Cognitive Control and Consequences of Multilingualism [1 ed.] 9789027266729, 9789027243720

Citation preview

Bilingual Processing and Acquisition

Cognitive Control and Consequences of Multilingualism

EDITED BY

2

John W. Schwieter

John Benjamins Publishing Company

Cognitive Control and Consequences of Multilingualism

Bilingual Processing and Acquisition (BPA) issn 2352-0531

Psycholinguistic and neurocognitive approaches to bilingualism/multilingualism and language acquisition continue to gain momentum and uncover valuable findings explaining how multiple languages are represented in and processed by the human mind. With these intensified scholarly efforts come thought-provoking inquiries, pioneering findings, and new research directions. The Bilingual Processing and Acquisition book series seeks to provide a unified home, unlike any other, for this enterprise by providing a single forum and home for the highestquality monographs and collective volumes related to language processing issues among multilinguals and learners of non-native languages. These volumes are authoritative works in their areas and should not only interest researchers and scholars investigating psycholinguistic and neurocognitive approaches to bilingualism/multilingualism and language acquisition but also appeal to professional practitioners and advanced undergraduate and graduate students. For an overview of all books published in this series, please see http://benjamins.com/catalog/bpa

Editor John W. Schwieter

Wilfrid Laurier University and University of Greenwich

Editorial Advisory Board

Jeanette Altarriba, University at Albany, State University of New York Panos Athanasopoulos, University of Reading Laura Bosch, Universitat de Barcelona Kees de Bot, University of Groningen Yanping Dong, Guangdong University of Foreign Studies Paola Dussias, Pennsylvania State University Mira Goral, Lehman College, The City University of New York Jonathan Grainger, Aix-Marseille University Annette M.B. de Groot, University of Amsterdam Marianne Gullberg, Lund University Janet G. van Hell, Pennsylvania State University & Radboud University Nijmegen Roberto R. Heredia, Texas A&M International University Arturo E. Hernandez, University of Houston Walter J.B. van Heuven, University of Nottingham

Ludmila Isurin, Ohio State University Scott Jarvis, Ohio University Iring Koch, RWTH Aachen University Judith F. Kroll, Pennsylvania State University Ping Li, Pennsylvania State University Li Wei, UCL IOE Gary Libben, Brock University Brian MacWhinney, Carnegie Mellon University Jürgen M. Meisel, Universität Hamburg & University of Calgary Silvina A. Montrul, University of Illinois at UrbanaChampaign Loraine K. Obler, The City University of New York Johanne Paradis, University of Alberta Jason Rothman, University of Reading Norman Segalowitz, Concordia University Antonella Sorace, University of Edinburgh Bill VanPatten, Michigan State University Virginia Yip, The Chinese University of Hong Kong

Volume 2 Cognitive Control and Consequences of Multilingualism Edited by John W. Schwieter

Cognitive Control and Consequences of Multilingualism A descriptive and prescriptive analysis

Edited by

John W. Schwieter Wilfrid Laurier University & University of Greenwich

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of the American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

doi 10.1075/bpa.2 Cataloging-in-Publication Data available from Library of Congress: lccn 2016022105 (print) / 2016033971 (e-book) isbn 978 90 272 4372 0 (Hb) isbn 978 90 272 4373 7 (Pb) isbn 978 90 272 6672 9 (e-book)

© 2016 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Company · https://benjamins.com

Table of Contents Acknowledgments

ix

About the editor

xi

About the contributors

xiii

Part I. Introduction Cognitive and neurocognitive implications of language control and multilingualism John W. Schwieter & Andrea Hadland

3

Part II.  Cognitive control and multilingualism chapter 1

Bilingualism, executive control, and eye movement measures of reading: A selective review and re-analysis of bilingual vs. multilingual reading data Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

11

chapter 2

Listening with your cohort: Do bilingual toddlers co-activate cohorts from both languages when hearing words in one language alone? Susan C. Bobb, Laila Y. Drummond Nauck, Nicole Altvater-Mackensen, Katie Von Holzen & Nivedita Mani

47

chapter 3

The role of executive function in the perception of L2 speech sounds in young balanced and unbalanced dual language learners Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

71

chapter 4

Are cognate words “special”? On the role of cognate words in language switching performance Mikel Santesteban & Albert Costa

97

 Cognitive control and consequences of multilingualism

chapter 5

Action speaks louder than words, even in speaking: The influence of (no) overt speech production on language switch costs Andrea M. Philipp & Iring Koch

127

chapter 6

Influence of preparation time on language control: A trilingual digit-naming study Julia Festman & Michela Mosca

145

chapter 7

When L1 suffers: Sustained, global slowing and the reversed language effect in mixed language context Ingrid Christoffels, Lesya Ganushchak & Wido La Heij

171

chapter 8

Effects of cognitive control, lexical robustness, and frequency of codeswitching on language switching John W. Schwieter & Aline Ferreira

193

chapter 9

The locus of cross-language activation: ERP evidence from unbalanced Chinese-English bilinguals Taomei Guo & Chunyan Kang

217

chapter 10

Syntactic interference in bilingual naming during language switching: An electrophysiological study Antoni Rodriguez-Fornells & Thomas F. Münte

239

chapter 11

Multi-component perspective of cognitive control in bilingualism Julia Morales, Carlos J. Gómez-Ariza & M. Teresa Bajo

271

Part III.  Consequences of multilingualism chapter 12

The bilingual advantage in the auditory domain: New directions in methodology and theory Julia Ouzia & Roberto Filippi

299



Table of Contents 

chapter 13

Executive functions in bilingual children: Is there a role for language balance? Anat Prior, Noa Goldwasser, Rotem Ravet-Hirsh & Mila Schwartz

323

chapter 14

Home language usage and executive function in bilingual preschoolers Sibylla Leon Guerrero, Sara Smith & Gigi Luk

351

chapter 15

Cognitive mechanisms underlying performance differences between monolinguals and bilinguals John G. Grundy & Kalinka Timmer

375

chapter 16

Time course differences between bilinguals and monolinguals in the Simon task Manjunath Narra, Andrew Heathcote & Matthew Finkbeiner

397

chapter 17

Top down influence on executive control in bilinguals: Influence of proficiency Ramesh Kumar Mishra & Niharika Singh

427

Index

451

Acknowledgments The hard work of the contributors in this book is to be admired and commended. I am very thankful to them for helping me to build a first book-length volume ­specifically addressing multilingual language control and the consequences of m ­ ultiple ­languages in one mind. I am also grateful to Andrea Hadland, my research assistant and ­co-author of the Introduction, for her excellent work in helping prepare a c­ amara-ready ­manuscript for publication. Finally, I am extremely appreciative of the reviewers who helped to anonymously evaluate and comment on the studies in this volume. Their expert advice has most certainly helped to develop and sharpen the quality of the research featured in this book. These peer reviewers include: Jon Andoni Duñabeitia Pilar Archila-Suerte Cristina Baus Susan Bobb Francesca Branzi Emily L. Coderre Julia Festman Roberto Filippi John G. Grundy Jason W. Gullifer Taomei Guo Roberto R. Heredia Eve Higby Matthew Hilchey Iva Ivanova Iring Koch Ramesh Kumar Mishra

Sibylla Leon Guerrero Jared A. Linck Gigi Luk Rhonda McClain Julia Morales Michela Mosca Thomas F. Münte Manjunath Narra Andrea M. Philipp Anat Prior Antoni Rodriguez-Fornells Mikel Santesteban Ana Schwartz Elizabeth Schotter Niharika Singh Gretchen Sunderman Debra Titone

About the editor John W. Schwieter is an associate professor of Spanish and linguistics and Faculty of Arts teaching scholar at Wilfrid Laurier University in Canada and a visiting professor of applied linguistics in the Centre for Applied Research and Outrearch in Language Education at the University of Greenwich in England. His research interests span cognitive and sociocultural perspectives of bilingualism/multilingualism, language acquisition, language teaching and learning, and translation processes. He is the founding general editor of the Bilingual Processing and Acquisition book series (John Benjamins Publishing) and his recent books include: The Handbook of Translation and Cognition (Wiley-Blackwell, forthcoming); The Cambridge Handbook of Bilingual Processing (Cambridge University Press, 2015); Psycholinguistic and Cognitive Inquiries into Translation and Interpreting (John Benjamins Publishing, 2015); The Development of Translation Competence: Theories and Methodologies from Psycholinguistics and Cognitive Science (Cambridge Scholars Publishing, 2014); and Innovative Research and Practices in Second Language Acquisition and Bilingualism (John ­Benjamins Publishing, 2013).

About the contributors The contributors to this book volume are based at institutions and research centres in 13 countries including Australia, Canada, China, England, France, Germany, India, Israel, Italy, the Netherlands, Poland, Spain, and the United States. Brief bios about the contributors can be found below.

Introduction John W. Schwieter (see “about the editor” above) Andrea Hadland is a research assistant to the editor of this volume. During her undergraduate work in languages, she held the President’s Gold Scholarship at Wilfrid ­Laurier University. She is currently working on her second degree in teacher education at Wilfrid Laurier University in Canada.

Chapter 1 Debra Titone is a professor in the Department of Psychology at McGill University in Canada. Veronica Whitford is a Fonds Québécois de la recherche sur la nature et les technologies (FQRNT) postdoctoral fellow in the Department of Psychology and Brain and Mind Institute at The University of Western Ontario in Canada. Agnieszka Lijewska is an assistant professor of linguistics in the Faculty of English at Adam Mickiewicz University in Poznan in Poland. Inbal Itzhak is a Postdoctoral Research Associate at McGill University in Canada.

Chapter 2 Susan C. Bobb is an assistant professor of psychology at Gordon College in the United States and was a Dorothea Schlözer  post-doctoral  fellow  in the Psychology of Language Research Group at the University of Göttingen in Germany at the time the data from the present study were collected. Laila Y. Drummond Nauck was an undergraduate student in the Psychology of Language Research Group at the University of Göttingen in Germany at the time the data from the present study were collected.

 Cognitive control and consequences of multilingualism

Nicole Altvater-Mackensen is a research fellow in the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig, Germany. Katie Von Holzen is a post-doctoral fellow in the Laboratoire Psychologie de la Perception UMR 8158 at the Université Paris Descartes/CNRS in France and was a doctoral student in the Psychology of Language Research Group, University of Göttingen in Germany at the time the data from the present study were collected. Nivedita Mani is a professor of language acquisition in the Psychology of Language Research Group at the University of Göttingen in Germany.

Chapter 3 Pilar Archila-Suerte is a postdoctoral fellow in the Department of Psychology at the University of Houston in the United States. Her research area is in the development of speech perception in bilingual children and adults. She uses behavioural and neuroimaging methods to investigate perceptual learning and neuroplasticity. Brandin A. Munson is a graduate student in the Laboratory for the Neural Bases of Bilingualism in the Department of Psychology at the University of Houston in the United States. His research focuses on phonologically-based language learning through immersive virtual environments. Arturo E. Hernandez is a professor of psychology at the University of Houston in the United States. His Laboratory for the Neural Bases of Bilingualism investigates perceptual and executive function processes in English monolinguals and Spanish-English bilinguals through tasks of speech perception, non-verbal switching, virtual learning, and creative art.

Chapter 4 Mikel Santesteban is a Ramón y Cajal research fellow in the Bilingual Mind research group at the University of the Basque Country in Spain. His main research focuses on the cognitive mechanisms involved in bilingual language production at both the lexical and the syntactic levels of representation. Albert Costa is an ICREA research professor and the head of the Speech Production and Bilingualism research group in the Center of Brain and Cognition at the Universitat Pompeu Fabra in Spain.

Chapter 5 Andrea M. Philipp is a senior lecturer of psychology at RWTH Aachen University in Germany. Iring Koch is a professor of psychology at RWTH Aachen University in Germany.



About the contributors 

Chapter 6 Julia Festman is head of the diversity and inclusion research group at the University of Potsdam in Germany. Michela Mosca is a PhD student at the Potsdam Research Institute for Multilingualism (PRIM) at the University of Potsdam in Germany.

Chapter 7 Ingrid Christoffels is a researcher in the Centre for Expertise on Vocational Education and Training in The Netherlands, and associated to the Leiden Institute of Psychology at Leiden University in The Netherlands. Lesya Ganushchak is a researcher in the Leiden Institute of Psychology, a researcher in the Leiden University Centre for Linguistics, and a researcher in Education and Child Studies in the Faculty of Social and Behavioural Sciences at Leiden University in The Netherlands. Wido La Heij is an associate professor in the Leiden Institute of Psychology and the Leiden Institute for Brain and Cognition at Leiden University in The Netherlands.

Chapter 8 John W. Schwieter (see “about the editor” above). Aline Ferreira is an assistant professor of Hispanic and Portuguese linguistics at the University of California Santa Barbara in the United States.

Chapter 9 Taomei Guo is a research professor of cognitive neuroscience and a faculty member in the State Key Laboratory of Cognitive Neuroscience and Learning at Beijing Normal University in China. Chunyan Kang is a PhD candidate in the State Key Laboratory of Cognitive Neuroscience and Learning at Beijing Normal University in China.

Chapter 10 Antoni Rodriguez-Fornells is a research professor at the Catalan Institution for Research and Advanced Studies (ICREA) and coordinator of the Cognition and Brain Plasticity Unit (IDIBELL-University of Barcelona) in Spain.

 Cognitive control and consequences of multilingualism

Thomas F. Münte is the chairman of the Department of Neurology and the director of the Institute of Psychology II at the University of Lübeck in Germany. His main research interest concerns the neural implementation of higher cognitive function including language.

Chapter 11 Julia Morales is a post-doc researcher in the Brain, Mind and Behavior Research ­Center at the University of Granada in Spain. Carlos J. Gómez-Ariza is an associate professor in the Department of Psychology at the University of Jaén in Spain. M. Teresa Bajo is a professor of experimental psychology and a head of the Memory and Language Research Group in the Brain, Mind and Behavior Research Center at the University of Granada in Spain.

Chapter 12 Julia Ouzia is a PhD student in the Multilanguage and Cognition Lab at Anglia Ruskin University in the United Kingdom. Roberto Filippi is the director of the Multilanguage and Cognition Lab and a senior lecturer of psychology at Anglia Ruskin University in the United Kingdom.

Chapter 13 Anat Prior is a senior lecturer in the Faculty of Education at the University of Haifa in Israel. Noa Goldwasser received her MA from the Faculty of Education at the University of Haifa in Israel and is currently Director of the Learning Center in Naharia Amal Psagot High School in Israel. Rotem Ravet-Hirsh received her MA from the Faculty of Education at the University of Haifa in Israel and is currently an LD diagnostician and lecturer at the Ort Israel Schools Network in Israel. Mila Schwartz is head of the Language Department and MA Program at Oranim ­Academic College of Education in Israel.

Chapter 14 Sibylla Leon Guerrero is a Ph.D. student at the Harvard Graduate School of E ­ ducation in the United States.



About the contributors 

Sara Smith is an assistant professor of human development and women’s studies at California State University, East Bay in the United States. Gigi Luk is an associate professor of education at the Harvard Graduate School of Education in the United States.

Chapter 15 John G. Grundy is a postdoctoral research fellow of cognitive neuroscience at York University in Canada. Kalinka Timmer is a postdoctoral research fellow with the Speech Production and Bilingualism Research Group at the Universitat de Pompeu Fabra in Spain.

Chapter 16 Manjunath Narra is a doctoral student in the Department of Cognitive Science at Macquarie University in Australia. His research investigates the effect of bilingualism on the time course of stimulus information processing in cognitive control tasks using ‘reach-to-touch’ paradigm and transcranial magnetic stimulation. Andrew Heathcote is a research chair at University of Tasmania (80%) and U ­ niversity of Newcastle (20%) in Australia where he founded the Tasmanian Cognition Laboratory (www.TasCL.org) and the Newcastle Cognition Laboratory (www.NewCL.org), respectively. His current research focuses on human memory and skill acquisition, and on the neural and cognitive processes that enable people to make rapid choices. He is an Associate Editor for Journal of Mathematical Psychology since 2011 and for Psychonomic Bulletin & Review since 2014. Matthew Finkbeiner is an associate professor in the Department of Cognitive Science at Macquarie University in Australia and a co-director of the Perception in Action Research Centre. His research focuses primarily on nonconscious processes. To investigate this, he uses the masked priming paradigm with several different dependent measures, including reaction times, ERPs, TMS, and motion capture.

Chapter 17 Ramesh Kumar Mishra is an associate professor of cognitive science and chair in the Center for Neural and Cognitive Sciences at the University of Hyderabad in India. His recent book Interaction between Attention and Language Systems in Humans: A Cognitive Science Perspective has been published by Springer. Niharika Singh is a PhD student in the Centre of Behavioural and Cognitive Sciences at the Allahabad University in India.

part i

Introduction

Cognitive and neurocognitive implications of language control and multilingualism John W. Schwieter1,2 & Andrea Hadland1 1Wilfrid

Laurier University, Canada / 2University of Greenwich, England

This section presents introductory remarks and important issues at the forefront of studying the cognitive control of multiple languages. These themes are subsequently elaborated on and tested in the following chapters which present original empirical data and/or explore new directions and implications for future research. Together, they shed light on the complicated nature of multilingual language control and demonstrate important implications for a research area which continues to take hold and establish itself as an effective way to study human cognition.

1.  Introduction With the explosion of research on cognitive control processes and the consequences (both advantages and disadvantages) of multilingualism, it is timely and appropriate for a volume to specifically showcase this expansion and spark new debates and research directions. Significant strides forward have been made in the last few decades in and future aspirations for subsequent research promise to continue breaking ground. Such is the main objective of the current book volume: To bring together original studies and commentaries on the state of the field along with well-informed suggestions on future research trajectories. Cognitive control and consequences of multilingualism explores the fascinating human mind by looking at specific domains across which it facilitates cognitive behaviour such as attention, language processing, and memory. Through new perspectives and original studies using innovative methodologies, this book sheds light on the underpinnings of multilingual language control and the notion of a cognitive advantage that may be a consequence of multilingualism. By exploring prominent themes in multilingual language control for both comprehension and production among adults and children alike, this book presents ground-breaking implications for further research. The book is divided into two main parts: the first focuses on the cognitive control of multiple languages in one mind; and the second reviews and explores consequences multilingualism has on cognition.

doi 10.1075/bpa.2.01int © 2016 John Benjamins Publishing Company



John W. Schwieter & Andrea Hadland

2.  Cognitive control and multilingualism The first main section of this book, Part II: Cognitive Control and Multilingualism, explores cognitive control and multilingualism. This section discusses the cognitive demands entailed by the knowledge of more than one language. Through traditional and contemporary methods such as the Simon and flanker tasks and other picture-­ naming tasks, findings are demonstrated related to co-activation of language and language control. Language-switching tasks, where multilinguals are required to switch between their first and second language, are presented in this section to explore cognitive control in multilinguals. Prominent models are included to explain previous findings related to the inhibition of languages. These models and research are then used to enhance and challenge our knowledge about multilingual language control and the implications this entails on human cognition in general. There is also an emphasis on executive function within bilinguals and multilinguals, and the difference between the two. Chapter 1, written by Titone, Whitford, Lijewska, and Itzhak, reviews the cognitive demands entailed to multilinguals. The researchers review the body of knowledge on the distinction between bilinguals and multilinguals and present a re-analysis of previously published data, illustrating the importance of this distinction in future research. Chapter 2 by Bobb, Nauck, Altvater-Mackensen, Von Holzen, and Mani discuss co-activation of languages and language control in bilingual children during word recognition and production. By testing German-English children on a crossmodal priming paradigm, the findings supported the ability for an in-the-moment adaptation and a high level of mental agility, leading to future work considering different age groups. Archila-Suerte, Munson, and Hernandez in Chapter 3 present an fMRI study investigating neural activity induced by L2 speech syllables in areas of the brain associated with executive function typically recruited by bilinguals in cognitive control tasks. The results suggested that children utilized different brain areas to process L2 speech syllables depending on the balance or imbalance of the proficiency of their languages. In Chapter 4, Santesteban and Costa use a language switching task to explore the extent to which different attentional strategies depend on the cognate status of words. Their findings suggested that the cognate status of words does not facilitate language switching nor does it alter the lexical selection mechanisms implicated during production. In Chapter 5, Philipp and Koch turn to an exploration of whether articulationrelated processes are necessary to produce switch costs in language production tasks, using a go/no-go signal delay. The results demonstrated that the overt articulation of a response (for example in go trials) was critical for language switch costs to be observed. Chapter 6 presents an investigation of the influence of preparation time on



Cognitive and neurocognitive implications of language control and multilingualism

trilingual language control by Festman and Mosca. The results indicated that multilinguals’ language control system is a highly adaptive system. In Chapter 7, Christoffels, Ganushchak, and La Heij study the “reverse language effect” in the language switching paradigm by addressing the consequences of mixed language use on later L1 production. Their experiments provide a contribution to the literature of bilingual control, providing evidence for a strong role of sustained control processes in bilingual language production, and calling for future investigation into whether this inhibition presented is at the lexical level or task schema level. Schwieter and Ferreira in Chapter 8 present the effects of individual differences on word production when switching between two languages. The results suggested that cognitive control and L2 lexical robustness had an effect on language switching, but only in limited cases, implying the importance of future studies viewing multilingualism as a dynamic procedure with several dimensions in order to better represent the multilingual experience. Guo and Kang in Chapter 9 investigate the locus of cross-language activation and the lexical selection mechanism in bilingual language production. Using a picture naming paradigm and a go/no-go paradigm, their findings supported a language nonspecific hypothesis of lexical selection in which language similarity aids in determining the locus of lexical selection. In Chapter 10, Rodriguez-Fornells and Münte use ERPs to evaluate the degree of cross-language interference when accessing syntactic information in a go/no-go naming task in a group of fluent Spanish-German bilinguals. The results show that bilinguals could not completely suppress the gender representation of the ­non-target language during covert language production, but the authors suggest that further ERP studies should compare blocked and mixed-language conditions in order to gain a more complex understanding of the cognitive control mechanisms involved. ­Chapter 11 presents experiments that imply that bilingual language control involves coordination between the components of a complex cognitive network. The authors, Morales, Gómez-Ariza, and Bajo, suggest that future research in language control should investigate the role of individual differences in executive functioning, language control, and selection mechanisms.

3.  Consequences of multilingualism Part III: Consequences of multilingualism introduces the ongoing “bilingual advantage debate” which hypothesizes that cognitive benefits accompany multilingualism, specifically regarding executive and cognitive control abilities. The studies in this section investigate how the practice of speaking two languages in everyday life can affect attention, memory, and performance. Some of the studies also compared bilinguals with





John W. Schwieter & Andrea Hadland

other bilingual groups to show important modulating factors involved in the bilingual experience such as age and level of exposure. The studies in this section help to provide an in-depth look at the cognitive and neurocognitive consequences of multilingualism. Chapter 12 by Ouzia and Filippi reviews mixed results from previous research supporting both advantages and disadvantages for bilinguals when it comes to interference control. They argue that any disadvantages that have been previously reported can be linked back to processing differences between native and non-native speakers of a language. The authors emphasize the need for future research that focuses on developmental changes over time, including across several domains. Prior, ­Goldwasser, Ravet-Hirsh, and Schwarz in Chapter 13 compare bilingual children of two age groups with monolingual peers and investigate inhibition and cognitive flexibility. The findings add important implications to the ongoing investigation of the possible relationship between bilingualism and enhanced executive function, suggesting that only the demands posed by balanced bilingualism and strong competition between the two languages might lead to executive function advantages, specifically related to inhibitory control abilities. Chapter 14 by Leon Guerrero, Smith, and Luk examines whether variation in nondominant language use moderates the developmental trajectory of executive function in bilingual preschoolers and seeks to understand how development of executive function changes with age and with different features of bilingual experience. Their results suggested that daily bilingual use moderates preschoolers’ development in executive function. Future work will need to examine how bilingualism might enhance executive functions at different points in time. Grundy and Timmer in Chapter 15 present possibilities for untangling the underlying mechanisms of executive function contributing to performance differences between monolinguals and bilinguals on cognitive control tasks. They suggest that disengagement of attention from previous information is an important mechanism to consider, and that future studies should collect more extensive linguistic background information since many linguistic factors, such as code-switching frequency and different socio-linguistic environments, may explain differential findings. In Chapter 16 by Narra, Heathcote, and Finkbeiner, investigate whether the “bilingual advantage” is due to bilinguals being better at ignoring task-irrelevant information, better at activating task-relevant information, or a combination of the two. The data collected is suggestive of a more efficient and dynamic attentional control system for bilinguals compared to monolinguals. Mishra and Singh conclude the book in Chapter 17 with a study asking whether bilingualism modulates oculomotor behaviour in a Stroop task with a simultaneous working memory load. The results suggested that bilingualism can indeed modulate executive control. This was especially apparent among highly proficient bilinguals who may possess a general executive control advantage that makes them faster on both conflict and non-conflict trials.



Cognitive and neurocognitive implications of language control and multilingualism

4.  Conclusion The studies in this book clearly demonstrate that the future is bright for research on multilingual language control and the cognitive/neurocognitive consequences of multilingualism. These studies and discussions have the potential to significantly move forward the body of knowledge on how languages are controlled and represented in the multilingual mind and the cognitive and neurocognitive consequences this entails. Perhaps just as important, they also clearly demonstrate – and in many cases, make direct claims for – a need for future work in these areas. It is our hope that C ­ ognitive control and consequences of multilingualism will help to clarify previous inconsistent findings and spark the next generation of research on the cognitive control and ­consequences of multilingualism.



part ii

Cognitive control and multilingualism

chapter 1

Bilingualism, executive control, and eye movement measures of reading A selective review and re-analysis of bilingual vs. multilingual reading data* Debra Titone1, Veronica Whitford2, Agnieszka Lijewska3 & Inbal Itzhak1 1McGill 3Adam

University, Canada / 2The University of Western Ontario, Canada / Mickiewicz University, Poland

This chapter selectively reviews the literature on bilingual language processing, with a special focus on the link to executive control, eye movements during reading, and differences between two different groups that are often lumped together: bilinguals (i.e., individuals who know two languages) and multilinguals (i.e., individuals who know more than two languages). To this end, we first discuss ideas about the cognitive demands associated with knowing more than a single language. We then review how eye movement reading research has clarified two important consequences of knowledge and use of more than one language: (1) cross-language activation and its relation to executive function and (2) weakened local (i.e., word-level) and global (i.e., text-level) aspects of reading performance. Finally, we review what is currently known about the bilingual vs. multilingual distinction, and present a re-analysis of previously published data (Whitford & Titone, 2016) exploring the effects of bilingual vs. multilingual status on natural reading in both younger and older adults. Although preliminary, these findings, along with the growing literature reviewed here from other domains,

*  This research was supported by the following grants to Debra Titone: Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Award; Canadian Institutes of Health Research (CIHR) Catalyst Grant in Aging; the Canadian Foundation for Innovation (CFI); and the Centre for Research on Brain, Language and Music (CRBLM). We also gratefully acknowledge additional support from a Fonds Québecois de recherche sur la nature et les technologies (FQRNT) masters scholarship and an NSERC doctoral scholarship awarded to Veronica Whitford.

doi 10.1075/bpa.2.02tit © 2016 John Benjamins Publishing Company

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

illustrate the importance of taking the bilingualism/multilingualism distinction into account when trying to understand the cognitive implications of knowing more than one language.

1.  Introduction A rich history of work has used eye movements to examine language processing in monolingual readers (for reviews, see Rayner, 1998, 2009; Rayner, Pollatsek, Ashby, & Clifton, 2012). Relatively fewer studies have capitalized on eye movements to examine language processing in bilingual readers (i.e., individuals who know exactly two languages), and fewer still have done so specifically for multilingual readers (i.e., ­individuals who know more than two languages). This imbalance is striking given that the number of people who are bilingual or multilingual world-wide are at least as n ­ umerous as the number of people who are monolingual (Grosjean, 2010). Thus, the challenges bilinguals and multilinguals collectively face in terms of language and cognitive processing represent the norm rather than the exception globally. In this chapter, we first discuss ideas about the cognitive demands associated with knowing more than one language. We then selectively review how eye movement reading research has clarified two important consequences of knowledge and use of more than one language: (1) cross-language activation and its relation to executive function and (2) weakened local (i.e., word-level) and global (i.e., text-level) aspects of reading performance. Finally, we review what is currently known about the bilingual vs. multilingual distinction, and present a re-analysis of previously published data from our group (Whitford & Titone, 2016) that examines the effects of bilingual vs. multilingual status on natural reading in both younger and older adults. With respect to this re-analysis, we specifically explore how executive control might modulate differences between bilingual and multilingual readers by comparing younger adults, who presumably operate at peak executive control efficiency, and older adults, who operate with reduced executive control efficiency due to normal age-related changes in cognition (e.g., Burke, 1997; Burke & Shafto, 2004; Campbell, Grady, Ng, & Hasher, 2012; Darowski, Helder, Zacks, Hasher, & Hambrick, 2008; S. Martin, Brouillet, ­Guerdoux, & Tarrago, 2006; Salthouse & Meinz, 1995; Titone, Prentice, & Wingfield, 2000). As will be seen, the results suggest important differences between bilingual and multilingual readers as a function of age, and thus, by proxy, executive control efficiency.

2.  The cognitive demands of knowing more than one language The cognitive demands associated with the bilingual (or multilingual) experience has been much discussed in recent years, due to the now controversial hypothesis that such



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

demands can lead to global neuroplastic changes in cognition more generally among bilinguals (Bialystok, Craik, Klein, & Viswanathan, 2004; Bialystok & Martin, 2004; Costa, Hernandez, Costa-Faidella, & Sebastian-Galles, 2009; de Bruin, ­Treccani,  & Della Sala, 2015; Gathercole et al., 2014; Gold, Kim, Johnson, Kryscio, & Smith, 2013; Ivanova & Costa, 2008; Kousaie & Phillips, 2012; Paap, 2014; Paap & Greenberg, 2013; Prior & MacWhinney, 2010). This chapter does not directly address that larger issue (for greater detail, see Baum & Titone, 2014; Titone & Baum, 2014); however, we focus on the mechanistic underpinnings of this question, that is, the ways in which the bilingual (or multilingual) experience may promote or engender certain kinds of cognitive demands during language processing. There are at least two ways that knowledge and use of multiple languages differ from knowledge and use of only one language (e.g., De Groot, 2011; De Groot & ­Christoffels, 2006; Green, 2011; Green & Abutalebi, 2013; Kroll, Dussias, Bogulski, & Kroff, 2012). First, bilinguals and multilinguals almost always experience some form of local (i.e., word or phrase-specific) linguistic or conceptual ambiguity when processing language (e.g., words or phrases that look or sound the same, translation equivalents that have similar or different meanings across languages). Many studies show evidence of cross-language activation and inhibition during language production and comprehension (Blumenfeld & Marian, 2011; Christoffels, Firk, & Schiller, 2007; ­Dijkstra, 2005; Green, 2011; Guo, Liu, Misra, & Kroll, 2011; Kroll, Bobb, Misra, & Guo, 2008; Macizo, Bajo, & Martin, 2010; M. C. Martin, Macizo, & Bajo, 2010; Misra, Guo, Bobb, & Kroll, 2012; Pivneva, Mercier, & Titone, 2014). Moreover, these demands vary as a function of task- and participant-related factors that include: the particular language task under study; the relative degree of L1 and L2 knowledge or proficiency; cross-language or within-language cues from the language context; and inherent differences between the L1 and L2 in question (Blumenfeld & Marian, 2011; Dijkstra, Miwa, Brummelhuis, Sappelli, & Baayen, 2010; Kroll et al., 2012; Libben & Titone, 2009; M ­ arian & Spivey, 1999; Mercier, Pivneva, & Titone, 2014; Schwartz & Kroll, 2006; Titone, Libben, Mercier, Whitford, & Pivneva, 2011; Van Assche, Duyck, & ­Hartsuiker, 2012; van Hell & De Groot, 2008; van Hell & Tanner, 2012). This is not to say that monolinguals do not face cognitive demands during language processing, indeed, they certainly do at multiple levels of language, for example, lexical ambiguities (e.g., Duffy, Kambe, & Rayner, 2001; Kambe, Rayner, & Duffy, 2001; Meyer & Federmeier, 2008); syntactic ambiguities (e.g., Folk & Morris, 2003; Kjelgaard, Titone, & Wingfield, 1999; Pauker, Itzhak, Baum, & Steinhauer, 2011; Sturt, Scheepers, & Pickering, 2002; Titone et al., 2006); and referential ambiguities (e.g., Arnold, Eisenband, Brown-Schmidt, & Trueswell, 2000; Ito & Speer, 2008; Itzhak & Baum, 2015). However, the crucial point is that bilinguals or multilinguals face exactly what monolinguals face with respect to within-language ambiguity, and also face cross-language ambiguity. The second possible difference between bilinguals or multilinguals compared to monolinguals is the need to regulate global activation of two or more language systems

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

to anticipate upcoming communicative demands (e.g., deliberately suppressing knowledge of English prior to entering the door of a party involving French monolinguals). On this point, many studies show evidence of anticipatory global inhibition, particularly during production (Abutalebi, Tettamanti, & Perani, 2009; Green, 1998; Guo et al., 2011; Kroll et al., 2008; Meuter & Allport, 1999; Misra et al., 2012; ­Pivneva, Palmer, & Titone, 2012; von Studnitz & Green, 2002). Moreover, there may be asymmetries in how easy it is to globally inhibit the L1 compared to the L2, with L1 global inhibition being more difficult than L2 global inhibition (e.g., Meuter & ­Allport, 1999). Again, this is not to say that monolinguals do not face similar challenges, indeed they could, for example, with respect to communicative register (Paradis, 2000), that is, changing the global way one speaks as a function of the interlocutor (e.g., one’s child vs. one’s boss). However, again, the crucial point is that bilinguals do all that monolinguals do in this regard, in addition to the more extreme demand of globally suppressing whole language systems.1 Such distinctions in language control sub-processes cohere with an influential model of bilingual language production (Abutalebi & Green, 2008; Green, 1998), recently applied to bilingual comprehension (Pivneva et al., 2014). This model, called the inhibitory control model, posits that bilingual language control, which is necessary to regulate local and global cross-language activation, is part and parcel of the domaingeneral executive control system (e.g., Abutalebi & Green, 2007). It also posits that executive control demands incurred in any bilingual or multilingual language processing situation will vary as a function of L2 (or L3, L4, etc.) knowledge or proficiency. For example, when L2 proficiency is low, L2 language processing is more controlled and less automatic (see also Favreau & Segalowitz, 1983; Segalowitz, 2010; Segalowitz & Hulstijn, 2005), thus, making the job of L2 processing generally more difficult, and also paving the way for cross-language intrusions of the L1 onto the L2. Both situations would necessitate greater inhibitory control. In contrast, when L2 proficiency is high, L2 language processing is more automatic, cross-language intrusions would be less likely, and thus, there would be less of a demand for inhibitory control. Moreover, high L2 proficiency might lead to increased L1 processing effort because of weakened links

.  Interestingly, local and global language control may have domain-general analogues in the executive control literature, namely, proactive and reactive control (Braver, 2012; Colzato et al., 2008; Morales, Gomez-Ariza, & Bajo, 2013; Morales, Yudes, Gomez-Ariza, & Bajo, 2015; Zhang, Kang, Wu, Ma, & Guo, 2015). Accordingly, proactive control enables us to maintain goal-relevant information in anticipation of future demands, for instance, when bilinguals globally inhibit knowledge of one language in one communicative setting, and shift attention to another language that is relevant to that setting. In contrast, reactive control acts as a late correction mechanism for high competition stimuli encountered in real time (e.g., when bilinguals inhibit semantically incompatible meanings that arise in the moment when ­comprehending or speaking).



