Scholarly article on topic 'Unravelling the Genetics of Ischaemic Stroke'

Unravelling the Genetics of Ischaemic Stroke Academic research paper on "Languages and literature"

CC BY
0
0
Share paper
Academic journal
PLoS Med
OECD Field of science
Keywords
{""}

Academic research paper on topic "Unravelling the Genetics of Ischaemic Stroke"

The Handbook of Language Teaching

The Handbook of Language Teaching Edited by Michael H. Long and Catherine J. Doughty © 2009 Blackwell Publishing Ltd. ISBN: 978-1-405-15489-5

Blackwell Handbooks in Linguistics

This outstanding multi-volume series covers all the major subdisciplines within linguistics today and, when complete, will offer a comprehensive survey of linguistics as a whole.

Already published:

The Handbook of Child Language

Edited by Paul Fletcher and Brian MacWhinney

The Handbook of Phonological Theory Edited by John A. Goldsmith

The Handbook of Contemporary Semantic Theory Edited by Shalom Lappin

The Handbook of Sociolinguistics Edited by Florian Coulmas

The Handbook of Phonetic Sciences

Edited by William J. Hardcastle and John Laver

The Handbook of Morphology

Edited by Andrew Spencer and Arnold Zwicky

The Handbook of Japanese Linguistics

Edited by Natsuko Tsujimura

The Handbook of Linguistics

Edited by Mark Aronoff and Janie Rees-Miller

The Handbook of Contemporary Syntactic Theory

Edited by Mark Baltin and Chris Collins

The Handbook of Discourse Analysis

Edited by Deborah Schiffrin, Deborah Tannen, and Heidi E. Hamilton

The Handbook of Language Variation and Change

Edited by J. K. Chambers, Peter Trudgill, and Natalie Schilling-Estes

The Handbook of Historical Linguistics

Edited by Brian D. Joseph and Richard D. Janda

The Handbook of Language and Gender

Edited by Janet Holmes and Miriam Meyerhoff

The Handbook of Second Language Acquisition

Edited by Catherine J. Doughty and Michael H. Long

The Handbook of Bilingualism

Edited by Tej K. Bhatia and William C. Ritchie

The Handbook of Pragmatics

Edited by Laurence R. Horn and Gregory Ward

The Handbook of Applied Linguistics Edited by Alan Davies and Catherine Elder

The Handbook of Speech Perception

Edited by David B. Pisoni and Robert E. Remez

The Blackwell Companion to Syntax, Volumes I-V

Edited by Martin Everaert and Henk van Riemsdijk

The Handbook of the History of English

Edited by Ans van Kemenade and Bettelou Los

The Handbook of English Linguistics

Edited by Bas Aarts and April McMahon

The Handbook of World Englishes

Edited by Braj B. Kachru, Yamuna Kachru, and Cecil L. Nelson

The Handbook of Educational Linguistics Edited by Bernard Spolsky and Francis M. Hult

The Handbook of Clinical Linguistics

Edited by Martin J. Ball, Michael R. Perkins, Nicole Muller, and Sara Howard

The Handbook of Pidgin and Creole Studies

Edited by Silvia Kouwenberg and John Victor Singler

The Handbook of Language Teaching

Edited by

Michael H. Long and Catherine J. Doughty

WILEY-BLACKWELL

A John Wiley & Sons, Ltd., Publication

This edition first published 2009

© 2009 Blackwell Publishing Ltd except for editorial material and organization © 2009 Michael H. Long and Catherine J. Doughty

Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell's publishing program has been merged with Wiley's global Scientific, Technical, and Medical business to form Wiley-Blackwell.

Registered Office

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom

Editorial Offices

350 Main Street, Malden, MA 02148-5020, USA

9600 Garsington Road, Oxford, OX4 2DQ, UK

The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

For details of our global editorial offices, for customer services, and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.

The right of Michael H. Long and Catherine J. Doughty to be identified as the authors of the editorial material in this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

The handbook of language teaching / edited by Michael H. Long and Catherine J. Doughty.

p. cm. — (Blackwell handbooks in linguistics) ISBN 978-1-4051-5489-5 (hardcover : alk. paper) 1. Language and languages— Study and teaching—Handbooks, manuals, etc. 2. Second language acquisition— Handbooks, manuals, etc. I. Long, Michael H. II. Doughty, Catherine. P51.H3265 2009 418.0071—dc22

2009008867

A catalogue record for this book is available from the British Library.

