Aspects of Multilingual Aphasia 9781847697554

Table of contents :
Contents
Contributors
Introduction
Part 1: Broad Considerations
1. The Study of Bilingual Aphasia: The Questions Addressed
2. Bilingual Aphasia: Neural Plasticity and Considerations for Recovery
Part 2: Assessment and Treatment
3. What Do We Know About Assessing Language Impairment in Bilingual Aphasia?
4. Morphological Assessment in Bilingual Aphasia: Compounding and the Language Nexus
5. The Clinical Management of Anomia in Bilingual Speakers of Spanish and English
6. Generalization in Bilingual Aphasia Treatment
7. Cross-Language Treatment Effects in Multilingual Aphasia
8. Language Deficits, Recovery Patterns and Effective Intervention in a Multilingual 16 Years Post-TBI
Part 3: Bilingual Language Phenomena
9. Bilingual Aphasia and Code-Switching: Representation and Control
10. Grammatical Category Deficits in Bilingual Aphasia
11. Language Choice in Bilingual Aphasia: Memory and Emotions
12. Acquired Dyslexia and Dysgraphia in Bilinguals Across Alphabetical and Non-Alphabetical Scripts
Part 4: Language Pairs
13. Morphosyntactic Features in the Spoken Language of Spanish-English Bilinguals with Aphasia
14. Non-Word Jargon Produced by a French-English Bilingual
15. Number-Processing Deficit in a Bilingual (Chinese-English) Speaker
16. A Case Study of a Bidialectal (African-American Vernacular English/Standard American English) Speaker with Agrammatism
Part 5: Cultural Context
17. Aphasia, Language and Culture: Arabs in the US
18. Towards Cultural Aphasiology: Contextual Models of Service Delivery in Aphasia
Index

Citation preview

Aspects of Multilingual Aphasia

COMMUNICATION DISORDERS ACROSS LANGUAGES Series Editors: Dr Nicole Müller and Dr Martin Ball, University of Louisiana at Lafayette, USA While the majority of work in communication disorders has focused on English, there has been a growing trend in recent years for the publication of information on languages other than English. However, much of this is scattered through a large number of journals in the field of speech pathology/ communication disorders, and therefore, not always readily available to the practitioner, researcher and student. It is the aim of this series to bring together into book form surveys of existing studies on specific languages, together with new materials for the language(s) in question. We also envisage a series of companion volumes dedicated to issues related to the crosslinguistic study of communication disorders. The series will not include English (as so much work is readily available), but will cover a wide number of other languages (usually separately, though sometimes two or more similar languages may be grouped together where warranted by the amount of published work currently available). We envisage being able to solicit volumes on languages such as Norwegian, Swedish, Finnish, German, Dutch, French, Italian, Spanish, Russian, Croatian, Japanese, Cantonese, Mandarin, Thai, North Indian languages in the UK context, Celtic languages, Arabic and Hebrew among others. Full details of all the books in this series and of all our other publications can be found on http://www.multilingual-matters.com, or by writing to Multilingual Matters, St Nicholas House, 31–34 High Street, Bristol BS1 2AW, UK

Aspects of Multilingual Aphasia Edited by Martin R. Gitterman, Mira Goral and Loraine K. Obler

MULTILINGUAL MATTERS Bristol • Buffalo • Toronto

Library of Congress Cataloging in Publication Data A catalog record for this book is available from the Library of Congress. Aspects of Multilingual Aphasia/Edited by Martin R. Gitterman, Mira Goral and Loraine K. Obler. Communication Disorders Across Languages: 8 Includes bibliographical references and index. 1. Aphasia. 2. Language disorders—Treatment. 3. Bilingualism. 4. Multilingualism. I. Gitterman, Martin R. II. Goral, Mira. III. Obler, Loraine K. RC425.A85 2012 616.85'52–dc23 2012009137 British Library Cataloguing in Publication Data A catalogue entry for this book is available from the British Library. ISBN-13: 978-1-84769-754-7 (hbk) Multilingual Matters UK: St Nicholas House, 31–34 High Street, Bristol BS1 2AW, UK. USA: UTP, 2250 Military Road, Tonawanda, NY 14150, USA. Canada: UTP, 5201 Dufferin Street, North York, Ontario M3H 5T8, Canada. Copyright © 2012 Martin R. Gitterman, Mira Goral, Loraine K. Obler and the authors of individual chapters. All rights reserved. No part of this work may be reproduced in any form or by any means without permission in writing from the publisher. The policy of Multilingual Matters/Channel View Publications is to use papers that are natural, renewable and recyclable products, made from wood grown in sustainable forests. In the manufacturing process of our books, and to further support our policy, preference is given to printers that have FSC and PEFC Chain of Custody certification. The FSC and/or PEFC logos will appear on those books where full certification has been granted to the printer concerned. Typeset by Techset Composition Ltd., Salisbury, UK. Printed and bound in Great Britain by the MPG Books Group.

To my mother, Leila, and the memory of my father, David Marty To my parents, Ruth and Yitzhak Goral Mira To my students who become my colleagues Loraine And to all the people with aphasia who have contributed time and effort to our projects over the years Marty Gitterman, Mira Goral and Loraine Obler

Contents

Contributors Introduction Martin R. Gitterman, Mira Goral and Loraine K. Obler

xi xix

Part 1: Broad Considerations 1

The Study of Bilingual Aphasia: The Questions Addressed Loraine K. Obler and Youngmi Park

2

Bilingual Aphasia: Neural Plasticity and Considerations for Recovery Daniel Adrover-Roig, Karine Marcotte, Lilian C. Scherer and Ana Inés Ansaldo

3

16

Part 2: Assessment and Treatment 3

What Do We Know About Assessing Language Impairment in Bilingual Aphasia? Swathi Kiran and Patricia M. Roberts

35

4

Morphological Assessment in Bilingual Aphasia: Compounding and the Language Nexus Gary Libben

51

5

The Clinical Management of Anomia in Bilingual Speakers of Spanish and English Maria L. Muñoz

69

6

Generalization in Bilingual Aphasia Treatment Kathryn Kohnert and Michael Peterson

vii

89

viii A spec t s of Mult ilingual Aphasia

7

Cross-Language Treatment Effects in Multilingual Aphasia Mira Goral

8

Language Deficits, Recovery Patterns and Effective Intervention in a Multilingual 16 Years Post-TBI Daly Sebastian, Usha Dalvi and Loraine K. Obler

106

122

Part 3: Bilingual Language Phenomena 9

Bilingual Aphasia and Code-Switching: Representation and Control Alessandra Riccardi

141

10 Grammatical Category Deficits in Bilingual Aphasia Yasmeen Faroqi-Shah

158

11 Language Choice in Bilingual Aphasia: Memory and Emotions Carmit Altman, Mali Gil and Joel Walters

171

12 Acquired Dyslexia and Dysgraphia in Bilinguals Across Alphabetical and Non-Alphabetical Scripts Maximiliano A. Wilson, Karima Kahlaoui and Brendan S. Weekes

187

Part 4: Language Pairs 13 Morphosyntactic Features in the Spoken Language of Spanish-English Bilinguals with Aphasia José G. Centeno 14 Non-Word Jargon Produced by a French-English Bilingual Nicole Müller and Zaneta Mok 15 Number-Processing Deficit in a Bilingual (Chinese-English) Speaker Nancy Eng 16 A Case Study of a Bidialectal (African-American Vernacular English/Standard American English) Speaker with Agrammatism Jean E. Jones, Martin R. Gitterman and Loraine K. Obler

207 224

242

257

Content s

ix

Part 5: Cultural Context 17 Aphasia, Language and Culture: Arabs in the US Reem Khamis-Dakwar and Karen Froud 18 Towards Cultural Aphasiology: Contextual Models of Service Delivery in Aphasia Claire Penn Index

275

292 307

Contributors

Daniel Adrover-Roig is a Professor of Language Pathology and Intervention at the Balearic Islands University (Spain). He specialized in psychology at the Universities of Würzburg (Germany), Salamanca (Spain) and the Universitat de les Illes Balears. During his PhD studies, he explored the neuroelectrical correlates of cognitive control in healthy aging using electroencephalography under the supervision of Dr Barceló. Since beginning his postdoctoral stay at the Institut Universitaire de Gériatrie de Montréal (Canada), supervised by Dr Ana Inés Ansaldo, he has been most interested in the influence of bilingualism on cognitive control and neurodegeneration, especially on the relationship between the bilingual mind and cognitive flexibility. Carmit Altman drafted this paper while she was a postdoctoral fellow at Lehman College and CUNY’s Graduate Center. Her PhD dissertation from Bar Ilan University examined English-Hebrew, Russian-Hebrew and Georgian-Hebrew immigrants in a Bilingual Autobiographical Memory paradigm and served as a basis for the chapter in this volume. She holds positions at Bar Ilan University (in the School of Education) and at Hadassah Academic College, Jerusalem, in the Bilingual Masters Program of the Communication Disorders Department. Her current research involves healthy bilingual aging and narrative analysis in aphasia and cross-language effects in maturing multilinguals. Ana Inés Ansaldo is a researcher and Assistant Professor in the Faculty of Medicine at the University of Montreal. She is also the Director of the Laboratoire de Plasticité Cérébrale, Communication et Vieillissement at the Institut Universitaire de Géraitrie de Montréal. Dr Ansaldo combines neuroimaging and behavioral approaches to disentangle the neural mechanisms underlying brain plasticity in normal and impaired language processing, studying monolingual and bilingual populations. Dr Ansaldo is funded by several national and international granting agencies and is a member of the Academy of Aphasia. José G. Centeno, PhD, CCC-SLP, is Department Chair and Associate Professor in the Department of Communication Sciences and Disorders, xi

xii A spec t s of Mult ilingual Aphasia

St John’s University, New York. Dr Centeno, a bilingual speech-language pathologist, studies the theoretical and professional bases that support the services for monolingual Spanish/bilingual Spanish-English speakers with aphasia, a large clinical population in the United States. Dr Centeno’s research areas include morphosyntactic expressive impairments in monolingual Spanish/bilingual Spanish-English speakers with aphasia, the clinical application of bilingualism principles to aphasia management in SpanishEnglish bilinguals, and issues in service delivery in culturally and linguistically diverse aphasia programs. Dr Centeno, a frequent invited presenter at national and international conferences, has participated in state, national and international professional committees on multicultural issues in communication disorders. Usha Dalvi is a Senior Lecturer and Head of the Department of Speech Language Pathology at Ali Yavar Jung National Institute for Hearing Handicapped, Mumbai, India. She has worked for over 35 years in various domains of communication disorders. She is the first recipient of the International Bridge School/International Society for Augmentative and Alternative Communication (ISAAC) ‘Teacher in Residence’ award. She has been funded by UNICEF and by the Science and Technology Mission Mode Project from the Government of India. Her core research areas are child language disorders, neurogenic speech and communication disorders, swallowing disorders and augmentative and alternative communication. She has published many national and international research papers in these areas. Nancy Eng, PhD, CCC-SLP, works with Chinese-English speakers who present with a range of communication disorders. She is currently an Associate Professor at Hunter College, The City University of New York and is on the doctoral faculty at the Graduate Center, CUNY. She teaches courses in language acquisition, psycholinguistics, bilingualism, as well as reading and dyslexia. Her research interests include language learning and language loss in both children and adults who are exposed to more than one language, reading ability and disability in biscript readers, neurogenic language loss in bilingual populations, and language attrition in internationally adopted children. She has presented lectures both nationally and internationally on these topics. Yasmeen Faroqi-Shah, PhD, CCC-SLP, is an Associate Professor in the Department of Hearing and Speech Sciences, University of Maryland. She received her doctorate from Northwestern University, Illinois, and her Bachelors and Masters degrees from the All India Institute of Speech and Hearing, India. Her research focuses on mechanisms of language production in speakers who have aphasia and/or who are bilingual. She also conducts research on the effects of language training on brain activity.

Contr ibutors

xiii

Karen Froud holds a PhD in theoretical linguistics from University College London, and is currently Associate Professor of Speech-Language Pathology and Neuroscience & Education at Teachers College, Columbia University. She is the director of the Neurocognition of Language Lab at Teachers College, where she conducts research into the neural correlates of multilingualism and communication disorders across the lifespan. She trained as a speechlanguage therapist in the UK, and has held research positions at the University of London and Massachusetts Institute of Technology (MIT). Dr Froud has developed expertise in training graduate students to work with multicultural populations and has founded the Transcultural SpeechLanguage Pathology Program in Cambodia. Mali Gil is head of the Speech and Language Department, Loewenstein Rehabilitation Center (Israel) with interests in aphasia, apraxia and most recently AAC (Augmentative and Alternative Communication) intervention. She has developed a language computerized program for intervention in aphasia and, with colleagues, has adapted the PALPA (Psycholinguistic Assessment of Language Processing in Aphasia) into Hebrew. In bilingual aphasia she has published work on cross-language effects of treatment. Martin R. Gitterman is Professor Emeritus of Speech-Language-Hearing Sciences at Lehman College and The Graduate Center, The City University of New York and is currently Ombuds Officer (Visiting Professor) at the Graduate Center. He served for six years as Chair of the Department of Speech and Theatre at Lehman College and six years as Executive Officer of the PhD Program in Speech-Language-Hearing Sciences at the Graduate Center. He was elected to membership in The Academy of Aphasia in 1997. He has published in the areas of neurolingusitics, aphasia, second language acquisition, bilingualism and applied linguistics. Mira Goral, PhD, CCC-SLP, is a Professor of Speech-Language-Hearing Sciences at Lehman College and the Graduate Center of The City University of New York. She also holds an appointment at the Harold Goodglass Aphasia Research Center of the Boston University School of Medicine. She completed her BA in Linguistics at Tel Aviv University and her PhD in Neurolinguistics at the Graduate Center, The City University of New York. She has published in the areas of multilingualism, aphasia, language attrtition and language and cognition in aging. She has been an invited speaker at international conferences and has received National Institutes of Health funding for her research. Jean E. Jones, PhD, CCC-SLP, TSSH, has 35 years of experience as a speechlanguage pathologist working with culturally and linguistically diverse populations. She received her PhD from the Graduate Center, The City University of New York. For the past 30 years she has worked as a consultant, therapist,

xiv A spec t s of Mult ilingual Aphasia

clinical supervisor and mentor. She has expertise in adult speech and language disabilities as well as in the area of preschool children with special needs. She has more than 25 years of experience as a clinical supervisor in mentoring and training undergraduate and graduate students seeking Clinical Competence Certification from the American Speech-LanguageHearing Association. Her special interest is in the area of African-American Vernacular English and bidialectalism, with a focus on enabling professionals and students to gain and maintain cultural competence in providing services. She is an active member of the New York State Speech-Language Hearing Association, the National Black Association for Speech-Language and Hearing and the American Speech-Language-Hearing Association. Karima Kahlaoui holds a PhD in Psychology from Nice University, France. She has completed a postdoctoral fellowship at the Centre de Recherche de l’Institut Universitaire de Gériatrie de Montréal, University of Montreal, Canada. Her research focuses on the semantic processing of words across hemispheres and semantic memory and its relationship with aging. In order to investigate these topics, Dr Kahlaoui uses behavioral methods, eventrelated potentials (ERPs) and near infrared spectroscopy (NIRS) techniques. She is currently working as a clinical neuropsychologist at the Institut Philippe-Pinel de Montréal and as Assistant Professor at the Speech Therapy School of the University of Montreal. Reem Khamis-Dakwar holds a PhD in Speech-Language-Pathology from Teachers College, Columbia University, and is currently Assistant Professor in the Department of Communication Sciences and Disorders at Adelphi University. Dr Khamis-Dakwar is the director of her Neurophysiology in Speech-Language Pathology Lab, where she conducts research into the neural correlates of linguistic processing and representation in specific sociolinguistic situations, such as in Arabic diglossia and understanding the neural correlates of functional changes related to speech-language pathology treatment and new learning. She is an expert in speech-language service provision for culturally and linguistically diverse populations, especially Arabic-speaking communities. Swathi Kiran, PhD, CCC-SLP, is an Associate Professor in the Department of Speech and Hearing Sciences at Boston University. Her research interests focus on lexical semantic treatment for individuals with aphasia, bilingual aphasia and neuroimaging of brain plasticity following a stroke. Her papers have appeared in journals such as Brain and Language, Aphasiology, Journal of Speech Language and Hearing Research and American Journal of Speech-Language Pathology. Her work is funded by the National Institutes of Health and American Speech Language Hearing Foundation awards, including the New Investigator Grant, the New Century Scholar’s Grant and more recently, the Clinical Research Grant.

Contr ibutors

xv

Kathryn Kohnert, PhD, CCC, is a Professor in Speech-Language-Hearing Sciences at the University of Minnesota, Minneapolis. She is also a bilingual speech-language pathologist and has been a certified member of the American Speech-Language-Hearing Association (ASHA) since 1992. Dr Kohnert’s research program is designed to advance understanding of language and cognition in diverse populations. Her publications include more than 70 articles, intervention curricula and review chapters as well as the recently published book Language Disorders in Bilingual Children and Adults. She is an ASHA Fellow, a past Associate Editor for the Journal of Speech, Language, and Hearing Research and the recipient of many research, teaching and service awards. Gary Libben’s research focuses on investigations of the representation and processing of multimorphemic words in the mind across populations and languages. His publications include The Representation and Processing of Compound Words (Libben & Jarema (eds), Oxford University Press, 2006), Core Perspectives on the Mental Lexicon (Jarema &Libben (eds), Elsevier, 2007) and The Nature of Language (Libben, Wiley Blackwell, forthcoming). He is a Fellow of the Royal Society of Canada, former President of the Canadian Linguistics Association and co-founder of the journal The Mental Lexicon. Libben was founding director of the University of Alberta’s Centre for Comparative Psycholinguistics and director of the project Words in the Mind, Words in the Brain. He is currently Professor and Vice President of Research at Brock University, Canada. Karine Marcotte graduated as a speech-language pathologist and began working as a clinician at Hôpital du Sacré-Cœur de Montréal with patients suffering from aphasia and other acquired communicative disorders. After three years as a clinician, she started her PhD in biomedical sciences at the University of Montreal and the Institut Universitaire de Gériatrie de Montréal, under the supervision of Dr Ana Inés Ansaldo. Her main focus of research as a doctoral student is the neural basis of recovery from chronic aphasia and primary progressive aphasia using fMRI, although she is also interested in second-language learning in healthy populations. Zaneta Mok is a PhD candidate at the University of Louisiana at Lafayette in the Applied Language and Speech Sciences Program. Her primary research interest is in the area of neurogenic communication disorders, including both aphasia and dementia, particularly in the context of culturally and linguistically diverse populations. Her research focuses on using clinical linguistics to describe the functional and interactional consequences of language disorders, and she has authored several publications in this area in national and international peer-reviewed journals and conferences. Nicole Müller is a Professor of Communicative Disorders at the University of Louisiana at Lafayette, and holds a Hawthorne Endowed Professorship.

xvi A spec t s of Mult ilingual Aphasia

Among her research interests are bilingualism and neurogenic and agingrelated impairments of cognition and communication, such as dementia and aphasia. She gained degrees at the Universities of Bonn (Germany) and Oxford (UK) and has been at University of Louisiana Lafayette since 2000. She co-edits (with Martin J. Ball and Thomas W. Powell) the international journal Clinical Linguistics and Phonetics and (with Martin J. Ball) the book series Communication Disorders Across Languages (Multilingual Matters). Maria L. Muñoz, PhD, CCC-SLP, is an Associate Professor and Coordinator of the Emphasis in Bilingual Speech-Language Pathology at Texas Christian University. She received her doctorate from the University of Texas as a participant in the Multicultural Leadership Training Program and completed a post-doctoral fellowship at the University of Arizona. She teaches and supervises in the areas of neurogenic disorders of cognition and language and multicultural issues in communication sciences and disorders. She conducts research and publishes on topics related to pedagogy, treatment outcomes and aphasia, and aphasia in Spanish/English bilinguals. Loraine K. Obler, PhD, is a Distinguished Professor at the CUNY Graduate Center, with appointments in both Speech-Language-Hearing Sciences and Linguistics. As well, she consults to her Language in the Aging Brain Lab at the Harold Goodglass Aphasia Research Center of the Boston University School of Medicine and to the Bilingual SLP MA program at Hadassah Academic College, Jerusalem where she heads the Steering Committee. She has co-authored articles and books on her areas of interest: neurolinguistics, bilingualism and the brain, cross-language study of aphasia and language in aging. Her books include: The Bilingual Brain: Neuropsychological and Neurolinguistic Aspects of Bilingualism (with Martin Albert, Academic Press, 1978), Bilingualism Across the Lifespan: Aspects of Acquisition, Maturity, and Loss (with Kenneth Hyltenstam, Cambridge University Press, 1989) and Clinical Communication Studies in Spanish Speakers: From Research to Clinical Practice (with José Centeno and Raquel Anderson, Multilingual Matters, 2007). Youngmi Park is a doctoral student in the Program in Speech-LanguageHearing Sciences at the Graduate Center, The City University of New York. She is a speech-language pathologist who has worked with people with neurogenic communication disorders and dysphagia. She also teaches undergraduate courses and has supervised student clinicians at Queens and Lehman Colleges, CUNY. Her primary research interests include sentence comprehension, bilingualism, agrammatism and the efficacy of aphasia therapy. Specifically, Ms Park is interested in the production of different verb types in people with non-fluent aphasia and changes of verb production patterns after aphasia therapy. She is also researching effects of syntactic complexity and

Contr ibutors

xvii

temporal ambiguity on sentence comprehension in normal aging and people with neurogenic communication disorders using eye-tracking methods. Claire Penn, PhD, is a researcher working at the University of Witwatersrand in Johannesburg, South Africa. She currently holds the position of Chair of the Department of Speech Pathology and Audiology and directs the Health Communication Research Unit which studies intercultural interactions in health care contexts. She is a member of the aphasia committee of the International Association of Logopedics and Phoniatrics and serves on the editorial board of Aphasiology. She has travelled and taught in a variety of countries and remains actively involved with clinical work. Michael Peterson is a graduate student in the professional MA Program in Speech-Language-Hearing Sciences at the University of Minnesota, Minneapolis. His background in cognitive psychology and linguistics led to his focus on neurogenic communication disorders and bilingual issues in speech-language pathology. As a speech-language pathologist, his primary interests are in the assessment and treatment of acquired communication disorders in diverse populations. Alessandra Riccardi holds a PhD in Linguistics from the University of Perugia, Italy (2004). She completed postdoctoral training at the Centre de Recherche de Gériatrie de l’Université de Montréal and the Center for Comparative Psycholinguistics of the University of Alberta (Canada). Her research interests focus on bilingual aphasia, language selection and control in multilinguals, and psycholinguistic aspects of morphological processing. Patricia M. Roberts, PhD, SLP(C), CCC-SLP, is an Associate Professor in the School of Rehabilitation Sciences at the University of Ottawa (Canada). She obtained her PhD from the Université de Montréal and has also studied in the United States and France. Her research interests stem directly from her prior clinical career as a speech-language pathologist and include the impact of bilingualism on many aspects of assessment and treatment; treatment efficacy in aphasia, naming and how best to assess it, the verbal fluency task and the nature of semantic categories. Lilian Cristine Scherer is an Associate Professor at the Pontifical Catholic University of Rio Grande do Sul (PUCRS) in the Language and Linguistics Department. She completed her PhD in Medicine at the University of Montreal and the Centre de Recherche de l’Institut Universitaire de Gériatrie de Montréal on Neuropsycholinguistics. She conducted her research on the cerebral correlates of aging and bilingualism on syntax and discourse processing. Her research interests are bilingualism, aging, second-language acquisition and the adaptation and elaboration of language assessment tools.

xviii A spec t s of Mult ilingual Aphasia

Daly Sebastian is an emerging researcher in the study of multilingual aphasia. She is a postgraduate student in Audiology and Speech Language Pathology at Ali Yavar Jung National Institute for Hearing Handicapped (AYJNIHH), Mumbai under Maharashtra University of Health Sciences (MUHS), India. She completed her Bachelors at Marthoma College of Special Education, India. Her key research areas include evidence-based practices in neurogenic communication disorders, cross-linguistic studies in communication disorders, motor speech disorders, neurogenic dysphagia, auditory processing disorders and speech perception. Joel Walters is Dean, Faculty of Humanities at Bar-Ilan University. He works on the interface of psycholinguistic and sociolinguistic aspects of bilingualism in early childhood and in mature adults with and without language impairments. His book, Bilingualism: The Sociopragmatic-Psycholinguistic Interface was published by Erlbaum/Routledge in 2005. Brendan Weekes is Chair Professor in Communication Science at the University of Hong Kong. He is an experimental psychologist and investigates the cognitive processes used for communication. His work is focused on the memory mechanisms that are used in normal and impaired language processing (aphasia). He uses cognitive models of language processing to guide his research in the areas of reading, writing and speech production and uses an experimental approach to study cognitive processes including behavioral, brain imaging and cross-linguistic methods. Professor Weekes is also a clinical psychologist and he carries out research with adults and children who have clinical communication disorders caused by dementia, dyslexia and poor reading comprehension in English, Chinese and other languages. These projects focus on cognitive aspects of communication disorders and how to treat impaired cognitive processes using behavioral and imaging methods. His research on communication disorders is part of a larger program of work that links research with clinical practice in Communication Science. He is Director of the Communication Science and Disorders Laboratory at the University of Hong Kong. Maximiliano A. Wilson is a postdoctoral fellow at the Centre de Recherche de l’Institut Universitaire de Gériatrie de Montréal, University of Montreal, Canada. He obtained his PhD in Health Sciences at the University of Buenos Aires, Argentina. He is interested in the study of lexical and semantic processing in normal adults and neuropsychological populations, such as monolingual and bilingual patients with aphasia and individuals with semantic dementia. He is particularly interested in the involvement of the anterior temporal lobes in exception word reading. He uses behavioral and brain imaging (fMRI) techniques to study normal and impaired language processing. Dr Wilson is Assistant Professor at the Speech Therapy School of the University of Montreal.

Introduction Martin R. Gitterman, Mira Goral and Loraine K. Obler

Planning this book on bilingual/multilingual aphasia required making some fundamental decisions. First, it should be noted that the terms multilingual and bilingual are being used interchangeably here to refer to more than one language. Thus, the title of the book, Aspects of Multilingual Aphasia, does not necessarily imply reference to more than two languages, although, of course, it does not exclude such cases. One decision to be made was whether to focus exclusively on aphasia in individuals who speak more than one language, or whether to include cross-language studies of aphasia in monolingual individuals pertinent to bilingualism. Despite our interests in the cross-language study of aphasia (e.g. Menn & Obler, 1990) we decided to limit our focus to multilinguals with aphasia in this volume because of the unique knowledge that can be gained from such studies and since there was more than enough new work to warrant that scope. Apart from that decision, we decided that this book should not be very narrowly focused. Rather, the scope would encompass chapters with diverse foci, as is made evident by the division of chapters into five separate, although somewhat overlapping, sections. Books with a narrower focus (e.g. Centeno et al., 2007, with a focus on speakers of Spanish; Paradis, 1987, devoted to assessment) have also been valuable additions to the literature on aphasia. It is our hope that Aspects of Multilingual Aphasia provides researchers, clinicians and other interested professionals with a thought-provoking and broad-based look at current work in aphasia in individuals with knowledge of two or more languages. The chapters in this volume include both theoretical overviews and experimental research. While the collection focuses on current research, some historical perspective is also provided. Many issues at the forefront of research some 40 years ago, to select just one historical period, such as the hypothesized increased role of the right hemisphere in organizing language in bilinguals, have been replaced by ‘more fine-grained questions’ (as noted

xix

xx A spec t s of Mult ilingual Aphasia

by Obler and Park in Chapter 1 of this volume). It is certainly the case that some of the widely cited works published at that time (e.g. Albert & Obler, 1978) have played a role in motivating research leading to many of the more recent findings. In recent years research in multilingual aphasia has continued to benefit greatly from advances in neuroimaging, which is certain to play an ever-increasing role in the years ahead. The breadth of material in this volume is one of its most evident attributes. Included is the assessment and treatment of multilinguals with aphasia (Chapters 3–8). Fabbro (1999: 54) aptly asserts, ‘The end of the 1970s saw renewed interest in the study of bilingual and polyglot aphasia and, at the same time, a raising of awareness that bilingual aphasia could no longer be assessed as naïvely and approximately as had been done from the end of the 19th century up to 1950.’ Recent work, reflecting this growing concern with appropriate assessment, is found in this volume, as is current work on treatment, including, for example, the study of generalization (i.e. carry-over effects) from treatment in one language (the treated one) to another language (the untreated one) (e.g. Chapter 6). Readers will also find research on a range of languages (and language pairs, Chapters 13–16) as well as an examination of bilingual aphasia from diverse perspectives (Chapters 9–12). Included, for example, is a case study of a bidialectal participant (Chapter 16). The study of the new field of bidialectalism (Siegel, 2010) is seen as an integral (and essential) component of research aimed at a comprehensive understanding of multilingual aphasia. Of note, Menn et al. (1995: 3) state, ‘Bi-dialectal speakers are a majority in some areas of most countries, for regional dialects of the standard national languages have histories going back hundreds of years and are of immense social and political importance.’ Also highlighted in this volume is the critical role of cultural awareness in understanding the dynamic of multilingual aphasia (Chapters 17 and 18). The level of importance of cultural context is widely recognized by both researchers and clinicians. Holland and Penn (1995: 152), for example, in reference to therapy, state, ‘Cultural factors are at least as important as linguistic factors in cross-linguistic therapy, if not more so.’ The chapters in this volume focusing on cultural factors provide details supportive of this assertion. The book is divided into five sections, each designating a particular topic within the study of multilingual aphasia. In Part 1 (‘Broad Considerations’) the chapters provide a fundamental overview of issues in multilingual aphasia. These chapters serve as a useful introduction to the issues covered in the remainder of the book. Obler and Park (Chapter 1, ‘The Study of Bilingual Aphasia: The Questions Addressed’) provide a chronologically organized historical perspective looking at questions that guided research on bilingual aphasia dating back to the 19th century. The emphasis is on current research (e.g. language switching, cross-language generalization following therapy). Questions likely to be addressed in the years ahead are also noted. AdroverRoig, Marcotte, Scherer and Ansaldo (Chapter 2, ‘Bilingual Aphasia: Neural

Introduc t ion

xxi

Plasticity and Considerations for Recovery’) cover a range of issues (language organization in the brain, language switching, recovery patterns, assessment and treatment). They aim to show the interrelationship among these areas, noting relevant, though not always consistent, research findings. Looking to the future, the authors advocate research on the application of neuroimaging in therapy programs. In Part 2 (‘Assessment and Treatment’) the chapters focus on diverse aspects of clinical intervention and the closely linked phenomenon of assessment. The chapters, while clearly relevant to the clinician, are not aimed at providing specific guidelines for the practicing clinician. Rather, they are intended to highlight what current research suggests about both assessment and treatment and, by extension, where future research is needed. Kiran and Roberts (Chapter 3, ‘What Do We Know About Assessing Language Impairment in Bilingual Aphasia?’) emphasize the complexities involved in assessment. These complexities are outlined with reference to proficiency ratings (by both patients and their family members) and to the type of information needed in order to ascertain (with reasonable confidence) pre-stroke proficiency in given languages. While recognizing the value of standard tests in assessing bilinguals with aphasia, they also report on the promise of other assessment tests in measuring patient performance and progress during recovery. Libben (Chapter 4, ‘Morphological Assessment in Bilingual Aphasia: Compounding and the Language Nexus’) espouses the merits of appropriate assessment of language performance in the area of morphology. He describes the components of assessment tests designed (and continuing to be designed) in various languages as part of the Morphological Assessment Project, suggesting that such assessment tests will have added value as a performance measure in cases where morphological therapy is used in treating bilinguals with aphasia. Muñoz (Chapter 5, ‘The Clinical Management of Anomia in Bilingual Speakers of Spanish and English’) addresses lexical retrieval from a theoretical perspective in order to establish a framework for the discussion of assessment and treatment of anomia. An extensive list (with explanatory details) of assessment measures of lexical access is provided. Kohnert and Peterson (Chapter 6, ‘Generalization in Bilingual Aphasia Treatment’) discuss research on generalization following treatment in multilinguals with aphasia, indicating that potential impairment in overall cognitive performance (including in non-linguistic domains) is an important consideration in designing a treatment plan for multilinguals with aphasia. Both crossdomain and cross-language generalization are noted as components of an ideal treatment plan. Goral (Chapter 7, ‘Cross-Language Treatment Effects in Multilingual Aphasia’) presents four studies, each focusing on the relationship between language proficiency and cross-language generalization subsequent to treatment in a multilingual individual with aphasia. The findings are examined in the context of current thought in neurolinguistics and psycholinguistics. Clinical implications are noted as well. Sebastian, Dalvi and

xxii A spec t s of Mult ilingual Aphasia

Obler (Chapter 8, ‘Language Deficits, Recovery Patterns and Effective Intervention in a Multilingual 16 Years Post-TBI’) describe a case study of a multilingual participant who received treatment for aphasia (resulting from an accident at age 16) in his L1 (Marathi) and L2 (English). Treatment (including Melodic Intonation Therapy) which started 16 years after the participant’s TBI is shown, nevertheless, to lead to substantial improvement. In Part 3 (‘Bilingual Language Phenomena’) a particular aspect of language performance (e.g. code-switching) is central in each chapter. The phenomena included in Part 3 are not necessarily restricted to performance in bilinguals, but, for our purposes, they are discussed in the context of bilinguals, and particularly bilinguals with aphasia. Riccardi (Chapter 9, ‘Bilingual Aphasia and Code-Switching: Representation and Control’) examines the phenomenon of code-switching, which is evident in both bilinguals with aphasia (including, in some cases, pathological switching) and healthy bilinguals. Reference is made to competing theories and models as well as to neuroimaging research in an attempt to explain the underlying dynamic of code-switching and its application to bilinguals with aphasia. Faroqi-Shah (Chapter 10, ‘Grammatical Category Deficits in Bilingual Aphasia’) reports on studies of the noun-verb dissociation (with greater difficulty in the retrieval of verbs) in both monolinguals and bilinguals (in both of their languages) with aphasia. Some findings suggest that the difficulty with verbs (relative to nouns) might be greater in bilinguals than in monolinguals. Faroqi-Shah argues that findings must, at this point, be considered far from definitive. Altman, Gil and Walters (Chapter 11, ‘Language Choice in Bilingual Aphasia: Memory and Emotions’) examine the interplay between language and aspects of autobiographical memory. The participants in their study, healthy Hebrew-English bilinguals and a Hebrew-English bilingual with aphasia, manifest both marked differences and areas of consistent performance, depending on the task. Wilson, Kahlaoui and Weekes (Chapter 12, ‘Acquired Dyslexia and Dysgraphia in Bilinguals Across Alphabetical and Non-Alphabetical Scripts’) examine the possible influence of type of script on language processing. With a focus on dyslexia and dysgraphia in bilinguals, the chapter highlights the difficulties encountered in drawing generalizations as well as in using current theoretical models as a framework for explaining these phenomena. In Part 4 (‘Language Pairs’) each chapter uses two specific languages (or dialects) as the basis for addressing the topic of aphasia. Centeno (Chapter 13, ‘Morphosyntactic Features in the Spoken Language of Spanish-English Bilinguals with Aphasia’) provides an in-depth explanation of the morphosyntactic structure of both Spanish and English. Details based on the limited amount of information currently available on the characteristics of agrammatism in Spanish-English bilinguals are presented. Noted, for example, are similarities in their discourse and that found in monolinguals with agrammatism. Müller and Mok (Chapter 14, ‘Non-Word Jargon Produced by a

Introduc t ion

xxiii

French-English Bilingual’) analyze the utterances of a French-English bilingual with aphasia, whose output consists largely of non-word jargon. Sounds and sound sequences/patterns produced by the participant are outlined and contrasted with what is found in the French and English sound inventory. Future research is encouraged as a follow-up to this descriptive analysis. Eng (Chapter 15, ‘Number-Processing Deficit in a Bilingual (Chinese-English) Speaker’) studies performance in a Chinese (Cantonese)-English bilingual participant with aphasia, with a focus on number and non-number words. Although dissociation is found between performance on these two wordtypes, performance on number words in both languages is similar, suggesting a deep semantic base for numbers across the two languages. Jones, Gitterman and Obler (Chapter 16, ‘A Case Study of a Bidialectal (AfricanAmerican Vernacular English/Standard American English) Speaker with Agrammatism’) report a case study of a bidialectal participant with agrammatism. A method is outlined for determining which utterances are truly agrammatic in this speaker of two structurally similar, yet clearly independent, dialects. The importance of dialect awareness in assessing patients with aphasia is made evident. In Part 5 (‘Cultural Context’) the chapters focus on the critical role of cultural factors in gaining a comprehensive understanding of patients with aphasia. That is, individuals with aphasia must be seen as part of a particular cultural context, not solely as isolated individuals with a particular medical condition. Khamis-Dakwar and Froud (Chapter 17, ‘Aphasia, Language, and Culture: Arabs in the US’) discuss the link between culture and the delivery of effective clinical services to clients with aphasia in the Arab-American community (noted to be a heterogeneous community). Socio-political factors are also addressed. Improved assessment of individuals with aphasia in the ArabAmerican community, it is argued, necessitates the development of more appropriate linguistic measures. Penn (Chapter 18, ‘Towards Cultural Aphasiology: Contextual Models of Service Delivery in Aphasia’) stresses the need for greater attention to cultural factors in working with patients with aphasia. Areas where an increased focus on cultural context would be useful include considerations for which language to use in therapy and expectations for possible cross-language generalization/transfer effects following treatment. While using South Africa as a primary reference point in the chapter, Penn emphasizes that the issues raised have relevance of a more universal nature. Aspects of Multilingual Aphasia, while suggesting answers to a number questions raised by researchers and clinicians, also makes clear the many issues in need of additional research. Such is the desired outcome of research in this field, and others as well. The study of aphasia in multilingual individuals is a rapidly evolving field. This book, with contributions from many of the leading researchers in aphasiology, is aimed at providing a substantive snapshot on a range of issues in multilingual aphasia at this point in its evolution.

xxiv A spec t s of Mult ilingual Aphasia

Acknowledgements A book like this requires substantial work. We thank our colleagues at Multilingual Matters, all the contributors to the volume, and, for help with the index, Eve Higby, Jungna Kim, Anmo Kim, and JungMoon Hyun.

References Albert, M.L. and Obler, L.K. (1978) The Bilingual Brain: Neuropsychological and Neurolinguistic Aspects of Bilingualism. Orlando, FL: Academic Press. Centeno, J.G., Anderson, R.T. and Obler, L.K. (eds) (2007) Communication Disorders in Spanish Speakers: Theoretical, Research and Clinical Aspects. Clevedon: Multilingual Matters. Fabbro, F. (1999) The Neurolinguistics of Bilingualism: An Introduction. Hove: Psychology Press. Holland, A.L. and Penn, C. (1995) Inventing therapy for aphasia. In L. Menn, M. O’Connor, L.K. Obler and A. Holland (eds) Non-Fluent Aphasia in a Multilingual World (pp. 144–155). Amsterdam: John Benjamins. Menn, L. and Obler, L.K. (eds) (1990) Agrammatic Aphasia: A Cross-Language Narrative Sourcebook (3 Volumes). Amsterdam: John Benjamins. Menn, L., O’Connor, M., Obler, L.K. and Holland, A. (1995) Non-Fluent Aphasia in a Multilingual World. Amsterdam: John Benjamins. Paradis, M. (1987) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum. Siegel, J. (2010) Second Dialect Acquisition. Cambridge: Cambridge University Press.

Part 1 Broad Considerations

1

The Study of Bilingual Aphasia: The Questions Addressed Loraine K. Obler and Youngmi Park

Introduction Educated 19th-century Europeans were well aware that Europe was home to many multilinguals as well as monolinguals. Of course the early debates about language organization in the brain that observation of aphasia patterns permitted intended to address the questions of universals of human language representation. Thus it followed that questions about language patterns in bilinguals and multilinguals would be among the issues that scholars addressed. In this chapter, we consider the questions that have been asked concerning bilingual and multilingual aphasia from that period to the present. The first focus of interest concerned the question of which language returned first, and whether it was the first or most-used language (Freud, 1891; Pitres, 1895). Such reports provided the literature with interesting individual case studies and series of case studies well into the 20th century. As understanding of lateral dominance for language grew in the mid20th century, the question arose of whether multilingualism resulted in more crossed aphasia (aphasia in a right-hander resulting from a righthemisphere lesion) than was seen in monolinguals with aphasia (e.g. Gloning & Gloning, 1965). By the late 1970s, more fine-grained questions were being asked about the patterns of impairments across the languages of multilingual individuals with aphasia, the patterns of recovery from their aphasia and the factors that determined these (Obler & Albert, 1977; Paradis, 1977). With the expansion of treatments for aphasia in the later 20th century came a set of studies of the effectiveness of such treatments in the languages of the multilingual (Fredman, 1975; and Watamori & Sasanuma, 1978, among the first). These have led to a recent series of more experimental 3

4

Par t 1: Broad Considerat ions

studies on the topic in the 21st century, some of which also ask whether enhanced brain activation can be seen post-therapy. In what follows we review the chronology of questions asked in the research on multilingual aphasia, considering the implicit as well as explicit questions addressed in each era. We focus at somewhat greater length on the more recent studies that have been less frequently reviewed, concluding with those questions that may be on the cutting edge of our topic.

Studies of Recovery Patterns In this section we cover the questions of the late 19th century and first threequarters of the 20th century concerning which language returns first, what patterns of impairment and recovery are seen, and where in the brain a mechanism for switching between languages can be found. The general rule for impairment patterns in aphasia is that there is the same kind of aphasia in all the languages, and the degree of impairment is proportional to the degree of proficiency the patient had in each language prior to the aphasia-producing incident. However, today, as in prior centuries, there have been cases where this rule did not hold, and these are the ones that provoked the question of what determines recovery patterns. While there must have been substantial numbers of multilinguals whose aphasia impairment and recovery were unremarkable, we now have a literature of over 200 cases where differential recovery was reported. The two explanations that dominated the first 80 years of studying this question were whether the first language would return first, as Ribot predicted in his book Disease of Memory (1881) and Freud asserted in his book On Aphasia (1891). The alternative that Pitres proposed was that the language that had been used most around the time of the accident was the one most likely to recover better (Pitres, 1895). At one point it appeared that Pitres’ rule held with greater than chance accuracy, at least in individuals with onset of aphasia before age 60 (Obler & Albert, 1977), but in a later analysis of additional cases the picture became murkier. So, it would seem that the answer to that question is that neither the rules of Pitres nor of Ribot hold with greater than chance accuracy (Obler & Mahecha, 1991). However, in the early–mid 20th century, a new alternative was added to the question, asking to what extent emotional attitudes to one’s languages played a role in predicting relative impairment. Authors such as Minkowski (1963) and Krapf (1955) reported single-case studies to make the point that an unusual pattern of recovery appeared to be linked to particular emotional attachments or dislikes. Only with Paradis’ (1977) taxonomy did the questions concerning differential impairment and recovery take a step back from simplistic causal ones to consider the complexity of impairment patterns as they play out over the course of recovery. Much more detailed analyses of language patterns

The Study of Bilingual Aphasia : The Quest ions Addressed

5

emerged as linguists and speech-language pathologists joined neurologists in the study of multilingual aphasia. The questions became ones about what specific pattern was seen among the six patterns that Paradis detailed in his 1977 article and others that he reported with colleagues in later articles (e.g. Paradis, 2001). In Fabbro (1999), we see that these have led to the further classification of interesting phenomena in the field (e.g. recovery of a neverpreviously spoken language such as Classical Greek, or of a language previously used only for religious purposes, such as Latin). Fabbro (1999) also briefly addresses questions such as what happens in bilingual aphasia in children.

Lateral Dominance in Multilinguals Having addressed these two localizationist questions, the next we will cover is whether multilinguals have more right-hemisphere involvement in language than monolinguals. By the mid-20th century, lateral dominance was a major focus of study in the field of neurolinguistics. Two techniques, although imperfect, had evolved to study it non-invasively in healthy individuals: tachistoscopic visual recognition and dichotic-listening techniques. This may have been what prompted Gloning and Gloning (1965) to write about the small series of cases they saw in Austria of multilinguals who had aphasia resulting strictly from right-hemisphere lesions, a phenomenon that the previous literature – also virtually entirely from Europe, one must note – had reported was extremely rare. A few follow-up studies found the question interesting, and reported a small series of such crossed aphasia. However, when consecutive series of patients coming to a neurology clinic were systematically studied, the number of instances of crossed aphasia proved not to be disproportionately large in multilinguals (Charlton, 1964; April & Tse, 1977). Nevertheless, the question of additional right-hemisphere involvement in multilinguals, either for language generally, or for first or later-learned, more or less proficient second languages (L2s), remains unresolved. Indeed, the level of analysis employed in these early lateral dominance studies is quite coarse relative to the greater specificity of lesion-site or processing-site that imaging techniques afforded in the late 20th and early 21st centuries.

The Language Switch: Where Is It? One knows that healthy multilinguals are able to speak in the language of those they are speaking with, switching languages or avoiding doing so when appropriate. The ability to do so must lie in the brain, so by the logic of neurolinguistics, one has only to look at a group of patients with problems limited to inappropriate switching in order to discover what region is

6

Par t 1: Broad Considerat ions

involved in switching. The only problem with this logic arose when the cases that were reported showed no overlap whatsoever in lesion location. Parietal, temporal and frontal lobe sites have been linked to the phenomenon in the earlier literature (e.g. Fabbro et al., 1997). A recent functional neuroimaging study has shown crucial brain areas for language switching by asking what regions underlie the behavior of people with aphasia who show what we term ‘inappropriate language switching’, that is, what Abutalebi et al. (2009: 143) term ‘interference from the non-target language’ that occurred during spontaneous speech and naming tasks.1 Abutalebi et al. (2009) provided language therapy to a bilingual (L1: Spanish/L2: Italian) with aphasia due to a cerebral hemorrhage affecting the subcortical area (the left lenticular nucleus and surrounding areas) at three months post-onset. Before the beginning of the treatment, L1 (Spanish) was better than L2 (Italian) during spontaneous speech and naming tasks. The language treatment was provided in L2 (Italian). The first treatment phase lasted for six weeks and this treatment focused on naming which was deficit-specific. The second four-week treatment phase which focused on improving global language abilities then followed. Improvement of the treated language (L2, Italian) was observed and performance of L2 became better than that of L1 after the first treatment phase, and, finally, L2 naming skills became normal at the end of the second phase. At the beginning of the treatment, between-language errors (inappropriate language switching) from L1 (the better language, Spanish) to L2 (Italian) were often observed when testing L2, but switching from L2 to L1 was not observed during L1 testing. However, with improving L2 naming skills, after the two treatment phases, such inappropriate switching became uncommon when naming pictures in L2 (Italian, the treated language). However, the participant unintentionally did this form of switching by naming in L2 when required to name in L1 (Spanish, the untreated language). This behavioral result, wherein a shift in the inappropriate switching was linked to language improvement, was observed in their functional magnetic resonance imaging (fMRI) data as well. Before the treatment phases began, the pattern of brain activity was similar for L1 and L2. However, neural reorganization was found only in L2 after language treatments. When the participant performed better in L1 than L2, brain regions associated with language control, monitoring and detecting potential errors, such as the prefrontal cortex, the caudate and the anterior cingulate cortex (Abutalebi & Green, 2007, 2008; Crinion et al., 2006), were activated during L1 production. By contrast, activation of only the anterior cingulate cortex was observed during L2 production. However, after the first treatment phase, although the left caudate was still activated in L1, no other areas were activated. In contrast, activation of the prefrontal cortex, the caudate and the anterior cingulate cortex was now observed for L2. Then, no further activation of these language control areas was found for L1 after the end of the treatment

The Study of Bilingual Aphasia : The Quest ions Addressed

7

phases. Abutalebi et al. (2009) postulate that this functional shift of control areas from L1 to L2 resulted in the different language-switching pattern evident in the shifting recovery pattern of the patient’s two languages. While the phenomenon of switching still remains of great interest, it is fair to say that the aphasia literature has not resolved it, nor have those undertaking neuroimaging with non-brain-damaged individuals resolved it either (but see Abutalebi & Green, 2007, 2008; Hernandez et al., 2000; Hernandez et al., 2001). This is most likely due to the fact that not a region but a diffuse system – probably all or part of the system governing executive functions – is involved in giving the multilingual the ability to appropriately address interlocutors. It may, however, also be due to the differing definitions of what constitutes a problem with ‘switching’. Switching impairment may mean inability to switch to a non-L1 language, as is often seen in dementia, or it may mean borrowing a word from one language to use in another as a strategy – deliberate or not – to deal with anomia as, one might argue, in the Abutalebi et al., 2009 case discussed above. These two forms of switching are commonly seen. Another possibility, however, would be violation of codeswitching norms between languages. In such a case, the patient might switch at a syntactic juncture different from where switching occurs in healthy participants, or within a word, or NOT switch in conversation where switching would be expected. To our knowledge such switching problems have not been reported in multilinguals with aphasia.

