of instruction that have addressed learning difficulties like this by drawing learners’ attention to subtle crosslinguistic differences.

3.3.2  L1 Frequency Effects in L2 Lexical Processing Taking a somewhat different approach to documenting and theorizing crosslinguistic influence in L2 learning, one line of research has investigated cross-​ language activation as a way to understand the potential nature and extent of cross-​language relationships in the multilingual mind (see Chapter 2; for reviews, see Bobb & Kroll, 2018; Palma & Titone, 2020). In this line of inquiry, researchers have investigated the extent to which a speaker’s knowledge of other languages might be activated or stimulated when a different language is being used (Kroll et al., 2012). Indeed, as discussed in Chapter 2, there is good reason to expect such co-​activation since many models of L2 learning propose extensive transfer of L1 knowledge in L2 learning, suggesting some type of relationship among a speaker’s known languages (e.g., MacWhinney, 2005; O’Grady, 2015). In lexical processing, this question has been investigated by studying how speakers process words that include some type of translation equivalence or shared property in L1 and L2. For example, to investigate the extent to which L1 representations might be activated during L2 lexical processing, Thierry and Wu (2004, 2007) asked Chinese-​speaking learners of English and English native speakers to judge whether word pairs were semantically related (e.g., post –​mail) or not (e.g.,

6

66  Studies of L2 Development

train –​ham). Half of the experimental items additionally included a repeated (or shared) character when translated into Chinese (the L1 of the L2 learners). For example, even though “train” and “ham” are not semantically related in English, their Chinese translations share a common character (Hou Che 火车 “train”, Hou Tui 火腿 “ham”). This design was used to understand if Chinese speakers processed English word pairs differently when the word pairs concealed character repetition in Chinese. If there were processing differences between these two types of word pairs, this finding could suggest co-​activation of L1 knowledge during L2 processing. Results from the event-​related potential (ERP) technique, which measures brain activity as an index of neurocognitive processing (for review, see Morgan-​ Short, 2014), showed that English native speakers and Chinese-​speaking learners of English did not perform the same. Even though both groups demonstrated a main effect for semantic relatedness, the concealed Chinese character effect influenced performance in the L2 speaker group only. According to the authors, this finding suggests “lexical activation of the native language during an experiment involving only second language stimuli” (Thierry & Wu, 2007, p. 12534). This is because Chinese speakers showed sensitivity to character repetition even though they were reading English words only. Not only do these findings suggest very close relationships among a speaker’s different languages, but they also suggest that use of one language stimulates or activates the other known languages. Jiang et al. (2020) extended this line of research by investigating the emergence and storage of L2 lexical representations. They investigated what co-​activation of L1 during L2 lexical processing might suggest about the nature and organization of lexical knowledge in the multilingual mind. For example, is it the gradual emergence of lexical links between L1 and L2 knowledge sources that leads to activation of L1 during L2 use or is L1 knowledge somehow more integral to the development of L2 knowledge? Jiang et al. (2020) addressed this question by integrating information about L1 frequency into their study design because “if L1 translations are not involved in L2 word recognition, or if their activation is a by-​product of L2 word recognition, the frequency of L1 translations should not affect L2 word recognition” (Jiang et al., 2020, p. 219). Thus, if the processing of L2 words appeared to be influenced by the frequency of L1 equivalents, this might suggest a more integrated relationship between L1 and L2 rather than L1 and L2 being separate sources of knowledge that are connected via lexical links. Here, the important question is this: to what extent does L2 knowledge emerge from L1 knowledge? In this study, participants were two groups of Chinese-​speaking learners of English and a group of English native speakers. The learner groups differed in terms of “length of immersion” (ibid., p. 220), with one group recruited in China (with minimal immersion experience in the target language) and one group recruited in the United States (with comparatively greater immersion experience

6

67

6

Studies of L2 Development  67

ranging from one to 45 months). The main data collection tool was a lexical decision task (for review, see Jiang, 2013) in which participants saw a string of letters and had to decide whether it was an English word or not. The stimuli included English words that were “matched for frequency and length but differed in the frequency of their Chinese translations” (Jiang et al., 2020, p. 220). For example, the English word “research” was found to have a mean frequency of 21.6 per million, but its Chinese translation (yanjiu) was found to have a mean frequency of 810.3 per million. In contrast, “memory” and its Chinese translation (jiyi) were just as frequent in both languages, 22 and 26.9 per million, respectively. Word frequencies were determined using frequency lists provided by the Beijing Language Institute (1986) for Chinese and Brysbaert and New (2009) for American English. Using this design, the authors investigated whether the speed and accuracy of processing English words varied as a function of their translation frequency in Chinese. The results showed that speed of processing English words in the L2 learner groups was influenced by the frequency of the Chinese translations: Reaction times were faster for English words with high frequency Chinese translations, but slower for English words with low frequency Chinese translations. No meaningful differences were found between the L2 groups, potentially indicating that “immersion experience” was not a key factor in explaining word recognition.The authors concluded that “this finding suggested that L1 translations were activated and their activation was more than a by-​product of L2 word recognition. Instead, they affected L2 word recognition” (Jiang et al., 2020, p. 223). These findings were interpreted in terms of “a verification or checking procedure whereby an activated word is checked against its L1 translation” (ibid., p. 228). Under this view, the meaning of an L2 word can only be identified following a verification/​checking procedure against its L1. This entails that the accessibility of L1 lexical items can influence the speed of L2 processing, especially since high frequency lexical items tend to be processed/​accessed faster than low frequency lexical items.This is one explanation for why L1 frequency appeared to play a role in L2 word recognition. Taken together, this line of research about potential relationships between L1 and L2 has shown to be a catalyst for empirical research and theorization in our field (e.g., De Houwer & Ortega, 2018). This is because such research is designed to advance what we know about the nature and emergence of L2 knowledge as well as how different aspects of (language) knowledge and experience are connected. Findings that L2 processing can be influenced by factors like translation equivalents and distributional properties of the L1 input point to an even closer integration of L1 and L2 than we might initially have thought. Indeed, these ideas are captured in some of the L2 learning models reviewed in Chapter 2, especially those which forefront critical roles for prior language knowledge and experience in L2 development (e.g., Ellis, 2006a; MacWhinney, 2005).

68

68  Studies of L2 Development

3.3.3  Inhibitory Control and L2 Lexical Processing Just as studies of co-​activation have shaped what we know about the nature and organization of cross-​language relationships in the multilingual mind, an equally important contribution to this body of knowledge involves investigating the ways in which speakers select between their languages and theorizing the cognitive processing mechanisms that support this ability (Meuter & Allport, 1999; Poulisse, 1999; see Chapter 2). Informed by previous research that has repeatedly demonstrated co-​activation of a speaker’s languages, even in initial L2 learning (Bice & Kroll, 2015), it is now understood that language learning involves not only the creation and/​or development of language representations but also the creation and/​or development of cognitive processing mechanisms that allow for specific language representations to be accessed and selected (Calabria et al., 2018; Green, 1998). In other words, multilinguals need to be able to select between their known languages at any given time. This line of research seeks to better understand how speakers carry out and manage language selection. In the field of psycholinguistics, a common approach to investigating language selection has involved the study of switch costs (Costa & Santesteban, 2004; Declerck & Philipp, 2015; Meuter & Allport, 1999). A switch cost is typically operationalized as the amount of time (in milliseconds) it takes a speaker to switch from one language to a different language (e.g., from L1 to L2, L2 to L1). In Timmer et al. (2019), for instance, a picture naming task forced speakers to regularly switch between languages by having them name pictures in either Catalan or Spanish. A solid or dashed line border indicated whether the picture should be named in Spanish or Catalan (see also Costa et al., 2006; Linck et al., 2008). Among unbalanced bilinguals and/​or L2 learners, smaller switch costs are understood to indicate more optimized language selection mechanisms (but see Bobb & Wodniecka, 2013), including, but not limited to, greater abilities in applying and reversing inhibitory control (see Chapter 2). In other words, learners with optimized language selection mechanisms appear better able to switch between languages, as reflected in more accurate performance and shorter reaction times (when compared to non-​switching contexts). Linck et al. (2012) contributed to this line of research by probing the relationships among inhibitory control and language switching, informed by the understanding that high levels of inhibitory control are one explanation for efficient switch cost behaviors. As discussed in Chapter 2, this is because inhibitory control is thought to be a critical cognitive mechanism in language selection. Linck et al. (2012) examined this question among speakers of three languages (English, French, and Spanish) to understand how inhibitory control is related to a speaker’s ability to use and switch between multiple languages at various levels of expertise.

68

69

68

Studies of L2 Development  69

Participants were English-​speaking learners of French and Spanish in Ontario (Canada), enrolled in advanced French and intermediate Spanish language courses. Self-​ratings of proficiency showed that the participants identified as English dominant in reading, writing, speaking, and listening, followed by French and then Spanish. As such, Linck et al. (2012) categorized the study participants as L1 English, L2 French, and L3 Spanish speakers. The study’s main data elicitation tasks were a picture-​naming task (e.g., Costa & Santesteban, 2004; Schwieter & Sunderman, 2008) and the Simon task (e.g., Bialystok, et al., 2004). First, the picture-​naming task included ten black and white drawings of nouns (e.g., pencil, house, car, dog). For each test item (or trial), a picture appeared on the computer screen with either a blue, red, or yellow background. Participants were instructed to the name the picture they saw on the computer screen.The background color of the picture served as a cue for which language to use when naming the picture: blue background for English, red background for French, and yellow background for Spanish. For example, if the computer screen showed a picture of a house on a red background then the image should be named in French (la maison “the house”), but if the background was yellow it should be named in Spanish (la casa “the house”). The trials were grouped into switch and non-​switch trials involving all language combinations. For example, one of the nonswitch trials included picture naming in English only (i.e., all pictures in that trial appeared with a blue background) and one of the switch trials included picture naming in English and French (i.e., some of the pictures in that trial appeared with a blue background and some with a red background). The Simon task (Simon & Rudell, 1967) was selected as the measure of inhibitory control. On the computer screen, a series of red and blue boxes were presented in one of three locations: center, left of center, or right of center. Participants indicated the color of the box by pressing a button on the left or on the right. This design required participants to ignore the location of the box and focus on the color of the box only. In matched trials, the location of the box and the button press matched (e.g., the box was on the left of the screen and required the left button to be pressed). In mismatched trials, the location of the box and the button press did not match (e.g., the box was on the right of the screen and required the left button to be pressed). Response times were expected to be longer on mismatched trials than on matched trials due to a mismatch between the stimulus on the screen and the button press. The magnitude of the differences (in reaction times) between matched and mismatched trials is called the Simon Effect, “an indicator of an individual’s ability to inhibit the prepotent tendency to respond based on the (task-​irrelevant) location of the stimulus” (Linck et al., 2012, p. 655). In short, the Simon Task can provide information about a speaker’s ability to inhibit dominant responses (e.g., Linck & Weiss, 2015; Poarch & van Hell, 2012). Overall, Linck et al.’s (2012) results showed that in the picture naming task participants responded faster in English (L1) than in French (L2) or Spanish (L3).

70

70  Studies of L2 Development

This means that on seeing a picture of a house, for example, it took speakers less time to name the picture in their native language, followed by their next dominant language, and then their weakest language. In terms of switching between languages, switch costs were smaller in L3 than the more dominant L2 and L1. This result was predicted since switches are understood to be less costly when they involve a weaker language (e.g., L3). In the Simon task, results showed longer RTs in the mismatched condition than in the matched condition. This means that, as predicted, participants were slower to respond when there was a mismatch between the stimulus and the button press. One interpretation of this is that the mismatched condition required participants to inhibit the prepotent response. Lastly, it was found that better performance on the Simon task (i.e., inhibition control ability) predicted smaller switch costs when switching into the L1, but not when switching into L2 or L3. Taken together, Linck et al.’s exploration of language performance, cognitive processing, and the relationships among them indicated an important relationship between speakers’ inhibitory control abilities and language switching involving L1. In particular, the findings suggest that language selection abilities (as evidenced by switch costs) appear related to a more general cognitive ability to manage and resolve conflict (as evidenced by performance on the Simon task). The finding that aspects of general cognition (e.g., an ability to inhibit a prepotent response) appeared related to language performance are consistent with claims that language learning and use involve cognition more generally (Bybee, 2010;Tomasello, 2003). It is also important to note that inhibitory control appeared to play a greater role in switches involving L1, but less so for switches involving the L2 and L3. One reason for this is that L1 is likely to be the more dominant, prepotent response that requires inhibition. In addition to providing information about language switching and theorizing the potential cognitive processing mechanisms involved, the current line of research can inform how we think about crosslinguistic influence in L2 learning. This is because it broadens our understanding of the negative effects of crosslinguistic influence to additionally include difficulties with language selection. That is, then, just as some accounts of crosslinguistic influence in L2 research have explained L2 learning difficulties in terms of missing or underdeveloped L2 representations, it is likely that underdeveloped language selection mechanisms can play an equally important role in explaining L2 learning (see Darcy et al., 2016; McManus, 2021). For example, negative instances of crosslinguistic influence could be explained by difficulties associated with selecting the appropriate processing routine and/​or inhibiting the dominant one. In Chapter 4, we review a small but growing body of research that has sought to reduce the negative effects of crosslinguistic influence by providing learners with extensive language switching practice as a means to develop language selection mechanisms. Here, difficulties associated with language selection appear to be one explanation for the negative effects of crosslinguistic influence in L2 performance.

70

71

70

Studies of L2 Development  71

3.3.4  Changes to L1 Performance Even though a primary focus of crosslinguistic influence research has investigated the extent to which L1 experience can shape L2 learning (see Chapter 1), a small body of research has also examined other directions of crosslinguistic influence, including how L2 learning can influence L1 use (e.g., Dussias & Sagarra, 2007). These lines of research are needed to develop more comprehensive understandings about the ways in which learning a new language might lead to changes in a speaker’s broader language system. In this vein, Kroll et al. (2002) investigated how classroom experience with French influenced the speed and accuracy of lexical processing in L2 as well as in L1. As previously noted, this is an under-​investigated line of research, but it has important consequences for how we think about multilingualism. One reason for this is that it can help us to better understand cross-​language relationships in the multilingual mind, both in terms of how these emerge as well as how such relationships are maintained and change over time. In Chapter 1, we noted that our field needs more research that examines the different directions of crosslinguistic influence. This is because an underexplored idea up until relatively recently was that crosslinguistic influence is one-​directional, but recent lines of inquiry are beginning to force researchers to revisit this claim. To contribute to this understanding, Kroll et al. (2002) compared picture naming in L1 and L2 and translation among English-​speaking learners with different amounts of classroom exposure to French. These tasks were selected to understand the fluency or ease with which learners were able to access lexical information in each of their languages as well as produce the phonology of the target item. Participants were split into two groups based on the number of years they had studied French. Learners in the “low” group had studied French for five or fewer years (group mean = 3.4 years of French study), while learners in the “high” group had studied French for five or more years (group mean = 8.4 years of French study). Participants completed four tasks: word naming in English, word naming in French, translation from English to French, and translation from French to English. Each task was presented in language blocks so that word naming in English and French was not mixed (i.e., the task did not require learners to regularly switch between languages, as in Linck et al., 2012, for example). That is, learners completed word naming in English and then word naming in French. In all tasks, participants saw a word on a computer screen and were instructed to either name it in a specific language or to translate it. Results showed that overall picture naming in L2 was slower than in L1, indicating that irrespective of the amount of classroom experience, both groups named pictures in L1 at a faster rate than they did pictures in L2. This was predicted since these speakers were classroom L2 learners and they were expected to be better users of their L1. In terms of between-​g roup comparisons, results showed that picture naming in L2 was faster and more accurate in the high group than in the

72

72  Studies of L2 Development

low group. For L1 picture naming, however, while no between-​g roup differences were found for accuracy, RTs were faster in the high exposure group (see also Bice & Kroll, 2015). Firstly, this result that the groups differed in terms of how quickly they named pictures in their L1 was unexpected since all learners were L1 speakers of English. Second, Kroll et al. suggested this finding shows that L2 learning can affect the fluency of L1 use. In terms of the translation tasks, both groups performed faster when translating from L2 to L1 than from L1 to L2. Overall, translation speed was faster in the high group than in the low group. Although the results for performance in the translation and the L2 picture naming tasks were expected and matched the study’s predictions, performance in the L1 picture naming task was not expected. The authors suggested that L2 learning may have influenced or changed the ways in which L1 knowledge is stored, accessed, and/​or assembled. L2 learning may have led to the re-​organization of L1 representations or, alternatively, the creation of L2 knowledge may have triggered the development of new and possible different access and assembly procedures to existing L1 knowledge. Over time, experience using L1 and L2 allows these new routines to become consolidated. In sum, therefore, knowledge creation, storage, as well as new access and assembly procedures triggered by L2 learning can introduce changes into a speaker’s broader language system (as detected through slower reaction times, for example), but use of both languages over an extended period of time may allow the system to adapt to these new changes. This is one explanation for why picture naming in L1 appeared faster among learners in the high exposure group. Indeed, comparisons of picture naming in L1 and L2 indicated larger differences in the low group (151 ms) than in the high group (93 ms). Observations that the high group performed more quickly relative to the low group could also suggest re-​organization of the entire language system, including the creation of between-​ language connections, that was potentially more developed than in the low group. Related to this, Dussias and Sagarra (2007) suggest that while L2 learning might lead to the creation of separate L1 and L2 systems, at least initially, these systems may begin to merge over time and with usage. As a result, convergence of L1 and L2 systems could be one other explanation for the faster performance in the high exposure group. Taken together, these results, along with those from Dussias and Sagarra (2007), suggest changes to L1 performance in light of L2 learning.They provide evidence that the L1 knowledge and experience speakers bring to the task of L2 learning is neither static nor inflexible. Indeed, research showing that the whole language system is sensitive to patterns of usage gives good reason for L2 researchers to begin incorporating measures of L1 use into studies of L2 development. Whether evidence of changes to L1 performance might additionally provide information about the types of knowledge that L2 speakers build is an empirical question (e.g., L1 changes might be more detectable for L2 knowledge that is procedural or automatic than declarative). A growing body of research indicating very

72

73

72

Studies of L2 Development  73

close relationships between L1 and L2 suggests this to be a worthy area for future research. An additional task for future research would be to investigate these questions with longitudinal designs to understand changes in learners’ L1 systems over the course of L2 learning. Cross-​sectional designs can provide a broad perspective about the ways in which L2 experience can influence L1 performance, but more information about the trajectories of these changes are needed to better understand how and in what ways L2 and L1 use change.

3.4  L2 Learning and Phonology Our aim in this section is to review some of the main trends in L2 learning research about the role of crosslinguistic influence in understanding and explaining phonological development among L2 speakers. This section includes L2 research about the ways in which aspects of L1 production can influence L2 pronunciation, the connections among perception and production abilities in L2 with regard to inhibitory control, the extent to which L1–​L2 similarities and differences can explain L2 learning outcomes, and how L1 and L2 sound systems might be represented. Taken together, these topics connect with the models of L2 learning and the theoretical concepts discussed in the Chapter 2 while additionally demonstrating that studies of phonological development are critical for understanding the ways in which prior knowledge and experience can shape L2 learning.

3.4.1  Inhibitory Control and L2 Performance As we have seen throughout this book, aspects of cognitive processing (e.g., attention, inhibition) are not only helpful for understanding and explaining aspects of L2 learning, but they have also provided an important motivator of empirical research and theorization in the field. In L2 phonology, research involving cognitive processing has investigated how working memory (e.g., Reiterer et al., 2011), attention (e.g., Safronova & Mora, 2013) and musical ability (e.g., Christiner & Reiterer, 2013), for example, can help us to better understand the routes and rates of L2 learning. In line with this research and consistent with work in morphosyntactic and lexical learning, studies of L2 phonology have shown inhibitory control to be a key cognitive process in understanding L2 performance, including language switching abilities and instances of negative crosslinguistic influence (see Linck et al., 2014). It is interesting to note that while Green’s (1998) inhibitory control model was initially proposed to explain lexical processing, it has since been productively applied to other aspects of language use. In one line of this research, Darcy et al. (2016) investigated the extent to which optimized inhibitory control led to more accurate perception and production of L2 segments (vowels and consonants). This research builds on previous studies of inhibitory control and L2 learning by asking whether individuals who appear better able to manage and resolve competition also demonstrate higher levels of L2 performance. This

74

74  Studies of L2 Development

is because optimized language selection mechanisms are thought to reduce the negative effects of crosslinguistic influence. To address this question, Darcy et al. (2016) compared L2 performance in a variety of tasks from Spanish-​speaking learners of English in Spain and English-​ speaking learners of Spanish in the United States. Self-​ratings of L2 abilities in speaking, listening, reading, and writing showed very similar L2 ratings across the two learner groups (mean rating = 4 or “good” on a five-​point scale). L2 proficiency was also tested using the X_​Lex test (Meara, 2005), a test of receptive lexical knowledge. In the X_​Lex test, words are presented one by one from a range of frequency bands, from more frequent to less frequent. Learners indicated whether the word they saw was an actual word. L2 ability in phonological processing was tested using an ABX categorization task for perception (Gottfried, 1984), a delayed sentence repetition task for production (Trofimovich & Baker, 2006), and a retrieval-​induced inhibition task for inhibitory control (Lev-​Ari & Peperkamp, 2013). First, in the ABX task, learners heard three stimuli in a row and had to choose if the last token (X) sounded more similar to the first token (A) or the second token (B). The stimuli were trisyllabic nonwords in Spanish (e.g., [saˈɾeβo]) and English (e.g., [səˈʃi:dən]). The stimuli were carefully designed so that nonnative vowel and consonant contrasts for Spanish-​speaking learners of English were native contrasts for English-​speaking learners of Spanish and vice versa. For example, the vowel contrast /​i-​ ɪ/​is a native contrast in English but it is a non-​native contrast in Spanish (see Darcy et al., 2016, p. 748). Second, for production, the delayed sentence repetition task provided learners with a question (the prompt; e.g., Did anyone get the job?) and then an answer (the response; e.g., Yes, they chose a brilliant person). The prompt was presented again, and learners had to repeat the answer they had previously heard. The first hearing of the prompt and response was accompanied by on-​screen text, but the second hearing was not. Third, for inhibitory control, the retrieval-​induced inhibition task required learners to memorize six words in their L1 after seeing their visual presentation on the computer screen. The six words were from three different categories (vegetables, occupations, animals). Participants practiced half of the words from two of the categories (e.g., tomato, nurse) by typing them out several times. Participants were then tested on their recognition of the practiced words, the unpracticed words, and new distractor words.This task was designed so that practicing some of the words would increase the activation level of those words and, at the same time, lead to inhibition of the unpracticed items. Overall, results showed that the most difficult type of L2 contrasts for learners to perceive and produce were those that were not native to the L1. This finding played out in both L2 groups irrespective of the specific L1. Taking these results, the authors examined the relationship between phonological processing and inhibitory control.This analysis examined whether performance on the inhibitory

74

75

74

Studies of L2 Development  75

control task appeared related to L2 abilities in production and perception (thus following a similar approach to Linck et al., 2012). Results showed that performance in the perception task correlated negatively with inhibition scores, indicating that higher inhibition scores patterned with more targetlike vowel and consonant perception. However, no relationship was found between inhibition and vowel production accuracy. Taken together, these results indicate relationships among different components of phonological ability and cognitive processing. Relationships with inhibitory control appeared more detectable in perception than in production. One question going forward is how this difference might be accounted for. The authors suggest that perception and production might not be governed by the same processing mechanisms. In particular, L2 production is thought to require inhibition of dominant motor (articulatory) plans from L1, but perception might not (assuming no role for sub-​vocalic repetition in perception). It is also possible that the use of non-​words in the perception task and the use of real words embedded in sentences in the production task played a role. That is, then, the perception task required learners to respond to made-​up words, but the production task required learners to respond to real words presented in discourse.These different task designs would be an additional factor to consider in future work. One way to address this could be to use real words in the perception task. In addition, the authors suggest that an inhibition task that tapped motor-​based inhibition, such as the Simon task, may lead to a different patterning of results than other types of inhibition tasks. In terms of a broad-​level interpretation of the study’s findings, Darcy et al. suggest that an ability to keep L1 and L2 systems apart with inhibitory control is one interpretation of these findings (but see MacWhinney (2005) and “decoupling”). This is because enhanced inhibitory control mechanisms can facilitate language selection. The more able speakers are to keep L1 and L2 systems separated with inhibitory control, the better able they will be to regulate language selection, thus enhancing L2 use. In other words, enhanced levels of inhibitory control could facilitate L2 learning by allowing learners to manage and keep separate their different language systems. Indeed, this interpretation is consistent with a growing narrative in the field of L2 learning that suggests cognitive processing mechanisms like inhibitory control appear to be closely related to language performance in a range of skills. In this study, target-​like L2 use in perception was related to inhibition. In terms of advancing this research program, replication studies can make an important contribution by incorporating additional and/​or different tasks (e.g., non-​words in the production task) and different designs (including longitudinal research designs) to understand the nature of these effects on learning (see Porte & McManus, 2019). Research should also consider the extent to which inhibitory control might be amenable to improvement through practice (see also Chapter 4). At this time, however, it seems that further exploration of the effects of cognitive processing effects on L2 learning is needed, including, for example,

76

76  Studies of L2 Development

the extent to which cognitive processing training is possible and can be beneficial for L2 learning.

3.4.2  Understanding Crosslinguistic Similarities Since one aim of crosslinguistic influence research is to understand the ways in which differences between a linguistic feature in L1 and L2 might lead to specific types of learning difficulty, agreed techniques for the identification of crosslinguistically (dis)similar language features are important. This is a critical methodological consideration given that claims about L2 learning difficulties tend to depend on the extent to which a structure, meaning, or linguistic feature (e.g., morpheme, phoneme) is similar or not in L1 and L2. In many cases, the magnitude of cross-​language similarity is thought to be an indicator of potential learning difficulty (e.g., Best & Tyler, 2007; Escudero, 2005; Lado, 1957). Up to this point, we have seen different proposals for how we might determine cross-​ language similarity (see Chapter 2, in particular). For example, drawing on the UCM, Tolentino and Tokowicz (2014) categorized L2 target features as “similar”, “dissimilar”, and “unique” based on the meanings expressed by the same linguistic cues in L1 and L2 (e.g., using “s” to pluralize singular nouns in Spanish and English was considered to be a cross-​language similarity, as in perro –​perros “dog-​dogs”). In L2 phonological learning, Flege (1987, 1988, 1992, 1995; see also Flege & Bohn, 2021) proposed categorizations of “identical”, “similar” and “new”, which include similarities with the UCM. Sounds that are identical in L1 and L2 are claimed to be relatively easy to learn and can be facilitated by positive transfer. For “similar” and “new” sounds, L2 learning is thought to be influenced by a learner’s ability to detect the phonetic dissimilarity. For “new” sounds, detecting an L1–​L2 difference can lead to the creation of a new phonological category. This suggests that the ability to perceive sound differences is a likely first step in the creation of new knowledge and for learning to take place (or at least begin). Learning “similar” sounds in the L2 is thought to be the most difficult. Clearly, these are ideas that have great potential to explain and predict L2 learning difficulties and are proposals that are well-​represented in many models of L2 phonological learning and use (e.g., Best & Tyler, 2007; Major & Kim, 1996). One part of this puzzle, however, is how to determine whether a sound in L1 and L2 is similar, dissimilar, or new. Not only is this necessary to advance theory building in the field, but doing so in a robust and reliable manner is essential for making predictions about L2 learning. Yang et al. (2020) contributed to this line of research by using well-​established methods in the field for classifying L2 sounds relative to the L1 (IPA comparison, acoustic difference, and feature redeployment). Using these methods, Yang et al. classified Russian voiced stops as dissimilar L2 sounds for Chinese learners. A perceptual assimilation task was then used to validate these claims, followed by perceptual discrimination and production tests to

76

7

76

Studies of L2 Development  77

examine L2 learning (performance on the perception and production tests is not discussed here, see Yang et al., 2020). If we begin by considering how cross-​language similarity is determined, Yang et al. (2020) raise the important point that “there has been little discussion focusing specifically on the independent measurement of cross-​linguistic phonetic similarity” (ibid., p. 3). Even though researchers have used a number of methods for identifying L2 sounds as “new” relative to the L1, there has been little systematic exploration of the validity and reliability of these measures. Yang et al. set out to investigate this issue. First, International Phonetic Alphabet (IPA) transcription is understood to be a common method for defining the nature of a cross-​linguistic similarity (following Flege, 1987, 1988, 1992), which is often used in studies of L2 phonological learning (e.g., Ingham, 2014; Larson-​Hall, 2004; Major, 1987). An L2 phone is identified as “dissimilar” when its transcription uses IPA symbols that do not exist in L1. Second, acoustic difference is thought to be “a rather concrete scale for assessing cross-​linguistic similarity” (p. 4, see also Flege, 1992) that functions with vowels by comparing the multidimensional space of their first and second formants (F1, F2). Third, feature deployment is a method that assesses the distinctive features underlying L2 sounds. It is understood that L1 features that are deployed in a different way or if the L2 phone uses features that do not exist in L1, then these types of L2 phones could lead to L2 learning difficulty (see Brown, 1998, 2000; Flege, 1988; McAllister et al., 2002). Using these measures, Yang et al. (2020) determined that the Russian voiced stops /​b d g/​should be classified as “new” sounds for Chinese learners. Importantly for Yang et al. (2020), these classifications were verified using a perceptual assimilation test (Strange & Shafer, 2008). This type of test can be useful for understanding the perceived relationship between L1 and L2 sounds because learners are required to choose the L1 sound that is most similar to the L2 sound and the rate the degree of similarity between the sounds (see Flege, 1991; Levy, 2009). This is an important consideration since, according to Flege, L2 speakers need to detect a dissimilarity for learning to take place. In the perceptual assimilation test, participants heard a Mandarin or Russian consonant, then selected the Mandarin consonant transcribed in Pinyin that was most similar to the sound they heard, and then rated their degree of similarity on a scale of 1–​5. If the sound was nonassimilable, participants were instructed to select “none”. Yang et al. recruited Chinese-​speaking learners of L2 Russian in China who were split into low and high groups based on prior experience with Russian. The low group had spent an average of two years learning Russian in China, whereas the high group had been learning Russian in China for an average of four years and had spent one year abroad in Russia. Results showed that “over 80% of both Russian stops /​b d g/​and /​p t k/​were assimilated to the Mandarin voiceless unaspirated stops /​p t k/​” in both groups of Russian learners (p. 9). According to previous research, an L2 sound that is assimilated to an L1 sound more than

78

78  Studies of L2 Development

70% of the time is said to be categorized as an L1 sound (see Faris et al., 2018; Tyler et al., 2014). The results on the perceptual assimilation task thus indicated that the two types of Russian stops were perceived as highly similar sounds to the corresponding Mandarin sounds, contrary to the linguistic measures of IPA, acoustic difference, and feature deployment. Since performance in a perceptual assimilation test arrived at a different conclusion about linguistic (dis)similarity compared to the linguistic measures, it is worth noting that Yang et al.’s findings do not mean that we should question the Speech Learning Model (see Flege, 1987, 1988, 1992, 1995). This is because the model makes predictions for L2 learning based on the extent to which a speaker can detect a difference between L1 and L2. Detecting a difference is therefore a necessary condition for learning to take place. The linguistic measures of dissimilarity (e.g., IPA, acoustic difference) do not include information about the extent to which speakers may or may not be able to detect a difference. We therefore have a potential mismatch about the ways in which L1–​L2 similarity can be operationalized and the predictions of a particular theory. The present study is therefore a helpful example that reminds us to think critically about the ways in which we go about determining cross-​language similarity in studies of crosslinguistic influence. Indeed, previous L2 research has provided important lessons for why predictions based on linguistic description alone might not always be borne out in L2 learning experiments (see Lado, 1957). Other aspects of learning, informed by learning science and psychology, for example, can also play an important role in determining L2 learning difficulty (e.g., see Ellis, 2006a; Spinner & Gass, 2019). Yang et al.’s (2020) study is therefore a useful reminder for us to be critical of the ways in which we operationalize key aspects of our study designs.