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

between word forms and concepts in the L1 (Bialystok, 2001; Bialystok, Luk, Peets, & Yang, 2010; Gollan, Montoya, Cera, & Sandoval, 2008; Gollan et al., 2011; Ivanova & Costa, 2008; Michael & Gollan, 2005; Whitford & Titone, 2012) or because of an increased likelihood of intrusions of L2 knowledge onto L1 processing. In sum, bilingual or multilingual language processing likely leads to qualitatively and quantitatively different patterns of executive control demands than monolingual language processing. Moreover, individual differences in language knowledge and ability among bilinguals and multilinguals may also have systematic effects on executive control demands. Thus, it stands to reason that if one were to consider possible ways in which the bilingual and multilingual experience differ quantitatively or qualitatively from each other, as we do later in this chapter, one might also expect executive control demands to vary during bilingual vs. multilingual language processing. In the next section, we set aside this issue temporarily, and first selectively review the eye movement approach to studying reading, and what is currently known about the processes involved in bilingual (or multilingual) reading as revealed by eye movement measures.

3.  Eye movements & bilingual reading Reading, a crucial life skill developed over many years of formal instruction and practice, is vitally important to virtually all domains of modern life. Unlike the human capacity for spoken language which is evolutionarily old, reading is a more recent human invention that has only been in existence for a few thousand years (discussed in ImmordinoYang & Deacon, 2007). Reading studies using eye movement recordings first emerged in the late 1800s (e.g., Huey, 1908), and have led to a substantial and important body of research (reviewed in Rayner, 1998, 2009; Rayner et al., 2012). This literature is one of the most mature and highly developed sub-disciplines of psychology, in no small part due to the pioneering work of Keith Rayner, who founded and nurtured this field (Clifton et al., 2016). Before turning to what we know about bilingual reading from eye movement studies, we first provide a short primer on why eye movement studies are particularly informative for uncovering the processes involved in reading.

3.1  A primer on the use of eye movements to study reading At its most basic level, reading involves a series of eye movements called saccades, which bring new printed information onto the fovea (i.e., the area of the visual field where vision is sharpest). Saccades are separated by brief pauses called fixations, during which visual information is obtained for detailed linguistic processing (see for example, Liversedge, Gilchrist, & Everling, 2011; Radach & Kennedy, 2013; Rayner, 1998, 2009; Rayner et al., 2012). Fixations and saccades can be progressive, that is, forward moving in the normal direction of reading or regressive, that is, backward

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

moving against the normal direction of reading, which reflect re-reading earlier portions of a text. The systematic analysis of fixation durations (among other measures), aggregated in various ways detailed below, allows us to make inferences about the cognitive processes that occur when people read. These cognitive processes could reflect many aspects of linguistic processing simultaneously: lower-level analysis of a word’s orthographic, phonological, and morphological properties; the comprehension of multiword sequences; syntactic structure; and discourse-level semantic integration. Although the eye movement approach largely assumes that what happens at a fixation reflects attention directed at that fixation (i.e., the Eye-Mind Hypothesis; Just & Carpenter, 1980), attention may not necessarily be directly linked to where the eyes are fixated. Thus, fixation durations could also reflect low-level processing (e.g., word length, word shape, letter features) that occurs beyond the fovea, that is, in the parafoveal region (i.e., 1–5 degrees of visual angle from fixation). A long history of eye movement research has identified a perceptual or attentional span during reading, which extends 3–4 characters to the left and 14–15 characters to the right of fixation in skilled readers of left-to-right orthographies, such as English (e.g., McConkie & Rayner, 1975; Rayner & Bertera, 1979; Rayner & McConkie, 1976; Rayner, Well, & Pollatsek, 1980). The perceptual span is functionally reversed in skilled readers of right-to-left orthographies, such as Arabic, Hebrew, and Urdu (Jordan et al., 2014; Paterson et al., 2014; Pollatsek, Bolozky, Well, & Rayner, 1981). The perceptual span mediates eye movement control during reading, especially in the selection of upcoming saccadic targets, and trades-off with the ease of textual processing: more parafoveal information is extracted from the right of fixation when the text is easier to process (see Henderson & Ferreira, 1990). Eye movements during reading are naturalistic behaviors that simultaneously reflect many different cognitive and linguistic factors. At the highest level, eye movement reading measures can be local or global, that is, based on fixations of target-words or aggregated over all the words within an entire sentence or text, respectively. Eye movement measures can also be early or late, that is, whether they reflect the initial stages of reading (e.g., low-level orthographic and lexical processing) or later stages of reading (e.g., semantic integration and meaning revision), respectively. With respect to local reading measures that isolate the earliest stages of processing, researchers generally examine what people’s eyes do when they encounter words for the first time, that is, on the first pass. These measures include first fixation duration (the duration of time the eye first lands on a word irrespective of how many fixations are ultimately made on that word); single fixation duration (the duration of a fixation on a word when it was fixated exactly once during first pass); gaze duration (the summed duration of all first-pass fixations on a word); word skipping probability (the likelihood of not fixating a word, which could indicate that the word was



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

sufficiently processed during a prior fixation); and forward saccade length (how far the eyes move forward from the current fixation on the very next saccade, which could indicate whether there was enough attentional resources left over at a given fixation to parafoveally process upcoming text). First-pass fixation durations tend to be very short (e.g., 250 ms), and they are known to vary systematically as a function of many linguistic variables that include word length, frequency, age of acquisition (AoA), and contextual predictability (reviewed in Abbott, Angele, Ahn, & Rayner, 2015). With respect to local reading measures that isolate later stages of processing, eye movement researchers generally examine what readers’ eyes do beyond the first pass. These measures include go-past-time (the duration of time from when the reader first looks at a word until moving past it, which would include whether they had to look back to earlier portions of a sentence to help make sense of that word); total reading time (the total duration of time people spend looking at a word, which includes both first-pass and second-pass fixations); second-pass time (the same as total reading time but excluding all first pass fixations); regression out probability (the proportion of times people have to look back earlier in a text to make sense of a given word); and r­ egression in probability (the proportion of times people return to this target word from later regions of a sentence). Indeed, many of the same linguistic factors that affect the earliest stages of processing are also studied using later measures, particularly to the extent that such low-level effects trickle up to impact higher-level comprehension processes, such as contextual integration and meaning revision. Finally, in addition to the array of early and late local (i.e., target-word-based) measures described above, one can also globally aggregate eye movement measures across entire sentences or passages to assess processing ease or difficulty at this macrolevel. Global reading measures include average reading rate (normally 200–400 words per minute for skilled readers); average saccade length (normally 7–9 characters); average fixation duration (normally 200–250 ms); average percent regressive saccades (normally 10–15% of all saccades); and total reading time (which varies greatly by the length of the text over which this is measured), again, across an entire sentence/passage of text (Rayner, 1998, 2009). Such measures, which can also isolate early or late processing depending on what is being aggregated, are modulated by text difficulty, participants’ reading ability, and participants’ reading goals (e.g., thorough reading for comprehension vs. scanning for key words). Specifically, increased text difficulty, reduced reading ability, and more thorough reading generally result in: slower reading rates, shorter saccade lengths, longer fixation durations, more regressions, and longer total reading times (Rayner, 1998, 2009). The potential for many highly informative measures with which to constrain theories of language processing constitutes a major advantage of this method, as one can

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

build a very rich story about language as it unfolds over time.2 Other advantages of the eye movement approach include higher ecological validity compared to other methods often used within psycholinguistics, such as lexical decision tasks, which require overt, artificial decisions (Clifton et al., 2016; Rayner, 1998, 2009, 2014).3 Indeed, participants in eye movement studies complete tasks that mirror everyday language use (e.g., sentence or passage reading, spoken language comprehension) while their eye movements are recorded with a camera. Moreover, because of the high temporal precision associated with eye tracking (e.g., state of the art models sample the location of gaze 1000 times per second), one can implement sophisticated manipulations, such as gaze contingent methods, which enable experimental display changes based on where a reader is currently fixating (reviewed in Schotter, Angele, & Rayner, 2012). Such gaze-contingent techniques, which include the moving window (McConkie & Rayner, 1975), boundary (Rayner, 1975), and fast priming paradigms (Lee, Rayner, & Pollatsek, 1999; Sereno & Rayner, 1992), have been essential for characterizing the nature of parafoveal processing across different linguistic levels (i.e., phonological, semantic, syntactic). The vast majority of eye movement studies have investigated the cognitive processes underlying monolingual (L1) reading; however, a growing number of studies have also investigated the cognitive processes involved in bilingual L2 reading (­Altarriba, Kambe, Pollatsek, & Rayner, 2001; Altarriba, Kroll, Sholl, & Rayner, 1996; Balling, 2013; Bultena, Dijkstra, & van Hell, 2014; Friesen & Jared, 2007; Hoverstern & Traxler, 2016; Titone, Libben, Mercier, Whitford, & Pivneva, 2011). This work pursues core questions that have long been asked using traditional cognitive and psycholinguistic tasks. Of relevance here, very little work in this area has distinguished between reading performance for bilinguals (who know exactly two languages) vs. multilinguals (who know more than two languages) –­a point we will focus on in later sections. However, setting this issue aside for now, we first focus on three specific issues within .  However, it is rare for any given eye movement study to report all possible eye movement measures at once in the interests of parsimony, limiting oneself to the linguistic level at which a particular study is aimed, and the fact that many of these measures co-vary by definition, as they reflect the same underlying eye movement behavior (e.g., more regressions into a region should, by definition, inflate total reading time). .  For example, many other approaches present isolated words as stimuli, despite the fact that people normally process words embedded in sentences or more extended texts. Moreover, many studies use artificial, response-based tasks that may be removed from real-life linguistic experience. These include one of the most popular paradigms used in visual word recognition, the lexical decision task, wherein participants decide whether a string of letters represents a real word or not in a given language. Other examples are semantic decision tasks (e.g., is the word presented animate or not?) and grammaticality decision tasks for sentence processing (e.g., is this a grammatically correct sentence or not?). Language users rarely, if ever, engage in this sort of language processing outside the laboratory.



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

the bilingual eye movement literature that link to our prior section on the factors influencing bilingual language control: Cross-language activation during bilingual reading; the comprehension of language-unique words during bilingual reading; and global aspects of bilingual reading performance.

3.2  E  ye movement studies of cross-language activation during bilingual reading The earliest eye movement studies of bilingual reading investigated code-switching and syntactic processing (e.g., Altarriba et al., 1996; Frenck-Mestre & Pynte, 1997). However, a core question in the most recent eye movement literature on bilingual reading is the degree to which people access linguistic knowledge in a language-selective manner, that is, only the target language is activated or a language non-selective manner, that is, all languages are initially co-activated (De Groot, Delmaar, & Lupker, 2000; Duyck, 2005; Nievas & Mari-Beffa, 2002; Simpson & Krueger, 1991). As previously discussed, the potential for cross-language activation presumably impinges on the degree to which executive control may be recruited during bilingual language processing. Cross-language activation has often been examined by comparing how people process words that overlap across their known languages vs. matched control words that exist in one language only. One class of cross-language words are interlingual homographs, which share orthography but not semantics across languages (whether these words share phonology is variable). For example, CHAT is a casual talk in English vs. a cat in French. Although interlingual homographs share orthography across languages, they are generally recognized more slowly than language-unique matched control words, which share neither form nor meaning across languages – a finding termed interlingual homograph interference. However, interlingual homograph effects can vary as a function of experimental stimuli and task demands (Dijkstra, Grainger, & van Heuven, 1999; Dijkstra, Timmermans, & Schriefers, 2000; Lemhofer & Dijkstra, 2004). For example, some studies have found facilitatory interlingual homograph effects (De Groot et al., 2000; Dijkstra, Van Jaarsveld, & Brinke, 1998), which may be attributable to the fact that these words have a higher form-frequency, and thus, may be relatively easy to orthographically decode. Another class of cross-language words are cognates, which share both orthography and semantics across languages (whether these words share phonology is, again, variable). For example, ORANGE refers to a fruit and/or colour in both English and French. Because of this dual overlap, cognates are recognized more rapidly than language-unique matched control words – a finding termed cognate facilitation. Cognate effects may also arise in what is referred to as non-identical cognates, where a reduced degree of orthographic overlap leads to a relatively smaller cognate facilitation effect (Balling, 2013; Comesaña et al., 2015; Dijkstra et al., 2010; Lemhofer & Dijkstra, 2004).

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

Duyck, Van Assche, Drieghe and Hartsuiker (2007) were among the first to use eye tracking to examine cognate processing during L2 reading. Specifically, Dutch-English bilinguals read semantically neutral L2 (English) sentences containing form-identical (e.g., RING) and non-identical cognates (e.g., SCHIP-SHIP). The authors observed cognate facilitation (shorter reading times for RING vs. COAT) for both early and late eye movement measures, which increased with greater cross-language orthographic overlap. Thus, consistent with work using other methods, bilinguals activate both L1 and L2 representations when reading exclusively L2 sentences. In a later study on L1 reading by the same group, Van Assche, Duyck, Hartsuiker, and Diependaele (2009) had Dutch-English bilinguals read cognates and matched control words in L1 (Dutch) sentences, again when there was no semantic bias towards a word. For example, Ben heeft een oude OVEN/LADE gevonden tussen de rommel op zolder [Ben found an old OVEN/DRAWER among the rubbish in the attic]. Like the L2 reading study, cognate facilitation occurred for early eye movement measures (the authors did not report late eye movement reading data). These results suggest that bilinguals co-activate both the L1 and L2, even when reading the more dominant, and presumably fluent L1 (see also Cop, Keuleers, Drieghe, & Duyck, 2015, for similar conclusions from an eye movement study of whole story reading). Our group has also used eye movements to investigate bilingual reading for both cognates and interlingual homographs embedded in semantically biased sentence contexts (Libben & Titone, 2009). This work was inspired by Schwartz and Kroll (2006), who investigated a similar issue with standard cognitive tasks. In Libben and Titone (2009), French-English bilinguals read cognates, interlingual homographs, or language-unique matched control words embedded in L2 (English) sentences with high vs. low semantic constraint (for cognates, such as JUNGLE, the high/low constraint sentences were as follows: “When they were on a safari, they saw an enormous JUNGLE that was dark and scary” vs. “When they were on their trip, they saw an enormous JUNGLE that was dark and scary”; for homographs, such as COIN (i.e., money in ­English vs. corner in French), the high/low constraint sentences were as follows: “Because she knew the change was counterfeit, the brown coloured COIN was thrown out” vs. “Because she knew it was worthless, the brown coloured COIN was thrown out”).4 The results showed cross-language activation for both cognates and homographs; however, this interacted with the semantic bias of the sentence for some eye movement measures. For neutral sentences, cognate facilitation and interlingual

.  Of note, we selected homographs that were more frequent in French than in English (e.g.,  manger: food in French vs. a farm trough in English) to increase the likelihood of a cross-language effect in the neutral context, such that we could assess whether a semantically biased context would successfully attenuate that effect.



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

homograph interference occurred for both early and late eye movement measures (first fixation or gaze duration vs. total reading time, respectively). For semantically biased sentences, facilitation and interference effects were only observed for early, but not late eye movement measures. Thus, bilinguals initially activated both languages, but rapidly capitalized on the semantic bias toward the target language in a way that reduced cross-language activation. Also of note, post-hoc analyses of these data revealed that cognate facilitation but not homograph interference varied as function of L2 proficiency: greater L2 proficiency related to less cognate facilitation (a point to which we return later). Extending this work to L1 reading, Titone et al. (2011) used the same materials, but also examined the impact of task demands and L2 age of acquisition. Here, English-French bilinguals read the same cognates, interlingual homographs, and matched language-unique control words, embedded in the same sentences as those used in Libben and Titone (2009). However, this time, bilinguals read sentences in their L1 (English). To examine task demands, Titone et al. (2011) manipulated the language of filler sentences across two otherwise similar experiments. In Experiment 1, filler sentences were presented exclusively in the task language (English). In Experiment 2, a subset of the filler sentences was presented in the other language (French). We hypothesized that this manipulation would promote different executive control demands across the two experiments (i.e., by making it difficult to inhibit the non-target language in the French filler experiment), which in turn, could impact the degree of cognate facilitation and homograph interference observed in the experimental sentences. The results generally supported this hypothesis. In Experiment 1 (i.e., English sentences only), L1 readers showed greater cognate facilitation for early eye movement measures, but only when L2 was acquired at a young age. In Experiment 2 (English sentences mixed with French filler sentences), by contrast, early stage cognate facilitation was found for all readers (i.e., it was not limited to the participants whose L2 was acquired at a young age). With respect to interlingual homographs, there was no reliable interference for early eye movement measures in either Experiment 1 or 2. However, interlingual homograph interference occurred for late eye movement measures in both Experiments 1 and 2, though the effect was statistically larger in Experiment 2 when French filler sentences were present.5 Taken together, these results, along with those of Libben and Titone (2009), suggest that cognates and interlingual homographs are not mirror-image reflections of cross-language activation, as is often assumed. For example, cognate facilitation varies as a function of individual differences in L2

.  Interestingly, the effect of semantic context was reduced compared to Libben and Titone (2009), though this may have occurred because cross-language activation during L1 reading was more fragile.

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

ability and history (see in both Libben & Titone, 2009; and Titone et al., 2011), whereas interlingual homograph interference does not. Other potential differences between cognates and homographs also emerge when one investigates how individual differences in executive control modulate the degree of cross-language activation observed. For example, Pivneva et al. (2014) tested whether individual differences in executive control modulate cross-­language activation during L2 reading using the same materials as Libben and Titone (2009) in a different group of French-English bilinguals. In addition, participants in the ­Pivneva et al. (2014) study also completed a battery of language proficiency and executive control tasks. Consistent with the idea that homograph interference and cognate facilitation may be driven by different underlying cognitive processes, homograph interference but not cognate facilitation was modulated by individual differences in executive control among bilinguals. Specifically, bilinguals with greater executive control capacity (measured using a composite score derived from the executive control battery) showed less homograph interference than bilinguals with lesser executive control capacity, an effect that was observed while statistically accounting for individual differences in L2 ability. In contrast, cognate facilitation was only affected by individual differences in L2 proficiency (similar to our past two studies), while statistically controlling for executive control capacity. Thus, across the studies reviewed from our group, cognate facilitation relates to how well bilinguals know and fluently use their L2 but not to executive control capacity. In contrast, interlingual homograph interference is exactly the reverse – it is sensitive to executive control capacity but not L2 proficiency. Accordingly, we would argue, as have others (e.g., Davis et al., 2010; Midgley, Holcomb, & Grainger, 2011), that this difference is due to the fact that cognate facilitation reflects a kind of word frequency effect (i.e., cognates are functionally more frequent than matched language-unique words), particularly for cognates that have identical forms across two languages (functional frequency may be higher for partial cognates in terms of their sub-word units). In contrast, interlingual homographs, which also share form and should hypothetically enjoy a processing advantage on that basis alone (and in some studies they do), are nevertheless usually more difficult to process because of cross-language semantic conflict, which draws upon executive control to regulate. In sum, eye movement studies of cross language activation among bilinguals suggest that many factors drive the extent to which cross-language effects are observed, and for what reasons (e.g., reactive control, proactive control). One important conclusion from this work is that not all cross-language effects are the same. That is, semantic conflict is likely the driving force behind interlingual homograph interference, whereas differences in frequency-driven, bottom-up lexical activation may be the driving force behind cognate facilitation. This leads to more basic questions about how bilingual reading, irrespective of cross-language activation, is affected by trade-offs in language



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

experience, and ensuing differences in relative L1 or L2 language entrenchment, a topic to which we now turn.

3.3  The comprehension of language-unique words during bilingual reading As detailed above, an important consequence of bilinguals’ knowledge and use of two languages is that they regularly experience cross-language activation (across their L1 and L2). However, another important consequence that has been relatively underinvestigated is how their divided knowledge and language use impact the processing of language-unique words (across their L1 and L2). Bilinguals, as a group, vary continuously in terms of their relative amount of L1 and L2 experience. For example, some bilinguals, especially those who acquired both languages from birth (i.e., simultaneous bilinguals) may have roughly the same amount of experience in their L1 and L2. However, other bilinguals, especially those who acquired their L2 at a later point in time (i.e., sequential bilinguals) may have considerably less L2 than L1 experience. While it stands to reason that more L2 experience should relate to greater ease of L2 word processing (and L2 reading fluency more generally), it is less clear whether it should also affect ease of L1 word processing (and L1 reading fluency more generally). This is because the L1 is generally the more dominant language, and thus, may be relatively less sensitive to changes in ongoing L2 experience. However, assuming that the representation of language in the brain, like many other neurocognitive processes, is modulated by local fluctuations in input and exposure, L1 and L2 word processing (and reading processes more generally) should trade-off to some degree as a function of increased L2 experience. Over time, this trade-off could potentially lead to L1 attrition effects, which reflect a weakening or reduction in L1 abilities (see review by Schmid, 2013). These predictions are consistent with theories of bilingual language processing, such as the weaker links (Gollan et al., 2008) or frequency-lag hypothesis (Gollan et al., 2011), which build on lexical quality models reported in the monolingual comprehension literature (e.g., Andrews & Hersch, 2010; Perfetti, 2007; Perfetti & Hart, 2002). In particular, these theories posit that less overall L2 than L1 experience among bilinguals weakens the integration of different kinds of word-related knowledge (e.g., orthography, phonology, semantics) in memory, leading to reduced ease of L2 word processing (see also the Bilingual Interactive Activation Plus Model - BIA+, Dijkstra & Van Heuven, 2002 for a similar account). In other words, because bilinguals have less L2 than L1 experience, the quality of their L2 lexical representations is weakened, ultimately leading to reduced L2 lexical accessibility. Of note, although these theories do not explicitly predict how individual differences in L2 experience among bilinguals should impact L1 and L2 word processing, greater current L2 experience should, in

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

principle, lead to increased ease of L2 word processing, but also lead to decreased ease of L1 word processing.6 Consistent with these theories, a series of eye movement studies from our group have found that L1 and L2 word-level processing are indeed sensitive to individual differences in current L2 experience, although in young adulthood only. Of note, across all studies from our group, bilinguals with relatively high levels of current L2 exposure vs. those with relatively low levels of current L2 exposure did not significantly differ in terms of historical L2 usage patterns (e.g., L2 AoA, age of L2 reading fluency, etc.), thus, the results from these studies always reflect the impact of current L2 experience on L1 and L2 reading behavior. For example, Whitford and Titone (2012) found that greater current L2 exposure facilitated L2 word processing (indexed by smaller L2 word frequency effects for both early and late eye movement measures), but impeded L1 word processing (indexed by larger L1 word frequency effects for early eye movement measures only) in a large sample of younger adults (n = 117) during naturalistic L1 and L2 paragraph reading (see also Whitford, Pivneva, & Titone, 2016 for a ­re-analysis of Whitford & Titone, 2012 examining the non-linearity of bilingual word frequency effects). More recently, Whitford and Titone (accepted) examined whether these findings extended to bilinguals across the adult lifespan, that is, in younger (i.e., 18 to 30 years) and older adults (i.e., 60+ years). Corroborating their earlier work, they observed a trade-off in L1/L2 word processing among younger adults (n = 62), using a larger number of paragraphs (see also Cop et al., 2015 for larger L2 vs. L1 word frequency effects during bilingual whole story reading). Interestingly, however, the tradeoff in L1/L2 word processing was not found in older adults; individual differences in current L2 exposure had no significant impact on their L1 and L2 word processing during paragraph reading. This finding suggests that older adults’ lexical representations, which have benefitted from more life-long experience, may be relatively more entrenched, and consequently, relatively insensitive to local fluctuations in L2 input and exposure. Thus, eye movement measures of reading also enable us to track how bilinguals process language-unique words during naturalistic L1 and L2 reading, as a function of graded differences in L2 experience. In the next section, we turn to studies of global

.  Although not of direct relevance to the present work, theories of bilingual language ­processing also posit reduced ease of word processing in bilinguals (especially during L2 ­processing) compared to monolinguals, given their reduced overall language exposure. In other words, because bilinguals know two languages, they necessarily have less experience in each of their languages than monolinguals, who, by definition, have experience with one language only. This issue has been examined in recent eye movement work by Gollan et al. (2011) and Cop et al. (2015).



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

aspects of bilingual reading performance, with a particular focus on how graded differences in L2 experience modulate measures of L1 and L2 reading at this scaled-up, global level, which may arise from cross-language activation, or difficulty generating low-level activation for low frequency language-unique words.

3.4  Global aspects of bilingual reading performance Similar to what was described in the prior section, our group has also found that the trade-off in L1/L2 word processing as a function of increased current L2 experience scales up to impact global aspects of L1 and L2 sentence reading in a large sample of younger adults (n = 95). In particular, Whitford and Titone (2015) found that greater current L2 exposure facilitated L2 sentence reading (e.g., faster reading rates, shorter fixations), but impeded L1 sentence reading (e.g., slower reading rates, longer fixations) using gaze-contingent, moving window tasks (McConkie & Rayner, 1975; Rayner, 2014; Rayner & McConkie, 1976; Schotter et al., 2012; Whitford et al., 2013). These tasks also allowed them to assess the impact of current L2 exposure on the L1 and L2 perceptual span, that is, the amount of visual attention allocated into the parafovea. Although the breadth of the L1 and L2 perceptual span (approximately 10 characters to the right of fixation) was insensitive to graded differences in current L2 exposure, they found that L1 and L2 parafoveal processing for the smaller window conditions was. Specifically, greater current L2 exposure lead to greater L2 parafoveal processing, but reduced L1 parafoveal processing. Extending this approach to older adults, Whitford and Titone (2016) found that global aspects of L1 and L2 sentence reading (e.g., forward saccade lengths, number of regressions) also trade-off with greater current L2 experience in a manner similar to that found for matched younger adults. This trade-off, however, was not found for the L1 and L2 perceptual span nor for parafoveal processing of the smaller window conditions. These findings suggest that current L2 experience has a more limited influence on L1 and L2 reading in older adults, impacting global reading measures only. Although Whitford and Titone (accepted) found no significant effect of current L2 experience on L1 and L2 word-level reading in their prior study involving paragraph reading in older adults, the findings from their gaze-contingent, moving ­window study suggest that global aspects of L1 and L2 reading, which tap into linguistic processes beyond the lexical level (e.g., syntactic processing), may indeed be sensitive to local fluctuations in L2 input and experience. Of note, Whitford and Titone (2016) also found that older adult bilinguals experience age-related reductions in global aspects of reading performance (across the L1 and L2), indexed by slower reading rates, longer fixations, and more regressions. These findings are likely attributable to age-related decrements in cognitive

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

processing, including reduced working memory capacity and reduced inhibitory control (Burke, 1997; Burke & Shafto, 2004; Darowski et al., 2008; S. Martin et al., 2006; Salthouse  &  Meinz, 1995). Of relevance to the next section, and to the re-analysis of Whitford and Titone (2016) presented later, is that in all of the studies reviewed above from our group (and likely others), bilingual readers were not differentiated into whether they knew exactly two or more than two languages. As we will see in the next section, this failure to differentiate our participants may not be entirely justifiable given what we know about other language processing differences between bilingual and multilingual adults.

4.  A  re L1/L2 trade-offs in reading performance modulated by L3+ knowledge? Eye movement measures have deepened our understanding of language processing in bilingual readers. However, this work has largely followed the widely accepted, but potentially flawed tradition of defining bilinguals as users of two or more languages, where the bilingual/multilingual distinction is rarely acknowledged (see Green & Wei, 2014; but see De Angelis, 2007). In contrast to this tradition, some studies, particularly those from the L3 acquisition literature, have shown that multilinguals exhibit patterns of language processing that differ from bilinguals (De Bot, 2012; Falk & Bardel, 2011, 2012; Flynn, Foley, & Vinnitskaya, 2010; Garcia-Mayo & ­Rothman, 2012; Rothman, 2013; Rothman & Cabrelli Amaro, 2010; Wrembel, 2012). For example, when trilinguals speak or write in their least dominant language (i.e., L3), they tend to be more affected by their L2 than by their L1, even when L1 is their strongest language (Dewaele, 2001; Ecke, 2001; Gibson & Hufeisen, 2003; Williams & Hammarberg, 1998; for a review see De Angelis, 2007; Otwinowska-Kasztelanic, 2015). This is especially the case during the initial stages of L3 acquisition as shown by Wrembel (2010). When she investigated the influence of L1 (Polish) and L2 (German) on the acquisition of L3 (English) phonology, she found that L2-accentedness in participants’ L3 speech decreased with increasing proficiency in L3 (for similar evidence see Hammarberg & Hammarberg, 1993; Llama, Cardoso, & Collins, 2010; Wrembel, 2012). Past research on L3 acquisition has also revealed that the interaction between the languages known to multilinguals is additionally modulated by the degree to which the languages are typologically related (i.e., structurally similar in terms of lexical items, syntactic features, phonological make-up, etc.) or are perceived as similar by the speakers (Cenoz & Gorter, 2011; Dewaele, 2001; Ecke, 2001; Goral, Levy, & Kastl, 2007; Hall & Ecke, 2003; Rothman, 2010; Rothman & Cabrelli Amaro, 2010). For example, Giancaspro, ­Halloran, and Iverson (2015) found that when English (L1) speakers of Spanish (L2) and Spanish (L1) speakers of English (L2) were asked to complete grammaticality



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

judgement tasks in their L3 (Brazilian Portuguese), they were all influenced by the typologically similar Spanish rather than English, irrespective of the L1/L2 status of Spanish. Thus, these studies suggest that multilinguals exhibit patterns of language processing which are unobservable in bilingual speakers. As such, multilinguals should not be treated as bilinguals with multiple “L2s”, and the characteristics and interactions between different L2s should not be overlooked. Further support for the distinction between bilingual and multilingual language processing comes from language production and comprehension work investigating cognate facilitation with trilingual participants (Lemhöfer, ­Dijkstra, & Michel, 2004; Lijewska & Chmiel, 2015; Poarch & van Hell, 2012, 2014; Szubko-Sitarek, 2012, 2014). A particularly interesting pattern of cognate facilitation has been observed in a series of lexical decision tasks reported by Szubko-Sitarek (2014), showing that not all cognates are processed in a similar manner. Specifically, Szubko-Sitarek found that, L1-L3 (Polish-German) cognates lead to significant facilitation, whereas L2-L3 (EnglishGerman) cognates were recognized as rapidly as language-unique words when people read in their L3 (German). Szubko-Sitarek interpreted this pattern as evidence for the influence of learning experience on cognate processing, given that the participants’ L1 was the language of instruction during L3 learning. The observed pattern suggests that stronger links may develop between L1 and L3 than between L2 and L3 due to the concurrent use of L1 and L3 in learnign experience. Similar conclusions were drawn by Lijewska and Chmiel (2015) from a translation task, where Polish-German-English trilinguals translated L2-L3 cognates and L3 control words from the L3 into the L1 or L2. Here, significant L2-L3 cognate facilitation was found only during L3-to-L1 translation (and not in L3-to-L2 translation), which suggested an important role of learning experience in language processing (of note, here again all participants acquired their L3 via their L1). In sum, the above-reviewed work on multilingualism suggests that multilinguals may differ from bilinguals in a number of important ways. However, whether these differences are qualitative or quantitative in nature remains an open question that is beyond the scope of this chapter (for a more thorough discussion of this issue, see De Bot & Jaensch, 2015). Another open question is how bilingualism vs. multilingualism relates to differences in executive control capacity. One possibility is that knowledge of multiple languages would further reduce the integrity of one’s L2, thereby increasing the likelihood of cross-language activation. This is because multilinguals would be more likely to confuse their lesser known languages, especially if the languages are perceived as being structurally similar. Thus, the impact of increased knowledge an L3 (or subsequenct languages) may selectively impair the L2, particularly if the L2 is already weak. However, it is also possible that knowledge of an L3 (and beyond) might collude with knowledge of the L2 in a way that differentially interferes with the L1. Finally, it is possible that both of these conjectures are correct, in that additional knowledge of an

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

L3 (or more) might make processing in both the L1 and L2 more difficult. Of course, it is also possible that the sheer amount of L1, L2, and L3 knowledge may not be the only factor that increases executive control demands, but rather the manner in which the different languages are learned (i.e., if the L3 is learned via the L2 or the L1). In the following section, we make a preliminary attempt to put these ideas to the test, in a novel re-analysis of the Whitford and Titone (2016) eye movement reading data in bilingual vs. multilingual younger and older adults.