Bilingual Aphasia Therapy: What is the Directionality of Generalization? A recent review (Lorenzen & Murray, 2008) has covered the clinical implications of the literature on bilingual aphasia; here we consider the theoretical clinical questions. Usually one language is chosen when aphasia therapy is provided to bilinguals due to the limited number of bilingual speech-language pathologists and/or a patient and family’s preference or needs. Therefore, many researchers have investigated the question of whether cross-language generalization can result from the treated language to a non-treated language (e.g. Edmonds & Kiran, 2006; Gil & Goral, 2004; Goral et al., 2010; Miertsch et al., 2009) and some researchers have tested which language (either L1 or a later language) or language elements (e.g. cognates or non-cognates) should be targeted to maximize treatment gains in a non-treated language (e.g. Croft et al., 2011; Edmonds & Kiran, 2006; Kohnert, 2004). Additionally, recent studies on bilingual aphasia treatment efficacy have reported brain activation changes using fMRI along with language performance changes before and after aphasia therapy (e.g. Abutalebi et al., 2009; Marangolo & Rizzi, 2009; Meinzer et al., 2007).

8

Par t 1: Broad Considerat ions

Although bilingual aphasia therapy has resulted in improvement of language skills in bilinguals in general (Faroqi-Shah et al., 2010), inconsistent cross-language generalization patterns have been found so far. As Kohnert and Peterson (Chapter 6, this volume) report, some researchers have observed cross-language generalization after aphasia therapy (e.g. Gil & Goral, 2004; Miertsch et al., 2009), others have reported that cross-language generalization is limited to certain items or a certain language (i.e. to a more proficient language or to a less proficient one) (e.g. Edmonds & Kiran, 2006; Kohnert, 2004) and the remaining researchers (e.g. Abutalebi et al., 2009; Galvez & Hinckley, 2003; Meinzer et al., 2007) have reported no generalization from the trained language to an untrained language. The majority of studies have reported cross-language generalization as a result of aphasia therapy in one language. For example, in the study of Marangolo and Rizzi (2009), a Flemish- (L1) and Italian- (L2) speaking bilingual with aphasia with expressive language deficits received aphasia intervention in Italian at two months post-onset and the therapy lasted for six months. This participant showed a parallel recovery pattern during therapy and cross-language generalization was shown when his naming skills before and after treatment were compared. That is, although the treatment was provided in Italian, the same amount of improvement was found in both languages and there were no significantly different performance skills between the two languages at any time. Among studies reporting cross-language generalization, there are a few that have asked whether aphasia therapy sessions in different languages are helpful as well. For example, Gil and Goral (2004) provided a series of aphasia therapy sessions to a Russian- (L1) and Hebrew- (L2) speaking late bilingual with aphasia who exhibited expressive and receptive language deficits in both languages starting at two weeks post-onset of his aphasia. Crosslanguage generalization was found after a month of language therapy in Hebrew. The patient presented parallel recovery of both languages in the receptive language deficits. However, after a three-and-a-half month continuation of the aphasia therapy in Hebrew, his non-treated Russian showed greater improvement than the treated Hebrew. At this time he demonstrated naming difficulty in Russian and receptive language deficits in Hebrew. Subsequently, aphasia therapy was provided in Russian for six weeks. Again, cross-language generalization was observed and the participant presented only naming difficulty in both languages at the end of therapy. Similarly, Goral et al. (2011) describe how they offered JM, a patient with severe non-fluent aphasia, intensive therapy in two of his five languages consecutively, testing his naming in each language pre-therapy and after each therapy block to determine how relative proficiency would interact with generalization. They report that while performance after therapy in the most proficient language tested, Spanish, did not generalize, therapy in a less proficient one, English, generalized to his other less-proficient languages. As

The Study of Bilingual Aphasia : The Quest ions Addressed

9

well as proficiency, they observe, the fact that English was the language of the environment may also have contributed to JM’s particular success with English-language therapy, though it is harder to see how this might have influenced the generalization to other non-L1 languages. Although most studies have reported cross-language generalization from treated to untreated languages, some researchers report that such generalization was found when treatment targeted structurally similar linguistic elements (e.g. Goral et al., 2010; Laganaro & Overton-Venet, 2001; Kohnert, 2004). In Kohnert’s (2004) study, the effect on generalization of focusing on cognates in the two treated languages was demonstrated. In her study, a 62-year-old Spanish-English speaking man with severe non-fluent aphasia received aphasia therapy focusing on matching written words, completing sentences, writing names and confrontation naming. The participant’s naming abilities were measured at pre- and post-treatment and naming skills for cognates and noncognates in both languages were compared. For the first week therapy was provided in Spanish and for the second week it was provided in English. After each treatment phase within-language generalization from trained items to untrained items for both cognates and non-cognates was found. However, cross-language generalization was found only for cognates. Kohnert suggested that targeting common elements such as cognates that have similar structures and meanings in both languages that the bilingual uses would yield maximum initial treatment gains. Indeed, in the Goral et al. (2011) study discussed above, the authors addressed the question of cognate status, finding that in the nontreated languages better naming performance was seen on cognates (as compared to non-cognates) of the treated language. Additionally, other researchers ask which language (the non-dominant language or the dominant one), should be targeted for therapy to result in cross-linguistic generalization. Edmonds and Kiran (2006) conducted a single-subject experimental research design to answer this question. Three bilinguals (L1: Spanish, L2: English) with aphasia participated in this study and all received aphasia therapy to improve naming skills; one participant was a balanced bilingual and the other two participants were not balanced bilinguals pre-morbidly. The first participant received intervention in English and within- and cross-language generalization were evident. The second participant received aphasia therapy in English, which was her dominant language, and failed to show cross-language generalization, although withinlanguage generalization was observed. This second participant then received aphasia therapy in Spanish (her less dominant/proficient language) and demonstrated cross-language generalization to English, yet no within-language generalization was found. The third participant received aphasia therapy in Spanish first and showed cross-language generalization and no withinlanguage generalization, just like the results of the second phase of the second participant. Based on these observations, Edmonds and Kiran claimed that if aphasia therapy is provided to unbalanced bilinguals in their less proficient

10

Par t 1: Broad Considerat ions

language, cross-language generalization can be better facilitated than if the more proficient language is used. However, a recent study reported the opposite result. Croft et al. (2011) provided two phases of aphasia therapy to five Bengali- (L1, more proficient language) and English- (L2, less proficient language) speaking bilinguals. Participants’ scores after therapy in each language were compared to baseline. Three out of five instances showed cross-linguistic generalization when aphasia therapy was provided in the participant’s dominant (L1, more proficient) language. Therefore, further studies should be conducted to investigate the effect of language dominance on cross-language generalization after aphasia therapy. Unlike the studies reviewed above there have also been those in which no cross-linguistic generalization at all was reported. For example, Meinzer et al. (2007) studied a German-French speaking early bilingual with chronic aphasia. Before participating in the study this participant had received aphasia therapy in German for the first 32 months and only German was spoken by the patient’s family and friends. Before the experiment the participant performed better in German on the naming task. After a short-term intensive language therapy, called Constraint-Induced Aphasia Therapy, the patient’s naming skills in German (the treated language) improved, but no cross-linguistic generalization to naming performance was seen in French (the untreated language). Note that in this study, as in Abutalebi et al. (2009) which is discussed above in the section on switching, only naming abilities were tested. In sum, the main questions of direction of generalization of aphasia therapy remain unresolved. That is, while it appears that generalization does sometimes obtain, it does not always do so, and, when it does, the direction of generalization may be either from L1 to another language, or vice versa. As well, it can be from a more to a less proficient language or vice versa. Clearly, then, the factors that enter into explaining these diverse possibilities will continue to be asked.

Bilingual Aphasia Therapy: What Activation Patterns Are Seen? Recent studies have started to present behavioral data and fMRI data to determine the efficacy of bilingual aphasia treatment and the resulting brain activation patterns of both languages. For example, Marangolo and Rizzi (2009) reported cross-language generalization from L2 (Italian) to L1 (Flemish), and the same brain areas were activated in both languages before and after a two-week program of intensive aphasia therapy. However, increased activation in additional perilesional and homologous contralesional

The Study of Bilingual Aphasia : The Quest ions Addressed

11

areas was also found after intervention. In the case of Meinzer et al. (2007) mentioned above, the research team provided intensive Constraint-Induced Aphasia Therapy in German (L2). Before the treatment, better performance in German was found during the naming task and this was associated with increased brain activation more in German than in French. After the treatment, improved naming performance and increased brain activation was found in German (the treated language) and no cross-linguistic generalization to naming performance and unchanged brain activation in French were observed. Recall that a similar pattern of no treatment generalization was also reported in the study of Abutalebi et al. (2009), and the brain-imaging data there suggested shifting patterns of activation resulting from therapy. In sum, many studies have reported efficacy of bilingual aphasia therapy, although treatment designs were various in terms of language choice for treatment, treatment objectives and intensity of treatment. As well, characteristics of the patients treated, age of L2 acquisition, lesion size and location, and pre-morbid proficiency differed across studies, as Faroqi-Shah et al. (2010) noted. Thus, the areas that remain to be addressed are widely agreed upon: to examine factors promoting and constraining cross-language generalization from a treated to an untreated language (Kohnert, 2009) and to document recovery patterns of bilinguals with a combination of behavioral performance and brain imaging data.

Cutting-Edge Questions Now that aphasiology is complemented by neuroimaging in neurolinguistic study, questions pass back and forth among sub-areas of the field to search for converging evidence that may give us confidence in any given type of findings. For example, the question of lateral dominance for the languages of the bilingual or multilingual was studied in healthy individuals via tachistoscopic and dichotic-listening techniques soon after Gloning and Gloning first raised the issue in their study of patients with crossed aphasia (see Obler et al., 1982 for a summary and critique of this work). The questions of overlapping language areas for the languages of a multilingual have been addressed primarily by fMRI and Positron Emission Tomography (PET) (e.g. Kim et al., 1997, is an early study; see Simmonds et al., 2011, for a recent review of the sizable literature since that time), after a small somewhat earlier series employing cortical stimulation (e.g. Ojemann & Whitaker, 1978). Several questions that remain unresolved from the previous literature will, no doubt, continue to be explored, both in multilingual patients with aphasia, and in multilingual individuals with no language impairment. These include the questions concerning the language switch, and explanations for

12

Par t 1: Broad Considerat ions

patterns of differential impairment and recovery (and, we would advise, for parallel impairment and recovery when it obtains). The questions concerning cognate status in the bilingual or multilingual lexicon have been studied in healthy participants (e.g. de Groot & Nas, 1991; Cai et al., 2011), as well as in those with aphasia (Goral et al., 2006; Goral et al., 2011; Kohnert, 2004). Related questions of how the degree of similarity of orthographic systems contributes to degree of impairment and recovery have been asked and will continue to be refined (see Wilson et al., Chapter 12, this volume; and Eng, Chapter 15, this volume); those concerning degree of similarity of morphology and syntax have yet to be rigorously considered in the literature on bilinguals with aphasia (Fabbro, 2001; Ibrahim, 2008), though they have certainly been considered in the cross-language aphasia literature that queries how aphasia may manifest differently in different languages (e.g. Menn & Obler, 1990; Menn et al., 1995). The question of the interaction between aphasia and accent is just now being asked – not the question about putative ‘foreign-accent syndrome’, but, rather, the question of how it is that a second-language speaker of a given language can have a markedly stronger L1 accent in that language after an aphasia-producing stroke than before (Levy et al., forthcoming). As the field evolves we may expect to see further questions concerning behaviors specific to bilingualism. These would include questions like: What patterns of translation behavior do we see post-aphasia in individuals with prior expertise in translation or simultaneous interpretation? Also, we expect expansion of the literature on the ways in which bidialectal aphasia is similar to, and different from, multilingual aphasia, as treated in the chapter by Jones et al. in this volume (Chapter 16). We may also anticipate fine-grained discoveries of specific parallels in bilingual performance in people with aphasia, like the report of agrammatism despite preferredmodality differences in Sebastian et al. (Chapter 8, this volume), or selective problems with oral numbers across Chinese and English, as reported in Eng (Chapter 15, this volume).

Acknowledgements Thanks are due to Mira Goral for suggestions that made this a stronger chapter.

Note (1) Note that in the case reported by Sebastian et al. (Chapter 8, this volume) such switching is considered to be positive in that it helps the participant communicate better.

References Abutalebi, J. and Green, D. (2007) Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics 20, 242–275.

The Study of Bilingual Aphasia : The Quest ions Addressed

13

Abutalebi, J. and Green, D. (2008) Control mechanisms in bilingual language production: Neural evidence from language switching studies. Language and Cognitive Processes 23, 557–582. Abutalebi, J., Rosa, P.A.D., Tettamanti, M., Green, D.W. and Cappa, S.F. (2009) Bilingual aphasia and language control: A follow-up fMRI and intrinsic connectivity study. Brain and Language 109, 141–156. April, R.S. and Tse, P.C. (1997) Crossed aphasia in a Chinese bilingual dextral. Archives of Neurology 34, 766–770. Cai, Z.G., Pickering, M.J., Yan, H. and Branigan, H.P. (2011) Lexical and syntactic representations in closely related languages: Evidence from Cantonese-Mandarin bilinguals. Journal of Memory and Language 65, 431–445. Charlton, M. (1964) Aphasia in bilingual and polyglot patients: A neurological and psychological study. Journal of Speech and Hearing Disorders 29, 307–311. Crinion, J., Turner, R., Grogan, A., Hanakawa, T., Noppeney, U., Devlin, J.T., Aso, T., Urayama, S., Fukuyama, H., Stockton, K., Usui, K., Green, D.W. and Price, C.J. (2006) Language control in the bilingual brain. Science 312, 1537–1540. Croft, S., Marshall, J., Pring, T. and Hardwick, M. (2011) Therapy for naming difficulties in bilingual aphasia: Which language benefits? International Journal of Language and Communication Disorders 46, 48–62. de Groot, A.M. and Nas, G.L. (1991) Lexical representation of cognates and noncognates in compound bilinguals. Journal of Memory and Language 31, 90–123. Edmonds, L.A. and Kiran, S. (2006) Effect of semantic naming treatment on crosslinguistic generalization in bilingual aphasia. Journal of Speech, Language and Hearing Research 49, 729–748. Fabbro, F. (1999) The Neurolinguistics of Bilingualism: An Introduction. Hove, Sussex: Psychology Press. Fabbro, F. (2001) The bilingual brain: Bilingual aphasia. Brain and Language 79, 201–210. Fabbro, F., Peru, A. and Skrap, M. (1997) Language disorders in bilingual patients after thalamic lesions. Journal of Neurolinguistics 10, 347–367. Faroqi-Shah, Y., Frymark, T., Mullen, R. and Wang, B. (2010) Effect of treatment for bilingual individuals with aphasia: A systematic review of the evidence. Journal of Neurolinguistics 23, 319–341. Fredman, M. (1975) The effect of therapy given in Hebrew on the home language of the bilingual or polyglot adult aphasic in Israel. British Journal of Language and Communication Disorders 10, 61–69. Freud, S. (1891) Zur Auffassung der Aphasien (translated to English as On Aphasia by E. Stengel, 1953). New York: International University Press. Galvez, A. and Hinckley, J. (2003) Transfer patterns of naming treatments in a case of bilingual aphasia. Brain and Language 87, 173–174. Gil, M. and Goral, M. (2004) Nonparallel recovery in bilingual aphasia: Effects of language choice, language proficiency, and treatment. International Journal of Bilingualism 8, 191–219. Gloning, I. and Gloning, K. (1965) Aphasien bei Polyglotten: Beitrag zur Dynamik des Sprachabbaus sowie zur Lokalisationsfrage dieser Störungen. Wiener Zeitschrift für Nervenheilkunde 22, 362–397. Goral, M., Levy, E.S., Obler, L.K. and Cohen, E. (2006) Lexical connections in the multilingual lexicon. Brain and Language 98, 235–247. Goral, M., Levy, E.S. and Kastl, R. (2010) Cross-language treatment generalization: A case of trilingual aphasia. Aphasiology 24, 170–187. Goral, M., Rosas, J., Conner, P.C., Maul, K. and Obler, L.K. (2011) Effects of language proficiency and language of the environment on aphasia therapy in a multilingual. Journal of Neurolinguistics 1–14.

14

Par t 1: Broad Considerat ions

Hernandez, A.E., Dapretto, M., Mazziotta, J. and Bookheimer, S. (2001) Language switching and language representation in Spanish-English bilinguals: An fMRI study. Neuroimage 14, 510–520. Hernandez, A.E., Martinez, A. and Kohnert, K. (2000) In search of the language switch: An fMRI study of picture naming in Spanish-English bilinguals. Brain and Language 73, 421–431. Ibrahim, R. (2008) Performance in L1 and L2 observed in Arabic-Hebrew bilingual aphasic following brain tumor: A case constitutes double dissociation. Psychology Research and Behavior Management 1, 11–19. Kim, K.H.S., Relkin, N.R., Lee, K.M. and Hirsch, J. (1997) Distinct cortical areas associated with native and second languages. Nature 388, 171–174. Kohnert, K. (2004) Cognitive and cognate-based treatments for bilingual aphasia: A case study. Brain and Language 91, 294–302. Kohnert, K. (2009) Cross-language generalization following treatment in bilingual speakers with aphasia: A review. Seminars in Speech and Language 30, 174–186. Krapf, E.E. (1955) The choice of language in polyglot psychoanalysis. The Psychoanalytic Quarterly 24, 343–357. Laganaro, M. and Overton Venet, M. (2001) Acquired alexia in multilingual aphasia and computer-assisted treatment in both languages: Issues of generalization and transfer. Folia Phoniatrica et Logopaedica 53, 135–144. Levy, E.S., Goral, M., Castelluccio de Diesbach, C. and Law II, F. (2011) Stronger accent following a stroke: The case of a trilingual with aphasia. Journal of Clinical Linguistics and Phonetics 25, 815–830. Lorenzen, B. and Murray, L. (2008) Bilingual aphasia: A theoretical and clinical review. American Journal of Speech-Language Pathology 17, 299–317. Marangolo, P. and Rizzi, C. (2009) Parallel recovery in a bilingual aphasic: A neurolinguistic and fMRI study. Neuropsychology 23, 405–409. Meinzer, M., Obleser, J., Flaisch, T., Eulitz, C. and Rockstroch, B. (2007) Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI. Neuropsychologia 45, 1247–1256. Menn, L. and Obler, L.K. (eds) (1990) Agrammatic Aphasia: A Cross Language Narrative Sourcebook (Vols 1–3). Amsterdam: Benjamins. Menn, L., O’Connor, M.P., Obler, L.K. and Holland, A. (1995) Non-Fluent Aphasia in a Multilingual World. Amsterdam: Benjamins. Miertsch, B., Meisel, J.M. and Isel, F. (2009) Non-treated languages in aphasia therapy of polyglots benefit from improvement in the treated language. Journal of Neurolinguistics 22, 135–150. Minkowski, M. (1963) On aphasia in polyglots. In L. Halpern (ed.) Problem of Dynamic Neurology (pp. 119–161). Jerusalem: Hebrew University. Obler, L.K. and Albert, M.L. (1977) Influence of aging on recovery from aphasia in polyglots. Brain and Language 4, 460–463. Obler, L.K. and Mahecha, N. (1991) First language loss in bilingual and polyglot aphasics. In H. Seliger and R. Vago (eds) First Language Attrition: Structural and Theoretical Perspectives (pp. 53–66). Cambridge: Cambridge University Press. Obler, L.K., Zatorre, R.J., Galloway, L. and Vaid, J. (1982) Cerebral lateralization in bilinguals: Methodological issues. Brain and Language 15, 40–54. Ojemann, G.A. and Whitaker, H.A. (1978) Language localization and variability. Brain and Language 6, 239–260. Paradis, M. (1977) Bilingualism and aphasia. In H. Whitaker and A. Whitaker (eds) Studies in Neurolinguistics (Vol. 3) (pp. 65–121). New York: Academic Press. Paradis, M. (2001) Bilingual and polyglot aphasia. In Handbook of Neuropsychology (2nd edn) Vol. 3. Language and Aphasia (pp. 69–91). Amsterdam: Elsevier Science.

The Study of Bilingual Aphasia : The Quest ions Addressed

15

Pitres, A. (1895) Etude sur l’aphasie. Revue de Médecine 15, 873–899. Ribot, T. (1881) Les maladies de la mémoire. Paris: Alcan; (1982) Disease of Memory (English trans.). London: Kegan Paul, Trench & Co. Simmonds, A.J., Wise, R.J. and Leech, R. (2011) Two tongues, one brain: Imaging bilingual speech production. Frontiers in Psychology 2, 166. doi: 10.3389/fpsyg.2011.00166. Watamori, T.S. and Sasanuma, S. (1978) The recovery processes of two English-Japanese bilingual aphasics. Brain and Language 6, 127–140.

2

Bilingual Aphasia: Neural Plasticity and Considerations for Recovery Daniel Adrover-Roig, Karine Marcotte, Lilian C. Scherer and Ana Inés Ansaldo

Introduction Given the increasing number of bilingual and multilingual people around the world, bilingual aphasia has become a prominent research topic. This chapter discusses the literature on first (L1) and second (L2) language representations, focusing on pathological language switching (LS) and mixing; whereby emphasis is placed on the frontal-basal ganglia network for language control. The factors that affect different recovery patterns are also discussed, and an update of considerations for bilingual aphasia therapy is provided. The impact of word type and proficiency level in bilingual patients’ recovery from aphasia is included as well. Finally, future directions are suggested, with a special emphasis on longitudinal studies involving brain plasticity. Nowadays, more than half of the world’s population can be considered to be bilingual or multilingual (Fabbro, 2001a); consequently, there is an increasing likelihood that brain damage will cause bilingual aphasia. Overall, 40% of bilinguals with aphasia experience parallel recovery (in which both languages recover at a similar rate), 32% report experiencing a better recovery of L1, whereas 28% show better recovery in L2 (Fabbro, 1999). Professionals concerned with bilingual aphasia face a number of theoretical and clinical challenges. As Centeno (2005) notes, they include understanding the heterogeneity of bilinguals as a group and the varied patterns of post-stroke language recovery. Another clinical challenge concerns the low number of bilingual testing and therapeutic resources available. Thus, detailed assessment methods and effective treatments addressing the specificities of bilingual populations are urgently needed. 16

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

17

Beyond the reasons stated above, it is important to know how language is represented in the bilingual brain. There is evidence that the degree of overlap between L1 and L2 brain representations depends on age of acquisition (AoA) and language use (Kim et al., 1997; Wartenburger et al., 2003). As well, the potential specificities of the bilingual’s cognitive system have recently been highlighted (Bialystok et al., 2005, 2006, 2007); furthermore, it has been shown that mastering more than one language creates some cognitive advantages. An epidemiological study by Bialystok et al. (2007) reported that elderly bilinguals show a four-year delay in presenting Alzheimer’s disease (AD) signs; this might reflect better inhibitory capacities in bilinguals, as the result of a lifetime’s experience in dealing with two competing languages (Bialystok et al., 2006). Given that baseline cognitive capacities may differ between monolinguals and bilinguals, so might recovery patterns following aphasia. A large number of case reports (see Green, 2005, for a review) provide evidence of the heterogeneity of recovery patterns, and argue that factors such as pre-morbid degree of proficiency in either language, age and context of acquisition and use, distance between languages, type of motivation and the emotional values associated with each language may play a role in bilingual aphasia. Importantly, proficiency level and AoA have been found to be related to specific memory systems; in particular, the role of both explicit and implicit memory systems in bilingual language processing has recently been highlighted (Ullman & Pierpont, 2005). Given that lesions in the perisylvian cortical regions, the left basal ganglia and the cerebellum lead to impairments affecting automatic linguistic knowledge (Ullman & Pierpont, 2005), it is likely that late bilinguals with damage to these areas will show a greater impairment in L1 than early bilinguals, who would be expected to show the same kinds of deficits in L1 and L2. Conversely, lesions in the hippocampal system may result in a selective impairment of metalinguistic knowledge, since these brain regions are involved in declarative memory (Ullman, 2004; Ullman & Pierpont, 2005). According to Paradis (2008), in the latter case, word access difficulties could be bypassed by translation operations. This perspective predicts differential effects of brain damage on L1 and L2 as a function of AoA. In sum, language components that can be addressed by rules relate to implicit language knowledge, and thus are more selectively impaired when primary language areas are damaged. Conversely, damage to medial-temporal regions might instead affect metalinguistic capacities. In line with this claim, Zanini et al. (2004) showed implicit memory impairments in both aphasia and Parkinson’s disease (PD), and reported that bilinguals with PD showed greater syntactic impairments in their more automatically retrieved L1. Conversely, bilinguals with brain damage to medial-temporal regions (declarative, explicit memory), as in cases of AD, tend to show a retreat to their L1 (Hyltenstam & Stroud, 1989). A similar interpretation could be extracted following the results by Goral et al. (2006),

18

Par t 1: Broad Considerat ions

who suggested that non-native languages (L2 and L3) may cause greater interference during translation and might also recover to a lesser extent than L1, as shown after the evaluation of a trilingual person who suffered from global aphasia caused by a left fronto-temporo-parietal infarct. The authors also proposed that ‘the pre-morbid patterns of inter-language lexical connection can remain potent in the presence of aphasia’ (Goral et al., 2006: 245). Consequently, damage to these circuits is more likely to interfere with the processing of the late acquired second and other languages than with the processing of the native one. As suggested by Paradis (2004), individuals may have conscious, metalinguistic knowledge of their L2 that they do not possess for L1. If such metalinguistic knowledge is the source of an advantage for the differential recovery of L2 over L1, then the processing of L2 will differ from that of normal proficient bilinguals who are using implicit rules of the procedural system. In sum, languages that are differentially mastered might rely on different memory networks, but when proficiency in both L1 and L2 is similar, the representation of different languages in a bilingual may rely upon overlapping neural networks (Kim et al., 1997; Vingerhoets et al., 2003). In this regard, there is converging evidence of the overlapping representation of L1 and L2 lexicons, in both early and late bilinguals (see Klein et al., 2006). To conclude, the literature reviewed here points to the close cortical representation of equally mastered languages in bilinguals; brain insult may disrupt the delicate balance that allows the bilingual to simultaneously activate both languages, avoid interference between them and use them in appropriate conversational contexts. These cognitive capacities may be disrupted when brain damage affects the neural networks involved in cognitive control.

Cognitive Control Capacity in the Recovery from Aphasia Cognitive control refers to the capacity to actively maintain patterns of activity that represent goals and the means to achieve them (Miller & Cohen, 2001). It has been proposed that cognitive control mechanisms are essential to bilingualism and to language recovery (e.g. Abutalebi & Green, 2007; Green, 1986; Paradis, 1998, 2004). Language control is important because selective loss of a language may arise because of an inability to inhibit the schema for producing speech in one language rather than another (Green, 1986) or to raise the activation level of the alternative sufficiently (Paradis, 1998). Green’s (1986) model proposes that linguistic and control units are distinct components of the bilingual system. Linguistic units concern the lexical-semantic system itself and correspond to semantic, syntactic, phonological and morphological language levels. Control units share devices that allow for language choice, by a complex interplay of activation and inhibition,

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

19

resulting in spontaneous communication in either language, and translation between them (Fabbro et al., 2000; Green, 1998). As previously mentioned, the boosted control mechanisms of bilinguals might extend to their LS abilities. Price et al. (1999) reported that LS elicited activations in Broca’s area and the supramarginal gyri bilaterally, which they related to phonological recoding. Conversely, translation, but not switching, was associated with greater activity in the anterior cingulate cortex (ACC) and subcortical structures (Price et al., 1999), together with more widespread activations in the anterior insula, cerebellum and supplementary motor area, regions that Price et al. (1999) related to articulation. These results were not replicated in studies by Hernandez et al. (2000, 2001), who reported a dorsolateral frontal activation for LS tasks, with both verbal and non-verbal stimuli (Rubinstein et al., 2001), and argued that dorsolateral frontal cortex activation reflects the executive processing induced by LS. These results (Hernandez et al., 2000, 2001; Price et al., 1999) show that, depending on the switching task (naming vs translation), different control demands may lead to different activation patterns. Given that bilinguals must constantly deal with the issue of language control, they are considered to outperform monolinguals at tasks that require control-switching abilities. Using a Simon Task, Bialystok (2001) showed that bilingual children develop control abilities more fully than monolingual children. In a subsequent study on inhibition and switching, Bialystok et al. (2005) reported that faster responses in bilinguals were associated with increased activation in the left inferior frontal regions, which could reflect enhanced control processing involved in inhibitory tasks, as proposed by Green’s (1998) model of inhibitory control. Thus, the relation between recovery patterns, language representation and control mechanisms is complex (Abutalebi et al., 2009; Green et al., 2009). In this vein, it has been claimed that selective aphasia in one language, as well as pathological switching between languages, can be explained by difficulties in language control. As outlined by Paradis (1998), parallel recovery may reflect a similar degree of inhibition across languages, whereas antagonistic recovery could result from the limited inhibition of one language, which then shifts to the other language (Green, 1986). In this regard, it is important to note that language control areas may even fall outside the classical language areas (Abutalebi & Green, 2007; Crinion et al., 2006) and thus determine recovery patterns that cannot be explained by the traditional aphasia classifications. In a recent study that addressed problems affecting language control in bilingual individuals with parallel recovery of aphasia, Green et al. (2009) tested two right-handed, non-native English-speaking participants who showed a parallel recovery of both languages after stroke and a group of nonnative English-speaking controls. The bilingual aphasia patients and controls described a scene in English and in their first language and completed three

20

Par t 1: Broad Considerat ions

explicit conflict tasks. Two of these were verbal conflict tasks: a lexical decision task in English, in which individuals distinguished English words from non-words, and a Stroop task, in English and in their first language. The third conflict task was a non-verbal flanker task. The results showed that participants with aphasia were impaired in the picture description task in English and in their first language, but showed different patterns of impairment on the conflict tasks. For the participant with left subcortical damage, performance on the conflict tasks showed abnormally high conflict during the verbal tasks (lexical decision and Stroop), but not during the non-verbal flanker task. In contrast, for the participant with extensive left parietal damage, conflict was less abnormal during the Stroop task than in the flanker or lexical decision tasks. These results led the authors to conclude that two separate impairments associated with control mechanisms can be distinguished in parallel recovery, and highlight the importance of evaluating the precise nature of control problems in bilingual aphasia. The next section describes in more detail the deficits associated with impaired language control mechanisms in bilingual aphasia, specifically pathological LS and mixing.

Pathological switching in bilinguals with aphasia Brain damage may alter the capability for language control, which is a major issue for bilinguals, and general cognitive abilities may be recruited to control dual-language use (Hernandez et al., 2000). The ability to select one language or the other is a trademark of bilinguals and is modulated by proficiency (Green, 1998; Price et al., 1999). LS requires compliance with linguistic and pragmatic constraints in order to achieve efficient communication. Patients who show pathological language mixing typically use words of one language in the context of an utterance in another language (Fabbro, 2001b). By contrast, pathological LS is characterized by the alternation of languages across different utterances. With regard to pathological LS, Fabbro (2001a) reported the case of a patient with a lesion to the left ACC and to the frontal lobe (marginally involving the right ACC), who presented pathological switching between languages in the absence of any other linguistic impairment. Unlike pathological mixing, which typically occurs within the context of bilingual aphasia, pathological switching may be observed in isolation and concurrently with relatively preserved language processing. Fabbro argued that both pathological switching and mixing are related to the capacity to withdraw attention and resources from L2, and reflect the cognitive capacity to avoid L2 interference when using the native language. There are relatively few reports on pathological switching and mixing, but those few suggest that damage to specific brain circuits is at the source of this deficit. However, pathological LS is not an inherent characteristic of bilingual aphasia; for instance, Vilarino et al. (1997), who studied a group of

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

21

Galician-Castilian persons with aphasia, did not find any signs of pathological switching in the 29 men and 20 women who participated in their study. In a more recent study, Ansaldo et al. (2008) reported the case of a SpanishEnglish bilingual with aphasia who showed pathological mixing and switching in the context of anomic aphasia. The patient presented damage to the subcortical (i.e. basal ganglia) circuits that connect to the frontal lobe structures involved in language control (Abutalebi et al., 2008; Green, 1986). Ansaldo et al. (2008) reported preserved translation abilities, and developed a therapy approach – Switch Back through Translation (SBTT) – that proved to be effective at improving language control and communication abilities in this bilingual patient. The pattern described by Ansaldo et al. (2008) is similar to the one reported by Lalor and Kirsner (2001), who argue that this impairment probably results from damage to the basal ganglia-frontal circuits responsible for language control. In sum, Ansaldo et al. (2008) provided evidence for model-driven intervention in bilingual aphasia, by showing that SBTT not only contributed to overcoming pathological switching but also improved word retrieval in both languages. In the next subsection, we further discuss the role of several specific neural networks involved in pathological LS and mixing.

The frontal-basal ganglia network in language switching and mixing In the last decade, an increasing number of studies have discussed the role of the frontal-basal ganglia network in LS and mixing. While the role of the left basal ganglia in monolingual aphasia has been defined (Abutalebi & Green, 2007; Longworth et al., 2005), their relevance to bilingual aphasia is less clear. For instance, Muñoz et al. (1999) argued that damage to this network could be related to an increased dependence on both languages, as the authors found that in bilinguals with aphasia, the number of code-switching events observed during conversational discourse is significantly larger than in healthy bilinguals. The pioneer study by Aglioti and Fabbro (1993) first showed that damage to the left basal ganglia causes severe L1 production deficits, in comparison to much better preserved spontaneous speech and translation abilities in L2, highlighting the role of the basal ganglia in automatic language output. Later, Aglioti et al. (1996) reported on another case of bilingual aphasia, secondary to subcortical damage also involving the left basal ganglia, characterized by impaired translation from L2 to L1 but less impaired L1 to L2 translation capacity, which again showed that damage to the basal ganglia could affect automatic motor and cognitive performance. An additional piece of evidence comes from Mariën et al. (2005), who evaluated an early bilingual presenting with symmetric transcortical sensory aphasia in L1 and L2 after a posterior left thalamic haemorrhage, whose remission of spontaneous pathological LS and mixing was associated with

22

Par t 1: Broad Considerat ions

restored perfusion in the left frontal lobe and left caudate nucleus. Thus, a circuit composed of the left frontal lobe and basal ganglia seems to be crucially involved in the control of LS and mixing. Other authors, such as Avila et al. (2004), have highlighted the important role of the white matter connecting the basal ganglia to cortical structures for language articulation. Avila et al. (2004) reported on a polyglot female (L1 Spanish, and knowledge of French, English and Catalan, L2, L3, and L4, respectively, all of them learned after the age of 12) who showed a small infarct in the left corona radiata causing a Foreign Accent Syndrome (FAS) exclusively in her L1. This observation led the authors to conclude that control of phonetic patterns is mediated by different areas as a function of AoA. Furthermore, recent neuroimaging evidence has confirmed the role of two frontal-basal ganglia circuits, one projecting from the basal ganglia to BA44 (sustaining procedural memory-related functions), and the other projecting to BA45/47 (sustaining lexical-declarative memory and retrieval-selection) (Ullman, 2006). Taking projections into account, anterior structures (BA45/57) and the basal ganglia might be more active during operations involving both the selection and retrieval of lexical knowledge across linguistic communities (Fabbro, 2001b), as in the case reported by Ansaldo et al. (2008). To sum up, these case reports provide evidence for the importance of integrating neurocognitive and neurobiological models of bilingualism when planning interventions with bilinguals suffering from aphasia. The next sections are devoted to key ideas for intervention in cases of bilingual aphasia, specifically with regard to assessment issues and cross-linguistic transfer of therapy gains (CLT).

Some Considerations on the Assessment of Bilingual Aphasia The assessment of bilingual aphasia must take into consideration the factors that are known to modulate language representation in bilinguals, including pre-morbid proficiency, language use and language characteristics (Muñoz & Marquardt, 2003), to avoid attributing to aphasia what is in fact the result of pre-morbid bilingual features (Kiran & Tuchtenhagen, 2005). Moreover, lesion location may interact with AoA and method of acquisition to determine bilingual aphasia profiles. Hence, a native bilingual is likely to show similar aphasia patterns across languages, given that both languages are likely to have overlapping representations and rely upon implicit memory systems; conversely, late bilinguals may show different aphasia patterns across languages, depending upon the impact of lesion location on implicit and explicit memory circuits. Given that the method of acquisition is closely related to brain representation of languages (Paradis, 2001), lesion location may determine different recovery patterns across languages

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

23

depending on the type of setting (informal vs structured) in which each language was acquired. As for proficiency, pre-morbid differences may modulate recovery patterns and should not be interpreted as the result of aphasia. Recovery may also be influenced by motivational factors. For example, if L2 was learned when the subject moved to another country, the motivation for recovery may be higher, given that the L2 may have become the language of social and work life. Conversely, if the L2 was learned as a hobby, the motivation to recover it may be weaker. Pre-morbid use of each language is another factor that should be considered, as frequency of use relates to practice, which has been associated with the strength of brain connections (Veroude et al., 2010). Hence, it is likely that the more a language or a specific language modality (writing, reading, speaking or listening) is used, the stronger the connections within the underlying network are, which makes it easier to recruit brain networks for each language across modalities (Paradis, 2004). Moreover, the structure and characteristics of a given language may determine the type of aphasia symptoms. This implies that cross-language differences in word frequency, orthography rules, word structure complexity and syntactic rules may determine the occurrence of specific errors in each language (Paradis, 2001). In this context, Alexiadou and Stavrakaki (2006) reported the case of a fluent late Greek-English bilingual who showed syntactic and grammatical deficits in adverb placement only in English, and this was associated with the nature of the syntactic hierarchy in the two languages. Thus, as Paradis (2001) states, patients might perform significantly worse in one language because of the intrinsic nature of their languages. Different tools can be used in the assessment of bilingual aphasia, which should be adapted to the time elapsed after brain damage. Hence, screening is more adapted to the changing patterns that characterize the acute phase, whereas a more complete assessment should be carried out during the subacute phase, when language abilities become more stable. The BAT (Bilingual Aphasia Test, Paradis & Libben, 1987) is a comprehensive assessment tool that has been developed for several language pairs and constitutes a unique source of information regarding bilingual aphasia profiles. Formal language testing should be administered in every language spoken by a multilingual person with aphasia. This would allow the clinician to describe the processing abilities in each language and compare the impairment across languages; it would also provide cues for clinical management, indicating which language to treat and which goals should guide treatment (Lorenzen & Murray, 2008). Importantly, the mere translation of a test is not acceptable, as tests need to be culturally and linguistically grounded so that similar linguistic abilities are assessed across languages. This allows for the comparability of aphasia type and severity across languages (see chapters in this volume on assessment and treatment as well as on cultural context for related discussion).

24

Par t 1: Broad Considerat ions

A supplementary tool for bilingual aphasia assessment is an informal conversation with the person with aphasia. For example, data collected within an informal narrative situation can be analyzed together with formal test data (Centeno, 2009). Special attention should be paid to the person’s family members, who are important sources of information regarding the pre-morbid degree of bilingualism. Thus, counseling with the family and other intimates is essential, so as to inform them about the possibility of changing recovery patterns across languages over time (Lorenzen & Murray, 2008), and involve them in the decisional process regarding the language choice for therapy. Finally, the assessment of cognitive abilities is an important component of any bilingual aphasia assessment, since traumatic brain injury and vascular damage may result in specific cognitive disorders among bilingual speakers (Green, 2005). Very few cognitive assessment tools have been constructed or adapted to different languages, and thus assessing the cognitive implications of brain damage in bilinguals represents an additional challenge for clinicians working with brain-damaged bilinguals (see also Kohnert & Peterson, Chapter 6, this volume).

Key Ideas For Intervention: The Importance of Cross-Linguistic Transfer of Therapy Effects Decisions regarding language choice should not be grounded exclusively on the better-preserved language after aphasia but should also consider the relative importance attributed by patients and their families to either language. Similarly, the most proficient language may be less useful in a given environment (Kohnert, 2004). Although there is no absolute answer regarding which language should be treated, intervention plans have traditionally excluded one language (Green, 2005; Paradis, 2001, 2004). The arguments supporting a particular language choice include treating the most frequently used language before brain insult, treating the language of the environment, treating the most impaired or least impaired language, or treating the language with stronger emotional attachments. The need for language choice has been claimed to be particularly important in cases presenting pathological LS or mixing (Fabbro, 2001b), and some authors have even claimed that treating more than one language at the same time might inhibit global speech recovery and delay the recovery of all languages (Wald, 1961). However, there is evidence that stimulating both languages simultaneously can favor rehabilitation (Ansaldo & Marcotte, 2007; Kohnert, 2004). This seems particularly relevant in the case of proficient bilinguals with aphasia (Ansaldo & Marcotte, 2007), given that the bilingual language system is a unity that comprises two language codes in one system. Hence, although cross-linguistic links at the lexical, morphosyntactic and discourse levels established before the onset of aphasia may be partially damaged, therapy should exploit those

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

25

that are still intact. According to this perspective, disregarding either language in a person with aphasia could be considered to be analogous to forcing a monolingual person with aphasia to inhibit preserved aspects of language, with the purpose of improving the others. Furthermore, when aphasia is a consequence of damage to the left hemisphere, pragmatic abilities – which are generally well preserved in aphasia – may be useful compensatory means to circumvent impairments. Bilinguals with aphasia must deal with communication settings that range from completely monolingual to completely bilingual, depending upon the communication context, making bilingual intervention a reasonable approach. Thus, given that cross-linguistic equivalents share semantic representations, and given that these representations are linked to two phonological forms, it is likely that the stimulation of shared semantic knowledge favors the activation of the corresponding phonological representations and results in the improvement of word-finding abilities in both languages (Ansaldo & Marcotte, 2007). This approach seems particularly relevant to promote cross-linguistic transfer (CLT) of therapy effects, an economical means of optimizing intervention. In this regard, it is considered that areas of similarity between languages are the best targets for CLT. With the aim of replicating earlier reports on CLT, Goral et al. (2010) examined a trilingual speaker with mild chronic aphasia, who was treated in English (the patient’s L2); a first treatment block focused on morphosyntactic skills, and a second on speech rate. Measurements in the treated language (English) as well as in the two non-treated languages (Hebrew, L1, and French, L3) were collected after each block. An improvement in pronoun and gender agreement in the treated language (L2) as well as in the non-treated L3 was observed following the treatment block on English morphosyntactic skills. There was also an improvement in speech rate in English and in French following the second block, but no changes were observed in Hebrew. These results led the authors to conclude that the selective CLT from L2 to L3 resulted from the greater similarities between English and French than between English and Hebrew. Recent studies have produced equivocal results, since some of them showed a broad-based CLT, whereas others failed to do so (Kohnert, 2009). In two recent extensive reviews, Kohnert (2009) and Faroqi-Shah et al. (2010) argue that whether treatment gains in one language will transfer to an untreated language is not a simple ‘yes’ or ‘no’ issue. While Kohnert (2009) points out that the best intervention practices should consider the client’s previous language experiences as well as current and future communication needs, Faroqi-Shah et al. (2010) report that improved language performance may be restricted to the treated language (when L2 is treated), in both the expressive and receptive domains. Given that CLT can optimize therapy effects (see Faroqi-Shah et al., 2010, for a review), further studies are necessary to identify the best ways to achieve it. Finally, considering that in bilinguals the premorbid mode of communication is generally bilingual and not

26

Par t 1: Broad Considerat ions

monolingual (Centeno, 2005), the potential of bilingual therapy to promote CLT should be further explored. Edmonds and Kiran (2006) showed that being a balanced bilingual contributed to CLT of therapy effects, whereas no transfer was observed among unbalanced bilingual patients. More specifically, Edmonds and Kiran (2006) investigated the CLT of gains achieved with a semantic approach to treat naming deficits by examining three English-Spanish bilinguals with aphasia, who all received semantic therapy in Spanish (participant 1) and in English and Spanish (participants 2 and 3). Therapy effects were tested on treated items, untreated items and translations; the results showed that both within- and across-language therapy effects were related to pre-morbid language proficiency. In their study, participant 1, a pre-morbid balanced bilingual, showed within-language generalization in the treated language (Spanish) and generalization to the untreated English items. Conversely, participants 2 and 3 (who were more proficient in English) showed within-language generalization to semantically related items, but no CLT to the untreated Spanish items. Moreover, following treatment in Spanish, participants 2 and 3 exhibited no withinlanguage generalization but CLT to English (their dominant language). Although not all the participants were treated in both languages, these results suggest that, in the case of unbalanced bilinguals, training in the less dominant language may be more beneficial in facilitating CLT. However, generalization may have been favored by the semantic nature of the therapy, given that this approach was not compared to any other one. Another possibility is to target similarities between languages as a means of facilitating CLT. To do so, therapy based on the stimulation of cognates – words that share both form and meaning in the two languages – appears to be a reasonable approach. Several clinical studies report a cognate effect on CLT from the treated to the untreated language (Edmonds & Kiran, 2006; Fabbro, 2001a; Gil & Goral, 2004); however, other studies have failed to show such an effect (Meinzer et al., 2007), or even reported a negative effect (Abutalebi et al., 2009). One possible explanation for these inconsistencies could be the degree of proximity between languages, since an enhanced CLT across languages during picture naming treatments has been observed for cognates but not for non-cognates (Kohnert, 2004; Lalor & Kirsner, 2001; Roberts & Deslauriers, 1999). For instance, Galvez and Hinckley (2003) observed L1 improvements after intensive treatment in L1, together with improvements in L2 after more superficial treatment in L2 (the language of treatment was alternated weekly), without improvement in the untreated language and with no signs of CLT benefits. However, improvements in both languages after treatment in L1 have also been reported (Ansaldo & Marcotte, 2007; Gil & Goral, 2004), with gains in both languages and CLT in particular with cognates alone (Kohnert, 2004). In her study, Kohnert (2004) compared two therapy approaches applied to stimulate skill learning and generalization within and across cognitive-linguistic domains in a 62-year-old Spanish-English bilingual

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

27

man with severe non-fluent aphasia. Treatment 1 was a cognitive-based treatment that emphasized non-linguistic skills, such as visual scanning, categorization, and simple arithmetic, whereas treatment 2 was a lexically based treatment that trained with cognates and non-cognates. The subject showed modest gains in both Spanish and English following treatment 1, and improved naming for non-cognates as well as cognates within each language following treatment 2, although generalization of gains from Spanish to English was apparent only for cognates. Kohnert argued that exploiting crosslanguage lexical-semantic links might be a way to maximize treatment gains, especially in severe cases of bilingual aphasia (Paradis, 1998). All in all, the evidence suggests that when cognitive processes are targeted simultaneously in both languages (Lalor & Kirsner, 2001; Laganaro & Overton-Venet, 2001), transfer of therapy gains can be observed even in an untreated language, such as L4 (Filiputti et al., 2002). Finally, recent neuroimaging studies on CLT provide evidence of the training-dependent enhancement of activation patterns in bilinguals who recover from aphasia. Meinzer et al. (2007) examined an early bilingual (German-French) with chronic aphasia following a left-hemisphere stroke by exploring the neural correlates of overt picture naming in German and French at two different times: first, 32 months after the stroke, aiming to assess differential recovery as a function of the language therapy received (in German), and second, after an additional short-term intensive language training (also in German). The first assessment showed a pattern of selective naming recovery in German, concurrently with increased functional brain activation, as compared to French. Interestingly, following the additional short-term therapy in German, a bilateral reorganization of the neural substrates of naming was observed only in the treated language, whereas no changes in the activation patterns for the untreated language were evident. In contrast, Marangolo et al. (2009) reported the case of a highly proficient bilingual (Flemish, L1, and Italian, L2) woman with aphasia, who after six months of therapy in L2 showed CLT gains to the untreated L1, as she experienced a similar recovery in both languages. In accordance with the behavioral results, the same neural structures were involved in a naming task in both L1 and L2, before and after training. However, since there is great variability in the patterns of recovery from bilingual aphasia, further studies assessing the impact of therapy intensity on CLT are strongly recommended.