3.4.3  Connections between Perception and Production An ongoing line of research in L2 phonology that connects with some of the methodological matters previously discussed includes studies of how L2 abilities in production and perception emerge, are related, as well as how performance/​ training factors may play a role in their emergence (e.g., Huensch & Tremblay, 2015; Thorin et al., 2018). With respect to L2 production, a large body of research has investigated the factors that contribute to accentedness (or, having a “foreign” accent, see Goto, 1971; Flege, 1999). According to Best and Tyler (2007) and Flege (1995), difficulties with the perception of L2 sounds can prevent speakers from producing L2 sounds that are native-​like. In addition to questions about how we classify a sound as native-​like or not, the key here is that perception and production are thought to have a developmental relationship. Under this view of learning, L1 phonological categories are understood to “transfer” to L2. L1 phonological categories are then used “to perceive and then produce L2 sounds as a function of their similarity to L2 sounds” (Kartushina & Frauenfelder, 2014, p. 1). Two

78

79

78

Studies of L2 Development  79

points are important here. First, L1 phonology can play a critical role in L2 speech learning. Second, the ability to perceive L2 sounds is thought to have an important influence on L2 production. Important implications of this understanding are that (i) L1 sounds that are different or are absent in L2 will be the most difficult to learn and (ii) developing target-​like abilities in perception is a necessary condition for target-​like L2 production. However, as noted by Kartushina and Frauenfelder (2014), even though these accounts of L2 phonological learning are relatively common in L2 research, the evidence base underlying these accounts is mixed, including, for example, findings that show no relationship between perception and production (e.g., Peperkamp & Bouchon, 2011), weak relationships only (e.g., Levy & Law, 2010), accurate production but poor perception (e.g., Kassaian, 2011), and weak or no transfer of perception training to L2 production ability (e.g., Lopez-​Soto & Kewley-​Port, 2009). This state of knowledge points to an unclear relationship between perception and production in L2, even though their relationship is thought to have a strong theoretical basis. One reason for this could be that insufficient attention is given to the experience that speakers bring to the task of L2 learning. Kartushina and Frauenfelder (2014) addressed this issue by investigating what role L1 production plays in L2 production. In so doing, this study advances the idea that abilities in L2 production may vary as a function of L1 production abilities rather than L2 perception abilities.This study therefore ascribes a critical role to prior experience in L2 learning. To understand the role of L1 experience in L2 phonological learning, the authors examined the roles of L2 perception and L1 production in L2 production given that some previous research has suggested that L2 perception and L2 production abilities might not be as closely connected as initially thought. Participants were Spanish-​speaking learners of French, with an average of four years of classroom-​based experience learning French. Their L2 abilities in French were identified as intermediate (using the Common European Framework of Reference for Languages, see Council of Europe, 2001). The study involved three tasks: a repetition task, a naming task, and a reading task. In the repetition task, participants heard an isolated L2 vowel (e.g., /​i/​) and were asked to repeat it as correctly as possible. The naming task included two phases, familiarization and testing. First, participants were exposed to a visual and an auditory stimulus on a computer screen (e.g., a picture of a bed [“lit” in French], followed by the auditory presentation of the vowel /​i/​and the word lit, “bed”). In the testing phase, participants were instructed to produce the vowel contained in the picture (e.g., on seeing the picture of the bed, the target vowel was /​i/​, and learners were supposed to produce the vowel /​i/​). A reading task was used as an assessment of L1 production, in which participants read aloud a passage of text (“The Godson” by Leo Tolstoy). In terms of L2 perception performance, results showed that perceiving L2 vowel contrasts was most difficult for sounds not present or similar to those in L1.

80

80  Studies of L2 Development

This finding is in line with previous research that new contrasts are more difficult for L2 speakers to perceive. For L2 production, the results generally showed that contrasts that were difficult to identify were also difficult to produce. However, when individual variability was factored into the analysis (correlation and by-​ subject mixed-​effects regression analyses), no relationships were found between the ability to identify a sound and the ability to produce that sound. As a result, the authors concluded that these findings indicated no relationship between speakers’ abilities in L2 perception and L2 production. That is, then, “speakers’ identification accuracy did not predict their production accuracy” (ibid., p. 13). In terms of a role for L1 production, the findings showed that the phonetic properties of an individual’s L1 productions predicted L2 production accuracy. Comparisons between vowels in L1 and L2 production showed that the closer the L1 category was to the target vowel, the higher the production accuracy for the target vowel was. The authors propose that these results likely suggested “a shared inter-​ phonological space between languages” (p. 15). Overall, these findings, together with other studies designed to better understand the extent to which L2 perception and L2 production abilities might be related, suggest that attributing a role to prior experience (in this case, L1 production) may provide a missing component in our understanding of L2 speech learning. Although the findings appear to contrast with some previous research suggesting L2 perception and L2 production abilities are connected, the results are, at the same time, consistent both with a growing body of evidence that challenges this claim (e.g., Peperkamp & Bouchon, 2011) as well as cognitive theories of general learning (i.e., not speech learning only; e.g., Anderson, 1982; DeKeyser, 1997, 2017). Skill Acquisition Theory, for instance, predicts comparable performance in perception and production only if the knowledge underlying those skills is declarative (rather than procedural or automatic, for instance). This is because expertise in production is driven by practice that leads to skill specificity (Li & DeKeyser, 2017). This is one reason why research generally shows that the development of expertise in one skill (e.g., perception) does not benefit a different skill (e.g., production, see DeKeyser, 2017). It is also important to reflect on the types of data used in Kartushina and Frauenfelder’s (2014) study, especially since the L2 data were collected with tightly controlled tasks that elicited perception and production of specific sounds in isolation, but the L1 production data were collected from a story retelling (i.e., the same specific sounds but embedded in words and larger stretches of discourse). Collecting data in L1 and L2 that are more comparable (e.g., story retellings in L1 and L2) would be an additional way to understand the relationships between L1 experience and L2 speech learning. In addition, collecting L1 perception data would potentially help us to understand the contribution of L1 experience to L2 speech more fully. One way to implement these changes while allowing for comparison would be via replication, thus allowing future research to advance

80

81

80

Studies of L2 Development  81

knowledge and understanding about the role of prior experience in L2 speech learning.

3.4.4  Changes to L1 Perception As we have noted elsewhere in this chapter, a common focus of L2 learning research has involved study designs that seek to understand how crosslinguistic differences can influence L2 use. In Chapter 1, this was discussed in terms of the positive effects of crosslinguistic influence, which is generally understood to be when aspects of another language appear to facilitate the learning of an additional language. In our discussion of Kartushina and Frauenfelder (2014), for example, the existence of similar phonological categories in L1 was found to provide an important benefit for L2 production. This is one reason why investigating cross-​ language similarity is important in L2 research: It can advance knowledge and understanding about the ways in which a speaker’s different languages are stored and accessed. As noted in Chapter 2, many theories of L2 learning suggest that L1 knowledge is transferred and then adapted for L2, suggesting some level of independence between the languages. However, as discussed, we know very little about what happens next. For example, perhaps a speaker’s different knowledge sources begin to converge, as suggested by Dussias and Sagarra (2007), or maybe these knowledge sources remain independent but connected therefore requiring the recruitment of language selection mechanisms to support language use (Green, 1998). Other accounts were also discussed in which there might be no need for transfer at all (see Chapter 1). Applied to L2 phonological learning, one line of research contributing to this question has investigated phonemes that are common to both languages and do not differ markedly in their phonetic-​acoustic realization (e.g., /​f/​or /​s/​in German and English). Using the perceptual learning paradigm (e.g., Eisner & McQueen, 2005; Norris et al., 2003,), in which speakers listen to real words and non-​words, with some of the real words containing phonetic realizations that are atypical, Schuhmann (2016) investigated the extent to which highly similar phonemes in L1 and L2 might be represented separately or not. Previous research has found that this type of training can lead to changes in a phoneme’s category boundary. In this way, Schuhmann (2016) investigated the extent to which altering phonetic boundaries in one language might influence perception in a different language for phonemes that appear highly similar across languages (see also Schuhmann, 2014). If changes across languages were found following training in one language only, this might suggest shared representations across languages (see also Hartsuiker & Pickering, 2008). Participants were English-​speaking learners of German and German-​speaking learners of English. All learners self-​reported intermediate to advanced L2 proficiency. The study involved an exposure phase followed by a categorization phase.

82

82  Studies of L2 Development

First, participants completed an auditory lexical decision task in which they were exposed to English words and non-​words that contained /​s/​-​stimuli (e.g., dinosaur, embassy) and /​f/​-​stimuli (e.g., daffodil, microphone). The aim of this phase was to provide learners with exposure to unusual pronunciations of the /​f-​s/​phoneme contrast. This was achieved by changing the pronunciations of either the /​s/​-​ stimuli (condition A) or the /​f/​-​stimuli (condition B) while keeping the other stimuli unmodified. Unusual /​s/​-​stimuli included “a [?fs] mixture” in places of /​s/​ and unusual /​f/​-​stimuli included “a [?fs] mixture” in place of /​f/​(see Schuhmann, 2016, p. 4). These mixtures were created by recording two pronunciations of the stimuli, one with /​s/​and one with /​f/​as in [lɛgəsi] and [lɛgəfi] for “legacy”, and then mixing the /​s/​and /​f/​segments using Praat (Boersma & Weenik, 2011). In the categorization phase, participants identified phones on a fricative continuum as either /​f/​or /​s/​in English and German using a forced-​choice task. In this design, the exposure involved English words only, but the categorization task included phonemes in English and German. Overall, the findings showed that following exposure to atypical sounds in English, both L1 and L2 speakers of English demonstrated shifts in their perception of English /​f/​and /​s/​sounds. Furthermore, exposure to English words also changed speakers’ perceptual categorization skills in German, even though no training in German was provided. Taken together, these results indicate that short-​term, intensive exposure to unusual sounds can lead to changes in L1 and L2 speakers’ perceptual categories that extended to other languages that shared very similar phonemes. Schuhmann (2016, p. 11) interpreted these findings to suggest “that the relationship between phonetically highly similar sounds in listeners’ L1 and L2 likely consists of separate yet interconnected mental representations for speech perception. Under this account, phoneme contrast representations between two languages are not statically either separate or merged, but interrelated and dynamic in nature” (for similar results in L2 syntactic processing, see Hopp, 2016). Indeed, these findings suggest that speakers’ language representations are malleable and can be adapted, as suggested by usage-​based accounts of language and development (Bybee, 2010). The results also suggest very close connections between a speaker’s linguistic representations for other known languages, to the extent that changes in one set of representations may also affect other representations in the language system, assuming that multiple language systems exist.

3.5  Conclusion In this chapter, we have reviewed some key lines of research about crosslinguistic influence in L2 learning. Our approach has been to identify strands of research important for understanding current questions and approaches in the field, rather than providing exhaustive accounts. Our review has demonstrated that despite differences in the types of linguistic phenomena being studied (e.g., morphosyntax,

82

83

82

Studies of L2 Development  83

vocabulary, or phonology), many common themes can be identified. Perhaps one of the most important lines of research reviewed here is work that advances new directions about the ways in which a speaker’s language systems are connected. As noted in Chapter 2, this is an important component of future theory-​building in the field.Work in morphosyntax, for example, showed that L2 experience can play a role in reshaping how speakers process information in their native language, and related research indicated that L2 learning can influence the speed and accuracy of L1 processing. Taken together, this research forces us to think more holistically about the multilingual language system and the ways in which new learning experiences can lead to changes in a speaker’s “established” language systems. These findings go to the heart of some of the questions discussed in Chapter 1 about the nature of language and of language learning. Other lines of research indicate important connections with cognitive processing (see Calabria et al., 2018; Kroll & Bialystok, 2013), some of which constitute an important focus of attention in psycholinguistics research but that are relatively under-​explored in the field of SLA (but see Darcy et al., 2016; Gass et al., 2013). Drawing on the ICM (Green, 1998), for example, a growing body of evidence is pointing to important connections between the management and resolution of conflict and language processing. This research suggests that inhibitory control can play a key role in allowing a speaker to inhibit dominant responses during L2 use and thus better regulate language selection. In Chapter 4, we take some of these claims forward and examine the ways in which L2 instructional research might be able to build on these ideas (see McManus, 2021). We have also seen further empirical evidence on the nature of L1–​L2 relations (e.g., Jiang et al., 2020; Thierry & Wu, 2007). These studies have investigated the ways in which L1 and L2 knowledge sources appear connected, be that at the level of form, meaning, or both. These insights directly feed into how we think about “transfer” and development. They also have important implications for how we think about the ways in which a speaker’s languages are integrated. For example, these lines of research could help us to theorize and test ideas about the ways in which a speaker’s language systems are connected and how these connections emerge. In Chapter 4, we explore how findings like these about crosslinguistic influence, informed by both theories of L2 learning and empirical L2 research, have been used to support learning and use. This body of work specifically investigates how instruction can draw on what we know about L2 learning and integrate these insights into the instruction (e.g., drawing learners’ attention to aspects of L2 that might be difficult to perceive). In comparison to the work reviewed in this chapter, work in crosslinguistic influence and L2 instruction is relatively more novel and under explored, but it points to a promising avenue of future research since it allows us to develop instructional techniques that are evidence-​based to improve L2 learning in clear ways.

84

4 INSTRUCTED L2 LEARNING

4.1  Introduction In this chapter, we review studies that have investigated the ways in which instruction informed by L2 research about crosslinguistic influence can be used to support the learning of difficult target features. This work has shown that evidence-​based language instruction, especially that grounded in research findings and theory about crosslinguistic influence, can benefit L2 learning in a variety of skills, including production and comprehension, as measured in tests assessing online and offline language abilities. Our review begins with a discussion of some key debates and ideas in the field of instructed SLA, including potential roles for (explicit) instruction and explicit knowledge in L2 development (for an overview of the field of instructed SLA research, see Loewen, 2020). We then move to a review of L2 instructional research in morphosyntax, vocabulary, and phonology. The specific aim of this review is to highlight the ways in which instruction has been used to address learning difficulties created by L1–​L2 differences. We end this chapter by bringing together these different lines of research and providing an overview of what is currently known about the effectiveness of instruction for addressing learning difficulties resulting from crosslinguistic influence.

4.2  Instructional Effects in L2 Learning A common discussion point in the field of instructed L2 learning concerns if and in what ways instruction can help improve a speaker’s use of an additional language (e.g., Kang et al., 2019; Norris & Ortega, 2000; Thomson & Derwing, 2015). Research designed to address this question is necessary because it not only pushes us to better understand the socio-​cognitive processes of learning (e.g., can DOI: 10.4324/9780429341663-4

84

85

84

Instructed L2 Learning  85

learning potential be maximized and, if it can, in what ways?), but this research has the potential to influence language teaching in the form of best or recommended practices (e.g., Glisan & Donato, 2016; VanPatten, 2017). One example of this includes the use of (instructed) L2 research to shape language teaching policy and the provision of language instruction, such as the American Council on the Teaching of Foreign Language’s Guiding Principles for Language Learning. Here, the United States’ largest membership organization for language educators draws on published research about L2 learning and instruction to “identify what is effective in language learning and provide guidance to educators and learners alike”, leading to recommended practices such as use of the target language at least 90% of the time and incorporating authentic texts and communicative tasks into language learning (see ACTFL, 2021). These are just some of the reasons for why an understanding of instructional effects in L2 learning is important (see also DeKeyser & Prieto Botana, 2019; Sato & Loewen, 2019; Tyler et al., 2018). In terms of defining and describing the boundaries of research about instructional effectiveness and L2 learning, Loewen (2020) proposed that a key aim of this research should be to develop understanding about “how the systematic manipulation of the mechanisms of learning and/​or the conditions under which they occur enable or facilitate the development and acquisition of an additional language” (pp. 2–​3, see also; Housen & Pierrard, 2005). Otherwise said, instructed L2 research involves documenting and theorizing the specific ways in which instruction can lead to improvements in how speakers use an additional language. In essence, this means that when “instructional method A” is found to be more effective than “instructional method B” for improving L2 performance, we want to be asking questions that pinpoint and refine understanding about why “instructional method A” was more effective. As researchers and educators, we know that the effectiveness of a particular instructional method will vary in a range of contexts and conditions because of factors like the complexity of the target feature, the experience the learner brings to the task of learning a new language, how we assess learning, as well as what the learning context looks like (DeKeyser & Prieto Botana, 2019; Leow, 2015). Even though our aim in this chapter is not to unpack the ways in which we might go about determining instructional effectiveness (for a review of considerations, see De Graaff & Housen, 2009; Norris & Ortega, 2000), we do want to highlight two considerations because they are relevant to our broader discussion about crosslinguistic influence and instructed L2 learning. First, instructed L2 research must ask questions about what it means for an instructional method to be effective.This means that we must ask ourselves: effective in terms of what? It is possible for a particular method to be effective in a variety of different ways. One way to look at this question could be: what type of L2 ability does the method improve? For example, “method A” might lead to general improvement in writing (but not speaking) or it might be more focused on improving comprehension (but not production) of a specific morpheme or

86

86  Instructed L2 Learning

phoneme. The effectiveness of a particular instructional method may also depend on the prior experience of the L2 speaker (e.g., beginning L2 learners, advanced L2 learners). Asking questions about what a particular method is effective for can help us tailor instruction to address particular learning difficulties. We will return to this issue later in the chapter when we begin to think about why instruction informed by L2 research about crosslinguistic influence can be useful and why it can be particularly beneficial for certain types of learning difficulties. In short, when we study instructed L2 learning, we want to be thinking about what the most appropriate tools might be for addressing learning difficulties created by crosslinguistic influence. Second, instructed L2 research must study if there might be specific components or designs of the instruction that are making it effective. Work in Processing Instruction (VanPatten, 2002) is one example of this line of inquiry as a means to better understand the instructional component(s) that appear to support learning the most (for review, see Benati & Lee, 2015). Some research for example, has studied what role explicit information (or metalinguistic explanation) about the target feature plays in improving L2 performance. This has been done by comparing structured input practice (a type of comprehension practice, see VanPatten, 2004) with and without explicit information (e.g.,VanPatten & Oikkenon, 1996). Even though some studies have found that explicit information can be beneficial (e.g., Fernández, 2008; Henry et al., 2009), others have found that it seems to provide few additional benefits (e.g., Benati, 2005; Sanz & Morgan-​Short, 2004). One way to explain these mixed conclusions is as follows: the effectiveness of explicit information for improving L2 abilities appears related to the complexity of the target feature as well as the experience of the learner. In short, the effectiveness of a specific instructional component (e.g., explicit information) is moderated by a variety of factors (DeKeyser & Prieto Botana, 2015). Despite some research casting doubts on the potential benefits of explicit information in L2 learning, a good amount of research has shown that it can provide a range of benefits, including enhancing the effectiveness of the practice (Henry et al., 2009) and developing more nuanced understandings of the target feature (e.g., Lantolf & Poehner, 2008; Tyler, 2012). If we focus on the ways in which explicit information can develop greater and more nuanced understandings of the target feature, this means that explicit information can help learners to better understand patterns, exceptions, new meanings for words, and so on.This is because explicit information engages learners in some type of explanation about specific aspects of the target language (Roehr-​Brackin, 2018). This can involve explanations about form, such as how to conjugate a particular tense, about meaning, like an explanation of a new or different concept, as well as explanations about form-​meaning mappings such as when a form that exists in L1 (e.g., articles) is used to express a new/​different meaning in L2 (e.g., grammatical gender).What remains less clear, however, is to what extent an enhanced understanding of language and usage by itself improves L2 performance. For example, as we have seen,

86

87

86

Instructed L2 Learning  87

some approaches to instruction suggest that providing learners with explanations about language is not always necessary (e.g.,VanPatten, 2015), while others suggest that it can play an important role (e.g., Lantolf & Zhang, 2017), especially when accompanied with opportunities for meaningful language use (DeKeyser, 2017). However, even though disagreements exist about whether explicit information can benefit L2 performance, the nature and complexity of the target feature plays a critical role in these claims. For example, it seems that functionally complex linguistic features, with multiple form-​meaning mappings, and that are perhaps less available in the input, are more difficult to learn through exposure alone. Indeed, these types of target features appear to benefit from explicit instruction the most (see Norris & Ortega, 2000; Spada & Tomita, 2010). A last point related to this discussion of instructional effects is that some research suggests that the type of knowledge created through explicit instruction may not be durable or that it might be qualitatively different from the type of knowledge required for actual language use (e.g., Krashen, 1985; VanPatten, 2014). For example, explanations about language can help develop declarative or metalinguistic knowledge about language, but that type of knowledge might not be helpful or useable during interaction. As a result, researchers have tested these claims by examining whether instructed contexts might be conducive to the development of more automatic language abilities, or whether these contexts can only lead to slow and effortful language use (e.g., Suzuki et al., 2019). As one might expect, there is evidence on both sides of the debate. Sometimes instruction can lead to more automatic language performance, whereas other times it might only lead to slow and effortful language use. The type and amount of practice is thought to play an important role here, with research suggesting that different types of practice are desirable. Taken together, explicit instruction has been shown to play an important role in L2 development, but its effectiveness can be moderated by a range of other contextual and learner factors, meaning that it will not always lead to the same effects. Some techniques are better at developing conceptual understanding, while others are better at developing more automatic language abilities. Instruction that seeks to develop both abilities (conceptual understanding, more automatic language use) is particularly desirable. In the next section, we take forward some of these ideas with a view to thinking through some of the reasons why explicit instruction and explicit knowledge might be particularly helpful for addressing the negative effects of crosslinguistic influence.

4.3  Explicit Instruction One of the key features of explicit instruction is that it seeks to improve L2 performance with instructional techniques designed to promote awareness and conscious noticing/​attention of the target feature (Leow, 2015; Roehr-​Brackin, 2018). In short, this means that learners are mostly aware that they are involved in some

8

88  Instructed L2 Learning

The past simple tense is one of the tenses most used to talk about events in the past. It does refer to finished actions and events. Very often the English past simple tense ends in -ed: I invited John for lunch I played tennis with Paula When you talk about a finished time in the past, the English past simple tense is often accompanied by a temporal adverb. Yesterday I smoked 20 cigarettes

DO NOT RELY ON THE TEMPORAL ADVERB TO UNDERSTAND WHEN THE ACTION TAKES PLACE AS SOMETIME [sic] YOU CAN HEAR A SENTENCE WITHOUT THE TEMPORAL ADVERB. YOU MUST PAY ATTENTION TO THE TENSE ENDING TO UNDERSTAND WHEN THE ACTION TAKES PLACE. IN THE CASE OF DESCRIBING PAST EVENTS PAY ATTENTION TO THE ENDING OF THE VERB: -ed

FIGURE 4.1  Explicit

Information Used in Benati (2005) for Teaching the Simple Past

type of learning. A common approach to explicit instruction involves providing speakers with some type of explicit information (or metalinguistic explanation) about language or a language feature. Although studies vary in how they deliver and structure explicit information, they tend to use explanations about the target feature’s structure (e.g., how to identify it, how to form it), how it might be used, as well as other tips that could be helpful in learning and using the target feature. For example, Figure 4.1 is an extract of the explicit information provided in Benati (2005, p. 92). In that study, participants were Greek-​and Chinese-​speaking high school students (aged 12–​13 years) who were beginning learners of English, and the target feature was the English Simple Past. The explicit information shown in Figure 4.1 includes a number of features that are worth noting. For instance, it describes with examples some of the ways in which the Simple Past is used in English (e.g., to talk about events in the past with finished actions/​events) and gives some tips on how to use it (e.g., pay attention to the verb ending). This approach contrasts with implicit instruction that seeks to improve learning without explicit awareness raising and/​or consciousness. R. Ellis et al. (2009, p.16) describes implicit instruction as “provid[ing] learners with experience of specific exemplars of a rule or pattern while they are not attempting to learn it (e.g., they are focused instead on meaning)”. In Bell (2017), for example, short stories with comprehension questions alongside crosswords were used to provide exposure to specific target features without drawing attention to specific grammatical features. In sum, therefore, the main difference between explicit and implicit instruction is learning with and without awareness. However, as many reviews and studies show, determining whether

8

89

8

Instructed L2 Learning  89

and to what extent learners are aware during learning can be a difficult claim to substantiate, especially among instructed learners (Rebuschat, 2013; Rogers et al., 2015). An important question going forward for the present discussion, therefore, is to what extent can instruction address the negative effects of crosslinguistic influence? Although research has mostly examined this question using explicit methods, this does not necessarily mean that implicit instructional techniques might not also be helpful.This is essentially an empirical question. Of the research conducted using explicit techniques, one motivation for using explicit methods is that previous L2 research has shown that exposure alone appears insufficient for overcoming the negative effects of crosslinguistic influence (see Chapter 3). Although some studies suggest that lack of exposure to those specific features might explain the learning difficulty, the learning issue appears more complex than an absence of input (see Ellis, 2006a). For example, Zhao and MacWhinney (2018) investigated article usage among Chinese Mandarin learners of English. As the authors note, a lack of exposure to English articles alone is an unlikely explanation for this learning difficulty given that articles are very frequent in the input (e.g., the cat, a cat, those cats). In learning situations like this, explicit instruction has the potential to effectively support development because it is designed to increase learners’ awareness of aspects of the L2 that might be missed due to prior knowledge and experience. In short, one reason why explicit instruction has the potential to address learning difficulties created by crosslinguistic influence is that it can draw learners’ attention to the source of the learning difficulty.

4.4  Evidence-​Based L2 Instruction: L1 Use in L2 Learning One area of instructed L2 research that continues to grow is the development evidenced-​based instructional methods (e.g., Sato & Loewen, 2019; Tyler et al., 2018). One reason for this is that a core aim of L2 research is to understand the routes and rates of L2 learning, including the extent to which specific conditions, contexts, and factors influence learning. Indeed, instructional research has designed pedagogical techniques to address the cause of particular learning difficulties. Importantly, the design of these instructional techniques is informed by empirical research findings and theories about L2 learning. Drawing on patterns of L1 use as a pedagogical tool makes an important contribution to this line of evidence-​based, theoretically driven, instructional L2 research. This is because the starting point of the instructional design includes many of the observations about crosslinguistic influence and L2 learning discussed in Chapter 2, such as claims that the cognitive learning mechanisms used in L2 learning are the same ones that have been committed over time to efficient and automatic L1 use. As a result, instruction informed by crosslinguistic influence takes as its starting point the prior experience learners bring to the task of L2 learning in order to address specific L2 learning difficulties. For example, if research

90

90  Instructed L2 Learning

evidence indicates that a lifetime of L1 processing subject–​object information has led to the development of automatic knowledge that word order can serve as a reliable cue for this meaning, then L2 learning that uses word order to express a different or additional meaning will most likely result in L2 learning difficulties. Not only is this a prediction of the UCM and other models of L2 learning, but this is a finding that has been reported in many empirical studies of L2 learning (see MacWhinney, 2005). Instruction that directly addresses this learning difficulty by targeting its underlying cause (e.g., effects of learned attention and blocking) has been shown to improve L2 learning outcomes. It is important to note, therefore, that when we discuss using L2 research evidence about crosslinguistic influence to support L2 learning, we are specifically referring to the development of pedagogical techniques that are sensitive to the prior language knowledge and experiences of learners. Such approaches to language instruction thus appeal to a type of usage-​based language instruction in the sense that the instruction is sensitive to the ways in which learners’ use of their other known languages can influence new language learning. In the remainder of this chapter, we review some of the ways in which instructional research has drawn on the prior linguistic knowledge and experience of learners in designing pedagogical techniques to address L2 learning difficulties.We focus on three strands of this research, concerning the learning of morphosyntax, vocabulary, and phonology, thus complementing our treatment of empirical studies of L2 learning presented in Chapter 3. Our approach starts by describing some of the ways in which researchers have addressed crosslinguistic influence in instruction before discussing the findings of these specific studies. This is helpful so that we can get a better idea of the rationale and motivation of specific instructional techniques without being influenced by the findings of a particular study. One reason for taking this approach (i.e., better understanding the motivation for the instructional design before talking about the findings) is that a study’s findings could be influenced by factors that do not directly bear on the instruction (e.g., sample size, appropriacy of outcome tests). To be clear, our aim is not to provide an exhaustive review of research on this topic, but to synthesize the main ideas in these lines of research and summarize what is currently known.

4.5  L2 Instruction and Morphosyntax In this section, we begin our review by focusing on studies that have investigated how instruction informed by empirical research evidence about crosslinguistic influence and L2 theory can address learning difficulties resulting from L1–​L2 differences. We focus on three lines of research that have investigated different aspects of morphosyntax: Spada et al. (2005) for possessive determiners in L2 English, Tolentino and Tokowicz (2014) for a range of morphosyntactic features in L2 Swedish, and McManus and Marsden (2017, 2018, 2019a, 2019b) for aspect in L2 French (see also McManus 2019b, 2021). In addition to examining different

90

91

90

Instructed L2 Learning  91

morphosyntactic foci, these studies also differ in that they investigate learning with different populations, namely experienced classroom learners, young children in immersion classroom contexts, and complete beginners. They also incorporated different instructional components to address the negative effects of crosslinguistic influence.

4.5.1  Explicit Information about L1–​L2 Differences Research by Spada and Lightbown represents an early line of investigation that explored the potential of instruction informed by speakers’ prior language knowledge and experiences to address longstanding L2 learning difficulties (see also Kupferberg, 1999; Kupferberg & Olshtain, 1996). This research builds on Spada and Lightbown’s earlier work in crosslinguistic influence and instructional effectiveness, including the potential for L2 research to inform instructional effectiveness (see Lightbown & Spada, 2013; Spada, 1997). In this instructional study, Spada et al. (2005) used explicit instruction to address longstanding learning difficulties among French speakers in the use of possessive determiners (e.g., his cat, her ball) and question formation in English. To illustrate the specific instruction developed by Spada et al. (2005), we focus on possessive determiners. Crosslinguistic influence was hypothesized to be a key explanation for the source of this learning difficulty because French possessive determiners (e.g., sa, son “her/​his”) agree with the gender of the possessed entity (e.g., a noun), but in English possessive determiners agree in natural gender with the possessor. In short, French and English do not use possessive determiners in the same ways, largely because French marks grammatical gender but English does not. This means that we can translate his as either sa or son depending on the grammatical gender of the possessed entity (e.g., his cat = son chat, but his mouse = sa souris). Critically, the choice of sa over son depends on the grammatical gender of the possessed entity in French, but not in English. Thus, even though both languages use possessive determiners, the rules for their usage are not the same. Given that explicit instruction can be an effective means for increasing learners’ awareness of a target feature (see Roehr-​Brackin, 2018; Spada & Tomita, 2010), Spada et al. hypothesized that explicit instruction designed to draw learners’ attention to the nature of L1–​L2 differences could address this learning difficulty. Spada et al.’s instructional design included explicit information about L1–​L2 differences followed by communicatively oriented practice that involved producing English possessive determiners. The following explicit information was provided to inform learners about how to use possessive determiners in English: Ask “whose is it?” If it belongs to a man or boy, use his. If it belongs to a woman or girl, use her. Spada et al., 2005, p. 210

92

92  Instructed L2 Learning

Spada et al.’s (2005) instruction was delivered to classroom learners aged 11–​12 years, who were not necessarily linguistically aware or able to use and interpret meta-​language. These learners were also enrolled in communicatively oriented language immersion programs in Canada in which attention to grammar and focus on form was not common. It was therefore appropriate for the instruction to be short, easy to understand, and useable (see also Kasprowicz & Marsden, 2018). In addition to classroom discussions, posters were fixed to classroom walls that repeated the explicit information with illustrations and example sentences. In terms of classroom activities that sought to apply and rehearse this newly provided explicit information, learners carried out meaning-​oriented practice requiring the use of English possessive determiners in a variety of games, including, for example, a guessing game where learners had to guess who was being described in sentences like “his hair is short and his T-​shirt is yellow”. Learners received corrective feedback during classroom activities that practiced the target feature. Using this design, Spada et al. (2005) set out to understand the relative effectiveness of explicit instruction with and without contrastive explicit information about L1–​L2 differences. This was achieved by including two classes that each received instruction about the target feature. One group included an explicit focus on L1–​L2 differences (contrastive group) and one did not (non-​contrastive group). The study included a pretest-​posttest design. Instruction was delivered during regular class time by “highly proficient speakers of English [… who] were ESL specialists and had taught intensive classes before” (Spada et al., 2005, p. 209), using classroom materials created by the research team. The instruction lasted approximately 3.5 hours over four weeks (30 minutes per day, three times per week). To evaluate the effectiveness of the instruction on learning, three tests were developed to assess learners’ knowledge and use of possessive determiners: a passage correction test, an oral production test, and meta-​talk interviews. A general proficiency test was administered after the intervention to determine how comparable the groups were in terms of listening and reading comprehension in L2. Both groups showed improvement in all tasks, indicating more accurate performance after instruction in terms of being able to (i) identify and correct possessive determiner errors as well as (ii) more appropriately use possessive determiners to describe sets of pictures. The meta-​talk interviews indicated an awareness of the rules for using possessive determiners. These results indicated promising findings about the effectiveness of instruction about L1-​L2 differences for reducing the negative effects of crosslinguistic influence. However, the research team noted that they were not able to draw conclusions about the extent to which instruction that included a contrastive component was more or less effective than instruction without. This is because discussion with the teachers after the study indicated that the non-​contrastive group most likely received instruction about possessive determiners in L1 and L2 outside of the experimental lessons. As a result, it is not possible to conclude that the non-​contrastive group did not receive instruction about L1–​L2 differences.