5.  D  oes bilingualism vs. multilingualism matter during reading?: A re-analysis of Whitford and Titone (2015a) To address the question of how knowledge of an L3 (or more) may differentially impact the L1 or L2, we examined whether readers’ bilingual vs. multilingual status influences L1 and L2 reading performance. Moreover, in a preliminary attempt to link executive control capacity to these interactions, we examined whether and how this influence changes in younger compared to older adults, under the general assumption (well supported in the cognitive aging literature) of age-related declines in executive control (e.g., Burke, 1997; Burke & Shafto, 2004; Campbell et al., 2012; Darowski et al., 2008; S. Martin et al., 2006; Salthouse & Meinz, 1995; Titone et al., 2000). To this end, we re-analyzed Whitford and Titone’s (2016) data. Recall, Whitford and Titone’s (2016) work elucidated how individual differences in L2 usage impact L1 and L2 reading in bilinguals across the adult lifespan; however, several of their younger (21 out of 31) and older adults (14 out of 31) had knowledge and use of multiple languages. Of note, all bilinguals had French as a first-language (L1), and English as a second-language (L2). The language breakdown was as follows: two languages (French and English: 10 younger and 17 older adults); three languages (French, English, and mostly Spanish, followed by German, and Italian: 16 younger and 11 older adults); and 4+ languages (French, English, and a combination of Spanish, German, Italian, Portuguese, Arabic, Russian, Polish, Greek, and Mandarin: 5 younger and 3 older adults). Participant characteristics as a function of bilingual/multilingual status are presented in Tables 2 (younger adults) and 3 (older adults). Given the past literature reviewed above on bilingual reading and that on l­ anguage processing in multilingualism, we formulated two hypotheses about the outcome of this re-analysis. First, we expected that L2 reading performance would be more affected by the knowledge of additional foreign languages than L1 reading performance, primarily because the latter would be more practiced and entrenched than the former, and because knowledge of an L3 and beyond would be more likely to dilute the impact of an L2 on L1 reading. Second, we expected that the negative impact on L2 reading would be greater as the number of additional foreign languages grew, for example, from just three known languages, to four or more. Finally, we expected that any effects arising from



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

knowledge of an L3 (or L4, etc.) on L1 and L2 reading performance would be exacerbated as a function of whether a multilingual reader in question had limited executive control capacity, that is, whether they were older compared to younger readers. Table 1.  Younger adult characteristics Language Status Bilingual (n = 10) [mean (S.D.)]

Multilingual (n = 21) [mean (S.D.)]

Age (years)

22.60 (3.69)

22.19 (3.50)

Education (years)*

16.60 (1.35)

15.24 (1.61)

L2 AoA (years)

8.30 (3.80)

8.10 (3.97)

Age of L2 Fluency (years)

15.40 (4.53)

14.67 (6.54)

L1

60.00 (9.13)

59.52 (12.03)

L2

39.00 (8.43)

39.29 (11.43)

1.04 (0.07)

1.04 (0.09)

Current language exposure (% time)

Objective language proficiency measure1 L2/L1 RT ratio 1Based

Note: on a ratio derived by dividing correct L2 reaction times by correct L1 reaction times to speeded lexical animacy judgment tasks. *p < .05; all other ps > .34

Table 2.  Older adult characteristics Language Status Bilingual (n = 17) [mean (S.D.)]

Multilingual (n = 14) [mean (S.D.)]

Age (years)

69.53 (5.33)

70.14 (7.69)

Education (years)

15.62 (3.17)

16.11 (2.86)

L2 AoA (years)

9.44 (4.79)

10.11 (7.28)

Age of L2 Fluency (years)

18.06 (10.49)

16.21 (11.84)

L1

65.12 (22.86)

60.50 (21.34)

L2

34.88 (22.86)

38.07 (20.33)

1.00 (0.07)

1.01 (0.07)

Current language exposure (% time)

Objective language proficiency measure1 L2/L1 RT ratio 1Based

Note: on a ratio derived by dividing correct L2 reaction times by correct L1 reaction times to speeded lexical animacy judgment tasks. All ps > .34

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

Assuming that divided language experience shapes how people represent and process language, one possibility is that multilinguals differ from bilinguals in terms of L1 and L2 reading. Thus, to address this issue, we re-analyzed Whitford and Titone’s (2016) data as a function of bilingual vs. multilingual status. Of note, in this study participants read syntactically-simple sentences (e.g., He visits a new country each year on vacation) in their L1 and L2 using the gaze-contingent, moving window paradigm. The specific questions that we pursue in this re-analysis are the following. First, to what extent does reading performance vary as a function of bilingual vs. multilingual status when L2 usage levels and L2 age of acquisition are statistically controlled? Secondly, does the relationship between the bilingual/multilingual status and eye movement behavior in reading specifically impact L1 reading, L2 reading, or both? Finally, does this relationship vary for younger and older adults who differ with respect to their absolute language experience and general executive control capacity? To address these questions, we used linear mixed effects (LME) models to analyze the data via the lme4 package (Bates, 2007; Bates & Sarkar, 2006) within R (version 2.13.1) (Baayen, 2008; R Development Core Team, 2010). Significant effects were those that exceeded a t value of 1.96. Given that ­Whitford and Titone (2016) used a gazecontingent moving window paradigm, they were able to analyze a variety of global eye movement measures (e.g., reading rate, forward fixation durations, and forward saccade lengths), as well as estimates of the perceptual span. However, for the sake of brevity, we only report the findings for reading rate (in number of words per minute) for the full-text reading (i.e., not for the window conditions). Fixed factors included: age group (older vs. younger adults); language (L1 vs. L2); and number of languages known. All fixed factors were deviation coded, which means that all levels of the factor were compared to the grand mean. Control predictors included: current L2 exposure (continuous; z-scored); L2 AoA (continuous; z-scored); and counter-balancing order. Random factors included: participants (random intercept and random slope adjustment for language) and items (random intercept only). Consistent with Whitford and Titone (2016), we found a significant effect of age group (b = 53.92, SE = 8.98, t = 6.00), where reading rates were slower for older than for younger adults (143 vs. 197 words per minute), as well as a significant effect of language (b = –11.84, SE = 3.75, t = –3.16), where reading rates were faster in the L1 than in the L2 (177 vs.163 words per minute). Interestingly, we also found a significant two-way interaction between age group and language status (b = 38.91, SE = 17.94, t = 2.17), where being multilingual related to reduced reading rates in older adults, but faster reading rates in younger adults. Of note, this effect did not interact with language (b = –9.93, SE = 15.00, t = –0.66), thus, we must conclude that is was equally applicable to the L1 and the L2 (see Figures 1 and 2).



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading  Age Group x Language Status Reading Rate (words/minute)



Bilingual Multilingual

     Younger Adults

Older Adults

Figure 1.  Graphical representation of the interaction between Age Group and Language ­Status. Means and standard error bars are plotted

Reading Rate (words/minute)

Age Group x Language Status x Language (L, L)      

L1

L2

Bilingual

L1

L2

Multilingual

Younger Adults

L1

L2

Bilingual

L1

L2

Multilingual

Older Adults

Figure 2.  Graphical representation of the interaction between Age Group, Language Status, and Language. Means and standard error bars are plotted

To determine how the number of languages known to participants might influence these effects, we performed a follow-up analysis that included number of languages (continuous; z-scored) as a fixed effect. Here, again, consistent with Whitford and Titone (2016), we found a significant effect of age group (b = 50.29, SE = 8.93, t = 5.63), where reading rates were slower for older adults than for younger adults (143 vs. 197 words per minute), as well as a significant effect of language (b = –12.87, SE = 3.56, t = –3.62), where reading rates were faster in the L1 than in the L2 (177 vs.163 words per minute). Moreover, we found a significant two-way interaction between language and number of languages (b = –10.72, SE = 3.48, t = –3.07), where knowing 4 languages related to faster reading rates, particularly during L1 reading (a finding likely driven by the younger

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

adults). See Figure 3 for a graphical depiction of these effects. Notably, we also found a significant two-way interaction between age group and number of languages (b = 24.16, SE = 8.93, t = 2.71), where knowing more languages related to reduced reading rates in older adults, but faster reading rates in younger adults. This was again true of both L1 and L2 bilingual reading. See Figure 4 for a graphical depiction of these effects. Given that the three-way interaction between age group, number of languages, and language did not reach significance (b = –9.96, SE = 7.13, t = –1.40), we must again conclude that this effect was, again, true of both L1 and L2 bilingual reading (see Figure 5). Taken together, these re-analyses suggest that reading rates of bilingual participants differed from those of multilingual participants. This indicates that reading performance indeed varies as a function of bilingual vs. multilingual status when L2 usage levels and L2 age of acquisition are statistically controlled. Given that the impact of Language (L1, L2) x Number of Languages Reading Rate (words/minute)



  

    

L1

L2

Figure 3.  Graphical representation of the interaction between Language and Number of ­Languages. Means and standard error bars are plotted Age Group x Number of Languages Reading Rate (words/minute)



  

    

Younger Adults

Older Adults

Figure 4.  Graphical representation of the interaction between Age Group and Number of ­Languages. Means and standard error bars are plotted



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

Reading Rate (words/minute)

Age Group x Number of Languages x Language (L1, L2)      

L1

L2

L1



L2 

Younger Adults

L1

L2 

L1

L2 

L1

L2 

L1

L2 

Older Adults

Figure 5.  Graphical representation of the non-significant interaction between Age Group, Number of Languages, and Language. Means and standard error bars are plotted

individual differences in L3+ usage was comparable for reading in L1 and in L2, we can conclude that the bilingual/multilingual status impacts L1 and L2 reading in a similar fashion. Somewhat unexpectedly, the results showed that multilingual status was related to increased reading rates for younger adults, but decreased reading rates in older adults. This shows that the relationship between the bilingual/multilingual status and eye movement behavior in reading varies for younger and older adults who differ with respect to their absolute language experience and general executive control capacity. But why would increased usage of an L3 (or L4, etc.) facilitate L1 and L2 reading for young adults, but impede the same for older adults? Although we do not have a clear answer to this question, one possibility is that managing the co-activation of multiple languages (as opposed to just two languages) in early adulthood, when readers are operating at peak executive control capacity, might further stimulate the use of executive control during comprehension, which in turn, improves global reading performance across both the L1 and the L2. However, when readers are not operating at peak executive control capacity, as is the case in older adulthood, managing the co-activation of multiple languages (as opposed to just two), might be especially taxing, leading to declines in reading performance across the L1 and the L2. This explanation is admittedly speculative in nature; however, regardless of the exact cognitive mechanism at work, the results clearly show that knowledge of an L3 (or more) is not inconsequential with respect to L1 and L2 reading.

6.  General conclusions In this chapter, we discussed the cognitive demands associated with knowledge of more than one language. Next, we introduced eye-tracking research of reading which,

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

in contrast to other methods often used in psycholinguistics, is more ecologically valid, and has high temporal resolution to conduct more nuanced experimental manipulations. We showed how eye-tracking research has deepened our understanding of the dynamics of cross-language activation during L1 and L2 reading in both early and late adulthood. Subsequently, we discussed the importance of distinguishing between bilingualism and multilingualism, as knowledge of three or more languages may lead to patterns of language processing that differ from those observed in speakers of only two languages. Lastly, to empirically test this hypothesis, we re-analyzed recent eye movement data from our group (Whitford & Titone, 2016) involving younger and older adults. Our re-analysis explored differences among bilinguals and multilinguals in terms of reading performance. Indeed, the findings revealed that knowing more languages (i.e., greater multilingualism) was related to stronger reading performance, as indexed by faster average reading rates; however, in early adulthood only. Conversely, in late adulthood, knowing more languages (i.e., greater multilingualism) was related to reduced reading performance, as indexed by slower average reading rates. These findings suggest that reading performance across multiple languages may be enhanced by greater executive control capacity in early life, but weakened by reduced executive control capacity in later life. While the current study focused on global aspects of reading performance in bilinguals and multilinguals, future work should also examine how bilingual vs. multilingual status impacts local aspects of reading performance (discussed in greater detail towards the beginning of the chapter). In particular, an examination of local, early-stage reading (e.g., first fixation duration, gaze duration) may shed more light on the nature of lexical access in bilinguals vs. monolinguals. This would help determine whether factors found in production or response-based, single word recognition tasks also affect the initial stages of language processing during natural reading. As well, an examination of local, late-stage reading (e.g., number of regressions into a word, total reading time) might potentially reveal differences in higher-order, post-lexical integration between bilinguals and multilinguals. Finally, given the observed differences in reading behavior in bilinguals vs. multilinguals as a function of age (and by proxy, executive function performance), future work should more closely examine how differences in executive control modulate reading across these language groups. Thus, the present re-analysis suggests that what we know from past research with bilinguals may have to be qualified by the presence or absence of knowledge and use of more than two languages. As discussed earlier in this chapter, cross-language activation is sensitive to factors that hinge on executive control capacity. Given that age and bilingual/multilingual status interact with reading performance, studies should



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

also examine more closely how cross-language activation is modulated by executive control and number of known languages. There are additional compelling avenues one could pursue to understand how language processing differs between individuals who know two languages and those who know more than two. One possibility is the study of formulaic language in bilinguals (Carrol & Conklin, 2014, 2015; Cieslicka, 2013; Cieslicka & Heredia, 2011; S­ iyanova-Chanturia, Conklin, & Schmitt, 2011; Titone, Columbus, Whitford, ­Mercier, & Libben, 2015; Titone & Libben, 2014). In particular, it would be interesting to examine whether formulaic language is processed differently when idiomatic expressions or fixed phrases exist only in one of a multilingual’s languages, as opposed to a situation when lexically and semantically equivalent phrases exist across two or more languages of a given individual. It would also be beneficial to utilize eye tracking methods that explore syntactic processing in multilingualism (Clahsen & Felser, 2006; Dussias, ­Valdés Kroff, Guzzardo Tamargo, & Gerfen, 2013; ­Frenck-Mestre & Pynte, 1997; Keating, 2009). Such studies could shed light on how multilinguals disambiguate syntactically complex structures that may differ across their known languages, or how they assign gender to nouns, especially in cases when multilinguals’ languages differ in clause attachment preferences or in patterns of grammatical gender assignment. Finally, other kinds of eye movement protocols could be used to investigate the impact of bilingualism vs. multilingualism for language domains other than reading. For example, the dynamics of bilingual spoken word recognition could also be using the visual world task, wherein eye movements are monitored as people hear spoken language while viewing a linked to a visual display. In existing work investigating bilinguals, researchers have focused mostly on the activation of within- or cross-language competitor words (Blumenfeld & Marian, 2011, 2013; Canseco-Gonzalez et al., 2010; Ju & Luce, 2004; Marian & Spivey, 2003; Mercier, ­Pivneva, & Titone, 2016; Pivneva et al., 2014; Shook & Marian, 2013; Weber & Cutler, 2004). Moreover, the visual world paradigm could also be used to assess how bilingual vs. multilingual experience affects processing of foreign-accented speech (Kang, Rubin, & ­Pickering, 2010; RomeroRivas, M ­ artin, & Costa, 2015; Weber, Broersma, & ­Aoyagi, 2011; Wrembel, 2012). To conclude, we hope that this chapter provides a useful, albeit selective, overview of the different lines of research germane to questions about bilingualism, executive control, and language processes as they mostly pertain to reading. We also hope that it will be an interesting source of evidence that the distinction between bilingualism and multilingualism can be important. Indeed, we believe that an improved understanding of how language processing links to executive control processes would benefit from additional work, using both eye movement methods and other techniques, that better addresses the specific mechanisms by which knowing multiple languages presents special challenges over and above only knowing two.

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak

References Abbott, M., Angele, B., Ahn, Y., & Rayner, K. (2015). Skipping syntactically illegal the previews: The role of predictability. Journal of Experimental Psychology: Learning Memory and ­Cognition, 41(6),1703–14.  doi: 10.1037/xlm0000142 Abutalebi, J., & Green, D. (2007). Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics, 20(3), 242–275.

doi: 10.1016/j.jneuroling.2006.10.003

Abutalebi, J., & Green, D. (2008). Control mechanisms in bilingual language production: Neural evidence from language switching studies. Language and Cognitive Processes, 23(4), 557–582.  doi: 10.1080/01690960801920602 Abutalebi, J., Tettamanti, M., & Perani, D. (2009). The bilingual brain: Linguistic and non-­ linguistic skills. Brain and Language, 109(2–3), 51–54.  doi: 10.1016/j.bandl.2009.04.001 Altarriba, J., Kambe, G., Pollatsek, A., & Rayner, K. (2001). Semantic codes are not used in integrating information across eye fixations in reading: Evidence from fluent spanish-english bilinguals. Perception and Psychophysics, 63(5), 875–890. doi: 10.3758/BF03194444 Altarriba, J., Kroll, J., Sholl, A., & Rayner, K. (1996). The influence of lexical and conceptual constraints on reading mixed-language sentences: Evidence from eye fixations and naming times. Memory & Cognition, 24(4), 477–492.  doi: 10.3758/BF03200936 Andrews, S., & Hersch, J. (2010). Lexical precision in skilled readers: Individual differences in masked neighbor priming. Journal of Experimental Psychology: General, 139, 299–318.

doi: 10.1037/a001836.

Arnold, J., Eisenband, J., Brown-Schmidt, S., & Trueswell, J. (2000). The rapid use of gender information: Evidence of the time course of pronoun resolution from eyetracking. ­Cognition, 76(1), B13-B26.  doi: 10.1017/S1366728911000733 Baayen, R. (2008). Analyzing linguistic data: A practical introduction to statistics using r. ­Cambridge, UK: Cambridge University Press. Balling, W. (2013). Reading authentic texts: What counts as cognate? Bilingualism: Language and Cognition, 16, 637–653. Bates, D. (2007). Linear mixed model implementation in lme4. Unpublished manuscript. Bates, D., & Sarkar, D. (2006). Me4: Linear mixed-effects models using s4 classes. Accessed from: http://cran.R-project.org. Baum, S., & Titone, D. (2014). Moving toward a neuroplasticity view of bilingualism, executive control, and aging. Applied Psycholinguistics, 35(5), 857–894. doi: 10.1017/ S0142716414000174

Bialystok, E. (2001). Bilingualism in development: Language, literacy, and cognition. New York, NY: Cambridge University Press.  doi: 10.1017/CBO9780511605963 Bialystok, E., Craik, F., Klein, R., & Viswanathan, M. (2004). Bilingualism, aging, and cognitive control: Evidence from the Simon task. Psychology and Aging, 19(2), 290–303.

doi: 10.1037/0882-7974.19.2.290

Bialystok, E., Luk, G., Peets, K., & Yang, S. (2010). Receptive vocabulary differences in monolingual and bilingual children. Bilingualism-Language and Cognition, 13(4), 525–531.

doi: 10.1017/S1366728909990423

Bialystok, E., & Martin, M. (2004). Attention and inhibition in bilingual children: Evidence from the dimensional change card sort task. Developmental Science, 7(3), 325–339.

doi: 10.1111/j.1467-7687.2004.00351.x



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

Blumenfeld, H., & Marian, V. (2011). Bilingualism influences inhibitory control in auditory comprehension. Cognition, 118(2), 245–257.  doi: 10.1016/j.cognition.2010.10.012 Blumenfeld, H., & Marian, V. (2013). Parallel language activation and cognitive control during spoken word recognition in bilinguals. Journal of Cognitive Psychology, 25, 547–567.

doi: 10.1080/20445911.2013.812093

Braver, T. (2012). The variable nature of cognitive control: A dual mechanisms framework. Trends in Cognitive Sciences, 16(2), 106–113.  doi: 10.1016/j.tics.2011.12.010 Bultena, S., Dijkstra, T., & van Hell, J. (2014). Cognate effects in sentence context depend on word class, L2 proficiency, and task. The Quarterly Journal of Experimental Psychology, 67, 1214–1241.  doi: 10.1080/17470218.2013.853090 Burke, D. (1997). Language, aging, and inhibitory deficits: Evaluation of a theory. Journals of Gerontology Series B-Psychological Sciences and Social Sciences, 52(6), P254-P264.

doi: 10.1093/geronb/52B.6.P254

Burke, D., & Shafto, M. (2004). Aging and language production. Current Directions in Psychological Science, 13(1), 21–24.  doi: 10.1111/j.0963-7214.2004.01301006.x Campbell, K., Grady, C., Ng, C., & Hasher, L. (2012). Age differences in the frontoparietal cognitive control network: Implications for distractibility. Neuropsychologia, 50(9), 2212–2223. doi: 10.1016/j.neuropsychologia.2012.05.025 Canseco-Gonzalez, E., Brehm, L., Brick, C., Brown-Schmidt, S., Fischer, K., & Wagner, K. (2010). Carpet or carcel: The effect of age of acquisition and language mode on bilingual lexical access. Language and Cognitive Processes, 25(5), 669–705.  doi: 10.1080/01690960903474912 Carrol, G., & Conklin, K. (2014). Getting your wires crossed: Evidence for fast processing of L1 idioms in an L2. Bilingualism-Language and Cognition, 17(4), 784–797.

doi: 10.1017/S1366728913000795

Carrol, G., & Conklin, K. (2015). Cross language lexical priming extends to formulaic units: Evidence from eye-tracking suggests that this idea ‘has legs’ Bilingualism: Language and Cognition. Cenoz, J., & Gorter, D. (2011). Focus on multilingualism: A study of trilingual writing. Modern Language Journal, 95(3), 356–369.  doi: 10.1111/j.1540-4781.2011.01206.x Christoffels, I., Firk, C., & Schiller, N. (2007). Bilingual language control: An event-related brain potential study. Brain Research, 1147, 192–208.  doi: 10.1016/j.brainres.2007.01.137 Cieslicka, A. (2013). Do nonnative language speakers chew the fat and spill the beans with different brain hemispheres? Investigating idiom decomposability with the divided visual field paradigm. Journal of Psycholinguistic Research, 42(6), 475–503.

doi: 10.1007/s10936-012-9232-4

Cieslicka, A., & Heredia, R. (2011). Hemispheric asymmetries in processing L1 and L2 idioms: Effects of salience and context. Brain and Language, 116(3), 136–150.

doi: 10.1016/j.bandl.2010.09.007

Clahsen, H., & Felser, C. (2006). Grammatical processing in language learners. Applied Psycholinguistics, 27(1), 3–42.  doi: 10.1017/S0142716406060024 Clifton, C., Ferreira, F., Henderson, J., Inhoff, A., Liversedge, S., Reichle, E., & Schotter, E. (2016). Eye movements in reading and information processing: Keith Rayner’s 40-year legacy. Journal of Memory and Language, 86, 1–19.  doi: 10.1016/j.jml.2015.07.004 Colzato, L., Bajo, M., van den Wildenberg, W., Paolieri, D., Nieuwenhuis, S., La Heij, W., & Hommel, B. (2008). How does bilingualism improve executive control? A comparison of active and reactive inhibition mechanisms. Journal of Experimental Psychology-Learning Memory and Cognition, 34(2), 302–312.  doi: 10.1037/0278-7393.34.2.302

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak Comesaña, M., Ferre, P., Romero, J., Guasch, M., Soares, A., & Garcia-Chico, T. (2015). Facilitative effect of cognate words vanishes when reducing the orthographic overlap: The role of stimuli list composition. Journal of Experimental Psychology-Learning Memory and ­Cognition, 41(3), 614–635.  doi: 10.1037/xlm0000065 Cop, U., Keuleers, E., Drieghe, D., & Duyck, W. (2015). Frequency effects in monolingual and bilingual natural reading. Psychonomic Bulletin and Review, 22(5), 1216–1234.

doi: 10.3758/s13423-015-0819-2

Costa, A., Hernandez, M., Costa-Faidella, J., & Sebastian-Galles, N. (2009). On the bilingual advantage in conflict processing: Now you see it, now you don’t. Cognition, 113(2), 135–149.  doi: 10.1016/j.cognition.2009.08.001 Darowski, E., Helder, E., Zacks, R., Hasher, L., & Hambrick, D. (2008). Age-related differences in cognition: The role of distraction control. Neuropsychology, 22(5), 638–644.

doi: 10.1037/0894-4105.22.5.638

Davis, C., Sanchez-Casas, R., Garcia-Albea, J., Guasch, M., Molero, M., & Ferre, P. (2010). Masked translation priming: Varying language experience and word type with spanish– english bilinguals. Bilingualism: Language and Cognition, 13(02), 137–155.

doi: 10.1017/S1366728909990393

De Angelis, G. (2007). Third or additional language acquisition. Clevedon, UK: Multilingual Matters. De Bot, K. (2012). Rethinking multilingual processing: From a static to a dynamic approach. In J. Cabrelli Amaro, S. Flynn, & J. Rothman (Eds.), Third Language Acquisition in Adulthood (pp. 79–94). Amsterdam, The Netherlands: John Benjamins.  doi: 10.1075/sibil.46.07bot De Bot, K., & Jaensch, C. (2015). What is special about L3 processing? Bilingualism: Language and Cognition, 18(2), 130–144.  doi: 10.1017/S1366728913000448 de Bruin, A., Treccani, B., & Della Sala, S. (2015). Cognitive advantage in bilingualism: An example of publication bias? Psychological Science, 26(1), 99–107.  doi: 10.1177/095679761 4557866

De Groot, A. (2011). Language and cognition in bilinguals and multilinguals: An introduction. New York, NY: Psychology Press. De Groot, A., & Christoffels, I. (2006). Language control in bilinguals: Monolingual tasks and simultaneous interpreting. Bilingualism: Language and Cognition, 9(2), 189–201.

doi: 10.1017/S1366728906002537

De Groot, A., Delmaar, P., & Lupker, S. (2000). The processing of interlexical homographs in translation recognition and lexical decision: Support for non-selective access to bilingual memory. The Quarterly Journal of Experimental Psychology, 53(2), 397–428.

doi: 10.1080/713755891

Dewaele, J. (2001). Activation or inhibition? The interaction of L1, L2 and L3 on the language mode continuum. In U. Jessner, B. Hufeisen, & J. Cenoz (Eds.), Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives (pp. 69–89). Clevedon, UK: Multilingual Matters. Dijkstra, T. (2005). Bilingual visual word recognition and lexical access. In J. Kroll & A. de Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp. 179–201). New York, NY: Oxford University Press. Dijkstra, T., Grainger, J., & van Heuven, W. (1999). Recognition of cognates and interlingual homographs: The neglected role of phonology. Journal of Memory and Language, 41(4), 496–518.  doi: 10.1006/jmla.1999.2654



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

Dijkstra, T., Miwa, K., Brummelhuis, B., Sappelli, M., & Baayen, R. (2010). How cross-language similarity and task demands affect cognate recognition. Journal of Memory and Language, 62(3), 284–301.  doi: 10.1016/j.jml.2009.12.003 Dijkstra, T., Timmermans, M., & Schriefers, H. (2000). On being blinded by our other language: Effects of task demands on interlingual homograph recognition. Journal of Memory and Language, 42, 445–464.  doi: 10.1006/jmla.1999.2697 Dijkstra, T., & van Heuven, W. (2002). The architecture of the bilingual word recognition system: From identification to decision. Bilingualism: Language and Cognition, 5(3), 175–197. Dijkstra, T., Van Jaarsveld, H., & Brinke, S. (1998). Interlingual homograph recognition: Effects of task demands and language intermixing. Bilingualism: Language and Cognition, 1, 51–66.  doi: 10.1017/S1366728998000121 Duffy, S., Kambe, G., & Rayner, K. (2001). The effect of prior disambiguating context on the comprehension of ambiguous words: Evidence from eye movements. In D. Gorfein (Ed.), On the consequences of meaning selection: Perspectives on resolving lexical ambiguity (pp. 27–43). Washington, DC: American Psychological Association.  doi: 10.1037/10459-002 Dussias, P., Valdés Kroff, J., Guzzardo Tamargo, R., & Gerfen, C. (2013). When gender and looking go hand in hand. Studies in Second Language Acquisition, 35, 353–387.

doi: 10.1017/S0272263112000915

Duyck, W. (2005). Translation and associative priming with cross-lingual pseudohomophones: Evidence for nonselective phonological activation in bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(6), 1340–1359.  doi: 10.1037/0278-7393.31.6.1340 Duyck, W., Van Assche, E., Drieghe, D., & Hartsuiker, R. (2007). Visual word recognition by bilinguals in a sentence context: Evidence for nonselective lexical access. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33, 663–679.  doi: 10.1037/0278-7393.33.4.663 Ecke, P. (2001). Lexical retrieval in a third language: Evidence from errors and tip-of-the-tongue states. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic aspects of L3 acquisition: Psycholinguistic perspectives (pp. 90–114). Clevedon, UK: Multilingual Matters. Falk, Y., & Bardel, C. (2011). Object pronouns in german L3 syntax: Evidence for the l2 status factor. Second Language Research, 27(1), 59–82.  doi: 10.1177/0267658310386647 Falk, Y., & Bardel, C. (2012). The L2 status factor and the declarative/procedural distinction. In J. Cabrelli Amaro, S. Flynn, & J. Rothman (Eds.), The L2 status factor and the declarative/ procedural distinction (pp. 61–78). Amsterdam, The Netherlands: John Benjamins.

doi: 10.1075/sibil.46.06bar

Favreau, M., & Segalowitz, N. (1983). Automatic and controlled processes in the 1st-language and 2nd-language reading of fluent bilinguals. Memory & Cognition, 11(6), 565–574.

doi: 10.3758/BF03198281

Flynn, S., Foley, C., & Vinnitskaya, I. (2010). The cumulative-enhancement model for language acquisition: Comparing adults’ and children’s patterns of development in first, second and third language acquisition of relative clauses. International Journal of Multilingualism, 1, 3–16.  doi: 10.1080/14790710408668175 Folk, J., & Morris, R. (2003). Effects of syntactic category assignment on lexical ambiguity resolution in reading: An eye movement analysis. Memory & Cognition, 31(1), 87–99.

doi: 10.3758/BF03196085

Frenck-Mestre, C., & Pynte, J. (1997). Syntactic ambiguity resolution while reading in second and native languages. Quarterly Journal of Experimental Psychology Section a-Human Experimental Psychology, 50(1), 119–148.  doi: 10.1080/027249897392251

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak Friesen, D., & Jared, D. (2007). Cross-language message- and word-level transfer effects in bilingual text processing. Memory & Cognition, 35, 1542–1556.  doi: 10.3758/BF03193489 Garcia-Mayo, M., & Rothman, J. (2012). L3 morphosyntax in the generative tradition: The initial stages and beyond. In J. Cabrelli Amaro, S. Flynn, & J. Rothman (Eds.), Third language acquisition in adulthood. Amsterdam, The Netherlands: John Benjamins.