Future Directions and Concluding Remarks One promising avenue to examine specific language therapy effects on CLT is to combine functional magnetic resonance imaging (fMRI) studies with dynamic causal modelling (DCM) techniques (Friston et al., 2003). DCM gathers information on networks beyond the activation sites, providing a

28

Par t 1: Broad Considerat ions

more complete picture of the neural substrates subserving a specific ability than traditional fMRI activation studies. For instance, Abutalebi et al. (2009) showed that improved performance in the treated language was associated with increased activation in language processing areas, as well as increased connectedness of the control (ACC and head of caudate, among others) and language processing networks for the treated language. The study by Abutalebi et al. (2009) shows that DCM holds great promise for investigating recovery patterns and the effects of specific language treatment within and across languages in bilingual aphasia. After the years of debate regarding the efficacy of language and therapy, today there is compelling evidence suggesting that intensive therapy specifically targeted at the linguistic dysfunction is a key to aphasia recovery (Bhogal et al., 2003; Pulvermüller et al., 2001). In monolinguals, recovery from aphasia has been related to functional reorganization involving the right homologues of damaged left-hemisphere regions, the reactivation of intact areas in the dominant left hemisphere, or both (Kohnert, 2009). Bilingual aphasia awaits the exploration and description of the neuroplastic patterns that characterize its recovery. Given its multifaceted nature, targeting a broad range of variables, from motivation to pragmatics, metalinguistics and AoA, will allow both researchers and clinicians to gain more understanding of how language is processed in a healthy brain and recovered after neurological insult. For instance, several authors have proposed transcranial magnetic stimulation (TMS) as a safe non-invasive technique that can be used to induce or enhance neuroplastic changes (Antal et al., 2001). This is done by causing depolarization or hyperpolarization in the neurons of the brain. TMS uses electromagnetic induction to induce weak electric currents with a rapidly changing magnetic field; this can cause activity in specific or general parts of the brain with minimal discomfort, allowing the functioning and interconnections of the brain to be studied. Thus, the use of emerging techniques, such as TMS and DCM will allow focusing on the impact of language distance and therapy approach as factors that might modulate CLT. These appear to be good candidates for future research in this field.

References Abutalebi, J. and Green, D.W. (2007) Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics 20 (3), 242–275. Abutalebi, J., Annoni, J.M., Zimine, I., Pegna, A.J., Seghier, M.L., Lee-Jahnke, H., Lazeyras, F., Cappa, S.F. and Khateb, A. (2008) Language control and lexical competition in bilinguals: An event-related fMRI study. Cerebral Cortex 18 (7), 1496–1505. Abutalebi, J., Rosa, P.A., Tettamanti, M., Green, D.W. and Cappa, S.F. (2009) Bilingual aphasia and language control: A follow-up fMRI and intrinsic connectivity study. Brain and Language 109 (2–3), 141–156.

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

29

Aglioti, S., Beltramello, A., Girardi, F. and Fabbro, F. (1996) Neurolinguistic and follow-up study of an unusual pattern of recovery from bilingual subcortical aphasia. Brain 119, 1551–1564. Aglioti, S. and Fabbro, F. (1993) Paradoxical selective recovery in a bilingual aphasic following subcortical lesions. NeuroReport 4 (12), 1359–1362. Alexiadou, A. and Stavrakaki, S. (2006) Clause structure and verb movement in a GreekEnglish bilingual patient with Broca’s aphasia: Evidence from adverb placement. Brain and Language 96, 207–220. Ansaldo, A.I. and Marcotte, K. (2007) Language switching and mixing in the context of bilingual aphasia. In J.G. Centeno, R.T. Anderson and L.K. Obler (eds) Studying Communication Disorders in Spanish Speakers: Theoretical, Research, and Clinical Aspects (pp. 12–21). Clevedon: Multilingual Matters. Ansaldo, A.I., Marcotte, K., Scherer, L. and Raboyeau, G. (2008) Language therapy and bilingual aphasia: Clinical implications of psycholinguistic and neuroimaging research. Journal of Neurolinguistics 20 (3), 242–275. Antal, A., Nitsche, M.A. and Paulus, W. (2001) External modulation of visual perception in humans. NeuroReport 12, 3553–3555. Avila, C., Gonzalez, J., Parcet, M.A. and Belloch, V. (2004) Selective alteration of native, but not second language articulation in a patient with foreign accent syndrome. NeuroReport 15 (14), 2267–2270. Bhogal, S.K., Teasell, R. and Speechley, M. (2003) Intensity of aphasia therapy, impact on recovery. Stroke 34 (4), 987–993. Bialystok, E. (2001) Bilingualism in Development: Language, Literacy, and Cognition. New York: Cambridge University Press. Bialystok, E., Craik, F.I., Grady, C., Chau, W., Ishii, R., Gunji, A. and Pantev, C. (2005) Effect of bilingualism on cognitive control in the Simon task: Evidence from MEG. NeuroImage 24 (1), 40–49. Bialystok, E., Craik, F.I. and Ruocco, A.C. (2006) Dual-modality monitoring in a classification task: The effects of bilingualism and aging. Quarterly Journal of Experimental Psychology 59 (11), 1968–1983. Bialystok, E., Craik, F.I. and Freedman, M. (2007) Bilingualism as a protection against the onset of symptoms of dementia. Neuropsychologia 45 (2), 459–464. Centeno, J.G. (2005) Working with bilingual individuals with aphasia: The case of a Spanish-English bilingual client. Perspectives on Communication Disorders and Sciences in Culturally and Linguistically Diverse Populations 12, 2–7. Centeno, J.G. (2009) Serving bilingual patients with aphasia: Challenges, foundations, and procedures. Revista de Logopedia, Foniatría y Audiologia 29, 30–36. Crinion, J., Turner, R., Grogan, A., Hanakawa, T., Noppeney, U., Devlin, J.T., Aso, T., Urayama, S., Fukuyama, H., Stockton, K., Usui, K., Green, D.W. and Price, C.J. (2006) Language control in the bilingual brain. Science 312 (5779), 1537–1540. Edmonds, L.A. and Kiran, S. (2006) Effect of semantic naming treatment on crosslinguistic generalization in bilingual aphasia. Journal of Speech and Language Hearing Research 49 (4), 729–748. Fabbro, F. (1999) The Neurolinguistics of Bilingualism: An Introduction. Hove: Psychology Press. Fabbro, F. (2001a) The bilingual brain: Bilingual aphasia. Brain and Language 79 (2), 201–210. Fabbro, F. (2001b) The bilingual brain: Cerebral representation of languages. Brain and Language 79 (2), 211–222. Fabbro, F., Skrap, M. and Aglioti, S. (2000) Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery and Psychiatry 68 (5), 650–652.

30

Par t 1: Broad Considerat ions

Faroqi-Shah, Y., Frymark, T., Mullen, R. and Wang, B. (2010) Effect of treatment for bilingual individuals with aphasia: A systematic review of the evidence. Journal of Neurolinguistics 23 (4), 319–341. Filiputti, D., Tavano, A., Vorano, L., De Luca, G. and Fabbro, F. (2002) Nonparallel recovery of languages in a quadrilingual aphasic patient. International Journal of Bilingualism 6, 395–410. Friston, K.J., Harrison, L. and Penny, W. (2003) Dynamic causal modelling. NeuroImage 19 (4), 1273–1302. Galvez, A. and Hinckley, J. (2003) Transfer patterns of naming treatment in a case of bilingual aphasia. Brain and Language 87 (1), 173–174. Gil, M. and Goral, M. (2004) Nonparallel recovery in bilingual aphasia: Effects of language choice, language proficiency, and treatment. International Journal of Bilingualism 8, 191–219. Goral, M., Levy, E.S., Obler, L.K. and Cohen, E. (2006) Cross-language lexical connections in the mental lexicon: Evidence from a case of trilingual aphasia. Brain and Language 98 (2), 235–247. Goral, M., Levy, E.S. and Kastl, R. (2010) Cross-language treatment generalisation: A case of trilingual aphasia. Aphasiology 103 (1–2), 203–204. Green, D.W. (1986) Control, activation, and resource: A framework and a model for the control of speech in bilinguals. Brain and Language 27 (2), 210–223. Green, D.W. (1998) Mental control of the bilingual lexico-semantic system. Bilingualism, Language and Cognition 1, 67–81. Green, D.W. (2005) The neurocognition of recovery patterns in bilingual aphasics. In J.F. Kroll and M.B. de Groot (eds) Handbook of Bilingualism: Psycholinguistic Perspectives (pp. 516–530). New York: Oxford University Press. Green, D.W., Grogan, A., Crinion, J., Ali, N., Sutton, C. and Price, C.J. (2009) Language control and parallel recovery of language in individuals with aphasia. Aphasiology 24 (2), 188–209. Hernandez, A.E., Martinez, A. and Kohnert, K. (2000) In search of the language switch: An fMRI study of picture naming in Spanish-English bilinguals. Brain and Language 73 (3), 421–431. Hernandez, A.E., Dapretto, M., Mazziotta, J. and Bookheimer, S. (2001) Language switching and language representation in Spanish-English bilinguals: An fMRI study. NeuroImage 14 (2), 510–520. Hyltenstam, K. and Stroud, C. (1989) Bilingualism in Alzheimer’s dementia: Two case studies. In K. Hyltenstam and L.K. Obler (eds) Bilingualism across the Lifespan. Aspects of Acquisition, Maturity, and Loss (pp. 202–226). Cambridge: Cambridge University Press. Kim, K.H.S., Relkin, N.R., Lee, K.M. and Hirsch, J. (1997) Distinct cortical areas associated with native and second languages. Nature 388, 171–174. Kiran, S. and Tuchtenhagen, J. (2005) Imageability effects in normal Spanish-English bilingual adults and in aphasia: Evidence from naming to definition and semantic priming tasks. Aphasiology 19, 315–327. Klein, D., Zatorre, R.J., Chen, J.K., Milner, B., Crane, J., Belin, P. and Bouffard, M. (2006) Bilingual brain organization: A functional magnetic resonance adaptation study. NeuroImage 31 (1), 366–375. Kohnert, K. (2004) Cognitive and cognate-based treatments for bilingual aphasia: A case study. Brain and Language 91 (3), 294–302. Kohnert, K. (2009) Cross-language generalization following treatment in bilingual speakers with aphasia: A review. Seminars in Speech and Language 30 (3), 174–186. Laganaro, M. and Overton-Venet, M. (2001) Acquired alexia in multilingual aphasia and computer-assisted treatment in both languages: Issues of generalisation and transfer. Folia Phoniatrica et Logopaedica 53 (3), 135–144.

Bilingual Aphasia : Neural Pl ast ic it y and Considerat ions for Recover y

31

Lalor, E. and Kirsner, K. (2001) The representation of ‘false cognates’ in the bilingual lexicon. Psychonomic Bulletin and Review 8 (3), 552–559. Longworth, C.E., Keenan, S.E., Barker, R.A., Marslen-Wilson, W.D. and Tyler, L.K. (2005) The basal ganglia and rule-governed language use: Evidence from vascular and degenerative conditions. Brain 128 (Pt 3): 584–596. Lorenzen, B. and Murray, L.L. (2008) Bilingual aphasia: A theoretical and clinical review. American Journal of Speech-Language Pathology 17 (3), 299–317. Marangolo, P., Rizzi, C., Peran, P., Piras, F. and Sabatini U. (2009) Parallel recovery in a bilingual aphasic: A neurolinguistic and fMRI study. Neuropsychology 23 (3), 405–409. Mariën, P., Abutalebi, J., Engelborghs, S. and De Deyn, P.P. (2005) Pathophysiology of language switching and mixing in an early bilingual child with subcortical aphasia. Neurocase 11 (6), 385–398. Meinzer, M., Obleser, J., Flaisch, T., Eulitz, C. and Rockstroh, B. (2007) Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI. Neuropsychologia 45 (6), 1247–1256. Miller E.K. and Cohen J.D. (2001) An integrative theory of prefrontal cortex function. Annual Reviews in Neuroscience 24, 167–202. Muñoz, M.L. and Marquardt, T.P. (2003) Picture naming and identification in bilingual speakers of Spanish and English with and without aphasia. Aphasiology 17 (12), 1115–1132. Muñoz, M.L., Marquardt, T.P. and Copeland, G. (1999) A comparison of the code switching patterns of aphasic and neurologically normal bilingual speakers of English and Spanish. Brain and Language 66 (2), 249–274. Paradis, M. (1998) Pragmatics in Neurogenic Communication Disorders. Oxford: Pergamon Press. Paradis, M. (2001) Bilingual and polyglot aphasia. In R.S. Berndt (ed.) Handbook of Neuropsychology (2nd ed., Vol. 3, pp. 69–91). Amsterdam: Elsevier Science. Paradis, M. (2004) A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins Publishing. Paradis, M. (2008) Language and communication disorders in multilinguals. In B. Stemmer and H.A. Whitaker (eds) Handbook of the Neuroscience of Language (pp. 341– 349). Boston, MA/Amsterdam: Academic Press/Elsevier. Paradis, M. and Libben, G. (1987) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum Associates. Price, C.J., Green, D.W. and von Studnitz, R. (1999) A functional imaging study of translation and language switching. Brain 122 (12), 2221–2235. Pulvermüller, F., Neininger, B., Elbert, T., Mohr, B., Rockstroh, B., Koebbel, P., Elbert, T. and Taub, E. (2001) Constraint-induced therapy of chronic aphasia after stroke. Stroke 32 (7), 1621–1626. Roberts, P. and Deslauriers, L. (1999) Picture naming of cognate and non-cognate nouns in bilingual aphasia. Journal of Communication Disorders 32, 1–23. Rubinstein, J.S., Meyer, D.E. and Evans, J.E. (2001) Executive control of cognitive processes in task switching. Journal of Experimental Psychology: Human Perception and Performance 27, 763–797. Ullman, M.T. (2004) Contributions of memory circuits to language: The declarative/ procedural model. Cognition 92 (1–2), 231–270. Ullman, M.T. (2006) Is Broca’s area part of a basal ganglia thalamocortical circuit? Cortex 42, 480–485. Ullman, M.T. and Pierpont, E.I. (2005) Specific language impairment is not specific to language: The procedural deficit hypothesis. Cortex 41 (3), 399–433. Veroude, K., Norris, D.G., Shumskaya, E., Gullberg, M. and Indefrey, P. (2010) Functional connectivity between brain regions involved in learning words of a new language. Brain and Language 113 (1), 21–27.

32

Par t 1: Broad Considerat ions

Vilarino, I., Prieto, J.M., Robles, A., Lema, M. and Noya, M. (1997) A study of bilingual Galician-Castilian aphasic patients. Revista de Neurología 25 (144), 1165–1167. Vingerhoets, G., Van Borsel, J., Tesink, C., van den Noort, M., Deblaere, K., Seurinck, S., Vandemaele, P. and Achten, E. (2003) Multilingualism: An fMRI study. NeuroImage 20 (4), 2181–2196. Wald, I. (1961) Problema afazii polyglotov. In Voprosy Kliniki i Patofiziologii Afazii (pp. 140–176). Moscow. Wartenburger, I., Heekeren, H.R., Abutalebi, J., Cappa, S.F., Villringer, A. and Perani, D. (2003) Early setting of grammatical processing in the bilingual brain. Neuron 37 (1), 159–170. Zanini, S., Tavano, A., Vorano, L., Schiavo, F., Gigli, G.L., Aglioti, S.M. and Fabbro, F. (2004) Greater syntactic impairments in native language in bilingual Parkinsonian patients. Journal of Neurology, Neurosurgery, and Psychiatry 75 (12), 1678–1681.

Part 2 Assessment and Treatment

3

What Do We Know About Assessing Language Impairment in Bilingual Aphasia? Swathi Kiran and Patricia M. Roberts1

Introduction Several advances have been made in the assessment of bilingual aphasia to better meet the needs of bilingual (and multilingual) adults who have a stroke and resulting aphasia. There is growing awareness of the need to assess these patients in a valid manner. However, it is not a simple task, for several reasons including the complexity of bilingualism and great range of possible levels across different language modalities. Additional factors include pre-stroke language proficiency, balance of proficiency across languages, age of acquisition, prior education/work history, phonological structure of each language, nature of aphasia type, severity, site of lesion and size of lesion. This chapter provides insights into the challenges faced by theoreticians and clinicians in understanding how to address the needs of bilingual patients with aphasia. The first section provides an overview of the problem by presenting a framework for examining the complex nature of assessment in bilingual aphasia. The second section discusses approaches to assess prestroke language proficiency and provides practical examples on how to do so. The third and final section provides some suggestions on utilizing alternate assessments that take into account each patient’s individual pre-stroke proficiency in the two languages. 35

36

Par t 2: A ssessment and Treatment

Challenges in the Assessment of Bilingual Patients Models of non-damaged bilingual processing In research, models of language processing and production are often used to develop experimental tasks. Studies examining bilingualism have been guided by theoretical models in word comprehension (Dijkstra et al., 2005; Schulpen et al., 2003), lexical-semantic processing (Kroll & Stewart, 1994; Kroll et al., 2005), and syntactic processing (Heredia & Altarriba, 2002) to name a few. Debates continue about appropriate and accurate models for different types of speakers and new data is gradually leading to more specific and more powerful models. For instance, one influential model of bilingual word production is the Revised Hierarchical Model (RHM) (Kroll & Stewart, 1994), which specifies connections between both L1 and L2 and a central conceptual system. These connections differ in their strengths as a function of fluency in L1 relative to L2. In another model, De Groot’s distributed model (de Groot, 1992), word meanings are represented as a set of features and the overlap in the meaning of two translations is determined by the number of features that are shared. There is less distinction between the L1 and L2 maps in the distributed features model when compared to the RHM but this model specifies the precise mapping of words-concepts that vary by imageability, frequency, category, ambiguity and so forth. The two models differ in their interpretation of the mechanisms underlying the influence of various psycholinguistic factors on lexical-semantic processing in bilingual individuals. Importantly, these and other models have generated studies that continue to expand our knowledge of language processing in bilinguals. In addition, our understanding of processes involved in language comprehension and production in aging populations has also increased considerably. Recent studies have begun to explore the interactions between aging and aspects of bilingualism that could be relevant to bilingual aphasia (Goral, 2004; Goral et al., 2008). A better understanding of language attrition, changes in processing strategies and even changes in the brain regions involved in language in older adults will be relevant to aphasia, given the age groups at risk for strokes. While these studies provide important points of comparison in bilingual aphasia, it is not clear how relevant they may be for the interpretation of patterns of impairment in aphasia. Aphasia flattens out differences in speed that are used to infer steps in a given task. Certainly, aphasia changes the processing ‘routes’ posited by the models. It is often said that no two patients with aphasia show exactly the same pattern (what is preserved and what is impaired). This is doubly true for bilingual patients. The individual variability influencing patterns of recovery is generally seen as the result of the interaction of factors that include, but are not limited to: pre-stroke language proficiency, balance of proficiency across languages, ages of acquisition of

A ssessing L anguage Impair ment in Bilingual Aphasia

37

each language, patterns of language use pre-stroke, phonological structure of each language, combined with the factors responsible for the great case-bycase differences seen across monolingual patients: aphasia type, aphasia severity, site of lesion and size of lesion. The other issue related to models of bilingual language processing is the problem of control (e.g. Abutalebi, 2008; Green, 1998). Much is known about the role of executive function and inhibition of the non-target language during specific tasks (e.g. Festman et al., 2010). If some of the problems in bilingual aphasia are related to selection and/or selective attention deficits caused by disruption to executive abilities, then the behavioral impairment will depend on both the nature of the patient’s language deficits and on the degree and nature of the often subtle cognitive deficits that affect control processes. With so many interacting factors, the assumptions that underlie models of normal processing are not met, and it may be too early in our study of bilingual aphasia to attempt to apply these models of unimpaired bilingual systems to real clinical cases (Roberts, 2010). Before clinical applications are possible, we need to develop and thoroughly test models of bilingual aphasia (e.g. Kiran et al., submitted).

Variability in bilingual speakers The assessment of bilingual speakers is made difficult by the great range of abilities. As many have pointed out (e.g. Hakuta, 1986; Roberts, 2001), bilingualism is a continuum, with each person at some point along that continuum for the different language modalities of auditory comprehension, verbal expression, reading and written expression. There is no common, predictable pattern or time course for acquisition and no common set of ‘things’ (words, syntactic structures) that each person learns. Grosjean’s Complementarity Principle (Grosjean, 1998) reminds us that bilinguals do not have ‘two complete sets’ of everything that a unilingual speaker has in their only language: only vocabulary and syntax needed for specific situations are acquired in each language, as it becomes necessary to know them. These two facts (greater differences between bilingual speakers than between unilingual speakers of the same languages and the very idiosyncratic and unpredictable acquisition process for bilingual speakers) are often noted in the literature. Roberts (1998, 2001) has pointed out the lack of available tests and the flawed or, in some cases, unknown psychometric properties of many of the available tests. What are the consequences of what Roberts (2008) has dubbed ‘the test mess’ for clinicians confronted with the very concrete problem of how to assess their 10 am Spanish-English speaking patient? Lorenzen and Murray (2008) repeat Roberts’ (1998) call for the development of better tests (and therapies) stating: ‘there has been limited

38

Par t 2: A ssessment and Treatment

development and critical evaluation of assessment and treatment protocols designed to meet the unique characteristics and needs of bilingual clients with aphasia’ (Lorenzen & Murray, 2008: 312).

A Functional Framework for Assessment Normative or functional frameworks In her classic textbook on neuropsychological assessment, Lezak (1995) identifies several types of frameworks commonly used as the basis for test design and interpretation. Although she was not writing about bilingual speakers, her framework for assessment seems pertinent. There are species-wide performance expectations for some abilities, including language. For these, it is appropriate to apply norms, to compare each patient to an expected level of performance. Norms and norm-referenced tests can be used for all cognitive functions and skills that follow a common course of development, that are usually fully developed long before adulthood, and that are taken for granted as part and parcel of the normal adult behavioral repertory. Speech is a good example . . . anything less than an acceptable performance in an adult raises the suspicion of impairment. (Lezak, 1995: 99) This is true, but only for unilingual speakers. As Lezak notes, ‘as a general rule normative standards are only appropriate when the function or skill or capacity that is being measured is well within the capability of all intact adults and does not vary greatly with age, sex, education, or general mental ability’ (Lezak, 1995: 100). A key point in what we are proposing in this chapter is that a normative framework is not appropriate for the assessment of bilingual adults with aphasia. There is no possible ‘norm’ or pre-stroke level of knowledge that can be assumed. Empirical studies in three pairs of languages demonstrate that even aiming for a very low ceiling, as in the Bilingual Aphasia Test (Paradis & Libben, 1987) does not ensure that ‘all speakers’ will perform at or above 90%. The Bilingual Aphasia Test includes items and subtests where this is not true for individuals without aphasia (Ivanova & Hallowell, 2009; Munoz & Marquardt, 2008; Roberts et al., submitted). Some individuals also fail to perform at the low ceiling level set for the Pyramids and Palm Trees Test (Gudayol-Ferre et al., 2008). Even tests that are norm-referenced require modification to allow for regional dialects, and it is by no means a given that each clinician would have the required expertise to navigate words such as ‘straw’: ‘pitillo’ in Venezuela, ‘pajilla’ or ‘popote’ in many other countries. Likewise, ‘zipper’ is known as ‘cierre’ in Uruguay, ‘cierrede cremallera’ in Argentina, Spain, Peru and Uruguay,

A ssessing L anguage Impair ment in Bilingual Aphasia

39

and as ‘zíper’ in Cuba, Puerto Rico and Panama. These examples also demonstrate that, depending upon the dialect spoken, a given stimulus word can be a cognate (for a bilingual speaker) or not. The cognateness of words in a pair of languages can affect task difficulty and cognitive processing strategies (Colomé & Miozzo, 2010; Lalor & Kirsner, 2001; Roberts & Deslauriers, 1999). Depending on the dialect spoken, a picture name may have one, two or three syllables. In cases where the behavior to be assessed does not have a uniform, species-wide expected level, Lezak recommends something she calls ‘individual comparison standards’. Lezak states, ‘A first step in measuring cognitive deficits in an adult is to establish – or estimate, when direct information is not available – the patient’s premorbid performance level for all of the functions and abilities being assessed’ (1995: 101).

Estimating pre-stroke language proficiency There are a number of reasons to want to estimate pre-stroke proficiency. Paradis’ classifications of impairment in bilingual aphasia (Paradis, 1989, 1993, 2000) (note, not recovery, but degree of impairment) rely on having accurate information about pre-stroke proficiency. Planning treatment also requires estimating what the patient was able to do pre-morbidly, since it is inappropriate for the clinician to fall into the role of language teacher, working on things that were not in the patient’s pre-stroke repertoire. There are several ways to estimate a patient’s level of mastery of each language. Roberts (2008) proposes four types: self-ratings of proficiency, ratings by family members, acquisition history and patterns of use for each language. We will consider each of these in turn. Each of these four types is incomplete and imperfect and can only make a meaningful contribution when integrated with other measures. There is no way to know the accuracy of any given self-report or a family member’s memory of past habits. Using several sources and asking more than one question will provide a clearer and more accurate profile. For example, if an American patient or research volunteer says they lived in Europe for two years, but says they speak only English, this needs further exploration. Perhaps they have forgotten to mention knowing a basic level of another language learned during this time OR perhaps they learned no other language and the original statement is accurate. A second example: if someone says they began learning a language at age 20 and also claims to have native-like proficiency in this language, this is a very unlikely combination. Perhaps they had quite a bit of exposure to that language prior to age 20 – via popular music, movies, TV, reading, or friends, and when they say ‘began to learn at age 20′ they mean that was the age when they took their first formal language learning class. Alternatively, perhaps their self-rating is unrealistically high. Do they speak with an accent? Can they understand speakers with different accents (one of the signs of very

40

Par t 2: A ssessment and Treatment

high levels of auditory comprehension abilities)? If they have native-like abilities (seven out of seven), this normally is due to having spent quite a bit of time living with and using a language. Look for this in their language history by asking if they lived in a city or country where this language was widely used; whether they attended French (or other language) immersion schools for a number of years, etc. The diagnostic interview skills all clinicians use to obtain a valid case history, including identifying and exploring contradictions or gaps in the history, are very useful in obtaining a valid picture of a person’s level of bilingualism.

Self-ratings Patients who can understand a numerical rating scale and the description of what the end points mean can use a five-point or seven-point scale. An alternative version of the same task is a line where the patient indicates where they place their pre-stroke abilities: one end of the line being ‘no ability’ and the other end being ‘perfect, like a native speaker’. While many studies of individuals without aphasia use this type of self-rating and it has been shown to correlate, on a group level with performance on a number of tasks (e.g. Kohnert et al., 1998; Roberts et al., 2002; Videsott et al., 2010), it cannot be used as an absolutely accurate ‘score’ for individuals, especially for those with aphasia.

Ratings by family members The emotional impact of the aphasia, faulty memories and the unknown validity of such ratings for individual cases are reasons to ask family members to perform similar ratings. As pointed out by Roberts (2001, 2008), the fact that ratings are influenced by each person’s own level of proficiency makes it useful to ask more than one family member to rate the patient’s previous levels. The patient’s spouse and an adult child, or anyone very familiar with the patient, can provide valuable ratings using the same numerical or visual scale used with the patient. Explore any contradictions through discussions with the patient and the family. Some patients claim relatively low levels of pre-stroke ability (that do not match what the family reports and do not line up logically with the patterns of language use), apparently to lessen the shock of the aphasia. By claiming little pre-stroke ability, they can console themselves that relatively little has been lost. Other patients appear to overestimate their pre-stroke abilities, and seem to feel the degree of loss very acutely.

Language acquisition history It is important to understand at what age each language was introduced in the environment. This is because of correlations between age (and amount) of exposure and the likely proficiency level. For example, a person could be exposed only to Spanish between ages zero to three (indicating that

A ssessing L anguage Impair ment in Bilingual Aphasia

41

Spanish is their native language, L1) however, by age nine, the exposure to English may increase with a resulting balanced exposure in each language. This example highlights the fact that acquisition occurs over time, and not at a specific age. We recommend ‘age of first exposure’ or ‘age acquisition began’ rather than the commonly used, but misleading ‘age of acquisition’.

Patterns of use for each language Another important piece of information is the degree of use of each language at different stages during education, family life and work. As illustrated in the previous example, a person may be exposed to Spanish for three quarters of their time during the early years and once the individual starts elementary school, the exposure to English may increase, resulting in equal exposure to both languages. This may change further during adult years where the work environment may determine what language is used. In addition, in many countries, the medium of instruction may not always be the preferred language of the individual. Or the language of instruction may be the formal, standard version of a language such as Arabic or French, and very different from the colloquial dialect spoken by the patient and/or family. Obtain this information systematically. Some questionnaires for research purposes ask for a great deal of information, including such things as a time log to be filled in for several days documenting which language the patient was using hour by hour during the day (e.g. Kiran et al., 2010), while others, more feasible in clinical practice, ask for less information. As is always the case, compromises must be made between the most thorough and (potentially) sensitive/valid set of questions and the time available to complete questionnaires, the impact of fatigue or impatience with lengthy forms or interviews. An interesting question is the relative importance of the patterns of use in childhood or in the 10 or so years before the stroke or across all decades. It is not clear whether any one of these time periods is more relevant to understanding the patient’s deficits, prognosis and treatment needs. Until there is research addressing this question, it is probably best to gather information on all time periods. Finally, it is important to get an estimate of the individual’s usage of each language in terms of time spent in each language (calculated as a percentage) during a typical weekday and weekend. To illustrate the value of obtaining a comprehensive picture of each individual’s pre-morbid language skills, we provide contrasting examples of four Spanish-English bilingual patients with aphasia tested in our laboratory using a fairly detailed language questionnaire (Kiran et al., 2010). This questionnaire obtains information about the period of age of language acquisition, and a proportion of language exposure in hearing, speaking and reading domains during the entire lifetime for each individual. A weighted average of

42

Par t 2: A ssessment and Treatment

the proportion of exposure across the lifespan in the three domains is obtained for each language. Next, participants fill out a detailed educational history form where they are asked to provide the language of instruction in each language and the language that they and their peers prefer during conversations. Also, participants estimate the time spent conversing in each language hour by hour during a typical weekday and typical weekend. A weighted average of this score reflects the proportion of time spent conversing in one language versus another language. Finally, participants are required to self-rate their proficiency in each language in terms of their ability to speak and understand the language in formal and informal situations. Again, an average proportion score in each language reflects participants’ perception of their own language proficiency. Based on this questionnaire, data from four patients AB, CD, EF and GH are shown in Table 3.1. AB began learning English at approximately age six and was exposed to English in elementary and high school but was exposed to more English than Spanish and spent more time conversing in English than Spanish prior to his stroke. CD, on the other hand, began learning English later in life (around 17 years), was educated in Spanish in elementary school and exposed to some English when in high school. This patient’s learning history indicates that CD was likely exposed to more Spanish (60%) than English (40%) and spoke more Spanish (100%) than English during a typical day. Both AB and CD, however, considered themselves to be equally proficient in the two languages reflecting an apparent mismatch between actual language exposure and perceived sense of proficiency in a language. If CD began learning English after age 16, it is extremely unlikely that he attained equal mastery of English and Spanish. This is the kind of contradiction mentioned earlier in the chapter; further exploration with follow-up questions is necessary, such as ‘what city/country did you grow up in (to see if English would have been spoken by many people)? did you ever listen to English TV or radio? If so, starting around what age (to see if the age of first exposure was, in fact, much younger than 16 years old)?’ EF also began learning English after childhood (approximately 20 years) but was not formally educated beyond elementary school. Consequently, this patient, with limited education in Spanish, never learned to read or write in English. Prior to the stroke, EF reported greater exposure to Spanish than English and also reported spending more time conversing in Spanish (78%) than in English (22%). Finally, GH learned both languages early in life and was educated in both languages until college. GH reported exposure to more English (61%) than Spanish (39%) during her life and also spent more time conversing in English (53%) than Spanish (46%). Even though GH appears to be a more balanced bilingual than the other three patients, based on exposure to and use of the two languages, she estimated her own pre-stroke proficiency in English to be higher than in Spanish. This discrepancy drives home the

A ssessing L anguage Impair ment in Bilingual Aphasia

43

Table 3.1 Proficiency self-ratings and test scores from four sample patients that illustrate the difficulty in assessing impairment scores. See text for details. Patient AB

Patient CD

Patient EF

Patient GH

Age began learning language English

6 years

17 years

20 years

0 years

Spanish

0 years

0 years

0 years

0 years

Estimated language exposure over lifetime English

75%

40%

28%

61%

Spanish

25%

60%

72%

39%

Education English

Elementary/ High school high school

Spanish

No formal education

No formal Elementary/ education middle/high

Elementary/ Elementary high school

Elementary/ middle/high

Average time spent conversing in language during weekday/weekend English

69%

0%

22%

53%

Spanish

31%

100%

78%

46%

Self rated proficiency (on a 5-point scale) English

100%

100%

86%

93%

1. Speaking

5

5

5

5

2. Listening

5

5

5

5

3. Reading

5

5

3

5

4. Writing

5

5

3

5

100%

100%

86%

73%

1. Speaking

5

5

5

4

2. Listening

5

5

5

4

3. Reading

5

5

3

4

4. Writing

5

5

3

3

English

53%

31%

37%

6%

Spanish

6%

33%

43%

5%

Spanish

Scores on BNT

Comprehension BAT (semi-complex commands) English

90%

80%

70%

70%

Spanish

100%

80%

80%

80%

Notes: BNT: Boston Naming Test; BAT: Bilingual Aphasia Test. The BAT and the BNT have not been shown to be of equal difficulty in English and Spanish. Scores must be compared and interpreted accordingly.

44

Par t 2: A ssessment and Treatment

point we made earlier, that self rating of language proficiency cannot be the only information on bilingualism obtained from a patient. Clearly, in this case, self assessment of language proficiency is not completely reflective of that individual’s learning history, background and current and pre-stroke usage of the languages in different contexts. Self-ratings of language proficiency need to be supplemented with estimates of language use and background information about acquisition and domains of use. Additionally, a contentious debate in the field of aphasia has always been the relationship between lesion, deficit and behavior. It is still unclear whether a specific lesion in a certain brain region in the left hemisphere results in a specific, observable linguistic deficit (Caplan, 2004). Further, in bilingual aphasia, the question of to what extent language impairment is influenced by the nature of recovery patterns has been the focus of quite a number of studies that report a greater loss in one language than another (differential impairment, relative to estimated pre-stroke proficiency) (Fabbro et al., 2000; Galvez & Hinckley, 2003; Laganaro & Overton Venet, 2001; Lalor & Kirsner, 2001; Meinzer et al., 2007; Moretti et al., 2001). It is however, impossible to completely understand to what extent impairment and the subsequent recovery of language are influenced by that individual’s language knowledge in each of the two languages. It is likely that L2 acquisition and level of pre-stroke proficiency may determine the extent of poststroke deficit and ultimate rehabilitation outcomes, but we have very little data to either support or refute this assumption (Kiran & Roberts, 2010). The patients described above and in Table 3.1 demonstrate the variability in language background and, consequently, the difficulty in making any assumptions about the nature of impairment without understanding prestroke language knowledge. For instance, based on the BNT naming scores, patient AB appears to have a more severe naming impairment in Spanish compared to English, however, there are several possible explanations for the higher score in English. These include: (1) unequal difficulty of the test itself in English and Spanish; (2) parallel impairment which has affected the two languages roughly equally, leaving English vocabulary knowledge much better than Spanish, as it was before the stroke; (3) a mismatch between the items tested and the patient’s knowledge (based on topics and activities experienced in each language and the lack of formal education in Spanish). A lower score on the Spanish BNT test does not necessarily indicate a greater impairment (loss due to aphasia) in Spanish. Likewise, even though CD and GH appear to have similar scores on the English and Spanish BNT test, their pre-stroke proficiency is quite different in the two languages. It is possible that the relative loss is more in the stronger language than the weaker language. In a preliminary assessment of 21 patients with bilingual aphasia, we have found an interesting relationship between pre-stroke proficiency and post-stroke impairment in naming skills (Gray & Kiran, submitted).

A ssessing L anguage Impair ment in Bilingual Aphasia

45

Which Assessments Can Reliably Measure Impairment in Bilingual Aphasia? Understanding the relationship between pre-stroke proficiency and poststroke impairment is essential to make an accurate diagnosis of language impairment in bilingual aphasia. First, as pointed out earlier, the individual’s language status prior to stroke can never be known precisely; we can only estimate likely pre-stroke levels. In addition, many standard published tests lack age, socio-cultural and language norms. Given these issues, we suggest two alternate assessment tasks that have the advantage of being less constrained in the way that standard published tests are. In these relatively open-ended tasks, patients will be able to demonstrate their abilities in a different way than in closed tasks, where there is only one correct answer. We make these recommendations recognizing that the two proposed tasks (verbal fluency and narration) do not replace standard aphasia tests that sample a much wider range of behaviors. These tasks have been useful in our own work. Published research supports their validity – in other contexts than the one we are proposing. We look forward to seeing studies testing their relevance, validity and sensitivity to changes in ability during recovery of bilingual patients with aphasia, in both their stronger and their weaker language. Verbal Fluency: The verbal fluency task has two commonly used forms: semantic-based and phonetic/letter-based. It goes by different names, including category definition, COWAL, the Set Test, the Thurstone Word Fluency Test. Roberts and Le Dorze (1997) pointed out several features of this test that make it suitable for use with bilingual patients. These tests have been applied widely to examine lexical semantic access in normal aging and neurologically impaired individuals (e.g. Standish et al., 2007). In recent years, studies of the verbal fluency task in several languages have appeared, including Spanish (Rosselli et al., 2000; Rosselli et al., 2002), French (Raoux et al., 2010; Roberts & Le Dorze, 1997; Sauzéon et al., 2004), Hebrew (Kavé, 2006), Finnish (Hänninen et al., 1995; Pekkala et al., 2009), and Dutch (Oberg & Ramirez, 2006; Schmand et al., 2008; Van der Elst et al., 2006). Some studies have reported differences between monolinguals and bilinguals (Rosselli et al., 2000) with a few studies reporting that bilinguals produce significantly fewer responses in all tasks, including letter, proper name fluency and semantic generation (Gollan et al., 2002). Other studies have reported differences in semantic organization and clustering strategies in different semantic categories across languages (Roberts & Le Dorze, 1997). The verbal fluency task allows clinicians and researchers to quickly do two things: (1) to provide a number, a score in each language which provides a rough indication of the lexical production abilities in each language, poststroke (note, however, that this is not the same thing as drawing conclusions

46

Par t 2: A ssessment and Treatment

about what degree of loss each language has undergone); (2) to measure change over time, keeping in mind that there will be a slight improvement for some patients due to a practice effect (Roberts & Le Dorze, 1994). Because the semantic and phonemic/letter versions of the verbal fluency task draw on different processing and may activate different regions of the brain, we recommend testing at least two semantic categories and at least two letters/ initial sounds (see Roberts & Tougas, 2012 for more on this). Narrative Output: Examining narrative output may also be another approach to obtaining an unconstrained sample of language output from patients that can be easily implemented in two languages (Muñoz & Marquardt, 2004; Pavlenko, 2008). We recommend not constraining the patient to a single picture, where they must use certain nouns and verbs to adequately describe the picture, and where there is a high risk of eliciting lists of nouns. This task has been used in a body of studies of monolingual aphasia (Ernest-Baron et al., 1987; Saffran et al., 1989). Procedural discourse is part of some cognitive assessments for dementia patients. Instead of a specific task, a choice from the following may be offered: (a) making a common food (tea, omelet, cookies, paella, etc.); (b) how to play a sport the patient indicates they know (do not ask the Peruvian immigrant who has been in Canada for two months to explain ‘all about hockey’); (c) tell me all about your job (if the patient used both languages at work) or your best trip (real or imagined). Narrating a story from a set of pictures (as in the six pictures used in the Bilingual Aphasia Test) is another option for this type of task. An improvement over a single story (which requires specific vocabulary about birds, nests, climbing, hospital, etc.) would be to present patients with three or four pictured stories, from which they must narrate two. As with the verbal fluency task, a sample that asks for more than one topic is likely to give a more valid picture of the patient’s abilities than a single task. Narrative tasks (with or without pictures as the starting point) allow measures of syntactic, morphosyntactic, lexical-semantic, pragmatic and discourse abilities. There are several scoring/rating systems for bilingual narratives in children (e.g. Fiestas & Pena, 2004; Lofranco et al., 2006). We need more studies of this task in the context of bilingual adults with no disorders (Foster & Tavakoli, 2009) and bilingual adults with aphasia (Muñoz & Marquardt, 2004).

Conclusion To summarize, the field of bilingual aphasia has evolved considerably in the last decade; however, diagnosis and treatment of bilingual aphasia are still

A ssessing L anguage Impair ment in Bilingual Aphasia

47

based more on clinical intuitions and trial and error than they are on research. As we have discussed, pre-stroke knowledge, post-stroke impairment and recovery are inter-related in bilingual aphasia making it difficult to examine any one issue without considering the influence of and interaction among multiple factors. In this chapter, we provided suggestions to measure and evaluate proficiency and impairment in individuals with bilingual aphasia as a first step towards laying out a road map for a systematic and inclusive approach to assess and treat individuals with bilingual aphasia. We hope that the suggestions in this chapter will be tested in future studies.

Note (1) The author order is alphabetical. Both authors contributed equally to this chapter.

References Abutalebi, J. (2008) Neural aspects of second language representation and language control. Acta Psychogica 128, 466–478. Caplan, D. (2004) The neuro in cognitive neuropsychology [Comment/Reply]. Cognitive Neuropsychology 21 (1), 17–20. Colomé, A. and Miozzo, M. (2010) Which words are activated during bilingual word production? Journal of Experimental Psychology: Learning, Memory, and Cognition 36, 96–109. doi: 10.1037/a0017677 de Groot, A. (1992) Determinants of word translation. Journal of Experimental Psychology: Learning, Memory and Cognition 18, 1001–1018. Dijkstra, T., Kroll, J.F. and de Groot, A.M.B. (2005) Bilingual visual word recognition and lexical access. In Handbook of Bilingualism: Psycholinguistic Approaches (pp. 179–201). New York: Oxford University Press. Ernest-Baron, C., Brookshire, R. and Nicholas, L. (1987) Story structure and retelling of narratives by aphasic and non-brain-damaged adults. Journal of Speech & Hearing Research 30 (1), 44–49. Fabbro, F., Skrap, M. and Aglioti, S. (2000) Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery, and Psychiatry 68 (5), 650–652. Festman, J., Rodriguez-Fornells, A. and Munte,T.J. (2010) Individual differences in control of language interference in late bilinguals are mainly related to general executive abilities. Behavioral and Brain Functions 6 (5). Online at http://www.behavioraland brainfunctions.com/content/6/1/5 Fiestas, C.E. and Pena, E.D. (2004) Narrative discourse in bilingual children: Language and task effects. Language Speech and Hearing Services in the Schools 35 (2), 155–168. Foster, P. and Tavakoli, P. (2009) Native speakers and task performance: Comparing effects on complexity, fluency, and lexical diversity. Language Learning 59 (4), 866–896. Galvez, A. and Hinckley, J. (2003) Transfer patterns of naming treatment in a case of bilingual aphasia. Brain and Language 87 (1), 173–174. Gollan, T.H., Montoya, R.I. and Werner, G.A. (2002) Semantic and letter fluency in Spanish-English bilinguals. Neuropsychology 16 (4), 562–576. Goral, M. (2004) First-language decline in healthy aging: Implications for attrition in bilingualism. Journal of Neurolinguistics 17 (1), 31–52. Goral, M., Libben, G., Obler, L.K., Jarema, G. and Ohayon, K. (2008) Lexical attrition in younger and older bilingual adults. Clinical Linguistics & Phonetics 22 (7), 509–522.

48

Par t 2: A ssessment and Treatment

Gray, T. and Kiran, S. (submitted) A theoretical account of lexical-semantic naming deficits in bilingual aphasia. Manuscript submitted for publication. Green, D.W. (1998) Schemas, tags and inhibition. Bilingualism: Language and Cognition 1 (2), 100–104. Grosjean, F. (1998) Studying bilinguals: Methodological and conceptual issues. Bilingualism: Language and Cognition 1, 131–149. Gudayol-Ferre, E., Lara, J.P., Herrera-Guzman, I., Bohm, P., Rodes, E., Ansaldo, A.I. and Peña-Casanova, J. (2006) Semantic memory as assessed by the Pyramids and Palm Trees Test: The impact of sociodemographic factors in a Spanish-speaking population. Journal of the International Neuropsychological Society 14 (1), 148–151. Hakuta, K. (1986) Mirror of Language. New York: Basic Books. Hänninen, T., Hallikainen, M., Helkala, E.-L., Reinikainen, K.J., Soininen, H., Mykkänen, L., Laakso, M., Pyörälä, K. and Riekkinen, P.J. Sr. (1995) A follow-up study of age-associated memory impairment: Neuropsychological predictors of dementia. Journal of American Geriatrics Society 43 (9), 1007–1015. Heredia, R. and Altarriba, J. (eds) (2002) Bilingual Sentence Processing. Amsterdam: Elsevier. Ivanova, M.V. and Hallowell, B. (2009) Short form of the Bilingual Aphasia Test in Russian: Psychometric data of persons with aphasia. Aphasiology 23, 544–556. Kavé, G. (2006) The development of naming and word fluency: Evidence from Hebrew speaking children between ages 8 and 17. Developmental Neuropsychology 29 (3), 493–508. Kiran, S., Grasemann, U., Sandberg, C. and Miikkulianen, R. (submitted) Simulating bilingual aphasia: A novel computational model. Manuscript submitted for publication. Kiran, S., Pena, E., Bedore, L. and Sheng, L. (2010) Evaluating the relationship between category generation and language use and proficiency. Paper presented at the Donostia Workshop on Neurobilingualism, 30th September–2nd October, San Sebastian. Kiran, S. and Roberts, P.M. (2010) Semantic feature analysis treatment in SpanishEnglish and French-English bilingual aphasia. Aphasiology 24 (2), 231–261. Kohnert, K.J., Hernandez, A.E. and Bates, E. (1998) Bilingual performance on the Boston Naming Test: Preliminary norms in Spanish and English. Brain and Language 65, 422–440. Kroll, J.F. and Stewart, E. (1994) Category interference in translation and picture naming: Evidence for asymmetric connection between bilingual memory representations. Journal of Memory and Language 33 (2), 149–174. Kroll, J.F., Tokowicz, N. and de Groot, A.M.B. (2005) Models of bilingual representation and processing: Looking back and to the future. In Handbook of Bilingualism: Psycholinguistic Approaches (pp. 531–553). New York: Oxford University Press. Laganaro, M. and Overton Venet, M. (2001) Acquired alexia in multilingual aphasia and computer-assisted treatment in both languages: Issues of generalisation and transfer. Folia Phoniatrica and Logopedia 53 (3), 135–144. Lalor, E. and Kirsner, K. (2001) The role of cognates in bilingual aphasia: Implications for assessment and treatment. Aphasiology 15 (10), 1047–1056. Lezak, M. (1995) Neuropsychological Assessment (3rd edn). New York: Oxford University Press. Lofranco, L.A., Peña, E.D. and Bedore, L.M. (2006) English language narratives of Filipino children. Language, Speech, and Hearing Services in the Schools 37 (1), 28–38. Lorenzen, B. and Murray, L.L. (2008) Bilingual aphasia: A theoretical and clinical review. American Journal of Speech Language Pathology 17 (3), 299–317. Meinzer, M., Obleser, J., Flaisch, T., Eulitz, C. and Rockstroh, B. (2007) Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI. Neuropsychologia 45 (6), 1247–1256.