92

93

92

Instructed L2 Learning  93

Taken together, these results indicate some of the ways that instruction can incorporate information about L1 usage into L2 learning as a way to address specific learning difficulties resulting from crosslinguistic influence. Although the specific effects of the contrastive instruction remain unclear, this study can be seen as an important one for stimulating other lines of research that have sought to tailor instruction to address specific learning difficulties resulting from crosslinguistic influence.

4.5.2  Instructional Effectiveness for Different Types of Target Features Tolentino and Tokowicz (2014) used computer-​based language instruction, which is one way to address the limitations in Spada et al. (2005) as well as a way to enhance the comparability between the instructional conditions (see also Lucas, 2020; McManus & Marsden, 2017). Tolentino and Tokowicz (2014) investigated whether explicit instruction informed by previous L2 research about crosslinguistic influence and theoretical accounts of L2 learning (MacWhinney, 2005) could address specific L2 learning difficulties. In particular, the authors investigated “whether particular instruction methods were more effective for teaching the types of morphosyntactic features that least benefit from positive transfer: those that are dissimilar in L1 and L2 and those that are unique to L2” (p. 280). In other words, this study investigated connections between (i) type of crosslinguistic influence and (ii) type of instruction. To achieve this aim, three types of instruction were provided to English-​speaking learners of Swedish. Three different types of L2 features were selected in terms of their crosslinguistic similarities with English. First, a target feature was determined to be “similar” if it existed in English and Swedish and was used in the same way. Demonstrative determiner–​noun agreement was classified as “similar”, for example, since English and Swedish use suffixes on demonstratives and nouns for pluralization (e.g., där dockorna “those puppies”). Second, a target feature was identified as “dissimilar” if it existed in both languages but was used/​instantiated differently. Definiteness in noun phrases was identified as dissimilar because while English and Swedish both express this meaning, English expresses definiteness on determiners (e.g., a boy runs vs. the boy runs) but Swedish additionally indexes definiteness as a noun suffix (e.g., en pojke springer “a boy runs” vs. pojken springer “the boy runs”). Lastly, a target feature was identified as “unique to the L2” if it existed in Swedish but not in English. An example of this target feature is grammatical gender agreement among the definite article and the adjective because English has no grammatical gender but Swedish does (e.g., en tung docka, “a cumbersome puppy”). See Tolentino and Tokowicz (2014) for more information about these target features. Following the UCM, these three types of target feature were expected to lead to different types of learning difficulty and were expected to respond to instruction in different ways. Target features identified as “similar” were predicted to be

94

94  Instructed L2 Learning

the least difficult to learn because transfer from L1 would facilitate L2 use. In contrast, target features identified as “dissimilar” and “unique to the L2” were predicted to be more difficult since L1 transfer would not facilitate L2 learning. Furthermore, “unique to the L2” features were thought to be easier to learn than “dissimilar” because of “high cue strength”, meaning these features exhibited high availability and high reliability in the input (see Chapter 2). Three types of instruction were developed to understand which, if any, appeared the most effective for improving learning given the nature of these L1–​L2 similarities and differences.That is, might a specific instructional technique be better suited to address particular types of learning difficulty? This is one example of instructional research designed to understand the ways in which instruction might be tailored to address the nature of the learning difficulty (see also McManus & Marsden, 2019a). Three types of instruction were provided: Control, Salience, Rule & Salience. The control group was presented with a pair of Swedish sentences on a computer screen. Participants were instructed to “pay attention to any grammatical patterns and were monitored to ascertain that they repeated the sentence pair aloud once after hearing it pronounced” (ibid., p. 292). The authors described instruction in the Salience group as follows: The Salience group was exposed to the same sentences but structured in pairs that contrasted two instantiations of a given morphosyntactic feature and included bold blue highlighting of the morphemes at the point of agreement (e.g., en pojke springer, “a boy runs” vs. pojken springer, “the boy runs”) Ibid., p. 287 Only Swedish sentences were shown (i.e., without English glosses). The Rule & Salience group was exposed to the same Swedish sentences as in the Salience treatment (i.e., sentences that contrasted a specific target feature with bold highlighting) but with additional “metalinguistic explanations of the morphosyntactic rules underlying the L2 feature” (ibid.). This study included a pretest, three posttests and lasted eighteen days with a variety of outcome measures, including, measures of aptitude and an operation span test. Because all learners were complete beginners, the instruction included vocabulary training in addition to the different grammar treatments. All learners completed the same vocabulary training to ensure that everyone had been exposed to the same type of vocabulary. In the vocabulary training, an English word was presented aurally and visually on a computer screen. Each English word presented was then followed by its Swedish translation. Participants were instructed to listen to the word pairs and then repeat the Swedish translation aloud twice. All participants completed this training for 35 pairs of words. Grammar training was administered at three different points over the 18 days (days 1, 3, and 18) and lasted 40 minutes in total. Grammar learning was assessed using grammaticality judgment tasks that presented a Swedish sentence on the screen and either

94

95

94

Instructed L2 Learning  95

asked learners to judge its grammaticality using a four-​point scale (at pretest) or a two-​point scale (at the posttests). Correct/​incorrect feedback was provided at the posttests, but not at the pretest. Performance on the grammaticality judgment tasks showed that all conditions resulted in some type of learning. In particular, target features that were similar in L1 and L2 improved in all groups, suggesting that the provision of input may have been sufficient for learning target features that are similar in L1 and L2. In terms of target features that are different in L1 and L2 and absent in L1 but present in L2, the findings showed improvement in both the Salience and Rule & Salience groups. Although these results show positive impacts of instruction on L2 learning outcomes, it must be noted, however, that it is difficult to tease apart the specific contribution of the Salience and Rule & Salience instructional treatments since (i) corrective feedback was provided in the posttests and (ii) the judgment tasks used at pretest and posttest were designed differently (for discussion, see McManus, 2020). This means that improvement in the Salience and Salience & Rule groups after the instruction may have been a result of the corrective feedback provided in the posttest judgment tasks, a result of the instruction, or a combination of both. Despite these limitations, however, the study’s findings show some consistencies with previous research on the role of instruction in addressing L2 learning difficulties associated with the negative effects of crosslinguistic influence. In line with Spada et al. (2005), for instance, it seems that instruction that (i) draws learners’ attention to specific aspects of the learning difficulty, with highlighting and explanations, for instance, and (ii) provides some type of explicit information can play an important role in improving L2 performance. Indeed, it is important to note that providing learners with input alone appeared insufficient for the learning of target features with L1–​L2 differences. In addition, these findings indicate that instruction can also play a beneficial role for complete beginners of a new language. In sum, accumulating evidence to date suggests a positive role for using attention raising techniques, such as salience and/​or meta-​linguistic information, to increase learners’ awareness of crosslinguistically different target features.

4.5.3  Adding Practice in L1 and Explicit Information about L1 to L2 Instruction Building on growing consensus that providing explicit information about crosslinguistic differences can be one way to address the negative effects of crosslinguistic influence, McManus and Marsden (2017, 2018, 2019a, 2019b; McManus 2019a, 2021) investigated if increasing learners’ sensitivity to form-​ meaning mappings in L1 and L2 could address well-​documented learning difficulties with past time aspectual forms in L2 French. In contrast to Spada et al. (2005) and Tolentino and Tokowicz (2014), however, this research involved experienced classroom learners. One difference here, therefore, is that McManus and Marsden’s learners had developed considerable experience with French over the course of

96

96  Instructed L2 Learning

their L2 studies (approximately ten years), but still struggled to use the Imparfait to express imperfective viewpoint aspect meanings. The target feature was viewpoint aspect (Smith, 1997) in L2 French, which has been shown to be a late-​acquired feature (Bartning & Schlyter, 2004; Kihlstedt, 2002, 2015; see also Chapter 3), due to functional complexity and complex L1–​L2 differences in how L1 and L2 express the same meanings (Ayoun, 2013; Howard, 2005; McManus, 2013, 2015). Briefly, this target feature is understood to be difficult to acquire because of L1–​L2 similarities and differences in how English and French map the same aspectual meanings to forms. On the one hand, English and French both use verbal morphology to express the same viewpoint aspect meanings (e.g., ongoingness, elle jouait au foot “she was playing football”), but the precise meanings mapped to individual verbal forms are different. English maps habituality and perfectivity to one form (the Simple Past, see Table 4.1 for examples) and ongoingness to a separate form. In contrast, French maps ongoingness and habituality to one form (Imparfait) and perfectivity to a separate form (Passé Composé).This means that English and French do not map meaning to form in consistent ways (McManus, 2013, 2015; McManus & Marsden, 2017). This L1–​L2 difference is understood to be a critical source of the learning difficulty leading English-​speaking learners to use the perfective Passé Composé to express habituality and the Imparfait to express ongoingness because of how these meanings are expressed in English. Difficulty in using the Imparfait is well-​documented in the literature and represents a persistent problem even after considerable exposure to the target language (Howard, 2005; Kihlstedt, 2002, 2015). McManus and Marsden (2017, 2018, 2019a, 2019b; McManus, 2019a, 2021) investigated the extent to which explicit instruction about these form-​meaning differences could address this well-​documented learning difficulty among English-​ speaking learners of French. The design of the instruction was motivated by SLA research findings about the ways in which L1–​L2 differences influence the learning and use of the French Imparfait.Three instructional treatments were designed. First, TABLE 4.1  Form-​meaning Mappings for Viewpoint Aspect in French with English

Translations Viewpoint Language

Example sentence

Perfective French Elle a joué au foot (hier) English translation She played football (yesterday)

Aspect form Passé Composé Simple Past

Habitual

French English

Elle jouait au foot (chaque jeudi) She played football (every Thursday) She used to play football (every Thursday) She would play football (every Thursday)

Imparfait Simple Past Used to Would

Ongoing

French English

Elle jouait au foot (hier) She was playing football (yesterday)

Imparfait Past Progressive

96

97

96

Instructed L2 Learning  97

the “L2-​only” group received explicit information about the meanings expressed by French aspect forms (Imparfait, Présent, Passé Composé) followed by extensive comprehension practice that required learners to interpret the aspectual meaning of French sentences (e.g., elle jouait au foot “she was playing football” vs. elle a joué au foot, “she played football”).The “L2+L1” group received the same explicit information about French and practice interpreting French sentences, but this group also received two instructional components about L1: explicit information about L1 form-​meaning mappings for viewpoint aspect and interpretation practice of L1 sentences. These L1 components were added to increase learners’ awareness about how English and French express the same viewpoint aspect meanings. L1 practice was added so that learners engaged in the same type of practice as they were completing in L2 (i.e., interpreting aspect forms) and allowed them to apply the explicit information about English. While some previous research has included explicit information about L1 (e.g., Spada et al., 2005), no previous research has used practice in L1. As a result, this instruction was designed to address the negative effects of crosslinguistic influence by increasing learners’ awareness of how L1 expressed viewpoint aspect, using both explicit information and practice. The “L2+L1prac” group received the same instruction as the L2+L1 group, but without explicit information about L1. That is, this group received explicit information about L2 with interpretation practice of French and English sentences. Removing the L1 explicit information component allowed for its specific contribution to learning to be better understood. McManus and Marsden’s design of the explicit information was different from that of the Spada et al. (2005) and Tolentino and Tokowicz (2014) in that it sought to clarify the meanings of aspectual forms rather than provide rules for how to use them.That is, then, it focused on the connections between form and meaning. It was also designed following a concept-​based approach (Tyler, 2012) in that the explicit information introduced and clarified aspectual meanings (habituality, ongoingness) before discussing how these meanings are expressed in French (L2-​ only, L2+L1prac) or in French and English (L2+L1). This involved making the conceptual meaning clear before introducing the specific verbal forms used to express these meanings. For example, a short video of a man eating an apple was used to illustrate the concept of ongoingness (see Figure 4.2). The video showed the apple being eaten bite by bite, but the apple is never fully eaten. After watching this video, learners were asked how they would describe what they had just seen. Learners’ attention was drawn to the unfinished event and the ways in which English and French express such an event (for full description of the explicit instruction, see McManus and Marsden, 2017). In terms of the practice, all learners completed the same comprehension practice of L2 sentences. The amount of comprehension practice conducted in the L1 was small for both the L2+L1 and L2+L1prac conditions. The L1 instructional components followed the same design principles as the L2 explicit information and the L2 practice. In this way, learners had the opportunity to apply the explicit

98

98  Instructed L2 Learning

FIGURE 4.2 Video

Stills Depicting Concept of Ongoingness from McManus and

Marsden (2017)

information about L1 to the interpretation of L1 sentences. Table 4.2 provides an overview of the treatment components received in the L2-​only and L2+L1 instructional conditions for session one about ongoingness in present versus past (for more detail, see McManus & Marsden, 2017). The study design included a pretest, immediate posttest, and a delayed posttest (administered six weeks later) and included a variety of outcomes measures: self-​ paced reading and judgment tasks in reading and listening (McManus & Marsden, 2017, 2018), oral production of picture-​based narratives (McManus & Marsden, 2019a), and an oral sentence completion task with source of knowledge probes (McManus, 2019a). Other analyses also examined performance during the training. For example, McManus and Marsden (2019b) analyzed how the accuracy, speed, and stability of L2 performance changed over the course of the practice. In McManus (2021), the focus was on understanding in what ways switching between L1 and L2 in the “L2+L1” and the “L2+L1prac” groups contributed to learning. Overall, the findings indicated important differences among the treatments. Here we briefly discuss two sets of findings: (i) knowledge from the outcomes tests at the pretest and posttests and (ii) knowledge from analyses of performance during the intervention. First, in terms of L2 performance at the pretest and posttests, results showed between-​group differences that related to the type of instruction received. The judgment tests showed improvement for all treatment groups between pretest and posttest, but these gains were maintained at the delayed posttest in the “L2+L1” group only. A similar pattern of results was found in the self-​paced reading task, where improvement at the delayed posttest was detected in the “L2+L1” group only (McManus & Marsden, 2017, 2018). In terms of oral production, only the “L2+L1” group maintained gains at the delayed posttest. These results indicated that while the instruction was immediately beneficial in all treatment groups, this improvement was only sustained for instruction that included explicit information about L2 and L1 and comprehension practice of L1 and L2 sentences.

98

9

Instructed L2 Learning  99 TABLE 4.2  Description of the Treatment Components Used in Session 1: Ongoingness

(Present vs. Past)

Pre-​practice explicit information

L2 components

L1 components

[Watch a six-​second video of man eating an apple. The apple is never fully eaten.] To describe this you could say:

[Same video as L2-​only treatment]

Il mange une pomme Or Il mangeait une pomme The difference between these two is: Il mange = ongoing action RIGHT NOW Il mangeait = ongoing action IN THE PAST The ends of the verbs distinguish between an ongoing action in the present versus past e.g. [Four verbs presented in pairs, aurally and in writing]:

98

To describe this you could say: He is eating an apple Or He was eating an apple The difference between these two is: “he is eating” = ongoing action RIGHT NOW “he was eating” = ongoing action IN PAST To identify ongoing meaning in the present versus the past, you need to focus on the auxiliary. Look/​listen out for “is” or “was” to indicate whether it is an ongoing action taking place RIGHT NOW (present) or it is one IN THE PAST (past).

Présent Imparfait RIGHT NOW IN PAST regarde [ʀəgaʀd] regardait [ʀəgaʀdɛ] Practice

96 French items (48 listening, 48 reading). Aim: Identify whether an ongoing event is taking place:

Additional 32 English items (16 listening, 16 reading). Aim: Identify whether an ongoing event is taking place:

“MAINTENANT” (right now) or “DANS LE PASSÉ” (in the past) Example (English glosses not provided): Il… (1) fait du shopping (“is shopping”) (2) faisait du shopping (“was shopping”)

“RIGHT NOW” or “IN THE PAST” Example: He… (1)  is eating a sandwich (2)  was eating a sandwich (continued)

01

100  Instructed L2 Learning TABLE 4.2  Cont.

Explicit information given immediately after incorrect responses during practice

L2 components

L1 components

After incorrectly responding “MAINTENANT”:

After incorrectly responding “RIGHT NOW”:

“NOTE:The IMPARFAIT expresses an ongoing event DANS LE PASSÉ, not an ongoing event taking place MAINTENANT”

“The present tense in English (“is +ing”) and in French expresses the same meaning: ongoing action taking place RIGHT NOW”

After incorrectly responding “DANS LE PASSÉ”:

After incorrectly responding “IN THE PAST”:

“REMEMBER:The present tense in “The past tense in English (“was French expresses an ongoing event +ing”) is the same as the taking place MAINTENANT; not an Imparfait in French (-​ait).They ongoing action DANS LE PASSÉ” both express an ongoing action IN THE PAST”

Second, in terms of performance during the practice, an often-​neglected analysis in instructional research, McManus and Marsden (2019b) reported that the speed, accuracy, and stability of performance on the L2 items changed over time. This analysis was carried out by examining the accuracy and speed of performance on each L2 practice item over the course of the entire instruction. Perhaps most interestingly, these analyses showed that among learners in the “L2+L1” group, accuracy was initially less accurate and slower when compared to the other treatments, but performance in the “L2+L1” group improved over time. However, in the “L2-​only” and “L2+L1prac” conditions, accuracy and speed were mostly constant over the course of the practice. These results indicated that the nature of the instruction had different effects on L2 performance over time, thus complementing performance on the posttest measures but also providing fine-​ grained information about the trajectories of learning that led to that posttest evidence. Findings that L2 performance was initially slower and less accurate among learners who received explicit information about L1 suggested that the provision of instruction about L1 led to different learning trajectories. L2 performance in the “L2+L1” group was much slower and less accurate at the beginning of the practice but L2 performance became more accurate and faster over time. McManus (2021) additionally examined how switching between L1 and L2 may have contributed to these findings, including the extent to which language switching may have been moderated by type of explicit information received.This analysis compared language switches in the “L2+L1prac” and “L2+L1” conditions

01

10

01

Instructed L2 Learning  101

(since only these groups completed comprehension practice in L2 and L1). The results showed that the accuracy and speed of L1–​L2 switches in the “L2+L1prac” condition remained mostly constant over time. In the L2+L1 condition, however, accuracy was lower and switch speed was slower at the beginning of each practice session, but both signatures of performance improved over time (i.e., performance became more accurate and quicker with greater amounts of practice). Furthermore, these analyses indicated that the direction of the switch was important. In the “L2+L1” group, switching from L2 into L1 was initially slow but got quicker with increasing amounts of switching practice, whereas switching from L1 into L2 was largely constant. In line with our earlier discussions from Chapters 2 and 3 about inhibitory control and language switching, this finding was interpreted as follows: switching from L2 into L1 is more difficult than switching from L1 into L2. This is because switching into L2 requires inhibition of L1 and then switching back requires the previously applied inhibition to be reversed. These results indicated that explicit information about L1 plus practice benefitted performance because the L1 explicit information clarified L1–​L2 differences, and the practice allowed these processing differences to be implemented. Taken together, these analyses of performance during the practice suggest that L2+L1 learners performed differently during the training (initially less accurate and slower performance that improved as a function of the practice), which then resulted in greater learning gains at the posttests. In sum, this line of research indicates that instruction tailored to the nature of the crosslinguistic learning difficulty can play an important role in improving L2 performance. The findings suggest that explicit information about L1 was an important component in understanding L2 performance in a variety of outcome measures. Indeed, as discussed in McManus (2019a), awareness of form-​meaning mappings in L1 played an especially important role, indicating that increasing learners’ sensitivity to the nature of form-​meaning mappings in L1 and L2 is critical for addressing L2 learning difficulties resulting from crosslinguistic influence.

4.5.4 Summary In this section, we have reviewed three lines of research that have each addressed the negative effects of crosslinguistic influence using explicit information and practice, but with important differences. In Spada et al. (2005), the pre-practice explicit information was contrastive, but in McManus and Marsden (2017) it was non-​contrastive. By contrastive versus non-​contrastive, this means that Spada et al. (2005) drew explicit contrasts between L1 and L2, but McManus and Marsden (2017) delivered the pre-practice explicit information about L2 and L1 separately. Contrastive explicit information was used to draw learners’ attention to specific similarities and differences between L1 and L2. Non-​contrastive explicit information serves the function of providing information about the target features

012

102  Instructed L2 Learning

in L1 and L2, specifically the ways in which they express the same meanings. In Tolentino and Tokowicz (2014), no explicit information about L1 was provided. Although the pre-practice explicit information in McManus and Marsden (2017) was non-contrastive, learners may have drawn their own contrasts. The purposes of including non-​contrastive EI was that the source of the learning problem hypothesized to be due to a lack of understanding how their L1 expressed viewpoint aspect, rather than a lack of awareness of how L1 and L2 are different. It seems that a relevant question going forward could be whether explicit information that includes a contrastive component is helpful for L2 learning or whether clarifying the form-​meaning connections in the respective languages is just as effective. In terms of the practice, this was meaning-​based in Spada et al. (2005) because learners used the target forms to express specific communicative functions (e.g., identifying someone in the classroom). In Tolentino and Tokowicz (2014), learners repeated the sentence stimuli, which did not require learners to attend to the meaning implications of the morphosyntactic features in the sentences. In McManus and Marsden (2017, 2019a) practice was meaning-​based because it required learners to select the meaning of a sentence using the fixed options provided. For example, in one of the practice sessions, on seeing elle jouait au foot “she was playing football”, learners had to select whether the sentence expressed an event from the past or something taking place right now. In Spada et al. and Tolentino and Tokowicz (2014), practice was in L2 only, but additional L1 practice was included in McManus and Marsden (2017). The inclusion of L1 practice was helpful arguably because it allowed learners to apply the explicit information about L1 to the interpretation of L1 sentences, which forced learners to attend to the meaning implications of L1 sentences. Taken together, this body of research indicates that instruction that draws learners’ attention to aspects of the L1 that cause learning difficulties is helpful for improving L2 performance in a range of skills, including in offline and online performance. In the following sections we review the extent to which these ideas about crosslinguistic influence and L2 instruction have been implemented in studies of vocabulary and phonology.

4.6  L2 Instruction and Vocabulary In this section, we review three lines of research that have addressed L2 learning difficulties for vocabulary by integrating instruction about the L1: Laufer and Girsai (2008), Zhao and Macaro (2016), and Pulido and Dussias (2020). Participants in these studies included Hebrew-​speaking learners of English, Chinese-​speaking learners of English, and English-​ speaking learners of Spanish, and the target features under study included single words and multi-​word units, concrete and abstract words, and phrasal verbs (or verb–​noun collocations). As in the previous section, our aim is to understand each study’s approach to instruction and

012

103

012

Instructed L2 Learning  103

the specific ways that speakers’ prior knowledge and experience was used as an instructional tool to support L2 learning.

4.6.1  Contrastive Analysis and Translation Laufer and Girsai (2008) extended an important line of research in language teaching that investigated why providing classroom learners with comprehensible input and meaning-​focused tasks alone did not appear to reliably improve learners’ L2 abilities (Lyster, 1994; Swain, 1988). Discontent with instructional approaches at that time led to a renewed interest in the potential benefits of encouraging learners to attend to form (R. Ellis, 2001; Long, 1991; Norris & Ortega, 2000). However, even though form-​focused instruction had been a large focus of research interest in grammatical learning, few studies had explored this approach for vocabulary learning. Laufer and Girsai (2008) investigated whether form-​focused instruction could also be used to improve L2 learners’ use and knowledge of vocabulary. At that time, other methods had been used to teach vocabulary, including, for example, glosses, dictionary use, negotiation of meaning, but findings for these methods had led to mixed conclusions. For example, some instructed L2 research examined vocabulary learning by comparing text reading with and without dictionary support (Luppescu & Day, 1993; Knight, 1994). These studies found that providing learners with a dictionary facilitated L2 vocabulary learning. In other words, providing some type of instructional support in addition to exposure benefitted vocabulary learning. Laufer and Girsai (2008) extended this line of research by examining whether L2 learning could be facilitated by encouraging learners to draw on their existing language knowledge as a learning resource (in contrast to using dictionaries or other learning devices). In short, Laufer and Girsai (2008) developed a “kind of instruction which leads to learners’ understanding of the similarities and differences between their L1 and L2 in terms of individual words and the overall lexical system” (p. 696). Furthermore, L2 instruction that draws on learners’ existing L1 knowledge was proposed as a means to address “pervasive influence of L1 on L2 learning” (p. 697), especially when concepts are expressed differently in L1 and L2, such as with phrasal verbs in L2 but not in L1 (Granger, 1998; Nesselhauf, 2003, 2005; see also Chapter 3) or concepts that are expressed with multi-​word units in L2 but not in L1 (Jiang, 2002, 2004). Indeed, these types of differences where L1 and L2 express the same meanings in different ways is known to constitute a major learning obstacle in L2 learning, as discussed in Chapter 3. To address this well-​documented learning problem, Laufer and Girsai (2008) compared the effectiveness of three types of instruction on L2 learners’ abilities to (i) provide the form of the target word in response to their L1 translation equivalents (active recall) and (ii) provide the meaning of the target words (passive recall). The target words were single words (e.g., glean, opulent, lavish) and collocations (e.g., settle scores, present a problem, derive pleasure), selected because they

104

104  Instructed L2 Learning

were unknown to the participants (as indicated through pretesting; see Laufer & Girsai (2008) for full list of target items). The instruction was carried out during regular classroom time and was described as being incidental (or implicit), meaning that learners were not told that they were participating in a vocabulary learning study. The three different instructional conditions were all preceded by the same reading task titled “Clinton’s biography”. Participants were instructed to read a passage of text and to answer 13 true or false statements. No dictionaries or reference materials were provided, but participants could ask the teacher for the meaning of a word. This reading passage stage was followed by three different types of instructional treatment. In the meaning-​focused instruction condition, participants received two tasks. First, a reading comprehension task that included specific questions about the text that required rereading (e.g., why is Arkansas mentioned in the text?). Second, a pair/​ group discussion task related to the text. In the non-​contrastive form-​focused instruction condition, participants com­ pleted two form-​focused tasks. First, a multiple-​choice, meaning recognition task in which participants were given a multi-​word unit and were asked to select its meaning from three options. For the multi-​word unit hit the headlines, for instance, learners selected one of the following three options: became hot news, caused headaches, was first in line. Second, a fill-​in-​the-​blanks exercise provided learners with a “word bank” and they had to fill in the blanks using the appropriate words. The teacher discussed the answers in a class discussion. Lastly, in the contrastive analysis and translation condition, participants com­ pleted two translation tasks and received brief explicit contrastive instruction. First, learners translated sentences from English (L2) into Hebrew (L1), such as will the book meet the expectations of the publishers?. After the teacher had collected the translations, learners then translated the same sentences from Hebrew into English. The translations could be completed individually or in small groups. After completing the task, the teacher provided corrective feedback in a class discussion of the translations, which also included brief explicit contrastive instruction about L1–​L2 differences: …in the case of collocations, [the teacher] pointed out that although the nouns had equivalent translations in Hebrew, the verbs that collocated with them were totally different (e.g. hit the headlines is “break into headlines” in Hebrew, meet expectations is “answer expectations”). [The teacher] suggested that learners should be cautious not to provide automatic verb translations, as this would result in unacceptable English collocations. Laufer & Girsai, 2008, p. 705 On the day after instruction, learners were tested on their ability to translate Hebrew words and phrases into English (active recall). Then, after a short delay,

104

105

104

Instructed L2 Learning  105

learners provided Hebrew translations or English explanations of the same words and phrases (passive recall). The same tests were then administered one week later. Results showed between-​ g roup differences both shortly after instruction (posttest) and then one week later (delayed posttest), with higher accuracy in the “contrastive analysis and translation” group. Performance in the “non-​contrastive focus on form” group was the least accurate of all the groups. Between-​g roup comparisons indicated that the “contrastive analysis and translation” group outperformed the other two instructional conditions in both single word and multi-​word items at both the posttest and delayed posttest. Because a pretest was used to determine words that learners did not know, which then was the focus of the instruction, these findings indicate that engaging learners in translation activities followed by explicit information about L2 and L1 was the most beneficial for learning both single word and multi-​word units, known to be difficult to learn due to crosslinguistic influence. Although some benefits were found for form-​ focused instruction, these were mostly found in passive recall, thus suggesting a benefit for receptive language knowledge. Because explicit instruction incorporating translation appeared to be the most beneficial for L2 learning, Laufer and Girsai suggest that one explanation for this finding could be that the translation activities plus corrective feedback drew learners’ attention to form-​meaning differences between L1 and L2 in ways that the L2-​focused tasks did not. Indeed, this is similar to the claims made in McManus and Marsden (2017, 2018) in that practice involving L1 was more effective than L2 practice only for this target feature because it required learners to process L1 sentences in ways that they might not regularly do (i.e., think about the meaning implications of L1 verb forms, see also McManus, 2019a, 2021). While findings like these do appear to be contributing to a developing narrative that using meaning-​focused tasks to develop awareness and sensitivity to how L1 and L2 are similar and/​or different, it should also be noted that the outcome measures used in Laufer and Girsai (2008) involved translation only. It is possible that more accurate performance was found in the “contrastive analysis and translation” condition because these learners had completed practice activities using translation whereas the other instructional conditions included no translation. An outcome task that did not involve translation is one way to address this point, as would analysis of performance during the practice activities (as in DeKeyser, 1997; McManus, 2021; McManus and Marsden, 2019b). Further research is needed to determine to what extent instruction involving contrastive analysis and translation benefits learners’ abilities to translate only or whether this type of instruction can also lead to improved L2 performance in general.

4.6.2  Instruction Involving Translation and L2 Description The focus of Zhao and Macaro’s (2016) study was to investigate whether the learning of abstract and concrete nouns may be influenced by type of explanation.