doi: 10.1075/sibil.46.04pil

Gathercole, V., Thomas, E., Kennedy, I., Prys, C., Young, N., Guasch, N., … Jones, L. (2014). Does language dominance affect cognitive performance in bilinguals? Lifespan evidence from preschoolers through older adults on card sorting, simon, and rnetalinguistic tasks. Frontiers in Psychology, 5.  doi: 10.3389/fpsyg.2014.00011 Giancaspro, D., Halloran, B., & Iverson, M. (2015). Transfer at the initial stages of L3 brazilian portuguese: A look at three groups of english/spanish bilinguals. [Article]. Bilingualism, 18(2), 191–207.  doi: 10.1017/S1366728914000339 Gibson, M., & Hufeisen, B. (2003). Investigating the role of prior foreign language knowledge: Translating from an unknown into a known foreign language. In J. Cenoz, B. Hufeisen & U. Jessner (Eds.), The multilingual lexicon (pp. 87–102). Dordrecht, The Netherlands: Springer.  doi: 10.1007/978-0-306-48367-7_7 Gold, B., Kim, C., Johnson, N., Kryscio, R., & Smith, C. (2013). Lifelong bilingualism maintains neural efficiency for cognitive control in aging. Journal of Neuroscience, 33(2), 387–396.  doi: 10.1523/JNEUROSCI.3837-12.2013 Gollan, T., Montoya, R., Cera, C., & Sandoval, T. (2008). More use almost always means a smaller frequency effect: Aging, bilingualism, and the weaker links hypothesis. Journal of Memory and Language, 58(3), 787–814.  doi: 10.1016/j.jml.2007.07.001 Gollan, T., Slattery, T., Goldenberg, D., Van Assche, E., Duyck, W., & Rayner, K. (2011). ­Frequency drives lexical access in reading but not in speaking: The frequency-lag hypothesis. Journal of Experimental Psychology: General, 140(2), 186–209.  doi: 10.1037/a0022256 Goral, M., Levy, E., & Kastl, R. (2007). Cross-language treatment generalization: A case of trilingual aphasia. Brain and Language, 103(1–2), 203–204.  doi: 10.1016/j.bandl.2007.07.116 Green, D. (1998). Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition, 1, 67–81.  doi: 10.1017/S1366728998000133 Green, D. (2011). Language control in different contexts: The behavioral ecology of bilingual speakers. Frontiers in Psychology, 2.  doi: 10.3389/fpsyg.2011.00103 Green, D., & Abutalebi, J. (2013). Language control in bilinguals: The adaptive control hypothesis. Journal of Cognitive Psychology, 25(5), 515–530.  doi: 10.1080/20445911.2013.796377 Green, D., & Wei, L. (2014). A control process model of code-switching. Language Cognition and Neuroscience, 29(4), 499–511.  doi: 10.1080/23273798.2014.882515 Grosjean, F. (2010). Bilingual: Life and reality: Cambridge, MA: Harvard University Press.

doi: 10.4159/9780674056459

Guo, T., Liu, H., Misra, M., & Kroll, J. (2011). Local and global inhibition in bilingual word production: fMRI evidence from chinese-english bilinguals. Neuroimage, 56(4), 2300–2309. doi: 10.1016/j.neuroimage.2011.03.049 Hall, C., & Ecke, P. (2003). Parasitism as a default mechanism in l3 vocabulary acquisition. In J. Cenoz, B. Hufeisen & U. Jessner (Eds.), The multilingual lexicon (pp. 71–85). Dordrecht, The Netherlands: Springer.  doi: 10.1007/978-0-306-48367-7_6 Hammarberg, B., & Hammarberg, B. (1993). Articulatory resetting in the acquisition of new languages. Phonum, 2, 61–67.  doi: 10.1037/0278-7393.16.3.417



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

Henderson, J., & Ferreira, F. (1990). Effects of foveal processing difficulty on the perceptual span in reading: Implications for attention and eye movement control. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16(3), 417–429. Hoverstern, L., & Traxler, M. (2016). A time course analysis of interlingual homograph processing: Evidence from eye movements. Bilingualism: Language and Cognition. Huey, E. (1908). The psychology and pedagogy of reading. New York, NY: The Macmillan Company. Immordino-Yang, M., & Deacon, T. (2007). An evolutionary perspective on reading and reading disorders? In K. Fischer, J. Bernstein, & M. Immordino-Yang (Eds.), Mind, brain, and education in reading disorders. Cambridge, UK: Cambridge University Press. Ito, K., & Speer, S. (2008). Anticipatory effects of intonation: Eye movements during instructed visual search. Journal of Memory and Language, 58(2), 541–573.  doi: 10.1016/j. jml.2007.06.013

Itzhak, I., & Baum, S. (2015). Misleading bias-driven expectations in referential processing and the facilitative role of contrastive accent. Journal of Psycholinguistic Research, 44(5), 623–650.  doi: 10.1007/s10936-014-9306-6 Ivanova, I., & Costa, A. (2008). Does bilingualism hamper lexical access in speech production? Acta Psychologica, 127(2), 277–288.  doi: 10.1016/j.actpsy.2007.06.003 Jordan, T., Almabruk, A., Gadalla, E., McGowan, V., White, S., Abedipour, L., & Paterson, K. (2014). Reading direction and the central perceptual span: Evidence from arabic and english. Psychonomic Bulletin and Review, 21(2), 505–511.  doi: 10.3758/s13423-013-0510-4 Ju, M., & Luce, P. (2004). Falling on sensitive ears: Constraints on bilingual lexical activation. Psychological Science, 15(5), 314–318.  doi: 10.1111/j.0956-7976.2004.00675.x Just, M., & Carpenter, P. (1980). A theory of reading: From eye fixations to comprehension. Psychological Review, 87(4), 329–354. Kambe, G., Rayner, K., & Duffy, S. (2001). Global context effects on processing lexically ambiguous words: Evidence from eye fixations. Memory & Cognition, 29(2), 363–372.

doi: 10.3758/BF03194931

Kang, O., Rubin, D., & Pickering, L. (2010). Suprasegmental measures of accentedness and judgments of language learner proficiency in oral english. Modern Language Journal, 94(4), 554–566.  doi: 10.1111/j.1540-4781.2010.01091.x Keating, G. (2009). Sensitivity to violations of gender agreement in native and nonnative spanish: An eye-movement investigation. Language Learning, 59(3), 503–535.

doi: 10.1111/j.1467-9922.2009.00516.x

Kjelgaard, M., Titone, D., & Wingfield, A. (1999). The influence of prosodic structure on the interpretation of temporary syntactic ambiguity by young and elderly listeners. ­Experimental Aging Research, 25(3), 187–207.  doi: 10.1080/036107399243986 Kousaie, S., & Phillips, N. (2012). Ageing and bilingualism: Absence of a “bilingual advantage” in stroop interference in a nonimmigrant sample. Quarterly Journal of Experimental Psychology, 65(2), 356–369.  doi: 10.1080/17470218.2011.604788 Kroll, J., Bobb, S., Misra, M., & Guo, T. (2008). Language selection in bilingual speech: Evidence for inhibitory processes. Acta Psychologica, 128(3), 416–430.

doi: 10.1016/j.actpsy.2008.02.001

Kroll, J., Dussias, P., Bogulski, C., & Kroff, J. (2012). Juggling two languages in one mind: What bilinguals tell us about language processing and its consequences for cognition. In B. Ross (Ed.), Psychology of learning and motivation, (pp. 229–262). Oxford, UK: Elsevier.

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak Lee, H., Rayner, K., & Pollatsek, A. (1999). The time course of phonological, semantic, and orthographic coding in reading: Evidence from the fast-priming technique. Psychonomic Bulletin and Review, 6(4), 624–634.  doi: 10.3758/BF03212971 Lemhofer, K., & Dijkstra, T. (2004). Recognizing cognates and interlingual homographs: Effects of code similality in language-specific and generalized lexical decision. Memory & Cognition, 32(4), 533–550.  doi: 10.3758/BF03195845 Lemhofer, K., Dijkstra, T., & Michel, M. (2004). Three languages, one echo: Cognate effects in trilingual word recognition. Language and Cognitive Processes, 19(5), 585–611.

doi: 10.1080/01690960444000007

Libben, M., & Titone, D. (2009). Bilingual lexical access in context: Evidence from eye movements during reading. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(2), 381–390.  doi: 10.1037/a0014875 Lijewska, A., & Chmiel, A. (2015). Cognate facilitation in sentence context? Translation production by interpreting trainees and non-interpreting trilinguals. International Journal of Multilingualism, 12, 1–18.  doi: 10.1080/14790718.2014.959961 Liversedge, S., Gilchrist, I., & Everling, S. (2011). The Oxford handbook of eye movements. Oxford, UK: Oxford University Press. Llama, R., Cardoso, W., & Collins, L. (2010). The influence of language distance and language status on the acquisition of L3 phonology. International Journal of Multilingualism, 7(1), 39–57.  doi: 10.1080/14790710902972255 Macizo, P., Bajo, M., & Martin, M. (2010). Inhibitory processes in bilingual language comprehension: Evidence from spanish-english interlexical homographs. Journal of Memory and Language, 63(2), 232–244.  doi: 10.1016/j.jml.2010.04.002 Marian, V., & Spivey, M. (1999). Activation of russian and english cohorts during bilingual spoken word recognition. Proceedings of the Twenty First Annual Conference of the Cognitive Science Society, 349–354. Marian, V., & Spivey, M. (2003). Bilingual and monolingual processing of competing lexical items. Applied Psycholinguistics, 24(2), 173–193.  doi: 10.1017/S0142716403000092 Martin, M., Macizo, P., & Bajo, M. (2010). Time course of inhibitory processes in bilingual language processing. British Journal of Psychology, 101, 679–693.  doi: 10.1348/000712609X480571 Martin, S., Brouillet, D., Guerdoux, E., & Tarrago, R. (2006). Inhibition and resource capacity during normal aging: A confrontation of the dorsal-ventral and frontal models in a modified version of negative priming. Encephale-Revue De Psychiatrie Clinique Biologique Et Therapeutique, 32(2), 253–262. McConkie, G., & Rayner, K. (1975). The span of the effective stimulus during a fixation in reading. Perception & Psychophysics, 17, 578–586.  doi: 10.3758/BF03203972 Mercier, J., Pivneva, I., & Titone, D. (2014). Individual differences in inhibitory control relate to bilingual spoken word processing. Bilingualism-Language and Cognition, 17(1), 89–117.  doi: 10.1017/S1366728913000084 Mercier, J., Pivneva, I., & Titone, D. (2016). The role of prior language context on bilingual spoken word processing: Evidence from the visual world task. Bilingualism: Language and Cognition. Meuter, R., & Allport, A. (1999). Bilingual language switching in naming: Asymmetrical costs of language selection. Journal of Memory and Language, 40(1), 25–40.

doi: 10.1006/jmla.1998.2602

Meyer, A., & Federmeier, K. (2008). The divided visual world paradigm: Eye tracking reveals hemispheric asymmetries in lexical ambiguity resolution. Brain Research, 1222, 166–183.



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

doi: 10.1016/j.brainres.2008.05.033 Michael, E., & Gollan, T. (2005). Being and becoming bilingual: Individual differences and consequences for language production. In J. Kroll & A. de Groot (Eds.), The handbook of bilingualism: Psycholinguistic approaches. New York, NY: Oxford University Press. Midgley, K., Holcomb, P., & Grainger, J. (2011). Effects of cognate status on word comprehension in second language learners: An ERP investigation. Journal of Cognitive Neuroscience, 23(7), 1634–1647.  doi: 10.1162/jocn.2010.21463 Misra, M., Guo, T., Bobb, S., & Kroll, J. (2012). When bilinguals choose a single word to speak: Electrophysiological evidence for inhibition of the native language. Journal of Memory and Language, 67(1), 224–237.  doi: 10.1016/j.jml.2012.05.001 Morales, J., Gomez-Ariza, C., & Bajo, M. (2013). Dual mechanisms of cognitive control in bilinguals and monolinguals. Journal of Cognitive Psychology, 25(5), 531–546.

doi: 10.1080/20445911.2013.807812

Morales, J., Yudes, C., Gomez-Ariza, C., & Bajo, M. (2015). Bilingualism modulates dual mechanisms of cognitive control: Evidence from ERPs. Neuropsychologia, 66, 157–169.

doi: 10.1016/j.neuropsychologia.2014.11.014

Nievas, F., & Mari-Beffa, P. (2002). Negative priming from the non-selected meaning of the homograph. British Journal of Psychology, 93, 47–66.  doi: 10.1348/000712602162445 Otwinowska-Kasztelanic, A. (2015). Cognate vocabulary in language acquisition and use. Bristol, UK: Multilingual Matters. Paap, K. (2014). The role of componential analysis, categorical hypothesising, replicability and confirmation bias in testing for bilingual advantages in executive functioning. Journal of Cognitive Psychology, 26(3), 242–255.  doi: 10.1080/20445911.2014.891597 Paap, K., & Greenberg, Z. (2013). There is no coherent evidence for a bilingual advantage in executive processing. Cognitive Psychology, 66(2), 232–258.  doi: 10.1016/j.cogpsych.2012.12.002 Paradis, M. (2000). Generalizable outcomes of bilingual aphasia research. Folia Phoniatrica Et Logopaedica, 52(1–3), 54–64.  doi: 10.1159/000021513 Paterson, K., McGowan, V., White, S., Malik, S., Abedipour, L., & Jordan, T. (2014). Reading direction and the central perceptual span in urdu and english. PLoS ONE, 9(2). Pauker, E., Itzhak, I., Baum, S., & Steinhauer, K. (2011). Effects of cooperating and conflicting prosody in spoken english garden path sentences: ERP evidence for the boundary deletion hypothesis. Journal of Cognitive Neuroscience, 23(10), 2731–2751.

doi: 10.1162/jocn.2011.21610

Perfetti, C. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies of Reading, 11, 357–383.  doi: 10.1080/10888430701530730 Perfetti, C., & Hart, L. (2002). The lexical quality hypothesis. In L. Verhoeven, C. Elbro, & P. Reitsma (Eds.), Precursors of functional literacy (pp. 189–213). Amsterdam, The ­Netherlands: John Benjamins.  doi: 10.1075/swll.11.14per Pivneva, I., Mercier, J., & Titone, D. (2014). Executive control modulates cross-language lexical activation during L2 reading: Evidence from eye movements. Journal of Experimental Psychology-Learning Memory and Cognition, 40(3), 787–796.  doi: 10.1037/a0035583 Pivneva, I., Palmer, C., & Titone, D. (2012). Inhibitory control and L2 proficiency modulate bilingual language production: Evidence from spontaneous monologue and dialogue speech. Frontiers in Psychology, 3.  doi: 10.3389/fpsyg.2012.00057 Poarch, G., & van Hell, J. (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.  doi: 10.1016/j.jecp.2012.06.013

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak Poarch, G., & van Hell, J. (2014). Cross-language activation in same-script and different-script trilinguals. International Journal of Bilingualism, 18(6), 693–716.

doi: 10.1177/1367006912472262

Pollatsek, A., Bolozky, S., Well, A., & Rayner, K. (1981). Asymmetries in the perceptual span for israeli readers. Brain and Language, 14(1), 174–180.  doi: 10.1016/0093-934X(81)90073-0 Prior, A., & MacWhinney, B. (2010). A bilingual advantage in task switching. Bilingualism: Language and Cognition, 13(2), 253–262.  doi: 10.1017/S1366728909990526 R Development Core Team. (2010). R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing Radach, R., & Kennedy, A. (2013). Eye movements in reading: Some theoretical context. Quarterly Journal of Experimental Psychology, 66(3), 429–452.  doi: 10.1080/17470218.2012.750676 Rayner, K. (1975). The perceptual span and peripheral cues in reading. Cognitive Psychology, 7(1), 65–81.  doi: 10.1016/0010-0285(75)90005-5 Rayner, K. (1998). Eye momvement in reeading and information processing: 20 years of research. Psychological Bulletin, 124, 372–422.  doi: 10.1037/0033-2909.124.3.372 Rayner, K. (2009). Eye movements and attention in reading, scene perception, and visual search. Quarterly Journal of Experimental Psychology, 62, 1457–1506.

doi: 10.1080/17470210902816461

Rayner, K. (2014). The gaze-contingent moving window in reading: Development and review. Visual Cognition, 22(3–4), 242–258.  doi: 10.1080/13506285.2013.879084 Rayner, K., & Bertera, J. (1979). Reading without a fovea. Science, 206(4417), 468–469.

doi: 10.1126/science.504987

Rayner, K., & McConkie, G. (1976). What guides a reader’s eye movements? Vision Research, 16(8), 829–837.  doi: 10.1016/0042-6989(76)90143-7 Rayner, K., Pollatsek, A., Ashby, J., & Clifton, C. (2012). Psychology of reading. New York, NY: Psychology Press. Rayner, K., Well, A., & Pollatsek, A. (1980). Asymmetry of the effective visual field in reading. Perception & Psychophysics, 27(6), 537–544.  doi: 10.3758/BF03198682 Romero-Rivas, C., Martin, C., & Costa, A. (2015). Processing changes when listening to foreignaccented speech. Frontiers in Human Neuroscience, 9.  doi: 10.3389/fnhum.2015.00167 Rothman, J. (2010). On the typological economy of syntactic transfer: Word order and relative clause high/low attachment preference in L3 brazilian portuguese. International Review of Applied Linguistics (IRAL), 48, 245–273. Rothman, J. (2013). Cognitive economy, non-redundancy and typological primacy in l3 acquisition: Evidence from initial stages of l3 romance. In S. Baauw, F. Dirjkoningen, & M. Pinto (Eds.), Romance languages and linguistic theory (pp. 217–248). Amsterdam, The ­Netherlands: John Benjamins.  doi: 10.1075/rllt.5.11rot Rothman, J., & Cabrelli Amaro, J. (2010). What variables condition syntactic transfer? A look at the L3 initial state. Second Language Research, 26(2), 189–218.  doi: 10.1177/0267658309349410 Salthouse, T., & Meinz, E. (1995). Aging, inhibition, working-memory, and speed. Journals of Gerontology Series B-Psychological Sciences and Social Sciences, 50(6), P297-P306.

doi: 10.1093/geronb/50B.6.P297

Schmid, M. (2013). First language attrition. WIRE’s Cognitive Science, 4, 117–123.

doi: 10.1002/wcs.1218

Schotter, E., Angele, B., & Rayner, K. (2012). Parafoveal processing in reading. Attention, Perception, and Psychophysics, 74, 5–35.  doi: 10.3758/s13414-011-0219-2



Chapter 1.  Bilingualism, executive control, and eye movement measures of reading 

Schwartz, A. I., & Kroll, J. (2006). Bilingual lexical activation in sentence context. Journal of Memory and Language, 55(2), 197–212.  doi: 10.1016/j.jml.2006.03.004 Segalowitz, N. (2010). The cognitive bases of second language fluency. New York, NY: Routledge. Segalowitz, N., & Hulstijn, J. (2005). Automaticity in bilingualism and second language learning. In J. Kroll & A. De Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp. 371–388). New York, NY: Oxford University Press. Sereno, S., & Rayner, K. (1992). Fast priming during eye fixations in reading. Journal of Experimental Psychology: Human Perception and Performance, 18(1), 173–184.

doi: 10.1037/0096-1523.18.1.173

Shook, A., & Marian, V. (2013). The bilingual language interaction network for comprehension of speech. Bilingualism: Language and Cognition, 16(2), 304–324.

doi: 10.1017/S1366728912000466

Simpson, G., & Krueger, M. (1991). Selective access of homograph meanings in sentence context. Journal of Memory and Language, 30(6), 627–643.  doi: 10.1016/0749-596X(91)90029-J Siyanova-Chanturia, A., Conklin, K., & Schmitt, N. (2011). Adding more fuel to the fire: An eyetracking study of idiom processing by native and non-native speakers. Second Language Research, 27(2), 251–272.  doi: 10.1177/0267658310382068 Sturt, P., Scheepers, C., & Pickering, M. (2002). Syntactic ambiguity resolution after initial misanalysis: The role of recency. Journal of Memory and Language, 46(2), 371–390.

doi: 10.1006/jmla.2001.2807

Szubko-Sitarek, W. (2012). Cognate facilitation effects in trilingual word recognition. Studies in Second Language Learning and Teaching, 1, 189–208. Szubko-Sitarek, W. (2014). Multilingual lexical recognition in the mental lexicon of third language users (2015 edition). New York, NY: Springer. Titone, D., & Baum, S. (2014). The future of bilingualism research: Insufferably optimistic and replete with new questions. Applied Psycholinguistics, 35(5), 933–942.

doi: 10.1017/S0142716414000289

Titone, D., Columbus, G., Whitford, V., Mercier, J., & Libben, M. (2015). Contrasting bilingual and monolingual idiom processing. In R. Heredia & A. Cieślicka (Eds.), Bilingual figurative language (pp. 171–207). Cambridge, UK: Cambridge University Press.

doi: 10.1017/CBO9781139342100.011

Titone, D., Koh, C., Kjelgaard, M., Bruce, S., Speer, S., & Wingfield, A. (2006). Age-related impairments in the revision of syntactic misanalyses: Effects of prosody. Language and Speech, 49(1), 75–99.  doi: 10.1177/00238309060490010501 Titone, D., & Libben, M. (2014). Time-dependent effects of decomposability, familiarity and literal plausibility on idiom priming: A cross-modal priming investigation. The Mental Lexicon, 9(3), 473–496.  doi: 10.1075/ml.9.3.05tit Titone, D., Libben, M., Mercier, J., Whitford, V., & Pivneva, I. (2011). Bilingual lexical access during L1 sentence reading: The effects of L2 knowledge, semantic constraint, and L1-L2 intermixing. Journal of Experimental Psychology-Learning Memory and Cognition, 37(6), 1412–1431.  doi: 10.1037/a0024492 Titone, D., Prentice, K., & Wingfield, A. (2000). Resource allocation during spoken discourse processing: Effects of age and passage difficulty as revealed by self-paced listening. Memory & Cognition, 28(6), 1029–1040.  doi: 10.3758/BF03209351 Van Assche, E., Duyck, W., & Hartsuiker, R. J. (2012). Bilingual word recognition in a sentence context. Frontiers in Psychology, 3.

 Debra Titone, Veronica Whitford, Agnieszka Lijewska & Inbal Itzhak Van Assche, E., Duyck, W., Hartsuiker, R., & Diependaele, K. (2009). Does bilingualism change native-language reading? Cognate effects in a sentence context. Psychological Science, 20(8), 923–927.  doi: 10.1111/j.1467-9280.2009.02389.x Van Hell, J., & De Groot, A. (2008). Sentence context modulates visual word recognition and translation in bilinguals. Acta Psychologica, 128(3), 431–451.doi: 10.1016/j.actpsy.2008.03.010 Van Hell, J., & Tanner, D. (2012). Second language proficiency and cross-language lexical activation. Language Learning, 62, 148–171.  doi: 10.1111/j.1467-9922.2012.00710.x Von Studnitz, R., & Green, D. (2002). The cost of switching language in a semantic categorization task. Bilingualism: Language and Cognition, 5(3), 241–251. Weber, A., Broersma, M., & Aoyagi, M. (2011). Spoken-word recognition in foreign-accented speech by L2 listeners. Journal of Phonetics, 39(4), 479–491.  doi: 10.1016/j.wocn.2010.12.004 Weber, A., & Cutler, A. (2004). Lexical competition in non-native spoken-word recognition. Journal of Memory and Language, 50(1), 1–25.  doi: 10.1016/S0749-596X(03)00105-0 Whitford, V., O’Driscoll, G., Pack, C., Joober, R., Malla, A., & Titone, D. (2013). Reading impairments in schizophrenia relate to individual differences in phonological processing and oculomotor control: Evidence from a gaze-contingent moving window paradigm. Journal of Experimental Psychology-General, 142(1), 57–75.  doi: 10.1037/a0028062 Whitford, V., Pivneva, I., & Titone, D. (2016). Eye movement methods to investigate bilingual reading. In R. Heredia, J. Altarriba, & A. Cieślicka (Eds.), Methods in bilingual reading comprehension research. Cambridge, UK: Cambridge University Press. Whitford, V., & Titone, D. (2012). Second-language experience modulates first- and second-­ language word frequency effects: Evidence from eye movement measures of natural paragraph reading. Psychonomic Bulletin & Review, 19(1), 73–80. Whitford, V., & Titone, D. (2015). Second-language experience modulates eye movements during first- and second-language sentence reading: Evidence from a gaze-contingent moving window paradigm. Journal of Experimental Psychology: Learning, Memory, and Cognition, 41, 1118–1129.  doi: 10.1037/xlm0000093 Whitford, V., & Titone, D. (2016). Eye movements and the perceptual span during first- and second-language sentence reading in bilingual older adults. Psychology and Aging, 31, 58–70. Whitford, V., & Titone, D. (accepted). The effects of word frequency and word predictability during first- and second-language paragraph reading in bilingual older and younger adults. Williams, S., & Hammarberg, B. (1998). Language switches in L3 production: Implications for a polyglot speaking model. Applied Linguistics, 19(3), 295–333.  doi: 10.1093/applin/19.3.295 Wrembel, M. (2010). L2-accented speech in L3 production. International Journal of Multilingualism, 7(1), 75–90.  doi: 10.1080/14790710902972263 Wrembel, M. (2012). Foreign accentedness in third language acquisition: The case of L3 English. In J. Cabrelli Amaro, S. Flynn, & J. Rothman (Eds.), Third language acquisition in adulthood (pp. 281–310). Amsterdam, The Netherlands: John Benjamins.  doi: 10.1075/sibil.46.16wre Zhang, H., Kang, C., Wu, Y., Ma, F., & Guo, T. (2015). Improving proactive control with training on language switching in bilinguals. Neuroreport, 26(6), 354–359.

doi: 10.1097/WNR.0000000000000353

chapter 2

Listening with your cohort Do bilingual toddlers co-activate cohorts from both languages when hearing words in one language alone? Susan C. Bobb1,2, Laila Y. Drummond Nauck2, Nicole Altvater-Mackensen3,2, Katie Von Holzen4,2 & Nivedita Mani2 1Gordon

College, United States / 2University of Göttingen, Germany / Planck Institute for Human Cognitive and Brain Sciences, Germany / 4Université Paris Descartes, France 3Max

Bilingual children, like bilingual adults, co-activate both languages during word recognition and production. But what is the extent of this co-activation? In the present study, we asked whether or not bilingual preschool children activate a shared phonological cohort across languages when hearing words only in their L1. We tested German-English children on a cross-modal priming paradigm. To ensure co-activation of languages, children first heard a short code-switch story. Compared to a monolingual control group, bilingual children in Experiment 1 showed only partial sensitivity to the L1 cohort. Bilingual children who did not hear the code-switch story (Experiment 2) showed priming effects identical to the monolinguals in Experiment 1. Results indicate that under single-language contexts, German-English bilingual preschoolers do not activate the non-target language cohort during word recognition but instead restrict cohort activation to the language of input. In contrast, presentation of the non-target language in the code-switch story appears to shift cohort activation and increase L2 activation, suggesting a highly flexible language system that is in tune to the broader linguistic context. We consider mechanisms of bilingual language control that may enable bilingual toddlers to limit cross-language phonological activation.

1.  Introduction One of the most compelling areas of inquiry in bilingual research has been the question of whether lexical access is fundamentally language-selective or not. In other

doi 10.1075/bpa.2.03bob © 2016 John Benjamins Publishing Company

 Susan C. Bobb et al.

words, can the unintended language be turned off in order to enable the bilingual to function as a monolingual? Research from multiple modalities has now provided evidence for language non-selectivity, suggesting that both of a bilingual’s languages are active to some extent at all times (e.g., visual: Dijkstra & Van Heuven, 1998; auditory: Marian & Spivey, 2003; Weber & Cutler, 2004; production: Hermans, Bongaerts, De Bot, & Schreuder, 1998; sign: Morford, Wilkinson, Villwock, Piñar, & Kroll, 2011). Under several accounts of language control, language co-activation leads to lexical competition between languages that subsequently needs to be resolved for the bilingual to fluently use one language alone. It is this conflict resolution that is thought to be at the root of an observed restructuring of the mind and brain that has enduring cognitive consequences for the bilingual (for a recent discussion see Kroll & Bialystok, 2013). Several chapters in this volume address these broader cognitive implications of cross-language activation. In the present chapter, we take a closer look at the organization of the developing bilingual mental lexicon with an eye towards examining the scope of language co-activation and language control and how a bilingual might limit the degree of activity of the non-target language. In light of the multitude of studies on adult bilingual language processing, only surprisingly few studies have investigated cross-language activation in bilingual children. Two recent studies show that bilingual children, similar to bilingual adults, co-activate both languages during word production and auditory word recognition (Poarch & Van Hell, 2012; Von Holzen & Mani, 2012). In Von Holzen and Mani (2012), GermanEnglish bilingual children’s recognition of L1 words was enhanced when preceded by an L2 word that overlapped in phonology [slide – Kleid “dress”, where slide and Kleid rhyme]. This facilitation effect might be caused by pure phonological overlap in the sounds of the L1 and L2 labels (i.e., the final phonemes of slide and Kleid overlap) and does not necessarily require lexical activation in both languages. Crucially, recognition was impaired when the rhyme relationship existed through translation [leg – Stein “stone”, where the German translation of leg is Bein, which rhymes with Stein]. This interference effect implies lexical activation of both L1 and L2 words and could only have been observed in this second condition if leg activated its translation. The results of Von Holzen and Mani (2012) provide strong support for lexical-level language coactivation in bilingual children. The evidence suggests that co-activation is not purely phonological, activating only sound information, but is also lexical, activating the mental representation of words. But what is the extent of this lexical co-activation? How far does activation spread? To address this question, we look at language cohorts, or words that start with the same consonant, both within and across languages. For instance, from Von Holzen and Mani (2012) we know that at the lexical level, a word in one language will activate its translation equivalent in the other language (see English leg – German Bein above). In the current study, we are interested in whether a word in one language also activates a phonologically



Chapter 2.  Listening with your cohort 

similar word (e.g., leg – leaf), and perhaps an entire cohort of similar overlapping words (e.g., leg – leaf – lake – lemon – lettuce) both within and across languages. This question is by no means trivial, particularly with respect to language and cognitive control in bilinguals. If the number of activated words across languages extends beyond translations to encompass a broad network of words, the potential competition within and across languages increases significantly, as does the work the bilingual child needs to do to resolve cross-language competition (e.g., Marian & Spivey, 2003) in order to fluently process the intended language. While recent accounts have pointed out that the complexity of the bilingual experience that leads to observed bilingual advantages in executive function cannot be reduced to competition resolution alone (e.g., Kroll & Bialystok, 2013), the “brain exercise” involved in negotiating the demands of two active languages certainly contributes to a constellation of factors, including language exposure, executive control, and metalinguistic ability, that shape the bilingual brain. In the present study, we focus on one particular lexical characteristic, namely the phonological onset of words, to examine the scope of bilingual lexical access and its consequences for auditory word recognition in bilingual children. According to onset-based models such as the Cohort Model, the organization of words centers on initial phonological overlap between lexical entries (Marslen-­ Wilson & Welsh, 1978). For example, words such as dog and doll both share the same phonological onset [d] and are therefore predicted to share a closer phonological relationship with each other than with a word such as table. Indeed, evidence from the adult literature suggests that phonological similarity is one of the organizing principles in the mental lexicon. Particularly relevant to the present study is the finding that adults can be slower to identify a word (e.g., jail) when previously exposed to a phonologically similar word (e.g., jar) compared to an unrelated word (e.g., phone) (Goldinger, Luce, Pisoni, & Marcario, 1992; Radeau, Morais, & Segui, 1995; Slowiaczek & Hamburger, 1992). Recent research finds that young first language (L1) learners, similar to adults, are sensitive to a word’s phonological cohort. Mani and Plunkett (2011) found that prior activation of phonologically related words influences 24-month-olds’ word recognition. Toddlers saw an image of a concrete and familiar object, which was followed by two images presented side-by-side. One of the images (i.e., the target) was labeled, and the proportion of looks to the target was compared to looks to the distracter image. The prime image and target word/image were either related phonologically or unrelated. Compared to unrelated prime-target pairs (e.g., bowl-cow), target identification was impaired following a phonologically related prime (e.g., bowl-bike). To investigate the locus of this interference effect, Mani and Plunkett further manipulated the target’s cohort size, that is the number of words in a typical child’s vocabulary that start with the same phonological onset. In English, [b] is a large cohort, with many words in a toddler’s vocabulary starting with [b] (e.g., ball, book, baby, bib, bottle, etc.). In contrast, [p] is a small cohort (e.g., potato, pot, puppy). They found

 Susan C. Bobb et al.

that by 24 months of age, monolingual toddlers showed impaired word recognition for large target cohorts compared to small target cohorts (Mani & Plunkett, 2011; but see Mani & Plunkett, 2010). To explain the effect, Mani and Plunkett argue that large cohorts inherently have more potential lexical candidates with the same onset that compete for selection than small cohorts, making it more difficult to select the target when it comes from a large cohort than a small cohort. This leads to impaired target recognition as indexed by reduced target looking. An important caveat to this finding is that Mani and Plunkett (2010) used an identical paradigm with 18-month-olds and found facilitation and not interference effects, suggesting that a critical vocabulary size has to be built before cohort activation influences lexical access. We will discuss this idea more in the discussion. The results of Mani and Plunkett (2011) strongly suggests that, just as in adults, phonological properties are used as a central basis for organizing the emerging lexicon of words once toddlers have reached a larger vocabulary size. But how would such an organizing principle work for bilingual children? For bilinguals, does hearing a word in the L1 activate a shared cohort in the second language (L2)? It is possible that activation stops once it has reached a single related word. But activation may cascade beyond a single word to the rest of the phonological cohort. Similar to results by Mani and Plunkett (2010), do we see developmental changes where toddlers need to reach a critical L2 vocabulary size before being influenced by cohort sizes? Answers to these questions can provide us with fundamental information on the structure of the bilingual lexicon. At the same time, this line of inquiry serves as a tool to investigate the nature of cross-language lexical activation: above and beyond asking whether single words or translation equivalents are accessed in both lexicons (e.g., Von Holzen & Mani, 2012), we can now test the scope of relative language activation.