A ssessing L anguage Impair ment in Bilingual Aphasia

49

Moretti, R., Bava, A., Torre, P., Antonello, R.M., Zorzon, M., Zivadinov, R. and Cazzato, G. (2001) Bilingual aphasia and subcortical-cortical lesions. Perceptual and Motor Skills 92 (3, Pt 1), 803–814. Muñoz, M.L. and Marquardt, T.P. (2004) The influence of language context on lexical retrieval in the discourse of bilingual speakers with aphasia. Journal of Multilingual Communication Disorders 2 (1), 1–17. Muñoz, M.L. and Marquardt, T.P. (2008) The performance of neurologically normal bilingual speakers of Spanish and English on the short version of the Bilingual Aphasia Test. Aphasiology 22 (1), 3–19. Oberg, G. and Ramirez, M. (2006) Cross-linguistic meta-analysis of phonological fluency: Normal performance across cultures. International Journal of Psychology 41, 342–347. Paradis, M. (1989) The Bilingual Aphasia Test. Hillsdale, NJ: Lawrence Erlbaum Associates (out of print). Online at http://www.mcgill.ca/linguistics/research/bat/ Paradis, M. (1993) Evaluacion de la afasia en los bilingues. Barcelona: Masson. Paradis, M. (2000) Generalizable outcomes of bilingual aphasia research. Folia Phoniatrica and Logopedia 52 (1–3), 54–64. Paradis, M. and Libben, G. (1987) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum Associates. Pavlenko, A. (2008) Narrative analysis. In The Blackwell Guide to Research Methods in Bilingualism and Multilingualism (pp. 311–325). Malden: Blackwell Publishing. doi: 10.1002/9781444301120.ch18 Pekkala, S., Goral, M., Hyun, J., Obler, L.K., Erkinjuntti, T. and Albert, M.L. (2009) Semantic verbal fluency in two contrasting languages. Clinical Linguistics & Phonetics 23 (6), 431–445. Raoux, N., Le Goff, M., Auriacombe, S., Dartigues, J-F. and Amieve, H. (2010) Fluences verbales sémantiques et littérales: Normes en population générale chez des sujets âgés de 70 ans et plus issus de la cohorte PAQUID. Revue Neurologique 166, 594–605. Roberts, P.M. (1998) Bilingual aphasia: Some answers and more questions. Aphasiology 12 (2), 141–146. Roberts, P.M. (2001) Aphasia assessment and treatment in bilingual and multicultural populations. In R. Chapey (ed.) Language Intervention Strategies in Adult Aphasia (4th edn, pp. 208–232). Baltimore, MD: Lippincott, Williams & Wilkins. Roberts, P.M. (2008) Issues in assessment and treatment in bilingual and multicultural populations. In R. Chapey (ed.) Language Intervention Strategies in Adult Aphasia (5th edn, pp. 245–276). Baltimore, MD: Lippincott, Williams & Wilkins. Roberts, P.M. (2010) Three things you might not know about bilingual aphasia treatment. Keynote presentation at the International Aphasia Rehabilitation Conference, Montreal, June 27–29, 2010. Roberts, P.M. and Tougas, E. (2012) Evocation lexicale problèmes associés aux tâches et aux directives du Bilingual Asphasia Test (BAT) [Translate title: Verbal fluency problems with the tasks and instructions in the Bilingual Aphasia Test. Poster presented at the 80th annual conference of the Association Francophone pour le Savoir, Montreal. Roberts, P.M. and Deslauriers, L. (1999) Picture naming of cognate and non-cognate nouns in bilingual aphasia. Journal of Communication Disorders 32 (1), 1–23. Roberts, P.M. and Le Dorze, G. (1994) Semantic verbal fluency in aphasia: A qualitative and quantitative study in test-retest conditions. Aphasiology 8 (6), 569–582. Roberts, P.M. and Le Dorze, G. (1997) Semantic organization, strategy use, and productivity in bilingual semantic verbal fluency. Brain and Language 59 (3), 412–449. Roberts, P.M., MacKay, I.R.A. and Flege, J.E. (2002) Lexical and syntactic errors in Italian-English bilinguals. Brain and Cognition 48 (2–3), 513–516. Roberts, P.M., Umunosa, S. and Tougas, E. (submitted) Commentary on the short form of the Bilingual Aphasia Test. Manuscript submitted for publication.

50

Par t 2: A ssessment and Treatment

Rosselli, M., Ardila, A., Araujo, K., Weekes, V.A., Caracciolo, V., Padilla, M. and OstroskySolis, F. (2000) Verbal fluency and repetition skills in healthy older Spanish-English bilinguals. Applied Neuropsychology 7 (1), 17–24. Rosselli, M., Ardila, A., Salvatierra, J., Marquez, M., Matos, L. and Weekes, V.A. (2002) A cross-linguistic comparison of verbal fluency tests. International Journal of Neuroscience 112 (6), 759–776. Saffran, E.M., Berndt, R.S. and Schwartz, M.F. (1989) The quantitative analysis of agrammatic production: Procedure and data. Brain and Language 37 (3), 440–479. Sauzéon, H., Lestage, P., Raboutet, C., N’Kaoua, B. and Claverie, B. (2004) Verbal fluency output in children aged 7–16 as a function of the production criterion: Qualitative analysis of clustering, switching processes, and semantic network exploitation. Brain and Language 89, 192–202. Schmand, B., Groenink, S.C. and van den Dungen, M. (2008) Letter fluency: Psychometric properties and Dutch normative data [Dutch]. Letter fluency: Psychometrische eigenschappen en Nederlandse normen. Tijdschrift voor Gerontologie en Geriatrie 39 (2), 64–76. Schulpen, B.R., Dijkstra, T., Schriefers, H.J. and Hasper, M. (2003) Recognition of interlingual homophones in bilingual auditory word recognition. Journal of Experimental Psychology: Human Perception and Performance 29 (6), 1155–1178. Standish, T.I.M., Malloy, D.W., Cunje, A. and Lewis, D.L. (2007) Do the ABCS 135 short cognitive screen and its subtests discriminate between normal cognition, mild cognitive impairment and dementia. International Journal of Geriatric Psychiatry 22, 189–194. Van Der Elst, W., Van Boxtel, M.P.J., Van Breukelen, G.J.P. and Jolles, J. (2006) Normative data for the animal, profession and letter M naming verbal fluency tests for Dutch speaking participants and the effects of age, education and sex. Journal of the International Neuropsychological Society 12, 80–89. Videsott, G, Herrnberger, B., Hoenig, K., Schilly, E., Grothe, J., Wiater, W., Spitzer, M. and Markus, K. (2010) Speaking, in multiple languages: Neural correlates of language proficiency in multilingual word production. Brain and Language 113, 103–112.

4

Morphological Assessment in Bilingual Aphasia: Compounding and the Language Nexus Gary Libben

There is a rare opportunity provided by considering morphology, the bilingual lexicon and aphasia together. It is an opportunity to gain true insight into what I will call the language nexus, that which binds seemingly disparate language functions and representations into an integrated whole. Multimorphemic words constitute the domain in which we can best understand the language nexus. Multimorphemic words have a dual nature: on the one hand, they function as semantically encapsulated structures with unique meanings and unique ways of patterning with other words to form sentences, texts and conversations. On the other hand, these words have substructures which themselves are meaningful and thus allow words to be linked in our minds and in our behaviors as sets of multidimensional properties. The language nexus brings together all of these. This chapter is focused on words with such substructure – words composed of more than one meaningful unit. This would include prefixed words such as ‘repave’, suffixed words such as ‘pavement’ and compound words such as ‘pavestone’. Words such as these are at the center of the language nexus because they are the results of creativity over generations of language use which builds the pattern-rich mental lexicon whose regularities, both in form and meaning, allow language users to create yet more structures and to benefit from more connections. Morphological processing thus involves storage, organization, access and online computation. Moreover, every act of morphological processing seems to encompass both the receptive and productive dimensions of language. I discuss the key features of morphological processing and the principles that allow morphological processing ability to be assessed in a clinical setting with bilingual participants. I claim that key principles of morphology can be 51

52

Par t 2: A ssessment and Treatment

targeted in bilingualism by concentrating on compound words. These are the most basic of morphological structures, they allow the greatest range of morphological creativity and they are the most comparable across the world’s languages. Thus, they are the most easily and profitably applicable to the understanding of key morphological deficits in bilingual aphasia. In the final part of this chapter, I discuss the development of a test instrument, the Morphological Assessment Project (MAP) that is designed to test for core aspects of morphological ability. This approach to morphological assessment can use the cross-linguistic comparability of compound processing to assess key aspects of morphological ability among bilinguals and advance our understanding of the manner in which morphological processing is affected by bilingualism.

Three Kinds of Morphology Morphology is both the branch of linguistics that examines the patterns by which words are formed from other words and those patterns themselves. Morphological patterns are typically divided into three kinds: inflectional morphology, derivational morphology and compounding. Inflectional morphology concerns primarily grammatical features that form parts of words. In English, verbs can be marked through inflectional morphology for person, number and tense, so that the sentence ‘He listens’ is the grammatically and morphologically native-like way of indicating that the subject (‘he’) is third person singular and that the action denoted by the verb takes place in the present. In the English sentence ‘He listened’, the form of the verb indicates a past tense event, but there is no marking on the verb for either person or number. In this way, the English system of inflectional morphology is said to be impoverished. There may be a fundamental difference between inflectional morphology and other forms of morphology. The reason for this is that inflectional morphology does not really create new words. Rather, it creates particular versions of words that can be used to meet particular grammatical, discourse or semantic needs. Derivational morphology, in contrast, does create new words. But it does so in a manner that allows new words to be linked to parent words, sibling words and, eventually, lexical offspring. It is this that gives it its dual nature. On the one hand, derived words can be seen as variants of their parent and sibling words. For example, for a speaker of English to function effectively as a competent user of morphology, an adjective such as ‘wide’ must be easily linkable to its related verb form ‘widen’, its adverb form ‘widely’ and its noun form ‘width’. Derivational forms may also be linked to one another as parts of compounds. So, for example, ‘bandwidth’ is a noun, in which ‘-width’ is the morphological head. Because the morphological element ‘-width’ functions as a distinct morphological entity as the head of that compound

Mor phological A ssessment in Bilingual Aphasia

53

(Libben, 2010), it becomes still further removed from its derivational relatives. This is because it is tied to the internal morphological and thematic structure of the compound. For this reason of thematic structure, the compound ‘bandwidth’ cannot be changed through derivation to ‘*bandwiden’ (as in ‘*He tried to increase the bandwidth by bandwidening it’). This brings us to compounding as the third type of morphological process and the one which will serve as the focus of this discussion. As noted by Jackendoff (2002, 2009), compounds may be linguistic fossils and as such the most fundamental of morphological processes (Libben, 2006). There is a way in which compounding is the least restricted of all word formation processes. In English, for example, virtually any noun may combine with any other noun to form a new compound word. Indeed, most compounds are formed of noun-noun combinations, but new compound words can also be created by combining adjectives and nouns, verbs and nouns, verbs and adjectives, prepositions and verbs, etc. Thus, a language that contained just 100 lexical items could, in principle, create from them 9900 new compounds. This, of course, is not what happens in practice. Most meanings created by such exhaustive combinations would not be needed for communicative purposes in a language. Perhaps, more importantly, there is the observation that compound words are typically created by exchanging one element of the compound based on a well established morphological family (Singh, 2006; Singh & Dasgupta, 1999). In this way ‘flashlight’ is potentially a member of two families – one based on ‘-light’ that includes ‘searchlight’ and ‘penlight’ and the other based on ‘flash-’ that includes the members ‘flashbulb’ and ‘flashdrive’.

Compound Words and the Nature of Morphological Processing As a means of targeting the essence of morphological ability and as a means to do so in a bilingual context, compound words offer the key advantage of having a high degree of comparability across languages, but with interesting points of variability. Even the basic type of compound, that which is composed of just two major constituents, shows substantial variability across languages. Compounds may be composed of two elements which themselves are free-standing words, as in the English word ‘scrapbook’. This is the dominant pattern in English, because monomorphemic roots can be used as freestanding words in sentences quite easily. Many other languages (e.g. Italian), however, require roots to have inflectional affixes. In such languages, compounds are more accurately seen as composed of roots, rather than words. In some languages (e.g. Polish), compound words have an obligatory linking element between the two constituents. In languages such as German, a number of different linking elements may be used, and part of native speaker

54

Par t 2: A ssessment and Treatment

language competence is the knowledge of which linking elements must be used with which compound constituents. It is important to note that linking elements in languages such as German and Dutch are related historically to affix forms that can attach to the first constituent (indicating, e.g. plurality or possession). Synchronically, however, these linking elements have very limited semantic function. In contrast, the prepositional elements that are often found within compounds in languages such as French (e.g. ‘pomme de terre’; literally ‘apple of ground’) are often perceived by native speakers to have real semantic content. Another key way in which languages differ in their compounding is in terms of the ordering of constituents. In Hebrew, for example, the word for potato is also composed of the morphemes ‘earth’ and ‘apple’. This is also the case in Dutch and some dialects of German. The difference, however, is that in the Germanic languages, the ordering of constituents is earth+apple. In Hebrew, it is apple+earth. The compound words have exactly the same meanings. A good deal of evidence regarding the ways in which compound processing is affected in aphasia comes from work in Italian. Delazer and Semenza (1998) showed that compounds can be selectively impaired in aphasia and errors with compound words very often involve the creation of neologisms which themselves are compounds. Mondini et al. (2004) investigated verbnoun compounds, finding a particular sensitivity to the verb element. This is in accordance with the general vulnerability of verbs in aphasia and thus evidence that compound words retain the grammatical features of their constituent elements in the mind. While patterns of compound processing impairment in aphasia show considerable variability, they present a strong case for the internal morphological structure of compounds playing a role in online processing. In addition to the Italian findings noted above, such effects have been reported for German (Blanken, 2000), English (Badecker, 2001) and Finnish (Mäkisalo et al., 1999). The key evidence for morphological processing is the tendency of persons with aphasia to omit and/or substitute individual morphological elements of compounds. This constitutes evidence that compound processing in aphasia makes reference to the morphological structure of the compound word. Effects of morphological structure in online compound processing have also been documented across a number of experimental paradigms (Libben, 2010; Marelli et al., 2009), including in magnetoencephalography (MEG) studies of unimpaired populations (Fiorentino & Poeppel, 2007). It is important to note, however, that the presence of morphological effects does not constitute evidence against whole-word processing as well. Indeed, as is discussed below, Libben et al. (2004) argue for a complex interplay in aphasia between the mental representations of compound words as wholeword units and their mental representations as morphologically structured organizations of constituents.

Mor phological A ssessment in Bilingual Aphasia

55

Four Morphological Processing Abilities In this section I describe the set of core abilities that might provide the link between constructs such as morphological structured representations and assessable behaviors. In this context, it is important to highlight the fact that morphological ability spans production and comprehension and extends to the processing of both existing and novel words. An additional important consideration, in my view, is that morphology not only allows for the creation of new words but also for the means by which they can be organized in the mind, as a result of use, along multiple dimensions including meaning and grammatical function.

The ability to repeat, comprehend and produce multimorphemic words with appropriate semantics and syntactic properties We might begin with the apparently uncontroversial claim that morphological processing must involve the ability to comprehend and produce multimorphemic words. This is indeed the core ability. It is important to point out, however, as a matter of orientation, that whether or not a word is multimorphemic is a psychological matter, not necessarily simply a linguistic one. Some examples may help to illustrate this point. Most Americans will know that Philadelphia is a city of brotherly love. But certainly not all will have access to the constituent morphemes of the name. Is it then perhaps monomorphemic for them? Consider another example: some Americans with an interest in etymology will perhaps know that Cincinnati derives its name from that of a Roman general, named for his ‘curly hair’. Some might also know that Ohio derives its name from the Iroquoian words for ‘good river’. But it is not at all obvious that any of this knowledge will be of much use in the online lexical processing of English. Breaking down the word Ohio into ‘ohi’ and ‘yo’ is only of value if the language user knows and can use some other words which have those morphological elements. Additionally, this morphological knowledge will only be of real use if there is a sufficient number of language users within a linguistic community so that the morphological elements can be used productively. It seems clear that these rather complex matters affect how we can characterize even apparently simple acts of lexical repetition. Indeed, we must also note that difficulties with multimorphemic lexical processing, even in the domain of repetition, lead us toward the conclusion that the language user is engaging in morphological processing. If a language user is treating multimorphemic words as though they are monomorphemic, then they should only be difficult to process by virtue of properties such as abstractness, length and frequency that apply to both monomorphemic and multimorphemic words. If, however, performance effects that are linkable to morphology emerge above and beyond such monomorphemic effects, then

56

Par t 2: A ssessment and Treatment

this would be evidence of morphological processing. As we will see below, this expands into a mosaic of balancing components. If a language user is engaged in automatic and obligatory access to the morphological components of words, then there is the risk that this access may overpower whole-word comprehension. Because of semantic drift and because of the semantic under specification of morphological combinations, it is very difficult to predict whole-word meanings from the meanings of morphological constituents. Once both the whole-word meaning and the constituent meanings are known, however, this task becomes considerably easier. Take the word ‘foot’ as an example: The role that it plays in the words ‘footbridge’, ‘football’, ‘clubfoot’ and ‘crowfoot’ seems quite straightforward once you know what those multimorphemic words mean. If, however, you do not, it is quite a challenge to predict their meanings. What follows from this is that morphological processing and wholeword processing are activities that could, in principle, be in competition. That competition might require that the results of morphological processing be deactivated for effective multimorphemic word comprehension and production to occur. Failure to do so would impede whole-word comprehension. This failure to deactivate was at the core of the morphological processing deficit exhibited by RS (Libben, 1993, 1998), a right-handed female who at the age of 22 suffered a brain hemorrhage in the temperoparietal area of the left hemisphere. RS typically blended whole word and constituent meanings for compounds so that her paraphrase of a compound such as ‘butterfly’ as a ‘pretty yellow fly’ indicated a blending of the properties of the whole word (the prettiness of butterflies) and the meaning of constituents (the yellowness of the constituent ‘butter’ in ‘butterfly’). This pattern was seen across a range of semantically opaque compounds. In related studies Buchanan et al. (2003) and Libben et al. (2004) discuss two deep dyslexic patients who provide additional evidence of the role that morphological deactivation may play in the normal processing of multimorphemic words. Libben et al. (2004) conclude that the normal processing of multimorphemic words may be characterized as having two major components: The first activates as much as possible rather indiscriminately, including the representations and meaning of all morphological elements and substrings as well as the representations and meanings of whole words. This ensures that the comprehension system will be as flexible as possible and maximally able to accommodate the presence of novel multimorphemic strings in the visual or auditory input. I have referred to this as the Maximization of Opportunity Principle (Libben, 2006). Maximization of Opportunity leads, however, to the activation of a superset of representations. This generates the need for a second processing component. This component provides a post-recognition cleanup process that is achieved through the deactivation of morphological components. In normal language processing, this deactivation can lead to elevated response times for

Mor phological A ssessment in Bilingual Aphasia

57

semantically opaque stimuli. For impaired language users, the failure of deactivation results in miscomprehension and misproduction.

The ability to understand novel multimorphemic constructions The ability to understand novel multimorphemic structures seems to trade on two abilities. The first appears to be a process of morphological decomposition which is automatic and obligatory and serves to isolate morphological units and morphological substrings. The second ability involves the determination of how both substrings may relate to one another so that an interpretation of a novel string can be determined. Although it is conceivable that this knowledge could be related to a form of grammatical declarative knowledge, recent exemplar-based approaches suggest that generalizations concerning the properties of a particular morphological pattern derive from experience with sets of words already known. What remains unclear is whether, in online processing, the actual families (e.g. all compound words that are headed by ‘berry’) are activated in the processing of a single member of the morphological family (e.g. ‘blueberry’).

The ability to produce novel multimorphemic constructions The ability to produce novel multimorphemic constructions in a manner that is consistent with the morphological patterns of a language, so that the probable interpretation accords with the interpretation of the language producer may very well be the highest level of morphological competence. In order to do this, the language user must draw upon the knowledge of morphological substrings and their relations to other words, word meanings and patterns of lexical usage in the language. Although this seems as though it might be a rather exotic ability, it is, in fact, the engine of lexical innovation that results in language change in the morphological domain. In languages in which compounding is particularly productive, language users routinely create new compound forms to meet this or that discourse, legal, or (quite often) bureaucratic need. It is for this reason that measures of morphological productivity (e.g. Baayen, 1992) can use the number of single occurrences of words in a corpus employing a particular morpheme in a multimorphemic word as a measure of that morpheme’s productivity. Measures of the morphological productivity of individual speakers are less well established in the literature. Second language speakers often show a considerable amount of morphological productivity, but this is often as a coping mechanism for the existence of lexical gaps in their own vocabularies and not necessarily lexical gaps in the language. Moreover, it is not always the case that their morphological productivity is based on the knowledge of morphological substrings and their relations to other words and word meanings in a manner that is consistent with the morphological patterns of the

58

Par t 2: A ssessment and Treatment

language, so that the probable interpretation accords with the interpretation of the language producer. Finally, there is the matter of the extent to which this ability is variable within the population. It seems that the loss of morphological productivity is associated with native speaker productivity decline in endangered languages (perhaps because of the need, alluded to above, for a community of speakers with well developed morphological systems in the particular language). But even within language communities that are characterized by high levels of education, literacy and lexical innovation, we expect that some speakers produce novel structures rather infrequently, and others could be described as ‘Morphological Poets’. It follows from this that, in the assessment of morphological ability in aphasia, special attention must be taken in making assumptions concerning the pre-aphasic levels of individuals’ morphological productivity.

The ability to employ morphological patterns within the language in the organization of vocabulary knowledge This is likely the most difficult aspect of morphological ability to test. It is probably, however, the most important, because it gets at the core advantage of morphological ability in human language. In principle, languages could do fine having absolutely no morphological word formation processes. Every new word would have a phonetic form, understood to be an arbitrary pairing of form and meaning. In any case the lexicon would indeed, in the sense of Bloomfield (1933), be a repository for all the exceptions in a language. Word formation morphology brings systematicity and cross-referencing ability to that potential list of exceptions. Compounds sharing a head noun can be expected to have something in common, so that if one already knows words such as ‘flashlight’, ‘penlight’, ‘searchlight’, ‘spotlight’ and ‘porchlight’, then a novel form such as ‘doorlight’ should be somewhat easier to interpret – certainly much easier than it would be if it were attached to an entirely novel string such as ‘volmor’. Clearly morphological patterns can be of great benefit in the organization and retrieval of words in the mind. The ability to make use of this organization in online processing is a core component of morphological processing ability. Returning to the concept of a language nexus, it is important also to note that although we have discussed the features of morphological processing as involving either comprehension or production, in fact both are involved in all morphological acts. It is because of this that we see morphology as being particularly revealing of the language nexus. It is a creative activity that at once draws on stored representations, but also contributes to the enrichment of the mental lexicon, both in terms of volume (i.e. the number of words known) and also in terms of the richness of connectivity. Morphological structures serve the needs of multiple indexing in dimensions for which linguists and psycholinguists have not yet found names.

Mor phological A ssessment in Bilingual Aphasia

59

Morphology, Homogeneity and the Bilingual Lexicon In the section above I have outlined what I consider to be the four core abilities associated with morphological processing. These are summarized in Table 4.1. As can be seen in the table, the four abilities refer to overarching cognitive activities rather than to specific representations for specific languages. If morphological abilities bring together language representation and processing in the language nexus there are substantial consequences for our conceptualizations of morphological processing among bilinguals and the impairment of this ability in bilinguals with aphasia. Specifically, it is important to address whether an English-French bilingual, for example, should be seen as possessing two morphological systems and abilities related to them, or a single, more general, morphological ability. Even at the outset, it seems clear that the answer must be both. Compound words can be very helpful to the understanding and assessment of the interaction of such general and language-specific abilities. In contrast to compounding, inflectional morphology, as we have discussed above, is intertwined with the syntax of a language. Moreover, the complexity and nature of inflectional morphology varies greatly across languages. Consider, for example, the differences between Finnish and English. Finnish is a case-rich language, whereas, in English, case is only marked on a few pronominal forms. Thus, it is not at all clear that when considering inflectional deficits or processing in Finnish and English the Table 4.1 Core characteristics of morphological processing ability Ability

Description

(a) Processing existing words The ability to repeat, comprehend and produce multimorphemic words with appropriate semantics and syntactic properties. (b) Understanding new words The ability to access familiar substrings from within an unfamiliar multimorphemic word and use their position and meaning patterns to develop appropriate novel word interpretations. (c) Producing new words

The ability to produce novel multimorphemic constructions that are understood in the manner intended and are well formed with respect to the morphological patterns of the language.

(d) Lexical organization

The ability to employ morphological patterns to build lexical associations and lexical families.

60

Par t 2: A ssessment and Treatment

same types of things are really being considered. In terms of compounding, however, Finnish and English are much more comparable (Bertram & Hyönä, 2003). Languages differ also quite considerably in the complexity of their derivational systems. However, the hierarchy articulated by Dressler (2006), namely that if a language has derivation it will have compounding, but not vice versa, reinforces the advantages of using compounding to understand the fundamentals of morphology in the bilingual domain. The emphasis in Table 4.1 on overarching abilities, and the selection of compounding as the domain in which those abilities will be examined, might lead one to wonder whether any language-specific features associated with morphological ability are likely to surface. One language-specific aspect of morphological ability concerns those components of the abilities that are directly related to vocabulary size. Indeed, all abilities are in some way dependent on vocabulary size. This is, of course, true for both monolinguals and bilinguals (a term used here for convenience to refer to persons with stable functional ability in more than one language). However, we must also expect that a bilingual might experience vocabulary size challenges in a nondominant language that a native speaker might not. An additional language-specific factor that may result from skill inequalities across a bilingual’s languages concerns the morphological knowledge alluded to in all of the skills summarized in Table 4.1. This inequality may favor either a dominant or non-dominant language. Non-native speakers of a language may have very well developed morphological knowledge, perhaps as a result of formal metalinguistic training in a non-dominant language, or perhaps as the result of the need to engage in online morphological analysis of whole-word forms that are not part of the bilingual’s recognition or production vocabulary. On the other hand, it is also very likely, even for the same bilinguals, that the richness of morphological associations within the non-dominant language is less well developed. This brings us to the question of the core of morphological ability. Although a bilingual will show language-specific vocabulary size differences, lexical richness differences, lexical association differences and morphological knowledge differences across languages, it seems to me that there are very good reasons to expect that the bilingual is in possession of a single morphological processing system. The first reason comes from the mounting evidence of non-selectivity in lexical functioning in language. A good deal of evidence has come to us from research on interlingual homographs. These are words such as ‘pain’, which is a word in both English and in French. Its meaning in French is ‘bread’. Similarly, the string ‘glad’ is a word in both English and Dutch. Its meaning in Dutch is ‘slippery’. Current findings suggest that both meanings of interlingual homographs are routinely activated in word recognition (see Obler & Goral, 2007, for a review of this literature and Libben

Mor phological A ssessment in Bilingual Aphasia

61

& Titone, 2009, for evidence of non-selectivity attenuation in sentence processing). Studies by Dijkstra and colleagues (e.g. Dijkstra et al., 2000; Dijkstra et al., 1999; Dijkstra et al., 1998) have documented evidence for the conceptualization of the bilingual lexicon as a single store with language specific tags as represented in the BIA + model (Dijkstra & Van Heuven, 2002). The notion of an integrated lexical store for the languages spoken by a bilingual will, as Obler and Goral (2007) note, place differential pressure on the system, depending on a number of factors, including the distance between languages along dimensions such as orthography and morphology. This is a matter taken up by Libben (2000) in the formulation of the Homogeneity Hypothesis. This proposal, as it relates to morphology in bilingualism, is that there will be pressure on the morphological system to resolve into a single homogenous system. This will mean, inevitably, that the bilingual morphological processing system will evolve to meet all of the needs of each language, and so will be more complex than that of a monolingual. To take a non-linguistic example, consider the fact that some North American cities have one-way streets. Others do not. It would be expected that ‘fluent’ drivers in cities having one-way streets would develop memory representations that would include both location and direction for each street. Drivers in other cities would need only to encode street location (because all streets are two-way streets). And, now, what about drivers that have experience in both kinds of cities? The homogeneity perspective claims that they would encode all streets in the same manner so that a single system could be maintained that would be adequate to the more complex system. Thus, all streets in the city with no one-way streets would be encoded as ‘goes in both directions’. Consider now, as a parallel case, compounding in English and Italian. All English compounds are head-final, so that a ‘tablecloth’ must be a type of ‘cloth’, not a type of ‘table’. Italian, however, contains both head-initial compounds and head-final compounds. Morphological ability in Italian, then, must include processing of headedness direction. In English this is not required. We expect that the homogenous morphological system for the Italian-English bilingual will involve a superset system in which headedness direction is represented (redundantly) for English compound words. One would expect that the Italian-English bilingual would redundantly assign ‘non-realized grammatical gender’ to English nouns in the same way and for the same reason. If the homogeneity hypothesis is indeed correct, then there are substantial consequences for the assessment of morphological ability across populations, including bilingual persons with aphasia. The MAP is designed to provide a framework for such an assessment. In the section below its applicability to the testing of bilinguals with aphasia is discussed with special reference to how compound words may provide the domain of comparability in which participants and morphemes may act as their own controls.

62

Par t 2: A ssessment and Treatment

The Morphological Assessment Project (MAP) The MAP is designed to investigate core aspects of morphological processing by targeting a small number of tasks and a very constrained set of stimuli. The MAP uses compound words as the testing core. The English version of the MAP has been normed with speakers of Canadian English and has been employed as a subtest of the Alberta Functional Assessment Battery (ALFAB) (Westbury, 2007). The core MAP tasks are listed in Table 4.2. A key feature of the approach to morphological assessment in the MAP is that the test instrument is not designed for language comprehensiveness. Rather, it is designed to target the processing of multimorphemic words (compounds in particular) in a manner that allows the individual test taker to function as his or her own control. It also allows the components of multimorphemic words to act as their own controls. The key to this control is a stimulus set in which individual morphological constituents serve as their own controls in a monomorphemic baseline condition and across categories of multimorphemic stimuli. In the English version of the MAP this is done by building the core stimulus set around triplet sets of monomorphemic words to create a core stimulus list of 230 items. The principles behind the stimulus list construction can be seen by examining Table 4.3, which lists five of the English MAP triplets. As can be seen in Table 4.3, the core stimuli are created so that each of the monomorphemic words can be tested on their own and also tested within compounds. The stimuli have been chosen so that each will occur in both initial and final position of a bi-constituent compound. This creates the three sets of compounds created by the AB, BC and CA orderings of monomorphemic stimuli. The triplet constitutes the smallest configuration set size that allows constituents to be in each position without having any core compound stimuli be reversed. As the arrangement of stimuli in Table 4.3 indicates, the test design principles allow monomorphemic core stimuli to act as their own controls. In

Table 4.2 Core tasks of the Morphological Assessment Project Task

Morphological ability addressed

(1) Repetition

Processing existing words

(2) Hyphenation decision

Lexical organization

(3) Definition judgment

Processing existing words Understanding new words

(4) Morphological fluency

Processing existing words Lexical organization Producing new words

Mor phological A ssessment in Bilingual Aphasia

63

Table 4.3 Core MAP monomorphemic English stimuli and their use in three compound orderings Mono A Mono B

Mono C

Compound AB

Compound BC

Compound CA

board

room

key

board-room

room-key

key-board

Back

side

kick

back-side

side-kick

kick-back

Case

work

book

case-work

work-book

book-case

Fire

water

gun

fire-water

water-gun

gun-fire

Foot

bridge

club

foot-bridge

bridge-club

club-foot

this way we can assess whether, for example, if an individual person with aphasia has no difficulty repeating the word ‘key’, they can also repeat its morphological variants ‘key-’ and ‘-key’ as the first and second elements of the compounds ‘key-board’ and ‘room-key’ respectively. This contrasting of free-standing words and embedded morphemes brings a number of questions to the foreground. The first is the question of how well a particular test participant might be able to repeat the string if it were embedded in another type of morphological structure. To address this question as part of the test design structure, all core MAP stimuli are also tested in two inflected forms. Suffixed by -ed and suffixed by -s. The ability to do this (for English) trades on the possibility in English for noun-verb conversion (or zero-affixation, as the phenomenon is sometimes characterized). Thus, a participant is also tested with the stimuli ‘keys’ and ‘keyed’. Each core stimulus is tested additionally in two derivational instantiations. For English this involves both prefixed and suffixed forms. To illustrate this structure, Table 4.4 shows the instantiation of each of the core stimuli in their prefixed and derivationally suffixed forms. As is evident in the discussion above, the ability to create tightly controlled multimorphemic stimulus sets that can act as their own controls in such situations presents a stimulus design challenge when we are also looking to achieve comparability across languages for bilingual testing. However, Table 4.4 Each of the A, B and C monomorphemic stimuli from Table 4.3 in their prefixed and derivationally suffixed test forms Prefixed A

Der-suf A

Prefixed B

Der-suf B

Prefixed C

Der-suf C

aboard

boarder

anteroom

Roomy

re-key

keyless

aback

backward

aside

Siding

re-kick

kicker

encase

casement

re-work

Worker

pre-book

bookish

misfire

fireproof

re-water

Watery

un-gun

gunner

forefoot

footage

re-bridge

Bridger

re-club

clubber

64

Par t 2: A ssessment and Treatment

if compounds are used as the core stimuli the magnitude of the problem is diminished considerably. Currently the MAP is being developed in both German and French, languages that have rather different morphological properties. A Hebrew version is also planned. Thus, although it would be more challenging to meaningfully compare English, French, Hebrew and German inflectional and derivational processing directly, their compounding can be compared directly. Bilingual modules for language pairs are under development. These modules will track and exploit cognate structure across languages and the translation equivalence of compound constituents. Bilingual administration of the MAP also allows for combined statistical analysis across languages. In each case the MAP allows for derivational and inflectional forms to provide satellite tests within each language. This configuration is represented in Figure 4.1. The compound stimuli that form the core of the MAP serve as the basis for comparison to other compound types within the core tasks listed in Table 4.2. The key characteristics of these tasks are described briefly below. The repetition task addresses the most basic morphological ability – the ability to repeat multimorphemic words. As has been discussed above, simple word repetition can be revealing of underlying morphological dynamics. The repetition task may be subdivided, depending on the profile of the test taker, into modality-specific repetition tasks, including copying and cross-modal repetition. In the hyphenation decision task participants are presented with multimorphemic stimuli that are broken through hyphenation at appropriate morpheme boundaries and at points within morphemes. Performance on the task provides a baseline of morphological recognition within the language.

L A N G 1

L A N G 2

Roots A

Compounds AB

Inflected A

Derived A

Roots B

Compounds BC

Inflected B

Derived B

Roots C

Compounds CA

Inflected C

Derived C

Roots A

Compounds AB

Inflected A

Derived A

Roots B

Compounds BC

Inflected B

Derived B

Roots C

Compounds CA

Inflected C

Derived C

Figure 4.1 The bilingual MAP configuration. Primary interlingual comparisons are made through compounds

Mor phological A ssessment in Bilingual Aphasia

65

Table 4.5 An example stimulus from the novel word definition subtest. Items of this subtest are built from the core compound list Category

Stimulus

Definition

Plausible

Linestream

A stream that flows in a straight line.

Category change

Linestream

To flow in a straight line.

Real word reversal

Linestream

To make something more efficient.

Plausible

Siderooom

A room at the side of a house or a building.

Category change

Siderooom

To live in one side of a house.

Novel word reversal

Siderooom

The side of a room.

The definition judgment task addresses the ability of participants to assess definitions of both existing and novel stimuli. Each novel compound stimulus is presented with three kinds of definitions: a plausible definition, one that requires a result from a lexical category change and one that corresponds to a real word in the language (or a word of another language for bilingual participants). Note in the examples provided in Table 4.5 that the novel compounds are formed from the core monomorphemic stimuli in a manner such that half are reversals of real stimuli compounds and half are not. The final core MAP task, morphological fluency, is a deceptively simple technique of assessing the ability of participants to access and produce multimorphemic words that contain particular morphemes. In testing with non-aphasic participants this task shows substantial performance variation among participants. The basic instructions for the task are ‘say (or write) as many words as you know that start (or end) with the following’ (e.g. ‘key’?). The novel word production component involves the instructional variant ‘Now, how many new words can you make up that start/end with ?’ Examples are given for each variant of the task. The morphological fluency task allows compound constituent productivity and availability to be contrasted for each morphological position. Prefixes and suffixes serve as comparisons, so that the instruction ‘Name as many words as you can that end with “-ing”’ generates a dataset that provides a measure of verb retrieval ability in English. A five minute time limit per stimulus has shown itself sufficient to exhaust lexical selection. For existing compound forms lexical retrieval is typically exhausted within 60 seconds with native speaker control participants.

Summary and Next Steps In this chapter I have laid out a view of morphological processing that places it at the core of language ability. I have used the term language nexus to

66

Par t 2: A ssessment and Treatment

describe this core. It is distinguished by the fact that it involves the creative manipulation of units of meaning in a language in order to communicate new concepts that can be understood in real time and in a manner that feeds back to enrich the lexical memory stores of the language participants. If the new morphological creation gains a foothold in the language it will then spread through the lexical acquisition process to become part of the lexical memory stores of speakers far removed in time and space from the context in which the original creative morphological act took place. Moreover, the new morphological creation will add to the system of morphological patterns within the language that will serve as templates for new acts of morphological creativity. Morphological ability is the ability possessed by an individual speaker of a language to make use of the meaningful substrings within a word in a manner that enables the comprehension of new forms while maintaining properties of established words. It involves the ability to produce novel forms and, perhaps most importantly, it involves the ability to exploit morphological patterns within a language. I have argued that it is very likely that the key characteristics of bilingual morphological systems are integrated into a homogenous system that is more intricate and extensive than that of a monolingual speaker. For bilingual persons with aphasia this suggests that morphological therapy with one language can have significant improvement effects on the other. Nault (2010) has presented evidence in English for how targeted morphological therapy that is focused on compound words has general ameliorative effects on both trained and untrained multimorphemic words. Extending the logic of this approach suggests that comparable beneficial effects could be obtained by working across the languages of a bilingual person with aphasia. However, to create the platform for the measurement of such effects in the bilingual morphological domain, an assessment instrument is required that will target core morphological abilities. The MAP is designed to create such a platform. It uses compound words as the morphological core of assessment and relatively simple tasks that probe the processing of existing multimorphemic words, the creation of and comprehension of new words and the ability to extract morphological substrings from words and employ morphological patterns within the language. It is anticipated that the morphological assessment enabled by this instrument can serve to deepen our understanding of bilingualism, language impairment in aphasia and the manner in which representations and processes interact in the language nexus.

References Baayen, R.H. (1992) Quantitative aspects of morphological productivity. Yearbook of Morphology 1991, 109–149. Badecker, W. (2001) Lexical composition and production of compounds: Evidence from errors in naming. Language and Cognitive Processes 16, 337–366.

Mor phological A ssessment in Bilingual Aphasia

67

Bertram, R. and Hyönä, J. (2003) The length of a complex word modifies the role of morphological structure: Evidence from eye movements when reading short and long Finnish compounds. Journal of Memory and Language 48, 615–634. Blanken, G. (2000) The production of nominal compounds in aphasia. Brain and Language 74, 84–102. Bloomfield, L. (1933) Language. New York: Henry Holt. Buchanan, L., McEwen, S., Westbury, C. and Libben, G. (2003) Semantic and semantic error: Implicit access to semantic information from words and nonwords in deep dyslexia. Brain & Language 84, 65–83. Delazer, M. and Semenza, C. (1998) The processing of compound words: A study in aphasia. Brain and Language 61, 54–62. Dijkstra, T., Van Jaarsveld, H. and Ten Brinke, S. (1998) Interlingual homograph recognition: Effects of task demands and language intermixing. Bilingualism: Language and Cognition 1, 51–66. Dijkstra, T., Grainger, J. and Van Heuven, W.J.B. (1999) Recognition of cognates and interlingual homographs: The neglected role of phonology. Journal of Memory and Language 41, 496–518. Dijkstra, T., De Bruijn, E., Schriefers, H. and Brinke, S.T. (2000) More on interlingual homograph recognition: Language intermixing versus explicitness of instruction. Bilingualism 3 (1), 69–78. Dijkstra, T. and Van Heuven, W. (2002) The architecture of the bilingual word recognition system: From identification to decision. Bilingualism: Language and Cognition 5, 175–197. Dressler, U.W. (2006) Compound types. In G. Libben and G. Jarema (eds) The Representation and Processing of Compounds Words (pp. 23–44). Oxford: Oxford University Press. Fiorentino, R. and Poeppel, D. (2007) Compound words and structure in the lexicon. Language and Cognitive Processes 22 (7), 953–1000. Jackendoff, R. (2002) Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford: Oxford University Press. Jackendoff, R. (2009) Compounding in the parallel architecture and conceptual semantics. Handbook of compounding. In R. Lieber and P. Stekauer (eds) The Oxford Handbook of Compounding (pp. 105–128). Oxford: Oxford University Press. Libben, G. (1993) A case of obligatory access to morphological constituents. Nordic Journal of Linguistics 16, 111–121. Libben, G. (1998) Semantic transparency in the processing of compounds: Consequences for representation, processing, and impairment. Brain and Language 61 (1), 30–44. Libben, G. (2000) Representation and processing in the second language lexicon: The homogeneity hypothesis. In J.A. Archibald (ed.) Second Language Grammars (pp. 228–248). Boston, MA: Blackwell Press. Libben, G. (2006) Why study compounds? An overview of the issues. In G. Libben and G. Jarema (eds) The Representation and Processing of Compound Words (pp. 1–21). Oxford: Oxford University Press. Libben, G. (2010) Compounds, semantic transparency, and morphological transcendence. Linguistische Berichte Sonderheft 17, 212–232. Libben, M. and Titone, D.A. (2009) Bilingual lexical access in context: Evidence from eye movement recordings during L1 reading. Journal of Experimental Psychology: Language, Memory, and Cognition 35 (2), 381–390. Libben, G., Buchanan, L. and Colangelo, A. (2004) Morphology, semantics, and the mental lexicon: The failure of deactivation hypothesis. Logos and Language 4 (1), 45–53. Mäkisalo, J., Niemi, J. and Laine, M. (1999) Finnish compound structure: Experiments with a morphologically impaired patient. Brain and Language 68, 249–53.

68

Par t 2: A ssessment and Treatment

Marelli, M., Crepaldi, D. and Luzzatti, C. (2009) Head position and the mental representation of nominal compounds. The Mental Lexicon 4 (3), 430–454. Mondini, S., Luzzatti, C., Zonca, G., Pistarini, C. and Semenza, C. (2004) The mental representation of verb-noun compounds in Italian: Evidence from a multiple singlecase study in aphasia. Brain and Language 90, 470–477. Nault, K. (2010) Morphological therapy protocol. Unpublished PhD Dissertation. University of Alberta, Edmonton, Alberta. Obler, L.K. and Goral, M. (2007) Bilingual lexica. In G. Jarema and G. Libben (eds) Core Perspectives on the Mental Lexicon (pp. 181–202). Oxford: Elsevier. Singh, R. (2006) Whole word morphology. In K. Brown (ed.) Encyclopedia of Language & Linguistics (2nd edn, Vol. 13, pp. 578–579). Oxford: Elsevier. Singh, R. and Dasgupta, P. (1999) On So-Called Compounds. The Yearbook of South Asian Languages and Linguistics. New Delhi: DeGruyter. Westbury, C. (2007) The Alberta Language Function Assessment Battery [ALFAB]. Macintosh and Windows versions. Online at http://www.psych.ualberta.ca/westburylab/ publications.html

5

The Clinical Management of Anomia in Bilingual Speakers of Spanish and English Maria L. Muñoz

Introduction The ability to retrieve words, as modeled in psycholinguistics or cognitive neuropsychology, is a multistage process that involves spreading activation of the semantic and phonological systems (Dell et al., 1999). Fundamentally, this process of lexical retrieval is similar for unilingual and bilingual speakers; however, bilingual models must account for language selection, shared and disparate representation of multiple languages, proficiency, code-switching and translation (Costa, 2005). Anomia may result when neurological injury causes a breakdown of function in, or access to, one or more stages of the retrieval process (Raymer & Gonzalez Rothi, 2008). Research and clinical goals associated with anomia are similar for monolingual and bilingual speakers, and include the following: understanding the nature of the impairment and the implications for clinical management, characterizing the severity of the impairment, taking into consideration a variety of communicative demands and developing treatments that promote generalized improvement in lexical retrieval, be it functional or linguistic (Howard, 2000; Nickels, 2002; Roberts & Kiran, 2007). The purpose of this chapter is to provide a theoretical and clinical understanding of the manifestation, assessment and treatment of anomia in bilingual speakers of Spanish and English. First, I will provide a brief overview of the theoretical understanding of lexical retrieval in monolinguals and bilinguals. Second, I will consider how anomia manifests in monolingual and bilingual speakers. Clinical considerations will begin with an overview of available instruments for assessing lexical retrieval in Spanish and English. Finally, considerations in the treatment of Spanish speakers and bilinguals with anomia will be reviewed. 69

70

Par t 2: A ssessment and Treatment

Lexical Retrieval The ability to produce and comprehend words is critical to communication between speakers of a given language. The linguistic processes involved in manipulating even single words are complex. The brain of the speaker must formulate a thought, select a word form that best represents that thought, attach a phonological form to that thought and instruct the muscles to produce that form. The bilingual brain has the additional requirement to activate and/or recognize the appropriate language for the interaction, while inhibiting the non-target language (Costa, 2005). The process of lexical retrieval has received significant attention in both the unilingual and bilingual literature (La Heij, 2005). Cognitive psychological models of single word comprehension and production have been developed to account for the underlying linguistic process (Caramazza, 1997). The models capture the process of lexical retrieval as a series of interactive steps that reflect the relatively independent stages of the process. The complexity of the model reflects ‘necessary’ components that have been identified through studies of healthy and neurologically impaired speakers of a given language. Minimally, the process of lexical retrieval consists of three distinct components: modality-specific input, semantic/conceptual system and modality-specific phonological output (Raymer & Gonzalez Rothi, 2008). A fourth process which is accounted for in some models, particularly bilingual models, involves lexical selection (Dell et al., 1999). Speech production begins with the activation of conceptual/semantic representations which spreads to the lexical representation and then the phonological representations. Bilingual models must account for language selection in addition to the spreading activation needed for lexical retrieval. Costa (2005) presents a model in which a concept activates lexical representations in both languages. Activation then spreads to phonological forms in both languages, with stronger activation for the target language. The exact point at which a language is specified is unclear. Language selection may occur at the conceptual level (La Heij, 2005), or at a later stage in the process. Additionally, interactions between languages, for example, code-switching and translation, must be accounted for when modeling lexical retrieval in bilinguals. Lexical retrieval is influenced by a variety of stimulus factors, such as frequency, familiarity, age of acquisition and imageability (Raymer & Gonzalez Rothi, 2008). These stimulus variables differ for a bilingual’s two languages due to individual and linguistic factors (Goggin et al., 1994). Understanding these variables requires considering issues of proficiency and language history and the similarity between languages, particularly cognate status.

Anomia in Bilingual Speakers of Spanish and English

71

Anomia Anomia can be characterized as a breakdown in one or more levels of the lexical retrieval process (Howard, 2000). The level of impairment can be identified with evidence from patient performance on a variety of tasks that tap the different levels, effects of stimulus variables on performance (e.g. frequency vs word length), and the nature of the errors a patient produces (Raymer & Gonzalez Rothi, 2008). Similar types of errors, and patterns of errors associated with aphasia types, are produced by monolingual speakers of Spanish and English (Cuetos et al., 2002). Additionally, Spanish speakers may be more likely to exhibit morphemic verbal paraphasias, non-words that have been assembled using morphemes belonging to the language inventory (Ardila et al., 1989). In English, morphemic verbal paraphasias are most often seen in patients with Wernicke’s aphasia. These paraphasias may be more common in Spanish because of increased saliency of noun and verb inflections. As with English speakers, age of acquisition (AOA) is highly correlated with accuracy (the earlier the AOA the better the accuracy). According to a study by Cuetos et al. (2002) the strongest predictors of naming accuracy in Spanish were AOA, word frequency and visual complexity. Imageability, word length and animacy were weak predictors.