016

106  Instructed L2 Learning

Abstract and concrete nouns were investigated because of differences in their learnability. Abstract nouns (e.g., nostalgia, wholesome) are understood to be more difficult to learn and process than concrete nouns (e.g., chestnut, wardrobe). This is because a concrete word like dog refers to a perceptible, material entity that an adult speaker is more likely to have experienced, thus making it more imageable (Connell & Lynott, 2012; Reed & Dick, 1968). Imageability and perceptual experience are understood to be factors that can explain why some words are more difficult to learn than others. One consequence of this understanding is that speakers are hypothesized to draw on different learning mechanisms in learning abstract versus concrete nouns. Zhao and Macaro investigated this question by comparing the effectiveness of different types of explanations on vocabulary learning. The researchers examined in what ways providing translation equivalents from L1 versus explanations in L2 contributed to word learning. The use of translation equivalents was designed to investigate whether bringing learners’ prior knowledge and experience into the classroom could be used as a helpful tool to promote word learning. Using a pretest, immediate posttest, delayed posttest design, the study included two intervention groups and a control group. Instruction in both intervention groups included a reading activity followed by a teacher-​led group activity that included presentation of the target words. First, the reading activity involved learners working with texts adapted from Time magazine as well as completing a series of tasks (e.g., fill-​in-​the-​blanks) that used the words and phrases from the text. Critically, the reading activities included no instances of the target vocabulary items. After the reading exercises, target vocabulary items were presented in class using PowerPoint slides.The meanings of the words were explained through teachers’ L1 use or their L2-​only descriptions. For example: Target word: Medal L1 use: 奖牌 L2 description: Medal is a small metal thing. It is given to sports winners. Target word: Unparalleled L1 use: 绝无仅有的 L2 description: Unparalleled means having no equal, better or greater than any other. For example, he is unparalleled in this field. Examples from Zhao & Macaro, 2016, p. 96 The control group completed the same reading activity but received no explanations for the target words. Vocabulary knowledge was assessed using a modified version of the Vocabulary Knowledge Scale (Paribakht & Wesche, 1993), in which participants select one of the following options for each target word:

016

107

016

Instructed L2 Learning  107

I. I haven’t seen this word before. II. I have seen this word before, and I think it’s related to the category _​_​_​_​_​_​_​_​e.g., A. Food B. Art C. Action D. Emotion III. I know this word. It means _​_​_​_​_​_​_​_​_​_​_​_​_​_​ (Write your answer either in Mandarin or in English) From Zhao & Macaro, 2016, p. 86 Performance on the Vocabulary Knowledge Scale test before the intervention (pretest), immediately after the intervention (posttest), and then one week later (delayed posttest) showed: that, in both immediate and delayed recall of the target words’ meanings, the two intervention groups significantly outperformed the [control] group, and the L1 use group significantly outperformed the L2-​only use group in terms of learning concrete and abstract words. Zhao & Macaro, 2016, p. 89 Although accuracy at the delayed posttest had decreased from the posttest in the intervention groups, performance at delayed posttest was still more accurate than at the pretest. Performance in the control group showed limited change over time. Taken together, vocabulary items appeared to be better learned when teachers used L1 equivalents rather than L2 explanations. The authors suggested that L2 explanations can be more complex to process when compared to providing corresponding L1 items, especially for abstract nouns (e.g., nostalgia). This is because the L2 explanations themselves require processing and decoding and then a meaning must be inferred from the explanation. Providing the L1 equivalent, in contrast, helps establish a clearer link between an existing conceptual representation and the new L2 word. These findings suggest, therefore, that drawing on learners’ existing language knowledge and experience can facilitate the learning of both abstract and concrete words. These findings therefore make an important contribution to approaches to language instruction that draw on learners’ existing language knowledge and experience, including for abstract nouns which are well documented to be difficult to learn. Notwithstanding these insights, it should be noted that only one test of language knowledge was included in the study and it was the same test used at each of the pretest and the posttests. Completing the same test on three occasions within a short period of time may have promoted further learning by itself. Although a lack of improvement in the control group would suggest that the instruction

018

108  Instructed L2 Learning

was the likely contributor to the study’s findings, it should also be noted that only productive vocabulary knowledge was tested. Benefits in receptive knowledge may have been developed, especially since instruction involved presentation only without any production practice. Further research building on this study can develop our understanding about the usefulness of L1 translation as a potentially effective means of vocabulary instruction.

4.6.3 

Competition-​based Instruction

Pulido and Dussias (2020) investigated instructional effects on L2 collocation learning. In line with Gries (2013), collocations were defined as “combinations of words that are associated due to frequent co-​occurrence” (Pulido & Dussias, 2020, p. 652), such as “run a business” but not “carry a business” in English. L1–​L2 differences in the nature of collocational knowledge plays an important role in L2 learning (e.g., Granger, 1998; Nesselhauf, 2003, 2005), especially when the same meaning in L1 and L2 is achieved by combining different lexical items (e.g., “launder money” in English but blanquear dinero “whiten money” in Spanish). One consequence of this is that collocational knowledge in L1 and L2 competes for selection, thus requiring the application of language selection mechanisms to manage and regulate competition (e.g., Green, 1998, see also Chapters 2 and 3). Pulido and Dussias (2020) investigated this issue by targeting the cognitive mechanisms (inhibitory control) underlying lexical processing to understand whether this method reduced the negative effects of crosslinguistic influence in L2 vocabulary learning. The study involved a pretest, posttest, delayed posttest design that included two instructional conditions, an “L1-​interference” condition and an “unrelated” condition. Participants were Spanish-​speaking learners of English. Everyone completed the same “familiarization phase” (lasting approximately 20 minutes) but did different types of practice.The familiarization phase started by presenting a Spanish collocation on a computer screen (e.g., llevar negocio), which was followed by an equivalent collocation in English (e.g., run a business). Learners were requested to repeat the English collocation out loud and then type it out. Corrective feedback was provided on the accuracy of learners’ typing of the English collocation. In a second phase of the familiarization training, learners followed the same procedure but had to recall the Spanish equivalent of the English collocations. This process was completed for 45 different collocations. The practice involved forced-​choice trials that required selection of the target verb in the collocation. Each trial presented two verbs (the target and a distractor) and the associated noun in the collocation. The difference between the instructional conditions was that the distractor verb in the “L1-​interference” condition was an English translation of the verb used in the corresponding Spanish collocation (e.g., run –​carry –​business), whereas the distractor verb in the “unrelated” condition was semantically unrelated to the collocation in either language (e.g.,

018

019

018

Instructed L2 Learning  109

run –​touch –​business). Participants made a selection by saying out loud the appropriate verb. For example, in the “run –​touch –​business” trial, learners said aloud “run”. Each trial was presented nine times, totaling 405 trials per session. Each trial provided feedback by presenting the correct verb following each trial, irrespective of whether the response was correct or incorrect. Collocational knowledge was assessed at pretest, immediate posttest, and delayed posttest using a multiple-​choice test with confidence ratings. In this task, learners saw a Spanish collocation followed by the corresponding English collocation with the verb removed. Learners had to select the appropriate English verb from a list of four options. Participants were also asked to indicate the certainty of their responses using a five-​point scale (1 = no knowledge, 5 = certainty in the response): llevar un negocio -​ _​_​_​_​_​_​_​_​_​_​_​_​ a business (a) run  (b) carry  (c) walk  (d) bring From Pulido & Dussias, 2020, p. 658 Overall, the findings showed that the “L1-​interference” condition led to more accurate and faster performance in the multiple-​choice task, especially for collocations that were different in L1 and L2. According to the authors, “inducing interference through L1-​related distractors produced significantly greater accuracy for incongruent collocations” (p. 662), suggesting that requiring learners to attend to L1-​related distractors improved L2 abilities to a greater extent than attending to unrelated distractors. Learners who did not attend to L1-​related distractors showed fewer learning gains, indicating that incorporating some type of crosslinguistic effect into the training appeared to be beneficial for learning. Taken together, these findings build on a growing line of research involving incorporating different aspects of the L1 into instructional designs to reduce the negative effects of crosslinguistic influence. In this study, this was achieved by requiring learners to attend to translation equivalents in L1 and L2 for the same collocation. By actively selecting one verb over the other, learners likely engaged in some form of conflict management and resolution, which resulted in a learning payoff. Although other types of language assessment would have been helpful for understanding the nature of learners’ knowledge abilities (e.g., production, online processing), including tasks that did not induce a translation component (see also Laufer & Girsai, 2008), there is growing agreement that L2 learning can be enhanced by drawing on speakers’ L1 knowledge sources in different ways.

4.7  L2 Instruction and Phonology In this section, we review instructed research in L2 phonology that has drawn on evidence and theory about crosslinguistic influence to support L2 learning.

10

110  Instructed L2 Learning

In contrast to our discussions about morphosyntax and vocabulary (sections 4.5 and 4.6, respectively), fewer instructional studies of L2 phonology have been conducted, especially in addressing learning difficulties created by crosslinguistic influence.This observation is well noted in reviews of the field, indicating a broader concern that much less is known about instructional effectiveness and L2 speech in general (see Kennedy & Trofimovich, 2017; Loewen, 2020). To date, a number of explanations have been proposed to explain this paucity of instructional L2 research about phonology, including, for example, a strong instructional focus in the 1970s and 1980s on L2 reading and writing, the influence of Krashen’s (1985) ideas about the ineffectiveness of instruction, in general, as well as Purcell and Suter’s (1980) study that claimed no effect of instruction on accent (see Derwing & Munro, 2015). In addition, L2 speech researchers in the field of applied linguistics have been engaged in longstanding debates about the appropriate norms and targets in L2 phonology research, discussed in Levis (2005, 2020) as the nativeness principle and the intelligibility principle (see also Derwing & Munro, 2015). As a result, one reason for comparatively fewer studies of instructional studies in L2 phonology could be due to ongoing discussions about what the targets of this research should be. Given that concepts such as accentedness, intelligibility, and comprehensibility feature prominently in this line of research, it will be useful for us to define these terms, especially given that different studies have used these terms in different ways (for review, see Thomson, 2017). On the one hand, accentedness refers to L2 speech that “differs in some noticeable respects from native speaker pronunciation norms” (Munro & Derwing, 1995b, p. 289). Intelligibility and comprehensibility, on the other hand, are about the decoding or understanding of a message: intelligibility has been defined in terms of “the extent to which a speaker’s message is actually understood” (Derwing & Munro, 1995a, p. 76), while comprehensibility has more to do with how easy or difficult it is for a listener to understand a speaker’s message. As Thomson (2017) makes clear, however, many different definitions and operationalizations of these concepts exist, including cases where these terms have been interchangeably used (e.g., Isaacs & Trofimovich, 2012). Indeed, as recent research indicates, understanding how distinct these concepts are as well as the ways in which they may be related represents an active line of L2 research (e.g., Huensch & Nagle, 2021; Kennedy & Trofimovich, 2008; Munro & Derwing, 1995a, 1995b). Given this context, discussions about what the targets of phonology instruction should be as well as debates about key concepts in the field (e.g., intelligibility and comprehensibility) have shaped instructed L2 phonological research in important ways.This seems particularly apparent when compared to research in L2 morphosyntax and vocabulary, for example. Nonetheless, a small but informative body of research has examined this topic with reference to drawing on learners’ L1 experience and knowledge as a pedagogical tool to improve L2 learning outcomes,

10

1

10

Instructed L2 Learning  111

focusing on word and syllable structure/​timing (Crowther, 2015), segmentals (Saito, 2011), and consonant phonemes and allophones (Kissling, 2013).

4.7.1  Explicit Information about Word Stress in L1 and L2 Given the importance of accentedness, comprehensibility, and intelligibility in contemporary L2 phonology research, one line of instructional research has investigated the extent to which instruction can facilitate the perception of L2 speech. That is, can instruction target specific aspects of L2 speech to improve the comprehensibility and intelligibility of messages? Trofimovich and Isaacs (2012) is one example of this research that examined which linguistic features can make L2 speech more difficult to understand. In that work involving French-​speaking learners of English, different linguistic properties of speech were examined to understand how they contributed to comprehensibility and accentedness. One linguistic artefact found to be particularly influential was word stress error. This occurs when a speaker places stress on a different part of a word than would typically be expected (e.g., UN-​ha-​ppy instead of un-​HA-​ppy). Trofimovich and Issacs suggested that word stress errors can be caused by the different rhythmic properties of a speaker’s languages (see also Ladefoged, 2001). In this language pairing, English and French have different rhythmic properties. This means that when a syllable is stressed in English it receives greater energy than an unstressed syllable, which, in turn, leads to length differences between stressed and unstressed syllables. Such a contrast does not exist in French. A similar L1–​L2 difference has also been reported in other languages (e.g., Japanese) as an important source of L2 learning difficulty (see Saito, Trofimovich, & Isaacs, 2016). Crowther (2015) examined the extent to which explicit instruction could address this learning difficulty, understood to be a negative effect of prior language knowledge and experience. In this study, participants were Japanese-​speaking learners of L2 English. L1 Japanese speakers were selected because the structure of Japanese syllables differs from English. One reason for this is that Japanese is either V or CV (but English can range in complexity, e.g., V, CV, CCV, VCC). When Japanese syllables (or mora) are combined in a word, their resulting prosodic pattern includes no stress contrast among syllables. As such, word stress in Japanese (as in French) is different from English. The study design included a pretest, posttest, and delayed posttest using a within-​ subject design (i.e., no control or comparison group was included). Instructional effectiveness was assessed using word stress error rate as well as ratings of accentedness and comprehensibility. In terms of the instruction, two sets of materials were created that provided learners with explicit information about similarities and differences between Japanese and English. First, a four-​page worksheet was used that explained structural differences between English and Japanese for word and syllable structure and timing. This was used to “raise learners’ awareness of how structural differences

12

112  Instructed L2 Learning

could negatively affect production” (p. 33). Handouts were provided that listed common consonant clusters in word-​ initial and word-​ final position (from Eckman, 1991). Second, learners received a worksheet that identified differences between content and function words in English and the ways in which these types of words can influence stress within a sentence. In addition to the explicit information, learners completed practice in perception and production. In the perception practice, learners first identified whether an epenthetic vowel was present or not within a word. Participants were then asked to identify syllables that featured word stress and whether the word stress was correct or not. For the production practice, three types of activities were provided. First, learners read a paragraph that included words with epenthetic vowels within and at the end of words. Then, learners eliminated the epenthetic vowels from the paragraph and read it to a partner. The partner was tasked with answering four comprehension questions connected to the reading. Second, learners had to correct syllable stress in words with two, three, four, and five syllables. Learners then checked their answers and practiced saying the words aloud. Third, learners practiced the stress of content and function words. Learning was assessed using a spontaneous speech task that required participants to provide two pieces of information (e.g., what are you doing in Canada? and why have you chosen Canada?). Analysis was based on a 35–​45-​second speech sample from the spontaneous speech, with each sample being coded for word stress error rate (Trofimovich & Isaacs, 2012). Speech samples were additionally rated for accentedness and comprehensibility by three L1 speakers of English. Results showed that word stress error rates reduced over time, indicating a positive effect of awareness-​raising instructional techniques on L2 performance. The difference between pretest and posttest scores indicated a medium-​sized benefit for the instruction. The instructional benefits were also maintained two weeks later at the delayed posttest. In terms of accentedness and comprehensibility ratings, which indicate more global improvement resulting from the instruction, no meaningful change was found. The finding that the training appeared to offer few benefits for more general measures of performance (e.g., accentedness, comprehensibility) suggests a more localized impact of the training that improved word stress but not more general language abilities. In sum, this small-​scale study suggests some interesting avenues for further research that could be used to develop and better understand some of these initial findings. We should be careful about inferring too much due to a number of methodological choices in this study (e.g., no comparison group, small sample size), but these findings could potentially suggest benefits of instruction about crosslinguistic influence for improving L2 speech abilities. In the next two sections, we review two other studies about L2 phonology, but it should be noted that neither study specifically examined the effects of instruction about L1 in L2 learning. What they did do, however, was incorporate

12

13

12

Instructed L2 Learning  113

instruction about aspects of L1 into a broader treatment designed to support L2 phonological development.

4.7.2  Crosslinguistic Identification and Discrimination Training Also with a focus on Japanese, Saito (2011) studied eight English-​ specific segmentals /​æ, f, v, θ, ð, w, l, ɹ/​, known to be difficult for Japanese speakers to produce (Flege, 2003; Piske et al., 2001). One explanation for this learning difficulty in English is that these segmentals do not exist in Japanese. To address this learning difficulty created by L1–​L2 differences, Saito provided segmental-​ based instruction that focused on English sounds well documented to be difficult for Japanese speakers to learn. Although some connections between Saito (2011) and Crowther (2015) are evident in that both studies investigated pronunciation instruction among Japanese-​speaking learners, the main difference between the studies is that Crowther (2015) specifically investigated how raising learners’ awareness to aspects of the L2 that are different and similar to the L1 can benefit L2 performance. This was not a direct aim in Saito (2011). Saito’s (2011) instruction included two perception activities, identification and discrimination. The assumption guiding this instructional approach was that two components are needed for learning to take place: (i) awareness of the target feature and (ii) “a strong mental representation of these sounds with their fundamental phonetic traits” (Saito, 2011, p. 48). The identification instruction was designed to provide learners with “a clear account of formal properties of English-​ specific sounds one by one in sequence” (ibid.). Then, learners produced individual sounds following this instruction. In the discrimination component of the instruction, learners were “alerted to which Japanese sounds they might confuse with English sounds and asked to discriminate the target English sounds from the closest Japanese counterparts (e.g., /​æ/​vs /​a/​)” (ibid.). This was followed by production practice whereby learners produced the target English sound and a close Japanese sound with the aim of encouraging learners “to consciously make a contrast between [the English sound and the Japanese sound]” (ibid.). In addition to these activities involving explicit information of English-​specific segmentals and corresponding sounds in Japanese with controlled production, learners also completed three types of reading, (i) a segmental reading task, (ii) a word-​level reading task, and (iii) a sentence-​level reading task that involved sentences containing instances of the target sounds. For example, the sentence-​level reading task included four sentences that were loaded with words that contained the target segmentals (e.g., when do you think they are going to read letters? includes ten words in which six were loaded with target features). A picture description task was also included in which learners described two pictures as if they were explaining them to someone who had never seen them. Feedback was provided during the production activities, described as “objective feedback with the aid of

14

114  Instructed L2 Learning

the acoustic analysis software Praat” (ibid., p. 49). The instruction was provided in English or Japanese “depending on the extent of learners’ understanding” (ibid., p. 50) and lasted four hours over the course of four weeks. In terms of the research design, Saito used a pretest-​posttest design with two groups, an experimental group and a control group.While the experimental group completed the instruction, the control group completed individual study in the library and completed the pretest and posttest tasks only. For the analysis, listeners rated 20 second segments of learners’ L2 performance in the reading tasks for accentedness and comprehensibility. Results showed no changes over time or between the groups for accentedness, indicating that the instruction did not influence the accentedness of L2 speech. In terms of comprehensibility, ratings of performance in the sentence reading task were judged as being more comprehensible among learners in the experimental group but not in the control group. In the picture description task, however, no effects for time or group were found. Taken together, these results indicate that instruction on English-​ specific segmentals led to more comprehensible performance but no benefits for accentedness were found. Because the sentence-​reading task was also included in the instructional treatment, there may be a familiarity or practice effect from the instruction at play in the study’s findings (for discussion, see Segalowitz & Lightbown, 1999). In other words, evidence of improvement was only found in tasks that were used in both the instruction and the pretests and posttests. This could explain why improvement in the instructional condition was found. Using other types of tasks not used in the instruction are needed to understand the extent to which this is an instructional effect or a task familiarity effect. In addition, it would be helpful to understand how the training led to awareness of the English–​Japanese difference, if at all, especially since awareness raising was an aim of the training. That is, Saito (2011) claimed that awareness of the learning difficulty was needed for L2 abilities to improve, but awareness was not assessed at the posttest. Related to this, the extent to which instructional benefits extended to L2 perception should be investigated. Only production ability was tested even though a large part of the instruction involved perception training in identification and discrimination.

4.7.3  Explicit and Implicit Pronunciation Instruction Focusing on the acquisition of Spanish, Kissling (2013) investigated how the type of instruction could help learners to more accurately produce Spanish sounds that are known to be difficult to acquire and use by English speakers, even after considerable experience and exposure to Spanish. The study focused on eight consonantal phones [p, t, k, β, ð, ɣ, ɾ, r], which can be categorized as stop consonants (/​p, t, k/​), approximants (/​β, ð, ɣ/​), and rhotics (/​ɾ, r/​). English speakers’ production of these phones in L2 Spanish can lead to accented speech. For example,

14

15

14

Instructed L2 Learning  115

English speakers tend to produce Spanish /​p, t, k/​in syllable-​initial and stressed position (e.g., pasa, toro), with extended voice onset times and aspiration (Lord, 2005). In addition, English speakers have been shown to substitute the English sound /​ɹ/​for Spanish /​ɾ/​or /​r/​(e.g., in the Spanish words pero or perro), leading to accented speech (Face, 2006). Together, these sounds are understood to be difficult for English speakers to acquire and produce in L2 Spanish. Kissling (2013) compared the effectiveness of explicit pronunciation instruction with a more implicit condition to understand whether some of the learning gains documented in previous research may have been related to the explicitness of the instruction (e.g., Castino, 1996; Lord, 2005). In addition, this study investigated the extent to which the effectiveness of the instruction varied as a function of curricular level (or L2 experience) by comparing learners at three levels of instruction. The study included English-​speaking learners of L2 Spanish enrolled in introductory, intermediate, and advanced Spanish courses in the United States. A comparison group of Spanish native speakers was recruited. The study lasted six weeks and included a pretest, posttest, and delayed posttest. Spanish learners were randomly assigned to either an explicit phonetics instruction group or a control group. The explicit instruction included four computer-​delivered modules on (i) introduction to articulatory phonetics, (ii) voiceless stops /​p, t, k/​, (iii) approximants (/​β, ð, ɣ/​), and (iv) rhotics (/​ɾ, r/​).The modules included the following: explanations of grapheme–​phoneme correspondences, explanations and animated diagrams of the place and manner of articulation, explanations of differences of analogous Spanish/​English sounds and the phonological environments in which the sounds are produced in each language, and identification activities that required learners to identify Spanish and English sounds or their manner of articulation. Examples of the instruction are provided in Appendix A of Kissling (2013), which illustrate the nature of the explanations given. For instance, learners received meta-​ linguistic information about the different sounds, such as: Another group of consonants are fricative consonants. These sounds are created by forming only a partial obstruction in the articulatory tract, rather than a complete closure. In articulating fricative consonants, the air never stops completely but rather passes through this partial obstruction with friction. Kissling, 2013, p. 741 Assessments using short multiple-​choice questions were used to determine the extent to which learners had understood the explanations: In Spanish, the sounds /​b, d, g/​are considered: • •

Occlusive consonants Fricative consonants

16

116  Instructed L2 Learning

• •

Vowels Liquid consonants Fricative consonants are articulated by creating:

• • • •

A partial obstruction in the vocal tract An unimpeded flow of air through the vocal tract An unimpeded flow of air through the nasal cavity A complete obstruction in the vocal tract From Kissling, 2013, p. 742

At the end of each module, learners completed pronunciation practice activities in which they were required to listen to and then repeat a Spanish speaker producing sounds. No feedback was provided. In total, learners experienced at least ten tokens of each target phone. Learners in the control group also completed computer delivered modules, but without explicit instruction about phonetics. Learning was assessed using a production test that included a 28-​item list of words and phrases that learners read aloud, which included four tokens of each phone. Twenty of the words used in the list came from the first-​year textbook (e.g., ¿como? “what”, ¿que tal? “how are you?”, señorita “miss”, avenida “avenue”). The same test was used at the pretest, posttest, and delayed posttest, but the order of the words was different at each test point. Overall, the results showed more accurate production scores following instruction, but no meaningful differences were found between the explicit and implicit instructional conditions. The author concluded that instruction on phonetics of any kind was beneficial for L2 phonology, suggesting the explicit information itself may not be a key component to explaining L2 learning in this study. In light of these findings, an important consideration going forward is why no potential instructional differences were found between the conditions, especially given that one condition included explicit phonetics instruction and one did not.The study’s results suggested few additional benefits of explicit instruction on L2 learning outcomes. In line with previous work on this topic, we should consider how useful instruction about the “formal properties of sounds and their ­articulation” is to the use and production of language. It seems possible that the explicit instruction provided in this study (e.g., about the different labels of phones) is not helpful for usage. That is, does knowledge about how fricative consonants are created help production of these sounds? Furthermore, although this study did include some contrastive information about similar sounds in L1 and L2, the design did not allow us to tease apart their specific contribution to learning. In short, it seems important that we reflect on how the explicit information was presented, what it contained, and in what ways it has the potential to benefit use. Meta-​linguistic information about the formal properties of language might not be the most appropriate form of instruction if our aim is to improve usage.

16

17

16

Instructed L2 Learning  117

4.8  Synthesizing Evidence about Instruction and L2 Learning In this chapter, we have looked in some detail at a variety of studies that have investigated how instruction can play a role in helping improve L2 learners’ knowledge and use of an additional language. We have reviewed instructional studies in three main areas of language, morphosyntax, vocabulary, and phonology, thus complementing the review of L2 learning studies in Chapter 3. In this section, our aim is to move away from the detail of specific studies to think more globally about (i) what do we know and what can we say about the usefulness of instruction to support L2 learning difficulties caused by L1–​L2 differences, (ii) what can we take forward from these studies in terms of research design and assessments of learning, and (iii) in what ways do these studies enhance or refine what we know about crosslinguistic influence in L2 learning? As we noted at the beginning of this chapter, we should be clear that we are working with a relatively small evidence base when it comes to understanding instructional effects and crosslinguistic influence. Nonetheless, this research offers useful points of reflection going forward. If we take as our starting point the observation that explicit instruction can be helpful for improving L2 learning, one question we need to further explore is why this might be the case.That is, in what ways do these studies appear to indicate that some type of instruction focused on L1–​L2 differences improved L2 performance? On the whole, our review showed linguistic improvement following instruction. Comparisons between instructional groups help us to understand the effectiveness of a specific instructional condition relative to a different one, if a study is set up in that way.Together, these analyses of performance within the same group over time (e.g., between a pretest and a posttest) and between different groups at specific points in time (e.g., at a posttest) are what we typically use to determine whether a specific instructional method appears effective or not. With this understanding, the body of research reviewed indicates important benefits for instruction about L1–​L2 differences for improving L2 learning. These benefits have been found in a variety of target languages, among learners with different levels of exposure and experience with the L2 (e.g., beginners, intermediate learners, advanced learners), and using a variety of instructional techniques (e.g., explanations, production practice, identification). In sum, our field is accumulating different pieces of evidence about the ways in which instruction appears helpful for L2 learning. In the next section we reflect on some of the specific components of these studies that have contributed to these conclusions.

4.8.1  Research Designs and Assessments of Learning Even though a study’s research design and its methods of assessments have long been known to influence the nature of a study’s conclusions, methodological scrutiny has recently become an important focus of attention in our field (e.g.,

18

118  Instructed L2 Learning

Byrnes, 2013; Ellis, 2015; Porte & McManus, 2019). Work in L2 proficiency is an important case in point. For example, many different tests of proficiency are used in the field of SLA research (e.g., Oxford Placement Test, Goethe Placement Test, c-​tests) as well as other indexes too, such as instructional level, amount of time spent learning a language, self-​ratings of language ability (see Leclercq et al., 2014). While each of these tests is useful, each test also conceptualizes “proficiency” in different ways, which can lead to difficulties in making comparisons between studies and advancing what we know about the role of proficiency in L2 learning (e.g., Brown et al., 2018; McManus & Liu, 2020). Similar issues are at play when we try to make sense of the literature on instructional effectiveness. In this subsection we focus on two related aspects of this issue: design of the instruction and learning assessments. First, the design of the instruction is of paramount importance in instructed L2 research (see Mackey, 2017; McDonough, 2017; Plonsky, 2017). In this chapter we have seen many different approaches to addressing the negative effects of crosslinguistic influence with instruction. One major source of agreement between these studies is that L2 improvement can be attributed to the instruction. In Tolentino and Tokowicz (2014), for example, language improvement following salience marking (e.g., highlighting) and explanations is what led learners in the Salience and Rule group to outperform learners in the Control group. Similarly, in Zhao and Macaro (2016), the provision of L1 translation equivalents for L2 vocabulary items rather than L2 explanations accounted for L2 learning improvements in the L1 translation group. This observation is important for two reasons. First, to understand the specific contribution of a particular instructional technique, an instructional technique must be compared with a different instructional technique. When we compare “instruction method A” to “no instruction”, we can only conclude that some type of instruction is more beneficial than no type of instruction. While this is a move in the right direction, further research is needed to better understand what made the instruction effective. If we take Saito (2011) as an example, that study included two groups, an explicit pronunciation group and a control group. If we want to understand the extent to which specific aspects of that instruction improved learning outcomes (e.g., the production activities, crosslinguistic discrimination practice), then we would need to replicate that study and intentionally vary one aspect of the instruction (Porte & McManus, 2019). Indeed, Saito (2011) notes in his conclusion that the study can prime future pronunciation instruction research about potential interactions between instruction type and feedback. This is important because even though Saito’s study involved components designed to draw learners’ attention to aspects of L1 and L2 sounds that are different, we cannot infer from this study alone whether those design features led to the reported learning benefits. To do this, we would need to replicate the study by creating a new instructional condition that provided the same type of training but without the crosslinguistic components, for instance.

18

19

18

Instructed L2 Learning  119

One example of better understanding the specific contribution of a particular instructional technique is reported in McManus (2021). That study considered how the type of explicit information (about L2 vs. about L2 and L1) influenced language switching performance. This was achieved by keeping the comprehension practice the same and varying the type of explicit information provided to the different treatment groups (see also McManus & Marsden, 2017, 2018). One group received explicit information about form-​meaning mappings for habituality and ongoingness in French and a different group received explicit information about form-​meaning mappings for habituality and ongoingness in French and English. By intentionally varying the type of explicit information in different instructional conditions, it was possible to understand how these differences in the type of explicit information shaped learning outcomes. Had the study design included only one type of explicit information, it would be difficult to conclude that the type of explicit information played a role. Designing instruction with these goals in mind is needed if we are to understand claims about instructional effectiveness. There is more nuance in the effects of instruction than many of our current designs allow for and it is only through systematic replication with careful variable modification of previous instructional studies that we can begin to understand the issues at play. The second point that we need to consider is how we assess learning. In many studies the main (if not only) index of whether learning happened is via some type of test after the instruction. In fact, all of the studies reviewed here included some type of posttest assessment of language ability.There are a number of related issues to be discussed with respect to assessments of language knowledge in instructional studies, and we will address each in turn (see also Leow, 2015).The first is that very little attention is paid to performance during the teaching/​intervention, the second is connected with the type(s) of language ability assessed, and the third concerns how durable changes in language ability following instruction might be. To start with the type of data researchers analyze to draw conclusions about what has been learned and what still needs to be learned, there is a strong tendency in the field to focus on posttests only. That is, we mainly use performance at the posttest to determine whether the instruction improved L2 performance or not. As previously noted, very few studies analyze learner performance during the instruction (but see DeKeyser, 1997; McManus, 2021; McManus & Marsden, 2019b). One limitation of focusing on posttest assessments of knowledge is that while these measures might tell us about the state of learners’ abilities at that point in time, posttests tell us very little about how those abilities came to be.That is, by focusing on posttest data without looking at L2 performance over the course of the instruction, we are forced to infer about what happened during the instruction that led to the observed changes at the posttests. Even though this is a common approach in our field, it is important to be clear that posttest data provide only a very small insight into the nature of the learning puzzle. We must analyze both types of data in order to understand how specific aspects of the instruction

210

120  Instructed L2 Learning

contributed to learning. Another reason for analyzing performance during the intervention is that as researchers we put a lot of time and effort into designing the instruction. It makes sense that we would want to look at this data. In short, in order to make sense of L2 learning, we need to be looking at both performance during the instruction (process data) and performance after the instruction (product data). A major source of research methodology that also plays a role in the conclusions we draw is connected to the types of language data we collect. It is now well known that learners appear to perform better on some tasks rather than others, which can be explained by the type of language knowledge developed during the instruction. For example, Saito (2011) was clear that learners had to “become explicitly aware of the English-​specific segmentals /​æ, f, v, θ, ð, w, l, ɹ/​and establish a strong mental representation of these sounds with their fundamental phonetic traits” (p. 48) in order for learning /​development to take place. However, even though being “explicitly aware” was thought to be an important component for learning to happen, Saito (2011) included no assessments of awareness in the study. This means that we are unable to conclude whether this awareness was indeed needed for learning to take place. This is an example of how collecting different types of data can be helpful in order to provide a fuller understanding of the instruction. In order to circle round on this claim that awareness plays an important role, some type of awareness data are needed (see Rebuschat, 2013). For example, McManus (2019a) reported on an awareness test and how performance on that test helped explain performance on a variety of different language assessment measures in reading, listening, and speaking.This was to understand the claim that awareness of L1–​L2 form-​meaning differences facilitated the use of the L2 feature under study. Without measures about learners’ awareness, it is difficult to make progress in understanding its role in L2 learning. Related to this point, a number of studies reviewed in this chapter used a single test for assessing L2 learning. For example, a grammaticality judgment task in Tolentino and Tokowicz (2014), a word-​reading test in Kissling (2013), a vocabulary knowledge test in Zhao and Macaro (2016). Future research is needed that collects different kinds of language data. This is needed to avoid the limitations of a single language test (see Chaudron, 2003), but also to arrive at a fuller understanding of the types of language knowledge developed through instruction. Collecting a partial sample of L2 performance provides only a tiny snapshot of the types of language developed during the instruction. Lastly, a concern in our field involves claims about the durability (or retention) of learning beyond an immediate posttest (see Shintani, 2015). This concern is addressed at studies that include either no testing post instruction or only a small delay after instruction (e.g., three or four days). Studies are needed that include more extended delayed posttesing for a number of reasons. One aspect of this is to try and tease apart different types of language knowledge, such as those which are thought to be more explicit or declarative versus those that are more automatic.