2.  The Present Study In the present study, we used an adaptation of the inter-modal preferential looking (IPL) paradigm (Golinkoff et al., 1987) that allows for target priming (e.g., Styles, Arias-Trejo, & Plunkett, 2008). German-English bilingual toddlers heard an auditory prime word in their L1 German followed by a pair of images (target, distracter) and then heard the target label in their L1 German. Prime and target label either shared phonological onset (Bild “picture” – Baum “tree”) or not (Kind “child” – Baum “tree”). Images remained on-screen during which we assessed the proportion of time children looked to target versus distracter image. The method allowed us to manipulate cohort relationships between prime and target and, similar to some versions of the visual world paradigm (e.g., Huettig & McQueen, 2007), to gather data on looking behavior without explicit verbal instructions to children (e.g., “Look! A ball!”).



Chapter 2.  Listening with your cohort 

We asked whether or not bilingual toddlers would be sensitive to cohort sizes across both languages when hearing words only in their L1 German. To answer this question, we manipulated cohort size within and across languages. We tested toddlers on four cohorts of varying sizes: [g], [k], [b], and [p].

2.1  Hypothesis Based on Mani and Plunkett (2011), we predicted an interference effect for large target cohorts but a facilitation effect for small target cohorts (see above and discussion). Given that the experimental language was the L1 German, if only the cohort of the language of testing is activated during word recognition, then results should pattern according to German cohort sizes, with interference effects for the two large cohorts ([b], and [k]) and facilitation effects for the two small cohorts ([g] and [p]). If, however, cohorts are co-activated across languages, then the combined cohort size should matter. A combined cohort size would constitute a crucial difference for the [p] category. While [p] is a small German cohort, it forms a large combined German-English cohort. If only the cohort of the language of testing (L1) is activated during word recognition, [p] should pattern as a small cohort. If the combined cohort should matter, then [p] should pattern as a large cohort.

3.  Experiment 1 3.1  Method 3.1.1  Participants We tested a total of 43 monolingual and bilingual children for this experiment. We included 22 German-English bilingual toddlers (13 girls and 9 boys) from a local bilingual preschool in Göttingen, Germany, in our final analyses. Four additional children were tested but excluded on the basis of the recognition test (see below). Toddlers ranged in age from 21 to 61 months (M = 41.72 months). While this age range is larger than in monolingual IPL studies, it allowed us to test toddlers with a similar daily bilingual language exposure, a factor which was particularly important for the set of questions under investigation in the current study (see Von Holzen & Mani, 2012, for a similar approach). Children spent six to eight hours daily, five days a week, at the preschool and received equal amounts of instruction in English and German, with each classroom having a designated English-speaking teacher and a designated German-speaking teacher (modeled after the one-parent-one-language approach; Bain & Yu, 1980). We also tested a group of age- and gender-matched monolingual German toddlers to confirm the cohort effects that had previously been observed in English (Mani & Plunkett, 2011) for German. We included 21 children (11 girls and 10 boys) in the final

 Susan C. Bobb et al.

analyses; an additional child was tested but excluded due to external distractions during testing. Monolingual children came from the same wider German community as the bilingual group. Toddlers ranged in age from 21 to 61 months (M = 42.20 months). All children came from German-speaking homes and had no exposure to a second language. Children received a small book as a token of appreciation for participating in the study.

3.1.2  Stimuli Four cohorts were chosen based on word counts assessing the cohort size of words in two standardized German and English vocabulary assessment lists, the German FRAKIS (Fragebogen zur Frühkindlichen Sprachentwicklung; Szagun, Stumper, & Schramm, 2009) and the English MacArthur-Bates CDI (The MacArthur Communicative Development Inventories, Fenson et al., 1993). We followed the logic of Mani and Plunkett (2008) and assessed cohorts as small if they were half the size or smaller than the largest cohort [b]. Based on cohort sizes in English and German, [b] is one of the largest cohorts. The cohorts [g] and [k] fall somewhere in between, both in English and German. The [p] cohort, however, is a large cohort in English (though not as large as [b]), but a small cohort in German. See Figure 1 for an overview of the cohort sizes across languages. 120

English German

Number of words

100 80 60 40 20 0 p

g

k

b

Cohort

Figure 1.  Cohort sizes (number of words in the cohort) for English and German based on word counts in the FRAKIS (Szagun, Stumper, & Schramm, 2009) and the MacArthur-Bates CDI (Fenson et al., 1993).

In order to estimate bilingual children’s L2 (English) knowledge of words in the specific cohorts under investigation, we designed a recognition study in which children were tested on English words from each of the four cohorts. This estimate was important



Chapter 2.  Listening with your cohort 

because if children do not show recognition for words in the cohort, we cannot be sure whether the effects in the priming study are due to cohort effects or the children simply not knowing the words. The test was administered after the priming study. For both the priming and the recognition study, we consulted the FRAKIS (­Fragebogen zur Frühkindlichen Sprachentwicklung; Szagun, Stumper, & Schramm, 2009) to select age-appropriate concrete and imageable words from each of the four cohorts (b, p, g, k). We chose the final list of stimuli in discussion with one of the teachers at the preschool who indicated the words children would be familiar with in both ­English and German. Target and distracter were either both animate or both inanimate and were of the same grammatical gender to avoid a biasing context (e.g., Bobb & Mani, 2013). Target and distracter also did not share onset either within or across languages and were not cognates. Note that because many of the words learned early in ­German are cognates in English, the auditory prime stimuli did include several cognates (Garage “garage”, Garten “garden”, Papier “paper”, Polizei “police”). Because we were chiefly concerned with activating the phonological cohort, using cognates would, if anything, strengthen the cohort associated with a given prime by co-activating both the German and English cohort. A highly proficient German-English bilingual produced auditory stimuli both in German (for the priming study) and in English (for the recognition study) in a friendly and child-directed manner. The speaker shared the same language background as the children’s teachers so as to approximate the pronunciation of words children would hear in every day life at the preschool. Auditory stimuli were recorded using Adobe Audition software at a sampling rate of 44.1 kHz and edited post recording to eliminate minor clicks using PRAAT. For each target and distracter word, we found a representative digital photograph image and superimposed it on a neutral gray background. For the priming study, German auditory primes and their English translations were matched on frequency [German: Quasthoff, 2002; English: subtlexus, ­Brysbaert & New, 2009] across the four cohort groups (ps >.1). German target labels and their E ­ nglish translations were matched to German distracter labels and their English translations on frequency (ps >.1). Each child heard the same German primes and saw the same target-distracter pairings once during the experiment. For some children, a prime or target-distracter pair appeared in the related condition and for others in the unrelated condition. Across participants, however, a given stimulus appeared with equal frequency in the related and unrelated condition. The prime, target, and distracter items and their groupings by related and unrelated trials can be found in Appendix A. For the recognition study, we chose four previously unseen target words in E ­ nglish from each cohort with a similar degree of familiarity according to the FRAKIS. Each target word was matched to a distracter word of similar frequency. Distracters did not belong to one of the four cohorts, and none of the words had previously been seen in the experiment. As in the priming study, target and distracter on a given trial were

 Susan C. Bobb et al.

either both animate or both inanimate. The target and distracter items and their cohort groupings can be found in Appendix B. Note that past research with adults has shown that it is difficult to find effects of L2 activation under L1-only testing conditions (e.g., Weber & Cutler, 2004). Indeed in Von Holzen and Mani (2012), which showed language co-activation, children heard both languages within every trial in the form of an English auditory prime word and a German auditory target word. In order to enhance language co-activation, while still maintaining an L1-German only testing environment, we (1) tested participants during the school day in their bilingual environment while speaking English to them, and (2) we presented children with a code-switching story before the experiment proper. Children saw pictures on a computer screen and heard a story narrated by the native German speaker who produced the auditory stimuli. For the code-switch story, we designed sentences following the Matrix Language (ML) theory (Myers-Scotton, 2001). According to the ML theory, one of the languages provides the syntactic frame while words from the other language are slotted into this frame, creating sentences that use both languages but are nevertheless grammatical (e.g., Es ist her birthday “It is her birthday”). To avoid a language bias, sentences alternatingly used German or English as the matrix frame. None of the words used in the story occurred later in the experiment, neither the actual word nor its translation. One of the authors (L. Drummond Nauck) drew child-appropriate pictures to accompany the story. The complete story is available in Appendix C.

3.1.3  Procedure To encourage toddlers to be in a bilingual language mode (Grosjean, 1985), bilingual toddlers were tested on-site at the daycare during the course of their school day. The experimenter only spoke English to the children, as requested by the preschool teachers. Testing took place in a separate room in the preschool, where each child was tested individually. The child sat at a small desk facing a computer monitor placed 60 cm away at eye-level. Target and distractor pictures (17.5 × 13 cm each) appeared ­side-by-side on the screen. Two loud speakers, one on each side of the computer monitor, presented auditory target words. A video camera mounted centrally above the computer screen recorded the experimental session. Age-matched monolingual German toddlers were tested individually in a laboratory setting. Because the monolingual children did not know any English, the experimenter only spoke ­German with the children. Each child sat either alone or on the lap of a caregiver 100 cm away from a large TV screen (92 × 50 cm). Target and distractor pictures (27 × 20 cm each) appeared side-by-side on the screen. Two loud speakers, one on each side of the TV screen, presented auditory target words. Two video cameras mounted centrally above the TV screen recorded the experimental session.



Chapter 2.  Listening with your cohort 

The testing session was similar for bilingual and monolingual children. It began with the code-switch story. The code-switch story lasted three minutes. Monolingual children were exposed to the code-switch story in order to keep the experimental setup comparable across language groups. Following the code-switch story, we tested children on the priming test. For the priming test, each child was presented with 16 trials, 2 phonologically related and 2 unrelated trials per phonological cohort. Each trial lasted 3750 ms. The trial started with a fixation sign (+) that was followed by the presentation of the auditory prime after 100 ms. At 1700 ms into the trial, the fixation sign disappeared and two pictures appeared side-by-side followed immediately by the auditory target label at 1750 ms. Pictures remained on the screen until the end of the trial. Across participants, target and distracter pictures appeared equally often on the left or right side. After the priming test, only bilingual children completed an additional recognition test to assess their L2 English cohort knowledge and L2 proficiency. The recognition test consisted of 16 trials, 4 trials from each cohort category [b, p, g, k]. On each trial, two pictures appeared side-by-side and remained on screen the entire trial. Trials lasted 5000 ms. At 2500 ms, we presented the English auditory target label of one of the two pictures. The label divided the trial into two equal prenaming and postnaming phases of approximately 2500 ms each, which allowed us to compare looks to the target before and after the target was labeled. The target is considered correctly identified when looks to the target increase after target labeling.

3.1.4  Data coding The video data we collected from the children were coded offline frame-by-frame in 40 ms intervals by a trained coder using LOOK (Meints & Woodford, 2008). A ­second trained coder independently coded the data of 10% of the participants to assess inter-rater reliability (r = .99). For each phase of a trial, we then determined the proportion of time children looked to the target (PTL). We began analyzing a trial’s phase 240 ms after target word onset to ensure we included actual responses to the auditory stimuli (e.g., Mani & Plunkett, 2010; Swingley & Aslin, 2002). For the priming study, we were interested in the PTL after the onset of the target label to assess the impact of the prime on subsequent target recognition and therefore examined target looking behavior from 1990 to 3750 ms after the onset of the target label. For the recognition study, the phases of interest were pre- and post-target label onset in order to assess whether children’s target looking increased after hearing the label, which would indicate recognition of the target. The prenaming phase ran from 240 to 2500 ms and the postnaming phase from 2740 to 5000 ms. Each child’s looking times in each phase were then aggregated by condition creating a mean PTL per condition.

 Susan C. Bobb et al.

3.2.  Results 3.2.1  Bilingual recognition study In order to evaluate children’s knowledge of English words in the cohorts [b], [p], [g], and [k], we tested to see whether there was an overall recognition effect for the cohort words, that is whether children’s looks to target increased after hearing the English target label. Looking to target more post label indicates successful recognition and identification of the English target word. A paired samples t-test indicated that there was a general recognition effect, with more looks to target post-label compared to prelabel (t(25) = 3.216, p = .004). Looking at individual participants, however, 4 children failed to provide any data for one or more of the cohorts, suggesting limited knowledge of the given cohort and potentially low English vocabulary fluency. These children were excluded from subsequent analyses because we could not be sure that their L2 cohort knowledge and L2 proficiency was adequate enough to show priming effects (see Arias-Trejo & Plunkett, 2009, for a similar approach). 3.2.2  Bilingual priming study A 4 (cohort) × 2 (relatedness) repeated-measures analysis of variance (ANOVA) showed a marginal effect of cohort [F(3,63) = 2.66, p = .056, ηp2 = .112] and no main effect of relatedness [F(1,21) = 1.06, p = .314, ηp2 = .048]. Critically, there was a significant interaction between cohort and relatedness [F(3,63) = 3.13, p = .032, ηp2 = .13]. To assess proportion of looks to target in related versus unrelated trials for each cohort, planned comparisons using paired-sample t-tests investigated whether pairing a prime and target of the same cohort (“related” trials) facilitated or interfered with target looking compared to pairing a prime and target that did not share the same cohort (“unrelated” trials). Recall that large cohorts were predicted to show interference effects, while small cohorts were predicted to show facilitation effects. In line with a large cohort, children showed an interference effect for the [b] cohort, orienting significantly more to the target on unrelated than related trials [t(21) = –2.55, p = .019, d = –0.68]. None of the other comparisons were significant ([p]: t(21) = –.41, p = .682, d = –0.07; [g]: t(21) = –.187, p = .85, d = –0.03; [k]: t(21) = 1.76, p = .093, d = 0.4), suggesting that the prime did not substantially modulate target word recognition for words from these cohorts. Table 1 presents the priming effect (PTL related trials minus PTL unrelated trials) for each of the four cohorts that were tested. 3.2.3  Monolingual priming study A 4 (cohort) × 2 (relatedness) repeated-measures analysis of variance (ANOVA) showed no main effects of cohort or relatedness (Fs < 2). There was a significant interaction between cohort and relatedness [F(3,57) = 3.32, p = .026, ηp2 = .15]. In line



Chapter 2.  Listening with your cohort 

with a large cohort, children showed an interference effect for the [b] cohort, orienting significantly more to the target on unrelated than related trials [t(19) = –2.12, p = .048, d = –0.57]. In line with a small cohort, children showed a facilitation effect for the [p] cohort ([p]: t(19) = 2.05, p = .055, d = 0.51). None of the other comparisons were significant ([g]: t(19) = 1.56, p = .13, d = 0.53; [k]: t(19) = .462, p = .65, d = 0.16). Table 1 shows the means for related and unrelated trials in each condition. 0.25

Bilingual Code-Switch (Exp 1) Monolingual (Exp 1)

0.2

Bilingual No Code-Switch (Exp 2)

Size of Priming Effect

0.15 0.1 0.05 0 –0.05

b

k

g

p

–0.1 –0.15 –0.2 Cohort

Figure 2.  Size of Priming Effect (PTL related minus PTL unrelated) for each cohort across ­experiments. Positive differences indicate facilitation effects, with more looks to target in related than unrelated trials. Error bars indicate standard error of the mean.

3.3  Discussion of Experiment 1 The results of Experiment 1 show priming effects in the expected direction for ­German monolingual children for the largest cohort [b] and the smallest cohort [p]. ­Specifically, children showed an interference effect for [b] and a facilitation effect for [p]. The cohorts [g] and [k] did not show significant effects, most likely because their size had either exceeded or not yet met critical sizes for small/large cohorts (see Mayor  & ­Plunkett, 2014, for evidence that the number of known words influences priming effects). Their mean trends, however, were in the expected direction, with [k] patterning as a large cohort and [g] patterning as a small cohort. In contrast, bilingual children only showed a priming effect for the [b] category, which is a large cohort in both German and English. The [p] category, which is a small cohort in German but a large cohort in English and a large cohort when combined across languages, showed neither a facilitation nor an interference effect, similar to the [g] and [k] cohorts. The

 Susan C. Bobb et al.

results suggest that the [p] cohort size had exceeded the size of a small cohort but not yet reached the size of a large cohort, indicating the possibility that the magnitude of the cohort may have been in flux, either because the code-switch story had increased activation of the English lexicon to some degree or because children’s English vocabulary was still growing. Evidence from the adult bilingual literature supports the idea that bilinguals adjust activation levels of each language in response to the language environment. Elston-Güttler and colleagues (Elston-Güttler, Gunter, & Kotz, 2005) presented nativeGerman speaking participants with a silent movie that was narrated either in English or German. Following the movie, participants completed an L2-only semantic priming experiment in English in which they read sentences followed by words and had to decide whether the word was a real word or not e.g., “The woman gave her friend an expensive item” “poison”. In the critical condition, the final word of the sentence was a cross-lingual homograph (e.g., “The woman gave her friend an expensive gift” where gift is the German translation equivalent of “poison”) followed by English words that were the translation of the German meaning (poison) of the target homograph. Previous work had shown that interlingual homographs such as gift prime their translation when they are presented as isolated words (Elston-Güttler, 2000). Priming effects disappear, however, when homographs are presented in sentence contexts, presumably because the sentence context constrains activation of the unintended language. ElstonGüttler and colleagues found, in contrast to previous work, that participants were faster to respond to related targets in sentence contexts, but only if they first heard the German narration to the movie. These results suggest that the semantic context, coupled with task demands, modulates the relative activation of both languages. In effect, previously hearing German increased the relevance of German to the Englishonly experiment, leading to increased activation of the homograph’s German meaning (see also Elston-Güttler & Gunter, 2009, but see counter-evidence with single word environments in Paulmann, Elston-Güttler, Gunter, & Kotz, 2006). While the direction of the language testing in Elston-Güttler et al. (2005) was different from ours (testing was done in the L2), these results strongly indicate that the testing environment is crucial to the relative activation of a bilingual’s languages. To test the possibility that the code-switch story had changed cohort activation levels across languages, we tested bilingual children who completed the same priming study, but without first hearing the code-switch story. If the code-switch story lead to increased English cohort activation in Experiment 1, then not presenting the code-switch story should allow bilingual children in Experiment 2 to zoom into the German language of the experiment. If this is the case, we predicted that bilingual children should pattern like the monolingual German children in Experiment 1, showing interference effects for the [b] cohort and facilitation effects for the [p] cohorts.



Chapter 2.  Listening with your cohort 

4.  Experiment 2 4.1  Method 4.1.1  Participants We included 15 German-English bilingual toddlers (6 girls and 9 boys) in our final analyses. Children were from the same local bilingual preschool in Göttingen, ­Germany from which participants were recruited in Experiment 1. Toddlers ranged in age from 27 to 46 months (M = 37.73 months, SD = 5.27, Mdn = 38.00). An additional child was tested but excluded because of talking to the experimenter during testing. As in Experiment 1, toddlers received equal amounts of daily instruction in English and German, with each classroom having a designated English-speaking teacher and a designated German-speaking teacher.

4.1.2  Stimuli and Experimental Design. The same stimuli and experimental design were used as in Experiment 1. This group of participants only completed the priming test, without first hearing the code-switch story. The experimenter again only spoke English to the children. Following the priming test, children then completed additional tests for another experiment that is not reported here. 4.2  Results A 4 (cohort) × 2 (relatedness) repeated-measures analysis of variance (ANOVA) showed a significant main effect of cohort [F(3,42) = 6.38, p = .001, ηp2 = .313] but no main effect of relatedness (F < 1). Importantly, the interaction between cohort and relatedness was significant [F(3,42) = 5.92, p = .002, ηp2 = .297]. As in previous comparisons, planned paired-sample t-tests investigated whether pairing a prime and target of the same cohort (“related” trials) facilitated or interfered with target looking compared to when prime and target did not share the same cohort (“unrelated” trials). In line with a large cohort, children oriented significantly more to the target on unrelated than related trials for the [b] cohort [t(14) = –3.07, p = .008, d = –1.04]. Critically, for the [p] cohort, children oriented more to the target on related than unrelated trials [t(14) = 2.52, p = .025, d = 0.98]. This facilitation effect is in line with a small cohort and therefore activation of the L1 German cohort alone. While not significant, the [g] category means are in line with a small cohort [t(14) = 1.15, p = .267, d = 0.47], while [k] was in line with a large cohort [t(14) = –.429, p = .674, d = –0.11] (see Table 1 for condition means).

 Susan C. Bobb et al.

Table 1.  Mean PTL scores and their SD for related and unrelated conditions in each cohort across experiments for bilinguals (bi) and monolinguals (mono) b

g

rel M

un SD

M

k

rel SD

M

un SD

rel

un SD

M

un

M

SD

Bi. 0.41 0.17 0.53 0.18 0.47 0.12 0.48 0.19 (Exp 1)

0.5

0.15 0.44 0.15 0.54 0.17 0.55 0.13

0.4

M

rel

SD

Mono. 0.42 0.17 0.52 0.16 0.48 0.14 (Exp 1)

M

p

SD

M

SD

0.16 0.43 0.14 0.41 0.11 0.52 0.12 0.44 0.17

Bi. 0.42 0.12 0.55 0.11 0.57 0.19 0.49 0.15 0.58 0.14 (Exp 2)

0.6

0.14 0.69 0.18 0.53 0.16

4.3  Discussion of experiment 2 When not presented with a code-switch story before the experiment proper, ­German-English bilingual children patterned similar to the monolingual Germanspeaking children tested in Experiment 1. The results suggest that rather than combining cohorts from both languages, children restricted cohort activation to the language of input, the L1 German, and did not activate the phonological cohorts of the non-target language, the L2 English, during word recognition. These results qualify previous findings on language co-activation, suggesting there are limits to the extent of language co-activation and that perhaps only closely related words, but not entire cohorts, are co-activated. Importantly, children were tested in a bilingual preschool environment, which should have placed them in a bilingual testing mode (Grosjean, 1988). Because the wider language environment is German, however, the L2 English cohort may be active to a weaker extent, so that testing in the L1 did not reveal cohort co-activation.

5.  General discussion Previous bilingual work has provided strong support for the idea that both languages are active even when only one language is in use. While only a few studies exist to date, recent work with bilingual children has demonstrated a similar permeability to the emerging bilingual language system. The present experiments aimed to expand this area of inquiry by addressing the extent to which both languages are activated and whether lexical co-activation can be constrained. Experiment 1 showed that monolingual German children were sensitive to the German distribution of cohorts under the current experimental conditions, which included a code-switch story before the experiment proper. Hearing a member of a large cohort (in this study, [b]) interfered with



Chapter 2.  Listening with your cohort 

target recognition of a word from the same cohort. This is expected if activating members from a large cohort increases the search field of potential lexical candidates during word recognition, making it more difficult to hone in on the appropriate word (e.g., Luce & Pisoni, 1998). Hearing a member of a small cohort (in this study, [p]) facilitated subsequent target recognition of a word from the same cohort. According to the Cohort Model, this is because activation of the small cohort boosts activation of other members in the cohort, easing lexical access of the target word after hearing the prime. In contrast to the monolinguals, the bilingual children of Experiment 1 showed neither facilitation nor interference effects for the [p] cohort, suggesting that they were not sensitive to a German cohort distribution nor a combined German-English [p] cohort. This non-significant [p] result for the bilingual children of Experiment 1 suggests that compared to monolingual German children, bilinguals exposed to the code-switch story experienced a shift in their cohort distribution such that pre-lexical facilitative effects and lexical competitive effects cancelled each other out. Results from Experiment 2 confirmed that in the absence of the code-switch story, bilingual children from the same sample show cohort sensitivity analogous to German monolingual children. Taken together, the results from Experiments 1 and 2 indicate that for bilingual children from this sample, the extent of L2 activation can be constrained so that the L2 English is only minimally active during L1 German processing. Only when L2 activation is enhanced through a previous code-switch story does the overall cohort sensitivity begin to shift. In the present study, the experimenter always addressed the children in their L2 (English). It is interesting to note that not only the code-switch story but also speaking English to the bilingual children before test should in theory make it more likely that the children activate both languages (i.e., are in a bilingual mode). With the experimenter speaking English and the test being German, there is a real language switch; thus, children might tune into one language. Our results suggest that the code-switching story was more effective than the experimenter speaking (only) English, perhaps because code-switching is inherently bilingual and therefore better supports co-activation. The current results extend the limited body of research on bilingual children and language co-activation while also supporting previous findings in the adult bilingual literature (Elston-Güttler, Gunter, & Kotz, 2005). The present results suggest a flexible access to the language system that can adapt to a given linguistic context. Evidence at the sentence level with adults has shown that lexical access can be modulated at least to some extent by speaker accent and sentence context, reducing, but not eliminating, cross-language activation (e.g., Lagrou, Hartsuiker, & Duyck, 2013; see also Ju & Luce, 2004; Weber & Cutler, 2004). Our results suggest that the experimental settings similarly affect the relative lexical activation level of both languages. Under monolingual settings, the experimental environment acts to constrain activation of the non-target language. When the experimental settings emphasize the relevance of both languages,

 Susan C. Bobb et al.

we see a pattern of results suggestive of broader language co-activation that encompasses the phonological cohorts of both languages. This possibility is in-line with recent proposals, such as the Adaptive Control Hypothesis (ACH, Green & Abutalebi, 2013), which emphasizes interactional context with differing demands on control processes. While the ACH focuses on predictions for language production, it is a hypothesis based on conversational interaction and thus suitable to explain the present findings. Bilinguals increase cognitive control depending on their linguistic context so that control processes adapt to situational demands. Under this proposal, Green and Abutalebi contrast three interactional levels: (1) single-language, (2) dual-language, and (3) dense code-switching. Each context of language use increases demands on a subset of 7 identified control processes: goal maintenance, interference control, cue detection, selective response inhibition, task disengagement, task engagement, and opportunistic planning. A dual-language situation, for instance, requires recurrent interference control and goal maintenance that subsequently lead to adaptive changes on a behavioral and neural level. Our data also provide evidence for important constraints to the extent of lexical co-activation. With respect to the scope of activation, previous research with bilingual children has shown co-activation to the lexical level, encompassing cognates and rhyme words (Poarch & Van Hell, 2012; Von Holzen & Mani, 2012). Our results suggest that unless influenced by environmental conditions, bilingual children’s activation of both languages does not extend to entire cohorts. That is, hearing a word in one language does not necessarily co-activate words from the same cohort in the other language, at least when the dominant (L1) is in use. This finding is important, because limiting the number of co-activated words would decrease sources of cross-language competition that would subsequently need to be resolved. For bilinguals operating under single language settings, the cognitive demands of negotiating co-activated languages could thus be greatly reduced. An open question is whether the linguistic context constrains non-target activation so that activation never spreads to the cross-language cohort in the first place or whether cross-language cohort activation is subsequently reactively controlled via executive control networks. On the flip side, namely under situations that enhance activation of both languages such as our code-switch story, co-activation of more words in both languages would increase potential sources of language competition, placing higher demands on cognitive resources to resolve competition. An alternative explanation for why cross-language cohort effects may be tenuous, particularly without a testing context that encourages co-activation, is because the bilingual children had not developed a large enough English vocabulary. Although we used the same stimuli in Experiments 1 and 2, and all items were designated as typically known by an early age (18-month-olds according to the FRAKIS), it is possible that children had not yet reached a critical English vocabulary size to show competitive effects of the English cohort. Monolingual work supports the idea that children need to have reached a critical vocabulary size for lexical interference and



Chapter 2.  Listening with your cohort 

cohort effects to emerge. Mani and Plunkett (2010), using a comparable experimental paradigm to Mani and Plunkett (2011), showed facilitation rather than interference effects and no cohort effects for 18-month-old infants. They point to converging evidence from studies on semantic priming to argue for a developmental change in the emerging lexical structure in which organization moves from lexical islands to a lexico-semantic network as a child’s vocabulary increases. In smaller vocabularies, there may not be enough words to activate a strong enough competitive environment that would show effects of interference. Consistent with this idea, Mayor and Plunkett (2014) apply the TRACE model of word recognition (­McClelland & Elman, 1986) to provide converging evidence from computational modeling for age-related differences in infant spoken word recognition patterns. Specifically, they were able to account for effects of vocabulary growth and cohort competition on phoneme perception. With increasing lexicon size, particularly at the age of the vocabulary spurt (18–21 months), the pool of potential competitors and by extension cohort competitors increases dramatically (see also evidence from bilingual language production in children by Poarch & Van Hell, 2012). In future work, it will therefore be crucial to test older children with longer bilingual immersion and larger lexicons in both languages. While previous studies have identified cross-language activation in children, based on the present results, this co-activation does not appear to extend to entire language cohorts, at least not under all testing environments. By the time activation spreads to the non-target language, the intended language may have been selected, in effect shutting down further activation of the non-target language cohort. Such a possibility is most parsimonious with bilingual models such as BLINCS (Shook & Marian, 2013), which can simulate bilingual speech comprehension and captures the co-activation of cross-linguistic stimuli such as words that share phonological onset. When the model was presented with phonological stimuli, words from both languages became active, such as the English word road and the Spanish word ropa (meaning clothes). Within language, words with denser neighborhoods (i.e., words that differ from each other by only one letter like pill and pull) were activated more slowly and faced greater competition than words with sparser neighborhoods. These previous findings suggest that BLINCS could account for the present results as well. Future simulations will need to test cohort sizes across languages directly. It should be noted that the present interpretation moves from an assumption that the underlying language system is fundamentally permeable (cf. e.g., Kroll, Bobb, & Wodniecka, 2006), but can under certain circumstances limit the extent of such cross-language activation. Other models such as the bilingual model of lexical access (BIMOLA; ­Grosjean, 1988) do not share this assumption. Instead, two separate but interconnected language networks make up the bilingual lexicon. The present results cannot adjudicate between these alternatives but suggest, regardless of the language architecture, that language co-activation in early development can be constrained to one language cohort alone.

 Susan C. Bobb et al.

While the present study appears to have captured a shift in cohort sensitivity in response to a top-down cue (i.e., the experimental setting), future work using other methods sensitive to the time-course of processing such as event-related potentials (ERPs) may be able to document this change more concretely. Like Elston-Güttler et al., bilingual children should also be tested on their L2, which may increase the likelihood of cohort activation across languages. In this respect the present study’s approach to design was conservative, aiming to test children in their dominant language (German), which is known to make effects of language co-activation more difficult to observe (e.g., Weber & Cutler, 2004).

6.  Conclusion Our results underscore developing bilinguals’ sensitivity to their immediate discourse environment, suggesting the ability for in-the-moment adaptation and a high level of mental agility. By testing a young population of bilinguals, we show that this cognitive flexibility is evident at early stages of language development. By testing phonological cohorts, the current data also provide a further avenue for investigating the scope of lexical co-activation in bilinguals. Future work will need to consider different age groups and modalities in order to clarify the extent of language co-activation and its modulation under diverse linguistic and interactional contexts. While the implications for cognitive control discussed here have been necessarily speculative, our results point to exciting possibilities for exploring the cognitive consequences of a highly permeable and interactive language system.

References Arias-Trejo, N., & Plunkett, K. (2009). Lexical–semantic priming effects during nfancy. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1536), 3633–3647.

doi: 10.1098/rstb.2009.0146

Bain, B., & Yu, A. (1980). Cognitive consequences of raising children bilingually: “One Parent, One Language.” Canadian Journal of Psychology, 34, 304–313.  doi: 10.1037/h0081106 Bobb, S. C., & Mani, N. (2013). Categorizing with gender: Does implicit grammatical gender affect semantic processing in 24-month-old toddlers? Journal of Experimental Child Psychology, 115, 297–308.  doi: 10.1016/j.jecp.2013.02.006 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, 977–990.

doi: 10.3758/BRM.41.4.977

Dijkstra, T., & Van Heuven, W. (1998). The BIA model and bilingual word recognition. Localist Connectionist Approaches to Human Cognition, 189–225.