Assessment of Anomia Linguistic assessments of aphasia routinely include assessing naming function as anomia is almost always present. The assessment of lexical access should provide information regarding naming accuracy, patterns of strengths and weaknesses, types of errors produced and responsiveness to cuing (Raymer & Gonzalez Rothi, 2008). A thorough assessment of naming ability allows for formulation of hypotheses regarding the nature of the underlying deficit in the lexical retrieval process (i.e. the component[s] of the model) that is impaired. Hypothesis formulation relies heavily on establishing three sources of information (Whitworth et al., 2005): the effect of stimulus characteristics (i.e. frequency, word length, familiarity, age of acquisition, concreteness) on naming accuracy, the types of errors made on different tasks and the comparison of performance across tasks and modalities. Patterns of performance allow for predictions as to the nature of the breakdown, for example, phonological and/or semantic processes, or access to these processes. To meet the goals in assessing naming, tasks suggested for monolingual speakers with aphasia can be used to assess lexical retrieval in each language of bilinguals (e.g. compare recommendations made by Raymer & Gonzalez Rothi, 2008, with the tasks used to assess Spanish speakers with aphasia in a study by Martin et al., 1999). However, certain factors must be considered

72

Par t 2: A ssessment and Treatment

when assessing bilingual individuals. When assessing language processing in a single language the establishment of the appropriate language mode is critical (Grosjean, 1998). Bilingual contexts will activate both languages, even if one asks the patient to respond only in one language. Minimally, it is advisable to restrict interaction to a single language when testing that language. The nature of the stimuli that make up the naming task must be carefully considered, within and between languages. Stimulus characteristics (including frequency, familiarity, abstractness, complexity and age of acquisition) influence the ease and accuracy of naming (Raymer & Gonzalez Rothi, 2008). Understanding the equivalency between tests in multiple languages requires looking beyond the name of an item to examine the representation of that item in a given language and its appropriateness for use in that language. Additionally, the distribution of cognates (i.e. words in both languages with similar meaning and phonological form) and non-cognates among stimuli should be controlled and differences in performance on cognates/ non-cognates should be considered (Lalor & Kirsner, 2001). Cognates are typically easier to produce and less prone to error (Roberts & Deslauriers, 1999). Within a language, dialectal variation in the use of specific lexical items may impact its underlying features. For example, Zentella (1990) found similarities and differences in the preferred labels for common household items among Spanish speakers of Puerto Rican, Dominican, Columbian and Cuban descent living in New York. Lexical selection among bilinguals is further complicated by the incorporation of borrowed or loan words from both the recent and remote past (Cardenas, 1970). Interpretation of test results which include cross-linguistic comparisons should be done with caution given likely pre-morbid differences in proficiency (Kohnert et al., 1998; Muñoz & Marquardt, 2003). Detailed language background information should be obtained to infer likely pre-existing differences in language proficiency (Muñoz et al., 1999). Interpretations of differences between expressive and receptive tasks also should be made with caution. Bilinguals often comprehend more in their weaker language than they can produce to an extent greater than seen in the stronger language (Muñoz & Marquardt, 2003).

Tests of Lexical Access A variety of resources are available for testing lexical access in normal and neurologically impaired monolingual speakers of Spanish and English. There is sufficient diversity in the types of procedures available in Spanish and English to allow a clinician to evaluate semantic and phonological comprehension and expression as a means to generate and test hypotheses regarding the nature of the underlying impairment. The majority of instruments were designed for use with monolingual speakers of each language. When using

Anomia in Bilingual Speakers of Spanish and English

73

these instruments with bilingual speakers, the focus of assessment should be to understand how anomia manifests in a given language because crosslinguistic comparisons of type and severity of impairment are difficult due to factors that impact the performance of bilinguals and in turn the clinical interpretation of accuracy and errors. Clinicians must consider whether the item was known to the individual prior to the stroke and account for the influence of proficiency (Muñoz & Marquardt, 2003) and dialect. Cognates and non-cognates appearing on any test may need to be considered separately (Roberts & Kiran, 2007). While tests may be administered in both Spanish and English, it is unlikely the forms are equivalent with respect to crosslinguistic differences in stimulus variables such as age of acquisition, familiarity, word length and frequency (Roberts & Kiran, 2007). Below is a summary of instruments and stimuli that can be used in assessing lexical comprehension and expression on a variety of tasks, including confrontation naming. The evidence for the utility of most measures is limited to the initial studies done by researchers in the process of test development. However, the data are sufficient to support their use in evaluating a variety of lexical processing skills.

Confrontation naming One of the most commonly utilized methods for assessing lexical retrieval is the confrontation naming task; more specifically, naming nouns in response to a picture, drawing or object. Both clinical- and research-oriented measures are available to assess naming. Additionally, clinicians can select items from corpora of naming stimuli to develop informal tests of naming. The following is an overview of available formal and informal measures for assessing confrontation naming, and available supporting evidence.

The Boston Naming Test (BNT) Among formal and informal confrontation naming tasks (see Appendix), the most well known is the Boston Naming Test (Kaplan et al., 2001). The BNT is composed of 15 (short version) or 60 (standard version) black-line drawings ordered by item difficulty, as determined by the standardization of the English version. The Spanish version (Kaplan et al., 2005) differs from the English version only in that five of the 60 items have been substituted, presumably to account for issues of familiarity. The administration process provides information regarding responsiveness to phonemic cues, types of errors and recognition (out of four choices). The standard version should be administered in its entirety rather than in accordance with the recommended ceiling of eight consecutive errors because the order of item difficulty varies in Spanish for Spanish-English bilinguals (Kohnert et al., 1998) and for monolingual Spanish speakers (Allegri et al., 1997). Based on the performance of 200 monolingual Spanish speakers in

74

Par t 2: A ssessment and Treatment

Buenos Aires, Allegri et al. (1997) suggest a revised item order to reflect item difficulty in Spanish. Additionally, given that naming is significantly correlated with education, Allegri et al. (1997) suggest alternate cut-off scores for Spanish speakers based on education. The authors note that accuracy on items numbered 41 and above was particularly influenced by education.

Modifications of the BNT The Modified BNT for Spanish speakers (MBNTS) was developed by asking expert judges to rate the original 60 items on appropriateness and difficulty (Ponton et al., 1992). The authors then selected 30 of the most appropriate items and re-ordered them according to rated difficulty (easy to hard). The authors normed the test on 300 monolingual and bilingual speakers of Spanish in Los Angeles and provide norms stratified by age, education and gender. The MBNTS has been shown to discriminate between individuals with and without dementia (Marquez de la Plata et al., 2009). Additionally, a linear item difficulty was evident for speakers of Columbian and Castilian Spanish, suggesting its utility across a variety of Spanish dialects (Marquez de la Plata et al., 2009).

The Spanish Naming Test (SNT) The Spanish Naming Test (Ardila et al., 1994) consists of 15 items ordered according to item difficulty. Six of the items were taken from the BNT while the remaining items were selected specifically for Spanish speakers. The cuing and scoring system from the BNT are retained with minor modifications. The authors sampled 346 normal Spanish speakers and provide norms stratified by age and education. The SNT provides a familiar procedure that can be utilized as a quick test of confrontation naming with norms to interpret performance.

Modifications of the Snodgrass and Vanderwart (S & V; 1980) corpus The standardized set of 260 pictures of objects was selected to be highly picturable, from commonly studied semantic categories, and to be represented at their basic level (Snodgrass & Vanderwart, 1980). Norms for name agreement, familiarity and visual complexity are provided for each item based on the ratings and responses of a group of native speakers of American English. Subsequently, Sanfeliu and Fernandez (1996) developed comparable norms for a set of 254 S&V pictures utilizing the responses of 261 Spanishspeaking students in Spain. A statistical comparison of Spanish and English speakers revealed a small but significant difference in familiarity and perceptions of visual complexity. Comparable analyses were conducted for Mexican-Spanish speakers (Aveleyra et al., 1996) and Cuban-Spanish speakers (Manzano et al., 1997). While the original corpus was designed for research purposes, the published analyses of stimulus characteristics can assist in the development of measures of lexical processing that are balanced and comparable in English and Spanish. These tasks may be developed

Anomia in Bilingual Speakers of Spanish and English

75

informally, or clinicians may utilize confrontation naming tasks that have been developed by a variety of researchers.

The Texas Spanish Naming Test (TNT) The Texas Spanish Naming Test (TNT) is a 30 item naming test designed specifically for Spanish speakers (Marquez de la Plata et al., 2008). The authors selected culturally salient words while accounting for stimulus variables such as word frequency, familiarity and age of acquisition. The drawings of objects were selected from the S&V corpus accounting for the stimulus variables as related to Spanish speakers reported by Cuetos et al. (1999). Of the pool of stimuli generated, the 30 selected for the final version of the test were those determined to be the most psychometrically reliable items during a standardization process. The items are ordered by difficulty, and administration and scoring techniques match those of the BNT. Marquez de la Plata et al. (2008) provide evidence of the test’s validity and reliability as well as a list of the 30 items. Naming accuracy on the TNT discriminated effectively between groups of participants with and without dementia, though effects of age, education and acculturation were evident. Additionally, the TNT was found to demonstrate good internal consistency, and good patterns of item difficulty and discrimination for Spanish speakers residing in the United States, Columbia and Spain (Marquez de la Plata et al., 2009). The TNT appears to be a valid test for assessing confrontation naming across a variety of Spanish dialects.

The Bilingual English Naming Tool (BENT) The Bilingual English Naming Tool (BENT) is a test of English naming ability in bilingual (Spanish/English) individuals with varying degrees of proficiency in English, their second language (Casas et al., 2008). Sixty-seven bilingual adult speakers of Spanish and English (mean age of 29.8) were asked to name 299 pictures, of which approximately 50% were from the S&V corpus. After eliminating items with a ‘no response’ rate of 5% or higher and items with lower name agreement, 51 items remained (21 cognates, 30 noncognates). The BENT is one of the few tests designed specifically for bilinguals. As such, it is intended to assess confrontation naming in English for Spanish speakers with varying degrees of proficiency in their second language. However, the utility of the BENT for accurately determining the nature and severity of anomia, particularly in older adults, needs further investigation. The S&V corpus was originally designed from the perspective of objects relevant to the English-speaking culture of the United States, suggesting that some items may not be relevant for Spanish speakers or that items significant in Spanish-speaking cultures are not represented. Perez and Navalon (2003) created a corpus of 290 drawings specifically for use with Spanish speakers. Norms for name agreement, familiarity and visual complexity are provided for each item based on the ratings and responses of a group of native

76

Par t 2: A ssessment and Treatment

Spanish speakers from Spain. The corpus retains much in common with that of Snodgrass-Vanderwort. Some items were retained but more clearly recognizable drawings were substituted, while other items were omitted or added. While the test has not been modified for clinical use, the fact that the publisher has made the appendices available for download makes the set accessible; as a result, clinicians and researches can use the stimulus database to select stimuli in order to develop informal measures of lexical processing.

Picture pool for oral naming (PEDOI) The PEDOI is an online corpus of pictures from a variety of semantic categories standardized in eight European languages, Canadian French and Latin-American Spanish (Kremin et al., 2003). Measures of name agreement, visual complexity, familiarity, imageability and age of acquisition are available for 269 pictures. Cross-linguistic correlations were strong for all variables except name agreement, which appears to be the variable most influenced by personal and cultural experiences. A set of stimuli for which data is easily available in various languages (and dialects) facilitates the development of balanced tasks, and may be particularly useful for developing comparable tasks for use with individuals who speak more than two languages.

International Picture Naming Project (IPNP) The IPNP is an online resource of picture stimuli and related norms in seven languages designed for psycholinguistic research (Szekely et al., 2004). The corpus includes 520 drawings of objects and 275 actions with norms for age of acquisition, frequency, familiarity, goodness of depiction and visual complexity. An online database query system allows the user to search for specific items. The subset of items available to the general public online can be used to develop a controlled informal assessment of action and object naming. Additionally, the IPNP is one of the few resources available with information that can be useful in developing tasks to assess action naming.

Picture naming subtests Picture naming tasks are included as subtests of larger batteries available in English and Spanish (see Appendix). The Multilingual Aphasia Test (MAT; Benton et al., 2000) and the Cognitive Linguistic Quick Test (CLQT; HelmEstabrooks, 2001) include a 30-item and a 10-item picture naming test, respectively. Assessing primarily noun retrieval, both tests are criterion referenced and provide cut-off scores for interpreting performance. These subtests allow for a quick screening of naming abilities in the context of a broader language assessment. Subtest 51 of the Evaluación del Procesamiento Lingüístico en la Afasia (EPLA; Kay et al., 1995) is a 40-item picture naming task designed to be administered in various output modalities. The authors provide means and standard deviations for accuracy in each response modality (naming, writing, reading and repetition). Additionally, the subtest is designed to use error

Anomia in Bilingual Speakers of Spanish and English

77

analysis to generate hypotheses regarding the level of impairment, the direction of additional testing and appropriate treatment approaches.

Lexical comprehension and production While the picture naming task is the most commonly utilized means of assessing lexical retrieval, it may not be sufficient to adequately represent lexical strengths and weaknesses. For example, individuals with dementia tend to perform better on confrontational naming tasks than on generative naming tasks (Bayles & Boone, 1982). Picture naming tasks, such as the BENT, ask the individual to produce nouns that have a high familiarity. These tests may fail to capture increased difficulty, producing less familiar words or words from other grammatical categories such as verbs (Kim & Thompson, 2000). To avoid this limitation, and to better hypothesize the nature of the impairment to guide treatment, a variety of tasks should be considered.

Evaluación del Procesamiento Lingüístico en la Afasia (Kay et al., 1995) The EPLA is a translated and modified version of the Psycholinguistic Assessment of Language Processing in Aphasia (Kay et al., 1992). The EPLA is a criterion referenced measure composed of 60 subtests which can be used to assess expressive and receptive syntax and semantics across a variety of modalities. The lexical tasks are designed to assess phonological and semantic comprehension and expression in a variety of modalities (spoken, written, repetition, reading). Tasks to assess phonological processing include minimal pair discrimination, rhyme judgment, non-word and word repetition, and sound segmenting of initial and final phonemes. Semantic tasks include identification of semantic associates, word to picture matching and synonym judgment. The clinician is not expected to administer all subtests. Rather, judicious observation of a client’s responses and error types should lead to the selection of specific subtests which help to identify the nature of the lexical impairment. Means and standard deviations are provided for each subtest to facilitate the interpretation of the client’s performance. The standardization on which statistics are based consisted of 22 neurologically normal Spanish speaking adults (between 35–65 years old) with a ‘not-too-high’ level of education. Given the limited standardization sample, the comparisons should be made with caution, particularly when using the tests with bilinguals who were not educated in Spanish or for whom Spanish is a second language (Muñoz & Marquardt, 2003). Additionally, given that the language of the translation was Castilian Spanish, some adjustment may need to be made for dialect variations in phonology and vocabulary when administered to speakers of other Spanish dialects.

Evaluación de la Afasia y de Trastornos Relacionados (BDAE-3S) The BDAE-3S, a slightly modified Spanish translation of the Boston Diagnostic Aphasia Examination-3 (Goodglass & Kaplan, 2001), offers

78

Par t 2: A ssessment and Treatment

several subtests which assess lexical comprehension and production. Lexical comprehension can be assessed via a word-to-picture matching task using words from a variety of categories (body parts, nouns, colors, letters and numbers). A supplemental subtest facilitates the assessment of semantic knowledge by requiring the patient to match items in the same category (tools, foods, animals). Semantic knowledge also is assessed in the semantic probes subtest in which the client is presented with a word then asked yes/ no questions regarding its semantic features. In addition to the BNT, lexical production can be assessed using two subtests: naming to definition and category naming (letters, numbers, colors, verbs, animals, tools). As with the EPLA, the test was designed for use with monolingual speakers of Castilian Spanish. Some allowances and modifications may need to be made in administering the test to bilingual speakers and speakers of other Spanish dialects. Ratings and percentile ranks are provided to guide identification of type and severity of impairment. However, the standardization was taken from the BDAE-3. Standardization data were not collected for the Spanish version and these ranks should not be used due to potential test bias.

The Bilingual Aphasia Test (BAT; Paradis, 1987) The BAT is a criterion referenced measure designed to assess the relative recovery of a bilingual’s languages. Phonological tasks include verbal auditory discrimination and word and non-word repetition with lexical judgment. Semantic tasks include recognition of synonyms and antonyms, and matching words in the same semantic category. The BAT is particularly useful to explore deficits in phonological comprehension and expression given that the semantic subtests tend to include only a small number of items. Unique to the BAT are two translation subtests requiring recognition of the translation word, and spontaneous production of the translation equivalent. The translation subtests may be useful in examining the influence of one language on the retrieval of another language. Though the BAT is a criterion referenced test which assumes the typical bilingual would score at 100% accuracy (Paradis, 1987), research has demonstrated that performance on the BAT must take into account the influence of pre-morbid proficiency in each language (Munoz & Marquardt, 2008). The BAT is available online in a number of language pairs and dialects of Spanish and English.

Verbal fluency tasks The MAT (Benton et al., 2000) and the Neuropsychological Evaluation of Spanish Speakers (NESS; Ardila et al., 1994) utilize verbal fluency tasks which require individuals to name as many items as possible in one minute within a given semantic or phonological category (e.g. animals, fruits, F, A and S). The NESS provides norms for phonological and semantic category generation stratified by age and education. The MAT provides for the calculation of percentile ranks and interpretations for each rank in the English and

Anomia in Bilingual Speakers of Spanish and English

79

Spanish versions. These tests provide some of the only norms available in Spanish for a generative naming task.

Discourse The functional impact of anomia can be assessed in discourse contexts. Muñoz and Marquardt (2004) evaluated lexical retrieval using conversation samples gathered in monolingual (English and Spanish) and a bilingual (code-switching) contexts. Measures of interest included semantic accuracy (presence/absence of semantic errors), verbal disruptions (including fillers and pauses), and frequency counts of specific naming errors (e.g. verbal or literal paraphasias, circumlocutions, use of other language in monolingual contexts, among others). The participants demonstrated differences in naming accuracy, and patterns of naming errors, associated with pre-morbid proficiency, conversational context and nature of the impairment. For example, one participant demonstrated more naming difficulty in monolingual contexts due to the need to search for the correct word in the target language. In the bilingual context he would code-switch as needed with minimal effort.

Treatment of Anomia Treatment to improve lexical retrieval in one language can target the phonological and/or semantic bases of naming. The form of the therapy may depend on the responsiveness of the patient to specific cues, treatment goals and the hypotheses generated during assessment regarding the underlying nature of the naming deficit. Treatment may strengthen the hypothesized area of deficit, or improve intact processes as a means of bypassing the impaired processes (Raymer & Gonzalez Rothi, 2008). While the literature related to treatment of anomia in monolingual English speakers is relatively extensive (Nickels, 2002), research involving individuals who are Spanish speaking or bilingual speakers is sparse. Recovery in each language is likely influenced by a variety of factors including the nature of therapy (language, intensity and type), age of acquisition of each language, literacy and aphasia type and lesion (Roberts & Kiran, 2007). Some issues that can be addressed, at least to some extent, given the current evidence regarding the treatment of anomia in bilinguals with aphasia include generalization or effects across languages, cuing hierarchies and treatment of related processes.

Generalization across languages Of particular interest in measuring treatment outcomes for bilinguals with aphasia is understanding the extent to which, and under what circumstances, treatment in one language generalizes to treatment in another (see

80

Par t 2: A ssessment and Treatment

Kohnert & Peterson, Chapter 6 in this volume, for related discussion). Structuring treatment to take advantage of generalization can facilitate maximum outcomes and may be particularly useful for clinicians who speak only one of the patient’s languages. One consideration is the identification of patterns of generalization, particularly in relation to status of the language of treatment (e.g. first language [L1] or second language [L2]). To date, results have been mixed. Galvez and Hinkley (2004), in a case study of a Spanish/English bilingual with transcortical aphasia, found that the treated language improved with no measurable improvement in the untreated language, suggesting that no generalization occurred. Edmonds and Kiran (2006) found that proficiency in each language seemed to influence generalization effects. The authors sequentially treated each language of three bilinguals (Spanish/ English) with aphasia using a semantic mapping task. Participants with higher proficiency in English than Spanish demonstrated generalized improvement in English but not Spanish when treated in English. When treated in Spanish patients demonstrated no generalized improvement in Spanish, but improved naming for English equivalents and semantically related items. Participants who were equally proficient in English and Spanish demonstrated generalized improvement for the trained and untrained languages. Proficiency may also influence the amount of generalization from one language to the other. Hinckley (2003) studied a bilingual (Spanish/ English) patient with transcortical motor aphasia. The patient received treatment for anomia for seven weeks, alternating languages from one session to the next. Spanish (L1) improved at twice the level of English (L2), suggesting L1 benefited from treatment in both L1 and L2 to an extent greater than L2. The nature of the stimuli used in treatment may also influence the extent to which treatment generalizes across languages. Items with similar semantic and phonological properties in each language may be more likely to result in improved naming in both languages when only treated in one. Kohnert (2004) reported treating a bilingual (Spanish/English) patient with a cognate-based treatment. Twenty concrete nouns (10 cognates) were selected. When treatment was conducted in only one language, treatment of the cognates transferred to the non-target language more than treatment of non-cognates. Costa (2005) suggests that the advantage for cognates is due to stronger cross-language activation of the target and its translation due to the lexical and phonemic similarity. Likely, generalization of treatment across language is a complicated process that is impacted by a number of person-specific variables including premorbid proficiency, age of acquisition of each language and the nature and severity of the impairment.

Anomia in Bilingual Speakers of Spanish and English

81

Cuing hierarchies The effectiveness of cues to facilitate production of a word may vary in the first and second language depending on proficiency (see Table 5.1). Preliminary evidence for this conclusion comes from a study by Bond and Hein (1985). They report the case of a bilingual (Spanish/English) patient with aphasia. Spanish was his first and better language, English his second language acquired after the age of 21. As a treatment for anomia they asked the patient to name pictures. If the patient was unable to name the picture, the authors provided one of three cues: initial syllable, sentence completion, translation. In the Spanish context the initial syllable was most facilitative. In the English context the translation cue was most facilitative. Use of the translation cue takes advantage of the natural L1 activation evident when low proficiency bilinguals make lexical selections in L2. Activation and selection in a less proficient bilingual relies, to some extent, on L1 activation. The effectiveness of specific cues may depend on the nature of the semantic system in the specific language. Unlike English, articles are inflected to agree with the gender of the noun. Because of the article’s close relationship to the noun, it may serve as a facilitative cue in Spanish in a way that the English article does not (see Table 5.1). Additionally, Spanish and English differ in that Spanish is heavily multisyllabic while English has a higher occurrence of monosyllabic words. As such, a phonemic cue may have

Table 5.1 Examples of possible cuing hierarchies based on proficiency and language Language proficiency L1

L2

(1) Translation

(1) Sentence completion

(2) Sentence completion

(2) Initial syllable

(3) Initial syllable

(3) Translation cue

Language features Spanish

English

(1) Semantic cue

(1) Semantic cue

(2) Article cue (un/una, el/la)

(2) Rhyme cue

(3) Rhyme cue

(3) Phonemic cue

(4) Initial syllable cue

(4) Written word cue

(5) Written word cue

(5) Repetition

(6) Repetition Sources: Bond and Hein (1985), Raymer (2005) and Wambaugh (2003)

82

Par t 2: A ssessment and Treatment

less facilitative effects in Spanish where the smallest unit appears to be the syllable rather than the phoneme.

Treatment of related processes The treatment of underlying cognitive impairments has been shown to result in generalized improvement in language (Murray, 1999). In the case of bilinguals, treating cognition, a shared process, may result in improvement in both languages. Kohnert (2004) provided cognitive treatment for a bilingual patient (card sorting, math computations, visual number and letter searches). Though tasks were non-verbal, interactions with the participant were in Spanish the first month of treatment, and in English the second. Results indicated post-treatment improvement in both languages, as well as cognition (see Kohnert & Peterson, Chapter 6 in this volume, for related discussion).

Summary The clinical management of anomia in bilingual speakers of Spanish and English is ultimately a patient-specific process that is influenced by a variety of factors including the nature and severity of the anomia, pre-morbid language proficiency and language history. There are a number of existing measures, and resources for creating instruments, to assess lexical processing across modalities and to allow for hypothesizing the nature of the underlying impairment. A hypothesis of the nature of the impairment, in turn, can guide treatment planning. While additional research is needed to understand how best to treat anomia in both languages, preliminary evidence suggests methods to foster generalization of treatment effects across languages.

References Allegri, R.F., Mangone, C.A., Villavicencio, A.F., Rymberg, S., Taragano, F.E. and Baumann, D. (1997) Spanish Boston Naming Test norms. The Clinical Neuropsychologist 11 (4), 416–420. Ardila, A., Montanes, P., Caro, C., Delgado, R. and Buckingham, H.W. (1989) Phonological transformations in Spanish-speaking aphasics. Journal of Psycholinguistic Research 18 (2), 163–180. Ardila, A., Roselli, M. and Puente, A.E. (1994) Neuropsychological Evaluation of the Spanish Speaker. New York: Plenum Press. Aveleyra, E., Gomez, C., Ostrosky, F. and Rogalt, C. (1996) Adaptación de los estímulos no verbales de Sndograss y Venderwart en población hispanoparlante: criterios para la denominación concordancia de la imagen, familiaridad y complejidad visual. Revista Mexicana de Psicología 13, 5–19. Bayles, K.A. and Boone, D.R. (1982) The potential of language tasks for identifying senile dementia. Journal of Speech and Hearing Disorders 47, 210–217.

Anomia in Bilingual Speakers of Spanish and English

83

Benton, A., Hamsher, K.d.S. and Sivan, A. (2000) Multilingual Aphasia Examination. Iowa City: AJA Associates, Inc. Bond, S.L. and Hein, S.L. (1985) Word finding abilities in a bilingual anomic aphasic: A case study. The Journal of Neurological Communication Disorders 2 (1), 36–40. Caramazza, A. (1997) How many levels of processing are there in lexical access? Cognitive Neuropsychology 14 (1), 177–208. Cardenas, D.R. (1970) Dominant Spanish Dialects Spoken in the United States. New York: ERIC Clearinghouse for Linguistics. Casas, R., Calamia, M. and Tranel, D. (2008) A screening test of English naming ability in bilingual Spanish/English speakers. Journal of Clinical and Experimental Neuropsychology 30 (8), 956–966. Costa, A. (2005) Lexical access in bilingual production. In J.F. Kroll and A. De Groot (eds) Handbook of Bilingualism: Psycholinguistic Approaches (pp. 308–325). New York: Oxford University Press. Cuetos, F., Ellis, A.W. and Alvarez, B. (1999) Naming times for the Snodgrass and Vanderwart pictures in Spanish. Behavior Research Methods, Instruments, and Computers 31 (4), 650–658. Cuetos, F., Aguado, G., Izura, C. and Ellis, A.W. (2002) Aphasic naming in Spanish: Predictors and errors. Brain and Language 82, 344–365. Dell, G.S., Chang, F. and Griffin, Z.M. (1999) Connectionist models of language production: Lexical access and grammatical encoding. Cognitive Science 23 (4), 517–542. Edmonds, L. and Kiran, S. (2006) Effect of semantic naming treatment on crosslinguistic generalization in bilingual aphasia. Journal of Speech Language and Hearing Research 49 (4), 729–748. Galvez, A. and Hinckley, J.J. (2003) Transfer patterns of naming treatment in a case of bilingual aphasia. Brain and Language 87, 173–174. Goggin, J.P., Estrada, P. and Villarreal, R.P. (1994) Picture-naming agreement in monolinguals and bilinguals. Applied Psycholinguistics 15, 177–193. Goodglass, H. and Kaplan, E. (2001) The Boston Diagnostic Aphasia Examination (3rd edn). Philadelphia, PA: Lippincott Williams & Wilkin. Grosjean, F. (1998) Studying bilinguals: Methodological and conceptual issues. Bilingualism: Language and Cognition 1 (2), 131–149. Helm-Estabrooks, N. (2001) Cognitive Linguistic Quick Test. San Antonio, TX: The Psychological Corporation. Hinckley, J.J. (2003) Picture naming treatment in aphasia yields greater improvement in L1. Brain and Language, 171–172. Howard, D. (2000) Cognitive neuropsychology and aphasia therapy: The case of word retrieval. In I. Papathanasiou (ed.) Acquired Neurogenic Communication Disorders: A Clinical Perspective (pp. 76–99). London: Whurr Publishers. Kaplan, E., Goodglass, H. and Weintraub, S. (2001) Boston Naming Test. Philadelphia, PA: Lippincott Williams & Wilkins. Kaplan, E., Goodglass, H. and Weintraub, S. (2005) Test de Vocabulario de Boston. Madrid: Editorial Medica Panamericana. Kay, J., Lesser, R. and Coltheart, M. (1992) Psycholinguistic Assessments of Language Processing in Aphasia (PALPA). East Sussex: Lawrence Erlbaum. Kay, J., Lesser, R. and Coltheart, M. (1995) Evaluación del Procesamiento Lingüístico en la Afasia. East Sussex: Erlbaum (UK) Taylor and Francis. Kim, M. and Thompson, C.K. (2000) Patterns of comprehension and production of nouns and verbs in agrammatism: Implications for lexical organization. Brain and Language 74, 1–25. Kohnert, K. (2004) Cognitive and cognate-based treatments for bilingual aphasia: A case study. Brain and Language 91, 294–302.

84

Par t 2: A ssessment and Treatment

Kohnert, K.J., Hernandez, A. and Bates, E. (1998) Bilingual performance on the Boston Naming Test: Preliminary norms in Spanish and English. Brain and Language 65, 422–440. Kremin, H., Akhutina, T., Basso, A., Davidoff, J., de Wilde, M., Kitzing, P., Lorenz, A., Perrier, D., van der Sandt-Koenderman, M., Vendrell, J. and Weniger, D. (2003) A cross-linguistic data bank for oral picture naming in Dutch, English, German, French, Italian, Russian, Spanish, and Swedish (PEDOI). Brain and Cognition 53, 243–246. La Heij, W. (2005) Selection processes in monolingual and bilingual lexical access. In J.F. Kroll and A. De Groot (eds) Handbook of Bilingualism: Psycholinguistic Approaches (pp. 289–307). Oxford: Oxford University Press. Lalor, E. and Kirsner, K. (2001) The role of cognates in bilingual aphasia: Implications for assessment and treatment. Aphasiology 15 (10–11), 1047–1056. Manzano, M., Pineiro, A. and Reigosa, V. (1997) Estudio de las características de un conjunto de 260 figuras en sujetos adultos de habla hispana. Cognitiva 9, 29–64. Marquez de la Plata, C. Arango-Lasprilla, J., Alegrat, M., Moreno, A., Tárraga, L., Lara, M., Hewlitt, M., Hynan, L. and Cullum, M. (2009) Item analysis of three Spanish naming tests: A cross-cultural investigation. Neurorehabilitation 24 (1), 75–85. Marquez de la Plata, C., Vicioso, B., Hynan, L., Evans, H.M., Diaz-Arrastia, R., Lacritz, L. and Cullum, M.C. (2008) Development of the Texas Spanish Naming Test: A test for Spanish speakers. The Clinical Neuropsychologist 22, 288–304. Martin, P., Serrano, J.M. and Iglesias, J. (1999) Phonological/semantic errors in two Spanish-speaking patients with anomic aphasia. Aphasiology 13 (3), 225–236. Muñoz, M.L. and Marquardt, T.P. (2003) Picture naming and identification in bilingual speakers with and without aphasia. Aphasiology 17 (12), 1115–1132. Muñoz, M.L. and Marquardt, T.P. (2004) The influence of language context on lexical retrieval in the discourse of bilingual speakers with aphasia. Journal of Multilingual Communication Disorders 2 (1), 1–17. Muñoz, M.L. and Marquardt, T.P. (2008) The performance of neurologically normal bilingual speakers of Spanish and English on the short version of the Bilingual Aphasia Test. Aphasiology 22 (1), 3–19. Muñoz, M.L., Marquardt, T.P. and Copeland, G. (1999) A comparison of the codeswitching patterns of aphasic and neurologically normal bilingual speakers of English and Spanish. Brain and Language 66, 249–274. Murray (1999) Attention and aphasia: Theory, research, and clinical implications. Aphasiology 13 (2), 91–111. Nickels, L. (2002) Therapy for naming disorders: Revisiting, revising, and reviewing. Aphasiology 16 (10–11), 935–979. Paradis, M. (1987) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum Associates. Perez, M-A. and Navalon, C. (2003) Normas españolas de 290 nuevos dibujos: Acuerdo en la denominación, concordancia de la imagen, familiaridad, complejidad visual y variabilidad de la imagen. Psicología 24, 215–241. Ponton, M.O., Satz, P., Herrera, L., Young, R., Ortiz, F., D’elia, L., Furst, C. and Namerow, N. (1992) A modified Spanish version of the Boston Naming Test. The Clinical Neuropsychologist 3, 334. Raymer, A.M. (2005) Naming and word-retrieval problems. In L. LaPointe (ed.) Aphasia and Related Neurogenic Language Disorders (pp. 68–82). New York: Thieme. Raymer, A.M. and Gonzalez Rothi, L.J. (2008) Impairments of word comprehension and production. In R. Chapey (ed.) Language Intervention Strategies in Aphasia and Neurogenic Communication Disorders (pp. 607–629). Baltimore, MD: Lippincott, Williams & Wilkins. Roberts, P. and Deslauriers, L. (1999) Picture naming of cognate and non-cognate nouns in bilingual aphasia. Journal of Communication Disorders 32, 1–23.

Anomia in Bilingual Speakers of Spanish and English

85

Roberts, P. and Kiran, S. (2007) Assessment and treatment of bilingual aphasia and bilingual anomia. In A. Ardila and E. Ramos (eds) Speech and Language Disorders in Bilinguals (pp. 109–131). New York: Nova Science Publishers, Inc. Sanfeliu, M.C. and Fernandez, A. (1996) A set of 254 Snodgrass-Vanderwart pictures standardized for Spanish: Norms for name agreement, image agreement, familiarity, and visual complexity. Behavior, Research Methods, Instruments, and Computers 28 (4), 537–555. Snodgrass, J.G. and Vanderwart, M. (1980) A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology 6 (2), 174–215. Szekely, A., Jacobsen, T., D’Amico, S., Devescovi, A., Andonova, E., Herron, D., Lu, C., Pechmann, T., Pleh, C., Wicha, N., Federmeier, K., Gerdijikova, I., Gutierrez, G., Hung, D., Hsu, J., Iyer, G., Kohnert, K., Mehotcheva, T., Orozco-Figueroa, A., Tzeng, A., Tzeng, O., Arevalo, A., Vargha, A., Butler, A., Buffington, R. and Bates, E. (2004) A new on-line resource for psycholinguistic studies. Journal of Memory and Language 51, 247–250. Wambaugh, J. (2003) A comparison of the relative effects of phonological and semantic cuing treatments. Aphasiology 17 (5), 433–444. Whitworth, A., Webster, J. and Howard, D. (2005) A Cognitive Neuropsychological Approach to Assessment and Intervention in Aphasia: A Clinician’s Guide. New York: Psychology Press. Zentella, A.C. (1990) Lexical leveling in four New York City Spanish dialects: Linguistic and social factors. Hispania 73 (4), 1094–1105.

86

Par t 2: A ssessment and Treatment

Appendix Table A5.1 Tools for measuring lexical retrieval in Spanish and English Tests of confrontation naming Test

Subtests

Description

Boston Naming Test

Short version (15 items) Standard version (60 items)

The Spanish version is primarily a translation of the English version with the exception of five items. The Spanish version was developed based on Castilian Spanish. Some vocabulary items may need to be modified for use with other dialects.

Texas Spanish Naming Test (TNT)

A 30-item naming test designed specifically for Spanish speakers (Marquez de la Plata et al., 2008).

15 Item Spanish Naming Test (SNT)

Includes six English BNT items and nine items with high frequency in Spanish (Ardila et al., 1994).

290 Item Spanish Picture Naming Test

Provides pictures and stimulus characteristics for 290 items with cultural relevance for Spanish speakers (Perez & Navalon, 2003). Parallels work done in English by Snodgrass and Vanderwort (1980).

The Modified Boston Naming TestSpanish (MBNT-S)

A 30-item adaptation of the English BNT (Ponton et al., 1992).

Bilingual English Naming Tool (BENT)

A 51-item test designed to assess English naming in Spanish/ English bilinguals (Casas et al., 2008).

Picture Pool for Oral Naming

Individual items standardized for eight European languages, Canadian French and LatinAmerican (Kremin et al., 2003).

Anomia in Bilingual Speakers of Spanish and English

87

Table A5.1 (Continued) International Picture Naming Project

Action and object naming

Individual items standardized for various languages, including Mexican Spanish and American English (Szekely et al., 2004).

Test batteries with naming subtests Test

Subtests

Description

Evaluación de la Afasia y de Trastornos Relacionados

• Word comprehension • Word comprehension by categories • Semantic probes • Responsive naming

The Spanish version is primarily a translation of the English version with some small changes in vocabulary.

Boston Diagnostic Aphasia Examination-3 (BDEA-3)

• Category naming • Repetition

The Spanish version was developed based on Castilian Spanish. Some vocabulary items may need to be modified for use with other dialects.

Multilingual Aphasia Test

• Picture naming • Verbal fluency • Repetition

The Spanish version is primarily a translation of the English version with modifications for phonological differences.

PALPA/EPLA

• Minimal pair discrimination • Lexical comprehension • Rhyme judgment • Word to picture matching • Semantic judgment • Semantic associations • Written word to picture matching • Picture naming • Sound segmenting • Synonym judgment

Individual subtests may or may not be direct translations. In particular, phonologically based tasks differ across languages. Semantically based tasks in Spanish are often translations of the comparable English subtest. Versions are based on British English and Castilian Spanish. Modifications may need to be made when used with other dialects.

Bilingual Aphasia Test

• Verbal auditory discrimination • Semantic categories • Synonyms • Antonyms

The Spanish version is primarily a translation of the English version Letters with modifications for phonological differences. (continued)

88

Par t 2: A ssessment and Treatment

Table A5.1 (Continued) • Verbal fluency • Word repletion and lexical decision • Object naming • Translation: Word recognition • Translation of single words Cognitive Linguist Quick Test

10-item confrontation naming

Items are identical in English and Spanish.

Neuropsychological Evaluation of Spanish Speakers

• Spanish naming test • Verbal fluency • Repetition • Phonological discrimination

Designed specifically for Spanish speakers, though the many subtests draw heavily from existing English tests. Provides norms for subtests based on the performance of normal and/or brain damaged adults.

6

Generalization in Bilingual Aphasia Treatment Kathryn Kohnert and Michael Peterson

This chapter focuses on relationships between languages and basic cognitive skills in bilingual adults with aphasia, and the implications of these relationships for treatment. Aphasia is a common acquired primary language disorder, most often the result of focal damage to language processing centers in the left cerebral hemisphere. Individuals with aphasia are distinguished from those with other types of cognitive-communication disorders such as traumatic brain injury, right hemisphere disorder or dementia by their medical history, the nature and location of brain damage, as well as by the language and cognitive behavioral profile. Aphasia onset is sudden and most often the result of a stroke. Additional causes of aphasia include tumors, aneurysms, gunshot wounds or other focal trauma. Aphasia is characterized by deficits in the expression and/or comprehension of language in spoken and/or written modalities. Inefficiencies or weaknesses in the basic cognitive information processing system may exist alongside the more easily recognized language difficulties. Bilinguals are operationally defined as individuals with systematic experience and communicative need for two or more languages. The relative proficiency or ability in each of a bilingual’s languages is a product of multiple interacting factors, including age and context of acquisition, opportunities to develop and use each language, the social prestige associated with each language in the broader community as well as individual preferences, capacities and motivations. As with single-language speakers, damage to cortical or subcortical regions of the language-dominant hemisphere may produce aphasia in bilinguals. Given that both languages of bilingual adults are processed in overlapping or adjacent brain areas, each language will be affected to some extent in the vast majority of bilinguals with acquired aphasia. The clinical challenge is to create conditions that facilitate recovery or use of both/all of an individual’s languages in the most effective and efficient manner possible. The generalization of skills from a treated to an untreated language in bilinguals with aphasia would represent effective as well as efficient treatment. 89

90

Par t 2: A ssessment and Treatment

This chapter has four main sections. First, we discuss treatment outcomes in monolingual individuals with aphasia. We then present general findings from treatment studies investigating cross-language generalization in bilingual individuals with aphasia. In the third section we present cognitive characteristics of monolingual and bilingual individuals with aphasia. We conclude with a discussion of treatment procedures designed to promote recovery of two languages in bilinguals with aphasia, including strategies that may facilitate cross-language or cross-domain (cognition to language) generalization.

Treatment Efficacy and Generalization in Monolingual Aphasia Language is made possible by the anatomy and physiology of the human brain (see Hernandez et al., 2009, for review). Damage to certain parts of the brain may result in aphasia, which appears as sudden and chronic difficulty in producing or understanding language in spoken or written modalities. The parts of the brain affected in aphasia include those that have primary responsibility for processing language content (words and meaning) and form (rules for combining linguistic units). The sudden and chronic difficulty with language can severely restrict the individual’s ability to participate in a wide range of personal, interpersonal and vocational activities. Within any individual, the severity of language impairment will change over time. The acute stage is considered the first hours or days immediately following the stroke or incurred brain damage, when behavioral symptoms are most severe. In some cases this acute stage is also marked by rapid functional recovery. Once the individual’s neurological condition has stabilized, typically within hours or a few days of acute symptom onset, the brain’s natural recovery process begins. The defining characteristic of this post-acute stage is spontaneous neurological and functional recovery, generally lasting up to six months following the initial brain insult. During spontaneous recovery, brain swelling may be reduced, cells that were injured but not permanently damaged may recover and there may be some reorganization of neural functioning. Following this post-acute phase of spontaneous recovery, the individual enters the chronic life-long stage of aphasia or persisting language impairment. Continued attention to communication is an essential component to enhancing the quality of life for the person with chronic aphasia. Those with chronic aphasia vary markedly in the degree to which they are independent or need assistance with various aspects of daily living. In addition to the nature, amount and location of neurological damage, a number of other patient factors may influence long-term outcome in those with aphasia. Patient factors may include age, mental and physical health status, the interrelated factors of socio-economic, educational and occupational levels, as

Generalizat ion in Bilingual Aphasia Treatment

91

well as individual preferences, motivations, environmental resources and access to effective treatments. Patients with depression have worse recovery and greater cognitive impairment than peers without depression who have comparable brain damage. Post-stroke depression is prevalent, affecting an estimated 30–60% of the population with aphasia (van de Weg et al., 1999). Successful management of depression symptoms can significantly enhance long-term outcomes for individuals with aphasia (Small, 2004). There is also some evidence indicating that aphasia symptoms may be more severe for individuals with lower educational or occupational levels (e.g. Connor et al., 2001). The reasons for differences in aphasia outcomes related to income and educational levels are not clear. Differences may be attributed to a combination of factors including pre-morbid levels of language and literacy (with higher levels of both associated with richer neural activity) as well as overall health status and access to adequate health care. Factors external to the individual that affect long-term outcome in aphasia are the availability of other resources, including family support and access to quality treatment. Treatment is recommended to increase functional outcomes in aphasia. Treatment is defined here as planned, systematic actions implemented by a skilled, knowledgeable professional designed to improve language and communication outcomes in individuals with aphasia. In turn, improved speech and language functioning by the individual with aphasia increases his or her ability to participate in meaningful life activities, thereby improving life quality. Studies in the monolingual aphasia literature have documented positive effects on language, communicative functioning and life participation following many different types of behavioral interventions. Treatment effectiveness is typically determined by documenting change in client performance on specific tasks following participation in a training program. Learning or ‘acquisition’ may be demonstrated by testing the client on behaviors, tasks or stimuli that were directly trained. A more rigorous measure of change is to demonstrate the generalization of treated behaviors to conditions, tasks or stimuli that were not directly included in the treatment setting (Davis, 2000). A wide variety of behavioral treatments have been shown to benefit monolingual individuals with aphasia (e.g. Hinckley & Carr, 2005; Meinzer et al., 2005; Thompson et al., 2003; see Raymer et al., 2008, for review). At the same time, no single treatment approach has proven effective for all individuals with aphasia. Robey (1998) used meta-analysis to investigate treatment efficacy for monolingual adults with aphasia. A total of 55 studies that employed various types of behavioral interventions with monolingual English-speaking adults were included in this meta-analysis. The number of individuals with aphasia participating in each study ranged from four to 92. Study effects were analyzed according to patients’ recovery stage – acute, post-acute and chronic. The overwhelming result of this meta-analysis was that outcomes for individuals who received behavioral

92

Par t 2: A ssessment and Treatment

treatment were superior to those of individuals who did not receive treatment, at all stages of recovery. There is still too little information available to determine if some types of behavioral interventions are more effective than others (Robey, 1998). In summary, a wide variety of language-based treatments have been effective and resulted in generalization to other skills not directly trained in monolingual individuals with chronic aphasia. For bilingual individuals with aphasia, an additional issue to consider is generalization or transfer across languages, from a treated to an untreated language.

Generalization from a Treated to an Untreated Language in Bilingual Aphasia Evidence of improvement in an untreated language as well as the treated language would present an effective and efficient method to facilitate positive communication outcomes in bilingual aphasia. In a recent systematic review, Kohnert (2009) considered investigations of cross-language transfer or generalization in bilingual or multilingual individuals with primary acquired aphasia. Studies included in the review met the following criteria: (1) inclusion of one or more bilingual or multilingual individuals with primary aphasia; (2) a discussion of abilities in each language prior to and following treatment; and (3) original data reported in a peer-reviewed publication in English. An electronic search of Medline and PyschINFO databases followed by a hand search of articles yielded 12 publications meeting the search criteria. Two additional case studies have since been published which meet the original systematic search criteria. Of the 14 studies investigating cross-language generalization following treatment in bilingual aphasia, 12 are single subject or case studies (with a combined total of 14 participants). There were two group studies with a total of 70 participants (Fredman, 1975; Junqué et al., 1989). Results from these combined studies have been equivocal: three studies found no evidence of cross-language generalization while the remaining 11 report some evidence of improvement in an untreated language on some tasks, under certain conditions such as reading activities exploiting strategies common to the client’s two languages (Laganaro & Overton-Venet, 2001), semantic naming treatment administered in the client’s relatively weaker language (Edmonds & Kiran, 2006), or naming activities highlighting form and meaning overlap in the client’s two languages (Kohnert, 2004).