210

12

210

Instructed L2 Learning  121

Delayed posttesting can be helpful since declarative knowledge is understood to deteriorate over time (DeKeyser, 2009). Posttesting is one way to determine the nature of the language knowledge created from instruction. In addition, as discussed by Shintani (2015), delayed posttesting is helpful for distinguishing between artefacts of the instruction that learners may have remembered versus those that have been learned. The difference being that remembered information is forgotten over time, while information that has been processed and integrated with existing knowledge sources is less likely to be forgotten a few days or hours later. In short, delayed posttesting is needed to better understand the types of language knowledge created from the instruction as well as how well that knowledge has been retained. Taken together, we have discussed some of the ways in which the design of instructional studies is important for making claims about instructional effectiveness and learning. Although focusing on any one of these methodological aspects can help us to better understand the instruction, a more comprehensive approach is to pay close attention to the design of the instruction, the types of language assessment data used, and the incorporation of delayed posttesting.

4.9  Conclusion In this chapter, we have reviewed a number of studies that have investigated how instruction can be used to address some of the negative effects of crosslinguistic influence. As in Chapter 3, we focused on three areas of language: morphosyntax, vocabulary, and phonology. Our review focused on the instructional and methodological designs of studies in each of these areas before considering the study’s findings. This is because it is important to be clear about each study’s approach to instruction rather than being influenced by the findings only. One overarching conclusion from these studies is that encouraging learners to attend to similarities and differences between L1 and L2 represents an important instructional strategy for improving L2 abilities. One explanation for this is that awareness raising techniques designed to increase sensitivity to differences between L1 and L2 followed by practice plays an important role in leading to improved L2 abilities. At the same time, this is a relatively under-​researched line of inquiry, but it represents an important one if we are to better understand the ways in which instruction can play a facilitative role in L2 learning.

21

5 REFLECTIONS AND FUTURE DIRECTIONS

5.1  Introduction The final chapter of this book has two aims. The first is to briefly revisit the main ideas and findings about crosslinguistic influence and L2 learning as discussed in this book. The second aim is to discuss new directions in SLA research by critically reflecting on this body of knowledge. If we think back to the book’s opening chapter, we set ourselves the goal of reviewing three connected bodies of research about crosslinguistic influence and L2 learning.We did this in order to get a better understanding about some of the main theoretical accounts and findings about crosslinguistic influence that have shaped the field of SLA. Our review has highlighted three main ideas: (i) learning is experience-​driven, (ii) transfer is a core process in L2 learning, and (iii) there is competition between and within languages. We will look at each of these in turn. First, the experiences we bring to the task of learning a new language shape the routes and rates of new learning in fundamental ways. Indeed, this is no different from how we learn any new skill: our previous experiences influence our future behaviors. In terms of language learning, we can conceptualize of this previous experience in a variety of ways, including in linguistic, processing/​attentional, and contextual terms (see Chapter 2). Not only are these types of experience closely connected and co-​dependent, but they are constantly changing. One consequence of this is that the knowledge underpinning all types of language use, including L1 use, is not static or resistant to change. For example, attending to new information can lead to the development of new knowledge (or processing routines, O’Grady, 2015), which, in turn, can influence existing (e.g., L1) processing behaviors. Our review of changes to L1 processing as a function of L2 learning is one example of this (see Chapter 3). In addition, contextual factors DOI: 10.4324/9780429341663-5

21

123

21

Reflections and Future Directions  123

influence how we use and attend to language. Examples of this include multilingual contexts in which speakers regularly use different languages and switch between them (e.g., Green & Abutalebi, 2013) as well as classroom contexts that deliberately and systematically practice specific language skills (DeKeyser, 2017). Together, our cumulative experiences with language feed into rich networks of knowledge that shape how we use language. Otherwise said: language learning is an experience-​driven process. When we conceptualize language learning as driven by our experiences with language, we need to be clear that sometimes that experience will be helpful, but sometimes it will make new learning more difficult. For example, the linguistic cues used in L1 might not be used in the same ways in the new language (e.g., word order in English and German, see MacWhinney, 2005) and learned attention can bias which cues we attend to (see Ellis & Sagarra, 2010, 2011). Either way, prior experience is a major influence in learning. We must be sensitive to this observation in how we think about and investigate topics in L2 learning. Second, theories and empirical studies of L2 learning view “transfer” as a core learning process in the creation and development of L2 abilities. Indeed, as our reviews in Chapters 3 and 4 showed, it is difficult to navigate L2 research without encountering some discussion of transfer. As noted from the outset, a dominant interpretation of transfer is that it involves a process of copying and restructuring. That is, L2 knowledge starts out as a copy of L1 knowledge that is restructured through L2 exposure use (see Jarvis, 2016; Sharwood Smith & Truscott 2006, 2014). These two bodies or systems of knowledge (e.g., L1, L2) are hypothesized to be separate but connected in some way.This account is thought to explain why L2 performance can appear L1-​like and why learning a new language does not replace prior knowledge of another language (e.g., L1). Transfer therefore allows knowledge of multiple languages to coexist. At the same time, we have discussed that very little is understood about what transfer actually is, even though it is a common explanation for L2 performance in our field. For example, we know relatively little about how transfer might be triggered as well whether L2 knowledge remains connected with L1 knowledge after initial copying. As our review has demonstrated, there is indeed ample discussion about the “what” of transfer. That is, L2 researchers know what the (negative) effects of transfer look like and the field has developed methodological frameworks for identifying instances of transfer (e.g., Jarvis, 2000). In terms of theories, such as in the UCM, for example, “all aspects of the first language that can possibly transfer to L2 will transfer” (MacWhinney, 1997, p. 119). In comparison, PD takes a more selective position on transfer: “L2 learners transfer dominant processing routines, unless the cost of implementing those routines is less favorable in the second language than in the first language” (O’Grady, 2013, p. 271). In Chapter 2, we saw that both accounts conceptualize of transfer as a process of copying (and, in the UCM, restructuring). Problematically, however, our field can

214

124  Reflections and Future Directions

describe transfer, but it cannot explain transfer. This is a problem because many theories of L2 learning attribute critical roles to transfer. A pressing question for the field of L2 research is as follows: to what extent does our conceptualization of transfer as a process of copying and restructuring still makes sense? This question is particular important given a growing body of evidence that L2 learning appears to influence L1 use. As shown in Chapter 3, L2 research indicates very close relationships between L1 and L2, which has pushed the field to think more critically about transfer and the ways in which a speaker’s languages are related and/​or connected. For example, some proposals suggest that we see changes to L1 performance because L1 and L2 eventually converge over time (see Chapter 3; Dussias & Sagarra, 2007), but it could also be the case that L1 and L2 are tightly networked into a single system of language knowledge from the outset. Indeed, as previously noted, these are ideas that fit with connectionist approaches to language learning as well as proposals that L1 and L2 representations can be identified and selected by associating linguistic representations with language tags (e.g., Calabria et al., 2018). We will develop this discussion in Section 5.3, but, for now, we can say that transfer has long represented a key process in how we talk about L2 learning. New directions in L2 research are encouraging us to question these traditional ideas about transfer, including the extent to which transfer still appears to be a useful process for explaining the rates and routes of L2 learning. Third, competition and cooperation are important concepts in how we describe and explain L2 learning, typically (but not only) found in discussions about cross-​language competition.The claim here is that linguistic representations become activated through stimulation (via an external stimulus, for example). Activated representations then compete for selection in the same language or in different languages (e.g., Boot in German and English). The resting (or activation) level of a representation can determine how costly or likely it might be for a representation to be selected. Representations with higher activation levels tend to be less costly to select (and are also more dominant).This is one explanation for why L1 representations tend to win out over L2 representations, at least initially. Usage is a critical factor that can contribute to a representation’s activation level (see Shirai, 2019). The relevance of competition in L2 learning is that a speaker’s language system is constantly managing competition. In our review of the ICM (Green, 1998), we saw that the creation and/​or development of language selection mechanisms is one way to explain how a speaker can (i) manage cross-​language competition and (ii) switch (or select) among their languages, especially given the dominance of L1 representations. In this model, representations are associated with language-​specific tags and inhibitory control operates on language tags to manage cross-​language competition. Inhibitory control reactively suppresses activated L1 representations during L2 use, and vice versa during L1 use. Taken together, not only are these ideas well represented in studies of L2 learning, but they represent core ideas in our field. Indeed, as noted in Chapter 1,

214

125

214

Reflections and Future Directions  125

interest in crosslinguistic influence and L2 learning represents a major focus of attention in our field because adult L2 learners begin learning a new language as experienced and expert users of at least one other language. The challenge in L2 learning, therefore, is that the learner is seeking to process and produce a new language using a system that has been built and fine-​tuned for a different language (e.g., L1). Sometimes, this prior experience can be helpful, but sometimes it can slow down the L2 learning process. This is why attending to the experience speakers bring with them is critical to understanding L2 learning. This observation has been demonstrated in a variety of domains of L2 use (e.g., morphosyntax, vocabulary, and phonology), in both studies of what L2 learning looks like as well as the extent to which instruction that is sensitive to features such as prior experience and competition can support L2 learning. In the remaining sections of this chapter, we take the ideas of experience, transfer, and competition and review them more critically. In Section 5.2, we bring together the main points from Chapters 3 and 4 and think more about what this evidence base suggests about the ways in which (i) language knowledge appears to be stored, (ii) experience influences language processing and use, (iii) general cognition shapes language use, and (iv) how using instruction to attend to specific aspects of experience and cognition can support L2 learning. In Section 5.3, our aim is to reflect on the state of theorization in the field of SLA by reviewing what is currently known about crosslinguistic influence in L2 learning, paying particular attention to transfer as a core learning process. Then, in Section 5.4, we entertain new directions in crosslinguistic influence research, including some of the reasons for why L2 research needs to be looking at (i) the whole language system rather than just L2, (ii) how a focus on longitudinal accounts of L2 learning are needed to document and better understand the processes of L2 leaning, as well as (iii) how studies examining aspects of general cognition are needed to advance knowledge and understanding about language, cognition, and their relationships.

5.2  Synthesizing Crosslinguistic Influence, L2 Learning, and Instruction In our review of empirical work in L2 learning and instruction, we saw that a number of studies have advanced what is currently known about the architecture and nature of L2 knowledge by examining if and how connections with other languages (e.g., L1) might be established and in what ways these connections change over time (e.g., Jiang et al., 2020; Thierry & Wu, 2007). Similarly, in L2 instruction, we have seen how instruction that is sensitive to patterns of L1 usage can be helpful in reducing some of the negative effects of crosslinguistic influence (e.g., McManus & Marsden, 2017; Spada et al., 2005). Thus, studying changes in how speakers use their other languages can potentially advance theory building in L2 research. This is an important point since theory building in our field has traditionally concerned itself with understanding and explaining how L2 abilities

216

126  Reflections and Future Directions

emerge and change over time, without paying too much consideration to how other abilities might be amenable to change as a function of L2 learning.

5.2.1  Organizing and Restructuring Knowledge One line of research that has made important contributions to theorization about the emergence and development of cross-​language relationships has involved studying how L2 experience can change L1 lexical, syntactic, and phonological processing (e.g., Dussias & Sagarra, 2007; Kroll et al., 2002; Schuhmann, 2016). In Dussias & Sagarra (2007), for example, Spanish speakers living in Spain with minimal experience with English (the L2) processed Spanish sentences like Spanish L1 speakers, whereas Spanish speakers living in the USA and with greater amounts of immersion experience in an English-​speaking country performed like English L1 speakers. This finding indicated that greater experience with English influenced sentence processing in L1. According to the authors, this finding could suggest convergence of L1 and L2 systems. In terms of lexical processing, Kroll et al. (2002) found changes to the speed of L1 processing in light of L2 learning. Using a picture naming task, Kroll et al. found that beginning learners of French were slower to name pictures in their L1 than more experienced L2 learners (see also Bice & Kroll, 2015). Together, these results indicate that L2 learning can influence L1 behaviors. A relevant question for L2 research, therefore, is to what extent can findings like these help us to better understand L2 learning? Findings that L2 learning appears to influence a speaker’s broader language system suggest that L2 learning leads to some type of restructuring in how linguistic representations are stored or organized, including those in L1. At the same time, such evidence indicates changes in how knowledge is accessed and assembled in real-​time. Indeed, it is plausible that L2 learning involves change in both storage/​organization and access/​assembly. This is because when new connections emerge between existing and new representations, the language processor has to learn new ways to access and assemble information. On this latter point, just as L2 learning involves the creation of linguistic representations that are potentially distinct from those used in L1 processing, L2 learning likely also involves the creation and/​or development of language selection mechanisms (McManus, 2021). In line with the ICM, therefore, new and existing language representations need to be associated with language tags to facilitate language selection. If this is the case, L2 learning should therefore trigger a tagging process that includes associating L1 representations with L1 tags. As this discussion suggests, these are relatively speculative accounts for some of the reasons why we might observe changes to L1 in light of L2 learning. One area of considerable interest going forward, therefore, is better understanding why aspects of L1 performance are susceptible to change and what these changes suggest about the organization of L1 and L2 knowledge sources. One relatively clear outcome of this research is that L1 abilities are not static, even despite many

216

217

216

Reflections and Future Directions  127

years of daily L1 use. Work in sociolinguistics has demonstrated language change for a long time now (Bybee, 2010), but our field has not paid enough attention to this work in how we think about language learning. Furthermore, the direction of crosslinguistic influence is not monodirectional, from dominant to less dominant language. Instead, the relationships between a speaker’s languages are dynamic and closely connected. Furthermore, speakers are sensitive to different aspects of language usage, such as frequency, to such an extent that aspects of word frequency in one language influence performance in a different language (e.g., Jiang et al., 2020). Just these few observations from our review reinforce a need for more integrated conceptualizations of linguistic knowledge.

5.2.2 Experience Throughout this book we have seen the numerous ways in which experience can shape the routes and rates of L2 learning, including the prior language experience that speakers bring to the task of using a new language as well as the experience speakers have with using a new language to communicate. In addition, the influence of experience on language is not compartmentalized because experience with a particular language has been shown to influence the entire language system. As we have discussed, this understanding suggests a very tightly connected network of language knowledge rather than separate bodies of language knowledge. The entire language system thus appears to be sensitive to changes in usage across the board rather than patterns in L1 usage only influencing L2 knowledge. Indeed, these understandings drill home Grosjean’s (1989) message that bilinguals are not “two monolinguals in one person” (p. 3). Given that learning is understood to be an experience-​driven process, it is therefore not surprising that many models of L2 learning, including those discussed in Chapter 2, attribute critical roles to experience in theorizing L2 learning. For example, when we think about prior experience, theories such as the UCM and PD provide important roles for the linguistic knowledge that learners have already constructed and how this knowledge feeds into L2 use.The Associative-​Cognitive CREED provides an additional perspective on this issue by including roles for blocking and learned attention which bias new learning. Together, the linguistic knowledge and the processing/​attentional knowledge that speakers bring to the learning of a new language are connected and each exerts different influences on L2 learning. We saw an example of this from work in L2 predictive processing. Studies have shown L2 speakers can demonstrate linguistic knowledge of a new L2 feature (e.g., grammatical gender in Spanish), but then struggle using this feature during real-​time processing, depending on the type of experience they bring with them (e.g., Dussias et al., 2013). One explanation for this body of research is that prior processing experience biases the cues learners attend to the most, because prior experience has shown them to be informationally relevant. Instructional research shows how encouraging learners to notice or attend to

218

128  Reflections and Future Directions

these new cues can be helpful (e.g., Cintrón-​Valentín & Ellis, 2016; McManus & Marsden, 2017). That is, then, focusing on how learners use language is helpful (see also Tyler et al., 2018). This suggests that we need to develop instructional approaches that improve the ways in which learners process and attend to linguistic cues because the creation and development of linguistic knowledge is only part of the L2 learning puzzle (see also VanPatten, 2002, 2015). In other words, one way to incorporate experience into theories of L2 learning and approaches to L2 instruction is to focus on how speakers use language to communicate rather than prioritizing the construction of linguistic knowledge only. Indeed, just as prior experience is important to understanding L2 learning, so are the experiences speakers have using the L2. We need to be documenting what the language input looks like as well as how learners use the L2 to communicate. These accounts must then be integrated into instructional approaches. One line of this research captured in the ICM is that speakers who have more experience using their different languages in meaningful ways, including switching between languages, perform faster and more accurately in activities that require them to switch between languages (Green, 1998; see also Green & Abutalebi, 2013; Calabria et al., 2018). As this theoretical model suggests, one explanation for this behavior among unbalanced bilinguals and L2 learners is that better language selection mechanisms can result in enhanced abilities to switch between languages. A broader point here, though, is that practice is a necessary condition for the development of language ability (DeKeyser, 2017), be it switching between languages or something more specific such as learning to use article cues to anticipate upcoming information. Taken together, the field of SLA can advance what we know about the routes and rates of L2 learning by paying closer attention to experience, both prior and current experiences with language. Finding ways to integrate this experience into L2 instruction can be made possible by increasing sensitivity to how speakers use language to communicate. Usage-​based approaches to instruction are beginning to show some of the ways in which attending to how learners use L2 and L1 to communicate can facilitate L2 learning. For example, we saw in Chapter 4 how some research has integrated L1 practice in L2 instruction alongside explicit information designed to increase learners’ awareness about the specific meanings expressed by cues in L2 and L1 (McManus, 2019, 2021; McManus & Marsden, 2017). To date, our approaches to instruction have focused mostly on developing linguistic knowledge about L2 without sufficient attention to usage, both in L1 and L2 (Tyler & Ortega, 2018). Research showing close connections between L1 and L2 suggests that we need to integrate both dimensions into accounts of L2 learning and approaches to L2 instruction.

5.2.3 

Cross-​language Relationships

Throughout the book, we have seen accounts and studies of L2 learning that discuss the nature and emergence of cross-​language relationships, which point to a

218

219

218

Reflections and Future Directions  129

growing awareness that learning language, just like the learning of any other skill, involves the creation and development of connections with other aspects of knowledge, linguistic and non-​linguistic (Bybee, 2010; Tomasello, 2003). These insights point to a system of language knowledge that is fluid and highly networked. This is one reason why discussions of an integrated system of language knowledge seem to make more sense than conceptualizing of L2 learning in terms of the creation of a new body of language knowledge only, which may or may not converge with other types of (language) knowledge over time. Indeed, this type of development makes good sense if we think about how the field has changed and diversified over the decades.To date, theorization in the field has been centrally concerned with how speakers construct knowledge of a new language, without attending so much to connections with other languages and/​ or aspects of cognition, which can be traced back to the influences of linguistic theory at that time (Divjak, 2019). Now, L2 researchers are asking questions about the nature of L2 knowledge and its relationships with other languages. As previously noted, Jiang et al. (2020) found that aspects of L1 word frequency influenced L2 lexical processing, suggesting remarkably close associations between L1 and L2. Findings such as this indicate that L1 and L2 knowledge sources are probably much more closely integrated at multiple levels of representation, in line with connectionist models of learning (Shirai, 2019). Theories that conceptualize of L2 learning as the creation and development of nodes and connections likely entail that equal consideration should be paid not only to how L1 and L2 are connected, but how the language system establishes connections among nodes and also how cognitive mechanisms select particular nodes given their highly connected structure. Such a line of questioning goes to the heart of understanding learning. We want to understand not only what a speaker’s language ability looks like at a given time, but we also want to know what leads the system to look the ways that it does (Leow, 2015). For example, lexical items in L1 and L2 are understood to fire together on activation, but the language system needs a way to be able to select one over the other given this parallel activation. The field of SLA has discussed roles for inhibitory control and cooperation in explaining how the language processor might resolve such competition, but the field has not thought about how speakers learn to do this. One account is that speakers associate specific nodes with specific language tags and then language selection operates by using these tags (e.g., Green, 1998). Not well understood, however, is how this process comes to exist. For example, perhaps the creation of new knowledge triggers a tagging process in the whole system or perhaps nodes only become tagged when they become connected with a different type of knowledge. That is, maybe new knowledge is tagged once it becomes integrated into the system. In short, we are asking two questions: (i) how do speakers learn to associate specific nodes with language tags and (ii) how do speakers learn to select specific language nodes? Taken together, the field is accumulating different types of evidence about the extent to which cross-​language connections appear to emerge and develop

310

130  Reflections and Future Directions

across different levels of representation. This evidence base is relevant to our understanding of L2 learning because it encourages us to think about L2 learning more holistically, involving not just the creation and development of L2 knowledge, but also how this connects with existing abilities and sources of knowledge. An important task going forward, therefore, is to develop more nuanced and fine-​grained accounts of cross-​language relationships, including theorization about the ways in which such relationships emerge. Understanding such questions will necessarily require the use of longitudinal research designs that systematically document the trajectories of multilingual use. Such designs would allow us to better understand the signatures of knowledge creation and/​or restructuring, which, in turn, have the potential to more fully inform theories of learning (e.g., Solovyeva & DeKeyser, 2018; McManus, 2021; McManus & Marsden, 2019b). In addition, current approaches to instruction must more fully incorporate multilingualism and usage into the curriculum. That is, then, if we understand one characterization of L2 learning to be the creation and development of cross-​language connections, we should be looking at ways at incorporating use and expertise of other languages in the language classroom as a means to allow these connections to emerge and develop.

5.2.4 Instruction Evidence-​ based language instruction is one way to strengthen connections between SLA research and language teaching. One way to achieve this aim is by using the insights about learning difficulties as evidenced in empirical research studies to inform the design of teaching materials and approaches to instruction. In Spada et al. (2005), for example, learners were provided with explicit contrastive information about the target features and in McManus and Marsden (2017) learners received additional explicit information about L1 as well as practice that required learners to apply that explicit knowledge to the processing of L1 sentences. In our review, we noted a number of ways in which drawing learners’ attention to specific aspects of L1 and L2 usage appeared to benefit L2 learning. One area for further theorization will likely involve understanding why explicit information seems to have selective benefits for L2 learning. One argument advanced in Chapter 4 with regard to this question has to do with the nature of the learning difficulty. This is one reason why we must tailor the instruction to address the nature of the learning difficulty. Given that explicit knowledge tends to deteriorate over time, it is unlikely that possessing explicit knowledge of the target feature alone is responsible for some of the learning gains discussed in Chapter 4. Rather, explicit information draws learners’ attention to aspects of the L2 (and L1, when the instruction targets this) that would ordinarily have been missed due to the effects of learned attention and blocking. If, for instance, the L2 learning difficulty appears explainable in terms of a lack of exposure, it is unlikely that explicit instruction involving L1 would be helpful because this type of instruction

310

13

310

Reflections and Future Directions  131

is not designed to address a lack of exposure to the target feature. However, if the learning difficulty is created by crosslinguistic influence, instruction that draws learners’ attention to the nature of that crosslinguistic difference can be helpful. Furthermore, one reason for exploring instructional effectiveness in a more systematic way is that it has the potential for SLA research and language teaching to become more closely connected. This is because instructional studies informed by SLA research can feed back into theorization by testing the pedagogical implications of SLA research. That is, designing instruction to target specific learning difficulties has the potential to inform whether more work is needed on understanding the nature of the learning problem. So, if the L2 learning problem is claimed to exist because of a lack of awareness of the target feature, studies that examine how the provision of different types of awareness influence language use are needed to determine whether or not we have understood the nature of the learning difficulty. In sum, research using empirical findings about L2 learning have the potential to make an important contribution to understanding both (i) learning processes and (ii) instructional effectiveness.To date, some of this work has sought to address specific learning difficulties created by crosslinguistic influence, such as drawing learners’ attention to L1–​L2 differences or using explicit information and practice to increase sensitivity to the L1–​L2 difference. Going forward, instruction has an important role to play in better understanding (i) aspects of instructional effectiveness and (ii) whether our explanations for L2 learning difficulties seem to be justified.

5.3  Theorizing Crosslinguistic Influence At various points in this chapter and in our reviews of SLA research, we have reflected on the state of theorization about crosslinguistic influence in the field. Some of these discussions have served the purpose of highlighting points of agreement among different accounts of learning as well as some of the ways in which existing accounts diverge. In this section, we do not intend to revisit these differences and similarities but rather to reflect on the ways in which we might think about transfer differently. This is because the field’s conceptualization of transfer has remained relatively fixed since early proposals in the field (e.g., see Jarvis, 2016; Sharwood Smith & Truscottt, 2014). Early accounts of transfer conceptualized it as a process of copying and restructuring (see Sharwood Smith & Truscott, 2006, 2014). It is interesting to note that few other constructs in the field of SLA have remained as constant as transfer. For example, there is now a greater awareness of the ways in which usage and aspects of cognition shape L2 learning (e.g., Miller et al., 2018), use of learner corpora to investigate topics in L2 learning is growing (e.g., Tracy-​Ventura & Paquot, 2020), as well as the use of assessments that tap into real-​time language use (see Godfroid, 2019). At the same time, however, even though we have been talking about transfer for a long

312

132  Reflections and Future Directions

time and have been using transfer as an explanation for the routes and rates of L2 learning, it is surprising that so little is known about transfer. In a recent edited volume dedicated to exploring “new perspectives on transfer in second language learning”, for example, Odlin and Yu (2016) provide a comprehensive account of what the effects of transfer look like, how transfer has been investigated in the field of L2 acquisition, and some early lines of research in transfer and L2 learning (see also Jarvis, 2016). The authors begin their discussion of transfer with Odlin’s (1989) now-​seminal definition: Transfer is the influence resulting from the similarities and differences between the target language and any other language that has been previously (and perhaps imperfectively) acquired. Odlin, 1989, p. 27 In this book, as in others about crosslinguistic influence, we have repeatedly noted that linguistic differences between a speaker’s languages can constitute an important way to think about transfer or crosslinguistic influence in the field of L2 research. Indeed, this is often the case in empirical studies of L2 learning, such as those discussed in Chapter 3. One part of Odlin’s discussion of transfer is that it includes “influence resulting from”. This is a relevant point because it suggests a conceptualization of transfer as a product. Before thinking more about what this means as well as the likely implications of viewing transfer as a product of learning (rather than a learning process, for instance), Odlin and Yu (2016) note two other points. The first is that “no single theory really attempts to model the full complexity of the phenomenon of cross-​linguistic influence” (Odlin & Yu, 2016, p.9, see also Jarvis, 2016). The second is that “in the last three decades, cross-​ linguistic influence has been ascribed a greater role in theories of acquisition such as the competition model (e.g., MacWhinney, 2008), processability theory (e.g., Pienemann et al., 2005), and universal grammar (e.g., White, 2003)” (ibid.). We should briefly consider these observations about the field. In one sense, we might feel alarmed by the claim that the field of L2 research has yet to propose a theory that sets out to describe and explain transfer in L2 learning, especially since transfer is not a new concept in how we think about L2 learning (for historical overview, see Jarvis & Pavlenko, 2008). At the same time, we also know that transfer is a common component in many theories of L2 learning (see Gass et al., 2020), even if it is not yet a principal component of a single theory. Indeed, attributing a role to transfer in describing and explaining L2 learning is certainly a move toward more theorization of transfer. These are certainly positive takeaways, but we should also keep in mind that transfer has long constituted a key concept in our field. It is likely that in order to better understand L2 learning, we need to be thinking more fully and more critically about this key concept in our field. Perhaps one way to start a discussion of transfer is to think about whether the terminology we use for talking about it is helpful. Readers with some experience

312

13

312

Reflections and Future Directions  133

with this topic may have noticed that the labels “crosslinguistic influence” and “transfer” are sometimes used in apparently interchangeable ways. Indeed, Odlin and Yu (2016) introduce their review of transfer research with a note that crosslinguistic influence “is a synonym for transfer” (p. 1, see also Pavlenko & Jarvis, 2002). In this book, we have used the two labels in more specific ways: transfer as a learning process and crosslinguistic influence as an outcome or product of that learning process. In short, crosslinguistic influence is conceptualized as the result or outcome of transfer. The use of transfer in this way is similar to that used in the UCM and PD, in which transfer is used to talk about the cognitive processes that take an established piece of language knowledge (typically L1) and makes it useable for processing L2 data. By trying to separate transfer from crosslinguistic influence, we are in agreement with Leow (2015) that distinguishing between the processes and products of learning are a helpful way to provide theoretical and empirical clarity on key constructs in the field. Indeed, a lack of terminological and conceptual clarity is potentially one reason for the perceived lack of progress in advancing a theory of transfer. By separating out process from product, we will be better situated to ask questions about what the nature of L2 knowledge is (product) and how it came to look the way that it does (process). This would allow us to make specific hypotheses about what transfer (as a process) might actually involve.

5.4  Future Research Directions in Crosslinguistic Influence In this last section of the book, we discuss some future directions in SLA research with the aim to better understand crosslinguistic influence and L2 learning. In our discussion, we focus on three principal lines of research, as follows: (i) incorporating aspects of L1 use in studies of L2 learning, (ii) doing more longitudinal research, and (iii) investigating links between aspects of general cognition and language learning and use. First, as this chapter and discussions in Chapters 3 and 4 have highlighted, there is much to be gained from investigating how learning an additional language influences a speaker’s knowledge and use of other languages. We have noted some of this research and, in general, this aspect of multilingualism is not well understood. For example, we saw how greater experience in an L2 environment led L2 speakers to change the ways in which they processed sentences in their L1 (Dussias & Sagarra, 2007). At the same time, we saw how early experience with L2 learning affected the speed of picture naming in L1 (Kroll et al., 2002). Research in L2 lexical processing additionally showed that the frequency of a word in the L1 influenced how quickly and accurately its translation equivalent was processed in the L2 (Jiang et al., 2020). These are just some of the findings discussed in this book that have implications for how we think about cross-​language relationships. Findings like these, however, are not plentiful and this is one reason why L2 researchers should be looking at integrating more data about L1 use into studies of L2 learning. The insights provided by such study designs would allow us to

314

134  Reflections and Future Directions

better understand how cross-​language relationships emerge and change over time. One way to do this would be to collect similar, or at least comparable, data in L1 and L2. For example, Kroll et al. (2002) used a picture naming task whereby learners saw pictures and had to name them in L1 and L2 (see also Linck et al., 2012). McManus (2021) required learners to read the sentences in L1 and L2 that expressed the same meanings. In oral production, Huensch and Tracy-​Ventura (2017) used a picture-​based narrative task and asked learners to tell the story twice, once in L1 and once in L2. Methods like these have been used to understand performance in L2 comparative to L1. In cross-​sectional designs, data like these have been used to understand how L1 performance seems to vary as a function of L2 experience. Incorporating analyses of L1 ability into studies of L2 learning would be helpful for advancing what we know about L2 learning and how this new learning shapes L1 performance. Related to this, we also need to incorporate more longitudinal designs into our research. Despite repeated calls for such research designs in L2 learning, very few studies exist (see Ortega & Iberri-​Shea, 2005; Tracy-​Ventura & Huensch, 2018). Of course, there is very good reason why comparatively fewer longitudinal studies of L2 learning are conducted: They are expensive, participant attrition is a concern, and such studies require greater investment from the research team. In comparison, cross-​sectional studies often require meeting with a participant at one point in time. Despite the greater amount of effort required to conduct longitudinal research, however, some of the most important questions in our field cannot be fully answered with cross-​sectional research designs only. For example, our understanding of how L2 abilities develop requires us to trace learners’ L2 use over an extended period of time (e.g., Mitchell, Tracy-​Ventura, & McManus, 2017). It is thus difficult for us to understand why a speaker’s language abilities look the way they do using cross-​sectional designs. This is because cross-​sectional data provide a snapshot of an individual’s language ability. Data like these provide few insights into how and why that individual’s L2 performance looks the way it does. In short, if we want to better understand how L2 development proceeds, we need to broaden our methodological toolkit in ways that allow us to examine L2 performance over an extended period of time. Lastly, we have seen a small amount of research about L2 learning that investigated relationships between language use and cognitive processing. While work in working memory and L2 learning has represented an active line of SLA research in recent years (e.g., Juffs & Harrington, 2011; Wen & Li, 2019), other aspects of cognitive processing have received far less attention. Future research should develop this component of L2 research in order to document and theorize how additional language learning and cognition work together. For example, we discussed work by Darcy et al. (2016) who investigated relationships between inhibitory control and L2 phonological processing which suggested important relationships between this cognitive processing ability and L2 performance. Other aspects of cognitive processing that would be relevant to investigate include categorization and chunking.