Set in 10/12 Palatino by Graphicraft Ltd, Hong Kong Printed in Singapore

1 2009

Craig Chaudron (1946-2006)

Contents

List of Contributors x

Part I Overview 1

1 Language Teaching 3 Michael H. Long

Part II Social, Political, and Educational Contexts of

Language Teaching 7

2 The Social and Sociolinguistic Contexts of Language Learning

and Teaching 9

Sandra Lee McKay and Rani Rubdy

3 The Politics and Policies of Language and Language Teaching 26 Robert Phillipson and Tove Skutnabb-Kangas

4 History of Language Teaching 42 Diane Musumeci

Part III Psycholinguistic Underpinnings of Language Learning 63

5 The Language-Learning Brain 65 Alan Beretta

6 Sequences and Processes in Language Learning 81 Lourdes Ortega

7 The Importance of Cross-Linguistic Similarity in Foreign

Language Learning 106

Hakan Ringbom and Scott Jarvis

8 Cognitive-Psychological Processes in Second Language Learning 119 Robert M. DeKeyser

9 Optimizing the Input: Frequency and Sampling in Usage-Based

and Form-Focused Learning 139

Nick C. Ellis

Part IV Program Design 159

10 Bilingual and Immersion Programs 161 Jim Cummins

11 Heritage Language Programs 182 Silvina Montrul

12 Specific Purpose Programs 201 Ken Hyland

13 Study Abroad Research: Findings, Implications, and Future

Directions 218

Joseph Collentine

14 Less Commonly Taught Languages: Issues in Learning and

Teaching 234

Kira gor and Karen Vatz

15 Third Language Acquisition Theory and Practice 250 William P. Rivers and Ewa M. Golonka

Part V Course Design and Materials Writing 267

16 Foreign and Second Language Needs Analysis 269 James Dean Brown

17 Syllabus Design 294 Peter Robinson

18 Advances in Materials Design 311 Alan Waters

19 Corpora in Language Teaching 327 John Flowerdew

20 Technology-Enhanced Materials 351 David Brett and Marta Gonzalez-Lloret

Part VI Teaching and Testing 371

21 Methodological Principles for Language Teaching 373 Michael H. Long

22 Teaching and Testing Listening Comprehension 395 Larry Vandergrift and Christine Goh

23 Teaching and Testing Speaking 412 Martin Bygate

24 Teaching and Testing Reading 441 William Grabe

25 Learning to Read in New Writing Systems 463 Keiko Koda

26 Teaching and Testing Writing 486 Charlene Polio and Jessica Williams

27 Teaching and Testing Grammar 518 Diane Larsen-Freeman

28 Teaching and Testing Vocabulary Paul Nation and Teresa Chung 543

29 Teaching and Testing Pragmatics Carsten Roever 560

30 Task-Based Teaching and Testing John M. Norris 578

31 Radical Language Teaching Graham Crookes 595

32 Diagnostic Feedback in Language Assessment Antony John Kunnan and Eunice Eunhee Jang 610

33 Computer-Assisted Teaching and Testing Carol A. Chapelle 628

Part VII Teacher Education 645

34 Language Teacher Education 647 Renee Jourdenais

35 Diffusion and Implementation of Innovations 659 Kris Van den Branden

Part VIII Assessing and Evaluating Instruction 673

36 Current Trends in Classroom Research 675 Rosamond F. Mitchell

37 Issues in Language Teacher Evaluation 706 Kathleen M. Bailey

38 Investigating the Effects and Effectiveness of L2 Instruction 726 Rick de Graaff and Alex Housen

39 Program Evaluation 756 Steven J. Ross

Author Index 779

Subject Index 791

Contributors

Kathleen M. Bailey

Kathleen M. Bailey received her PhD from the University of California at Los Angeles. She is a professor of Applied Linguistics at the Monterey Institute of International Studies, where she has taught since 1981. In 1998-99 she was the President of the international TESOL association.

Alan Beretta

Alan Beretta is Professor of Linguistics at Michigan State University. His research is in neurolinguistics and has been published in such journals as Brain and Language, Cognitive Brain Research, and Aphasiology.

David Brett

David Brett worked in Italy as an ESL teacher for 10 years before becoming a researcher in English Linguistics at the University of Sassari. He has published and presented widely on New Technologies and Second Language Learning, with particular reference to pronunciation teaching. He has also held training workshops for language teachers on various aspects of technology-enhanced teaching, both in Italy and in other countries.

James Dean Brown

James Dean ("JD") Brown is Professor of Second Language Studies at the University of Hawai'i at Manoa. He has authored or co-authored numerous articles and books on topics as diverse as second language testing and quantitative research methods, language curriculum development, using surveys in language programs, teaching connected speech, and heritage language curriculum.

Martin Bygate

Martin Bygate is Professor in Applied Linguistics and Language Education at Lancaster University, UK. He has undertaken funded research and taught courses on oral language teaching and development. Principal publications are Speaking

(1987, Oxford University Press), Grammar and the Language Teacher (co-edited with A. Tonkyn and E. Williams, 1994, Prentice-Hall), Researching pedagogic tasks: Second language learning, teaching and testing (co-edited with P. Skehan & M. Swain, 2001, Pearson Educational Ltd), and, co-authored with Virginia Samuda, Tasks in second language learning (2008, Palgrave).