Chapter 2.  Listening with your cohort 

Elston-Güttler, K. E. An inquiry into cross-language lexical-conceptual relationships and their effect on L2 lexical processing, Ph.D. Dissertation, University of Cambridge, 2000. Elston-Güttler, K. E., & Gunter, T. C. (2009). Fine-tuned: Phonology and semantics affect first-to second-language zooming in. Journal of Cognitive Neuroscience, 21, 180–196.

doi: 10.1162/jocn.2009.21015

Elston-Güttler, K. E., Gunter, T. C., & Kotz, S. A. (2005). Zooming into L2: Global language context and adjustment affect processing of interlingual homographs in sentences. Cognitive Brain Research, 25, 57–70.  doi: 10.1016/j.cogbrainres.2005.04.007 Fenson, L., Dale, P. S., Reznick, J.S., Thal, D., Bates, E., Hartung, J.P., Pethick, S., & Reilly, J.S. (1993). The MacArthur Communicative Development Inventories: User’s guide and technical manual. San Diego, CA: Singular Publishing Group.  doi: 10.1037/0278-7393.18.6.1211 Goldinger, S. D., Luce, P. A., Pisoni, D. B., & Marcario, J. K. (1992). Form-based priming in spoken word recognition: The roles of competition and bias. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 1211.  doi: 10.1037/0278-7393.18.6.1211 Golinkoff, R. M., Hirsh-Pasek, K., Cauley, K. M., & Gordon, L. (1987). The eyes have it: Lexical and syntactic comprehension in a new paradigm. Journal of Child Language, 14, 23–45.

doi: 10.1017/S030500090001271X

Green, D. W., & Abutalebi, J. (2013). Language control in bilinguals: The adaptive control hypothesis. Journal of Cognitive Psychology, 25, 515–530.  doi: 10.1080/20445911.2013.796377 Grosjean, F. (1985). The bilingual as a competent but specific speaker-hearer. Multilingual and Multicultural Development, 6, 467–477.  doi: 10.1080/01434632.1985.9994221 Grosjean, F. (1988). Exploring the recognition of guest words in bilingual speech. Language and Cognitive Processes, 3, 233–274.  doi: 10.1080/01690968808402089 Hermans, D., Bongaerts, T., De Bot, K., & Schreuder, R. (1998). Producing words in a foreign language: Can speakers prevent interference from their first language?.  Bilingualism: ­Language and Cognition, 1, 213–229.  doi: 10.1017/S1366728998000364 Huettig, F., & McQueen, J. M. (2007). The tug of war between phonological, semantic and shape information in language-mediated visual search.  Journal of Memory and Language,  57, 460–482.  doi: 10.1016/j.jml.2007.02.001 Ju, M., & Luce, P. A. (2004). Falling on sensitive ears constraints on bilingual lexical activation. Psychological Science, 15, 314–318.  doi: 10.1111/j.0956-7976.2004.00675.x Kroll, J. F., & Bialystok, E. (2013). Understanding the consequences of bilingualism for language processing and cognition. Journal of Cognitive Psychology, 25, 497–514.

doi: 10.1080/20445911.2013.799170

Kroll, J. F., Bobb, S. C., & Wodniecka, Z. (2006). Language selectivity is the exception, not the rule: Arguments against a fixed locus of language selection in bilingual speech. Bilingualism: Language and Cognition, 9, 119–135.  doi: 10.1017/S1366728906002483 Lagrou, E., Hartsuiker, R. J., & Duyck, W. (2013). The influence of sentence context and accented speech on lexical access in second-language auditory word recognition.  Bilingualism: ­Language and Cognition, 16, 508–517.  doi: 10.1017/S1366728912000508 Luce, P. A., & Pisoni, D. B. (1998). Recognizing spoken words: The neighborhood activation model. Ear and Hearing, 19, 1–36.  doi: 10.1097/00003446-199802000-00001 Mani, N. & Plunkett, K. (2008). Phonological priming in infancy. Proceedings of the 30th Annual Meeting of the Cognitive Science Society, Washington, USA. Mani, N., & Plunkett, K. (2010). In the infant’s mind’s ear evidence for implicit naming in 18-month-olds. Psychological Science, 21, 908–913.  doi: 10.1177/0956797610373371

 Susan C. Bobb et al. Mani, N., & Plunkett, K. (2011). Phonological priming and cohort effects in toddlers. Cognition, 121, 196–206.  doi: 10.1016/j.cognition.2011.06.013 Marian, V., & Spivey, M. (2003). Competing activation in bilingual language processing: Withinand between-language competition. Bilingualism: Language and Cognition, 6, 97–115.

doi: 10.1017/S1366728903001068

Marslen-Wilson, W. D., & Welsh, A. (1978). Processing interactions and lexical access during word recognition in continuous speech. Cognitive Psychology, 10, 29–63.

doi: 10.1016/0010-0285(78)90018-X

Mayor, J., & Plunkett, K. (2014). Infant word recognition: Insights from TRACE simulations. Journal of Memory and Language, 71, 89–123.  doi: 10.1016/j.jml.2013.09.009 McClelland, J. L., & Elman, J. L. (1986). The TRACE model of speech perception.  Cognitive Psychology, 18, 1–86.  doi: 10.1016/0010-0285(86)90015-0 Meints, K., & Woodford, A. (2008). Lincoln Infant Lab Package 1.0: A new programme package for IPL, Preferential Listening, Habituation and Eyetracking. [www document: Computer software & manual]. 〈http://www.lincoln.ac.uk/psychology/babylab.htm〉 Morford, J. P., Wilkinson, E., Villwock, A., Piñar, P., & Kroll, J. F. (2011). When deaf signers read English: Do written words activate their sign translations?.Cognition, 118, 286–292.

doi: 10.1016/j.cognition.2010.11.006

Myers-Scotton, C. (2001). The matrix language frame model: Development and responses. Trends in Linguistics Studies and Monographs, 126, 23–58. Paulmann, S., Elston-Güttler, K. E., Gunter, T. C., & Kotz, S. A. (2006). Is bilingual lexical access influenced by language context?. NeuroReport, 17, 727–731.

doi: 10.1097/01.wnr.0000214400.88845.fa

Poarch, G. J., & Van Hell, J. G. (2012). Cross-language activation in children’s speech production: Evidence from second language learners, bilinguals, and trilinguals. Journal of Experimental Child Psychology, 111, 419–438.  doi: 10.1016/j.jecp.2011.09.008 Quasthoff, U. (2002). DeutscherWortschatz im Internet [Online database]. Leipzig, Germany: University of Leipzig. 〈http://www.wortschatz.uni-leizpig.de/〉. Radeau, M., Morais, J., & Segui, J. (1995). Phonological priming between monosyllabic spoken words. Journal of Experimental Psychology: Human Perception and Performance, 21, 1297–1311.  doi: 10.1037/0096-1523.21.6.1297 Shook, A., & Marian, V. (2013). The bilingual language interaction network for comprehension of speech. Bilingualism: Language and Cognition, 16, 304–324.

doi: 10.1017/S1366728912000466

Slowiaczek, L. M., & Hamburger, M. (1992). Prelexical facilitation and lexical interference in auditory word recognition.  Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 1239. Styles, S. J., Arias-Trejo, N., & Plunkett, K. (2008). Priming and lexical interference in infancy, In Love, B. C., McRae, K. & Sloutsky V.M., (Eds.), Proceedings of the 30th Annual Conference of the Cognitive Science Society, (pp. 651–656). Austin, TX: Cognitive Science Society, 651–656. Swingley, D., & Aslin, R. N. (2002). Lexical neighborhoods and the word-form representations of 14-month-olds. Psychological Science, 13, 480–484. Szagun, G., Stumper, B., & Schramm, S. A. (2009). Fragebogen zur frühkindlichen Sprachentwicklung (FRAKIS) und FRAKIS-K (Kurzform). Pearson. Von Holzen, K. & Mani, N. (2012). Language nonselective lexical access in bilingual toddlers, Journal of Experimental Child Psychology, 113, 569–586.  doi: 10.1016/j.jecp.2012.08.001 Weber, A., & Cutler, A. (2004). Lexical competition in non-native spoken-word recognition. Journal of Memory and Language, 50, 1–25.  doi: 10.1016/S0749-596X(03)00105-0



Appendix A

Appendix B

Chapter 2.  Listening with your cohort 

 Susan C. Bobb et al.

Appendix C Das ist Lynn. Sie ist very friendly und always has gute Laune. Sie lives in a blue house am Rande der Stadt. Today ist ein wunderschöner Tag. Lynn steht früh auf. She is excited und freut sich sehr.

Denn today ist ein besonderer Tag. Es ist her birthday. Sie wird five years old.

Sie hatte viele Wünsche for her birthday. Lynn hat sich a real horse gewünscht. But this is leider nicht possible. Dafür haben sie at home zu wenig Platz.

Als Lynn in the morning die Kerzen auspustet, singen her parents ihr ein tolles birthday song. They tell her. dass sie nach dem Frühstück einen Ausflug machen werden. Aber sie verraten nicht, where they are going to take her.

Lynn cannot wait for her surprise weil sie sehr gespannt ist und sich so sehr freut. Sie beeilt sich with breakfast. damit sie schnell los fahren können.



Chapter 2.  Listening with your cohort  Dann ist es endlich soweit. Her parents holen the bicycles. Zusammen machen sie sich auf den Weg und they ride through a forest and pass many Wiesen und Felder. Lynn überlegt die ganze Zeit, what the surprise for her could be about. When her parents tell her, dass sie hinter der nächsten Kurve endlich an ihrem Ziel ankommen, klopft her heart vor lauter Glück very fast. Lynn rides her bike so schnell wie noch nie. Sie rast um die Kurve und steht suddenly in the center of a big Reiterhof. Überall stehen beautiful horses, auf denen man reiten darf. When her parents arrive, nehmen sie Lynn in den Arm. Sie sagen ihr, dass she may choose a horse, auf dem sie fue whole afternoon reiten kann. Lynn hüpft vor Freude in die Luft. Sie klatscht her hands, weil sie so happy ist. Es ist die tollste Geburtstagsüberraschung, she could think of! Lynn sucht sich the most beautiful Pferd von allen aus – es sieht aus exactly the way the horse of her dreams looks like. Sie reitet darauf bis es abends dunkel wird. Dann erst fahren Lynn und ihre Eltern back home. At night, als Lynn in ihrem Bett liegt, ist sie von dem exciting Ausflug ganz müde. Tonight she will surely dream of the beautiful horse, auf dem sie heute geritten ist…

chapter 3

The role of executive function in the perception of L2 speech sounds in young balanced and unbalanced dual language learners* Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez University of Houston, United States

This fMRI study investigated neural activity evoked by L2 speech syllables in brain regions associated with executive function typically recruited by bilinguals in cognitive control tasks. The main areas examined were the bilateral anterior cingulate cortex, bilateral supplementary motor area, bilateral inferior frontal gyrus, and bilateral middle frontal gyrus. Based on the degree of discrepancy between L1 and L2 proficiency scores, twenty-nine children classified as balanced (smaller discrepancy) or unbalanced (larger discrepancy) dual language learners were matched for age, socioeducational background, years of education in L2, and L2 age of acquisition. Children passively listened to L2 syllables while a muted film was presented. The results showed that unbalanced learners had increased activity in multiple frontal regions bilaterally relative to balanced learners. Balanced learners showed increased activity in a region of the right temporal lobe. The results suggest that unbalanced learners who have more difficulty learning the second language engage regions of executive function to support the perception of L2 speech sounds.

1.  Introduction The goal of the present study was to investigate the role of executive function in the perception of second language (L2) speech sounds in young balanced and unbalanced

*  We give a special thanks to Madeleine Gorges for her comments and feedback on an earlier version of this manuscript. The data collected for this study was supported by grants from the National Institutes of Health (NIH) R21HD059103-01 and 1R03HD079873-01 as well as a grant from the Institute for Biomedical Imaging Science (IBIS) for Plasticity in Speech ­Perception in Early Bilingual Children.

doi 10.1075/bpa.2.04arc © 2016 John Benjamins Publishing Company

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

dual language learners. Much research has been conducted in the area of cognitive control in bilinguals through the examination of language switching (Abutalebi et al., 2007; Bialystok, 2009; Costa & Santesteban, 2004; Garbin et al., 2011; Hernandez, Dapretto, Mazziotta, & Bookheimer, 2001; Hernandez, Martinez, & Kohnert, 2000; Luk, Green, Abutalebi, & Grady, 2011) and non-verbal task switching (Prior & Gollan, 2011; Prior & Macwhinney, 2010). However, no links have been made between the research areas of executive function and speech perception in L2 learners. To date, we do not know how prefrontal regions involved in executive function interplay with temporal regions involved in auditory processing in tasks of L2 speech perception. In today’s growing bilingual society, many dual language learners report lower scores than their monolingual counterparts in standardized tests of academic achievement (Fernandez & Nielsen, 1986). These low scores appear to be the result of difficulties in acquiring and processing the L2 (Cummins, 1984; Fernandez & Nielsen, 1986). While difficulties in L2 learning can be attributed to factors such as language family, low socioeconomic status, low parental education, etc., it is known that for some individuals part of the difficulty in learning the L2 originates from deficits in speech sound perception. In monolinguals, for example, early perceptual experiences with speech sounds have been found to lay the groundwork for later vocabulary learning (Kuhl, 2008). That is, monolingual infants who neurally entrench to the speech sounds of the native language early on are more likely to have advanced language abilities at 24-months and 30-months than monolingual infants who entrench to the native language later on (Kuhl, Conboy, Padden, ­Nelson,  & Pruitt, 2005). Moreover, in specific language skills such as reading, it has been found that the basis of a reading difficulty is an impaired ability to accurately perceive speech sounds which lead to proper phonological awareness (Breier, Fletcher, ­Denton, & Gray, 2004; Breier et al., 2003; Shaywitz et al., 2004; Stanovich, Cunningham, & Cramer, 1984). Examining the involvement of executive function in sensory-based processes of speech perception in young dual language learners will help us to understand how the early establishment of two speech systems impacts overall language development, and perhaps even predict when L2 acquisition is likely to be successful and when it is likely to present concerns. In classifying bilinguals as “balanced” or “unbalanced”, balanced bilinguals are those who have similar levels of proficiency in both languages and unbalanced bilinguals are those who have a higher level of proficiency in one of the languages (De Groot, 2011). Unbalanced bilinguals are therefore dominant in one language and moderately weak in the other language. Given that most bilinguals live in linguistic environments that emphasize one of the languages, unbalanced bilinguals tend to be the norm and balanced bilinguals tend to be more rare (Kroll & De Groot, 2005). However, even in environments where two languages are regularly used, unbalanced bilinguals typically use one of the languages more often than the other (Grosjean & Li, 2012).



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

Behavioral studies investigating different aspects of language processing have found important differences between balanced and unbalanced bilinguals. In lexical processing, for example, it has been demonstrated that balanced bilinguals name more pictures correctly in both languages if the picture-naming task provides object images with cognates across languages (e.g. dart and dardo). Unbalanced bilinguals, on the other hand, only benefit from cognates when operating in the weak language because words similar to those from the strong language provide a reminder (Gollan, Fennema-Notestine, Montoya, & Jernigan, 2007). Word identification tasks have also shown that balanced bilinguals are equally fast and accurate at identifying words from both languages when different orthotactic rules are employed, whereas unbalanced bilinguals are slower and make more errors in L2 words than L1 words (Casaponsa, Carreiras, & Dunabeitia, 2014). In general, early bilingualism in balanced individuals has been shown to enrich the linguistic interdependence of vocabulary knowledge, such that deep lexical knowledge in one of the languages improves the lexical knowledge of the other language (Schwartz, 2014). These findings suggest that balanced bilinguals have enhanced lexical access to both languages compared to unbalanced bilinguals. Similarly, in behavioral studies of attention and language switching, balanced bilinguals outperform unbalanced bilinguals. For instance, deaf bilinguals are better able to successfully deal with changing target locations and to inhibit distractors than unbalanced bilinguals (Kushalnagar, Hannay, & Hernandez, 2010). Unbalanced bilinguals appear to have more difficulty suppressing the dominant language because they must strongly inhibit it in order to allow production in the weaker language (­Grosjean & Li, 2012). Voluntary language switching in everyday situations may enable unbalanced bilinguals to complement their speech and enrich their social interactions to operate like balanced bilinguals (Gollan & Ferreira, 2009). On the whole these studies suggest that balanced bilinguals are better than unbalanced bilinguals at keeping track of each of their languages and keeping interlingual interference at bay. Very few behavioral studies have looked at the processes of speech perception in groups of balanced or unbalanced bilinguals, fewer still in groups of balanced or unbalanced young dual-language learners. For a more extensive read on the processes of speech perception in bilingual children, see Archila-Suerte, Zevin, Ramos, and ­Hernandez, 2013. One study in a related area found that unbalanced bilinguals have better phonological awareness than balanced bilinguals, potentially suggesting that being dominant in one of the languages (L1 or L2) enhances the ability to analyze phonological information in the L2 (San Francisco, Carlo, August, & Snow, 2006). Another study of perception and production found that unbalanced bilinguals produce words with Voice Onset Times (VOT’s) weighting speech cues according to the language in which they are most dominant (Hazan & Boulakia, 1993). That is, the weak language was produced using speech cues from the dominant language.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

A reanalysis of a study conducted by Archila-Suerte and colleagues (2011) examined how L2 speech sounds were perceived by Spanish-English bilinguals. The authors found that discriminating between two qualitatively different L2 phonemic categories (/ɒ/ as in “hot” and /ʌ/ as in “hut”) was more difficult for unbalanced bilinguals who were dominant in the first language (L1) than for balanced bilinguals, independent of the age of acquisition of the L2. In other words, the group of early and late unbalanced bilinguals tended to intermix the English phonemic categories of /ɒ/ and /ʌ/, whereas the group of early and late balanced bilinguals were better at discriminating and categorizing the second-language phonemes. Contrary to the results of San F ­ rancisco et al. (2006), the reanalysis of Archila-Suerte et al. (2011) suggests that balanced bilinguals attend to the speech cues that provide relevant information for the correct categorization of L2 sounds, whereas unbalanced bilinguals do not. Unbalanced bilinguals therefore appear to have more difficulty distinguishing between two speech sounds of different categories in the non-dominant language. This impaired ability may be the result of a previously established perceptual filter created by the input of first language speech sounds (Best, McRoberts, & Goodell, 2001; Flege, 2003; Pallier, Bosch, & Sebastian-Galles, 1997), which modifies the way in which speech cues are weighted in the L2 (Holt & Lotto, 2006; Iverson, Hazan, & Bannister, 2005; Ylinen et al., 2010). The discrepancy in the results between San Francisco et al. (2006) and the reanalysis of Archila-Suerte (2011) may be attributed to differences in statistical methodology. To be precise, San Francisco et al. examined phonological awareness and Archila-Suerte examined categorical perception. While these two concepts are theoretically related, the tasks employed and statistical analyses conducted can lead to different interpretations of the data. Neuroimaging studies investigating neural processing differences between balanced and unbalanced bilinguals have been primarily focused around tasks of attention and language switching. Several studies report that frontal and parietal regions of the brain, known to be involved in executive processing, are recruited for language switching tasks (Abutalebi et al., 2007; Costa & Santesteban, 2004; Hernandez et al., 2000; Luk et al., 2011). For instance, the left dorsolateral prefrontal cortex (DLPFC) in the middle frontal gyrus (MFG) has been documented in studies of picture naming with balanced bilinguals switching between languages (Hernandez, 2009; Hernandez et al., 2001). Activity in the nearby region of the left inferior frontal gyrus (IFG) has also been reported in studies of language switching (Fabbro, Skrap, & Aglioti, 2000; Price, Green, & von Studnitz, 1999; Wang, Xue, Chen, Xue, & Dong, 2007). Others have extended this research and have proposed that the prefrontal cortex is engaged in general executive processing while the anterior cingulate gyrus (ACC) and caudate nucleus are recruited more specifically for language switching (Abutalebi, 2008). An fMRI study with English monolingual and Spanish-English bilingual children found that passively listening to English syllables evoked increased activity in



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

the bilateral MFG and bilateral inferior parietal lobule (IPL) of bilingual children (9–10 years) relative to monolingual children of the same age group (Archila-Suerte, Zevin, Ramos, & Hernandez, 2013). See (Hutzler, 2014; Poldrack, 2006) for a review on reverse inference. The authors found that L2 speech sounds elicit activity in regions of the prefrontal cortex – in addition to increased activity in expected regions of the temporal lobe (i.e., STG) – in bilingual children. However, it remains unclear whether different neural resources are recruited in response to L2 speech sounds given a participant’s level of proficiency in each of the languages, or given the level of discrepancy in proficiency between the languages (i.e., balanced and unbalanced proficiency). In the present study we analyze a subset of data from the same participant pool collected for Archila-Suerte et al. (2013), to investigate the role of executive function in L2 speech perception, among subgroups of young balanced and unbalanced dual language learners.1 Switching to the weak language (typically switching from L2 to L1; known as backward switching) has been associated with stronger activation of frontal regions like the MFG and ACC (Wang et al., 2007). This effect of switching to the weaker language has been observed in phonological priming tasks. Rodriguez-Fornells (2005) designed a tacit picture-naming task where monolinguals and Spanish-German bilinguals were given the first letter of the picture to name it. In half the trials, the name of the picture began with the same letter category (consonant or vowel) in both languages (i.e., congruent trials). In the other half of the trials, the name of the picture began with a different letter category in the two languages (i.e., incongruent trials). One of the languages was designated the target language. Participants were asked to tacitly name the picture in this language and to respond with a button press if the picture’s name began with a vowel but to withhold the response if the name began with a consonant. The results showed that in incongruent trials, where the phonology of the non-target language interfered with the response, bilinguals showed increased activity in the left DLPFC and supplementary motor area (SMA). The results of this neuroimaging study parallel and corroborate the findings of behavioral investigations, wherein unbalanced bilinguals appear to require increased mental effort to process the weak language through the recruitment of brain regions involved in executive function. The role of executive function in the processing of L2 speech sounds has not been the focus of researchers thus far; instead, numerous studies have emphasized preattentive processing in speech perception (Hickok & Poeppel, 2000; Hickok & Poeppel,

.  Even though participants in the current study came from the same participant pool as Archila-Suerte et al., 2013, they are labeled differently here. While participants are labeled as “bilingual children” in Archila-Suerte, et al., 2013, they are more appropriately labeled in the present study as “dual language learners”.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

2007; Joanisse, Robertson, & Newman, 2007; Joanisse, Zevin, & McCandliss, 2007). It is widely known that listening to speech sounds engages the posterior region of the left superior temporal gyrus, STG (Binder, 2000; Celsis et al., 1999; Liebenthal, Binder, Spitzer, Possing, & Medler, 2005) and middle temporal gyrus, MTG (Démonet et al., 1992; Zatorre, Evans, Meyer, & Gjedde, 1992). It is thought that the STG is primarily involved in sensory processing of acoustic-phonetic information, while the MTG is involved in the processing of semantic associations or auditory-to-meaning mappings needed for comprehension (Hickok & Poeppel, 2000; Hickok & Poeppel, 2007). Activity in the bilateral STG has been observed in monolinguals and bilinguals passively listening to English syllables (Archila-Suerte, Zevin, & Hernandez, 2015). Moreover, it has been found that monolinguals and bilinguals who are classified as advanced learners of speech, based on a phonetic training task, have more intense and widespread activity in the bilateral STG in response to listening to novel speech sounds than nonadvanced learners, suggesting that individual ability is a significant factor in speech learning (Archila-Suerte, Bunta, & Hernandez, 2014). Given that balanced and unbalanced bilinguals differ in the processing of numerous tasks, including lexico-semantic, attention, and language switching, we want to investigate how these groups additionally differ in the processing of L2 speech sounds. More importantly, we will do this with a sample of children who are classified as balanced or unbalanced dual language learners, based on the level of discrepancy in proficiency between the languages. We hypothesize that balanced learners who are highly proficient in both languages will recruit regions of the temporal lobe bilaterally in response to listening to L2 speech sounds, similar to the results observed for the group of advanced learners in Archila-Suerte, et al., 2014. This hypothesis is supported by studies that have demonstrated that speech sounds elicit increased activity in the bilateral STG in good learners of speech sounds (Archila-Suerte et al., 2014; Diaz, Baus, Escera, Costa, & SebastianGalles, 2008; Sebastián-Gallés et al., 2012; Wong, Perrachione, & Parrish, 2007). On the other hand we hypothesize that unbalanced learners, who are less proficient in L2 than L1, will recruit regions of the frontal lobe to support the processing of speech sounds through increased activity in regions associated with attention and working memory. This hypothesis is supported by the notion that unbalanced bilinguals show increased activity in regions involved in executive function in multiple language tasks (Casaponsa et al., 2014; Gollan et al., 2007; Kushalnagar et al., 2010; Schwartz, 2014). Even though unbalanced bilinguals are considered the norm in society (Kroll & De Groot, 2005), in the present study, unbalanced proficiency is equated with a certain level of difficulty in language learning given that both groups of children are matched on age, age of acquisition (AoA), percent language use, years of education, and socioeducational background. In summary, given the findings of previous studies and in order to investigate the role that executive function plays in the perception of L2 speech sounds in young



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

balanced and unbalanced dual language learners, we hypothesize that children who are unbalanced in their knowledge of the first and second language will recruit frontal regions associated with executive function in response to listening to L2 syllables. Specifically, we expect to see activity in the MFG, ACC, IFG, and the SMA bilaterally. On the other hand, we hypothesize that children who are relatively balanced in their knowledge of both languages will recruit regions involved in speech and auditory lexico-semantic processing, including the STG and MTG. To this end, region of interest (ROI) analyses were conducted in the abovementioned regions to compare brain activity elicited by L2 speech sounds between groups of young balanced and unbalanced dual language learners.

2.  Present study 2.1  Participants Twenty-nine typically developing children participated in this study (13 boys, 16 girls). All children learned Spanish as their native language and English as their L2. The majority of the children (75.9%) were born in the United States; therefore, Spanish was primarily learned at home and English was learned at school. On average, children began to acquire the L2 at around 4 years of age (SD = 0.31) and had received 4 years of education in the L2 (SD = 0.32) at the time of testing. Children in this study were between 6 and 10 years of age. All parents and children consented to the criteria stipulated to participate in the research study.

2.2  Procedure Children’s oral and receptive abilities in English and Spanish were assessed using the tests of picture vocabulary and listening comprehension from the Woodcock-Muñoz Language Proficiency Battery (Woodcock & Muñoz-Sandoval, 1995). Picture vocabulary consisted of naming objects and listening comprehension consisted of listening to incomplete sentences and filling in the blank with the correct target word. The level of difficulty in each test gradually increased from beginning to end. A language history questionnaire was used to collect information about AoA and the percentage of time spent using each language across different tasks (reading, watching TV, interactions with peers, interactions with family members, etc.). It also collected information about other experiences related to language learning (e.g., travel and residency outside the US). Upon completion of the language history questionnaire, the language assessment, and a metal screening form, a brain scan was conducted in a 3 Tesla Siemens MRI at the Human Neuroimaging Laboratory of Baylor College of Medicine in Houston, Texas.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

2.3  Outside the scanner speech recordings To indirectly assess children’s perceptual abilities with speech, children were asked to read out loud a list of English words that contained the same target vowels heard inside the scanner (i.e., /æ/, /ɒ/, /ʌ/). The words were recorded using a table microphone in a quiet room. Four native speakers of English were asked to judge the intelligibility of the words articulated by explicitly writing the word heard. The judges also rated the degree of foreign accent for each participant using a 5-point scale (1 = no foreign accent, 5 = strong foreign accent).

2.4  Stimuli and fMRI task design Stimuli were naturally recorded using a Sony MZ-NH800 mini-disc recorder and a Sony ECM-CS10 stereo microphone. Praat (Boersma, 2001) and Wavesurfer (Sjolander & Beskow, 2009) were used to normalize the audio and cut each sound file. The stimuli were the English syllables saf (/æ/, e.g., hat), sof (/ɒ/, e.g., hot), and suf (/ʌ/, e.g., hut). The fricatives /s/ and /f/ padded each phoneme at the beginning and at the end of the syllable to enhance the recognition of the vowel (Fernandez, Feijoó, Balsa, & Barros, 1998). The crowded vowel space in which the English phonemes /æ/, /ɒ/, and /ʌ/ reside tends to cause Spanish speakers to assimilate these sounds into the Spanish phonemic category /a/ (e.g., casa) (Flege, Munro, and Mackay, 1995). Therefore, Spanish speakers learning English must have acute perception to discriminate the English sounds from one another. The syllables were presented in pairs (e.g., saf-saf; sof-sof; saf-sof; saf-suf, and sof-suf) during a silent interval in the scanner. Children were instructed to watch a muted non-captioned movie (i.e., Planet Earth) for 25 minutes while the syllables played through the MRI headphones. Verbal information was not presented in the movie to prevent interference with the speech task. To ensure passive listening of the L2 sounds, children were not asked to press buttons to discriminate between syllables. Instead, children’s attention was held constant through the movie, which contained non-verbal images of animals and landscapes. A task of passive listening enabled us to examine basic perceptual processes while directing attentional engagement to an unrelated visual stimulus. Explicitly minimizing the use of executive processes like attention and working memory was important to prevent the conscious use of cognitive strategies that could lead to attending to extra-phonetic cues that may overestimate performance (Best, McRoberts, & Sithole, 1988; Zevin & McCandliss, 2005). We were particularly interested in understanding how executive function plays a role in tasks that are thought to be purely pre-attentive and perceptual. Using a task of overt attention to auditory stimuli would not help us meet the purpose of the study because it would explicitly demand conscious attention to the speech sounds and the results would simply reveal that brain areas of executive function are involved in tasks where such attention is already required. The present study is unique in that increased activity in brain areas involved in executive function is expected in response to tasks that do not explicitly require the engagement of such.



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

2.5  Neuroimaging acquisition parameters Whole-brain scans were performed with a 3.0 Tesla Magnetom Trio (Siemens, ­Germany) at the Human Neuroimaging Laboratory of Baylor College of Medicine in Houston, Texas. A total of 542 functional (T2*-weighted) images were acquired using a clustered volume acquisition (CVA) paradigm to quiet the scanner while the auditory stimuli were presented. An interleaved descending Echo Planar Imaging (EPI) sequence was employed with the following parameters: repetition time (TR) = 3s, TR delay (silent interval) = 1.42s, volume acquisition time (TA) = 1.58s, transversal slices per volume = 26, Echo time (TE) = 30ms; 5 mm thickness, 3.4 × 3.4 × 5.0 mm resolution, flip angle = 90 degrees, with the centermost slice aligned with the anterior commissure and posterior cingulate (AC–PC). High-resolution anatomical images used a T1-weighted Magnetization Prepared Rapid Gradient Echo (MPRAGE) sequence (TR= 1.2s, TE = 2.66 ms, 1 mm³ isotropic voxel size) reconstructed into 192 slices. Auditory stimuli were presented using the in-house software NEMO (Network Experiment Management Objects). This software synchronized the task with the scanner with millisecond accuracy.

2.6  FMRI data analysis Functional images were slice-time corrected, motion-corrected, aligned to anatomical scans, and normalized to MNI stereotaxic space. Spatial smoothing used an 8 mm full-width half maximum Gaussian Kernel. Children’s data was further inspected for intolerable motion with Artifact Detection toolbox (Whitfield-Gabrielli, 2010). Children’s brain scans that exceeded 1 mm of linear movement and 0.5 degree of angular movement were dropped from analyses; about 15% of data points were dropped for each child. Whole-brain analyses at the first level (within-participant, fixed-effects ­analyses) used a block design specification as the statistical model in SPM8 (Wellcome Trust Center for Neuroimaging, London). For each participant, a condition called “L2 speech sounds” combined the effects of all syllable pairs presented to reveal the neural activity evoked by L2 syllables. The condition of L2 speech sounds was contrasted with a baseline condition of silence. ROI analyses in frontal and prefrontal regions of the brain were conducted at the second level (between-groups, random-effects analyses) between balanced and unbalanced learners. The regions of interest were (1) the bilateral inferior frontal gyrus, (2) bilateral middle frontal gyrus, and (3) bilateral supplementary motor area, (4) bilateral anterior cingulate cortex, (5) bilateral superior temporal gyrus, and (6) middle temporal gyrus. These regions were selected because of their known involvement in higher-order cognitive and sensory processing. The regions were created in PickAtlas (Lancaster et al., 2000; Maldjian, Laurienti, Kraft, & Burdette, 2003) and the ROI analyses were conducted with an alpha of 0.05 as peak threshold and at least 20 contiguous voxels as extent threshold. Xjview 8.14 (Cui, Li, & Song, 2011) was used to visualize the results presented here.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

3.  Results 3.1  Participant groups By means of hierarchical clustering, children were classified as balanced or unbalanced dual language learners based on the absolute discrepancy between the level of proficiency of English and the level of proficiency of Spanish obtained from the Woodcock Language Proficiency assessment. Children with balanced proficiency (n = 16) had roughly the same level of proficiency in both languages, with each participant having less than 8 points of discrepancy between languages (English proficiency mean = 49.7 % correct, SD = 10.8; Spanish proficiency mean = 49.6 % correct, SD = 12.08). Children with unbalanced proficiency (n = 13) were more proficient in one language than the other, with each participant having between 10 and 29 points of discrepancy between languages (English proficiency mean = 37.7 % correct, SD = 13.03; ­Spanish proficiency mean = 53 % correct, SD = 12.8). On average, unbalanced dual language learners were significantly more proficient in Spanish than English (t(12) = 4.3, p  speech) showed that young unbalanced dual language learners do not have increased activity in any of the frontal



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

Children with high proficiency in L2 vs. Children with high proficiency in L1 2 1.5 Left superior temporal gyrus

1 0.5 0

Children with high proficiency in L1 vs. Children with high proficiency in L2

Left superior Frontal gyrus

3 2.5 2 1.5 1 0.5 0

Figure 2b.  Increased brain activity according to the direction of language imbalance. Children with higher proficiency in L2 than L1 show increased activity in the left STG. Children with higher proficiency in L1 than L2 show increased activity in the left SFG

regions involved in executive function (t = 0.05, k = 20).2 These results demonstrate that activity in frontal regions of unbalanced dual language learners is in response to the speech sounds and not the movie.