314

135

Reflections and Future Directions  135

Bybee (2010) discusses these cognitive processes as being critical to understanding language development and use. For example, categorization in language learning involves “similarity and identity matching that occurs when words and phrases and their component parts are recognized and matched to stored representations” (p. 7). In other words, categorization is an ability to recognize and match categories. It is possible that general abilities in categorization would have a positive benefit on L2 learning outcomes. Research investigating these types of questions has the potential to contribute to theory building in the field as well as approaches to L2 instruction. For example, if research suggests that speakers with better categorization skills learn more and at a faster rate, then one approach to L2 instruction could be to find ways that develop speakers’ categorization abilities to see in what ways that could support L2 learning. Taken together, we have come a long way in terms of our understanding of crosslinguistic influence and L2 learning.The field has seen a number of important theoretical and empirical developments that have advanced knowledge and understanding about the ways in which experience plays a role in shaping the routes and rates of L2 learning. Going forward, research that investigates L1 use alongside L2 use, incorporates longitudinal research designs, and includes measures of cognitive processing holds great potential for developing our understanding of crosslinguistic influence and L2 learning.

314

5.5  Conclusion In this book, we have reviewed three connected bodies of research about crosslinguistic influence and L2 learning to better understand theoretical and empirical work in this area. Our review has shown great diversity and breadth in the approaches used to document and understand crosslinguistic influence. In terms of theories, we have seen that while L1–​L2 differences and similarities are part of this puzzle, work has also attributed roles to the processing/​attentional behaviors that speakers develop as they fine-​tune how they use language to communicate. We have also seen that contextual factors also play a critical role in shaping the routes and rates of development. In Chapters 3 and 4, we reviewed some of the ways in which the predictions made by these theories have been borne out in studies of morphosyntactic, vocabulary, and phonological L2 learning. In this chapter, we took a step back from the detail of individual studies and reflected on the broader implications of their findings for the field. Together, our review has demonstrated that much progress has been made in how we think about and investigate crosslinguistic influence. At the same time, we have seen that much work remains, including, for example, thinking about how we might contribute to a more nuanced understanding of what transfer is. Expanding our research to include measures of L1 performance is one way to achieve this goal, especially since such data can provide us with a more holistic take on the entire language system, rather than just part of it.

136

RECOMMENDED READING

Bybee, J. (2010). Language, Usage and Cognition. Cambridge University Press. This volume presents a clear and accessible introduction to usage-​based linguistics, particularly recommended for readers with interests in the connections among language and cognition. Perhaps one of the most useful aspects of this book is that it introduces a variety of cognitive processes (e.g., memory, chunking, analogy) and systematically explores how these processes shape language structure and use. Cadierno, T., & Eskildsen, S. W. (Eds.). (2015). Usage-​Based Perspectives on Second Language Learning. De Gruyter. The authors of this edited volume bring together a variety of empirical studies and critical reflections on usage in L2 learning and teaching. The book offers accounts on the role of frequency and exposure in L2 learning, the development of interactional and constructional competence, as well as how ideas about usage-​based linguistics can be used in language teaching. Divjak, D. (2019). Frequency in Language: Memory, Attention and Learning. Cambridge University Press. Divjak presents a critical overview of language learning research, drawing attention to critical insights from experimental and corpus-​based work as well as some of the challenges that lay ahead in understanding cognition and language learning. One very useful aspect of this book is that it presents a clear and accessible account of usage-​based approaches to understanding language structure, language learning, and language use. Ellis, N. C., Römer, U., & O’Donnell, M. B. (2016). Usage-​Based Approaches to Language Acquisition and Processing: Cognitive and Corpus Investigations of Construction Grammar. Language Learning 66(s1), 1–​358. In this book, Ellis and colleagues describe a series of empirical studies about the acquisition and processing of language by L1 and L2 users. The collection brings together a number of critical insights about what the language input looks like, how learning and use unfold, and how the two work together.

136

137

136

Recommended Reading  137

Leow, R. P. (2015). Explicit Learning in the L2 Classroom: A Student-​Centered Approach. Routledge. This book presents a comprehensive overview of explicit learning in the classroom. The author provides an approach that is grounded in theory, empirical research findings, methodological perspectives, and model-​building. Of particular interests to readers of the current book, Leow focuses on awareness and attention in L2 learning, including conceptual elaboration of these key learning processes, as well as ways in which they can be studied in L2 research. Porte, G., & McManus, K. (2018). Doing Replication Research in Applied Linguistics. Routledge. This book presents a step-​by-​step guide to carrying out replication studies in the field of applied linguistics. It is written in a user-​friendly manner that requires no previous knowledge of replication. Roehr-​ Brackin, K. (2018). Metalinguistic Awareness and Second Language Acquisition. Routledge. Roehr-​Brackin presents an in-​depth overview of metalinguistic awareness in the field of SLA.This topic is addressed both in terms of child and adult language learning and provides a methodological synthesis for assessing and measuring metalinguistic awareness. In addition, the place of metalinguistic awareness in language education is discussed. Segalowitz, N. (2010). Cognitive Bases of Second Language Fluency. Routledge. This book explores fluency from a range of perspectives in L2 learning. In addition to discussing the methodological aspects of fluency, cognitive and neural correlates of fluency, as well as speech production and automaticity, this book provides L2 researchers with important information about a variety of processes in language learning. Shirai,Y. (2018). Connectionism and Second Language Acquisition. Routledge. Given that connectionism is a key area of current L2 research, this book will be an important source of information for readers with interests in cognitive science and language learning. Tokowicz, N. (2014). Lexical Processing and Second Language Acquisition. Routledge. This volume provides a comprehensive review of research on L2 lexical processing. Although some of this research was discussed in the current book, Tokowicz provides a fuller account that focuses more squarely on vocabulary both in terms of theory and empirical research. Tyler, A. E., Ortega, L., Uno, M., & Park, H. I. (Eds.). (2018). Usage-​Inspired L2 Instruction: Researched Pedagogy. John Benjamins. This edited volume brings together a variety of theoretical and empirical papers involving usage in language teaching. In addition to discussions about why and how usage can be integrated into the classroom to inform language teaching, the different chapters explore specific types of usage that appear to be the most useful (e.g., corpus). Yu, L., & Odlin, T. (Eds.). (2016). New Perspectives on Transfer in Second Language Learning. Multilingual Matters. Yu and Odlin bring together a variety of empirical studies and conceptual reviews of work in transfer in the field of SLA. This timely volume points to new and emergent directions in transfer research involving a range of understudied languages (e.g., Hindi, Tamil).

138

REFERENCES

Abutalebi, J., & Green, D. (2007). Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics, 20(3), 242–​275. https://​ doi.org/​10.1016/​j.jneuroling.2006.10.003. ACTFL. (2021). Guiding principles for language learning. Retrieved January 31, 2021, from www.actfl.org/​resources/​guiding-​principles-​language-​learning. Alonso, R. A. (2016). Crosslinguistic influence in second language acquisition. Multilingual Matters. Ambridge, B., & Lieven, E. V. M. (2011). Child language acquisition: Contrasting theoretical approaches. Cambridge University Press. Anderson, J. R. (1982). Acquisition of cognitive skill. Psychological Review, 89(4), 369. https://​doi.org/​10.1037/​0033-​295X.89.4.369. Andersen, R. (1983). Transfer to somewhere. In S. Gass and L. Selinker (eds.) Language transfer in language learning (pp. 177–​201). Newbury House.Auer, P., & Wei, L. (Eds.). (2008). Handbook of multilingualism and multilingual communication. Walter de Gruyter. Ayoun, D. (2013). The second language acquisition of French tense, aspect, mood and modality. John Benjamins. Bailey, N., Madden, C., & Krashen, S. D. (1974). Is there a “natural sequence” in adult second language learning? Language Learning, 24(2), 235–​243. https://​doi.org/​10.1111/​ j.1467-​1770.1974.tb00505.x. Bartning, I., & Schlyter, S. (2004). Itineraires acquisitionnels et stades de developpement en francais L2. Journal of French Language Studies, 14(3), 281–​299. https://​doi.org/​10.1017/​ S0959269504001802. Bates, E., & MacWhinney, B. (1981). Second Language Acquisition from a Functionalist Perspective: Pragmatic, Semantic, and Perceptual Strategies. Annals of the NewYork Academy of Sciences, 379, 190–​214. https://​doi.org/​10.1111/​j.1749-​6632.1981.tb42009.x. Bates, E., & MacWhinney, B. (1982). A functionalist approach to grammar. In E. Wanner & L. Gleitman (Eds.), Language acquisition: The state of the art (pp. 173–​218). Cambridge University Press.

138

319

138

References  139

Bates, E., & MacWhinney, B. (1987). Competition, variation, and language learning. In B. MacWhinney (Ed.), Mechanisms of language acquisition (pp. 157–​193). Lawrence Erlbaum. Bates, E., & MacWhinney, B. (1989). Functionalism and the Competition Model. In B. MacWhinney & E. Bates (Eds.), The crosslinguistic study of sentence processing (pp. 3–​76). Cambridge University Press. Beckner, C., Ellis, N. C., Blythe, R., Holland, J., Bybee, J., Ke, J., Christiansen, M. H., Larsen-​ Freeman, D., Croft, W., Schoenemann, T., & Five Graces Group (2009). Language is a complex adaptive system: Position paper. Language Learning, 59(S1), 1–​26. Beijing Language Institute (1986). A frequency dictionary of modern Chinese. Beijing Language Institute Press. Bell, P. K. (2017). Explicit and implicit learning: Exploring their simultaneity and immediate effectiveness. Applied Linguistics, 38(3), 297–​317. https://​doi.org/​10.1093/​applin/​ amv028. Benati,A. (2005).The effects of processing instruction, traditional instruction and meaning—​ output instruction on the acquisition of the English past simple tense. Language Teaching Research, 9(1), 67–​93. https://​doi.org/​10.1191/​1362168805lr154oa. Benati, A., & Lee, J. F. (2015). Processing instruction: New insights after twenty years of theory, research and application. International Review of Applied Linguistics in Language Teaching 53(2), 87–​275. https://​doi.org/​10.1515/​iral-​2015-​0004. Bernolet, S., & Hartsuiker, R. J. (2018). Syntactic representations in late learners of a second language. Bilingual Cognition and Language, 54, 205–​224. https://​doi.org/​10.1075/​ sibil.54.10ber. Bernolet, S., Hartsuiker, R. J., & Pickering, M. J. (2007). Shared syntactic representations in bilinguals: Evidence for the role of word-​order repetition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33(5), 931–​949. https://​doi.org/​10.1037/​ 0278-​7393.33.5.931. Best, C. T., & Tyler, M. D. (2007). Nonnative and second-​ language speech perception: Commonalities and complementarities. In O.-​S. Bohn & M. J. Munro (Eds.), Language experience in second language speech learning: In honor of James Emil Flege (pp. 13–​34). John Benjamins. https://​doi.org/​10.1075/​lllt.17.07bes. Bialystok, E., Craik, F. I., Klein, R., & Viswanathan, M. (2004). Bilingualism, aging, and cognitive control: Evidence from the Simon task. Psychology and Aging, 19(2), 290–​303. https://​doi.org/​10.1037/​0882-​7974.19.2.290. Bice, K., & Kroll, J. F. (2015). Native language change during early stages of second language learning. NeuroReport, 26(16), 966–​971. 10.1097/​WNR.0000000000000453. Bobb, S. C., & Kroll, J. F. (2018). Words on the brain: The bilingual mental lexicon. In D. Miller, F. Bayram, J. Rothman, & L. Serratrice (Eds.), Bilingual cognition and language (pp. 307–​324). John Benjamins. Bobb, S. C., & Wodniecka, Z. (2013). Language switching in picture naming: What asymmetric switch costs (do not) tell us about inhibition in bilingual speech planning. Journal of Cognitive Psychology, 25(5), 568–​585. https://​doi.org/​10.1080/​ 20445911.2013.792822. Boersma, P. & Weenink, D. (2011). Praat: Doing Phonetics by Computer. [Computer program]. Retrieved: www.praat.org/​. Bohnemeyer, J., & Swift, M. (2004). Event realization and default aspect. Linguistics and Philosophy, 27, 263–​296. https://​doi.org/​10.1023/​B:LING.0000023371.15460.43. Brown, A. V., Plonsky, L., & Teimouri, Y. (2018). The use of course grades as metrics in L2 research: A systematic review. Foreign Language Annals, 51(4), 763–​778.

410

140 References

Brown, C. A. (1998). The role of the L1 grammar in the L2 acquisition of segmental structure. Second Language Research, 14(2), 136–​193. https://​doi.org/​10.1191/​0267658 98669508401. Brown, C. A. (2000).The interrelation between speech perception and phonological acquisition from infant to adult. In J. Archibald (Ed.), Second language acquisition and linguistic theory (pp. 4–​63). Blackwell. Brysbaert, M., & New, B. (2009). Moving beyond Kučera and Francis: A critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behavior Research Methods, 41(4), 977–​990. https://​doi.org/​10.3758/​BRM.41.4.977. Bybee, J. (2008). Usage-​based grammar and second language acquisition. In P. Robinson & N. C. Ellis (Eds.), Handbook of cognitive linguistics and second language acquisition (pp. 226–​ 246). Routledge. Bybee, J. (2010). Language, usage, and cognition. Cambridge University Press. Byrnes, H. (2013). Editor’s introduction. Modern Language Journal, 97(1), 105–​108. Cadierno,T. (2004). Expressing motion events in a second language: A cognitive typological perspective. In M. Achard & S. Neimeier (Eds.), Cognitive linguistics, second language acquisition and foreign language pedagogy (pp. 13–​49). Mouton de Gruyter. Cadierno, T. (2010). Motion in Danish as a second language: Does the learner’s L1 make a difference? In Z. Han & T. Cadierno (Eds.), Linguistic relativity in SLA: Thinking for speaking (pp. 1–​33). Multilingual Matters. Cadierno, T. (2012). Thinking for speaking in second language acquisition. In C. Chapelle (ed.), The encyclopedia of applied linguistics (pp. 1–​5). Wiley-​Blackwell. Cadierno, T. (2017). Thinking for speaking about motion in a second language. In I. Ibarretxe-​Antuñano (Ed.), Motion and space across languages: Theory and applications (pp. 279–​300). Benjamins. https://​doi.org/​10.1075/​hcp.59.12cad. Calabria, M., Costa, A., Green, D. W., & Abutalebi, J. (2018). Neural basis of bilingual language control. Annals of the New York Academy of Sciences, 1426(1), 221–​235. https://​doi. org/​10.1111/​nyas.13879. Castino, J. (1996). Impact of a phonetics course on FL learners’ acquisition of Spanish phonology. Selecta: Journal of the Pacific Northwest Council on Foreign Languages, 17, 55–​58. Chaudron, C. (2003). Data collection in SLA research. In C. J. Doughty & M. H. Long (Eds.), The handbook of second language acquisition (pp. 762–​828). Wiley. Chen, P., Bobb, S. C., Hoshino, N., & Marian,V. (2017). Neural signatures of language co-​ activation and control in bilingual spoken word comprehension. Brain Research, 1665, 50–​64. https://​doi.org/​10.1016/​j.brainres.2017.03.023. Chomsky, N. (1965). Aspects of a theory of syntax. MIT Press. Christiansen, M., & Chater, N. (2016). Creating language. integrating evolution, acquisition, and processing. MIT Press. Christiner, M., & Reiterer, S. M. (2013). Song and speech: Examining the link between singing talent and speech imitation ability. Frontiers in Psychology, 4, 874. https://​doi.org/​ 10.3389/​fpsyg.2013.00874. Cintrón-​Valentín, M. C., & Ellis, N. C. (2016). Salience in second language acquisition: Physical form, learner attention, and instructional focus. Frontiers in Psychology, 7, 1284. https://​doi.org/​10.3389/​fpsyg.2016.01284. Collins, L., White, J., Trofimovich, P., Cardoso, W., & Horst, M. (2012). When comprehensible input is not comprehensive input: A multi-​dimensional analysis of instructional input in intensive English as a foreign language. In C. Muñoz (Ed.), Intensive exposure experiences in second language learning (pp. 66–​87). Multilingual Matters.

410

14

410

References  141

Comrie, B. (1976). Aspect: An introduction to the study of verbal aspect and related problems. Cambridge University Press. Comrie, B. (1985). Tense. Cambridge University Press. Connell, L., & Lynott, D. (2012). Strength of perceptual experience predicts word processing performance better than concreteness or imageability. Cognition, 125(3), 452–​ 465. https://​doi.org/​10.1016/​j.cognition.2012.07.010. Cook,V. (Ed.). (2003). Effects of the second language on the first. Multilingual Matters. Cornillie, F.,Van Den Noortgate, W.,Van den Branden, K., & Desmet, P. (2017). Examining focused L2 practice: From in vitro to in vivo. Language Learning & Technology, 21(1), 121–​145. Costa, A., & Santesteban, M. (2004). Lexical access in bilingual speech production: Evidence from language switching in highly proficient bilinguals and L2 learners. Journal of Memory and Language, 50(4), 491–​511. https://​doi.org/​10.1016/​j.jml.2004.02.002. Costa, A., Santesteban, M., & Ivanova, I. (2006). How do highly proficient bilinguals control their lexicalization process? Inhibitory and language-​specific selection mechanisms are both functional. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32(5), 1057–​1074. https://​doi.org/​10.1037/​0278-​7393.32.5.1057 Council of Europe. (2001). Common European framework of reference for languages: Learning, teaching, assessment. Cambridge University Press. Crowther, D. (2015). Using what you know: Can cross-​linguistic instruction improve L2 pronunciation? Concordia Working Papers in Applied Linguistics, 6, 27–​45. Cuetos, F., & Mitchell, D. C. (1988). Cross-​linguistic differences in parsing: Restrictions on the use of the Late Closure strategy in Spanish. Cognition, 30(1), 73–​105. https://​doi. org/​10.1016/​0010-​0277(88)90004-​2. Dąbrowska, E. (2012). Different speakers, different grammars: Individual differences in native language attainment. Linguistic Approaches to Bilingualism, 2(3), 219–​253. https://​ doi.org/​10.1075/​lab.2.3.01dab. Dąbrowska, E. (2013). Functional constraints, usage, and mental grammars: A study of speakers’ intuitions about questions with long-​ distance dependencies. Cognitive Linguistics, 24(4), 633–​665. https://​doi.org/​10.1515/​cog-​2013-​0022. Dąbrowska, E. (2014). Recycling utterances: A speaker’s guide to sentence processing. Cognitive Linguistics, 25(4), 617–​653. https://​doi.org/​10.1515/​cog-​2014-​0057. Darcy, I., Mora, J. C., & Daidone, D. (2016). The role of inhibitory control in second language phonological processing. Language Learning, 66(4), 741–​773. https://​doi.org/​ 10.1111/​lang.12161. Davies, M. (2002). Corpus del Español (100 million words). Retrieved March 13, 2021, from www.corpusdelespagnol.org. De Graaff, R., & Housen, A. (2009). Investigating the effects and effectiveness of L2 instruction. In M. Long & C. Doughty (Eds.), The handbook of language teaching (pp. 736–​755). Blackwell. https://​doi.org/​10.1002/​9781444315783.ch38. De Houwer, A., & Ortega, L. (Eds.). (2018). The Cambridge handbook of bilingualism. Cambridge University Press. Declerck, M., & Philipp, A. M. (2015). A sentence to remember: Instructed language switching in sentence production. Cognition, 137, 166–​173. https://​doi.org/​10.1016/​ j.cognition.2015.01.006. DeKeyser, R. M. (1997). Beyond explicit rule learning: Automatizing second language morphosyntax. Studies in Second Language Acquisition, 19(2), 195–​221. https://​doi.org/​ 10.1017/​S0272263197002040.

412

142 References

DeKeyser, R. M. (2009). Cognitive-​psychological processes in second language learning. In M. H. Long & C. J. Doughty (Eds.), The handbook of language teaching (pp. 119–​138). Blackwell Publishing Ltd. https://​doi.org/​10.1002/​9781444315783.ch8. DeKeyser, R. M. (2017). Knowledge and skill in ISLA. In S. Loewen and M. Sato (Eds.), Routledge handbook of instructed second language acquisition (pp. 15–​32). Routledge. DeKeyser, R. M., & Prieto Botana, G. (2015). The effectiveness of processing instruction in L2 grammar acquisition: A narrative review. Applied Linguistics, 36(3), 290–​305. https://​ doi.org/​10.1093/​applin/​amu071. DeKeyser, R. M., & Prieto Botana, G. (2019). Doing SLA research with implications for the classroom: Reconciling methodological demands and pedagogical applicability. John Benjamins. DeLong, K.A., Urbach,T. P., & Kutas, M. (2005). Probabilistic word pre-​activation during language comprehension inferred from electrical brain activity. Nature Neuroscience, 8(8), 1117–​1121. Derwing, T. M., & Munro, M. J. (2015). Pronunciation fundamentals: Evidence-​based perspectives for L2 teaching and research. John Benjamins Publishing Company. Dijkstra, T., Timmermans, M., & Schriefers, H. (2000). On being blinded by your other language: Effects of task demands on interlingual homograph recognition. Journal of Memory and Language, 42(4), 445–​464. https://​doi.org/​10.1006/​jmla.1999.2697. Divjak, D. (2019). Frequency in language: Memory, attention, and learning. Cambridge University Press. Durrell, M. (2011). Hammer’s German grammar and usage (5th ed.). Routledge. Dussias, P. E. (2001). Sentence parsing in fluent Spanish-​English bilinguals. In J. L. Nicol (Ed.), One mind, two languages: Bilingual language processing (pp. 159–​176). Blackwell Publishing. Dussias, P. E. (2003). Syntactic ambiguity resolution in L2 learners: Some effects of bilinguality on LI and L2 processing strategies. Studies in Second Language Acquisition, 25(4), 529–​557. https://​doi.org/​10.1017/​S0272263103000238. Dussias, P. E., & Sagarra, N. (2007). The effect of exposure on syntactic parsing in Spanish-​ English bilinguals. Bilingualism: Language and Cognition, 10(1), 101–​116. https://​doi.org/​ 10.1017/​S1366728906002847. Dussias, P. E., Beatty-​Martínez, A. L., & Perrotti, L. (2017). Susceptibility to interference affects the second and the first language. Bilingualism: Language and Cognition, 20(4), 681–​682. https://​doi.org/​10.1017/​S1366728916001024. Dussias, P. E., Valdes Kroff, J. R., Guzzardo Tamargo, R. E., & Gerfen, C. (2013). When gender and looking go hand in hand. Studies in Second Language Acquisition, 35(2), 353–​ 387. https://​doi.org/​10.1017/​S0272263112000915. Eckman, F. (1991). The structural conformity hypothesis and the acquisition of consonant clusters in the interlanguage of ESL learners. Studies in Second Language Acquisition, 13(1), 23–​41. https://​doi.org/​10.1017/​S0272263100009700. Eisner, F., & McQueen, J. M. (2005). The specificity of perceptual learning in speech processing. Perception & Psychophysics, 67, 224–​238. https://​doi.org/​10.3758/​ BF03206487. Ellis, N. C. (2015). Foreword. Language Learning, 65(S1), v–​vi. Ellis, N. C., Hafeez, K., Martin, K. I., Chen, L., Boland, J., & Sagarra, N. (2014).An eye-​tracking study of learned attention in second language acquisition. Applied Psycholinguistics, 35(3), 547–​579. https://​doi.org/​10.1017/​S0142716412000501. Ellis, N. C. (2006a). Language acquisition as rational contingency learning. Applied Linguistics, 27(1), 1–​24. https://​doi.org/​10.1093/​applin/​ami038.

412

413

412

References  143

Ellis, N. C. (2006b). Cognitive perspectives on SLA: The associative-​cognitive CREED. AILA Review, 19(1), 100–​121. Ellis, N. C., & Robinson, P. (2008). An introduction to cognitive linguistics, second language acquisition, and language instruction. In P. Robinson & N. C. Ellis (Eds.), Handbook of cognitive linguistics and second language acquisition (pp. 3–​24). Routledge. Ellis, N. C., & Sagarra, N. (2010). Learned attention effects in L2 temporal reference: The first hour and the next eight semesters. Language Learning, 60(s2), 85–​108. https://​doi. org/​10.1111/​j.1467-​9922.2010.00602.x. Ellis, N. C., & Sagarra, N. (2011). Learned attention in adult language acquisition: A replication and generalization study and meta-​analysis. Studies in Second Language Acquisition, 33(4), 589–​624. https://​doi.org/​10.1017/​S0272263111000325. Ellis, N. C., Römer, U., & O’Donnell, M. B. (2016). Usage-​based approaches to language acquisition and processing: Cognitive and corpus investigations of construction grammar.Wiley-​Blackwell. Ellis, R. (2001). Introduction: Investigating form-​focused instruction. Language Learning, 51(s1), 1–​46. https://​doi.org/​10.1111/​j.1467-​1770.2001.tb00013.x. Ellis, R., Loewen, S., Elder, C., Erlam, R., Philp, J., & Reinders, H. (2009). Implicit and explicit knowledge in second language learning, testing, and teaching. Multilingual Matters. Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a target letter in a nonsearch task. Perception & Psychophysics, 16(1), 143–​149. https://​doi. org/​10.3758/​BF03203267. Ervins, S. M., & Osgood, C. E. (1954). Second language learning and bilingualism. Journal of Abnormal and Social Psychology, 49, 139–​145. Escudero, P. (2005). Linguistic perception and second language acquisition: Explaining the attainment of optimal phonological categorization. PhD dissertation, Utrecht University. Face,T. L. (2006). Intervocalic rhotic pronunciation by adult learners of Spanish as a second language. In C. A. Klee & T. L. Face (Eds.), Selected proceedings of the 7th conference on the acquisition of Spanish and Portuguese as first and second languages (pp. 47–​58). Cascadilla Proceedings Project. Faris, M. M., Best, C. T., & Tyler, M. D. (2018) Discrimination of uncategorised non-​native vowel contrasts is modulated by perceived overlap within native phonological categories. Journal of Phonetics, 70, 1–​19. https://​doi.org/​10.1016/​j.wocn.2018.05.003. Fernández, C. (2008). Reexamining the role of explicit information in processing instruction. Studies in Second Language Acquisition, 30(3), 277–​305. https://​doi.org/​10.1017/​ S0272263108080467. Flege, J. (1987). The production of “new” and “similar” phones in a foreign language: Evidence for the effect of equivalence classification. Journal of Phonetics, 15(1), 47–​65. https://​doi.org/​10.1016/​S0095-​4470(19)30537-​6. Flege, J. (1991). Age of learning affects the authenticity of voice-​onset time (VOT) in stop consonants produced in a second language. Journal of the Acoustical Society of America, 89(1), 395–​411. https://​doi.org/​10.1121/​1.400473. Flege, J. (1999). Age of learning and second language speech. In D. Birdsong (Ed.), Second language acquisition and the critical period hypothesis (pp. 101–​131). Lawrence Erlbaum. Flege, J. (2003). Assessing constraints on second-​language segmental production and perception. In A. Meyer & N. Schiller (Eds.), Phonetics and phonology in language comprehension and production, differences, and similarities (pp. 319–​355). Mouton de Gruyter. https://​ doi.org/​10.1515/​9783110895094.319. Flege, J. E. (1988).The production and perception of foreign language speech sounds. In H. Winitz (Ed.), Human communication and its disorders: A review (pp. 224–​401). Ablex.

41

144 References

Flege, J. E. (1992). The intelligibility of English vowels spoken by British and Dutch talkers. In R. D. Kent (Ed.), Intelligibility in speech disorders: Theory, measurement, and management (pp. 157–​232). John Benjamins. https://​doi.org/​10.1075/​sspcl.1.06fle. Flege, J. E. (1995). Two procedures for training a novel second language phonetic contrast. Applied Psycholinguistics, 16(4), 425–​442. https://​doi.org/​10.1017/​S0142716400066029. Flege, J. E., & Bohn, O. S. (2021). The revised speech learning model (slm-​r). In Wayland, R. (Ed.), Second Language Speech Learning: Theoretical and Empirical Progress (pp. 3–​83). Cambridge University Press https://​doi.org/​10.1017/​9781108886901. Gass, S. M. (2017). Input, interaction, and the second language learner. Routledge. Gass, S. M., Behney, J., & Plonsky, L. (2020). Second language acquisition. An introductory course. Routledge. Gass, S. M., Behney, J. N., & Uzum, B. (2013). Inhibitory control, working memory, and L2 interaction. In K. Drozdzial-​Szelest & M. Pawlak (Eds.), Psycholinguistic and sociolinguistic perspectives on second language learning and teaching (pp. 91–​114). Springer. Gasser, M. (1990). Connectionism and universals of second language acquisition. Studies in Second Language Acquisition, 12(2), 179–​199. Glisan, E. W., & Donato, R. (2016). Enacting the work of language instruction: High-​leverage teaching practices. ACTFL. Godfroid, A. (2019). Eye tracking in second language acquisition and bilingualism: A research synthesis and methodological guide. Routledge. Goldberg, A. E. (2006). Constructions at work: The nature of generalization in language. Oxford University Press. Goldberg, A. E. (2019). Explain me this: Creativity, competition, and the partial productivity of constructions. Princeton University Press. Goldberg, A. E., & Casenhiser, D. M. (2008). Construction learning and second language acquisition. In N. Ellis & P. Robinson (Eds.), Cognitive linguistics and SLA (pp. 197–​215). Lawrence Erlbaum Press. Goto, H. (1971). Auditory perception by normal Japanese adults of the sounds “L” and “R.” Neuropsychologia, 9(3), 317–​323. https://​doi.org/​10.1016/​0028-​3932(71)90027-​3. Gottfried, A. W. (Ed.). (1984). Home environment and early cognitive development: Longitudinal research. Academic Press. Granger, S. (1998). Prefabricated patterns in advanced EFL writing: Collocations and formulae. In A. P. Cowie (Ed.), Phraseology: Theory, analysis, and applications (pp. 145–​160). Clarendon Press. Green,D.(1998).Mental control of the bilingual lexico-​semantic system.Bilingualism:Language and Cognition, 1(2), 67–​81. https://​doi.org/​10.1017/​S1366728998000133. Green, D. W., & Abutalebi, J. (2013). Language control in bilinguals: The adaptive control hypothesis. Journal of Cognitive Psychology, 25(5), 515–​530. 10.1080/​204459 11.2013.796377. Gries, S.T. (2013). 50-​something years of work on collocations:What is or should be next… International Journal of Corpus Linguistics, 18(1), 137–​166. https://​doi.org/​10.1075/​ ijcl.18.1.09gri. Gries, S. T. (2019). Ten lectures on corpus linguistics with R: Applications for usage-​based and psycholinguistic research. Brill. Grosjean, F. (1988). Exploring the recognition of guest words in bilingual speech. Language and Cognitive Processes, 3(3), 233–​274. 10.1080/​01690968808402089. Grosjean, F. (1989). Neurolinguists, beware! The bilingual is not two monolinguals in one person. Brain and Language, 36(1), 3–​15. https://​doi.org/​10.1016/​0093-​934X (89)90048-​5.