Post-acute stage treatment investigations Half of these 14 reports, including the two large group studies, investigated treatment effects in participants who were a few weeks to a few months

Generalizat ion in Bilingual Aphasia Treatment

93

post-aphasia onset, with no control group. In these cases it is not possible to separate cross-language generalization as a consequence of treatment from spontaneous recovery. At the same time, functional outcomes for monolingual individuals with aphasia who receive intense treatment during the three months immediately following brain injury are significantly greater than for those patients who do not receive behavioral treatment at this stage of recovery (Robey, 1998). With a single exception, studies investigating the generalization of treatment effects from a treated to an untreated language in the post-acute stage of bilingual aphasia have been positive. That is, treatment studies undertaken during the spontaneous recovery period indicate that there is improvement in both languages of bilingual individuals with aphasia following intensive treatment delivered only in one language. Although these studies showed improvement on pre- and post-treatment measures in both languages, gains were generally greater in the treated language, perhaps providing some indication of the therapeutic effect, even in the early stages of aphasia (cf. Robey, 1998). One recent single case study, however, found that gains were restricted to the treated language (Abutalebi et al., 2009). The client was a native Spanish speaker who had immigrated to Italy at age 16, attaining high levels of proficiency in both Spanish and Italian. At the age of 56 he suffered a hemorrhagic stroke in the left hemisphere, resulting in aphasia. At aphasia onset and prior to treatment, residual abilities were better in Spanish than Italian. Intensive treatment focusing on lexical skills was administered in Italian over a fourmonth period. Following treatment, results of both behavioral and neuroimaging measures showed significant improvement in Italian, the treated language, but not Spanish. The authors assert that there was a decline in the untreated language, although behavioral evidence also supports an interpretation of stabilization as opposed to regression in Spanish. On 14 of 22 subtests of the Bilingual Aphasia Test (BAT, Paradis, 1986) there was no change in Spanish raw scores across the four-month treatment period, with performance remaining at or near ceiling. There was improvement on one Spanish subtest and slight declines (mostly one or two raw points) on seven Spanish subtests. Raw subtest scores cannot be interpreted independently, particularly when the difference is minimal and could easily be explained by inter-rater reliability or normal variation in performance by clinical populations. What is clear is that there was a steep up-tick in performance in Italian, the treated language. In this case, Italian was not only the language of treatment but also the language used at home and in the community.

Chronic aphasia treatment investigations The literature includes seven studies with bilingual participants in the chronic stage of aphasia, at least six months post-aphasia onset and presumably

94

Par t 2: A ssessment and Treatment

beyond the period of rapid spontaneous neurological recovery and associated functional improvement. These studies reported on a total of nine bilingual or multilingual adults with chronic aphasia (see Kohnert, 2009, for review). Of these, there was evidence of qualified cross-language generalization – with the untreated language showing benefit under some conditions or on some tasks, but not others. For example, Edmonds and Kiran (2006) found that there was both within and across-language generalization following semantic-naming treatment in a participant who, prior to aphasia, had comparable proficiency levels in Spanish and English. For two other participants whose proficiency and use of English was much greater than Spanish prior to aphasia, these researchers found that cross-language generalization was present from the weaker language (English) to the stronger language (Spanish), but not the reverse (Edmonds & Kiran, 2006). In a recent study, Miertsch et al. (2009) investigated whether lexical and semantic training in one language could generalize to other languages in a multilingual person with chronic aphasia who was a native speaker of German with English and French as second and third languages, respectively. At the time of the investigation the client was more than two years post-aphasia onset; all previous treatment had been in German. Twice daily language training was provided in French for three weeks with a focus on word production and comprehension using picture cards. Following treatment there was improvement in lexical and syntactic measures in French (the treated language) as well as English (the untreated language), as measured by the BAT (Paradis, 1986). In contrast to Edmonds and Kiran (2006), there was no improvement to the strongest language (German), perhaps because the client’s pre-test score in German was already within the norm for the BAT. Gains seen in English when French was the language of treatment suggest qualified support for generalization from a treated language to an untreated language. The authors suggest generalization of treatment from French to English was facilitated by targeting a common semanticconceptual system. Laganaro and Overton-Venet (2001) investigated the effects of two different computer-based reading treatments with a 50-year-old SpanishEnglish bilingual who had chronic aphasia resulting from a gunshot wound. An AB-AB design was implemented, first in English and then Spanish, each lasting two weeks for a combined eight weeks of treatment. The first treatment facilitated whole-word reading using lexical decision, categorization and word association tasks; the second treatment was designed to facilitate phonological encoding through sound-to-print associations using letter and syllable assembly, homophony and rhyme judgment tasks on words and non-words. For tasks that relied on phonological encoding, including reading time, gains were specific to the treated language. In contrast, on the lexical decision measure both treated and untreated languages benefited from the computer-based intervention. Results supported the notion that

Generalizat ion in Bilingual Aphasia Treatment

95

cross-language transfer is possible when common processes are targeted but not when language-specific representations are the treatment focus. Kohnert (2004, Study 2) investigated the role of structural similarity on cross-language generalization during naming treatment with a SpanishEnglish bilingual with chronic aphasia. The study goal was to determine if there would be generalization from trained to untrained items for cognate and non-cognate word types within and across languages. Cognates are translation equivalents that also share phonological or orthographic form (e.g. teléfono-telephone); non-cognates are translation equivalents with little or no overlap in form (rey-king). Treatment in Spanish (the treated language) resulted in improved performance on trained as well as untrained cognates and non-cognates. However, cross-language generalization to untreated stimuli was evident only for cognates. Two investigations of cross-language generalization following treatment in bilingual individuals with chronic aphasia revealed no evidence of improvement in the untreated language. In both cases, improved performance was restricted to the treated language (Galvez & Hinckley, 2003; Meinzer et al., 2007). As with many of the other studies reviewed, these investigations focused on some aspect of word naming, perhaps because anomia is pervasive in all types and severities of aphasia. Divergent findings may be attributed in large part to differences in methodology, including the nature, focus and timing of treatment. However, it also seems that, collectively, these studies illustrate the many factors affecting bilingual aphasia outcomes. Studies point to interactions between client, language and environmental factors – including treatment – that may facilitate or constrain cross-language generalization. Client factors noted were the number of languages spoken and the order in which they were learned, tempered by patterns of daily use and attained proficiency levels in each language prior to aphasia onset as well as client motivation for training in a particular language. Language factors include the aspect of language trained and measured (lexical-semantics, grammar, reading) as well as the structural similarities in the bilingual’s two languages. Treatment factors include strategies, stimuli and general procedures used in training. Clearly, factors found to affect long-term functional outcomes in monolinguals noted in the previous section hold true for bilingual individuals with aphasia. Future investigations are needed to more precisely define the parameters of generalization from one language to another, both in terms of possibilities and limitations. In the meantime, results from combined studies on crosslinguistic generalization in bilingual aphasia can be used to inform clinical interactions in at least two important ways. First, by providing direct treatment in only one of a proficient bilingual’s or a multilingual’s two or more languages, we do not seem to be harming the other language(s). This is important as many clinicians do not speak all of the languages of their clients and, at the very least, need to be reassured that they are not hampering

96

Par t 2: A ssessment and Treatment

bilingual recovery when providing direct treatment in only a single language. That is, across studies, treatment in one language resulted in either positive or neutral outcomes for the untreated language. On the other hand, complete inattention to one language in the clinical setting combined with a recommendation to avoid using it outside the clinic may result in a long-term plateau, with persisting impairment greater than perhaps might be expected if this language was considered at some earlier point in post-aphasia planning (Meinzer et al., 2007). Second, the potential for facilitating improvement in two languages, with a single set of treatment procedures, may be increased when training is directed at areas of overlap. This overlap may be structural, for those languages that share common roots, or computational, for cognitive processes that support language more broadly. Also, cognitive inefficiencies or weaknesses may exist alongside the more easily recognized linguistic difficulties in bilinguals with aphasia and should be considered in treatment. Next, we discuss the clinical relevance of cognitive weaknesses in aphasia.

Cognitive Inefficiencies and Cross-Domain Associations in Aphasia By definition, aphasia encompasses a disproportionate interruption to the language system. It is not the case however, that language is the sole domain of impairment. Acquired inefficiencies in general cognitive functioning may affect the processing of non-linguistic as well as linguistic information. Chapey’s (1981: 31) definition of aphasia as ‘An acquired impairment in language and the cognitive processes which underlie language caused by organic damage to the brain’ underscores the dual deficit of language and cognition. In some cases cognitive weaknesses may be obvious, with symptoms including impaired orientation or memory, significant fatigue, or difficulty in sustaining attention during simple activities. More often, impairment in the general cognitive processing system is subtle and overlooked in aphasia in light of the much more observable language deficits (see also Chapey, 2001; Luria, 1966; McNeil & Pratt, 2001). A range of empirical findings in the monolingual literature reinforces the notion of aphasia as a primary, but not exclusive, language deficit. For example, Saygin and colleagues found that participants with chronic aphasia were much slower and less accurate in processing non-linguistic auditory and visual information as well as language stimuli, compared to both normal age-matched controls and participants with right hemisphere strokes (Saygin et al., 2003, 2004). Furthermore, investigations found a robust positive association between performance on language and non-linguistic tasks, suggesting that the two domains draw on common neural regions and or/cognitive processes. Helm-Estabrooks (2002) found that 11 of 13 participants with

Generalizat ion in Bilingual Aphasia Treatment

97

chronic mild to moderate aphasia performed in the impaired range on nonlinguistic tasks included in the Cognitive Linguistic Quick Test (CLQT; Helm-Estabrooks, 2001). Basic cognitive mechanisms may play an additional role in bilingual language processing. Proficient, neurologically intact bilinguals can use their languages one at a time or combine them as when code-switching or interpreting across languages. Each of these speaking conditions places different demands on the cognitive-communicative system. Even in monolingual speaking circumstances, there is always some demand on the cognitive system to prevent unintended cross-talk between the bilingual’s two languages. The presence of unintentional cross-talk would indicate some compromise to the general cognitive control system. A frequently documented phenomenon in bilingual and multilingual aphasia is cross-language intrusions or unintentional language mixing, also referred to as ‘pathological code switching’. Unintentional language mixing in aphasia appears random, is often viewed as undesirable, unintended and unstoppable by the speaker and therefore can be a source of considerable frustration. Unintended intrusions from the other language observed in bilingual individuals with aphasia provide clear evidence of some weakness in cognitive control (Abutalebi et al., 2009; Fabbro et al., 2000; Green, 1998; Green & Abutalebi, 2008; Kohnert, 2004). Recognition of cognitive weaknesses as part of the aphasia profile leaves open the possibility that, at least in some cases, treatment which strengthens underlying cognitive processing mechanisms will lead to gains in overall language ability, or provide a foundation for subsequent treatment directed at language. For example, Coelho (2005) reported that cognitive training (focusing on attention) improved reading outcomes in an English-speaking woman with chronic aphasia. Van Mourik and colleagues (1992) found that 13 of 17 monolingual individuals with severe to profound aphasia had deficits in non-verbal cognition such as impaired concentration or reduced visual processing skills. These clinical researchers recommended that clients with aphasia who have cognitive inefficiencies receive treatment focused directly at cognitive processing deficits as a prerequisite to language treatment. In a systematic single case treatment study, Kohnert (2004, Study 1) investigated the possibility of cross-domain generalization in a SpanishEnglish bilingual individual with chronic aphasia. Twelve sessions of cognitive treatment over a two-month period were conducted to determine if: (a) basic cognitive processing abilities were malleable to training, and (b) improved performance on skills in both languages would be observed following training. Specific training activities included card-sorting tasks to target perception and categorization skills; written single-digit math computations, visual searches to facilitate sustained and alternating attention, as well as several visual perception and speeded response tasks from commercially available neuro-rehabilitation software (Scarry-Larkin, 1994; Scarry-Larkin

98

Par t 2: A ssessment and Treatment

& Price, 2004). As this cognitive-based treatment was designed to emphasize non-linguistic information processing, no language training was provided. Following this non-linguistic cognitive treatment, performance on all trained non-verbal cognitive processing tasks improved in speed and/or accuracy. These results demonstrated malleability of cognitive processing skills and, consistent with reports in the monolingual aphasia literature, indicate that reinforcing basic information processing skills may be an important initial step in effectively treating the more obvious language deficits (cf. Van Mourik et al., 1992). Importantly, there were also improvements in both Spanish and English following this cognitive-based training. In Spanish, improvement was observed on four of five language measures administered immediately before and after treatment. In English, there were gains on all five language measures, providing at least some ‘proof of concept’ with respect to cross-domain generalization. Clearly this is an area in which additional studies are needed to determine when and how connections between cognition and language may be used to facilitate functional gains in both languages of bilinguals with aphasia. For present purposes cognitive-based treatments may provide a potential entry point for increasing cognitive as well as language abilities in bilingual clients with aphasia.

Holistic Models of Intervention and Bilingual Treatment Strategies Bilinguals with aphasia are distinguished from monolingual individuals with aphasia by their prior experience and proficiency in more than one language. The question this distinction most often raises for planning clinical actions with bilingual clients is: which language should be supported in treatment? In the mid-twentieth century a single language approach was advocated in bilingual aphasia (Wald, 1958 as cited in Paradis, 1983). It was reasoned, in the absence of direct empirical evidence, that attention to two languages would confuse the client or delay functional recovery in all languages. Although this question continues to be asked (e.g. Fabbro, 2001; Lorenzen & Murray, 2008) many bilingual clinical experts suggest that it be replaced with ‘How can we best support both/all languages in bilinguals with aphasia to increase meaningful participation in their diverse environments?’ (cf. Ansaldo et al., 2008; Ansaldo & Marcotte, 2007; Kohnert, 2008). This reframed question (from which language? to how can we support what’s needed?) represents a fundamental paradigm shift that is consistent with holistic models of intervention. Holistic intervention models require consideration of individuals’ language and communicative functioning with respect to environmental needs, past, present and future. When treatment improves the comprehension or production of specific aspects of one language, at the

Generalizat ion in Bilingual Aphasia Treatment

99

same time it isolates people from their communicative environments, the value of treatment is negated. For bilinguals, when treatment is designed to promote one language and systematically exclude the other, clients may become further isolated from people and activities for which the excluded language is needed. Despite limitations in the ability to access two different languages due to aphasia, typically there continues to be a robust need for these languages in the communicative environment. The systematic disregard of the individual’s life circumstances in treatment violates fundamental principles of holistic models of disability, including those advocated by the Life Participation Approach to Aphasia (LPAA Project Group, 2001) and the World Health Organization (2001). It is also true that bilingual individuals with aphasia can functionally recover two languages and should be supported in their efforts to do so. In addition, there is evidence that personal memories are encoded and recalled with greater emotional intensity in the language in which original events were lived (e.g. Schrauf & Rubin, 2000). The implication is that, by excluding a language when planning intervention, we may be unintentionally limiting access to past experiences for bilinguals with aphasia. Moreover, cross-language and cross-domain associations within the bilingual mind/ brain may serve as a springboard for effective and efficient aphasia treatment. At the same time, the empirical evidence on cross-language generalization in bilingual aphasia indicates that improvement in an untreated language may not spontaneously occur, independent of careful planning, at least in chronic aphasia. Even for bilinguals in the post-acute stage of aphasia, clinical actions which are mindful of the fundamental ‘bilingual’ nature of the individual may increase positive long-term outcomes in both languages. Consistent with holistic models of intervention, our perspective is that clinicians who work with bilingual individuals with aphasia have a professional mandate to facilitate communication in both or all languages needed for participation in meaningful life activities. Clinicians who do not speak both of their clients’ languages, as is often the case, may, through careful planning, implementation and collaboration, support the recovery and use of a language they do not speak (Kohnert, 2008). As part of a larger intervention plan, clinicians may employ ‘bilingual treatment strategies’ to promote cross-language or cross-domain generalization. Bilingual treatment strategies may be directed at the common cognitive or conceptual substrates of language, the cognitive-linguistic interface, or the structural overlap in the bilinguals’ two languages.

Cognitive or conceptual strategies As noted earlier, even though aphasia presents as a primary impairment in language, basic cognitive processes may also be damaged. (Refer to section Cognitive Inefficiencies and Cross-domain Associations in Aphasia.) In some cases,

100

Par t 2: A ssessment and Treatment

clinical researchers have recommended directly targeting these basic cognitive underpinnings as a prelude to direct language treatment (e.g. Van Mourik et al., 1992). Activities directed at shoring up the efficiency of basic attention, perception, categorization or problem-solving systems may, in turn, provide a foundation for greater subsequent gains in both languages of the bilingual individual with aphasia (cf. Kohnert, 2004). For example, environmental sounds may be employed in clinician-directed activities to increase attention to slight changes in signal form; visual searches or complex figure-ground pictures reminiscent of ‘I spy’ activities may be used to promote both visual attention and perception. Arithmetic problems, adapted to the client’s level, may be used to promote accuracy and efficiency in problem-solving. There are also several software and online resources available with many different activities to promote high levels of cognitive processing in adults (e.g. Lumosity, www.lumosity.com; Scarry-Larkin & Price, 2004). Activities using computer interface may be used initially in clinician-directed training and, eventually, for independent practice. For individuals with severe expressive aphasia, common conceptual underpinnings may be used to develop a single picture-based communication board. This board may be used to facilitate interactions with speakers of different languages (e.g. spouse and other family members, clinician, medical personnel).

Cognitive-linguistic strategies A possible strategy to increase cognitive control of the language system is structured translation activities – essentially using translation as a therapeutic strategy. For bilinguals with aphasia, translation tasks are a standard part of assessment (e.g. BAT, Paradis, 1986). This assessment, as well as information gained regarding unintentional cross-language intrusions during spontaneous speech, may be used to guide treatment activities at the cognitive-linguistic interface. Translation, moving meaning from one linguistic code into another, places demands on the cognitive system in addition to the inherent language demands. Training performance on this more complex task may have positive cascading effects on simpler tasks (cf. Kiran & Thompson, 2003; Thompson et al., 2003). Clinicians can develop a hierarchy of translation tasks. Stimuli may vary from high frequency simple object words to complex sentences. Forward and backward translation activities within or across modalities using live or computer interface may be used. Within the lexical level, translation tasks may be used to highlight any conceptual overlap between language systems to promote cross-language generalization. Picture dictionaries, online bilingual dictionaries, commercially available language software programs, and bilingual family members may be used to develop stimuli. Treatment procedures may also be directed at shaping and expanding the communicative intent of unintentional language switching. When

Generalizat ion in Bilingual Aphasia Treatment

101

other-language intrusions occur, the clinician may help the client use a form of inter-language circumlocution to facilitate lexical access in the other language, drawing on the common conceptual store. Circumlocution is a common therapeutic technique often used by clinicians to facilitate lexical access in monolingual individuals with anomia. The clinician may also counsel the individual and his or her family to increase understanding and reduce frustration associated with unintentional language switches. By shifting focus from the form of other language intrusions to their meaning, crosslanguage intrusions may continue to move the communicative interaction forward, versus bringing it to a halt (see Ansaldo & Marcotte, 2007, for discussion).

Structurally-based language strategies Another general strategic category to promote simultaneous improvement in both languages of bilinguals with aphasia is to exploit the structural overlap in the two languages. Laganaro and Overton-Venet (2001) found cross-linguistic generalization in a gentleman with bilingual aphasia when reading processes fundamental to both Spanish and English were trained. For bilingual individuals with significant difficulty in lexical-semantic retrieval, the clinician may select cross-linguistic cognates as training stimuli to ‘jump start’ the process (Kohnert, 2004). In other cases, attention to common aspects of sounds or words may be used. For example, Spanish and English share 20 different phonemes as well as noun-final morphological inflections. Structural strategies are possible when the bilingual’s two languages are historically linked. For very high functioning clients who also had high levels of pre-morbid proficiency in two languages, contrastive structural analysis may be used as a strategy to further facilitate the recovery of cross-linguistic links. Of course, languages also differ in important ways, and bilinguals use their languages for different purposes. Therefore, specific attention to each language will likely be needed to complement bilingual strategies. Commercially available software in languages other than those spoken by the clinician can be modified and used by the client to develop parallel home and clinic programs. As with other strategies described here, even when direct treatment is implemented in primarily one language, that spoken by the clinician, the bilingual nature of the client can be systematically considered in planning. In summary, the general goal of bilingual treatment strategies is to promote gains in functional communication skills in both languages of the bilingual client in the most efficient and effective way possible. This is done by structuring activities that highlight connections between basic cognitive processing skills that are needed for efficient language processing, by selecting intervention activities and stimuli that reinforce the links or overlap

102

Par t 2: A ssessment and Treatment

between the client’s two languages or some combination of these measures. The strategies introduced here are representative only. Critically, we recommend that bilingual strategies intended to promote cross-language or crossdomain generalization be included as part of a comprehensive intervention program. A comprehensive intervention program includes both direct and indirect treatment approaches, directed at language, cognitive, communicative and/or environmental areas of need, based on each client’s unique profile and social circumstances (Kohnert, 2008).

Conclusion and Future Directions The study of treatment efficacy for bilingual individuals with aphasia is very much in its infancy. Systematic investigation of cross-language and crossdomain relationships in bilinguals with aphasia is needed to advance clinical work and to refine our basic understanding of mediating factors on functional recovery following acquired brain injury. Available evidence suggests that, in some cases, it may be possible to structure treatment exercises to emphasize links between the individual’s two different languages. In this case, the goal is to provide training in one language (e.g. English) that will produce improvement in the treated as well as untreated language (e.g. English as well as Spanish). At the same time, results suggest that there may be a number of constraints on the generalization of intervention gains across languages. One of the most important factors may be the type of treatment employed. At present there is no definitive response to the question, ‘Does an untreated language benefit from treatment administered in a different language in bilinguals with aphasia?’ This is not surprising as there are many unknowns regarding factors affecting treatment outcomes even in the monolingual aphasia literature (e.g. Raymer et al., 2008). Based on evidence to date, it is likely that a revised set of more sophisticated questions will be needed to guide research. These questions may include: What specific treatment strategies are most successful in promoting benefits in both languages of bilingual speakers and why? Are there certain aspects of language that must be trained separately (using language-specific approaches) and others that are amendable to general language procedures? Is cross-language generalization following treatment differentially affected by aphasia severity or time post-onset? Based on the available evidence in both monolingual and bilingual aphasia, we discussed potential ways to facilitate improvement in two languages, with a single set of treatment procedures. We placed these bilingual treatment strategies within the context of holistic intervention models. Suggested strategies were directed at different areas of overlap in the dual-language cognitive-linguistic system of bilinguals with aphasia. Because there is no

Generalizat ion in Bilingual Aphasia Treatment

103

literature which compares the effects of different treatment strategies on outcomes in bilingual aphasia, the recommendations here are intended to be more provocative than prescriptive. Research is needed to validate or refute these recommendations. At the same time, recommended bilingual treatment strategies are consistent with available evidence in monolingual and bilingual aphasia, the much more robust evidence on non-selective access in neurologically intact bilingual adults, as well as highly interactive theories of language in monolingual and bilingual individuals.

References Abutalebi, J., Della Rosa, P.A., Tettamanti, M., Green, D.W. and Cappa, S.F. (2009) Bilingual aphasia and language control: A follow-up fMRI and intrinsic connectivity study. Brain and Language 109, 141–156. Ansaldo, A.I. and Marcotte, K. (2007) Language switching and mixing in the context of bilingual aphasia. In J. Centeno, L. Obler and R. Anderson (eds) Studying Communication Disorders in Spanish Speakers: Theoretical, Research, and Clinical Aspects (pp. 214–230). Clevedon: Multilingual Matters. Ansaldo, A.I., Marcotte, K., Scherer, L. and Raboyeau, G. (2008) Language therapy and bilingual aphasia: Clinical implications of psycholinguistic and neuroimaging research. Journal of Neurolinguistics 21, 539–557. Chapey, R. (1981) The assessment of language disorders in adults. In R. Chapey (ed.) Language Intervention Strategies in Adult Aphasia (pp. 31–84). Baltimore, MD: Williams & Wilkins. Chapey, R. (2001) Language Intervention Strategies in Aphasia and Related Neurogenic Communication Disorders (4th ed.). Baltimore, MD: Williams & Wilkins. Coelho, C.A. (2005) Direct attention training as a treatment for reading impairment in mild aphasia. Aphasiology 19, 275–283. Connor, L.T., Obler, L., Tocco, M., Fitzpatrick, P. and Martin, A. (2001) Effect of socioeconomic status on aphasia severity and recovery. Brain and Language 78, 254–257. Davis, G.A. (2000) Aphasiology: Disorders and Clinical Practice. Needham Heights, MA: Allyn and Bacon. Edmonds, L.A. and Kiran, S. (2006) Effect of semantic naming treatment on crosslinguistic generalization in bilingual aphasia. Journal of Speech, Language, and Hearing Research 49, 729–748. Fabbro, F. (2001) The bilingual brain: Cerebral representation of languages. Brain and Language 79, 211–222. Fabbro, F., Skrap, M. and Aglioti, S. (2000) Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery, & Psychiatry 68, 650–652. Fredman, M. (1975) The effect of therapy given in Hebrew on the home language of the bilingual or polyglot adult aphasic in Israel. International Journal of Language & Communication Disorders 10, 61–69. Galvez, A. and Hinkley, J. (2003) Transfer patterns of naming treatment in a case of bilingual aphasia. Brain and Language 87, 173–174. Green, D.W. (1998) Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition 1, 67–81. Green, D.W. and Abutalebi, J. (2008) Understanding the link between bilingual aphasia and language control. Journal of Neurolinguistics 21, 558–576. Helm-Estabrooks, N. (2001) Cognitive Linguistic Quick Test. San Antonio, TX: Harcourt Assessment.

104

Par t 2: A ssessment and Treatment

Helm-Estabrooks, N. (2002) Cognition and aphasia: A discussion and a study. Journal of Communication Disorders 35, 171–186. Hernandez, A., Hiscock, M. and Bates, E.A. (2009) The development of neural substrates of language over the lifespan. In K. de Bot and R.W. Schrauf (eds) Language Development Over the Lifespan (pp. 288–308). New York: Routledge. Hinckley, J. and Carr, T. (2005) Comparing the outcomes of intensive and non-intensive context-based aphasia treatment. Aphasiology 19, 965–974. Junqué, C., Vendrell, P., Vendrell-Brucet, J.M. and Tobeña, A. (1989) Differential recovery in naming in bilingual aphasics. Brain and Language 36, 16–22. Kiran, S. and Thompson, C.K. (2003) The role of semantic complexity in treatment of naming deficits: Training semantic categories in fluent aphasia by controlling exemplar typicality. Journal of Speech, Language, and Hearing Research 46, 608–622. Kohnert, K. (2004) Cognitive and cognate treatments for bilingual aphasia: A case study. Brain and Language 91, 294–302. Kohnert, K. (2008) Language Disorders in Bilingual Children and Adults. San Diego, CA: Plural. Kohnert, K. (2009) Cross-language generalization following treatment in bilingual aphasia: A review. Seminars in Speech and Language 30, 174–186. Laganaro, M. and Overton-Venet, M. (2001) Acquired alexia in multilingual aphasia and computer-assisted treatment of both languages: Issues of generalization and transfer. Folia Phoniatrica et Logopaedica 53, 135–144. Lorenzen, B. and Murray, L.L. (2008) Bilingual aphasia: A theoretical and clinical review. American Journal of Speech-Language Pathology 17, 299–317. LPAA Project Group (2001) Life participation approach to aphasia. In R. Chapey (ed.) Language Intervention Strategies in Aphasia and related Neurogenic Communication Disorders (pp. 235–246). Philadelphia, PA: Lippincott, Williams & Wilkins. Luria, A.R. (1966) Higher Cortical Functions in Man. New York: Basic Books. McNeil, M.R. and Pratt, S.R. (2001) Defining aphasia: Some theoretical and clinical implications of operating from a formal definition. Aphasiology 15, 901–911. Meinzer, M., Djundja, D., Barthel, G., Elbert, T. and Rockstroh, B. (2005) Long-term stability of improved language functions in chronic aphasia after constraint-induced aphasia therapy. Stroke 36, 1462–1466. Meinzer, M., Obleser, J., Flaisch, T., Eulitz, C. and Rockstroh, B. (2007) Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI. Neuropsychologia 45, 1247–1256. Miertsch, B., Meisel, J. and Isel, F. (2009) Non-treated languages in aphasia therapy of polyglots benefit from improvement in the treated language. Journal of Neurolinguistics 22, 135–150. Paradis, M. (1983) Readings on Aphasia in Bilinguals and Polyglots. Montreal: Didier. Paradis, M. (1986) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum. Raymer, A., Beeson, P., Holland, A., Kendall, D., Maher, L., Martin, N., Murray, L., Rose, M., Thompson, C.K., Turkstra, L., Altmann, L., Boyle, M., Conway, T., Hula, W., Kearns, K., Rapp, B., Simmons-Mackie, N. and Gonzalez Rothi, L. (2008) Translational research in aphasia: From neuroscience to neurorehabilitation. Journal of Speech, Language, and Hearing Research 51, 259–275. Robey, R.R. (1998) A meta-analysis of clinical outcomes in the treatment of aphasia. Journal of Speech, Language, and Hearing Research 41, 172–187. Saygin, A.P., Dick, F., Wilson, S.W., Dronkers, N.F. and Bates, E. (2003) Shared neural resources for processing language and environmental sounds: Evidence from aphasia. Brain 126, 928–945. Saygin, A.P., Wilson, S.M., Dronkers, N.F. and Bates, E. (2004) Action comprehension in aphasia: Linguistic and non-linguistic deficits and their lesion correlates. Neuropsychologia 42, 1788–1804.

Generalizat ion in Bilingual Aphasia Treatment

105

Scarry-Larkin, M.K. (1994) Attention and Memory, Vol. I. San Luis Obispo, CA: LocuTour Multimedia. Scarry-Larkin, M.K. and Price, E.D. (2004) Attention and Memory, Vol. II. San Luis Obispo, CA: LocuTour Multimedia. Schrauf, R.W. and Rubin, D.C. (2000) Internal languages of retrieval: The bilingual encoding of memories for the personal past. Memory and Cognition 28, 616–623. Small, S. (2004) A biological model of aphasia rehabilitation: Pharmacological perspectives. Aphasiology 18, 473–492. Thompson, C.K., Shapiro, L., Kiran, S. and Sobecks, J. (2003) The role of syntactic complexity in treatment of sentence deficits in agrammatic aphasia: The complexity account of treatment efficacy (CATE). Journal of Speech, Language, and Hearing Research 42, 690–707. van de Weg, F.B., Kuik, D.J. and Lankhorst, G.J. (1999) Post-stroke depression and functional outcome: A cohort study investigating the influence of depression on functional recovery from stroke. Clinical Rehabilitation 13, 268–272. Van Mourik, M., Verschaeve, M., Boon, P., Paquin, P. and Van Harkskamp, F. (1992) Cognition in global aphasia: Indicators for therapy. Aphasiology 6, 491–499. World Health Organization (2001) International Classification of Functioning, Disability and Health (ICF). Geneva, Switzerland: WHO. Online at http://www3.who.int/icf/ icftemplate.cfm.

7

Cross-Language Treatment Effects in Multilingual Aphasia Mira Goral

This chapter presents data from aphasia treatment with bilingual and multilingual speakers (henceforth multilinguals) and addresses the question whether current neurolinguistic and psycholinguistic theories of multiple language representation and processing account for the findings observed. Following a summary of current theories of multilingualism, four studies of treatment with multilinguals with aphasia are briefly presented. The results of the studies and their theoretical and clinical implications are discussed.

Theories of Multiple Language Representation and Processing Neurolinguistic data obtained from neuroimaging and lesion studies with multilinguals have been taken as evidence for two contrasting hypotheses about the representation and processing of more than one language in the adult brain. According to one hypothesis, predominantly overlapping neural networks are assumed to be associated with the multiple languages of an individual, and any differential activation during processing of one language versus another is attributed to mechanisms of language control rather than to language processing per se (Abutalebi & Green 2007, 2008; Green, 1998). Accordingly, the regions and networks associated with these processes of selective language activation and inhibition differ from the classic brain regions that have been associated with language processing in monolingual individuals. The alternative hypothesis argues for at least partially nonoverlapping neural networks associated with the different languages of multilingual speakers. Differential representation has been associated with separate networks for the first language (L1) and the second language (L2) within the language-dominant hemisphere (e.g. Giussani et al., 2007; 106

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

107

Ojemann & Whitaker, 1978). Differential processing of the languages of multilinguals has also been associated with two distinct memory systems, procedural versus declarative. Such theories distinguish the automatic processing of L1, assumed to be processed by implicit memory mechanisms, from the non-automatic processing of later learned, non-L1 languages, hypothesized to be processed by explicit memory systems (e.g. Paradis, 1994, 2009; Ullman, 2001, 2005). Psycholinguistic evidence corroborates the important role that mechanisms of control play in how multilinguals access and process their languages. Specifically, findings from psycholinguistic studies with multilinguals suggest that both or all languages are active during language processing and that a multilingual speaker has to actively inhibit or de-select one language and activate or select the other(s). Evidence for language non-selective processing includes cognate effects, as in faster recognition of cognate words (translation equivalents that share phonological or orthographic form, e.g. coffee and café in English and Spanish, respectively) than non-cognate translation equivalents (e.g. Dijkstra, 2005), and slower processing of ‘false friends’ (words that look like cognates but have different meanings in the two languages, e.g. coin in English and in French, where the word means ‘corner’) than words that differ in form and meaning (e.g. Dijkstra et al., 2005; Dijkstra et al., 1998). Furthermore, there are data suggesting that cross-language alternatives (cognate as well as non-cognate translation equivalents) get activated and compete for selection during language production tasks, such as picture naming, and that the target word is eventually selected as the non-target alternatives are inhibited (e.g. Costa et al., 1999; Kroll et al., 2008). Consistently, support for the need for language selection and inhibition can be found in studies that document a language-switching cost. This phenomenon demonstrates that when multilinguals need to produce words in two languages within the same task, performance on trials that appear following a language switch is affected by the switch. For example, longer response times in naming pictures of objects are seen for items that appear after a cue to switch the target language was provided than for items named in the same language of the preceding item (e.g. Kroll et al., 2006; Meuter & Allport, 1999). Moreover, the slowed response times in the switch conditions have demonstrated an interesting relation to relative language proficiency. Namely, it has been shown that when dominant bilinguals name objects in their L1, dominant language, their response times are slower following an item named in their weaker L2 than after a no switch trial. In contrast, no or small differences in response times for switched versus no switch items are evident when the dominant bilingual participants are naming in their weaker L2. Moreover, balanced bilinguals with high proficiency in both L1 and L2 do not show this directional asymmetry of the switching cost (e.g. Costa & Santesteban, 2004). These findings have been extended also to speakers of more than two languages, such as an asymmetric switching cost

108

Par t 2: A ssessment and Treatment

between a strong L1 and a weak L3, but with the exception of bilinguals switching between a strong L2 and a weak L3 (Costa et al., 2006). Whereas further study is needed to better understand the relations between proficiency and asymmetry of the switching cost, the majority of the findings to date can be explained by the notion that bilinguals who are using their weaker language actively inhibit their strong language. The scope and time course of this inhibition is fertile ground for further investigation.

Studies of Cross-Language Effects in Aphasia Treatment The neurolinguistic and psycholinguistic theories of multilingual language representation and processing reviewed above can inform our predictions regarding cross-language treatment generalization in multilingual aphasia. For example, the postulation of overlapping neuronal representation of multiple languages is consistent with the prediction that treating one language of multilingual speakers with aphasia should affect their other languages. Specifically, co-activation of multiple languages during processing would predict that treating one language should facilitate processing the others. Findings from treatment studies in multilinguals with aphasia published to date have yielded equivocal, confusing results (see Kohnert & Peterson and Obler & Park, Chapters 6 and 1, respectively, of this volume). For example, cross-language facilitation has been reported in a number of studies (e.g. Faroqi & Chengappa, 1996; Kohnert, 2004); a lack of facilitation has been found in other studies (e.g. Galvez & Hinckley, 2003; Meinzer et al., 2007); and mixed results, possibly depending on the direction of generalization examined and on a variety of additional variables (e.g. relative aphasia severity, age of language learning, relations among the languages) have also been reported (e.g. Edmonds & Kiran, 2006; Goral et al., 2010; Miertsch et al., 2009). An examination of the studies in which positive cross-language generalization is reported suggests that patterns of proficiency related facilitation and inhibition can account for at least some of the results obtained. For example, cross-language generalization was found for individuals with comparable levels of proficiency (the first participant reported in Edmonds & Kiran, 2006; Miertsch et al., 2009); and no generalization was found from treatment in English, a participant’s L2, to the untreated French, his first and more dominant language (Miller Amberber, in press). Moreover, the assumption that multilinguals inhibit their stronger language while processing weaker ones would lead to the prediction that under such circumstances, negative cross-language effects should be observed. Indeed, negative effects were reported in Abutalebi et al. (2009) in the participant’s L1 Spanish, following treatment in Italian, his L2 and weaker post-CVA (cerebrovascular accident) language.

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

109

The contribution of the relative proficiency of the treated and untreated languages to cross-language treatment effects is the focus of this chapter. In what follows, four single-subject studies of multilingual individuals with aphasia who received treatment in one or more of their languages are summarized. The data are examined in the context of the predictions outlined above, of positive cross-language effects between languages of comparable proficiency as well as from treatment in a stronger language to a weaker one, and of negative cross-language effects from treatment in a weaker language to a stronger untreated language. The first three cases received treatment that followed the principles of constraint-induced aphasia treatment (CIAT), in that verbal language production was encouraged in the context of an intense treatment schedule that focused on exchange of new information. In these three cases, treatment targeted sentence production. In the treatment administered with the fourth case, word retrieval was targeted in addition to sentence production, and the treatment did not specifically follow the principle of CIAT. In all four cases, the focus of the analysis presented here is morphosyntactic accuracy in connected language production (other aspects of performance following treatment are reported elsewhere) and evidence for facilitatory and inhibitory cross-language effects on sentence grammaticality was examined. The pattern that emerged supports the hypothesis that working in one language can benefit a non-treated language of comparable or lower current proficiency levels, but when the language of treatment is of lower proficiency than the untreated one, evidence for global and local inhibition of a stronger untreated language is found.

Four Single-Subject Treatment Studies in Multilingual Aphasia Case 1 History A man in his 40s enrolled in a treatment study seven years post a large left CVA infarct that extended to frontal, temporal and parietal areas and the basal ganglia. Hebrew, acquired from birth, was his dominant language in his youth. English was in his environment in the first years of life but was formally learned in childhood at school. He was later immersed in English during the years he lived in the US and it remained the language of his environment at the time of his stroke. French, the third language, was learned formally in late childhood, achieving high proficiency with immersion during the years the participant lived in French-speaking countries, and has continued to be the primary home language. The participant reported high proficiency in all three languages pre-CVA. Following the stroke, comparable impairment was observed in all three languages but he experienced the most

110

Par t 2: A ssessment and Treatment

difficulty with French and the least difficulty with Hebrew. He had received therapy in English in the years following his stroke and has been participating in aphasia conversation group sessions in English. He did not receive structured therapy in Hebrew and French. At the time of the study he presented with mild non-fluent aphasia with morphosyntactic and lexicalretrieval difficulties in sentence and discourse production, slow speech, hesitations and self-corrections. There was no marked apraxia or dysarthria, and no marked comprehension and cognitive impairment. Additional details about the participant’s language performance can be found in our previous studies with this participant (Altman et al., submitted; Goral et al., 2006; Goral et al., 2010).

Design The participant received two sequential treatment phases, both in English. Each treatment was administered for three weeks, three hours per week. Testing in all three languages was administered prior to and following each treatment. Three testing sessions were conducted at each testing time to assure stability and allow for calculation of change.

Treatment Treatment adhered to the principles of a modified constraint-induced therapy with emphasis on verbal production and exchange of new information (see Kempler & Goral, 2011, generative treatment). The first treatment focused on morphosyntactic accuracy of sentence production; the second on word-retrieval strategies. Both followed a loosely structured protocol.

Measures Data from picture-elicited sentences (reported in Goral et al., 2010) and personal narratives (Altman et al., submitted) were examined. In the sentence production task, the participant was presented with line drawings from the Sentence Production Program for Aphasia (Helm-Estabrooks & Nicholas, 2000) and was instructed to generate two sentences to describe what the people in the picture may be saying. In the narrative task, the participant was instructed to talk for a few minutes about one of three topics (e.g. a recent vacation) with minimal prompting. The measure reported throughout the chapter is the proportion of grammatical sentences out of total sentences produced.

Results Following treatment, there was an increase in the percentage of grammatical sentences the participant produced in English, the treated language, in the sentence elicitation task (from 52% to 62%) and in the narrative production task (from 70% to 89%). A similar increase was observed in the participant’s third language, French, which was not treated, in the sentence elicitation task (from 61% to 76%) and in the narrative production task (from 8% to 58%) (see Figure 7.1). In Hebrew, the non-treated L1, increased

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

111

35

Number of uerances

30 25 20 ungrammacal

15

grammacal

10 5 0 Pre English 1

Post English 1

Figure 7.1 Proportion of grammatical and ungrammatical sentences in French, L3, narratives following the first treatment in English (Case 1)

grammaticality was noted only in the narrative production task and only when the grammaticality of complex sentences was examined (from 55% to 80%); overall grammaticality of all sentences in Hebrew did not change in the sentence task or the narrative task, with high accuracy rates obtained before and after treatment. Little change was noted in any of the languages following the second treatment phase which did not focus on morphosyntactic production.

Summary Positive treatment effects were found in the treated language following treatment as well as in two non-treated languages: French, which exhibited lower post-CVA proficiency than the treated language, and Hebrew, which had comparable proficiency to the treated English.

Case 2 History A woman in her 40s enrolled in the treatment study 13 years post a left CVA. She acquired Persian as her first language and it continued to be the language at home after the family moved to Germany, where she acquired her second language. She achieved high proficiency in German, completing high school and college in Germany. She learned her third language, English, in late childhood at school, and achieved high proficiency while immersed in the language after moving to the US. She reported high proficiency in all three languages prior to her stoke, with the exception of limited literacy in Persian. At the time of the study she demonstrated mild aphasia in English, with a Western Aphasia Battery (WAB) Aphasia

112

Par t 2: A ssessment and Treatment

Quotient (AQ) of 78 and with morphosyntactic and lexical retrieval difficulties in sentence and discourse production. Her speech production was slow, with hesitations and self-corrections. She had no marked apraxia or dysarthria, and no marked comprehension or cognitive impairment. In German, she experienced greater difficulty in her language production than in English and self-rated her abilities lower than in English. Her least recovered language was Persian, with marked difficulty with both comprehension and production skills. (Additional details about her performance in Persian can be found in Conner et al., in preparation.) She did not receive language therapy in German or Persian prior to her enrollment in this study; indeed, she did not attempt to use these languages and reported no spontaneous improvement prior to the study.

Design The participant received three treatment blocks, one in each of her three languages. Testing was administered in all three languages, repeated over three days, prior to and following each treatment block: before and after the English treatment, after the Persian treatment, before and after the German treatment (there was no break between treatment in English and Persian and a month break prior to the third treatment block). Each treatment block comprised 35–40 hours. Treatment was provided in a relatively intense schedule of seven to eight hours a week.

Treatment Treatment adhered to the principles of a modified constraint-induced therapy with emphasis on verbal production and exchange of new information (see Kempler & Goral, 2011, generative treatment). All treatment blocks used similar language activities that focused on elicited sentences while describing action pictures. The first treatment block was administered in English and targeted structures including simple sentences, verb agreement and tense morphology, and prepositional use; the second block was administered in Persian and focused on elicited complete (simple) sentences; the third block was administered in German and focused on a variety of structures, including simple sentences, verb agreement and tense, articles and pronouns.

Measures Data from two measures, the picture description subtest of the Bilingual Aphasia Test (BAT, Paradis & Libben, 1987) and a personal narrative production task, were considered here; the analysis focused on proportions of grammatical sentences produced. Results are reported for all three languages.

Results In the picture description task, the participant increased her proportion of grammatical sentences in English following treatment in English (from 30% to 57%). This proportion decreased in English following the next treatment

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

113

Number of uerances

60 50 40 30 ungrammacal

20

grammacal

10 0 Pre English

Post English

Pre Persian

Post Pre Post Persian German German

Figure 7.2 Proportion of grammatical and ungrammatical sentences in English, L3, picture description following three treatment blocks (Case 2)

block, administered in Persian (from 57% to 32%), and increased again following treatment in German (from 24% to 37%) (see Figure 7.2). In Persian, the participant initially produced few sentences and virtually none were grammatical. Following treatment in Persian, she produced significantly more utterances (from 8 to 32) and more were complete sentences, but the proportion of grammatical sentences did not increase. There was no change in her Persian performance following English or German treatment. In German, the proportion of grammatical sentences increased following treatment in German (from 15% to 38%). There was no change following treatment in English and an increase following treatment in Persian (from 3% to 17%). In the narrative task, no substantial change was observed in English following any of the treatment blocks. In Persian, a significant increase in the number of sentences produced was evident following treatment in Persian as well as when performance was compared from before the treatment began to following the last treatment block; however, as in the picture description task, the proportions of grammatical sentences did not change, with the majority of the participant’s sentences in Persian being ungrammatical. In German, grammaticality increased following treatment in German (from 0% to 30%) as well as following the treatment in English (from 0% to 30%), but decreased following the intermediate block in Persian (from 20% to 6%).

Summary Positive within- and across-languages treatment effects on the production of grammatical sentences were observed in English and in German following treatment in English and in German. In contrast, following treatment in Persian (the most impaired language), performance in the more proficient, untreated languages decreased. This was true for English in the picture description task and for German in the narrative task.

114

Par t 2: A ssessment and Treatment

Case 3 History A bilingual man in his 30s enrolled in two blocks of treatment, one in each of his languages, three years post a left CVA. A native speaker of English, he was exposed to Hebrew since early childhood at home in the US and later improved his language production and literacy skills by taking language classes and during several extended stays in Israel. Pre-morbidly, his English skills were superior to his Hebrew skills and, similarly, at the time of the study he rated his Hebrew as more impaired than his English. In English, he presented with mild non-fluent aphasia with a WAB AQ 86.4, morphosyntactic and lexical retrieval difficulties in sentence and discourse production and good comprehension and cognitive skills. In Hebrew, he experienced more severe impairment with great difficulty in word retrieval and language production and mild comprehension difficulty. He had been receiving language therapy in English prior to participation in this study but no therapy in Hebrew.

Design The participant received two treatment blocks, the first in Hebrew and the second, approximately a year later, in English. Testing was administered in both languages, repeated over three days, prior to and following each treatment block. Each treatment block comprised 35–40 hours in an intense schedule of about 20 hours a week.

Treatment Treatment adhered to the principles of a modified constraint-induced therapy with emphasis on verbal production and exchange of new information (see Kempler & Goral, 2011, generative treatment). Both treatment blocks used similar language activities that focused on elicited language while describing action pictures. The first treatment, administered in Hebrew, focused on elicited single nouns and verbs as well as complete simple sentences; the second block, administered in English, elicited sentences and targeted structures including simple sentences, verb agreement and tense morphology, and pronoun use.

Measures The picture description subtest of the BAT (Paradis & Libben, 1987) was used to examine sentence grammaticality. As in the two previous cases, reported here is change in proportion of grammatical sentences. Preliminary results are presented for English; further analysis of data collected in English and in Hebrew is in progress.

Results The proportion of grammatical sentences in the picture description task decreased in English following treatment in Hebrew (from 44% to 38%) and increased following treatment in English (from 36% to 52%) (see Figure 7.3).

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

115

60

Number of uerances

50 40 30

ungrammatical grammatical

20 10 0 Pre Hebrew Post Hebrew treatment treatment

Pre English treatment

Post English treatment

Figure 7.3 Proportion of grammatical and ungrammatical sentences in English, L1, picture description following two treatment blocks (Case 3)

Summary Following treatment in Hebrew (the less proficient language), performance in the more proficient untreated English decreased.

Case 4 History A multilingual man in his 40s enrolled in the treatment study five years following a left CVA. A native of Barcelona, Spain, he acquired Catalan from birth and Spanish from early childhood, and learned French in early childhood and German and English in late childhood. He achieved high proficiency in Spanish and Catalan and moderate–high proficiency in the remaining languages. Following the stroke and prior to participation in the current study, he received language therapy in some of his languages. At the time of the study, the relative strength of his languages was maintained, with greater impairment in the less proficient languages than in Catalan and Spanish. His language production was characterized by slow, distorted speech, hesitations, morphosyntactic and lexical retrieval difficulties in sentence and discourse production and good comprehension and cognitive skills. For additional details about the participant’s history and performance see Goral et al. (in press).

Design Treatment was administered in Spanish and then in English, at a time when the participant was visiting the US. Testing was administered in all five languages prior to and following each treatment block. Each treatment block comprised 22–25 hours administered in a fairly intense schedule over a period of five weeks.

116

Par t 2: A ssessment and Treatment

Treatment Each treatment session comprised three activities, administered for object and action targets separately. The first activity was a modified semantic feature analysis (Boyle & Coelho, 1995). The second activity was a sentence generation task: using the same target words, the participant was instructed to verbally produce a sentence containing the target word. The third activity was a rapid naming task, in which the participant was presented with the pictures of the target words and instructed to name them as quickly as he could.