3.3.2  Balanced learners vs. unbalanced learners Balanced learners did not show increased activity in any of the preselected executive function regions of interest relative to unbalanced learners. Instead, balanced dual language learners showed increased activity in the right MTG relative to unbalanced learners (56 –8 –26). No activity was observed in the STG of either hemisphere in ­balanced learners. See Figure 2C and Table 2.

.  All ROI analyses were conducted with a set threshold of 0.05(t) for peak intensity and a minimum of 20 voxels(k) for magnitude. Uncorrected whole-brain analysis used a threshold of t = 0.01 and k = 20.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

Table 2.  Areas of brain activity in balanced and unbalanced L2 child learners in ­pre-selected regions of interest and areas of brain activity given the direction of language imbalance Children with unbalanced proficiency vs. Children with balanced proficiency Cluster ­magnitude

Peak T

MNI coordianate

Peak threshold p-value

Left inferior frontal gyrus

202

2.69

–48 16 –2

0.05

Left middle frontal gyrus

209

2.44

–34 48 20

0.05

Right middile frontal gyrus

28

2.31

36 52 24

0.05

32

1.98

40 48 0

Left anterior cingulate cortex

21

2.77

0 28 – 2

0.05

Right anterior cingulate cortex

77

3

40 30 – 2

0.05

Left supplementary motor area

124

3.22

0 12 64

0.05

30

2.49

–16 2 64

Right supplementary motor area

135

3.05

2 10 64

0.05

Children with higher proficiency in L2 > Children with higher proficiency in L1 Left superior temporal gyrus

60

2.19

–48 –34 –16

22

2.07

–50 –20 12

0.05

Children with higher proficiency in L1 > Children with higher proficiency in L2 Left superior frontal gyrus

48

33.34

–16 8 64

45

22.56

–14 38 50

0.05

Children with balanced proficiency vs. Children with unbalanced proficiency Right middle temporal gyrus

183

3.74

56 –8 –26

0.05

Children with balanced proficiency vs. Children with unbalanced proficiency 4 3 2 1 0



Right middle temporal gyrus

Figure 2c.  Increased brain activity in the middle temporal gyrus in balanced L2 child learners relative to unbalanced L2 child learners



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

3.3.3  Extension of analyses A sample of 13 English-speaking monolingual children (age M = 8.15, SD = 1.3) also completed the task. In this case, monolingual children listened to speech sounds in their native language (i.e., English). Uncorrected whole-brain analysis (t = 0.01, k = 20) showed that listening to speech sounds led to increased activity in the bilateral STG while viewing the movie led to increased activity in the occipital lobe and midbrain region of the superior colliculus in monolingual children. Activity was not observed in frontal regions associated with executive function in response to the speech sounds or the movie at the set threshold. These same whole-brain analyses of L2 speech vs. baseline and baseline vs. L2 speech were conducted with the sample of young dual language learners and the results were virtually the same. In English-speaking monolingual children 6 5 4 3 2 1 0 speech > baseline

3 2 1 0 baseline > speech

Figure 3.  Brain activity evoked by speech sounds and movie in monolingual children. The baseline condition consisted of watching a movie while the speech sounds were turned off. As expected, the bilateral superior temporal gyrus showed increased activity in response to native-language speech sounds and the occipital lobe showed increased activity in response to the movie

4.  Discussion This study examined whether brain areas known to be involved in executive processes are also engaged in the perception of speech. Through the use of a passive listening fMRI task, this study revealed that listening to L2 speech sounds engages brain areas associated with attention, working memory, and error monitoring in children who are unbalanced dual language learners. By contrast, children who have similar levels of proficiency in both languages engage sensory areas of perception and lexico-semantic processing in L2 speech perception. Groups of young balanced and unbalanced language learners of equivalent L1 and L2 percent use, L1 proficiency, socioeducational background, age, and AoA participated in this study. The groups only significantly differed in L2 proficiency. Therefore, differences in neural processing can only be attributed to the significant difference

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

in L2 proficiency, which is tied to the conceptualization of proficiency discrepancy between languages in the groups of balanced and unbalanced bilinguals. The results of the ROI analyses showed that listening to L2 speech sounds engages several regions of the prefrontal cortex in young unbalanced dual learners. Specifically, unbalanced learners showed increased activity in the left inferior frontal gyrus, bilateral middle frontal gyrus, bilateral anterior cingulate gyrus, and bilateral supplementary motor area. These areas have been respectively associated with response inhibition and selection (Moss et al., 2005; Swick, Ashley, & Turken, 2008; Tamm, Menon, & Reiss, 2002; Zhang, Feng, Fox, Gao, & Tan, 2004), working memory (Cohen et al., 1997; Courtney, Ungerleider, Keil, & Haxby, 1997; McCarthy et al., 1994), error monitoring (Botvinick, Cohen, & Carter, 2004; Carter et al., 1998; Van Veen, ­Holroyd, Cohen, Stenger, & Carter, 2004), planned motor action (Goldberg, 1985; Roland, Larsen, Lassen, & Skinhoj, 1980) and have been widely reported in studies of cognitive control in bilinguals during language switching tasks (Abutalebi, 2008; Hernandez et al., 2001; Price et al., 1999; Rodriguez-Fornells et al., 2005). Other areas such as the inferior and superior parietal lobule have also been implicated in processes of auditory working memory (Bushara et al., 1999), but the results of the current study did not show increased activity in either one of these two regions in balanced or unbalanced dual language learners listening to L2 syllables. The set of prefrontal regions that were significantly more active, in addition to the activity observed in the bilateral STG, potentially suggests that young unbalanced dual language learners use areas of executive function to facilitate the perception of L2 syllables. For unbalanced learners, perceiving L2 phonemes may require a controlled process that involves effortful attention. Through monitoring, maintenance of auditory information in working memory, and selection of relevant speech cues, unbalanced learners may improve their perception. Moreover, activity in the SMA suggests that unbalanced learners potentially plan to produce the syllables heard, maybe in an attempt to internally practice the articulation of the sounds. Post-hoc examination of the data showed that unbalanced learners also had increased activity in the left rolandic operculum, an area involved in articulatory processes (Biermann-Ruben, Salmelin, & Schnitzler, 2005; Brown et al., 2009). One possible interpretation is that unbalanced learners are continuously prepared to vocalize the L2 in order to be ready for when overt speech is required. This notion of preparedness for speech output is reasonable, given that unbalanced learners in our sample had thicker foreign accents than balanced learners in L2. Even though young balanced and unbalanced learners were shown to be equally able to intelligibly produce words in the L2, unbalanced learners were more likely to produce words with a thicker foreign accent. This global judgment of foreign accent demonstrates that unbalanced learners have more difficulty perceiving and producing speech sounds in the L2, in addition to the difficulty with overall L2 language proficiency.



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

The analyses conducted also showed increased activity in the right MTG of balanced learners relative to unbalanced learners. Activity in this area, primarily in the left hemisphere, has been associated with lexico-semantic processing (Démonet et al., 1992; Vandenberghe, Price, Wise, Josephs, & Frackowiak, 1996), and mappings of sound to meaning that give rise to comprehension (Hickok & Poeppel, 2007). The integration of linguistic information from visual and auditory modalities in the middle temporal gyrus renders this region significant for reading (Shaywitz et al., 2002; Simos et al., 2009). The MTG in the right hemisphere also appears to play a role in semantic processing, but to a lesser extent. For example, cerebral blood flow to the right MTG increases in response to discourse processing (St George, Kutas, Martinez, & Sereno, 1999), and in tasks of phonological naming in reading (Turkeltaub, Gareau, Flowers, Zeffiro, & Eden, 2003). Given these previous studies, it appears that balanced learners may be attending to larger suprasegmental cues in the speech signal, which could result in the addition of illusory meaning to the syllables heard. Therefore, balanced young dual language learners might be processing L2 syllables as meaningful words. A future study should address this possibility using more precise methodology, as the design of the current study was not developed to investigate whether or not balanced learners interpret certain types of syllables as words. Further evidence for the involvement of prefrontal regions in L2 speech processing comes from observed left SFG activity. Unbalanced children who were more proficient in the first language than in the second show more SFG activity, but children who were more proficient in the second language than in the first do not. Rather, unbalanced learners who were more proficient in the L2 than in the L1 showed activity in the left STG. Therefore, the direction of the linguistic imbalance (better proficiency in the L1 than the L2 or better proficiency in the L2 than the L1) influences which brain regions are recruited. Unbalanced learners, especially those with better proficiency in the L1 than the L2, appear to be recruiting prefrontal regions because they are having more difficulty learning the speech sounds of the L2. Two possible interpretations for the brain activity observed in balanced and unbalanced bilinguals in response to listening to the L2 speech sounds are (1) individual differences in speech perception abilities, and (2) quality of language use. For the first interpretation concerning individual ability in speech perception, it is possible that speech sounds in the L2 engage regions of the temporal lobe associated with speech and language processing in balanced learners, because the brains allow for more efficient processing of speech sound information in the L1 and the L2. For example, it has been reported that fast learners of speech sounds have greater white matter density in the left STG than slow learners (Golestani, Molko, Dehaene, LeBihan,  & ­Pallier, 2007; Golestani, Paus, & Zatorre, 2002) and that expert phoneticians have multiple or split left transverse gyri in the auditory cortex compared to non-experts (Golestani, Price, & Scott, 2011). It has also been found that highly proficient bilinguals

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

with acute perceptual skills for speech sounds have thinner cortex in the region of the STG and angular gyrus (Burgaleta, Baus, Diaz, & Sebastian-Galles, 2014). It is possible that balanced bilinguals who have adequate brain structure also have proper emergent function of the STG bilaterally that allows them to hear speech sounds in L1 and L2 better than unbalanced bilinguals. This interpretation seems appropriate in light of mounting evidence demonstrating that good perceivers and good learners of speech sounds show more intense and widespread activity of the STG (Archila-Suerte et al., 2014; Diaz et al., 2008; Gaab, Gaser, & Schlaug, 2006; Wong et  al., 2007). An individual ability for speech sound perception would predispose young balanced learners to develop more advanced linguistic skills in both languages. Future studies should directly assess the development of L1 and L2 speech perception and its effect on language development, as very little has been done to date. For the second interpretation related to quality of language input, it is possible that balanced language learners have a higher level of English (L2) proficiency than unbalanced language learners because they have more active and diverse socio-­ linguistic interactions while using the L2, which results in enhanced exposure that leads to enriched vocabulary in the L2. Alternatively, unbalanced learners may not be properly using their time in English to build more proficiency in the L2. If the quality of input in the L2 is poor, then extensive use of the language may not be enough for learning, as the linguistic atmosphere would be considered deficient. Future studies should be conducted to scrutinize the relationship between quality of input and speech perception in L2. Alternatively, future studies could be conducted to investigate the effect of executive function training on the improvement of L2 speech perception in young unbalanced dual language learners. It appears that balanced dual language learners who are more advanced in their knowledge of the L2 engage brain regions that are specifically designed to process linguistic information. According to the interactive specialization hypothesis of brain development proposed by Johnson (2001), regions of cortex become specialized to respond to a particular stimulus (e.g., speech sounds) through interactions between a broad number of brain regions and networks earlier in development. In the data presented here, a third potential interpretation is that the processing of L2 speech sounds requires the recruitment of prefrontal areas in unbalanced children who are still undergoing a prolonged period of L2 learning, whereas children with balanced proficiency appear to have already gone through the process of neural specialization and therefore engage expected regions of the temporal lobe to process speech sounds in the L2. In the fMRI task, participants were not asked to overtly respond to auditory stimuli. Therefore, the results suggest that children, who are typically thought to learn a second language easily, are actually exerting more effort to process L2 speech sounds than usually recognized. This additional effort (or lack thereof) in perceptual tasks



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

may help explain why some young dual language learners are more successful at L2 learning than others. Participants were classified as balanced or unbalanced dual language learners based on the level of discrepancy between the proficiency in their L1 and L2. Brain activity elicited by L2 speech sounds was compared between groups. Given this model, more intense activity in any given brain region was interpreted as increased engagement of the mental process typically ascribed to the area. Depending on the brain region, more intense activity may be the result of increased conscious effort or increased perceptual processing. For example, unbalanced participants (who had less proficiency and more difficulty perceiving and producing speech sounds in the L2) showed more intense activity in areas involved in executive function. Therefore, it is likely that these children go through a more laborious experience processing and learning L2 speech sounds. On the other hand, balanced participants show more intense activity in areas known to be involved in perceptual processing of speech. Therefore, it is likely that these children effectively rely on primary and secondary sensory areas to process speech. It is worth clarifying that the results presented here do not indicate that frontal regions are directly involved in speech perception. Instead, they suggest that regions of the frontal lobe (known to be involved in executive functions) support perception of speech sounds in L2 learning, likely through increased attention and extensive use of working memory. One limitation to our study is not having included L1 speech sounds to investigate how young Spanish-English dual language learners respond to L1. Future studies should compare brain activity evoked by L1 and L2 in children with different levels of proficiency in each language to corroborate the results presented here.

5.  Conclusion This fMRI study showed that children matched on age, years of education, socioeducational background, language use, and L1 proficiency among other variables, recruit different brain areas to process L2 speech syllables depending on their status of balanced or unbalanced proficiency between languages. Unbalanced dual language learners recruit bilateral frontal regions, whereas balanced dual language learners recruit a region of the temporal lobe in the right hemisphere. The results suggest that while unbalanced learners engage regions involved in executive function to support the processing of L2 speech sounds, balanced learners engage temporal regions already associated with linguistic processing. Moreover, when the directional imbalance between languages is lower proficiency in L1 and higher proficiency in L2, these unbalanced learners also recruit regions of the temporal lobe in the left hemisphere known to ­process acoustic-phonetic information, just like balanced learners.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez

References Abutalebi, J. (2008). Neural aspects of second language representation and language control. Acta Psychologica, 128(3), 466–478.  doi: 10.1016/j.actpsy.2008.03.014 Abutalebi, J., Brambati, S., Annoni, J., Moro, A., Cappa, S., & Perani, D. (2007). The neural cost of the auditory perception of language switches: An event-related functional magnetic resonance imaging study in bilinguals. Journal of Neuroscience, 27(50), 13762–13769.

doi: 10.1523/JNEUROSCI.3294-07.2007

Archila-Suerte, P., Bunta, F., & Hernandez, A. (2014). Speech sound learning depends on individuals’ ability, not just experience. International Journal of Bilingualism. Archila-Suerte, P., Zevin, J., Bunta, F., & Hernandez, A. (2011). Age of acquisition and proficiency in a second language independently influence the perception of non-native speech. Bilingualism: Language and Cognition, 15(1), 190–201.  doi: 10.1017/S1366728911000125 Archila-Suerte, P., Zevin, J., & Hernandez, A. (2015). The effect of age of acquisition, socioeducational status, and proficiency on the neural processing of second language speech sounds. Brain & Language, 141, 35–49.  doi: 10.1016/j.bandl.2014.11.005 Archila-Suerte, P., Zevin, J., Ramos, A., & Hernandez, A. (2013). The neural basis of non-native speech perception in bilingual children. NeuroImage, 67, 51–63.

doi: 10.1016/j.neuroimage.2012.10.023

Best, C., McRoberts, G., & Goodell, E. (2001). Discrimination of non-native consonant contrasts varying in perceptual assimilation to the listener’s native phonological system. J­ournal of the Acoustical Society of America, 109(2), 775–794.  doi: 10.1121/1.1332378 Best, C., McRoberts, G., & Sithole, N. (1988). Examination of perceptual reorganization for nonnative speech contrasts: Zulu click discrimination by English-speaking adults and infants. he Journal of Experimental Psychology: Human Perception and Performance, 14(3), 345–360.  doi: 10.1037/0096-1523.14.3.345 Bialystok, E. (2009). Bilingualism: The good, the bad, and the indifferent. Bilingualism: Language and Cognition, 12(1), 3–11.  doi: 10.1017/S1366728908003477 Biermann-Ruben, K., Salmelin, R., & Schnitzler, A. (2005). Right rolandic activation during speech perception in stutterers: A MEG study. NeuroImage, 25(3), 793–801.

doi: 10.1016/j.neuroimage.2004.11.024

Binder, J. (2000). The new neuroanatomy of speech perception. Brain, 123(12), 2371–2372.

doi: 10.1093/brain/123.12.2371

Boersma, P. (2001). Praat, a system for doing phonetics by computer. Glot International, 5(9/10), 341–345. Botvinick, M., Cohen, J., & Carter, C. (2004). Conflict monitoring and anterior cingulate cortex: an update. Trends in Cognitive Sciences, 8(12), 539–546.  doi: 10.1016/j.tics.2004.10.003 Breier, J., Fletcher, J., Denton, C., & Gray, L. (2004). Categorical perception of speech stimuli in children at risk for reading difficulty. Journal of Experimental Child Psychology, 88(2), 152–170.  doi: 10.1016/j.jecp.2004.03.003 Breier, J., Simos, P., Fletcher, J., Castillo, E., Zhang, W., & Papanicolaou, A. (2003). Abnormal activation of temporoparietal language areas during phonetic analysis in children with dyslexia. Neuropsychology, 17(4), 610–621.  doi: 10.1037/0894-4105.17.4.610 Brown, S., Laird, A., Pfordresher, P., Thelen, S., Turkeltaub, P., & Liotti, M. (2009). The somatotopy of speech: Phonation and articulation in the human motor cortex. Brain Cogn, 70(1), 31–41.  doi: 10.1016/j.bandc.2008.12.006



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

Burgaleta, M., Baus, C., Diaz, B., & Sebastian-Galles, N. (2014). Brain structure is related to speech perception abilities in bilinguals. Brain Struct Funct, 219(4), 1405–1416.

doi: 10.1007/s00429-013-0576-9

Bushara, K., Weeks, R., Ishii, K., Catalan, M., Tian, B., Rauschecker, J., & Hallett, M. (1999). Modality-specific frontal and parietal areas for auditory and visual spatial localization in humans. Nature Neuroscience, 2(8), 759–766.  doi: 10.1038/11239 Carter, C., Braver, T., Barch, D., Botvinick, M., Noll, D., & Cohen, J. (1998). Anterior cingulate cortex, error detection, and the online monitoring of performance. Science, 280(5364), 747–749.  doi: 10.1126/science.280.5364.747 Casaponsa, A., Carreiras, M., & Dunabeitia, J. (2014). Discriminating languages in bilingual contexts: The impact of orthographic markedness. Frontiers in Psychology, 5.

doi: 10.3389/fpsyg.2014.00424

Celsis, P., Boulanouar, K., Doyon, B., Ranjeva, J., Berry, I., Nespoulous, J., & Chollet, F. (1999). Differential fMRI responses in the left posterior superior temporal gyrus and left supramarginal gyrus to habituation and change detection in syllables and tones. NeuroImage, 9(1), 135–144.  doi: 10.1006/nimg.1998.0389 Cohen, J., Perlstein, W., Braver, T., Nystrom, L., Noll, D., Jonides, J., & Smith, E. (1997). Temporal dynamics of brain activation during a working memory task. Nature, 386(6625), 604–608.  doi: 10.1038/386604a0 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.  doi: 10.1016/j.jml.2004.02.002 Courtney, S., Ungerleider, L., Keil, K., & Haxby, J. (1997). Transient and sustained activity in a distributed neural system for human working memory. Nature, 386(6625), 608–611.

doi: 10.1038/386608a0

Cui, X., Li, J., & Song, X. (2011). Xjview: A viewing program for SPM. Retreived from www. alivelearn.net/xjview8. Cummins, J. (1984). Wanted: A theoretical framework for relating language proficiency to academic achievement among bilingual students. Language proficiency and academic achievement, 10, 2–19. De Groot, A. (2011). Language and cognition in bilinguals and multilinguals: An introduction. New York, NY: Psychology Press. Démonet, J.-F., Chollet, F., Ramsay, S., Cardebat, D., Nespoulous, J.-L., Wise, R.,…& ­Frackowiak, R. (1992). The anatomy of phonological and semantic processing in normal subjects (Vol. 115). Diaz, B., Baus, C., Escera, C., Costa, A., & Sebastian-Galles, N. (2008). Brain potentials to native phoneme discrimination reveal the origin of individual differences in learning the sounds of a second language. Proceedings of the National Academic of Sciences, 105(42), 16083–16088.  doi: 10.1073/pnas.0805022105 Fabbro, F., Skrap, M., & Aglioti, S. (2000). Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery & Psychiatry, 68(5), 650–652.  doi: 10.1136/jnnp.68.5.650 Fernandez, R., & Nielsen, F. (1986). Bilingualism and Hispanic scholastic achievement: Some baseline results. Social Science Research, 15(1), 43–70.  doi: 10.1016/0049-089X(86)90003-7 Fernandez, S., Feijoó, S., Balsa, R., & Barros, N. (1998). Recognition of vowels in fricative ­context. ICSLP(0452).

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez Flege, J. (2003). Assessing constraints on second language segmental production and perception. In N. Schiller & A. Meyer (Eds.), Phonetics and phonology in language comprehension and production. Berlin, Germany: Mouton de Gruyter. Flege, J. E., Munro, M. J., & MacKay, I. R. (1995). Effects of age of second-language learning on the production of English consonants. Speech Communication, 16(1), 1–26. Gaab, N., Gaser, C., & Schlaug, G. (2006). Improvement-related functional plasticity following pitch memory training. NeuroImage, 31(1), 255–263.  doi: 10.1016/j.neuroimage.2005.11.046 Garbin, G., Costa, A., Sanjuan, A., Forn, C., Rodriguez-Pujadas, A., Ventura, N.,…Avila, C. (2011). Neural bases of language switching in high and early proficient bilinguals. Brain & Language, 119(3), 129–135.  doi: 10.1016/j.bandl.2011.03.011 Goldberg, G. (1985). Supplementary motor area structure and function: Review and hypotheses. Behavioral and Brain Sciences, 8(4), 567–588.  doi: 10.1017/S0140525X00045167 Golestani, N., Molko, N., Dehaene, S., LeBihan, D., & Pallier, C. (2007). Brain structure predicts the learning of foreign speech sounds. Cereb Cortex, 17(3), 575–582.

doi: 10.1093/cercor/bhk001

Golestani, N., Paus, T., & Zatorre, R. (2002). Anatomical correlates of learning novel speech sounds. Neuron, 35(5), 997–1010.  doi: 10.1016/S0896-6273(02)00862-0 Golestani, N., Price, C., & Scott, S. (2011). Born with an ear for dialects? Structural plasticity in the expert phonetician brain. J Neurosci, 31(11), 4213–4220.

doi: 10.1523/JNEUROSCI.3891-10.2011

Gollan, T., Fennema-Notestine, C., Montoya, R., & Jernigan, T. L. (2007). The bilingual effect on Boston Naming Test performance. Journal of the International Neuropsychological Society, 13(02), 197–208. Gollan, T., & Ferreira, V. (2009). Should I stay or should I switch? A cost-benefit analysis of voluntary language switching in young and aging bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(3), 640–665.  doi: 10.1037/a0014981 Grosjean, F., & Li, P. (2012). The psycholinguistics of bilingualism. New York, NY: Wiley. Hazan, V., & Boulakia, G. (1993). Perception and production of a voicing contrast by FrenchEnglish bilinguals. Language and Speech, 36(1), 17–38. Hernandez, A. (2009). Language switching in the bilingual brain: What’s next? Brain and Language, 109(2), 133–140.  doi: 10.1016/j.bandl.2008.12.005 Hernandez, A., Dapretto, M., Mazziotta, J., & Bookheimer, S. (2001). Language switching and language representation in Spanish-English bilinguals: an fMRI study. NeuroImage, 14(2), 510–520.  doi: 10.1006/nimg.2001.0810 Hernandez, A., Martinez, A., & Kohnert, K. (2000). In search of the language switch: An fMRI study of picture naming in Spanish-English bilinguals. Brain & Language, 73(3), 421–431.  doi: 10.1006/brln.1999.2278 Hickok, G., & Poeppel, D. (2000). Towards a functional neuroanatomy of speech perception. Trends in Cognitive Sciences, 4(4), 131–138.  doi: 10.1016/S1364-6613(00)01463-7 Hickok, G., & Poeppel, D. (2007). The cortical organization of speech processing. Nature Reviews Neuroscience, 8(5), 393–402.  doi: 10.1038/nrn2113 Holt, L., & Lotto, A. (2006). Cue weighting in auditory categorization: implications for first and second language acquisition. The Journal of the Acoustical Society of America, 119(5 Part 1), 3059–3071.  doi: 10.1121/1.2188377 Hutzler, F. (2014). Reverse inference is not a fallacy per se: cognitive processes can be inferred from functional imaging data. NeuroImage, 84, 1061–1069.  doi: 10.1016/j.neuroimage.2012.12.075



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

Iverson, P., Hazan, V., & Bannister, K. (2005). Phonetic training with acoustic cue manipulations: a comparison of methods for teaching English /r/ – /l/ to Japanese adults. The Journal of the Acoustical Society of America, 118(5), 3267–3278.  doi: 10.1121/1.2062307 Joanisse, M., Robertson, E., & Newman, R. (2007). Mismatch negativity reflects sensory and phonetic speech processing. NeuroReport, 18(9), 901–905.  doi: 10.1097/WNR.0b013e3281053c4e Joanisse, M., Zevin, J., & McCandliss, B. (2007). Brain mechanisms implicated in the preattentive categorization of speech sounds revealed using FMRI and a short-interval habituation trial paradigm. Cereb Cortex, 17(9), 2084–2093.  doi: 10.1093/cercor/bhl124 Johnson, M. (2001). Functional brain development in humans. Nature Reviews Neuroscience, 2(7), 475–483.  doi: 10.1038/35081509 Kroll, J., & de Groot, A. (Eds.) (2005). Handbook of bilingualism: Psycholinguistic approaches. New York, NY: Oxford University Press. Kuhl, P. (2008). Linking infant speech perception to language acquisition: Phonetic learning predicts language growth. In P. McCardle, J. Colombo, & L. Freund (Eds.), Infant pathways to language: Methods, models, and research directions (pp. 213–243). New York, NY: Lafwrence Erlbaum Associates. Kuhl, P., Conboy, B., Padden, D., Nelson, T., & Pruitt, J. (2005). Early speech perception and later language development: Implications for the “critical period.” Language Learning and Development, 1(3), 237–264.  doi: 10.1080/15475441.2005.9671948 Kushalnagar, P., Hannay, H., & Hernandez, A. (2010). Bilingualism and attention: A study of balanced and unbalanced bilingual deaf users of American Sign Language and English. Journal of deaf studies and deaf education, 15(3), 263–273.  doi: 10.1093/deafed/enq011 Lancaster, J., Woldorff, M., Parsons, L., Liotti, M., Freitas, C., Rainey, L.,…& Fox, P. (2000). Automated Talairach atlas labels for functional brain mapping. Human Brain Mapping, 10(3), 120–131.  doi: 10.1002/1097-0193(200007)10:33.0.CO;2-8 Liebenthal, E., Binder, J., Spitzer, S., Possing, E., & Medler, D. (2005). Neural substrates of phonemic perception. Cerebral Cortex, 15(10), 1621–1631.  doi: 10.1093/cercor/bhi040 Luk, G., Green, D., Abutalebi, J., & Grady, C. (2011). Cognitive control for language switching in bilinguals: A quantitative meta-analysis of functional neuroimaging studies. Language and Cognitive Processes, 27(10), 1479–1488.  doi: 10.1080/01690965.2011.613209 Maldjian, J., Laurienti, P., Kraft, R., & Burdette, J. (2003). An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. NeuroImage, 19(3), 1233–1239.  doi: 10.1016/S1053-8119(03)00169-1 McCarthy, G., Blamire, A., Puce, A., Nobre, A., Bloch, G., Hyder, F.,…& Shulman, R. (1994). Functional magnetic resonance imaging of human prefrontal cortex activation during a spatial working memory task. Proceedings of the National Academy of Sciences, 91(18), 8690–8694.  doi: 10.1073/pnas.91.18.8690 Moss, H., Abdallah, S., Fletcher, P., Bright, P., Pilgrim, L., Acres, K., & Tyler, L. (2005). Selecting among competing alternatives: selection and retrieval in the left inferior frontal gyrus. Cerebral Cortex, 15(11), 1723–1735.  doi: 10.1093/cercor/bhi049 Pallier, C., Bosch, L., & Sebastian-Galles, N. (1997). A limit on behavioral plasticity in speech perception. Cognition, 64(3), B9-17.  doi: 10.1016/S0010-0277(97)00030-9 Poldrack, R. (2006). Can cognitive processes be inferred from neuroimaging data? Trends in Cognitive Sciences, 10(2), 59–63.  doi: 10.1016/j.tics.2005.12.004 Price, C., Green, D., & von Studnitz, R. (1999). A functional imaging study of translation and language switching, 122(12), 2221–2235.