41

145

41

References  145

Grosjean, F. (1997). The bilingual individual. Interpreting, 2(1/​2), 163–​187. https://​doi.org/​ 10.1075/​intp.2.1-​2.07gro. Grüter,T., Lew-​Williams, C., & Fernald, A. (2012). Grammatical gender in L2: A production or a real-​time processing problem? Second Language Research, 28(2), 191–​215. https://​ doi.org/​10.1177/​0267658312437990. Grüter, T., Rohde, H., & Schafer, A. J. (2017). Coreference and discourse coherence in L2: The roles of grammatical aspect and referential form. Linguistic Approaches to Bilingualism, 7(2), 199–​229. https://​doi.org/​10.1075/​lab.15011.gru. Hartsuiker, R. J., & Pickering, M. J. (2008). Language integration in bilingual sentence production. Acta Psychologica (Amst), 128(3), 479–​489. https://​doi.org/​10.1016/​ j.actpsy.2007.08.005. Henry, N., Culman, H., & VanPatten, B. (2009). More on the effects of explicit information in instructed SLA: A partial replication and a response to Fernández (2008). Studies in Second Language Acquisition, 559–​575. 10.1017/​S0272263109990027. Hernandez, A., Li, P., & MacWhinney, B. (2005). The emergence of competing modules in bilingualism. Trends in Cognitive Sciences, 9(5), 220–​225. https://​doi.org/​10.1016/​ j.tics.2005.03.003. Hinkel, E. (2015). Effective curriculum for teaching L2 writing: Principles and techniques. Routledge. Hopp, H. (2013). Grammatical gender in adult L2 acquisition: Relations between lexical and syntactic variability. Second Language Research, 29(1), 33–​56. https://​doi.org/​ 10.1177/​0267658312461803. Hopp, H. (2014). Constraining multiple grammars. Second Language Research, 30(1), 37–​40. https://​doi.org/​10.1177/​0267658313517837. Hopp, H. (2016). Learning (not) to predict: Grammatical gender processing in second language acquisition. Second Language Research, 32(2), 277–​307. https://​doi.org/​10.1177/​ 0267658315624960. Hopp, H. (2019). Cross-​linguistic influence in the child third language acquisition of grammar: Sentence comprehension and production among Turkish-​ German and German learners of English. International Journal of Bilingualism, 23(2), 567–​583. https://​ doi.org/​10.1177/​1367006917752523. Housen, A., & Pierrard, M. (Eds.). (2005). Studies in language acquisition: Investigations in instructed second language acquisition. Walter de Gruyter. Howard, M. (2005). Les contexts prototypiques et marquees de l’emploi de l’imparfait par l’apprenant du francais langue etrangere. In E. Labeau & P. Larrivée (Eds.), Nouveaux développements de l’imparfait (pp. 175–​197). Rodopi. Huensch, A. (2019). The pronunciation teaching practices of university-​level graduate teaching assistants of French and Spanish introductory language courses. Foreign Language Annals, 52(1), 13–​31. https://​doi.org/​10.1111/​flan.12372. Huensch, A. & Nagle, C. (2021).The effect of speaker proficiency on intelligibility, comprehensibility, and accentedness in L2 Spanish: A conceptual replication and extension of Munro and Derwing (1995a). Language Learning. https://​doi.org/​10.1111/​lang.12451. Huensch, A., & Tracy-​ Ventura, N. (2017). Understanding second language fluency behavior: The effects of individual differences in first language fluency, cross-​linguistic differences, and proficiency over time. Applied Psycholinguistics, 38(4), 755–​785. https://​ doi.org/​10.1017/​S0142716416000424. Huensch, A., & Tremblay, A. (2015). Effects of perceptual phonetic training on the perception and production of second language syllable structure. Journal of Phonetics, 52, 105–​120. https://​doi.org/​10.1016/​j.wocn.2015.06.007.

416

146 References

Hwang, H., Shin, J.-​A., & Hartsuiker, R. J. (2018). Late bilinguals share syntax unsparingly between L1 and L2: Evidence from crosslinguistically similar and different constructions. Language Learning, 68(1), 177–​205. https://​doi.org/​10.1111/​ lang.12272. Ingham, J. (2014). Perception, production, and perceptual learning in the second language: A study of perceptual learning by L1 Bengali speakers of L2 English. In H. H. Avilés, M. C. López, D. P. Hughes, et al. (Eds.), The 9th Lancaster University postgraduate conference in linguistics and language teaching (pp. 49–​69). Lancaster: Lancaster University. Ionin,T., Choi, S. H., & Liu, Q. (2021). Knowledge of indefinite articles in L2-​English: Online vs. offline performance. Second Language Research, 37(1), 121–​ 160. https://​doi.org/​ 10.1177/​0267658319857466. Isaacs, T., & Trofimovich, P. (2012). Deconstructing comprehensibility: Identifying the linguistic influences on listeners’ L2 comprehensibility ratings. Studies in Second Language Acquisition, 34(3), 475–​505. https://​doi.org/​10.1017/​S0272263112000150. Izquierdo, J., & Collins, L. (2008). The facilitative role of L1 influence in tense-​aspect marking: A comparison of Hispanophone and Anglophone learners of French. Modern Language Journal, 92(3), 350–​368. https://​doi.org/​10.1111/​j.1540-​4781.2008.00751.x. Jackson, C. (2008). Proficiency level and the interaction of lexical and morphosyntactic information during L2 sentence processing. Language Learning, 58(4), 875–​909. https://​ doi.org/​10.1111/​j.1467-​9922.2008.00481.x. Jarvis, S. (2000). Methodological rigor in the study of transfer: Identifying L1 influence in them interlanguage lexicon. Language Learning, 50(2), 245–​309. https://​doi.org/​ 10.1111/​0023-​8333.00118. Jarvis, S. (2010). Comparison-​based and detection-​based approaches to transfer research. Eurosla Yearbook, 10(1), 169–​192. https://​doi.org/​10.1075/​eurosla.10.10jar. Jarvis, S. (2016). Clarifying the scope of conceptual transfer. Language Learning, 66(3), 608–​ 635. https://​doi.org/​10.1111/​lang.12154. Jarvis, S., & Pavlenko, A. (2008). Crosslinguistic influence in language and cognition. Routledge. Jegerski, J. (2014). Self-​paced reading. In J. Jergerski & B.VanPatten (Eds.), Research methods in second language psycholinguistics (pp. 20–​49). Routledge. Jiang, N. (2002). Form-​ meaning mapping in vocabulary acquisition in a second language. Studies in Second Language Acquisition, 24(4), 617–​637. https://​doi.org/​10.1017/​ S0272263102004047. Jiang, N. (2004). Semantic transfer and development in adult L2 vocabulary acquisition. In P. Bogaards & B. Laufer (Eds.), Vocabulary in a second language: Selection, acquisition, and testing (pp. 101–​126). John Benjamins. https://​doi.org/​10.1075/​lllt.10.09jia. Jiang, N. (2013). Conducting reaction time research in second language studies. Routledge. Jiang, N., Li, M., & Guo, T. (2020). A tale of two frequency effects: Toward a verification model of L2 word recognition. Applied Psycholinguistics, 41(1), 215–​236. https://​doi.org/​ 10.1017/​S0142716419000481. Jordan, G. (2004). Theory construction in second language acquisition. John Benjamins. Juffs, A., & Harrington, M. (2011). Aspects of working memory in L2 learning. Language Teaching, 44(2), 137–​166. https://​doi.org/​10.1017/​S0261444810000509. Kaan, E. (2014). Predictive sentence processing in L2 and L1: What is different? Linguistic Approaches to Bilingualism, 4(2), 257–​282. https://​doi.org/​10.1075/​lab.4.2.05kaa. Kang, C., Ma, F., & Guo, T. (2018). The plasticity of lexical selection mechanism in word production: ERP evidence from short-​term language switching training in unbalanced Chinese-​English bilinguals. Bilingualism: Language and Cognition, 21(2), 296–​313. https://​ doi.org/​10.1017/​S1366728917000037.

416

147

416

References  147

Kang, E.Y., Sok, S., & Han, Z. (2019). Thirty-​five years of ISLA on form-​focused instruction: A meta-​analysis. Language Teaching Research, 23(4), 428–​453. https://​doi.org/​ 10.1177/​1362168818776671. Kartushina, N., & Frauenfelder, U. H. (2014). On the effects of L2 perception and of individual differences in L1 production on L2 pronunciation. Frontiers in Psychology, 5(1246), 1–​17. https://​doi.org/​10.3389/​fpsyg.2014.01246. Kasprowicz, R., & Marsden, E. (2018). Towards ecological validity in research into input-​ based practice: Form spotting can be as beneficial as form-​meaning practice. Applied Linguistics, 39(6), 886–​911. https://​doi.org/​10.1093/​applin/​amw051. Kassaian, Z. (2011). Age and gender effect in phonetic perception and production. Journal of Language Teaching and Research, 2(2), 370–​376. 10.4304/​jltr.2.2.370-​376. Keck, C., & Kim,Y. (2014). Pedagogical grammar. John Benjamins. Kellerman, E. (1995). Crosslinguistic influence: Transfer to nowhere? Annual Review of Applied Linguistics, 15, 125–​150. https://​doi.org/​10.1017/​S0267190500002658. Kempe,V. (1999). Processing of morphological and semantic cues in Russian and German. Language and Cognitive Processes, 14(2), 129–​171. 10.1080/​016909699386329. Kempe, V., & Brooks, P. J. (2016). Miniature natural language learning in L2 acquisition research. In G. Grañena, D. O. Jackson, & Y. Yilmaz (Eds.), Cognitive individual differences in L2 processing and acquisition (pp. 39–​65). John Benjamins. Kennedy, S., & Trofimovich, P. (2008). Intelligibility, comprehensibility, and accentedness of L2 Speech: The role of listener experience and semantic context. The Canadian Modern Language Review, 64(3), 459–​489. http://​dx.doi.org.ezaccess.libraries.psu.edu/​10.3138/​ cmlr.64.3.459. Kennedy, S., & Trofimovich, P. (2017). Pronunciation acquisition. In S. Loewen & M. Sato (Eds.), The Routledge handbook of instructed second language acquisition (pp. 260–​ 279). Routledge. Kihlstedt, M. (1998). La référence au passé dans le dialogue: Etude de l’acquisition de la temporalité chez des apprenants dits avancés de français. (Cahiers de la Recherche 6). Stockholms University. Kihlstedt, M. (2002). Reference to past events in dialogue: The acquisition of tense and aspect by advanced learners of French. In M. R. Salaberry & Y. Shirai (Eds.), The L2 acquisition of tense-​aspect morphology (pp. 323–​361). John Benjamins. https://​doi.org/​ 10.1075/​lald.27.14kih. Kihlstedt, M. (2015). Acquisition of the imparfait in L2 French in adults and children: The same or different? A comparative case study of Swedish adults and children in an immersion setting. Language, Interaction and Acquisition, 6(1), 74–​106. Kissling, E. M. (2013). Teaching pronunciation: Is explicit phonetics instruction beneficial for FL learners? Modern Language Journal, 97(3), 720–​744. https://​doi.org/​10.1111/​ j.1540-​4781.2013.12029.x. Klein, W. (1994). Time in language. Routledge. Klein, W., Li, P., & Hendriks, H. (2000). Aspect and assertion in Mandarin Chinese. Natural Language & Linguistic Theory, 18, 723–​770. https://​doi.org/​10.1023/​ A:1006411825993. Knight, S. (1994). Dictionary use while reading:The effects on comprehension and vocabulary acquisition for students of different verbal abilities. The Modern Language Journal, 78(3), 285–​299. Kohonen, T. (1990). The self-​organizing map. Proceedings of the IEEE, 78(9), 1464–​1480. 10.1109/​5.58325. Krashen, S. (1985). The input hypothesis: Issues and implications. Longman.

418

148 References

Kroll, J., Dussias, P., & Bajo, M. (2018). Language use across international contexts: Shaping the minds of L2 speakers. Annual Review of Applied Linguistics, 38, 60–​79. https://​doi. org/​10.1017/​S0267190518000119. Kroll, J., Dussias, P., Bice, K., & Perrotti, L. (2015). Bilingualism, mind, and brain. Annual Review of Linguistics, 1(1), 377–​394. https://​doi.org/​10.1146/​annurevlinguist-​030514-​124937. Kroll, J. F., & Bialystok, E. (2013). Understanding the consequences of bilingualism for language processing and cognition. Journal of Cognitive Psychology, 25(5), 10.1080/​ 20445911.2013.799170. Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33(2), 149–​174. https://​doi.org/​ 10.1006/​jmla.1994.1008. Kroll, J. F., Dussias, P. E., Bogulski, C. A., & Kroff, J. R.V. (2012). Juggling two languages in one mind: What bilinguals tell us about language processing and its consequences for cognition. Psychology of Learning and Motivation, 56, 229–​262. https://​doi.org/​10.1016/​ B978-​0-​12-​394393-​4.00007-​8. Kroll, J. F., Michael, E., Tokowicz, N., & Dufour, R. (2002). The development of lexical fluency in a second language. Second Language Research, 18(2), 137–​171. https://​doi.org/​ 10.1191/​0267658302sr201oa. Kupferberg, I. (1999). The cognitive turn of contrastive analysis: Empirical evidence. Language Awareness, 8(3–​4), 210–​222. https://​doi.org/​10.1080/​09658419908667130. Kupferberg, I., & Olshtain, E. (1996). Explicit contrastive instruction facilitates the acquisition of difficult L2 forms. Language Awareness, 5(3–​4), 149–​165. https://​doi.org/​ 10.1080/​09658416.1996.9959904. Labeau, E. (2005). Beyond the Aspect Hypothesis: Tense–​aspect development in advanced L2 French. EUROSLA Yearbook, 5(1), 77–​101. Ladefoged, P. (2001). Vowels and consonants: An introduction to the sounds of languages. Blackwell Publishers. Lado, R. (1957). Linguistics across cultures: Applied linguistics and language teachers. University of Michigan Press. Lantolf, J. P., & Poehner, M. E. (Eds.). (2008). Sociocultural theory and the teaching of second languages. Equinox. Lantolf, J. P., & Zhang, X. (2017). Concept-​based language instruction. In S. Loewen & M. Sato (Eds.), The Routledge handbook of instructed second language acquisition (pp. 146–​165). Taylor & Francis. Lardiere, D. (2009). Some thoughts on the contrastive analysis of features in second language acquisition. Second Language Research, 25(2), 173–​227. https://​doi.org/​10.1177/​ 0267658308100283. Larrañaga, P., Treffers-​Daller, J., Tidball, F., & Ortega, M. G. (2012). L1 transfer in the acquisition of manner and path in Spanish by native speakers of English. International Journal of Bilingualism, 16(1), 117–​138. https://​doi.org/​10.1177/​1367006911405577. Larsen-​Freeman, D. (2013). Transfer of learning transformed. Language Learning, 63, 107–​ 129. https://​doi.org/​10.1111/​j.1467-​9922.2012.00740.x. Larson-​Hall, J. (2004). Predicting perceptual success with segments: A test of Japanese speakers of Russian. Second Language Research, 20(1), 33–​76. https://​doi.org/​10.1191/​ 0267658304sr230oa.

418

419

418

References  149

Laufer, B., & Girsai, N. (2008). Form-​focused instruction in second language vocabulary learning: A case for contrastive analysis and translation. Applied Linguistics, 29(4), 694–​ 716. https://​doi.org/​10.1093/​applin/​amn018. Le Ha, P., & Baurain, B. (Eds.). (2011). Voices, identities, negotiations, and conflicts: Writing academic English across cultures (Vol. 22). Brill. Leclercq, P., Edmonds, A., & Hilton, H. (Eds.). (2014). Measuring L2 proficiency: Perspectives from SLA. Multilingual Matters. Lee, S. (2009). Interpreting scope ambiguity in first and second language processing: Universal quantifiers and negation [Unpublished doctoral dissertation]. University of Hawai’i at Manoa. Leow, R. P. (2015). Explicit learning in the L2 classroom: A student-​ centered approach. Routledge. Lev-​Ari, S., & Peperkamp, S. (2013). Low inhibitory skill leads to non-​native perception and production in bilinguals’ native language. Journal of Phonetics, 41(5), 320–​331. https://​ doi.org/​10.1016/​j.wocn.2013.06.002. Levis, J. M. (2005). Changing contexts and shifting paradigms in pronunciation teaching. TESOL Quarterly, 39(3), 369–​377. https://​doi.org/​10.2307/​3588485. Levis, J. M. (2020). Revisiting the Intelligibility and Nativeness Principles. Journal of Second Language Pronunciation, 6(3), 310–​328. https://​doi.org/​10.1075/​jslp.20050.lev. Levy, E. S. (2009). Language experience and consonantal context effects on perceptual assimilation of French vowels by American–​English learners of French. The Journal of the Acoustical Society of America, 125(2), 1138–​1152. https://​doi.org/​10.1121/​1.3050256. Levy, E. S., & Law, F. F., II. (2010). Production of French vowels by American-​English learners of French: Language experience, consonantal context, and the perception-​ production relationship. The Journal of the Acoustical Society of America, 128(3), 1290–​1305. 10.1121/​1.3466879 Li, M., & DeKeyser, R. (2017). Perception practice, production practice, and musical ability in L2 Mandarin tone-​word learning. Studies in Second Language Acquisition, 39(4), 593–​ 620. https://​doi.org/​10.1017/​S0272263116000358. Li, P., Farkas I., & MacWhinney, B. (2004). Early lexical development in a self-​organizing neural network. Neural Networks, 17(8–​9), 1345–​1362. https://​doi.org/​10.1016/​ j.neunet.2004.07.004. Lightbown, P. M., & Spada, N. (2013). How languages are learned (4th ed.). Oxford University Press. Linck, J. A., Hoshino, N., & Kroll, J. F. (2008). Cross-​language lexical processes and inhibitory control. The Mental Lexicon, 3(3), 349–​374. https://​doi.org/​10.1075/​ml.3.3.06lin. Linck, J. A., Osthus, P., Koeth, J. T., & Bunting, M. F. (2014). Working memory and second language comprehension and production: A meta-​ analysis. Psychonomic Bulletin & Review, 21(4), 861–​883. https://​doi.org/​10.3758/​s13423-​013-​0565-​2. Linck, J. A., Schwieter, J. W., & Sunderman, G. (2012). Inhibitory control predicts language switching performance in trilingual speech production. Bilingualism: Language and Cognition, 15(3), 651–​662. https://​doi.org/​10.1017/​S136672891100054X. Linck, J. A., & Weiss, D. J. (2015). Can working memory and inhibitory control predict second language learning in the classroom? Sage Open, 5(4), 1–​ 11. 10.1177/​ 2158244015607352. Liu, H., & Cao, F. (2016). L1 and L2 processing in the bilingual brain: A meta-​analysis of neuroimaging studies. Brain and Language, 159, 60–​73. https://​doi.org/​10.1016/​ j.bandl.2016.05.013.

510

150 References

Loewen, S. (2020). Introduction to instructed second language acquisition (2nd ed.). Routledge. Logan, G. D. (1994). On the ability to inhibit thought and action: A users’ guide to the stop signal paradigm. In D. Dagenbach & T. H. Carr (Eds.), Inhibitory processes in attention, memory, and language (pp. 189–​239). Academic Press. Long, M. (1991). Focus on form: A design feature in language teaching methodology. In K. De Bot, R. Ginsberg, & C. Kramsch (Eds.), Foreign language research in cross-​cultural perspectives (pp. 39–​52). John Benjamins. Lopez-​Soto, T., & Kewley-​Port, D. (2009, May). Relation of perception training to production of codas in English as a second language. In Proceedings of Meetings on Acoustics 157ASA (Vol. 6, No. 1, p. 062003). Acoustical Society of America. Lord, G. (2005). (How) can we teach foreign language pronunciation? On the effects of a Spanish phonetics course. Hispania, 88(3), 557–​567. Lucas, M. (2020). Cross-​linguistic influence and plural marking: Enhancing error recognition and written accuracy among Japanese EFL learners through web-​based awareness-​ raising. Language Awareness, 29(2), 155–​177. 10.1080/​09658416.2020.1737096. Luppescu, S., & Day, R. R. (1993). Reading, dictionaries, and vocabulary learning. Language Learning, 43(2), 263–​279. https://​doi.org/​10.1111/​j.1467-​1770.1992.tb00717.x. Lyster, R. (1994).The effect of functional-​analytic teaching on aspects of French immersion students’ sociolinguistic competence. Applied Linguistics, 15(3), 263–​287. https://​doi. org/​10.1093/​applin/​15.3.263. Mackey, A. (2017). Classroom-​based research. In S. Loewen & M. Sato (Eds.), The Routledge handbook of instructed second language acquisition (pp. 541–​561). Routledge. Mackey, A., & Gass, S. M. (2016). Second language research: Methodology and design. Routledge. MacWhinney, B. (1987). The competition model. In B. MacWhinney (Ed.), Mechanisms of language acquisition (pp. 249–​308). Lawrence Erlbaum Associates, Inc. MacWhinney, B. (1992). Transfer and competition in second language learning. In R. J. Harris (Ed.), Advances in psychology, 83. Cognitive processing in bilinguals (pp. 371–​390). North-​Holland. https://​doi.org/​10.1016/​S0166-​4115(08)61506-​X. MacWhinney, B. (1997). Second language acquisition and the Competition Model. In A. M. B. de Groot & J. F. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives (pp. 113–​142). Lawrence Erlbaum Associates Publishers. MacWhinney, B. (2005). A unified model of language acquisition. In J. F. Kroll & A. M. B. De Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp. 49–​67). Oxford University Press. MacWhinney, B. (2008). A unified model. In P. Robinson & N. C. Ellis (Eds.), Handbook of cognitive linguistics and second language acquisition (pp. 341–​371). Routledge. MacWhinney, B. (2012). The logic of the Unified Model. In S. Gass & A. Mackey (Eds.), The Routledge handbook of second language acquisition (pp. 211–​227). Routledge. MacWhinney, B. (2017). A unified model of first and second language learning. In M. Hickmann, E. Veneziano, & H. Jisa (Eds.), Sources of variability in first language acquisition: Languages, contexts, and learners (pp. 298–​312). John Benjamins. MacWhinney, B. (2019). Language attrition and the competition model. In M. S. Schmid & B. Kopke (Eds.), The Oxford handbook of language attrition (pp. 11–​ 17). Oxford University Press. MacWhinney, B., Leinbach, J.,Taraban, R., & McDonald, J. (1989). Language learning: Cues or rules? Journal of Memory and Language, 28(3), 255–​277. Major, R. C. (1987). Phonological similarity, markedness, and rate of L2 acquisition. Studies in Second Language Acquisition, 9(1), 63–​82. https://​doi.org/​10.1017/​S0272263100006513.

510

15

510

References  151

Major, R. C., & Kim, E. (1996).The similarity differential rate hypothesis. Language Learning, 46(3), 465–​496. https://​doi.org/​10.1111/​j.1467-​1770.1996.tb01244.x. Marsden, E., Thompson, S., & Plonsky, L. (2018). A methodological synthesis of self-​paced reading in second language research. Applied Psycholinguistics, 39(5), 861–​904. https://​ doi.org/​10.1017/​S0142716418000036. McAllister, R., Flege, J. E., & Piske, T. (2002). The influence of L1 on the acquisition of Swedish quantity by native speakers of Spanish, English, and Estonian. Journal of Phonetics, 30(2), 229–​258. https://​doi.org/​10.1006/​jpho.2002.0174. McDonough, K. (2017). Experimental research methods. In S. Loewen & M. Sato (Eds.), The Routledge handbook of instructed second language acquisition (pp. 562–​576). Routledge. McManus, K. (2013). Prototypical influence in second language acquisition: What now for the Aspect Hypothesis? International Review of Applied Linguistics in Language Teaching, 51(3), 299–​322. https://​doi.org/​10.1515/​iral-​2013-​0013. McManus, K. (2015). L1-​L2 differences in the acquisition of form-​meaning pairings:A comparison of English and German learners of French. Canadian Modern Language Review, 71(2), 51–​77. http://​dx.doi.org.ezaccess.libraries.psu.edu/​10.3138/​cmlr.2070.51 McManus, K. (2019a). Awareness of L1 form-​meaning mappings can reduce crosslinguistic effects in L2 grammatical learning. Language Awareness, 28(2), 114–​ 138. 10.1080/​ 09658416.2019.1620756. McManus, K. (2019b). Relationships between social networks and language development during study abroad. Language, Culture and Curriculum, 32(3), 270–​284. 10.1080/​ 07908318.2019.1661683. McManus, K. (2020). Crosslinguistic influence and L2 grammar learning: Proposed replications of Ellis and Sagarra (2011) and Tolentino and Tokowicz (2014). Language Teaching, 1–​9. doi:10.1017/​S0261444820000270. McManus, K. (2021). Examining the effectiveness of language-​switching practice for reducing cross-​language competition in L2 grammatical processing. Bilingualism: Language and Cognition, 24(1), 167–​184. McManus, K., & Liu, Y. (2020). Using elicited imitation to measure global oral proficiency in SLA research: A close replication study. Language Teaching, 1-​ 20. 10.1017/​ S026144482000021X. McManus, K., & Marsden, E. (2017). L1 explicit instruction can improve L2 online and offline performance. Studies in Second Language Acquisition, 39(3), 459–​492. https://​doi. org/​10.1017/​S027226311600022X. McManus, K., & Marsden, E. (2018). Online and offline effects of L1 practice in L2 grammar learning: A partial replication. Studies in Second Language Acquisition, 40(2), 459–​475. https://​doi.org/​10.1017/​S0272263117000171. McManus, K., & Marsden, E. (2019a). Using explicit instruction about L1 to reduce crosslinguistic effects in L2 grammar learning: Evidence from oral production in L2 French. The Modern Language Journal, 103(2), 459–​480. https://​doi.org/​10.1111/​ modl.12567. McManus, K., & Marsden, E. (2019b). Signatures of automaticity during practice: Explicit instruction about L1 processing routines can improve L2 grammatical processing. Applied Psycholinguistics, 40(1), 205–​234. https://​doi.org/​10.1017/​S0142716418000553. McManus, K., Mitchell, R., & Tracy-​Ventura, N. (2021). A longitudinal study of advanced learners’ linguistic development before, during, and after study abroad. Applied Linguistics, 42(1), 136–​163. https://​doi.org/​10.1093/​applin/​amaa003. Meara, P. (2005). LLAMA language aptitude tests:The manual. Swansea: Lognostics.

512

152 References

Meuter, R. F. I., & Allport, A. (1999). Bilingual language switching in naming: Asymmetrical costs of language selection. Journal of Memory and Language, 40(1), 25–​40. https://​doi. org/​10.1006/​jmla.1998.2602. Miller, D., Bayram, F., Rothman, J., & Serratrice, L. (Eds.). (2018). Bilingual cognition and language. John Benjamins. Mitchell, R. F.,Tracy-​Ventura, N., & McManus, K. (2017). Anglophone students abroad: Identity, social relationships, and language learning. Routledge. Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41(1), 49–​100. https://​doi.org/​10.1006/​cogp.1999.0734. Monsell, S. (1996). Control of mental processes. In V. Bruce (Ed.), Unsolved mysteries of the mind:Tutorial essays in cognition (pp. 93–​148). Psychology Press. Montrul, S. (2010). Current issues in heritage language acquisition. Annual Review of Applied Linguistics, 30, 3–​23. Morgan-​Short, K. (2014). Electrophysiological approaches to understanding second language acquisition: A field reaching its potential. Annual Review of Applied Linguistics, 34, 15–​36. https://​doi.org/​10.1017/​S026719051400004X. Munro, M. J., & Derwing, T. M. (1995a). Foreign accent, comprehensibility, and intelligibility in the speech of second language learners. Language Learning, 45(1), 73–​97. https://​doi.org/​10.1111/​j.1467-​1770.1995.tb00963.x. Munro, M. J., & Derwing, T. M. (1995b). Processing time, accent, and comprehensibility in the perception of native and foreign-​accented speech. Language and Speech, 38(3), 289–​ 306. https://​doi.org/​10.1177/​002383099503800305. Nation, I. S. P., & Webb, S. A. (2011). Researching and analyzing vocabulary. Heinle, Cengage Learning. Nesselhauf, N. (2003). The use of collocations by advanced learners of English and some implications for teaching. Applied Linguistics, 24(2), 223–​242. https://​doi.org/​10.1093/​ applin/​24.2.223. Nesselhauf, N. (2005). Collocations in a learner corpus. John Benjamins Publishing. https://​ doi.org/​10.1075/​scl.14. Nokes, T. J. (2009). Mechanisms of knowledge transfer. Thinking & reasoning, 15(1), 1–​36. https://​doi.org/​10.1080/​13546780802490186. Norris, D., McQueen, J. M., & Cutler, A. (2003). Perceptual learning in speech. Cognitive Psychology, 47(2), 204–​238. https://​doi.org/​10.1016/​S0010-​0285(03)00006-​9. Norris, J. M., & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quantitative meta-​analysis. Language Learning, 50(3), 417–​528. https://​doi.org/​10.1111/​ 0023-​8333.00136. O’Grady, W. (2012). Language acquisition without an acquisition device. Language Teaching, 45(1), 116–​130. https://​doi.org/​10.1017/​S026144481000056X. O’Grady,W. (2013).The illusion of language acquisition. Linguistic Approaches to Bilingualism, 3(3), 253–​285. https://​doi.org/​10.1075/​lab.3.3.01ogr. O’Grady, W. (2015). Processing determinism. Language Learning, 65(1), 6–​32. https://​doi. org/​10.1111/​lang.12091. O’Grady, W., Yamashita, Y., & Lee, S. Y. (2005). A note on canonical word order. Applied Linguistics, 26(3), 453–​458. https://​doi.org/​10.1093/​applin/​ami019. Odlin, T. (1989). Language transfer: Cross-​linguistic influence in language learning. Cambridge University Press.

512

513

512

References  153

Odlin, T. & Yu, L. (2016). Introduction. In L.Yu & T. Odlin (Eds.), New perspectives on transfer in second language learning (pp. 1–​16). Multilingual Matters. Ortega, L. (2009). Understanding second language acquisition. Routledge. Ortega, L., & Iberri-​Shea, G. (2005). Longitudinal research in second language acquisition: Recent trends and future directions. Annual Review of Applied Linguistics, 25, 26–​45. https://​doi.org/​10.1017/​S0267190505000024. Palma, P., & Titone, D. (2020). Something old, something new: A review of the literature on sleep-​related lexicalization of novel words in adults. Psychonomic Bulletin & Review, 28, 96–​121. https://​doi.org/​10.3758/​s13423-​020-​01809-​5. Paradis, M. (1981). Neurolinguistic organization of a bilingual’s two languages. In J. E. Copeland & P. W. Davis (Eds.), The seventh LACUS Forum (pp. 486–​ 494). Hornbeam Press. Paradis, M. (1993). Linguistic, psycholinguistic, and neurolinguistic aspects of “interference” in bilingual speakers: The activation threshold hypothesis. International Journal of Psycholinguistics, 9(2), 133–​145. Paribakht, T., & Wesche, M. (1993). Reading comprehension and second language development in a comprehension-​based ESL program. TESL Canada Journal, 11(1), 09–​29. https://​doi.org/​10.18806/​tesl.v11i1.623. Pavlenko, A., & Jarvis, S. (2002). Bidirectional transfer. Applied Linguistics, 23(2), 190–​214. https://​doi.org/​10.1093/​applin/​23.2.190. Peperkamp, S., & Bouchon, C. (2011). The relation between perception and production in L2 phonological processing. Interspeech, 12, 161–​164. Pienemann, M., Di Biase, B., Kawaguchi, S., & Håkansson, G. (2005). Processing constraints on L1 transfer. Handbook of bilingualism: Psycholinguistic approaches, 128–​153. Piske,T., MacKay, I., & Flege, J. (2001). Factors affecting degree of foreign accent in an L2: A review. Journal of Phonetics, 29(2), 191–​215. https://​doi.org/​10.1006/​jpho.2001.0134. Plonsky, L. (2017). Quantitative research methods in instructed SLA. In S. Loewen & M. Sato (Eds.), The Routledge handbook of instructed second language acquisition (pp. 505–​ 521). Routledge. Poarch, G. J., & van Hell, J. G. (2012). Executive functions and inhibitory control in multilingual children: Evidence from second-​language learners, bilinguals, and trilinguals. Journal of Experimental Child Psychology, 113(4), 535–​551. https://​doi.org/​10.1016/​ j.jecp.2012.06.013. Porte, G., & McManus, K. (2019). Doing replication research in applied linguistics. Routledge. Poulisse, N. (1999). Slips of the tongue: Speech errors in first and second language production. John Benjamins Publishing Company. Presson, N., Macwhinney, B., & Tokowicz, N. (2014). Learning grammatical gender: The use of rules by novice learners. Applied Psycholinguistics, 35(4), 709–​737. https://​doi.org/​ 10.1017/​S0142716412000550. Pulido, M. F., & Dussias, P. E. (2020). Desirable difficulties while learning collocations in a second language: Conditions that induce L1 interference improve learning. Bilingualism: Language & Cognition, 23(3), 652–​667. https://​doi.org/​10.1017/​ S1366728919000622. Purcell, E. T., & Suter, R. W. (1980). Predictors of pronunciation accuracy: A reexamination. Language Learning, 30(2), 271–​287. https://​doi.org/​10.1111/​j.1467-​1770.1980. tb00319.x. Rebuschat, P. (2013). Measuring implicit and explicit knowledge in second language research. Language Learning, 63(3), 595–​626. https://​doi.org/​10.1111/​lang.12010.