Measures The picture description subtest of the BAT (Paradis & Libben, 1987) was used to examine sentence grammaticality. As in the three previous cases, reported here is change in proportion of grammatical sentences. Results are presented for English and Spanish (the treated languages), as well as for French and German, the two untreated languages.

Results The proportions of grammatical sentences increased in English following treatment in English (from 0% to 50%) and did not change following treatment in Spanish. In Spanish, there was a small increase following treatment in Spanish (from 75% to 79%) and a decrease (from 79% to 40%) following treatment in English (see Figure 7.4). In the untreated languages, there was an increase in German following treatment in English (from 0% to 30%) and no change following treatment in Spanish; in French there was no significant change. 30

Number of uerances

25 20 15

ungrammatical grammatical

10 5 0 Pre Spanish

Post Spanish

Post English

Figure 7.4 Proportion of grammatical and ungrammatical sentences in Spanish, L2, picture description following two treatment blocks (Case 4)

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

117

Summary Increased grammaticality was observed in each of the treated languages; a positive cross-language effect was noted in German following treatment in a language of comparable proficiency, English. A negative cross-language effect was observed from the weaker language, English, to the stronger, Spanish, following treatment in English.

Summary of results Data from four participants with aphasia who received treatment in one or more of their languages revealed that, under certain circumstances, treatment benefit generalizes to the untreated languages of multilingual speakers. This finding, along with findings from previous studies of aphasia treatment with multilinguals (e.g. Abutalebi et al., 2009; Edmonds & Kiran, 2006), is consistent with theories that hypothesize that the neural networks associated with multiple languages in the adult brain are primarily overlapping and that multiple languages are co-activated during processing. Specifically, we found facilitatory cross-language effects between pairs of languages with approximately comparable proficiency. This was the case for the untreated L1 (Hebrew) following treatment in L2 (English) of the first participant; and for the untreated L4 (German) following treatment in L5 (English) of the fourth participant. Facilitatory effects were also observed from a stronger treated language to a weaker untreated one, as demonstrated in the untreated L3 (French) following treatment in the stronger L2 (English) of the first participant, and for the untreated L2 (German) following treatment in the most-recovered L3 (English) of the second participant. Additionally, we found support for the prediction that during processing in a weaker language, the more proficient languages are actively inhibited. This was seen in the negative cross-language effects from the treated (L1) Persian to the untreated L2 (German) and L3 (English) of the second participant; from the treated L2 (Hebrew) to the untreated L1 (English) of the third participant; and for the untreated Spanish (L2) following treatment in L5 (English) of the fourth participant. We note, however, that facilitation was observed for the untreated L3 (English) following treatment in L2 (German) of the second participant. It is possible that the proficiency difference between the two languages was not great or that similarities between the two languages lead to a facilitation pattern despite the proficiency difference. The results obtained from the cases described above are consistent with the role of relative language proficiency in determining the nature of crosslanguage effects following treatment. Many, albeit not all, multilingual individuals with aphasia who had achieved differing levels of proficiency in their languages prior to their stroke, demonstrate parallel impairment following the stroke, that is, a preserved relative proficiency among their languages

118

Par t 2: A ssessment and Treatment

(as in our fourth participant). However, multilinguals with aphasia may experience greater difficulty post-stroke in a language that was more proficient pre-stroke than in one that was less proficient before the stroke (as in our second participant) (for further discussion see Fabbro, 1999; Faroqi-Shah et al., 2010; Paradis, 2004). The patterns of cross-language treatment effects we observed here seem to generally follow the predicted inhibition from a weaker to a stronger language and facilitation from a stronger or comparable language when we consider the participants’ post-stroke proficiency, not their levels of proficiency prior to the stroke. For example, the first participant had high proficiency in all three languages prior to his aphasia onset but experienced relatively more impairment in his L3, French, following the stroke; he demonstrated benefit from the treated English to his more impaired language, French. Similarly, the second participant had higher proficiency in her L3 English and lower proficiency in her L1 Persian post her aphasia onset (in contrast to high proficiency in all her languages prior to her stoke) and demonstrated negative effects on the stronger languages following treatment in her more impaired Persian.

Conclusions At least one factor that appears to contribute to the presence or absence of positive cross-language effects following aphasia treatment in multilingual individuals is the relative post-stroke proficiency of the treated and untreated languages. Data reported here are consistent with the hypothesis that when multilingual individuals engage in language processing, all their languages are active and that when they focus their processing on their weaker language(s), their stronger language(s) are actively inhibited. In turn, this inhibition may result in the depressed performance of the stronger language(s) observed during post-treatment testing. The degree to which this effect is transient and likely to be reversed warrants further examination. Furthermore, whether the inhibition is local, that is, the effect is confined to the language components that were targeted during treatment or global, that is, all aspects of the inhibited language are inhibited, is currently an open question. Converging evidence from additional language treatment studies, as well as psycholinguistic studies assessing co-activation of multiple languages and neurolinguistic studies tracking neuronal activation during multiple language processing will allow for a deeper understanding of selective language activation and inhibition in multilinguals. The clinical implications of the findings reported here include the prediction that treating individuals in their weaker language may impede performance in their stronger one, at least in the short run. Of course, the decision of what language should be selected as the target of aphasia treatment is a complex one and typically involves a number of variables, such as the

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

119

language most useful for the clients in terms of their interlocutors and their treatment goals, as well as availability of clinicians (Albert & Obler, 1978; Adrover-Roig et al., Chapter 2, this volume; Lorenzen & Murray, 2008; Paradis, 2004). These considerations can be further informed by research evidence from treatment studies in multilingual speakers.

Acknowledgements I thank the four participants who enthusiastically participated in the treatment studies. I am grateful to my collaborators and to our research assistants, Carmit Altman, Peggy S. Conner, Leila Geramian, Dolors Girbau, Jessica Harris, Daniel Kempler, Violaine Lazecki, Erika Levy, Kristen Maul, Maryam, Naghibolhosseini, Loraine K. Obler, Keren Ohayon, Yana Pugach, Jason Rosas and Melissa Santander. A version of this paper was presented at the Academy of Aphasia meeting in Montreal, 2011, as part of the symposium ‘Bilingual Aphasia: Treatment & Recovery’ organized by Swathi Kiran. The work was supported, in part, by NIH grant DC-009792, and by Lehman College and the Graduate School and University Center of the City University of New York.

References Abutalebi, J. and Green, D. (2007) Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics 20, 242–275. Abutalebi, J. and Green, D.W. (2008) Control mechanisms in bilingual language production: Neural evidence from language switching studies. Language and Cognitive Processes 23 (4), 557–582. Abutalebi, J., Rosa, P.A.D., Tettamanti, M., Green, D.W. and Cappa, S.F. (2009) Bilingual aphasia and language control: A follow-up fMRI and intrinsic connectivity study. Brain and Language 109 (2–3), 141–156. Albert, M. and Obler, L.K. (1978) The Bilingual Brain: Neuropsychological and Neurolinguistic Aspects of Bilingualism. New York: Academic Press. Altman, C., Goral, M. and Levy, E.S. (in press) Integrated narrative analysis in multilingual aphasia: The relationship among narrative structure, grammaticality and fluency. Aphasiology. Boyle, M. and Coelho, C.A. (1995) Application of semantic feature analysis as a treatment for aphasic dysnomia. American Journal of Speech-Language Pathology 4 (4), 94–98. Conner, P.S., Neghibolhosseini, M., Maul, K.K. and Goral, M. (in preparation) Language change following 13 years of non-use in a Persian speaker with aphasia. Costa, A. and 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. Costa, A., Miozzo, M. and Caramazza, A. (1999) Lexical selection in bilinguals: Do words in the bilingual’s two lexicons compete for selection? Journal of Memory and Language 41 (3), 365–397. Costa, A., Santesteban, M. and Ivanova, I. (2006) How do highly proficient bilinguals control their lexicalization process? Inhibitory and language-specific selection mechanisms are both functional. Journal of Experimental Psychology: Learning, Memory, and Cognition 32 (5), 1057–1074.

120

Par t 2: A ssessment and Treatment

Dijkstra, A. (2005) Bilingual visual word recognition and lexical access. In J.F. Kroll and A.M.B. De Groot (eds) Handbook of Bilingualism: Psycholinguistic Approaches (pp. 179– 201). New York: Oxford University Press. Dijkstra, T., Van Jaarsveld, H. and Ten Brinke, S. (1998) Interlingual homograph recognition: Effects of task demands and language intermixing. Bilingualism: Language and Cognition 1, 51–66. Dijkstra, T., Moscoso del Prado Martín, F., Schulpen, B., Schreuder, R. and Baayen, H. (2005) A roommate in cream: Morphological family size effects on interlingual homograph recognition. Language and Cognitive Processes 20 (1–2), 7–41. Edmonds, L. and Kiran, S. (2006) Effects of semantic naming treatment on crosslinguistic generalization in bilingual aphasia. Journal of Speech Language and Hearing Research 49, 729–748. Fabbro, F. (1999) The Neurolinguistics of Bilingualism: An Introduction. Hove: Psychology Press. Faroqi, Y. and Chengappa, S. (1996) Trace deletion hypothesis and its implications for intervention with a multilingual agrammatic aphasic patient. Osmania Papers in Linguistics, 22–23. Faroqi-Shah, Y., Frymark, T., Mullen, R. and Wang, B. (2010) Effect of treatment for bilingual individuals with aphasia: A systematic review of the evidence. Journal of Neurolinguistics 23 (4), 319–341. Giussani, C., Roux, F-E., Lubrano, V., Gaini, S.M. and Bello, L. (2007) Review of language organisation in bilingual patients: What can we learn from direct brain mapping? Acta Neurochirurgica 149, 1109–1116. Galvez, A. and Hinckley, J.J. (2003) Transfer patterns of naming treatment in a case of bilingual aphasia. Brain and Language 87 (1), 173–174. Goral, M., Levy, E.S., Obler, L.K. and Cohen, E. (2006) Cross-language lexical connections in the mental lexicon: Evidence from a case of trilingual aphasia. Brain and Language 98, 235–247. Goral, M., Levy, E.S. and Kastl, R. (2010) Cross-language treatment generalization: A case of trilingual aphasia. Aphasiology 24 (2), 170–187. Goral, M., Rosas, J., Conner, P.S., Maul, K.K. and Obler, L.K. (in press) Effects of language proficiency and language of the environment on aphasia therapy in a multilingual. Journal of Neurolinguistics. Green, D.W. (1998) Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition 1, 67–81. Helm-Estabrooks, N. and Nicholas, M. (2000) Sentence Production Program for Aphasia. Austin, TX: Pro-Ed. Kempler, D. and Goral, M. (2011) A comparison of drill- and communication-based treatment for aphasia. Aphasiology 25, 1327–1346. Kohnert, K. (2004) Cognitive and cognate-based treatments for bilingual aphasia: A case study. Brain and Language 91, 294–302. Kroll, J.F., Bobb, S.C. and Wodniekca, 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. Kroll, J.F., Bobb, S.C., Misra, M. and Taomei, G. (2008) Language selection in bilingual speech: Evidence for inhibitory processes. Acta Psychologica 128, 416–430. Lorenzen, B. and Murray, L.L. (2008) Bilingual aphasia: A theoretical and clinical review. American Journal of Speech-Language Pathology 17 (3), 299–317. Meuter, R.F.I. and Allport, A. (1999) Bilingual language switching in naming: Asymmetrical costs of language selection. Journal of Memory and Language 40, 25–40. Meinzer, M., Obleser, J., Flaisch, T., Eulitz, C. and Rockstroh, B. (2007) Recovery from aphasia as a function of language therapy in an early bilingual patient demonstrated by fMRI. Neuropsychologia 45 (6), 1247–1256.

Cross-L anguage Treatment Ef fec t s in Mult ilingual Aphasia

121

Miertsch, M., Meisel, J.M. and Isel, F. (2009) Non-treated languages in aphasia therapy of polyglots benefit from improvement in the treated language. Journal of Neurolinguistics 22, 135–150. Miller Amberber, A. (in press) Language intervention in French-English bilingual aphasia: Evidence of limited therapy transfer. Journal of Neurolinguistics. Ojemann, G.A. and Whitaker, H.A. (1978) The bilingual brain. Archives of Neurology 35, 409–412. Paradis, M. (1994) Neurolinguistic aspects of implicit and explicit memory: Implications for bilingualism and SLA. In Implicit and Explicit Learning of Second Languages (pp. 393–419). London: Academic Press. Paradis, M. (2004) A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins. Paradis, M. (2009) Declarative and Procedural Determinants of Second Languages. Amsterdam: John Benjamins. Paradis, M. and Libben, G. (1987) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum Associates. Ullman, M.T. (2001) The neural basis of lexicon and grammar in first and second language: The declarative/procedural model. Bilingualism: Language and Cognition 4 (2), 105–122. Ullman, M.T. (2005) A cognitive neuroscience perspective on second language acquisition: The declarative/procedural model. In C. Sanz (ed.) Mind and Context in Adult Second Language Acquisition: Methods, Theory, and Practice (pp. 141–178). Washington, DC: Georgetown University Press.

8

Language Deficits, Recovery Patterns and Effective Intervention in a Multilingual 16 Years Post-TBI Daly Sebastian, Usha Dalvi and Loraine K. Obler

Introduction We report the case of Mr SC, a multilingual 32-year-old with aphasia resulting from traumatic brain injury (TBI) 16 years earlier. The case is of interest for several reasons. First, there was an unusual distribution of language abilities across his two most proficient languages as a result of the TBI: before the threemonth treatment he underwent with the first author most recently, what speech he produced was in his first language, Marathi, but any writing was in his second language, English. Second, the quality of his output was agrammatic in both languages in the modalities in which he produced them. Third, Melodic Intonation Therapy (MIT) proved remarkably successful in enabling him to initiate more complex speech in Marathi and to speak in English. After a description of the language situation in his environment, we report his medical and vocational history and describe his agrammatic output. We then turn to a description of the therapy that was undertaken, including bilingual MIT work and the outcome of the therapy. We conclude with a discussion of the implications of this case for issues that have been brought up in research on multilingual aphasia.

Language Background In India, children growing up in urban areas are invariably exposed to a multilingual and multicultural environment. As a result they learn and use 122

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

123

two or three languages in their earliest childhood. Additionally, in elementary school it is mandatory for the child to formally learn three languages: one being the medium of instruction, the second introduced in first grade (or first standard as it is called in India) and, in fifth grade, a third language is added. Of course, proficiency in their several languages varies substantially across the languages and across individuals. Generally most of the children who, like the individual we write about, Mr SC, attend English-medium schools, study the state language as their second language (in SC’s case this is his mother tongue) and can choose among several local or state languages for their third language. As this report is discussing a multilingual who knew Marathi as a first language (L1), English as his second language (L2), Hindi (L3) and some Gujarati (L4), a brief discussion about similarities and differences among these languages is warranted. Languages in India can be broadly classified into two categories: IndoAryan languages and Dravidian languages. About 80% of all Indians speak one of the Indo-Aryan groups of languages and around 18% of the Indian population speaks Dravidian languages. Indo-Aryan speakers spread across northern and central India, and most Dravidian speakers reside in South India. The 26 or more languages of these speakers are written in different, rarely mutually intelligible scripts; however, all of the scripts of these two major families in India share a common underlying system which originated from Brahmi script. Indic scripts constitute the most typical and most widespread examples of the alpha-syllabic scripts (Salomon, 2000). In the individual we discuss, the L1, L3 and L4 are Indo-Aryan languages and his L2, English, is a foreign language, albeit an Indo-European one, so that it shares certain distant historical roots. While there are quite a few words borrowed from the L2 that are routinely used in L1 (e.g. table, teacher, pen, pencil) there are few shared syntactic features in either the spoken or written forms of the language. In L2, for example, the sentence structure follows a SVO (subject-verb-object) pattern whereas in Indo-Aryan languages (SC’s L1, L3 and L4) the sentence structure follows an SOV (subjectobject-verb) pattern. In the case of his L1 and L3 there is a mutually shared script system which is based on the Devangari script (but historically – until texting became possible in recent years – none of the Indo-Aryan languages used a Roman script like that of English).

Case Report Mr SC, a right-handed accountant, was seen at age 32 for speech therapy to help him with speech, reading and writing in order to hold a job. Sixteen years previously he had received a traumatic brain injury (TBI) as the result of an accident on a train.

124

Par t 2: A ssessment and Treatment

He is a native speaker of Marathi (L1) as he was born in a Marathispeaking family and grew up in the Marathi-speaking community in the state of Maharashtra where it is the official language. He acquired English (his L2) from kindergarten. English remained the medium of instruction throughout his schooling. Up until the time of his accident, he used English in academic, professional, private and recreational settings. His family reports that pre-morbidly his proficiency in English was higher than Marathi. Indeed, we can assume that his ability to read and write in Marathi was competent, as he would have studied it as a course in school and seen it in his environment, but that his ability to read and write it was not as advanced as his ability to read and write English. As well, we should note, SC spoke Gujarati, the language of a neighboring state, with friends outside his home. He developed normally as a child and performed well in school. After his schooling ended at age 15, he entered a college to study a pre-degree course in business.

Medical history On 4 October 1993 (at the age of 16) SC fell from a fast-moving train, and was immediately rushed to a nearby hospital. At the time of admission he was in a coma. SC was operated on that day. Fractured fragments of his skull were removed via a left frontoparietal flap. Craniectomy was carried out to achieve subtemporal decompression and an extradural hematoma of about 25cc removed. In a subsequent surgical procedure carried out four days after the traumatic brain injury, another left frontal extradural hematoma was removed. The previous craniectomy was extended to the frontal region anteriorly and inferiorly to the temporal region. A left temporoparietal contusion was noted. The patient’s coma resolved after 19 days. At that time his medical condition was as follows:

Motor deficits He had a right-sided paresis which lasted for almost three months, after which motor function returned to normal.

Cognitive deficits While he had no post-traumatic amnesia, SC’s family reported that he had difficulties making decisions, and slow processing of auditory information was observed. There were no significant emotional setbacks reported except agitation when he could not express himself (this was observed during pre-treatment assessment and language therapy).

Linguistic abilities SC had severe expressive difficulties in the two languages he was tested in: Marathi and English. His auditory comprehension for both Marathi and English showed gradual improvement. Thus his receptive abilities showed

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

125

parallel recovery in his first and second languages, whereas, as we detail below, his productive abilities differed substantially across modalities. From the onset of his aphasia he had spoken only in Marathi and written only in English. He had little occasion to use his L3 and L4 post the TBI because he remained at home, requiring assistance with activities for daily living (ADL) and having little interaction with those outside his immediate family.

Post-morbid social and occupational history At the time the authors worked with him, SC was able to do all ADL on his own. He traveled independently by bus to familiar places like his workplace and relatives’ homes. One year post his injury, SC attempted to continue his academic studies but due to his language deficits was unable to cope with the program and discontinued on the advice of his neurologist who recommended that his parents enroll him in vocational courses. He also took up photography and has been an active member of The Photographic Association of India since age 21. From that time (1994) until 2009 his parents enrolled him in a variety of vocational training courses in which he was interested that were not crucially dependent on language abilities. These included courses in cooking, graphics and animation, computer software and hardware, travel and tourism, and gold appraisal. He was successful at a data entry job in 2008 but found working eight hours at a stretch painful. Following that, a threemonth course in accounts and inventory management software helped him get the job he began while in therapy with the authors.

Pre-language-intervention assessment SC reported for speech and language rehabilitation 16 years post his TBI in 2009. At the initial assessment, apraxia and dysarthria were ruled out and he was administered the Brief Test of Head Injury (BTHI; Helm-Estabrooks & Hotz, 1990), the Western Aphasia Battery (Kertesz, 1977) and the Bilingual Aphasia Test (BAT; Paradis & Vaid, 1987).

Brief Test of Head Injury (Helm-Estabrooks & Hotz, 1990) Administration of the BTHI showed major impairment in languagerelated but not visuospatial domains (see Table 8.1).

Western Aphasia Battery (Kertesz, 1977) Testing was done first in Marathi and then in English. Results revealed great discrepancies in verbal expression and reading and writing between the languages. Verbal tasks were comparatively better in Marathi whereas reading, reading comprehension and writing were better in English. His severe impairment in reading and writing is suggestive of alexia with agraphia in Marathi (see details in Table 8.2). Agrammatism was present in spoken

126

Par t 2: A ssessment and Treatment

Table 8.1 Pre- and post-treatment BTHI scores BTHI subtests (English)

Pre-treatment evaluations

Post-treatment evaluations

Orientation Following commands Linguistic organizations Reading comprehension Naming Memory Visuospatial skills

62.5% 50% 55.5% 75% 75% 54% 100%

70% 80% 70% 85% 90% 75% 100%

Table 8.2 Pre- and post-treatment WAB scores WAB Subtest

Marathi Pre-therapy

Spontaneous 4/10 speech Fluency 2/10 Auditory verbal 9.85/10 comprehension Repetition 0.4/10 Naming 3.8/10 Reading 0.5/10 Writing 1.3/10 Apraxia 60/60 Aphasia quotient 40.1/100

English Post-therapy

Pre-therapy

Post-therapy

7/10

0/10

5/10

6/10 9.9/10

0/10 9.1/10

3/10 9.7/10

5/10 6.8/10 2.4/10 3.4/10 60/60 69.4/100

0.5/10 1.5/10 6.8/10 6.75/10 60/60 22.2/100

4/10 2.2/10 8.5/10 8.6/10 60/60 47.8/100

Marathi and in written English. (It could not be observed in writing in Marathi or in speech in English because at this time he produced so little of these modalities in these languages.) Verbs, tenses, prepositions, person, number and gender markers, and possessives were sometimes present. The pattern of agrammatism was similar in both spoken responses in Marathi and written ones in English.

Example 1. WAB picture description SC’s written description of the WAB picture can be seen in Figure 8.1. The remaining items in the picture he named orally using nouns in Marathi (e.g. ‘pathang’ (kite)) and iconic gestures (Figure 8.1).

Example 2 The clinician narrated the following short story: ‘My uncle went to Italy. Every day he ate 6 eggs and he put on weight.’ After 30 minutes the patient was asked to write the story. Figure 8.2 shows what he wrote pre-treatment.

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

127

Figure 8.1 Pre-treatment written sample of the WAB picture description task

Figure 8.2 Pre-treatment written sample of the narration task Notes: First he wrote ‘Rome’ then self-corrected it

Bilingual Aphasia Test (Paradis & Vaid, 1987) The English-Hindi combination of the BAT was used. Hindi tasks were adapted to Marathi lexicon as the languages are orthographically and syntactically similar. SC was unable to perform on the translation subtest. Details are given in Table 8.3 (1 & 2). SC’s expressive abilities thus showed what is termed ‘differential recovery’ in the bilingualism aphasiology literature. His use of oral expression in Marathi had gradually increased from word level to limited simple sentences about his daily routine. However, at the time of testing his verbal expressions consisted predominantly of phrases with only content words – mostly nouns. In English, by contrast, he did not speak at all. The differential recovery thus was evident in opposing patterns vis-à-vis modality: in the written modality, SC’s English reading comprehension and writing were agrammatic but relatively good, while he could not perform on written-language tasks in Marathi.

Language Intervention SC was enrolled for approximately once-weekly hour-long speechlanguage therapy sessions for 12 sessions between December 2009 and April 2010. Then he had two additional sessions in the following month, after which therapy was discontinued due to his job schedule. His goal was to improve his functional communication so he could obtain and hold a better job. (Indeed, follow-up sessions twice in the fourth month were discontinued because he had applied, and was hired for, a new, better-paying job as an accountant.)

128

Par t 2: A ssessment and Treatment

Table 8.3(1) Pre- and post-treatment Bilingual Aphasia Test (BAT) scores (parts A and B) BAT sub tests (English)

Marathi (tasks adapted) Pre-therapy

Pointing 7/10 5/10 Simple and semicomplex commands Complex commands 3/5 Verbal auditory 12/18 discrimination Syntactic 62/87 comprehension Synonyms 2/5 Antonyms 3/5 Grammaticality 5/10 judgment Semantic acceptability 7/10 Word repetition 9/30 Naming 5/20 Reading comprehension 0/10 of words Reading comprehension NA of sentences

English Post-therapy

Pre-therapy

Post-therapy

9/10 8/10

9/10 10/10

10/10 10/10

4/5 15/18

3/5 10/18

4/5 14/18

78/87

68/87

79/87

4/5 4/5 8/10

3/5 3/5 4/10

5/5 4/5 6/10

10/10 20/30 12/20 4/10

8/10 2/30 0/20 8/10

10/10 6/30 4/20 10/10

7/10

10/10

NA

Table 8.3(2) Pre- and post-treatment scores for BAT part C (translation) BAT part C: translation

Pre-treatment

Post-treatment

Marathi to English English to Marathi

0/5 0/5

2/5 3/5

Several therapeutic approaches were taken to help him reach his goal. These included MIT to increase and improve oral expression, tasks based on the Helm Elicited Program for Syntactic Stimulation (HELPSS), a functional communication approach, and, to enhance word-retrieval specifically, semantic linking, a lexical neighborhood task and cueing techniques. Sessions were conducted bilingually, in Marathi and English, as the clinician was proficient in both. During therapy, as in the pre-treatment assessment, for some tasks SC spoke in Marathi and for some he wrote in English. That is, when he was not able to express himself in speech or when he thought that the listener was not able to understand him, he switched over

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

129

not only to a different modality but also to a different language. However, he never switched between the languages within a modality, except in the semantic-linking task where he sometimes spontaneously responded successively in the two different languages and modalities. Aiming to spend half of each session in Marathi and half in English, the therapist used both languages deliberately in treatment. For example, during one of the sessions, SC wrote ‘I have one brother’ in English then he was instructed to read that sentence aloud in English. After the reading task was completed, he was encouraged to say and then write the same sentence in Marathi.

Intervention procedure All the sessions followed the same structure and strategies, centering on different topics of discussion. Below, by way of example, we detail the treatment in the first session, which focused on the 2008 terror attack in Mumbai. Some of the topics discussed were his pilgrimage tour with his family to Shirdi, his family, his new laptop, a film he had seen, his favorite film stars, his school years, a visit to Kerala and soccer.

Selection of target sentence SC was asked to explain the terror attack in Mumbai. He tried to explain with a few Marathi words, facial expressions, gestures and he wrote some content words in English: Based on his responses the following target sentence was selected ‘Terrorists attacked major places in Mumbai.’

Example 3 In discussing the terror attack, SC was asked ‘Which are all the places that were attacked by the terrorists?’ His agrammatic response (with contraction of the word ‘Terminal’) is shown in Figure 8.3.

Use of HELPSS First the clinician explained the related event in a short narrative and ended it with the target sentence ‘Terrorists attacked major places in

Figure 8.3 Agrammatic written sample of discussion task on a terror attack

130

Par t 2: A ssessment and Treatment

Mumbai.’ Then the clinician repeated the same story and ended the story without the target sentence and the client was asked to produce the target sentence. He immediately wrote the content words of the sentence in English and spoke their translation equivalents in Marathi.

Use of Melodic Intonation Therapy (MIT) MIT (Albert et al., 1973) was incorporated to facilitate SC’s difficulty with verbal expression in English. The same target sentence was used (e.g. Terrorists attacked major places in Mumbai). The response to MIT was startling because, while he had produced no spoken words in English up to this point, on the first trial he was able to produce the target sentence in English quite fluently. Indeed, his success with MIT was so immediate that during the first two sessions the clinicians demonstrated how to use it at home. To generalize the principles of MIT outside the clinical situation, SC was guided to tap on his lap or the arm of his chair while speaking or to hum softly and speak, whichever he found convenient to do.

Focus on writing Once the patient uttered the target sentence in English using MIT principles, he was asked to say the same thing in Marathi and he did so, then he was instructed to write the sentence in Marathi. As he was not able to do that, the clinician wrote the sentence, providing a model, and told him to copy it. He copied the sentence exactly.

Focus on word retrieval Then the clinician asked him a few questions related to the topic of discussion. For example: ‘What was the other place attacked by the terrorist?’ When he was unable to retrieve that phrase, the technique of semantic linking was employed. The target was ‘Nariman Point’ so some semantically related nouns were given: for example, seashore, Queen’s Necklace (a wellknown tourist spot in southern Mumbai near to the target location, Nariman Point, itself an important tourist site). He was not able to retrieve the target word so the lexical neighborhood prompts were given: for example, ‘Tourists visit that place.’ As SC was still unable to retrieve ‘Nariman Point’, the clinician provided the verbal cue ‘point’. Then SC wrote the full place name, Nariman Point, correctly in English, however he was not able to say it. To facilitate the production of the spoken syllables ‘Na - ri - man - po int’ the clinician instructed SC to use the MIT principles (tapping and

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

131

humming). With that he was able to say the phrase easily. Most impressively, these and other words which were retrieved via MIT were then spontaneously produced later on during the session and again even after two weeks without the use of MIT.

Post-treatment evaluation on 19 June 2010 Outcome measurements were made via the same clinical tools that were used for pre-treatment testing (see Tables 8.1–3). Substantial improvement was evident in spontaneous speech, repetition and naming in L1 (Marathi). While he still had agrammatic output, SC’s sentence length had improved and the number of agrammatic errors was reduced. His verbal communication now included three behaviors which he had not used previously: (1) switching languages within modality, (2) using e-communication with friends and family in a mixture of English and Marathi (in English transliteration), (3) using cellular and landline telephones. Slight improvement was seen in reading and writing in Marathi, as SC had started to recognize Marathi letters and words. For example, he wrote the Marathi letter ‘ ’ and then spoke its correct pronunciation ‘p’ in English. Improvement was seen in translation abilities also in both directions at the morpheme and word level in writing, and in speech at the phrase and simple sentence (three to four words) level. A little improvement was observed in spontaneous speech in English. Naming, reading and writing aspects in English showed a modicum of improvement; only repetition was clearly improved, even though there was no direct work on repetition in therapy except during MIT. SC’s written English was still agrammatic post-treatment, but he included some verbs and generally wrote longer, fuller utterances.

Example 4. Post-treatment When, half an hour after he had heard them, he was asked to write the sentences ‘My uncle went to Italy. Every day he ate six eggs and put on weight’, SC now wrote the sentences as shown in Figure 8.4. His response was longer and more complete than the pre-treatment one seen in Figure 8.1, though it was still clearly agrammatic (e.g. ‘food of eggs’ for ‘he ate 6 eggs’). Here he produced verbal paraphasias, writing ‘Dad’ for ‘uncle’ and ‘Brazil’ for ‘Italy’.

Figure 8.4 Post-treatment written sample of narration task

132

Par t 2: A ssessment and Treatment

Figure 8.5 Post-treatment written sample of the WAB picture description task

Example 5. Post-treatment WAB picture description Also, SC was now mixing oral and written language, across his two languages, for better communication (see Figure 8.5). SC completed the last written phrase orally in Marathi by saying ‘khelto’ (play).

Agrammatism in Marathi Before the current treatment course, SC would only respond minimally as tasks required. Post-treatment he was still agrammatic, but would generate longer phrases and spontaneously make conversation points. For example, when he saw the picture of a bride he gestured that he attended somebody’s wedding, then the clinician asked whose wedding in Marathi. SC gestured that it was his teacher’s wedding and responded verbally as follows: ‘mi janarlo’. He wanted to convey ‘mi gelo’ (I went). His utterance was agrammatic because he mixed the tense markers in the last word. That is Root word – ‘ja’ = go Past tense – ‘gelo’ Future tense – ‘janar’

(1) (2) (3)

He mixed 2 and 3. Then he gestured that he held the bride’s veil. In Marathi he said: ‘pahale char hey, ani mi panch’ (first four and I five). His mother explained the event: even though that teacher had four brothers, they requested that SC join them in holding the veil. Thus SC wanted to convey that he was the fifth person to hold the veil along with the four brothers. Then, to continue, he gestured along with Marathi utterance ‘thi India nahiyethi’ (she India not now). When asked for an explanation, SC’s mother said that he wanted to convey ‘atha thi Indiath nahi’ (now she is not in India). Both SC and his family reported that his communication had improved outside of treatment. He is now markedly better able to communicate in Marathi with his colleagues, relatives and friends. In order to convey complex

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

133

linguistic information SC continues to use a multi-modality approach, varying the use of modalities from one situation to the next. For example, he might speak in Marathi, write in English and include gestures as well. Qualitative analysis indicates advances in his communication patterns. Also, his parents have now reported that he could read high frequency Marathi words. Thus, while he still is better in spoken Marathi than spoken English, and better in written English than written Marathi, the distance between the two languages in each modality is less than it was pre-treatment, and he mixes the two languages successfully to communicate better.

Discussion After the traumatic brain injury SC underwent at age 16, he displayed severe impairment in all his previously proficient languages. The authors’ clinical evaluation and history taken when he returned for therapy 16 years later revealed an atypical pattern of language deficits and recovery. A number of theoretical constructs contribute to our understanding this case. We touch first on how different organization of expressive and receptive skills may interact with SC’s modality-specific recovery. We then turn to notions of shared syntactic mechanisms to account for his agrammatic production in both languages. Next we argue that a number of the subtypes of impairment and recovery in bilingual aphasia pertain to his case. Finally, we consider how Paradis’ activation threshold hypothesis applies to SC’s case. We conclude with a discussion of the effects of intervention which may have called upon the right hemisphere and SC’s development of effective switching between languages to enhance his communication abilities. We have already commented above on SC’s case representing an interesting example of differential aphasia in the productive but not receptive modalities. This peculiar symptomatology is consistent with the fMRI findings of Kim et al. (1997), which suggested overlapping brain regions for healthy bilinguals’ languages in the superior temporal lobe for a comprehension task, whereas less overlap was seen in activation in the Broca’s area region on a task representing production skills (i.e. participants were asked to describe something to themselves). As Albert and Obler (1978) suggest, and as Paradis (1981) terms the ‘tripartite hypothesis’, identical items across the languages of the multilingual are likely represented by a shared neural substrate and those which are different have their own neural representations. That SC has agrammatism similarly in the oral modality (Marathi) and the written modality (English) raises the interesting issue of what one might presume is the existence of a common central syntactic mechanism for both the languages irrespective of modality. Problematic in making this claim, of course, are the substantial differences in syntax and morphosyntax between the two languages. Nevertheless, it is clear that SC was pretty severely

134

Par t 2: A ssessment and Treatment

agrammatic in both languages pre-treatment, and remains agrammatic, albeit markedly less so, in both languages, in whatever modality he employs, post-treatment. Further testing of the specific features of the agrammatism in both his best languages would be warranted to discover language-specific features that might be expected to be reflected in his relative strengths and weaknesses. Similar to SC’s original modality-specific pattern of language deficits, his recovery pattern also varied between the languages. The primary dichotomy in types of language recovery that has been identified in the literature on bilingual and multilingual speakers who have suffered aphasia is between parallel and non-parallel recovery (Albert & Obler, 1978; Paradis, 1977, 1983, 1998). In the most common case, that of parallel deficit and recovery patterns, both (or all) languages demonstrate similar deficits and appear to recover at a similar rate. Apart from his expressive agrammatism, we cannot say that SC showed evidence of parallel deficit and recovery, of course. In some cases of bilingual aphasia, however, one of the languages appears less affected and shows faster recovery, leading Paradis to taxonomize subtypes of non-parallel recovery (Paradis, 1977, 1993). The types most pertinent to SC’s case include successive recovery, wherein recovery of one language precedes that of the other, antagonistic recovery, in which recovery of one language seems to put recovery of an earlier recovered language into regression (Paradis, 1983), selective recovery, in which only one of the languages seems available for recovery, and unintentional language switching, in which bilingual speakers are unable to avoid switching between their languages (see Gil & Goral, 2004, for a good review of these recovery types). We would argue that SC’s original language picture, post his TBI, was that of selective recovery across modalities. While we have no evidence that what recovery he had before the current treatment showed an antagonistic effect, the results were those that might have arisen from a within-modality antagonism: that is, he only attempted to speak in Marathi and to write in English. Recall that even before the current treatment, SC would mix languages, but only across modalities; post our treatment we saw it as an improvement that he would attempt to mix languages within modality while producing speech or writing, as this enabled him to communicate better. Paradis’ (1987, 1993, 2004) activation threshold hypothesis, which accounts for non-parallel patterns of polyglot aphasia is also helpful in understanding SC’s recovery. Paradis hypothesized that each language component needs to reach a certain threshold of activation to be produced. Stimulation and frequent use of a component lowers its activation threshold and, conversely, inactivation raises that threshold. Moreover, frequent activation of a given component (e.g. a word) may reduce activation levels of competing candidates, including components from another language. Aphasia would

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

135

likely result in elevated thresholds for a language generally or for certain language components; such inactivation of these components would lead to high activation thresholds and production in that language would be difficult or impossible. Patterns of selective inactivation and threshold elevation can thus account for non-parallel patterns of language deficits in polyglots. SC’s deficits and recovery patterns can be explained according to this hypothesis. That is, post his TBI, his oral communication and stimulation were predominantly in Marathi. This should have reduced the activation threshold for oral expression in Marathi. By contrast, SC’s need for oral communication in English was limited and this may have raised the activation threshold for oral expression in English. Post TBI, recall, he read and wrote only in English. Even though this was likely due to his markedly better premorbid proficiency in written English (relative to spoken English, and relative to his lesser proficiency in written Marathi), we can presume that this may have further reduced the activation threshold for reading and writing in English. As to intervention, the two most effective aspects of the eclectic therapeutic approach undertaken by the first author under the supervision of the second both suggest the possibility that SC’s right hemisphere was engaged in his response to therapy. Response to MIT was striking for both of SC’s languages and semantic linking was very effective in both the languages. In conversation, post-treatment, SC would generate his own semantic linking, no longer needing the substantial cueing his parents had provided previously. For example, when he was searching for the place name Kumarakam, which is famous for fish and houseboat travel, he was able to retrieve ‘fish’ and then reported that he ate nice fried fish; he then mentioned boating in the backwaters, at which point he retrieved the word ‘Kumarakam’. Recent research shows that the right hemisphere may be particularly involved in semantic aspects of languages (e.g. St. George et al., 1999). One might speculate that had we been able to undertake functional neuroimaging with SC, we would have seen increased activation in his right hemisphere post-therapy, relative to pre-therapy, similar to that seen in Videsott et al. (2010). As well, as noted above, post-intervention testing showed the emergence of switching between the languages within a single modality, which was not present earlier. There has been much debate concerning ‘where’ in the brain ‘the switching mechanism is localized’. For example, Pötzl (1930) located the switch mechanism in the parietal lobe near and including the posterior section of the Sylvian fissure; Leischner (1948) suggested the supramarginal gyrus or the posterior parts of the second and third temporal convolution. A recent study suggests that language switching elicits greater activation in the right superior prefrontal cortex (BA9/10/32), left middle and superior frontal cortex (BA8/9/46) and right middle cingulam and caudate (BA11) (Wang et al., 2007). Whichever regions may be involved, one might argue that we have evidence from SC’s case that it can be ‘turned on’

136

Par t 2: A ssessment and Treatment

even 16 years post TBI to the improvement of his overall communication abilities. Alternately, the threshold for activation of spoken English or written Marathi may have been lowered via therapy, thus resulting in more possibility for switching. Perhaps the most striking and interesting aspect in SC’s case is the effectiveness of language intervention 16 years after the TBI. Quantitative and qualitative evaluation emphasize the fact that the language intervention accelerated a very slow recovery into a sudden and productive process. Trauma had been at the age of 16 years, but during and even after intervention, SC’s language ability appeared to improve daily. One possible explanation for SC’s marked progress would be a deblocking of whatever prevented his production of writing in Marathi and speaking in English previously (e.g. Weigl, 1970). A second explanation would be that our intervention had induced changes due to plasticity in the brain. Neuroplasticity motivates treatment of acquired brain injury, and we have seen such success in SC’s case that we wish we had been able to document the brain bases of the recovery via a technique like fMRI to see if in fact perilesional or contralateral areas are now engaged that were not before our treatment. Whatever the explanation, improvement in his language functions has markedly increased SC’s confidence and motivation level. As a result of his early treatment success with MIT, he applied for and got a regular job as an accountant with a fairly good salary. Another measure of the success of his recovery is that post-treatment, as is customary in India, his parents undertook to arrange his marriage.

Conclusion SC presents an interesting case of bilingual aphasia in that he initially presented a quite clear example of differential recovery in terms of his modality of expression, using exclusively Marathi in speech and English in writing, consistent with his pre-morbid proficiency in these modes. At the same time, his impairment was agrammatic in both languages, which must be considered evidence of ‘parallel’ impairment. His striking recovery thanks to MIT and the other techniques employed may be considered consistent with engagement of right-hemisphere compensation, brain plasticity and/or deblocking that may be argued to have reduced his activation thresholds for those language domains that were relatively severely impaired pre-treatment. In our experience, when patients report to clinic as late as SC sometimes they may hear ‘you are too late to the recovery process so we cannot expect further improvement’. SC’s case, however, suggests that intervention is worth attempting, perhaps particularly in a relatively young man with the motivation to improve and the external support system to enhance his ability to benefit from treatment.

L anguage Def ic it s, Recover y Pat ter ns and Ef fec t ive Inter vent ion

137

Acknowledgements We thank God almighty for providing us the strength to go ahead and complete this work. We are grateful and thankful to SC and his family for their cooperation and permission for us to write this paper. Thanks to Prof. R. Rangasayee, the Director of Ali Yavar Jung National Institute for the Hearing Handicapped (AYJNIHH), Mumbai, for permitting us to use the department resources for this project and for his valuable comments. Thanks to Ms Sadhana Relekar, Ms Meera Shah, Ms Susan Barla and Ms Mugda Mhatre for their helpful clinical support. Thanks too to Dr Martin Albert for advice on wording the medical history and to Ms JungMoon Hyun for substantial help in creating the manuscript. We extend our thanks to Dr Swapna Sebastian, Ms Gayithri Hattangadi, Dr Geetha Mukundan, Dr Medha Adhyaru and Dr Geeta Rao for their valuable suggestions and Ms Meenal Umrao for typing the documents. The first author worked with SC as a student clinician under the supervision of the second author. Together they generated early versions of this paper which the third author then helped them shape into the current version.

References Albert, M.L., Sparks, R.W. and Helm, N.A. (1973) Melodic intonation therapy for aphasia. Archives of Neurology 29, 130–131. Albert, M.L. and Obler, L.K. (1978) The Bilingual Brain: Neuropsychological and Neurolinguistic Aspects of Bilingualism. New York: Academic Press. Gil, M. and Goral, M. (2004) Nonparallel recovery in bilingual aphasia: Effects of language choice, language proficiency, and treatment. International Journal of Bilingualism. 8 (2), 191–219. Helm-Estabrooks, N. and Hotz, G. (1990) The Brief Test of Head Injury. Austin: Pro-Ed Publishers. Kertesz, A. (1977) The Western Aphasia Battery. New York: Grune and Stratton Publishers. Kim, K.H., Relkin, N.R., Lee, K.M. and Hirsch, J. (1997) Distinct cortical areas associated with native and second languages. Nature 388, 171–174. Leischner, A. (1948) Ueber die Aphasie der Mehrsprachigen. Archiv für Psychiatrie und Nervenkrankheiten 180, 731–775. Paradis, M. (1977) Bilingualism and aphasia. In H. Whitaker and H.A. Whitaker (eds) Studies in Neurolinguistics (Vol. 3, pp. 65–121). New York: Academic Press. Paradis, M. (1981) Neurolinguistic organization of a bilingual’s two languages. LACUS Forum 7, 486–494. Paradis, M. (ed.) (1983) Readings on Aphasia in Bilinguals and Polyglots. Montreal: Didier. Paradis, M. (1987) The Assessment of Bilingual Aphasia. Hillsdale, NJ: Lawrence Erlbaum. Paradis, M. (1993) Linguistic, psycholinguistic, and neurolinguistic aspects of ‘interference’ in bilingual speakers: The activation threshold hypothesis. International Journal of Psycholinguistics 9, 133–145. Paradis, M. (1998) Language and communication in multilinguals. In B. Stemmer and H. Whitaker (eds) Handbook of Neurolinguistics (pp. 417–430). San Diego, CA: Academic Press.

138

Par t 2: A ssessment and Treatment

Paradis, M. (2004) A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins. Paradis, M. and Vaid, J. (1987) Dvibhashai ka pratikshan (Hindi version of the Assessment of Bilingual Aphasia Test). Hillsdale, NJ: Lawrence Erlbaum. Pötzl, O. (1930) Aphasie und Mehrsprechigkeit. Zeitschrift für die gesamte Neurolgie und Psychiatrie 124, 145–162. Salomon, R. (2000) Typological observations on the Indic script group and its relationship to other alphasyllabaries. Studies in the Linguistic Sciences 30 (1), 87–103. St. George, M., Kutas, M., Martinez, A. and Sereno, M.I. (1999) Semantic integration of reading: Engagement of right hemisphere during discourse processing. Brain 122, 1317–1325. Videsott, G., Herrnberger, B., Klaus, H., Schilly, E., Grothe, J., Wiater, W., Spitzer, M. and Kiefer, M. (2010) Speaking in multiple languages: Neural correlates of language proficiency in multilingual word production. Brain and Language 113, 103–112. Wang, Y., Xue, G., Chen, C., Xue, F. and Dong, Q. (2007) Neural bases of asymmetric languages switching in second language learners: An ER-fMRI study. Neuroimage 35 (2), 862–870. Weigl, E. (1970) Neuropsychological studies of structure and dynamics of semantic fields with the deblocking method. In A.J. Greimas, R. Jakobson, M. Mayenowa, S.K. Saumjan, W. Steinitz and S. Zolkiewski (eds) Sign, Language, Culture (pp. 287–290). The Hague, Paris: Mouton.

Part 3 Bilingual Language Phenomena

9

Bilingual Aphasia and CodeSwitching: Representation and Control Alessandra Riccardi

We know from our daily experience that bilingual people1 can select the language most appropriate to the context and mainly to the interlocutor and can alternate languages within the same conversation with absolute ease. It is common in our multicultural societies to see different generations of the same community of immigrants conversing in two languages because both rely on the mutual intelligibility. Notwithstanding the apparent simplicity of the process, many things must be going on in the brain when it selects the target language or shifts from Lx to Ly or when elements from Lx are introduced into the base language of the moment. The issue of the selection and switching of the languages raises a question about the existence of a neuroanatomical switching device versus a distributed process comparable, in nature, to any other process leading an individual to make a choice between two competing elements. If the latter is correct, the selection and the switching between two different languages would not be any different from the choice between two registers or two synonyms within the same language. In turn, the issue of the selection and shift of a language is intimately related to how and where multiple languages are represented in the brain and which variables might influence this representation (age and manner of acquisition, frequency and recency of use). In the following paragraphs we shall review the literature on codeswitching, presenting some of the interpretations and theoretical frameworks developed to explain the functional and neuroanatomical mechanisms that govern the shift from one language to another. Though the comparison with the code-switching in a healthy population is methodologically unavoidable, the description of the sociolinguistic, pragmatic and 141

142

Par t 3: Bilingual L anguage Phenomena

grammatical characteristics of normal code-switching goes beyond the aim of this chapter.

Introduction According to Albert and Obler (1978) language mixing (including what some authors have referred to as code-switching and code-mixing) is present in only 7% of bilingual people with aphasia, most of whom have sensory aphasia. A smaller number of people with different lesion sites are reported to manifest pathological mixing: a person with bilateral trauma (Perecman, 1984), two right-brain damaged left-handed individuals and one with a left frontal tumor (Gloning & Gloning, 1965), one with a left frontal lesion (Fabbro et al., 2000), among a few others. It is crucial to verify whether the literature is a reliable description of the situation in reality and to double check whether the code-switching shown in the published cases was not motivated by the linguistic behavior of the professional during the test session, which usually goes neglected in the reports. To highlight what we are going to discuss moving forward, we have to provide the defining criteria of pathological code-switching. We consider an instance of mixing/switching as pathological when it is pragmatically inappropriate, that is, when the speaker selects a language not understood by the interlocutor and is no longer able to monitor and maintain control over the selection of the desired language. The types of switching in aphasic speech seem to be comparable to those in normal speech, as far as it concerns grammatical constraints though, in aphasic speech, the use of mixing increases (Muñoz et al., 1999).