 Pilar Archila-Suerte, Brandin A. Munson & Arturo E. Hernandez Prior, A., & Gollan, T. (2011). Good language-switchers are good task-switchers: Evidence from Spanish-English and Mandarin-English bilinguals. Journal of the International Neuropsychological Society, 17(4), 682–691.  doi: 10.1017/S1355617711000580 Prior, A., & Macwhinney, B. (2010). A bilingual advantage in task switching. Bilingualism: Language and Cognition, 13(2), 253–262.  doi: 10.1017/S1366728909990526 Rodriguez-Fornells, A., Lugt, A., Rotte, M., Britti, B., Heinze, H.-J., & Münte, T. (2005). Second language interferes with word production in fluent bilinguals: Brain potential and functional imaging evidence. Journal of Cognitive Neuroscience, 17(3), 422–433.

doi: 10.1162/0898929053279559

Roland, P., Larsen, B., Lassen, N., & Skinhoj, E. (1980). Supplementary motor area and other cortical areas in organization of voluntary movements in man. Journal of Neurophysiology, 43(1), 118–136. San Francisco, A., Carlo, M., August, D., & Snow, C. (2006). The role of language of instruction and vocabulary in the English phonological awareness of Spanish–English bilingual children. Applied Psycholinguistics, 27(2), 229–246.  doi: 10.1017/S0142716406060267 Schwartz, M. (2014). The impact of the First Language First model on vocabulary development among preschool bilingual children. Reading and Writing, 27(4), 709–732.

doi: 10.1007/s11145-013-9463-2

Sebastián-Gallés, N., Soriano-Mas, C., Baus, C., Díaz, B., Ressel, V., Pallier, C.,…& Pujol, J. (2012). Neuroanatomical markers of individual differences in native and non-native vowel perception. Journal of Neurolinguistics, 25(3), 150–162.  doi: 10.1016/j.jneuroling.2011.11.001 Shaywitz, B., Shaywitz, S., Blachman, B., Pugh, K., Fulbright, R., Skudlarski, P.,…& Gore, J. (2004). Development of left occipitotemporal systems for skilled reading in children after a phonologically- based intervention. Biological Psychiatry, 55(9), 926–933.

doi: 10.1016/j.biopsych.2003.12.019

Shaywitz, B., Shaywitz, S., Pugh, K., Mencl, W., Fulbright, R., Skudlarski, P.,…& Gore, J. (2002). Disruption of posterior brain systems for reading in children with developmental dyslexia. Biological Psychiatry, 52(2), 101–110.  doi: 10.1016/S0006-3223(02)01365-3 Simos, P., Pugh, K., Mencl, E., Frost, S., Fletcher, J., Sarkari, S., & Papanicolaou, A. (2009). Temporal course of word recognition in skilled readers: A magnetoencephalography study. Behavioural Brain Research, 197(1), 45–54.  doi: 10.1016/j.bbr.2008.07.038 Sjolander, K., & Beskow, J. (2009). Wavesurfer [Computer program] (Version 1.8.5). St George, M., Kutas, M., Martinez, A., & Sereno, M. (1999). Semantic integration in reading: engagement of the right hemisphere during discourse processing. Brain, 122(7), 1317–1325.  doi: 10.1093/brain/122.7.1317 Stanovich, K., Cunningham, A., & Cramer, B. (1984). Assessing phonological awareness in kindergarten children: Issues of task comparability. Journal of Experimental Child Psychology, 38(2), 175–190.  doi: 10.1016/0022-0965(84)90120-6 Swick, D., Ashley, V., & Turken, A. (2008). Left inferior frontal gyrus is critical for response inhibition. BMC Neurosci, 9, 102.  doi: 10.1186/1471-2202-9-102 Tamm, L., Menon, V., & Reiss, A. (2002). Maturation of brain function associated with response inhibition. Journal of the American Academy of Child & Adolescent Psychiatry, 41(10), 1231–1238.  doi: 10.1097/00004583-200210000-00013 Turkeltaub, P., Gareau, L., Flowers, D., Zeffiro, T., & Eden, G. (2003). Development of neural mechanisms for reading. Nature Neuroscience, 6(7), 767–773.  doi: 10.1038/nn1065 Van Veen, V., Holroyd, C. B., Cohen, J. D., Stenger, V. A., & Carter, C. S. (2004). Errors without conflict: Implications for performance monitoring theories of anterior cingulate cortex. Brain and Cognition, 56(2), 267–276.  doi: 10.1016/j.bandc.2004.06.007



Chapter 3.  The role of executive function in the perception of L2 speech sounds 

Vandenberghe, R., Price, C., Wise, R., Josephs, O., & Frackowiak, R. (1996). Functional anatomy of a common semantic system for words and pictures. Nature, 383(6597), 254–256.

doi: 10.1038/383254a0

Wang, Y., Xue, G., Chen, C., Xue, F., & Dong, Q. (2007). Neural bases of asymmetric language switching in second-language learners: An ER-fMRI study. NeuroImage, 35(2), 862–870. doi: 10.1016/j.neuroimage.2006.09.054 Whitfield-Gabrielli, S. (2010). Artifact detection and QA manual. Retreived from: 〈http://web. mit.edu/swg/art/art.pdf〉. Wong, P., Perrachione, T., & Parrish, T. (2007). Neural characteristics of successful and less successful speech and word learning in adults. Human Brain Mapping, 28(10), 995–1006.

doi: 10.1002/hbm.20330

Woodcock, R., & Muñoz-Sandoval, A. (1995). Woodcock-Johnson Language Proficiency BatteryRevised (Spanish). Itasca, IL: Riverside. Ylinen, S., Uther, M., Latvala, A., Vepsäläinen, S., Iverson, P., Akahane-Yamada, R., & Näätänen, R. (2010). Training the brain to weight speech cues differently: A study of Finnish secondlanguage users of English. Journal of Cognitive Neuroscience, 22(6), 1319–1332.

doi: 10.1162/jocn.2009.21272

Zatorre, R., Evans, A., Meyer, E., & Gjedde, A. (1992). Lateralization of phonetic and pitch discrimination in speech processing. Science, 256(5058), 846–849.  doi: 10.1126/science.1589767 Zevin, J., & McCandliss, B. (2005). Dishabituation of the BOLD response to speech sounds. Behavioral and Brain Functions, 1(1), 1–4.  doi: 10.1186/1744-9081-1-4 Zhang, J., Feng, C.-M., Fox, P., Gao, J.-H., & Tan, L. (2004). Is left inferior frontal gyrus a general mechanism for selection? NeuroImage, 23(2), 596–603. doi: 10.1016/j.neuroimage.2004.06.006

chapter 4

Are cognate words “special”? On the role of cognate words in language switching performance* Mikel Santesteban & Albert Costa

University of the Basque Country (UPV/EHU), Spain / ICREA/Universitat Pompeu Fabra, Spain One of the most remarkable abilities of bilingual speakers is that of keeping their two languages apart during speech production. Several researchers have argued that the attentional mechanisms responsible for this ability may vary depending on the L2-proficiency achieved by the bilinguals (Green, 1986; Costa & Santesteban, 2004). Here we use a language-switching task to further explore the extent to which these different attentional strategies depend on the cognate status of the words (Costa, Santesteban, & Caño, 2005). We also explore whether the production of cognates facilitates language-switching (Broersma & De Bot, 2006). Twenty-four low-proficient (L2-Learners) and twenty-four high-proficient (Bilinguals) Spanish-Catalan bilinguals performed a cued language-switching task including cognate and non-cognate words. Our results revealed the following: (a) responses to cognates were faster than to non-cognates; (b) the magnitude of the switching cost was similar regardless of the cognate status of either the preceding or the target word; (c) L2-Learners vs. Bilinguals showed different language switching cost patterns (replicating Costa & Santesteban, 2004). Overall, these findings suggest that the cognate status of words does not facilitate language switching and it does not alter the lexical selection mechanisms’ implicated during production.

* This research was supported by research grants from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant no. 613465-AThEME, as well as from the Basque Government (IT665-13), the Catalan ­Government (2014 SGR 1210) and the Spanish Government (FFI2014-55733-P, PSI201123033). Mikel Santesteban is supported by a Ramon y Cajal research fellowship (RYC-201314722). We thank two anonymous reviewers for their very helpful comments.

doi 10.1075/bpa.2.05san © 2016 John Benjamins Publishing Company

 Mikel Santesteban & Albert Costa

1.  Introduction Bilingual speakers are able to speak in any of their two languages with their interlocutors, and if these were also bilinguals, they can even switch between their two languages in the course of the same conversation. The main factors leading bilinguals to switch between their two languages can be very diverse (e.g., social and pragmatic consideration, linguistic typological distance, etc), and are still not fully determined (De Bot, Broersma & Isurin, 2009). One of the linguistic factors that have been suggested it might trigger the occurrence of language switching is the formal similarity between translation words (e.g., their cognate status; Broersma & De Bot, 2006; Clyne, 1967; 2003). Cognates are those translation words that are phonologically (and/or orthographically) similar in the two languages of a bilingual (e.g., the Spanish-English pair tren-train) while non-cognates are dissimilar (e.g., barco-ship). Cognate words are in comparison to non-cognates: (a) faster to learn and more resistant to forgetting (e.g., Lotto & de Groot, 1998), (b) faster to produce and to translate (Christoffels, Firk, & Schiller, 2007; Costa, Caramazza, & Sebastián-Gallés, 2000; Sánchez-Casas, Davis, & García-Albea, 1992; Strijkers, Costa, & Thierry, 2010), (c) more sensitive to cross-linguistic priming (De Groot & Nas, 1991; Van Hell & De Groot, 1998), and (d) elicit a more similar brain activity between languages (De Blesser, Dupont, Postler, et al., 2003). The cognate facilitation effects reveal that cognates lead to a higher level of co-activation of the bilinguals’ two lexicons as compared to non-cognates (see Costa, Santesteban, & Caño, 2005 for a discussion on the origin of cognate effects). Interestingly, since the special characteristics of cognates enhance cross-linguistic co-activation, Clyne (1967; 2003) and Broersma and De Bot (2006) proposed that cognates might trigger language switching during sentence production. Here we review the experimental evidence for these claims, and we further explore whether cognates also trigger (or facilitate) language switching in isolated word naming conditions. For that purpose, we explore whether the cognate status of words affect the language switching performance of low and high proficient bilinguals.

2.  Bilingual lexical selection and the cognate facilitation effect Present models of bilingual lexical access assume that during the course of lexicalization in one language (e.g., the L2), the semantic system activates in parallel the lexical nodes of the two languages (Kroll & Gollan, 2014; Runnqvist, Strickers, & Costa, 2014). Interestingly, if the two lexicons of a bilingual are activated, how do bilingual speakers avoid the activation of the lexical nodes of the non-response language? Some models of lexical access assume that the lexical selection mechanism is language-specific (Costa & Caramazza, 1999; Costa, Miozzo, & Caramazza, 1999), meaning that it only



Chapter 4.  Are cognate words “special”? 

considers the activation-levels of words in the intended language. Therefore, although words of the non-response language will be activated, they will not be able to interfere during lexical access (see Costa, 2005, for further details of this mechanism). According to the so-called inhibitory (or language non-specific selection) models, all activated lexical nodes are considered for selection, irrespective of the language to which they belong. Thus, successful selection of the proper lexical node is achieved through a reactive inhibition of lexical nodes belonging to the non-response language (Green, 1986, 1998; Meuter & Allport, 1999). This inhibition is proportional to the level of activation of the lexical items, with more activated items requiring more inhibition. Thus, a low-proficient bilingual speaking in her dominant first language (L1), would not require much inhibition to suppress the activation of her non-dominant L2, because the baseline level of activation of L2 items is assumed to be lower than that of L1 items. However, when speaking in L2, L1 items must be strongly inhibited to ensure selection of L2 items. Most evidence about the nature of bilingual lexical selection mechanisms comes from the picture-word interference and the language-switching paradigms (see Costa, 2005; Bobb & Wodniecka, 2013 for comprehensive reviews). Since here we will use the language-switching task, we will only review the evidence provided with this task. In the language-switching task, participants name pictures in L1 and L2, d ­ epending on a language cue (e.g., the colour of the picture). Sometimes, participants name the pictures in the same language used in the preceding trial (non-switch trials), while on others they do so in a different language (switch trials). The most common result in this task is the larger naming latencies for switch than for non-switch trials (the language-switching cost. The magnitude of the language-switching cost is supposed to stem from, at least, two sources: a) the time needed to change the task goal (name in L1, name in L2), and (b) the time needed to retrieve lexical representations inhibited in the previous trial (e.g., Meuter & Allport, 1999). We are interested in this second component of the switching cost, and specifically on its impact in the performance of low-proficient bilinguals. Low-proficient bilinguals usually show asymmetrical switching costs, with larger costs for the dominant than for the weaker language (e.g., Costa & Santesteban, 2004; Meuter & Allport, 1999). This asymmetrical switching cost has been interpreted as revealing the presence of inhibitory processes during bilingual lexical access, supporting the Inhibitory Control model (Costa & Santesteban, 2004; Green, 1998; Meuter & Allport, 1999). It is argued that, when speaking in the weak L2, the lexicon of the dominant L1 is inhibited in order to avoid cross-linguistic interference during lexical selection. When speakers switch to their L1, they need to overcome the strong inhibition applied before, resulting in considerably slower response latencies. In contrast, when naming in the dominant L1 the lexical entries corresponding to the weaker L2 are not inhibited so much (or at all). Thus, when a switch to L2 is required, the retrieval of L2 words will not be so hard.

 Mikel Santesteban & Albert Costa

Following this rationale, the Inhibitory Control model predicts that balanced bilinguals should show symmetrical language switching costs. Since they are high-proficient in their two languages, they should be inhibited to a similar extent. Costa and Santesteban (2004) and Costa, Santesteban and Ivanova (2006) confirmed this prediction and showed that high-proficient bilinguals show symmetric switching patterns regardless of the age at which the L2 was learned and of the similarities of the two languages of the bilinguals (see also Christoffels et al. 2007; Gollan & Ferreira; 2009; Martin, Strijkers, Santesteban, Escera, Hartsuiker, & Costa, 2013). Importantly, these bilinguals also showed symmetrical switching costs while switching between their dominant L1 or L2 and their much weaker L3. Based on the observed link between the switching cost patterns and L2 language proficiency, Costa and Santesteban (2004) argued that while low-proficient bilinguals rely on inhibitory control mechanisms, high-proficient bilinguals rely on non-inhibitory language-specific selection mechanisms. However, other studies showed that language switching cost patterns might also be linked to stimulus type (Finkbeiner, Almeida, Janssen, & Caramazza, 2006), ­grammatical difficulty of the task (Tarlowski, Wodniecka, & Marzecova, 2013) or task-related factors such as Cue-Stimulus Interval or Inter-Trial Time (Verhoef, ­ ­Roelofs, & Chwilla, 2009; but see Fink & Goldrick, 2015). Bobb and Wodniecka (2013) made an extensive review of all the participant-related or task-related factors known to modulate the asymmetrical/symmetrical language switching patterns, and warned researchers that the language switching paradigm might not be the best tool to reveal the use of inhibition. In fact, as we will mention in the General Discussion, alternative accounts suggested that the asymmetrical language switching costs might be reflecting persisting activation of the weaker language rather than inhibition (Phillip, Gade, & Koch, 2007). Nevertheless, in our studies we consistently showed that, keeping all task related factors unchanged (Costa & Santesteban, 2004; Costa et al., 2006) or minimally changed (250 ms post-stimulus delayed naming; Martin et al., 2013), L2 proficiency is one of the key factors in the modulation of the language switching patterns. Here we aim to replicate these effects of L2 proficiency on the language switching patterns of bilinguals. Additionally, we further explore the role the cognate status of words might play in the language-switching behaviour of bilingual speakers.

3.  Cognate effects on code-switching and language switching Based on language contact and code-switching corpora analyses of several types of bilingual populations, Clyne (1967; 2003) suggested that bilinguals seem to be more prone to switch languages in the vicinity of cognate words, and proposed that cognate words might trigger language switching. Broersma and De Bot (2006) further developed this proposal and analyzed the data of a corpus containing conversations between three Dutch-Moroccan Arabic bilingual speakers. Their results showed that language



Chapter 4.  Are cognate words “special”? 

switching occurred more frequently for words that immediately followed a cognate word or that were part of the same clause as the cognate than for words that did not have a cognate word in their vicinity. Broersma, Isurin, Bultena and De Bot (2009) replicated these findings in their analyses of corpora of interviews to elderly DutchEnglish bilinguals from New Zealand and Australia and a corpus of an interview with one Russian-­English bilingual from the United States. These findings revealed that cognate words trigger language switching regardless of their grammatical class and regardless of the amount of lexical overlap (i.e., number of cognates) and the typological similarity of the languages spoken by the bilinguals. Broersma and De Bot (2006) further elaborated Clyne’s proposal, and hypothesized that, since cognate words share most of their phonological form, the production of a cognate word might increase the activation of lexical entries of the other language more than the production of a non-cognate word. Hence, when a bilingual is speaking in language A and produces a word that is a cognate with a lexical entry in language B, the level of activation of language B increases, which might trigger a switch from language A to language B. Moreover, triggering words might influence not only following words, but any words within the sentence. Thus, cognates might trigger language switching at words preceding or following their position, or even at their own position. This triggering was assumed to take place at the lexical (lemma) level of representation. Thus, assuming that the cognate effects arise at the phonological level,1 the non-target lexicon is assumed to

.  Although we are favouring a phonological origin of the cognate effects, there is still some controversy in the literature about the origin of cognate effects, as some authors have suggested that the cognate facilitation effect might arise at either the semantic or the lexical levels of representation. In the former case, Van Hell and De Groot (1998) suggested that cognate effects originate as a consequence of the different semantic overlap between translation words: cognate translations share more semantic features than non-cognate translations. In this scenario, the cognate facilitation effect may arise either because the retrieval of semantic representations shared across languages is easier than that of non-shared representations (since they are retrieved more often), or because access to lexical representations from a shared semantic representation is faster than from non-shared ones. In the latter case, some authors have suggested that cognate words share a single lexical representation, while non-cognate translations are represented by two different lexical entries (Kirsner, Lalor, & Hird, 1993; Sánchez-Casas & García-Albea, 2005). Thus, the retrieval of a cognate word entails the retrieval of the same lexical representation in L1 and L2, while the retrieval of a non-cognate entails the retrieval of distinct lexical entries. This proposal could be easily adapted to current models of speech production for identical cognates with full phonological/orthographic overlap (e.g., piano in English and Spanish; see Peeters, Dijkstra, & Grainger, 2013), but it cannot be so easily adapted for the (probably most frequent) cognate words with partial phonological overlap (e.g., trumpet-trompeta in English and Spanish). However, note that solving which is the origin of the cognate effects is out of the scope of the present study (for further discussions on this issue, see Costa et al., 2005).

 Mikel Santesteban & Albert Costa

be activated through feedback activation from the phonological to the lexical level of representation (Costa et al., 2005; Goldrick, 2014). The “triggering hypothesis” was experimentally tested for the first time by ­Broersma (2011). In two language-switching experiments, medium/high-proficient bilinguals were required to switch between naming pictures in their L1-Dutch and L2-English either following a language-cue or freely at will. Since phonologically identical cognates were used as triggers, each trial comprised the triggering cognate or non-cognate control word preceded and followed by two non-cognate words that were used to create the switch/non-switch conditions (e.g., L1-trigger-L2, etc.). When the language switching was cued, participants showed faster switching times following a cognate than a non-cognate, regardless of the switching direction. When participants switched freely, they switched more frequently following a cognate than a non-­ cognate, but only when switching from L1 to L2. Broersma (2011) concluded that these results confirm Broersma and De Bot’s (2006) “triggering hypothesis”. However, evidence from intra-sentential code-switching studies does not provide clear experimental support to the “triggering hypothesis”. In two studies, Bultena, Dijkstra and Van Hell (2015a, 2015b) explored whether the presence of a cognate verb facilitates the comprehension and production (shadowing) of intra-sentential language switching. Dutch-English bilinguals were tested with self-pace reading (Bultena et al., 2015a) and sentence shadowing (Bultena et al., 2015b) paradigms on which they were asked to read or shadow L1-to-L2 or L2-to-L1 code-switched sentences. Results showed that the cognate status of the verb did not modulate the costs of switching languages in any direction (neither in reading times or in shadowing latencies) at the following sentence position. These results provide negative evidence for the “triggering hypothesis”. However, evidence from syntactic priming suggests that cognate words can enhance intra-sentential codeswitching (Kootstra, Van Hell, & Dijkstra, 2012). These authors showed that Dutch-English high-proficient bilinguals (but not lowproficient ones) tended to produce more intra-sentential (Dutch to English) language switches at the same sentence position as in the prime sentences when sentences contained a cognate than a non-cognate word. Although their aim was not to test the “triggering hypothesis”, there are another two studies that explored the role of cognate status on language switching. In a language-switching task containing cognates and non-cognates, Christoffels et al. (2007) explored the performance of German learners of Dutch. Participants showed larger cognate facilitation effects in L1 than L2, reversing the pattern they showed in a blocked naming task (naming only in L1-German or L2-Dutch). However, the cognate status of the words did not affect their language switching pattern: despite being low-proficient bilinguals, maybe due to the fact that they were immersed in an L2-environment, participants showed symmetrical language switching costs both when naming cognates and non-cognates. Interestingly, against the predictions of the “triggering hypothesis”,



Chapter 4.  Are cognate words “special”? 

the language switching costs were larger when participants switched to name a cognate rather than a non-cognate word. Finally, in a study aiming to explore the role of preparation time (i.e., cue-stimulus intervals) in language switching patterns, Verhoef et al., (2009) tested low-proficient Dutch-English bilinguals. Results showed asymmetrical language switching costs for short intervals (larger for L1 than L2) and symmetrical for long intervals. Replicating Christoffels et al. (2007), participants showed larger cognate facilitation effects in L1 than L2 and the asymmetric/symmetric language switching patterns were not affected by the cognate status of target words. However, unlike Christoffels et al., (2007), cognate words, as compared to non-cognates, did not enhance (or facilitate) the overall switching costs. Overall these findings reveal mixed evidence about the “triggering hypothesis” proposed by Broersma & De Bot (2006). In a study with high-proficient bilinguals that manipulated the cognate status of preceding words, Broersma (2011) showed that previously named cognates might facilitate language switching; However, in two studies with low-proficient bilinguals that manipulated the cognate status of the ­target word, Christoffels et al. (2007) showed that naming cognates makes switching more costly; and Verhoef et al. (2009) revealed no effects. Here we aim to further test the “triggering hypothesis” proposed by Broersma & De Bot (2006), with the main novelty that we explore simultaneously the effects of cognate status of the preceding word and of the target word. Additionally, we explore the role L2 proficiency may play in the occurrence of language switching “triggering” effects by testing both low- and high-proficient bilinguals.

4.  Present Study In this experiment, we explore the language-switching performance of two groups of speakers: Spanish native speakers who are learners of Catalan (L2-Learners) and ­Spanish-Catalan highly proficient bilinguals (Bilinguals). On the one hand, we explore to what extent the language switching task can be used to explore the language selection mechanisms of bilinguals. For that purpose, we will compare the language switching performance of low- and high-proficient bilingual speakers using the same experimental design as in Costa and Santesteban (2004). At the same time, we also put to test the predictions of the “triggering hypothesis” proposed by Broersma and De Bot (2006). For that purpose, we manipulate the following main variables: (a) the cognate status of the preceding word before a language switch; and (b) the cognate status of the target word. Regarding the role of L2-proficiency on the language switching performance of bilinguals, we expect to replicate the main findings of Costa and Santesteban (2004): asymmetrical switching costs for L2-Learners and symmetrical ones for Bilinguals.

 Mikel Santesteban & Albert Costa

Regarding the “triggering hypothesis”, we expect smaller switching costs after naming pictures with cognate as compared to non-cognate names, and maybe also when naming target pictures with cognate rather than non-cognate names. For the sake of completeness, we explore these effects in both low- and high-proficient bilinguals.

4.1  Method 4.2  Participants Forty-eight native speakers of Spanish took part in the experiment. Twenty-four participants were highly proficient bilinguals (Bilinguals), and twenty-four were learners of Catalan (L2-Learners). All the bilingual participants had learned Spanish as L1 but acquired Catalan as their L2 at a very early age (they were schooled in Catalan at the school, high-school and university). L2-Learners had been learning Catalan for an average of 1 year by means of formal training, and had been living in Barcelona for an average of 2.5 years before the time of testing. Therefore, they were regularly exposed to Catalan and Spanish, but they were low-proficient in Catalan (see Appendix A).

4.3  Materials Twenty pictures of common objects, half with non-cognate and half with cognate names were selected. The same design and procedure used in Costa and Santesteban (2004) was used. Participants were instructed to choose the language of the response according to the colour of the picture (red or blue). The assignment of colour cue to response language was counterbalanced across participants. Half of the ­participants were instructed that “red” indicated “respond in Spanish” and “blue” indicated “respond in Catalan”, and the other half received the reverse assignments. Pictures were presented in short sequences (“lists”) ranging in length unpredictably from 5 to 14 trials. There were two types of trials: (a) trials in which the language of the response (either L1 or L2) was the same as the trial immediately before (non-switch trials); (b) trials in which the language of the response (either L1 or L2) was different from that used on the preceding trial (switch trials). In order to explore the effects of cognate status of both the target word and of its preceding word we should make sure that the same number of switch and non-switch, L1 and L2, and cognate and non-cognate target trials are preceded by the same number of cognate and non-cognate trials. However, if we were to manipulate all these properties for each participant, the number of trials per condition would be relatively small. Thus, we decided to control for the cognate status of the target trial and its preceding trial in a different manner. For half of the participants in each group, cognate and non-cognate target items were preceded by trials of the same type (no-change of cognate status context; see Figure 1). That is, a cognate target trial was always preceded by another cognate trial, and a non-cognate target trial was always preceded by

Chapter 4.  Are cognate words “special”? 

Trial 1

Trial 2

Trial 3

Trial 4

Trial 5

Trial 6

Trial 7

Trial 8

Trial 9

L2-NS

L1-S

Trial 10

Trial 11

Change of cognate status context Trial Condition

L2-Initial L2-NS

L1-S

L1-NS

L1-NS

L2-S

L2-NS

L2-S

L2-NS

Transition

C

CtoNC

NCtoC

CtoNC

NCtoC

CtoNC

NCtoNC NCtoNC NCtoC

CtoNC

NCtoC

Trial Analysis

Filler

EXPT

EXPT

EXPT

EXPT

EXPT

Filler

Filler

EXPT

EXPT

EXPT

L2-NS

L2-NS

L1-S

L2-S

L2-NS

No-change of cognate status context Trial Condition

L2-Initial L2-NS

L1-S

L1-NS

L1-NS

L2-S

Transition

C

CtoC

CtoC

CtoC

CtoNC

NCtoNC NCtoNC NCtoNC NCtoNC NCtoNC NCtoC

Trial Analysis

Filler

EXPT

EXPT

EXPT

Filler

EXPT

EXPT

EXPT

EXPT

EXPT

Filler

Figure 1. Example of an 11 trial list structure presented to participants in the no-change and change of cognate status contexts. L1 = L1 Response; L2 = L2 Response; NS = No-switch trial; S = Switch trial; NC = Non-cognate word; C = Cognate word; NCtoC; Cognate trial preceded by a Non-cognate trial; CtoNC; Non-cognate trial preceded by a Cognate trial; NCtoNC; Non-cognate trial preceded by a Non-cognate trial; CtoC; Cognate trial preceded by a Cognate trial; Filler = Filler trial; EXPT = Experimental trial.

 Mikel Santesteban & Albert Costa

a non-cognate trial. The first trials of a list and all those target trials in which there was a change in the cognate status were regarded as fillers. For the other half of the participants, the reverse situation was constructed. That is, a cognate target trial was always preceded by a non-cognate trial, and a non-cognate target trial was always preceded by a cognate trial (change of cognate status context). The first trials of a list and all target trials in which there was a repetition of cognate status were regarded as fillers. This design allows us to perform two different types of Cognate Status analyses: One on which we focus on the Cognate Status of the Target Word (regardless the cognate status of the preceding word) and another one on which we focus on the Cognate Status of the Preceding Word (regardless the cognate status of the target word). In each of these analyses, for each group, there were 8 experimental conditions as a result of crossing the following variables: Cognate Status of the Target or Preceding Word (non-cognate vs. cognate); Type of Trial (non-switch vs. switch); and Response Language (L1 vs. L2). Each experimental condition had 71 trials. Thus, there were 284 non-switch trials and 284 switch trials (divided in 4 experimental conditions: L1 and L2 non-cognate and L1 and L2 cognate conditions). The remaining 382 trials of the non-switch conditions were left as filler trials. Cognates and non-cognates were represented evenly for the filler trials. The lists varied in the number of switch trials (from 0 to 4 switching trials). To prevent participants from developing strategies, there were some lists in which no switch trials were included. A total of 400 such lists were constructed. These lists varied in the specific sequence of: (a) L1 and L2 responses; and (b) switch and non-switch trials. Each participant was presented with 100 of such lists that varied in length (from 5 to 14 trials). In total, each participant was presented with 950 trials (70% non-switch and 30% switch). On half of the non-switch trials, participants were asked to name the picture in L1 (333 trials), and in the other half, in L2 (333 trials). The same applied for the switch trials. Therefore, participants used their L1 and L2 the same number of times during the course of the experiment (475 responses in each language). For each subject, 50 of the 100 randomly distributed lists started with a picture to be named in L1, and 50, with a picture to be named in L2. Each picture was presented 47 or 48 times. No picture could appear twice in a list, and there were at least two different items between the first and second appearance of the same picture across lists. The assignment of each specific picture to each trial was left random, but since it varied for each participant, overall each picture was almost equally distributed across experimental conditions. That is, for one participant, a given picture could be named more often in L1 than in L2 and could appear more often in a switch trial than in a non-switch trial. However, for another participant this distribution changed, given that the assignment of the pictures to the different trial types varied randomly across participants.



Chapter 4.  Are cognate words “special”? 

4.4  Procedure Participants were tested individually in a soundproof room. They were asked to name the pictures as fast as possible while trying not to make errors. They were informed that the language in which a given picture had to be named was determined by the colour in which the picture appeared, and that in a given list of trials, pictures with different colours could be presented. Before the experiment proper, participants were familiarized with the name of the pictures in the two languages. A list of trials had the following structure: (1) a blue or red circle along with the word CATALÀ (Catalan) or ESPAÑOL (Spanish) written below was presented for 2000 ms (the colour cue was counterbalanced for language across participants). This circle (and the word) indicated the language in which the first picture of the list had to be named; (2) the first picture of a list appeared and remained on the screen for 2000 ms or until participant’s response; (3) there was a blank interval of 1150 ms; (4) the next picture was presented, and the cycle was repeated until the end of the list; (5) after the presentation of the last picture of the list an asterisk was presented for 1000 ms, signalling the end of the list; and the next list started. The experiment started with the presentation of 6 training lists.

4.5  Data scoring and analysis Three types of responses were scored as errors: (a) production of names that differed from those designated by the experimenter; (b) verbal disfluencies (stuttering, utterance repairs, and production of nonverbal sounds that triggered the voice key); (c) recording failures and naming latencies faster than 250 ms or exceeding 2 SD from a given participant’s mean. The results from the two groups of participants (L2-Learners and Bilinguals) were first analysed separately and later in a joined analysis. Two different types of analyses were performed: The Cognate Status of the Target Word analysis (cognate vs. non-cognate status of the target item, regardless of the cognate status of the preceding item); and the Cognate Status of the Preceding Word analysis (cognate vs. non-cognate status of the item preceding the target, regardless of the cognate status of the target). Accuracy and RT data were analyzed with mixed logit and linear mixed effects regression models, respectively, using the R software. The lmerTest package was used to calculate the Estimates, t-values (for LME), z-values (for GLME) and p-values. Models included the binomial accuracy (correct vs. error) or the log transformed RT as dependent variables, with an intercept and participants and items as random effects (Jaeger, 2008; Baayen, Davidson, & Bates, 2008). The models included the following variables and their interactions as fixed effects: Cognate Status of the Target/Preceding Word (cognate vs. non-cognate), Type of Trial (non-switch vs. switch), and Response

 Mikel Santesteban & Albert Costa

Language (L1 vs. L2).2 All predictors were sum coded and centered. The maximal random effects structures justified by model comparison in the joined analyses were used for the analyses of L2-Learners and Bilinguals (see Tables 2 and 4).

4.6  Results and discussion 4.6.1  Cognate Status of the Target Word analysis 4.6.1.1  L2-Learners.  Erroneous responses (6% of the trials) were excluded from the analyses. In the error analyses, the models did not include random slopes. Results showed a significant Type of Trial effect (β = .090, SE = .037, z-value = 2.435, p = .014), revealing that switch trials (6.5%) were more prone to errors than non-switch trials (5.5%). Moreover, a significant Type of Trial by Response Language interaction was found (β = –.076, SE = .037, z-value = –2.057, p = .039), revealing larger language switching effects in L1 (1.9% effect) than in L2 (0.2% effect). Finally, a significant Cognate Status of the Target Word by Response Language interaction was also found (β = –.073, SE = .037, z-value = –1.979, p = .047), revealing larger cognate facilitation effects (fewer errors naming cognates than non-cognates) in L2 than in L1 (effects of 1,5% vs. –0.1%, respectively). In the analysis of naming latencies, the main effects of Cognate Status of the ­Target Word, Type of Trial, and Response Language were significant (see Tables 1 and 2). Participants were faster naming cognate (736 ms) than non-cognate words (813 ms), non-switch (738 ms) than switch trials (810 ms), and L2 than L1 words (758 vs. 790 ms). The Type of Trial by Response Language interaction was also significant,

.  Our experimental design also manipulated the Change vs. No-change of cognate status context between participants. However, a model including a cognate status change Context (Change vs. No-change) and its interactions with the other variables showed no main effects of Context in any of the reported analyses. The only significant effect involving the Context variable was a Context by Cognate Status of preceding word interaction observed in both groups of participants. This interaction revealed reversed cognate status patterns for both Bilinguals (β = .039, SE = .007, z-value = 5.500, p < .001), and L2-Learners (β = .047, SE = .008, z-value = 5.667, p < .001): while participants in the No-change of cognate status context responded faster when the previous word was a cognate than a non-cognate (main effects of −60 and −78 ms for Bilinguals and L2-Learners, respectively), the participants in the Change of cognate status context responded slower when the previous word was a cognate than a non cognate (main effects of 64 and 77 ms for Bilinguals and L2-Learners). But note that, due to the structure of our experimental design, for trials on which the preceding word was a cognate, non-cognate targets were produced, and viceversa (see trials with CtoNC and NCtoC in Figure 1). Hence, if we considered the cognate status of the target words instead of the preceding words, the inhibitory cognate effects observed in the Change of cognate status context would be showing similar facilitatory effects as in the No-Change of cognate status context. For the sake of simplicity, we do not further mention this variable in the reported analyses.



Chapter 4.  Are cognate words “special”? 

indicating that to switch into L1 was harder than to switch into L2 (the mean switching cost was of 86 vs. 59 ms for L1 and L2, respectively). No other interactions were significant (all ps > .1), revealing both similar cognate facilitation effects in L1 and L2 (91 vs. 64 ms), and similar switching cost for cognates and non-cognates (69 vs. 76 ms). Table 1.  Mean Reaction Time (Mean) and Error Percentage (% Error) for the L2-Learner and Bilingual groups tested in the Experiment. The data of the non-switch and switch trials in both L1 (Spanish) and L2 (Catalan) are reported by means of the Cognate Status of the Target Word. COGNATE STATUS OF THE TARGET WORD Non-Cognate

Cognate

Response Language

Response Language

L1 (Spanish)

L2 (Catalan)

L1 (Spanish)

L2 (Catalan)

Mean

% Error

Mean

% Error

Mean

% Error

Mean

% Error

792

5.2

758

6.1

702

5.7

700

4.9

L2-Learners Non-Switch Switch

879

7.5

822

6.6

787

7.1

753

4.9

Switching Cost

 87

1.3

 63

.5

 85

1.4

 54

.0

Non-Switch

763

4.1

718

2.9

694

3.2

672

2.4

Switch

823

5.4

785

4.9

748

4.9

726

3.7

Switching Cost

 60

1.3

 67

2.0

 54

1.7

 54

1.3

Bilinguals

Table 2.  Linear mixed models for the overall analysis of the response times per Group, Cognate Status of the Target Word, Response Language, and Type of Trial conditions Predictor

Estimate

SE

6.579

.021

t-value

p

BOTH GROUPS: (Intercept) Group

301.299