514

154 References

Reed, H. B., & Dick, R. D. (1968).The learning and generalization of abstract and concrete concepts. Journal of Verbal Learning and Verbal Behavior, 7(2), 486–​490. https://​doi.org/​ 10.1016/​S0022-​5371(68)80037-​4. Reiterer, S. M., Hu, X., Erb, M., Rota, G., Nardo, D., Grodd, W., Winkler, S., & Ackermann, H. (2011). Individual differences in audio-​ vocal speech imitation aptitude in late bilinguals: Functional neuro-​ imaging and brain morphology. Frontiers in Psychology 2(271), 1–​12. https://​doi.org/​10.3389/​fpsyg.2011.00271. Roberts, L., & Liszka, S. A. (2013). Processing tense/​ aspect-​ agreement violations on-​ line in the second language: A self-​paced reading study with French and German L2 learners of English. Second Language Research, 29(4), 413–​439. https://​doi.org/​10.1177/​ 0267658313503171. Roberts, L., & Liszka, S. A. (2020). Grammatical aspect and L2 learners’ online processing of temporarily ambiguous sentences in English: A self-​paced reading study with German, Dutch and French L2 learners. Second Language Research. https://​doi.org/​10.1177/​ 0267658319895551. Roberts, L., Gullberg, M., & Indefrey, P. (2008). Online pronoun resolution in L2 discourse: L1 influence and general learner effects. Studies in Second Language Acquisition, 30(3), 333–​357. https://​doi.org/​10.1017/​S0272263108080480. Robinson, P. (Ed.). (2001). Cognition and second language instruction. Cambridge University Press. Roehr-​ Brackin, K. (2020). Investigating explicit and implicit L2 knowledge and learning: Replications of Erlam (2005) and Roehr-​Brackin and Tellier (2019). Language Teaching, 1–​13. 10.1017/​S0261444820000415. Roehr-​Brackin, K., (2018). Metalinguistic awareness and second language acquisition. Routledge. Rogers, J. (2021). Input spacing in second language classroom settings: Replications of Bird (2010) and Serrano (2011). Language Teaching, 1–​10. 10.1017/​S0261444820000439. Rogers, J., Revesz, A., & Rebuschat, P. (2015). Challenges in implicit learning research: Validating a novel artificial language. In P. Rebuschat (Ed.), Implicit and explicit learning of languages (pp. 275–​300). John Benjamins. Rothman, J., & Slabakova, R. (2018). The generative approach to SLA and its place in modern second language studies. Studies in Second Language Acquisition, 40(2), 417–​442. https://​doi.org/​10.1017/​S0272263117000134. Safronova, E., & Mora, J. C. (2013). Attention control in L2 phonological acquisition. In A. Llanes Baró, L. Astrid Ciro, L. Gallego Balsà, & R. M. Mateus Serra (Eds.), Applied linguistics in the age of globalization (pp. 109–​114). Lleida: Edicions de la Universitat de Lleida. Saito, K. (2011). Examining the role of explicit phonetic instruction in native-​like and comprehensible pronunciation development: An instructed SLA approach to L2 phonology. Language Awareness, 20(1), 45–​59. 10.1080/​09658416.2010.540326. Saito, K., Trofimovich, P., & Isaacs, T. (2016). Second language speech production: Investigating linguistic correlates of comprehensibility and accentedness for learners at different ability levels. Applied Psycholinguistics, 37(2), 217–​240. https://​doi. org/​10.1017/​S0142716414000502. Sanoudaki, E., & Thierry, G. (2014). Juggling two grammars. In E. M.Thomas & I. Mennen (Eds.), Advances in the study of bilingualism (pp. 214–​230). Multilingual Matters. https://​ doi.org/​10.21832/​9781783091713-​013. Sanz, C., & Morales-​Front, A. (Eds.). (2018). The Routledge handbook of study abroad research and practice. Routledge.

514

15

514

References  155

Sanz, C., & Morgan-​Short, K. (2004). Positive evidence versus explicit rule presentation and explicit negative feedback: A computer-​assisted study. Language Learning, 54(1), 35–​78. https://​doi.org/​10.1111/​j.1467-​9922.2004.00248.x. Sato, M., & Loewen, S. (Eds.). (2019). Evidence-​ based second language pedagogy: A collection of instructed second language acquisition studies. Routledge. Schmid, M. S., & Köpke, B. (Eds.). (2019). The Oxford handbook of language attrition. Oxford University Press. 10.1093/​oxfordhb/​9780198793595.001.0001. Schmitt, N. (2010). Researching vocabulary: A vocabulary research manual. Palgrave Macmillan. 10.1057/​9780230293977. Schmitt, N., & Schmitt, D. (2020). Vocabulary in language teaching (2nd ed.). Cambridge University Press. Schmidt, R. W. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2), 129–​158. https://​doi.org/​10.1093/​applin/​11.2.129. Schuhmann, K. S. (2014). Perceptual learning in second language learners [Unpublished doctoral dissertation]. Stony Brook, NY: State University of New York. Schuhmann, K. S. (2016). Cross-​linguistic perceptual learning in advanced second language listeners. Proceedings of the Linguistic Society of America, 1(31), 1–​15. https://​doi.org/​ 10.3765/​plsa.v1i0.3731. Schwartz, B. D., & Sprouse, R. (1994). Word order and nominative case in nonnative language acquisition: A longitudinal study of (L1 Turkish) German interlanguage. In T. Hoekstra & B. D. Schwartz (Eds.), Language acquisition studies in generative grammar (pp. 317–​368). John Benjamins. Schwartz, B. D., & Sprouse, R. (1996). L2 cognitive states and the Full Transfer/​Full Access model. Second Language Research, 12(1), 40–​72. https://​doi.org/​10.1177/​ 026765839601200103. Schwartz, B. D., & Sprouse, R. (2020). In defense of ‘copying and restructuring’. Second Language Research. https://​doi.org/​10.1177/​0267658320975831. Schwieter, J. W., & Sunderman, G. (2008). Language switching in bilingual speech production: In search of the language-​specific selection mechanism. The Mental Lexicon, 3(2), 214–​238. https://​doi.org/​10.1075/​ml.3.2.06sch. Segalowitz, N. (2010). Cognitive bases of second language fluency. Routledge. Segalowitz, N., & Lightbown, P. M. (1999). Psycholinguistic approaches to SLA. Annual Review of Applied Linguistics, 19, 43–​63. https://​doi.org/​10.1017/​S0267190599190032. Sharwood Smith, M., & Truscott, J. (2006). Full transfer full access: A processing-​oriented interpretation. In S. Unsworth,T. Parodi, A. Sorace, & M.Young-​Scholten (Eds.), Paths of development in L1 and L2 acquisition (pp. 201–​216). Benjamins. https://​doi.org/​10.1075/​ lald.39.10sha. Sharwood Smith, M., & Truscott, J. (2014). The multilingual mind. Cambridge University Press. https://​doi.org/​10.1017/​CBO9781139644044. Shintani, N. (2015). The incidental grammar acquisition in focus on form and focus on forms instruction for young beginner learners. TESOL Quarterly, 49(1), 115–​140. https://​doi.org/​10.1002/​tesq.166. Shirai,Y. (2019). Connectionism and second language acquisition. Routledge. Simon, J. R. (1969). Reactions toward the source of stimulation. Journal of Experimental Psychology, 81(1), 174–​176. https://​doi.org/​10.1037/​h0027448. Simon, J. R., & Rudell, A. P. (1967). Auditory S-​R compatibility:The effect of an irrelevant cue on information processing. Journal of Applied Psychology, 51(3), 300–​304. https://​doi. org/​10.1037/​h0020586.

516

156 References

Slobin, D. I. (2004).The many ways to search for a frog: Linguistic typology and the expression of motion events. In S. Strömqvist & L.Verhoeven (Eds.), Relating events in narrative, Vol. 2.: Typological and contextual perspectives (pp. 219–​257). Lawrence Erlbaum Associates Publishers. Smith, C. S. (1997). The parameter of aspect. Kluwer. Solovyeva, K., & DeKeyser, R. (2018). Response time variability signatures of novel word learning. Studies in Second Language Acquisition, 40(1), 225–​239. https://​doi.org/​ 10.1017/​S0272263117000043. Spada, N. (1997). Form-​focused instruction and second language acquisition: A review of classroom and laboratory research. Language Teaching, 30(2), 73–​87. https://​doi.org/​ 10.1017/​S0261444800012799. Spada, N., Lightbown, P. M., & White, J. L. (2005). The importance of form/​meaning mappings in explicit form-​focused instruction. In A. Housen & M. Pierrard (Eds.), Investigations in instructed second language acquisition (pp. 199–​234). De Gruyter Mouton. Spada, N., & Tomita, Y. (2010). Interactions between type of instruction and type of language feature: A meta-​ analysis. Language Learning, 60(2), 263–​308. https://​doi.org/​ 10.1111/​j.1467-​9922.2010.00562.x. Spinner, P., & Gass, S. M. (2019). Using judgments in second language acquisition research. Routledge. Stahl, C.,Voss, A., Schmitz, F., Nuszbaum, M., Tüscher, O., Lieb, K., & Klauer, K. C. (2014). Behavioral components of impulsivity. Journal of Experimental Psychology: General, 143(2), 850–​886. 10.1037/​a0033981. Strange, W., & Shafer, V. L. (2008). Speech perception in second language learners: The re-​education of selective perception. In J. G. Hansen Edwards & M. L. Zampini (Eds.), Phonology and second language acquisition (pp. 153–​191). John Benjamins. https://​doi.org/​ 10.1075/​sibil.36.09str. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18(6), 643–​662. https://​doi.org/​10.1037/​h0054651. Sugaya, N., & Shirai, Y. (2007). The acquisition of progressive and resultative meanings of the imperfective aspect marker by L2 learners of Japanese: Transfer, universals, or multiple factors? Studies in Second Language Acquisition, 29(1), 1–​38. https://​doi.org/​ 10.1017/​S0272263107070015. Suzuki, Y. (2018). The role of procedural learning ability in automatization of L2 morphology under different learning schedules:An exploratory study. Studies in Second Language Acquisition, 40(4), 923–​937. https://​doi.org/​10.1017/​S0272263117000249. Suzuki,Y., Nakata, T., & DeKeyser, R. (2019). Optimizing second language practice in the classroom: Perspectives from cognitive psychology. The Modern Language Journal, 103(3), 551–​561. https://​doi.org/​10.1111/​modl.12582. Swain, M. (1988). Manipulating and complementing content teaching to maximize second language learning. TESL Canada Journal, 6(1), 68–​83. https://​doi.org/​10.18806/​tesl. v6i1.542. Tagliamonte, S., & Lawrence, H. (2000). “I used to dance, but I don’t dance now”: The habitual past in English. Journal of English Linguistics, 28(4), 324–​353. https://​doi.org/​ 10.1177/​007542420002800402. Talmy, L. (1985). Lexicalization patterns: Semantic structure in lexical forms. In T. Shopen (Ed.), Language typology and syntactic description (pp. 36–​149). Cambridge University Press. Talmy, L. (2000a). Toward a cognitive semantics (Volume 1, Concept Structuring Systems). MIT Press.

516

517

516

References  157

Talmy, L. (2000b). Toward a cognitive semantics (Volume 2, Typology and Process in Concept Structure). MIT Press. Thierry, G., & Wu, Y. J. (2004). Electrophysiological evidence for language interference in late bilinguals. NeuroReport, 15(10), 1555–​1558. 10.1097/​01.wnr.0000134214.57469.c2. Thierry, G., & Wu, Y. J. (2007). Brain potentials reveal unconscious translation during foreign-​language comprehension. Proceedings of the National Academy of Sciences, 104(30), 12530–​12535. https://​doi.org/​10.1073/​pnas.0609927104. Thomson, R. I. (2017). Measurement of accentedness, intelligibility, and comprehensibility. In O. Kang & A. Ginther (Eds.), Assessment in second language pronunciation (pp. 11–​29). Routledge. Thomson, R. I., & Derwing, T. M. (2015). The effectiveness of L2 pronunciation instruction: A narrative review. Applied Linguistics, 36(3), 326–​344. https://​doi.org/​10.1093/​ applin/​amu076. Thorin, J., Sadakata, M., Desain, P., & McQueen, J. (2018). Perception and production in interaction during non-​native speech category learning. The Journal of the Acoustical Society of America, 144(1), 92–​103. 10.1121/​1.5044415. Timmer, K., Calabria, M., & Costa, A. (2019). Non-​linguistic effects of language switching training. Cognition, 182, 14–​24. https://​doi.org/​10.1016/​j.cognition.2018.09.001. Tokowicz, N. (2014). Lexical processing and second language acquisition. Routledge. Tolentino, L. C., & Tokowicz, N. (2014). Cross-​language similarity modulates effectiveness of second language grammar instruction. Language Learning, 64(2), 279–​309. https://​ doi.org/​10.1111/​lang.12048. Tomasello, M. (2003). Constructing a language. Harvard University Press. Tracy-​Ventura, N., & Huensch, A. (2018). The potential of publicly shared longitudinal learner corpora in SLA research. In A. Gudmestad & A. Edmonds (Eds.), Critical reflections on data in second language acquisition, (pp. 149–​170). John Benjamins. https://​ doi.org/​10.1075/​lllt.51.07tra. Tracy-​Ventura, N., & Paquot, M. (Eds.). (2020). The Routledge handbook of second language acquisition and corpora. Routledge. Traugott, E. C., & Trousdale, G. (2013). Constructionalization and constructional changes. Oxford University Press. Trofimovich, P., & Baker, W. (2006). Learning second language suprasegmentals: Effect of L2 experience on prosody and fluency characteristics of L2 speech. Studies in Second Language Acquisition, 28(1), 1–​30. https://​doi.org/​10.1017/​S0272263106060013. Trofimovich, P., & Isaacs, T. (2012). Disentangling accent from comprehensibility. Bilingualism: Language and Cognition, 15(4), 905–​916. https://​doi.org/​10.1017/​S13667 28912000168. Tyler, A. (2012). Cognitive linguistics and second language learning: Theoretical basics and experimental evidence. Routledge. Tyler, A. E., & Ortega, L. (2018). Usage-​inspired L2 instruction: Some reflections and a heuristic. In A. E. Tyler, L. Ortega, M. Uno, & H. I. Park (Eds.), Usage-​inspired L2 instruction: Researched pedagogy (pp. 315–​ 321). John Benjamins. https://​doi.org/​10.1075/​ lllt.49.14tyl. Tyler, A. E., Ortega, L., Uno, M., & Park, H. I. (Eds.). (2018). Usage-​inspired L2 instruction: Researched pedagogy (Vol. 49). John Benjamins. https://​doi.org/​10.1075/​lllt.49. Tyler, M. D., Best, C. T., Faber, A., & Levitt, A. G. (2014). Perceptual assimilation and discrimination of non-​ native vowel contrasts. Phonetica, 71(1), 4–​ 21. 10.1159/​ 000356237.

518

158 References

Vanek, N. (2020). Changing event categorization in second language users through perceptual learning. Language Learning, 70(2), 309–​348. https://​doi.org/​10.1111/​ lang.12377. VanPatten, B. (2002). Processing instruction: An update. Language Learning, 52(4), 755–​803. https://​doi.org/​10.1111/​1467-​9922.00203. VanPatten, B. (Ed.). (2004). Processing instruction: Theory, research, and commentary. Lawrence Erlbaum Associates Publishers. VanPatten, B. (2014). The limits of instruction: 40 years after “interlanguage”. In Z. Han & E. Tarone (Eds.), Interlanguage: Forty years later, 39, (pp. 105–​126). John Benjamins. https://​doi.org/​10.1075/​lllt.39.07ch5. VanPatten, B. (2015). Foundations of processing instruction. International Review of Applied Linguistics in Language Teaching, 53(2), 91–​109. https://​doi.org/​10.1515/​ iral-​2015-​0005. VanPatten, B. (2017). While we’re on the topic: BVP on language, acquisition, and classroom practice. ACTFL. VanPatten, B., & Borst, S. (2012). The roles of explicit information and grammatical sensitivity in processing instruction: Nominative-​accusative case marking and word order in German L2. Foreign Language Annals, 45(1), 92–​109. https://​doi.org/​10.1111/​ j.1944-​9720.2012.01169.x. VanPatten, B., & Oikkenon, S. (1996). Explanation versus structured input in processing instruction. Studies in Second Language Acquisition, 18(4), 495–​510. https://​doi.org/​ 10.1017/​S0272263100015394. VanPatten, B., Keating, G. D., & Wulff, S. (Eds.). (2020). Theories in second language acquisition: An Introduction. Routledge. Votaw, M. C. (1992). A functional view of bilingual lexicosemantic organization. In R. J. Harris (Ed.), Cognitive processing in bilinguals, Vol. 83 (pp. 299–​321). Elsevier. https://​doi. org/​10.1016/​S0166-​4115(08)61502-​2. Wen, Z., & Li, S. (2019).Working memory in L2 learning and processing. In J.W. Schwieter & A. Benati (Eds.), The Cambridge handbook of language learning (pp. 365–​389). Cambridge University Press. White, L. (2003). Second language acquisition and universal grammar. Cambridge University Press. Wu, J., Kang, C., Ma, F., Gao, X., & Guo, T. (2018). The influence of short-​term language-​ switching training on the plasticity of the cognitive control mechanism in bilingual word production. Quarterly Journal of Experimental Psychology, 71(10), 2115–​2128. https://​doi. org/​10.1177/​1747021817737520. Yang, Y., Chen, X., & Xiao, Q. (2020). Cross-​ linguistic similarity in L2 speech learning: Evidence from the acquisition of Russian stop contrasts by Mandarin speakers. Second Language Research. https://​doi.org/​10.1177/​0267658319900919. Zhao, H., & MacWhinney, B. (2018). The instructed learning of form-​function mappings in the English article system. The Modern Language Journal, 102(1), 99–​119. https://​doi. org/​10.1111/​modl.12449. Zhao, T., & Macaro, E. (2016). What works better for the learning of concrete and abstract words:Teachers’ L1 use of L2-​only explanations? International Journal of Applied Linguistics, 26(1), 75–​98. https://​doi.org/​10.1111/​ijal.12080.

518

519

INDEX

Note: Page numbers in italics indicate figures, bold numbers indicate tables, on the corresponding pages. abstract and concrete nouns 105–​108 accentedness 110, 111, 113–​114 activation 14, 46–​47, 124 additional language learning: construction of new knowledge 4–​7; cross-​language relationships 11–​15; grammatical gender 4–​7; predictions for 26–​28, 33–​35, 38–​39, 42–​44, 43; prior experience brought to 7–​10; risk and support factors 25–​26; transfer 10–​11; use of term 19n1; see also studies of L2 development; theoretical models of L2 learning amelioration hypothesis 36, 38 American Council on the Teaching of Foreign Language’s Guiding Principles for Language Learning 85 aspect morphology 27, 53–​60, 95–​101 assessment of learning 119–​120 Associative-​Cognitive CREED: blocking 32; constructions 29; context, experience of and language use 29; emergent, language as 30; exemplar-​based abstraction 30; learned behavior 31–​32; predictions for L2 learning 33–​35; prior experience 30–​31, 127; rational, use of language as 29; theoretical basics 29–​32; transfer 30 availability, cue 22–​23

backtracking 37–​38 Bates, E. 20 Bell, P.K. 88 Benati, A. 88, 88 Bernolet, S. 48 Best, C.T. 78 blocking 32, 51–​53 Bybee, J. 135 categorization of information 4 cloned language knowledge 12 co-​activation 12–​13, 65–​67 cognitive mechanisms used for learning 4 cognitive processing, future research in 134–​135 collocations 108–​109 competition 14, 124; instruction 108–​109; Unified Competition Model (UCM) 24 comprehensibility 110, 111 constructions 3, 3, 29 context, experience of and language use 29 contrastive analysis and translation 103–​105 cooperation 14 cross-​language relationships 11–​15, 129–​130, 133–​134 crosslinguistic influence: defined 1; different ways of researching 1;

610

160 Index

directions of 15–​17, 16; explicit instruction in 17; future research in 133–​135; key research questions 1–​2; L1 use, changes in due to L2 learning 16; narrowness of research 15; negative effects of 42, 89; spider’s web metaphor 16, 16; theorizing 131–​133 crosslinguistic similiarities 76–​78, 81 Crowther, D. 111–​112, 113 cues: availability, cue 22–​23; crosslinguistic similarity 27; focusing 27–​28; morphological 22, 23; reliability, cue 23; strength, cue 24–​25, 26; syntactic cues 22; validity, cue 23 Darcy, I. 73–​75, 134 decoupling 26, 28 design of instruction 118–​119 domain-​general learning mechanisms 3–​4, 41 durability of learning 120–​121 Dussias, P.E. 6, 7, 60–​62, 72, 81, 108–​109, 126 Ellis, N.C. 30, 51–​53 Ellis, R. 88 emergent, language as 30 entrenchment 25, 26, 27 evidence-​based instruction 89–​90, 130 exemplar-​based abstraction 30 explicit instruction 17, 87–​89; see also instructed L2 learning Flege, J.E. 76, 78 focusing, cue 27–​28 form-​meaning mapping 53–​57, 55 Frauenfelder, U.H. 79–​81 future research in crosslinguistic influence 133–​135 gender, grammatical 4–​7 Girsai, N. 103–​105 grammatical gender 4–​7 Green, D.W. 13, 40–​41, 73 Gries, S.T. 108 Grosjean, F. 127 Guiding Principles for Language Learning (American Council on the Teaching of Foreign Languages) 85 Hartsuiker, R.J. 48 Hopp, H. 5 Huensch, A. 134

immersion 28 inhibitory control 13–​14; and L2 lexical processing 68–​70; phonological development 73–​76; theoretical models of L2 learning 46–​48 Inhibitory Control Model 13–​14; domain-​general learning mechanisms 41; language knowledge and experience 45–​46; language selection 47; language switching 44; predictions for L2 learning 42–​44, 43; selection of appropriate language 40–​41; tag specification 41; theoretical basics 40–​42 instructed L2 learning: accentedness 113–​114; assessment of learning 119–​120; collocations 108–​109; competition-​based instruction 108–​109; contrastive analysis and translation 103–​105; contrastive/​ non-​contrastive 101–​102; crosslinguistic identification and discrimination training 113–​114; design of instruction 118–​119; durability of learning 120–​121; effectiveness for different types of target features 93–​95; effects of 84–​87; evidence-​ based methods 89–​90; explicit instruction 87–​89, 88; feedback to research 131; form-​meaning mapping, sensitivity to 95–​101, 96, 98, 99–​100; as improving L2 learning 117–​121; L1-​L2 differences, explicit instruction in 91–​93; L1 use in L2 learning 89–​90; meaning-​based practice 102; morphosyntax 90–​102, 96, 98, 99, 100; phonology 109–​116; pronunciation instruction 114–​116; research design 117–​121; segmental-​based instruction 113–​114; selective benefits of explicit instruction 130–​131; theory building in research 130–​131; translation and L2 description 105–​108; types of language data collected 120; viewpoint aspect 96, 96–​101; vocabulary 102–​109; word stress 111–​112 integrated system of language knowledge 129–​130 intelligibility 110, 111 Isaacs, T. 111 Jiang, N. 66–​67, 129 Kartushina, N. 79–​81 Kissling, E.M. 114–​116, 120 knowledge and experience of language 45–​46

610

16

610

Index  161

Krashen, S. 110 Kroll, J. 12, 71–​73, 126, 134 language: constructions 3, 3; functionalist understandings of 3; theories of 2–​4 language control 13–​14 Larrañaga, P. 63–​65 Laufer, B. 103–​105 Lawrence, H. 23 learned attention 51–​53 learned behavior 31–​32 learning: cognitive mechanisms used for 4; domain-​general learning mechanisms 3–​4; see also L2 learning; studies of L2 development; theoretical models of L2 learning Leow, R.P. 133 Levis, J.M. 110 Lightbown, P.M. 91–​93 Linck, J.A. 68–​70 Liszka, S.A. 57–​60, 58 L2 learning: construction of new knowledge 4–​7; cross-​language relationships 11–​15; crosslinguistic influence in context of 2; grammatical gender 4–​7; impact on L1 performance 60–​62; predictions for 26–​28, 33–​35, 38–​39, 42–​44, 43; prior knowledge and experience brought to 7–​10; risk and support factors 25–​26; transfer 10–​11; use of term 19n1; see also instructed L2 learning; studies of L2 development; theoretical models of L2 learning longitudinal research 134 L1 performance: changes to L1 perception 81–​82; impact of L2 learning on 60–​62, 71–​73, 124, 126–​127 Macaro, E. 105–​108, 118, 120 MacWhinney, B. 14, 16, 20, 26, 27, 28, 89 Marsden, E. 95–​101, 102, 130 McManus, K. 27, 53–​54, 55–​57, 95–​101, 102, 119, 120, 130, 134 morphological cues 22 morphosyntax: blocking 51–​53; converging representations in L1 and L2 60–​62; form-​ meaning mapping, sensitivity to 95–​101, 96, 98, 99–​100; instructed L2 learning 90–​102, 96, 98, 99, 100; L1-​L2 differences, explicit instruction in 91–​93; online/​ offline performance comparison 58; remapping meaning to form 53–​57, 55 motion, vocabulary and 62–​65

network of knowledge representations, multilingual mind as 60 nouns, abstract and concrete 105–​108 Odlin, T. 10 O’Grady, W. 35 online/​offline performance comparison 57–​60, 58 past habituality 22–​23 path and motion, vocabulary and 62–​65 perception and production, connections between 78–​81 phonology: accentedness 110, 111, 113–​114; changes to L1 perception 81–​82; comprehensibility 110, 111; crosslinguistic identification and discrimination training 113–​114; crosslinguistic similiarities 76–​78; inhibitory control and L2 performance 73–​76; instructed L2 learning 109–​116; intelligibility 110, 111; paucity of research on 110; perception and production, connections between 78–​81; pronunciation instruction 114–​116; segmental-​based instruction 113–​114; word stress 111–​112 prior knowledge and experience: Associative-​Cognitive CREED 30–​31; attending to, importance of 125; blocking 51–​53; as complicating L2 learning 9; difficulties due to 123; future research 133–​134; learned attention 51–​53; positive impact on L2 learning 7–​9; reasons for impact of 9–​10; theory building in L2 research 127–​128; types of experience 122–​123 Processing Determinism: amelioration hypothesis 36, 38; backtracking 37–​38; performance grammar 36; predictions for L2 learning 38–​39; prior experience 127; routines 36–​37, 38; theoretical basics 35–​38; transfer 48–​49, 123; transfer calculus 37–​38, 39 production and perception, connections between 78–​81 pronunciation instruction 114–​116 Pulido, M.F. 108–​109 Purcell, E.T. 110 rational, use of language as 29 reliability, cue 23 remapping meaning to form 53–​57, 55

612

162 Index

research into L2 learning see studies of L2 development resonance 14, 25, 27–​28, 47 rich memory storage 4 risk and support factors, UCM and 25–​26 Roberts, L. 57–​60, 58 routines 36–​37, 38 Sagarra, N. 51–​53, 60–​62, 72, 81, 126 Saito, K. 113–​114, 118, 120 Satellite-​framed languages 63 Schuhmann, K.S. 81–​82 second language learning see additional language learning segmental-​based instruction 113–​114 Sharwood Smith, M. 10 Shintani, N. 121 Skill Acquisition Theory 28, 38, 80 Spada, N. 91–​93, 95, 101, 102, 130 speech, comprehensibility and intelligibility of 111; see also phonology Speech Learning Model 78 spider’s web metaphor 16, 16 strength, cue 24–​25 Stroop task 43, 43–​44 studies of L2 development: changes to L1 perception 81–​82; converging representations in L1 and L2 60–​62; crosslinguistic similiarities 76–​78; future research in crosslinguistic influence 133–​135; inhibitory control and L2 lexical processing 68–​70; inhibitory control and L2 performance 73–​76; learned attention 51–​53; L1 frequency effects in L2 lexical processing 65–​67; L1 performance, impact of L2 learning on 60–​62, 71–​73; morphosyntax, acquisition of 50–​62, 55, 58; online/​ offline performance comparison 57–​60, 58; path and motion 62–​65; perception and production, connections between 78–​81; phonological development 73–​82; remapping meaning in L2 63–​65; remapping meaning to form 53–​57, 55; selection of language 68–​70; vocabulary learning 62–​73 support factors 25–​26, 27 switch costs 68–​70 syntactic cues 22 Tagliamonte, S. 23 tag specification 41 tag suppression 14

theoretical models of L2 learning: activation 46–​47; Associative-​Cognitive CREED 29–​35; inhibition 46–​48; Inhibitory Control Model 40–​44; language knowledge and experience 45–​46; Processing Determinism 35–​39; transfer 48–​49; Unified Competition Model (UCM) 21–​28, 23 theorization about crosslinguistic influence 131–​133 theory building in L2 research: cross-​ language relationships 129–​130; evidence-​based instruction 130; experience 127–​128; instruction 130–​131; integrated system of language knowledge 129–​130; L2 experience 128; organizing and restructuring knowledge 126–​127; prior knowledge and experience 127–​128; selective benefits of explicit instruction 130–​131; usage-​based accounts of learning 128 Thierry, G. 12, 65–​66 Timmer, K. 68 Tokowicz, N. 76, 93–​95, 101–​102, 118, 120 Tolentino, L.C. 76, 93–​95, 101–​102, 118, 120 Tracy-​Ventura, N. 134 transfer 26, 27; alternative account for effects of 11; Associative-​Cognitive CREED 30; as copying and restructuring 10–​11; as core process in L2 learning 123; defined 10; Processing Determinism 48–​49; risk factor of 26; theoretical models of L2 learning 48–​49; theorization about 131–​133; understanding of 123–​124; Unified Competition Model (UCM) 24–​25 transfer calculus 37–​38, 39 Trofimovich, P. 111 Truscott, J. 10 Tyler, M.D. 78 Unified Competition Model (UCM): activation and 14, 47; aims 21; competition 24; crosslinguistic similiarities 76; cues 21–​24, 23; language knowledge and experience 46; origins 21; predictions for L2 learning 26–​28; prior experience 127; resonance 47; risk and support factors 25–​26; theoretical basics 21–​24, 23; transfer 24–​25, 48–​49, 123 usage-​based accounts of learning 20–​21, 128

612

613

612

Index  163

validity, cue 23 verbal morphology 54, 56 Verb-​framed languages 63 viewpoint aspect 54–​55, 55, 96, 96–​101 vocabulary: abstract and concrete nouns 105–​108; changes to L1 performance 71–​73; co-​activation 65–​67; collocations 108–​109; competition-​based instruction 108–​109; contrastive analysis and translation 103–​105; emergence and storage of L2 lexical representations 66–​67; form-​focused instruction 103–​105; inhibitory control and L2 lexical processing 68–​70; instructed L2

learning 102–​109; L1 frequency effects in L2 lexical processing 65–​67; path and motion 62–​65; remapping meaning in L2 63–​65; selection of language 68–​70; translation and L2 description 105–​108 web metaphor 16, 16 word stress 111–​112 Wu,Y.J. 12, 65–​66 Yang,Y. 76–​78 Zhao, H. 89 Zhao, T. 105–​108, 118, 120

164