Code-Switching and Bilingual Aphasia Research on bilingual aphasia is crucial for the understanding of the representation and interaction of multiple languages stored in the same brain and more broadly for the link between brain and language. The deficits caused by a brain injury can isolate and highlight some mechanisms or a part of a mechanism normally hidden in the physiological flow of the processing of language. Aphasia has been defined for this reason as ‘an experiment in nature’ (Menn & Obler, 1982: 6). In the present chapter we shall focus on the code-switching in braindamaged bilinguals to shed light on the nature of the pathological switching and, most of all, on the mechanisms and principles underlying the representation and selection of languages in non-brain-damaged bilinguals. We shall try to track a historical overview of the study of code-switching in aphasia starting from the modern period, then we will discuss the major theoretical

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

143

frameworks and the development of the research with the use of neuroimaging techniques.

Historical overview As told above, code-switching (CS) and code-mixing (CM)2 have deserved the recognition they have received over time as a normal linguistic behavior of the bilingual speaker. Most of the researchers, though, draw a clear-cut distinction between the two phenomena claiming, among other reasons, that while the introduction of a word from the embedded language into the matrix language (CM) is used also by monolinguals in the form of borrowing, CS is proper only of bilinguals. While CS/CM were long judged as a manifestation of lack of competence and accuracy of non-proficient bilinguals, nowadays, on the contrary, linguists speak of a grammar of CS/CM (Myers-Scotton, 1993; Poplack, 1980). Some authors (Poplack, 1998; Muysken, 1995) not only recognize CS as a highly demanding performance from both a cognitive and a linguistic point of view but also claim (Poplack, 1998) that the mastery of CS and CM is the key to defining a pure bilingual. Considering how CS in the non-brain-damaged population was judged in early reports, it is not difficult to intuit how it was interpreted in bilingual people with aphasia; it was almost always classified as pathological, part of the language disruption caused by the lesion. The dawn of a more systematic debate on the nature of the CS/CM in aphasia took place in the early 1980s on the pages of Brain and Language between Perecman and Grosjean. Perecman (1984) published an article starting with an interesting review of the literature reporting cases of polyglots with aphasia3 who showed language mixing and spontaneous translation. She claimed that while the mixing at the lexical level (CM) is seen in individuals with aphasia and non-brain-damaged populations, the utterance level mixing (CS) is almost always inappropriate and therefore part of the aphasic syndrome characterizing sensory aphasia. Perecman also presented different theories providing an interpretation of the nature of CS (and, consequently, of the representation of multiple languages in the brain). Some of the authors (Macnamara et al., 1968; Paradis, 1997) quoted in Perecman’s article support the idea of the switching mechanism as non-languagespecific but rather as a part of a more general skill that allows individuals to perform any task that involves switching or a selection between two alternatives. As far as the neuroanatomical localization of a potential switching mechanism is concerned, both anterior and posterior lesions are reported in the literature on CS. Albert and Obler (1978), however, make a distinction and classify the cases with posterior lesions, supporting therefore a posterior switching mechanism, as selective recovery4 rather than language mixing and

144

Par t 3: Bilingual L anguage Phenomena

the cases with anterior lesions as having CS/CM. This has brought Albert and Obler themselves to hypothesize that the shift from one language to another is not caused by a disruption of the switching mechanism but rather by frontal disinhibition (i.e. part of the manifestation of the frontal syndrome). Other scholars (Paradis, 1977, 1997) have interpreted the switching mechanism as a functional system, arguing against the existence of a neuroanatomical specific site. L’Hermitte et al. (1966) add also a pragmatic aspect to the functional interpretation, by distinguishing an unintentional CS, the consequence of the brain lesion that makes the desired language inaccessible, from an intentional CS, which is strategically used by the person with aphasia who is aware of his/her inability to retrieve the target item in the required language. In order to avoid a breakdown in communication, the tested person would use the language first available, disregarding the fact that it is not the one required by the test or the therapy session. Perecman (1984) found these theoretical approaches not powerful enough to account for all the recovery patterns reported in the literature and, furthermore, partially inconsistent with data. She adopted the Microgenetic Model (Brown, 1979) which interprets the processing of language as a continuum from thought to articulation through a series of cognitive levels. The cognitive progression gradually assumes the form of the language or, in other words, turns the thought into language up to its phonetic form. In this perspective, aphasia would show a linguistic deficit that mediates the corresponding cognitive stage unveiling exactly where the deficit occurs. Instances of CM would be explained by a momentary merge of two homologous stages of the two different and usually independent grammars5 of the languages spoken by the participant. Other experiments quoted by Perecman report left-brain-damaged people who are able to process conceptual information but fail when asked to elaborate on the same pieces of information in lexical forms (Zurif & Blumstein, 1978) and non-language-impaired, right-brain-damaged individuals with a reverse behavior: poor performance in a comprehension task on the conceptual relation between comparative adjectives but good comprehension when the same relation is expressed by means of lexical items. This would be evidence of the existence of a common pre-linguistic level and single, independent grammars, as many as the languages spoken by the individual. Spontaneous translation, that is, a not very frequent deficit consisting of a continuous, automatic and unsolicited translation from one language to another, would therefore be a pre-linguistic deficit, while CS a linguistic deficit. Grosjean (1985), reviewing Perecman’s article, analyzed the cases presented in her paper through the ‘Language Mode Theory’ which describes the linguistic behavior of a polyglot as a continuum whose two extremes are a totally monolingual speech mode. The speaker will be strictly monolingual (that is, positioned at one of the extremes of the continuum) when her/his interlocutor is monolingual but will change the mode along the continuum when her/his interlocutor is a polyglot, based on the conversational situation,

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

145

the topic and so on. In any case, the language mode does not presume a total deactivation of the language not in use, which would explain dynamic interference6 from the language not in use. With particular reference to the phenomena of CS in bilingual aphasia, Grosjean (1985) did not deny that some of the cases presented showed a pathological switching or a spontaneous translation. Instances of pathological switching are easy to identify because patients switch to a language not spoken by the interlocutor, violate the rules of CS, or mix during reading or talking to a monolingual, thus bringing about a breakdown in communication. However, regarding the other cases reported in Perecman’s (1984) article and in general in the literature, Grosjean (1985) cast doubts, first of all, about the environment of the therapy session. If the patient is aware of the fact that her/his examiner is bilingual, s/he will switch language when experiencing lexical retrieval difficulties in the base language in order to carry on the communication. Even if it is the case that the phenomena of CS and CM increase after brain injury, still the shift to another language might be a conscious communicative strategy to bypass the difficulties of retrieving the target item in the tested language. To diagnose the CS of an individual with aphasia as pathological (with the different therapeutic interventions and the potential theoretical interpretations that can derive from this), Grosjean (1985) warns practitioners and researchers to get clear and unambiguous information about the pre-morbid proficiency and the kind of knowledge in all the languages spoken by the patient and the modality (mainly attitude towards CS/CM) and frequency of use before the injury. Along with information about the linguistic history of the patient, particular attention should be paid to the linguistic behavior of the examiner during the testing session: an inadvertent shift from one language to another might trigger a switch in the patient as well (which commonly happens among healthy bilinguals). Grosjean (1989) argued that for most of the cases reported in the literature we do not know anything about the pre-morbid linguistic habits of the subject nor about the linguistic competence of the examiner and her/his behavior during the session test. The issue of the linguistic behavior of the examiners during the sessions with individuals with aphasia is repeatedly underlined also in Paradis’ work: he underlines how the only way to draw an unequivocal picture of the nature of CS is to have the patient tested in each one of his/her languages in a strict ‘monolingual mode’ by a monolingual examiner. Normally the linguistic behavior of the professional goes unnoticed and we risk interpreting a pragmatically appropriate answer as a manifestation of the speech deficit. The Bilingual Aphasia Test (BAT; Paradis et al., 1987), which is available in more than 60 language pairs, in fact aims at evaluating the residual skills on each level of grammar for each language known by the subject along with her/his capacity to translate bidirectionally from one language to another. As stated by Grosjean (1985) himself, he did not criticize Perecman’s claim that pathological switching can exist as part of the aphasic manifestation,

146

Par t 3: Bilingual L anguage Phenomena

but he argued that most of the cases reported might have been misinterpreted for lack of information about the pre-morbid language background of the patient and the conditions of the testing session.

Cognitive Models According to the Microgenetic Model, the aphasic deficits would correspond with a disruption to the pre-linguistic cognitive level which mediates a specific linguistic level. Other models focus more strictly on the concept of the linguistic representation to provide an explanation of the switching mechanism. Many of these models share the assumption that language access is non-selective, that is, each time we want to access an item in the target language, all the languages are activated (Abutalebi & Green, 2007; Kroll et al., 2008; for language-specific selection see Costa & Sansebastan, 2006; Costa & Caramazza, 1999). Bilinguals take longer to name objects than monolinguals (Mägiste, 1979) and this time course is interpreted as the amount of time necessary to solve the competition because all the languages are at first activated. The selection of a language or an item in a given language is the result of a competition in which an element reaches the threshold while the others are ‘prevented’ from doing so. Studies using interlingual homographs (Dijkstra et al., 1998; De Groot et al., 2000) suggest that individuals are slower to identify a word as an item of the target language if it has a homograph in the non-target language. For Dutch-English bilinguals reaction time (RT) is longer to identify the word ANGEL as an English word during the English lexical decision task than a control word because ANGEL has a Dutch homograph meaning sting that evidently gets activated. The main difference among theories exploring the issue of representation and control resides in the way the competition is solved: through different levels of activation of the linguistic systems (Grosjean, 1997; Paradis, 1997), through the inhibition of the nontarget language systems or their single components (Green, 1986; Dijkstra & van Heuven, 1998) or by increasing the activation of items in the output language (De Bot & Schreuder, 1993). What all these theories have in common is, first, the idea that brain injury does not destroy languages and, second, the intent to build up a framework powerful enough to account for both the different post-trauma patterns of recovery and the functioning of the normal system (including speech errors and mixing).

Inhibitory Control Model Green (1986) proposed a system characterized by three components: control, activation and resources. Control is exerted to monitor the selection of the appropriate language. The functional circuit control was extended in

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

147

Green (1998) with the specification of three loci of control: an executive locus to select and maintain the goal, a locus supporting the language task schemas (e.g. the selection of one language over the other or the translation from one language into the other) and third, a locus in the bilingual lexicosemantic system at the lemma level, where the selection of a word in the target language occurs on the basis of a language tag. Albert and Obler (1978) interpret the control as realized by tags attached to each word thereby indicating the language the item belongs to. Control is necessary in a bilingual system to activate one language over the other; activation is not sufficient however to select a language. Rather, to have the system work properly, the non-target language(s) must be inhibited. Control, in sum, is in charge of supervising the activation of the target language and the contemporary inhibition of the non-target language(s). The language system does not work as an on-off switch, that is, the non-target language is never completely deactivated, as experiments quoted above prove. This would also account for interference and slips of the tongue in neurologically healthy bilinguals under stress or fatigued conditions, and instances of CS whereas, for example, an expression or word in one language reaches the threshold before the correspondent one in the base language because it is more tightly linked to the concept in the speaker’s experience. The nonselective access, though, is given only for languages which are maintained active over time. Once more, the frequency of use proves crucial for the way languages are processed and for their reciprocal interaction. Languages unused for long periods (neither spoken nor heard) would not take part in the non-selective access. Green (1986) defines the state of the languages following the classification that Norman and Shallice (1980) and Shallice (1982) developed for nonverbal motor skills. A language can be selected (hence reach the output stage), active (it takes part in the ongoing process and competition for the output), or dormant (it is represented in the long-term memory but is not taking part in the ongoing process hence is not one of the competitors for the output). In order to activate the target language, inhibit the other(s) and exert control over the output, the system needs resources. More precisely, Green’s (1986) framework predicts two different kinds of suppressions of languages: an external suppression and an internal one. An external suppression of L2, for example, is obtained when spontaneously speaking L1. CS, in its normal use, does not require any external suppression, because the output system will select the item that first reaches the threshold. For translation from L2 into L1, the process is more complex because both the language systems are needed; in this case, L2 output must be suppressed by L1 output, otherwise a repetition rather than a translation of the message would be obtained. The suppression is, in this case, internal, because L2 output is suppressed within the system impeding the retrieval of words from L2. External and internal suppression are two different forms of control and they are sustained by

148

Par t 3: Bilingual L anguage Phenomena

different resources. Therefore it might be possible that one of them is impaired while the other is spared; this would provide an interpretation for alternate antagonism, that is, when a bilingual person with aphasia can speak only in L1 for a certain period and then only in L2 in a successive period, and paradoxical translation, that is, when a bilingual person with aphasia can speak for example only in L2 but can translate only from L2 into L1. The Inhibitory Control Model predicts that in antagonistic recovery, the unavailability of one of the two systems might be due to insufficient resources to inhibit the other one, L1 for example. It also predicts that the same kind of deficit should turn out even more clearly in performing a bilingual Stroop task in L2 with a color word in L1 because of the increasing interference. In parallel recovery, which is the most common recovery pattern, the extent to which the two languages recover over time is proportional to the decrease of reciprocal interference since the resources become adequate to all the requirements; in such a case the performance on a bilingual Stroop task should also improve.

The Activation Threshold Hypothesis Paradis (1987) developed a theoretical framework based on an analogy with the functioning of the neuron system (Kandel et al., 1991). The flow of information from neuron to neuron runs through action potentials, which have to reach a threshold to release in the interneuron spaces called synapses, the transmitter that brings the information (excitatory or inhibitory) to the receptor of another neuron. In turn, the receptor does not get activated until the incoming signals have reached its threshold. In the same way Paradis (1987) hypothesizes that a language, as well as one of its components or a single item, reaches its activation threshold through the amount of impulses it gets. The quantity of energies required to reach the threshold is a function of the frequency and recency of use. The more a language (or part of it) is used and the more recent is the last time it was activated, the more its threshold is lowered and therefore the more easily accessible it is. Over time the threshold is gradually raised until a large amount of activation is needed to stimulate it. This is the case of attrition, when a language has not been spoken for a considerably long time. The selection of a language or an item entails necessarily a choice among different competitors. In order to let only the target language/item reach the output, it is necessary to inhibit the competitors or, in Paradis’ (1987) terminology, to raise their activation threshold. Comprehension and production would have different activation thresholds since the amount of impulses necessary to activate the target from an external input as a visual or auditory one is lower than that necessary when a target is activated from within the system by recall or self-activation. Generally speaking, the threshold of comprehension does not have to be as

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

149

low as that of production to be stimulated. This would account for different behaviors in production and comprehension in both healthy speakers and those with aphasia: some subjects find it impossible to produce any utterances in a given language or cannot use a certain syntactic structure or retrieve a word but can understand the same language and its relative structures perfectly. In his model, Paradis (1987) does not map functional mechanisms onto specific cerebral areas but proposes that a target (language, structure or word) is activated when a sufficient amount of impulses reaches the neural substrate where the target is represented. We have presented so far some of the most meaningful and used functional models that propose an interpretation of representation and interaction of multiple languages in the brain. These factors provide a basis to explain the switching mechanism in normal and impaired conditions.

Neuroimaging Data A more insightful picture of the complex interaction of multiple languages in the brain and of the way one language is selected over another and control is exerted to perform different task schemas can be obtained by mapping behavioral data onto neuroimaging data. The techniques currently available to researchers provide information on the increase of blood and the electric activity in the areas involved in the execution of a task and can track the change of the neural area (Indefrey, 2006) and the density of white and grey matter (Mechelli et al., 2004) longitudinally during the language learning process. The investigation of the switching mechanism and the neuroanatomical localization of a specific device for it, if any, cannot disregard the issue of the representation of the languages. Data from aphasia and pathological switching, in particular, in fact, can be differently interpreted on the ground of the cognitive framework adopted to hypothesize where and how languages are represented. In a localizationist view, differential recovery (selective, antagonistic, paradoxical) is read as evidence that different languages are organized in different brain regions. More precisely, while there is general agreement in the literature that words share a common substrate, for grammar the positions vary and some scholars hold the view that grammar would be subserved by a different network from words (Ullman, 2001a). Some others hypothesize that the areas of representation are determined by the age of acquisition (Proverbio et al., 2004; Kim et al., 1997), wherein early and late bilinguals recruit different networks for L1 and L2. One proposal is Ullman’s (2001b) declarative/procedural model; according to his view, the grammar would be mediated by procedural memory while the lexicon by declarative memory.

150

Par t 3: Bilingual L anguage Phenomena

Procedural memory is supposed to be supported by the frontal-subcortical circuit, including the basal ganglia, while declarative memory by the temporo-parietal system; both the procedural and the declarative memory systems are in charge also of tasks other than language processing. Extending the system to the bilingual lexicon, Ullman (2001a) proposes that, L2 grammar, which is typically learned by explicit teaching, unlike L1 grammar, would be supported by declarative memory. As long as proficiency improves, L2 grammar becomes more and more automatized and its representation shifts towards the frontal regions converging with the network supporting L1. Ullman’s model is taken as the key to interpret cases with a frontal lesion presenting a selective recovery of L2 over L1 even with a very low pre-morbid level of L2. Paradis (1994, 2004) also accounts for selective recovery with L2 better recovered by addressing the two different memory systems that support the processing of L1 and L2. However, Paradis (1994) explains selective recovery patterns by assuming that the speaker may rely on the metalinguistic knowledge of L2 that, being learned explicitly, is part of the declarative memory system, rather than by assuming different anatomical networks subserving L1 and L2. The issue of different loci of representation or even of lateralization for different languages has been debated in the literature for many years, from Scoresby-Jackson (1867), who inferred that the impaired language must have been localized in the area affected by the lesion and the recovered language in the area spared by the lesion, to Minkowski (1927/1983), who rejected the idea of the necessity of postulating different cerebral organizations for different languages and imputed the differential recovery pattern to functional factors. Pötzl (1925) hypothesized the development of specialized centers for bilinguals that allow them to switch and translate from one language into another. Based on the fact that 10% of the cases reported in the literature on bilingual aphasia show aphasia following a right-hemisphere lesion, Albert and Obler (1978) and Galloway and Krashen (1980) hypothesize that the second languages would be, at least partially, represented in the right hemisphere. To an extent, a different anatomical organization for multiple languages leads implicitly to postulating a specific device to switch among languages. To date there is quite a consensus among researchers that a single network underlies the representation of more than one language (for a different position see Proverbio et al., 2004; Kim et al., 1997) though neuroimaging studies show a larger area of activation for L2 than for L1 processing (more extended areas than the classic BA 9, 46, 47). The differentiation has been found in the extension of the area of activation, in the density of the grey matter (Mechelli et al., 2004) and in an increase of the connectivity (Indefrey, 2006). According to the dynamic view of bilingualism, both the Convergence Hypothesis (Green, 2003) and the Single Network Hypothesis (Abutalebi & Green, 2007) claim that these changes should be analyzed longitudinally because they are a function of L2 proficiency. Holding that L2 is represented

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

151

and learned through the same network supporting L1, there is still a basis for hypothesizing that any new task requires a more intense activity and a larger area of activation (for a complete review of the studies supporting this view or arguing against it, see Abutalebi & Green, 2007). Furthermore, the network has to adapt to new demands presented by the structural properties of the new language, as, for example, in the case of a native speaker of English learning a tonal language such as Chinese, or a Chinese speaker learning a morphosyntactically rich language such as Italian, or learners processing languages with a different word order from their L1 (Jeong et al., 2007; Green et al., 2006). Since areas specialize over time to perform a task, at the beginning of the learning process new needs determine an expansion of the original area involved in the processing of that linguistic level or an increase in its activity, but as proficiency improves and the frequency of use intensifies, the new network converges with that of L1. In other words, high proficiency correlates with automatic processing; the more the proficiency increases, the less the processing needs control. The assumption of overlapping of L1 and L2 representations is not synonymous with the claim that a bilingual processes his/her L1 or L2 as a monolingual native speaker of that language. Notwithstanding the convergence of the areas, as Grosjean (1989) warns, a bilingual is not two monolinguals in one person. Neuroimaging investigations shed light on the fact that proficiency seems to be more crucial than age of acquisition but see Proverbio et al. (2004) for a different view. Sakai et al. (2004) carried out an fMRI longitudinal study on late bilingual Japanese twins acquiring English and showed that the region processing grammar in L2 converged exactly with that performing the same processing in L1. The specialization of the left inferior frontal gyrus (LIFG) in the process of learning regular and irregular English verbs (L2) was the same involved in L1 processing. These findings show that the same areas underlying L1 are engaged in the processing of homologous levels of L2 from the very early stages of the acquisition process. The same conclusions are drawn by Indefrey et al. (2005) who also conducted a longitudinal fMRI study on adult Chinese immigrants learning Dutch. In six months, the LIFG appeared to be recruited to process L2 syntax as well as L1 syntax (see also Abutalebi, 2008; Perani & Abutalebi, 2005). Golestani et al. (2006) obtained two other interesting findings; first, they found that peaks of activation of L1 and L2 converged as long as L2 proficiency increased, and second, a progressive involvement of the basal ganglia correlated to the increase of L2 proficiency. Since the basal ganglia and the prefrontal cortex, in general, are involved in controlled processing, these results suggest that an enhanced proficiency corresponds to an increase of automaticity in language processing leading to a final anatomical convergence between L1 and L2. The prefrontal activity is not specific for language processing but it is held to be the area in charge of cognitive control. For a complete review of event related potentials (ERP) studies investigating control in bilinguals see Moreno et al. (2008).

152

Par t 3: Bilingual L anguage Phenomena

The Switching Mechanism: A Specific Device or a Distributed Process? The issue of the organization of multiple languages is deeply connected to the switching mechanism. If we hold that languages are represented in the same network, the selective recovery or pathological switching cannot be imputed to the disruption of the tissues where the languages are represented but rather to a deficit of the control system. Experiments using the switching paradigm and data from bilingual aphasia have reported a stronger activation of the frontal area for bilinguals compared to monolinguals (Khateb et al., 2007; Rodriguez-Fornells et al., 2005; Rodriguez-Fornells et al., 2002) and for bilinguals tested in monolingual and mixed conditions (Abutalebi et al., 2007). When asked to name in L1 a picture in a mixed condition they showed a larger frontal activation than when naming the same picture in L1 in a monolingual context. Specifically, a more extended activation was found for the left prefrontal cortex, the anterior cingulated cortex and the left caudate nucleus (Crinion et al., 2006). The activation was even larger when testing the less proficient L2 in a mixed context. These findings suggest that prefrontal regions that are engaged in controlling competing behavioral responses are also responsible for language switching and selection. The increasing activation in a mixed context can indicate the effort to inhibit the non-target language especially when it is the stronger L1 and requires a higher switching cost (Meuter & Allport, 1999).

Data from aphasia Data from aphasia reporting pathological switching and mixing and selective recovery or fixation, that is, the incapability to switch from one language to another (see Fabbro, 1999, for a complete review) do corroborate the hypothesis that the lesion disrupts the selection/control system, that is, the activation/inhibition system, and not the representation of the languages. There are cases of pathological switching reported in literature caused by a lesion to the prefrontal area (Leemann et al., 2007; Marien et al., 2005; Abutalebi et al., 2000; Fabbro et al., 2000) and fixation (Aglioti et al., 1996; Aglioti & Fabbro, 1993) suffered from a lesion to the left prefrontal area, the basal ganglia, the left anterior cingulated, the head of the left caudate and the inferior parietal cortex. The dynamic view allows interpreting not only the two extreme patterns of recovery (i.e. pathological switching and fixation) but also the other varieties of recovery. A parallel recovery would involve balanced recovery of both the mechanisms of activation and inhibition: the non-target language gets the right amount of resources to be inhibited. A selective recovery would occur when the inhibition threshold of one language is raised so high by the lesion that the system cannot have any more control over the selection of that language. Pathological mixing and switching

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

153

would result from an impossibility of the control network to inhibit the languages selectively.

Conclusion Nowadays most of the population in the world uses more than one language in their daily life. As a direct consequence, the number of bilingual individuals with aphasia is proportionally increasing. According to Paradis (2001) on the basis of census data, 45,000 new cases are calculable every year in the US. Considering the flow of immigration that Europe has been experiencing in the last 15 years and the consequent increasing percentage of bilingual individuals in society, a remarkable number of these people are annually expected to suffer from aphasia. Research findings about the organization of multiple languages in the brain and on the mechanisms of their selection/control system are more and more urgent. This knowledge would facilitate predictions on the recovery pattern following a given lesion site and consequently setting a protocol of intervention to rehabilitate the language deficit. The first warning is for professionals and researchers to test the patients in a strictly monolingual environment, including monolingual examiners, to verify that the CS/CM shown by the patient is really part of the aphasic deficit and not a response to the examiner’s behavior or a strategy performed by the patient to make up for difficulties of retrieval in the tested language. To date data from behavioral and neuroimaging studies on both healthy populations and those with aphasia seem to support the view that all the languages are represented in the same anatomical network. Neuroimaging reveals more extended areas of engagement for processing L2 than L1 but this appears to be true only at the beginning of the L2 learning process; longitudinal studies, in fact, show a convergence between L1 and L2 networks as long as L2 proficiency increases. Another crucial finding seems to be the functional and anatomical independence between the language network and the selection/control network. This distinction would better explain cases such as that reported in Fabbro et al. (2000) where a patient presenting a tumor of the white matter in the left frontal lobe involving the anterior cingulate did not exhibit any aphasic symptoms either in L1 and L2 after the operation but manifested pathological switching. Disruption at the selection/control system correlates with lesions to the left prefrontal cortex, inferior parietal cortex, anterior cingulated cortex, left head caudate and basal ganglia. These circuits and mainly the prefrontal cortex are responsible for cognitive executive control, in general, while the basal ganglia are traditionally associated with motor control. They are engaged in monitoring and selecting competing behavioral responses. The task that a bilingual speaker has to face to select a language would not be any different from any other behavioral cognitive task where two conflicting

154

Par t 3: Bilingual L anguage Phenomena

alternatives compete to achieve the goal, therefore there would be no need to postulate the existence of a specific language switching device.

Notes (1) From now on for simplicity we will also call polyglot people, that is, people who speak more than two languages, bilingual. (2) Following the terminology proposed by Ritchie and Bhatia (1996), we will refer to CS as interphrasal language switch and to CM as intraphrasal switch. (3) For an exhaustive review of the literature on aphasia in bilinguals and polyglots, see Paradis (1977). (4) For the classification of the patterns of recovery in bilingual aphasics after a brain injury see Paradis (1977). (5) The Microgenetic Model hypothesizes a prelinguistic conceptual level with languageindependent and different, separate grammars for all the languages known by the speaker. (6) Interferences are called static when they are elements from all levels of one language fossilized in the other (for instance, a foreign accent); they are called dynamic when they occur in the base language of the moment only occasionally.

References Abutalebi, J. (2008) Neural aspects of second language representation and language control. Acta Psychologica 128, 466–478. Abutalebi, J. and Green, D. (2007) Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics 20, 242–275. Abutalebi, J., Miozzo, A. and Cappa, S.F. (2000) Do subcortical structures control language selection in bilinguals? Evidence from pathological language mixing. Neurocase 6, 101–106. Abutalebi, J., Annoni, J.M., Zimine, I., Pegna, A.J., Seghier, M.L., Lee Jahnke, H., Lazeyras, F., Cappa, S.F. and Khateb, A. (2007) Language control and lexical competition in bilinguals: An event-related fMRI study. Cerebral Cortex 18, 1496–1505. Aglioti, S. and Fabbro, F. (1993) Paradoxical selective recovery in a bilingual aphasic following subcortical lesion. Neuroreport 4, 1359–1362. Aglioti, S., Beltramello, A., Girardi, F. and Fabbro, F. (1996) Neurolinguistic and follow-up study of an unusual pattern of recovery from bilingual subcortical aphasia. Brain 119, 1551–1564. Albert, M.L. and Obler, L.K. (1978) The Bilingual Brain. New York: Academic Press. Brown, J.W. (1979) Language representation in the brain. In H.D. Steklis and M. Raleigh (eds) The Neurobiology of Social Communication in Primates (pp. 29–68). New York: Academic Press. Costa, A. and Caramazza, A. (1999) Is lexical selection in bilinguals language-specific? Further evidence from Spanish–English bilinguals and English–Spanish bilinguals. Bilingualism: Language and Cognition 2, 231–244. Costa, A. and Santesteban, M. (2006) The control of speech production by bilingual speakers. Bilingualism: Language and Cognition 9, 115–212 [Special Issue]. Crinion, J., Turner, R., Grogan, A., Hanakawa, T., Noppeney, U., Devlin, J.T., Aso, T., Urayama, S., Fukuyama, H., Stockton, K., Usui, K., Green, D.W. and Price, C.J. (2006) Language control in the bilingual brain. Science 312, 1537–1540. De Bot, K. and Schreuder, R. (1993) Word production and the bilingual lexicon. In R. Schreuder and B. Weltens (eds) The Bilingual Lexicon (pp. 191–214). Amsterdam: John Benjamins.

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

155

De Groot, A.M.B., Delmaar, P. and Lupker, S. (2000) The processing of interlexical homographs in translation recognition and lexical decision: Support for non-selective access to bilingual memory. Quarterly Journal of Experimental Psychology 52 (2), 397–428. Dijkstra, A. and Van Heuven, W.J.B. (1998) The BIA model and bilingual word recognition. In J. Grainger and A.M. Jacobs (eds) Localist Connectionist Approaches to Human Cognition (pp. 189–225). Mahwah, NJ: Erlbaum. Dijkstra, T., van Jaarsveld, H. and ten Brinke S. (1998) Interlingual homograph recognition: Effects of task demands and language intermixing. Bilingualism: Language and Cognition 1, 51–66. Fabbro, F. (1999) The Neurolinguistics of Bilingualism: An Introduction. Hove, UK: Psychology Press Ltd. Fabbro, F., Skrap, M. and Aglioti, S. (2000) Pathological switching between languages after frontal lesions in a bilingual patient. Journal of Neurology, Neurosurgery and Psychiatry, 68, 650–652. Galloway, L. and Krashen, S.D. (1980) Cerebral organization in bilingualism and second language. In R.C. Scarcella and S.D. Krashen (eds) Research in Second Language Acquisition (pp. 74–80). Rowley, MA: Newbury House. Gloning, I. and Gloning, K. (1965) Aphasien bei Polyglotten: Beitrag zur Dynamik des Sprachabbaus sowie zur Lokalisationsfrage dieser Ströngen. Wiener Zeitschrift für Nervenheilkunde, 22 (translated in Paradis, 1983: 681–716). Golestani, N., Alario, F-X., Meriaux, S., Le Bihan, D., Dehaene, S. and Pallier, C. (2006) Syntax production in bilinguals. Neuropsychologia 44, 1029–1040. Green, D.W. (1986) Control, activation and resource: A framework and a model for the control of speech in bilinguals. Brain and Language 27, 210–223. Green, D.W. (1998) Mental control of the bilingual lexico-semantic system. Bilingualism: Language and Cognition 1 (2), 67–81. Green, D.W. (2003) The neural basis of the lexicon and the grammar in L2 acquisition: The convergence hypothesis. In R. van Hout, A. Hulk, F. Kuiken and R. Towell (eds) The Interface between Syntax and the Lexicon in Second Language Acquisition (pp. 197–218). Amsterdam: John Benjamins. Green, D.W., Crinion, J. and Price, C.J. (2006) Convergence, degeneracy and control. Language Learning 56, 99–125. Grosjean, F. (1985) Polyglot aphasics and language mixing: A comment on Perecman (1984). Brain and Language 26, 349–355. Grosjean, F. (1989) Neurolinguistis, beware! The bilingual is not two monolinguals in one person. Brain and Language 36, 3–15. Grosjean, F. (1997) Processing mixed languages: Issues, findings and models. In A.M.B. De Groot and J.F. Kroll (eds) Tutorials in Bilingualism: Psycholinguistic Perspectives (pp. 225–254). Mahwah, NJ: Lawrence Erlbaum. Indefrey, P. (2006) A meta-analysis of hemodynamic studies on first and second language processing: Which suggested differences can we trust and what do they mean? Language Learning 56, 279–304. Indefrey, P., Hellwig, F., Davidson, D. and Gullberg, M. (2005) Native-like hemodynamic responses during sentence comprehension after six months of learning a new language. Poster presented at the 11th Annual Meeting of the Organization for Human Brain Mapping, Toronto, Canada. Jeong, H., Sugiura, M., Sassa, Y., Yokoyama, S., Horie, K. and Sato, S. (2007) Crosslinguistic influence on brain activation during second language processing: An fMRI study. Bilingualism: Language and Cognition 10, 175–187. Kandel, E.R., Schwartz, J.H. and Jessel, T.M. (1991) Principles of Neural Science. New York: Elsevier.

156

Par t 3: Bilingual L anguage Phenomena

Khateb, A., Abutalebi, J., Michel, C.M., Pegna, A.J., Lee-Jahnke, H. and Annoni, J-M. (2007) Language selection in bilinguals: A spatio-temporal analysis of electric brain activity. International Journal of Psychophysiology 65, 201–213. Kim, K.M., Relkin, N.R., Lee, K.M. and Hirsch, J. (1997) Distinct cortical areas associated with native and second languages. Nature 388, 171–174. Kroll, F.J., Bobb, S.C., Misra, M. and Guo, T. (2008) Language selection in bilingual speech: Evidence for inhibitory processes. Acta Psychologica 128, 416–430. Leemann, B., Laganaro, M., Schwitter, V. and Schnider, A. (2007) Paradoxical switching to a barely-mastered second language by an aphasic patient. Neurocase 13, 209–213. L’Hermitte, R., Hécaen, H., Dubois, J., Culioli, A. and Tabouret-Keller, A. (1966) Le Problème de l’Aphasie des Polyglottes: Remarques sur quelques observations. Neuropsychologia 4, 315–329. Macnamara, J., Krauthhammer, M. and Bolgar, M. (1968) Language switching in bilinguals as a function of stimulus and response uncertainty. Journal of Experimental Psychology 78, 208–215. Mägiste. E. (1979) The competing language systems of the multilingual: A developmental study of decoding and encoding processes. Journal of Verbal Learning and Verbal Behavior 18, 79–89. Marien, P., Abutalebi, J., Engelborghs, S. and De Deyn, P.P. (2005) Pathophysiology of language switching and mixing in an early bilingual child with subcortical aphasia. Neurocase 11, 385–398. Mechelli, A., Crinion, J.T., Noppeney, U., O’Doherty, J., Ashburner, J., Frackowiack, R. S. and Price, C.J. (2004) Neurolinguistics: Structural plasticity in the bilingual brain. Nature 431, 757. Menn, L. and Obler, L.K. (1982) Exceptional language data as linguistic evidence: An introduction. In L. Menn and L.K. Obler (eds) Exceptional Language and Linguistics (pp. 3–14). New York: Academic Press. Meuter, R.F.I. and Allport, A. (1999) Bilingual language switching in naming: Asymmetrical costs of language selection. Journal of Memory and Language 40, 25–40. Minkowski, M. (1983) A clinical contribution to the study of polyglot aphasia especially with respect to Swiss-German. In M. Paradis (ed.) Readings on Aphasia in Bilinguals and Polyglots (pp. 205–232). Montreal: Didier (original work published 1927). Moreno, M.M., Rodriguez-Fornells, A. and Laine, M. (2008) Event-related potentials (ERPs) in the study of bilingual language processing. Review. Journal of Neurolinguistics 21, 477–508. Muñoz, M.L., Marquandt, T.P. and Copeland, G. (1999) A comparison of the codeswitching patterns of aphasic and neurologically normal bilingual speakers of English and Spanish. Brain and Language 66, 249–274. Muysken, P. (1995) Codeswitching and grammatical theory. In L. Milroy and P. Muysken (eds) One Speaker, Two Languages (pp. 177–198). Cambridge: Cambridge University Press. Myers-Scotton, C. (1993) Duelling Languages: Grammatical Structure in Codeswitching. Oxford: Oxford University Press (Clarendon Press). Norman, D. and Shallice, T. (1980) Attention to Action: Willed and Automatic Control of Behavior. San Diego: Center for Human Information Program, University of California. Paradis, M. (1977) Bilingualism and aphasia. In H. Whitaker and H. Whitaker (eds) Studies in Neurolinguistics (Vol. 3, pp. 65–121). New York: Academic Press. Paradis, M. (1987) The neurofunctional modularity of cognitive skills: Evidence from Japanese alexia and polyglot aphasia. In E. Keller and M. Gopnik (eds) Motor and Sensory Processes of Language (pp. 277–289). Hillsdale, NJ: Lawrence Erlbaum. Paradis, M. (1994) Neurolinguistic aspects of implicit and explicit memory: Implications for bilingualism. In N. Ellis (ed.) Implicit and Explicit Learning of Second Languages (pp. 393–419). London: Academic Press.

Bilingual Aphasia and Code-Sw itching : Represent at ion and Control

157

Paradis, M. (1997) The cognitive neuropsychology of bilingualism. In A. De Groot and J. Kroll (eds) Tutorials in Bilingualism: Psycholinguistic Perspectives (pp. 331–354). Hillsdale, NJ: Lawrence Erlbaum. Paradis, M. (2001) Bilingual and polyglot aphasia. In R.S. Berndt (ed.) Handbook of Neuropsychology (2nd edn, Vol. 3, pp. 69–91) Language and Aphasia. Amsterdam: Elsevier Science. Paradis, M. (2004) A Neurolinguistic Theory of Bilingualism. Philadelphia, PA: John Benjamins. Paradis, M., Libben, G. and Hummel, K. (1987) Bilingual Aphasia Test. Hillsdale, NJ: L.E.A. Perani, D. and Abutalebi, J. (2005) The neural basis of first and second language processing. Current Opinion in Neurobiology 15, 202–206. Perecman, E. (1984) Spontaneous translation and language mixing in a polyglot aphasic. Brain and Language 23, 43–63. Poplack, S. (1980) Sometimes I’ll start a sentence in English y Termino en Español: Towards a typology of code-switching. Linguistics 18, 581–618. Poplack, S. and Meechan, M. (1998) How languages fit together in code-mixing. International Journal of Bilingualism 2(2), 127–138. Pötzl, O. (1925) Uber die parietal bedingte Aphasie und ihren Einfluss auf die Sprechen mehrerer Sprachen. Zeitschrift fuer die Gesamte Neurologie und Psychiatric 96, 100–124. Proverbio, A.M., Leoni, G. and Zani, A. (2004) Language switching mechanisms in simultaneous interpreters: An ERP study. Neuropsychologia 42, 1636–1656. Rodriguez-Fornells, A., Rotte, M., Heinze, H.J., Noesselt, T. and Muente, T.F. (2002) Brain potential and functional MRI evidence for how to handle two languages with one brain. Nature 415, 1026–1029. Rodriguez-Fornells, A., van der Lugt, A., Rotte, M., Britti, B., Heinze, H.J. and Muente, T.F. (2005) Second language interferes with word production in fluent bilinguals: Brain potential and functional imaging evidence. Journal of Cognitive Neuroscience 17, 422–433. Sakai, K.L., Miura, K., Narafu, N. and Muraishi, Y. (2004) Correlated functional changes of the prefrontal cortex in twins induced by classroom education of second language. Cerebral Cortex 14, 1233–1239. Scoresby-Jackson, R.E. (1867) Case of aphasia with right hemiplegia. Edinburgh Medical Journal 12, 696–706. Shallice, T. (1982) Specific impairments in planning. In D. Broadbent and L. Weiskrantz (eds) The Neuropsychology of Cognitive Function (pp. 199–209). London: The Royal Society. Ullman, M.T. (2001a) The neural basis of lexicon and grammar in first and second language: The declarative/procedural model. Bilingualism: Language and Cognition 4, 105–122. Ullman, M.T. (2001b) A neurocognitive perspective on language: The declarative/procedural model. Nature Reviews Neuroscience 2, 717–726. Zurif, E.B. and Blumstein, S. (1978) Language and the brain. In M. Halle, J. Bresnan and G. Miller (eds) Linguistic Theory and Psychological Reality (pp. 229–240). Cambridge, MA: MIT Press.

10 Grammatical Category Deficits in Bilingual Aphasia Yasmeen Faroqi-Shah

Difficulty in retrieving words is the most frequent linguistic symptom of aphasia (see Muñoz, Chapter 5, this volume, also for treatment of lexical retrieval). There are numerous reports of a proportionally greater difficulty with specific semantic or grammatical categories compared to others. A variety of such category specific deficits have been described, including but not limited to, deficits for proper names, tools, animals, open class words and abstract words. Over the past three decades in aphasiology, noun versus verb retrieval deficits are the most frequently reported instances of category specific deficits in monolingual persons (e.g. Berndt et al., 1997; Conroy et al., 2006; Kemmerer & Tranel, 2000; Zingeser & Berndt, 1990). An obvious contribution of category specific deficits is the insight they provide into the neurocognitive architecture of language representation in the adult human brain. The logic is that if two word categories are observed to dissociate consequent to focal brain damage, then these must be represented in (at least partially) distinct neural substrates. This interpretation of neurally distinct substrates for each word category is strengthened by the demonstration of double dissociations: patients with opposing patterns of selective impairment in one, but not the other word category (Shallice, 1988; Teuber, 1955). Given that nouns and verbs are universally distinct grammatical categories, noun-verb dissociations have generated considerable interest among neurolinguists and are taken as evidence that lexical organization in the brain is governed by grammatical class (as will be discussed later, several authors have questioned this assumption) (e.g. Caramazza & Hillis, 1991; McCarthy & Warrington, 1985; Zingeser & Berndt, 1990). Whether the noun-verb dissociation is manifested across languages in bi(multi)lingual1 speakers with aphasia is a relatively less explored question (but see Miozzo et al., 2010 for a review). If bilingual persons show parallel noun-verb dissociation in both their languages, it provides additional insight into the nature of lexical organization in bilingual brains. For instance, such a pattern may suggest that grammatical and/or semantic features are shared across languages. 158

Grammat ical Categor y Def ic it s in Bilingual Aphasia

159

The primary focus of this chapter is to provide an overview of the literature on noun-verb deficits in bilingual persons with aphasia. First, patterns of noun-verb deficits in monolingual aphasia are briefly discussed to provide a broader context for interpreting bilingual noun-verb deficits. Linguistic differences between nouns and verbs that serve as explanations for noun-verb dissociations are also considered. This is followed by a review and discussion of lexical deficits in bilingual aphasia. Finally, implications for clinical assessment of bilingual individuals with aphasia are discussed.

Differences between Nouns and Verbs A key grammatical difference between nouns and verbs is that verbs represent syntactically relevant information, such as subcategorization, which refers to words or phrases that co-occur in a sentence. Most verbs have multiple possible subcategorization frames which adds to their processing complexity in unimpaired individuals and those with aphasia (Grimshaw, 2000; Kim & Thompson, 2000, 2004). For example, the verb eat can appear in intransitive, transitive and passive frames (as in ‘We ate’, ‘She ate the banana’ and ‘The pie was eaten’). Although both nouns and verbs are morphologically marked in most languages, verbs tend to be morphologically more complex than nouns, often inflecting for a wider range of parameters such as tense, aspect, number, gender and mood (Shapiro & Caramazza, 2003; Vigliocco et al., 2006). Morphosyntactic differences aside, nouns and verbs also differ in their core conceptual properties. Nouns refer to entities that can be counted or individuated (O’Grady, 1997). This holds true not only for nouns referring to concrete objects, but also for mass (e.g. this water) and abstract (e.g. these facts) nouns. In contrast, verbs refer to events and actions, which are temporally transient (O’Grady, 1997). Differences between prototypical nouns (i.e. objects) and prototypical verbs (i.e. agentive actions) are evident along graded dimensions of imageability and concreteness, both of which are higher for nouns (Bird et al., 2000; Conroy et al., 2006). Nouns tend to have more sensory-perceptual attributes, while verbs tend to have higher functional-motoric associations (McCarthy & Warrington, 1985; Warrington & Shallice, 1984). Further, verbs are more likely to be polysemous than nouns (Gentner, 1981; Miller & Fellbaum, 1998). For example, to break not only refers to ‘cause something to become into fragments’ in transitive and intransitive senses, but also ‘to diminish’ (His fever broke), ‘to better’ (She broke the record), to go against (She broke the law) and over 20 other senses (Fellbaum, 1998). Another difference is that concrete nouns are known to be hierarchically organized into semantic categories and subcategories (e.g. animal → dog → poodle), while verbs share fewer and less transparent semantic features (Collins & Quinlan, 1969; Vinson & Vigliocco, 2002).

160

Par t 3: Bilingual L anguage Phenomena

Early reports of double dissociations of noun and verb retrieval in aphasia were interpreted as evidence of distinct neurocognitive networks for these two grammatical categories (e.g. Caramazza & Hillis, 1991; McCarthy & Warrington, 1985; Zingeser & Berndt, 1990). More recent evidence has consistently revealed that verb deficits are far more frequent than noun deficits in aphasia (Luzzatti et al., 2001; Mätzig et al., 2009). A meta-analysis of 280 published cases of noun-verb retrieval in aphasia conducted by Mätzig et al. (2009) revealed that 75% of all cases reported worse performance with verbs than nouns. This greater susceptibility of verbs to brain damage occurs despite the existence of numerous highly frequent verbs compared to nouns (e.g. see, look, do, make). Interestingly, when sensitive measures such as naming latencies are used, unimpaired participants also demonstrate greater difficulty with verbs compared to nouns (Bogka et al., 2003; Szekely et al., 2004). The greater vulnerability of verbs in unimpaired persons and those with aphasia (as well as other populations) suggests that verbs may be inherently more difficult than nouns, at least in terms of the cognitive processes required in typical experimental tasks such as picture naming (Mätzig et al., 2009). In order to control for inherent differences in imageability, concreteness and morphological complexity between nouns and verbs, a few studies controlled for one or more of these variables, or used experimental tasks such as sentence production (as opposed to picture naming) which impose greater parity along these variables (e.g. Bird et al., 2000, 2003; Crepaldi, et al., 2006; Luzzatti et al., 2001; Shapiro et al., 2000). For example, Crepaldi and colleagues (2006) compared verb and noun retrieval in picture naming versus sentence contexts in 16 Italian speakers with aphasia. In sentence contexts, imageability differences between nouns and verbs are minimized because both word types have to be retrieved from memory rather than from visually presented stimuli. Noun-verb differences were eliminated in 14 out of 16 participants in the sentence context. Other studies have found that, in at least a subset of participants with aphasia, a true verb deficit persisted beyond these experimental controls. That is, there are complex interactions between grammatical class, conceptual-semantic variables and experimental task that need to be considered while interpreting the cognitive implications of noun-verb dissociations (see also Bird et al., 2003; Mätzig et al., 2009). It must be pointed out that nouns are not always easier than verbs. In fact, as mentioned earlier, there are individuals with aphasia who are selectively more impaired for nouns than verbs (Caramazza & Hillis, 1991; Hillis & Caramazza, 1995; Zingeser & Berndt, 1990). Populations in which noun impairments are consistently reported include semantic dementia and dementia of Alzheimer’s type (e.g. Bird et al., 2000; Crutch & Warrington, 2008; Silveri et al., 2003). Often, noun disadvantaged patients demonstrate impairments in semantic or sensory aspects that are more crucial for noun retrieval compared to verb retrieval. These include impairments in object

Grammat ical Categor y Def ic it s in Bilingual Aphasia

161

knowledge, visual processing or modality specific representations (written versus oral) (e.g. Bird et al., 2000; Druks & Shallice, 2000). The left temporal lobe is involved in cognitive operations that are heavily relied on for noun retrieval. And, it is consistently found that selective noun impairments are associated with left temporal lobe damage (see Mätzig et al., 2009, for a review). In summary, several processing distinctions between nouns and verbs have been identified to explain why these two grammatical categories dissociate in aphasia. The greater propensity for verb deficits in aphasia and longer processing times for verbs by unimpaired individuals undermine arguments of a true double dissociation of grammatical categories, unless the numerous variables that distinguish nouns and verbs can be matched (e.g. imageability, frequency of usage, age of acquisition, morphological structure, polysemy and visual knowledge). Several authors argue that smaller magnitudes of verb disadvantage (e.g.