Carol A. Chapelle

Carol A. Chapelle, Professor of TESL/Applied Linguistics at Iowa State University, is Past President of the American Association for Applied Linguistics (2006-7), former editor of TESOL Quarterly (1999-2004), and co-editor of the Cambridge Applied Linguistics Series. Her books include Computer applications in second language acquisition: Foundations for teaching, testing, and research (2001, Cambridge University Press), English language learning and technology: Lectures on applied linguistics in the age of information and communication technology (2003, John Benjamins), Assessing language through technology (with Dan Douglas, 2006, Cambridge University Press), Building a validity argument for the Test of English as a Foreign Language (with Mary Enright & Joan Jamieson, 2007, Routledge) and Tips for teaching with CALL (2008, Pearson-Longman).

Teresa Chung

Mihwa Chung (Teresa) teaches at Korea University. She has published articles on technical vocabulary, the vocabulary of newspapers, and developing reading speed in a foreign language. Her PhD thesis from Victoria University of Wellington was on the methodology of developing lists of technical vocabulary and the role of technical vocabulary in technical texts.

Joseph Collentine

Joseph Collentine is Professor of Spanish at Northern Arizona University. He has published articles and research about study abroad, the acquisition of grammar, and corpus linguistics. He is currently the director of the Spanish Masters programs at NAU and the coordinator of the Spanish online program.

Graham Crookes

Graham Crookes is Professor, Department of Second Language Studies, University of Hawai'i at Manoa, where he is also Executive Director, ESL Programs. His most recent books are A Practicum in TESOL and Making a Statement: Values, Philosophies, and Professional Beliefs in TESOL (2003 and 2008, Cambridge University Press).

Jim Cummins

Jim Cummins is Professor and Canada Research Chair in the Curriculum, Teaching and Learning Department at the Ontario Institute for Studies in Education (OISE) at the University of Toronto. His research focuses on literacy development in multilingual school contexts, as well as on the potential roles of technology in promoting language and literacy development.

Robert DeKeyser

Robert DeKeyser (PhD, Stanford University) is Professor of Second Language Acquisition at the University of Maryland. His research is mainly on second language acquisition, with emphasis on cognitive-psychological aspects such as implicit versus explicit learning, automatization of rule knowledge, and individual differences and their interaction with instructional treatments. He has published in a variety of journals, including Studies in Second Language Acquisition, Language Learning, Language Testing, The Modern Language Journal, TESOL Quarterly, and AILA Review. He has contributed chapters to several highly regarded handbooks, and he recently published an edited volume with Cambridge University Press entitled Practice in a Second Language: Perspectives from Applied Linguistics and Cognitive Psychology (2007).

Catherine J. Doughty

Catherine J. Doughty is Senior Research Scientist and SLA Area Director at the Center for the Advanced Study of Language at the University of Maryland, and is an affiliate Professor of SLA at the University of Maryland.

Nick C. Ellis

Nick C. Ellis is Research Scientist at the English Language Institute and Professor of Psychology at the University of Michigan. His research interests include language acquisition, cognition, reading in different languages, corpus linguistics, cognitive linguistics, psycholinguistics, and emergentist accounts of language acquisition.

John Flowerdew

John Flowerdew is Professor of Applied Linguistics, Centre for Language Education Research, School of Education, University of Leeds. For many years he worked at the City University of Hong Kong. He has also worked in South America and the Middle East. As well as writing and editing a number of books, he has published widely in the leading Applied Linguistics, Language Teaching and Discourse Analysis journals, focusing on academic discourse, corpus linguistics, and English for Specific Purposes. His most recent book (with Lindsay Miller) is Second Language Listening (2005, Cambridge University Press). His most recent edited book (with Vijay Bhatia and Rodney Jones) is Advances in Discourse Studies (2008, Routledge).

Christine Goh

Christine Goh is Associate Professor of applied linguistics in the National Institute of Education, Singapore (Nanyang Technological University). Her interests are in listening and speaking development, and the role of metacognition in L2 learning. She has authored many international journal articles and book chapters on listening research and teaching methodology for listening.

Ewa M. Golonka

Ewa M. Golonka holds a PhD in Russian Linguistics and Second Language Acquisition from Bryn Mawr College. She has taught Russian, linguistics, and SLA at various universities. Currently, she is an Assistant Research Scientist at the University of Maryland Center for Advanced Study of Language.

Marta Gonzalez-Lloret

Marta Gonzalez-Lloret has taught at the Spanish division of the LLEA department at the University of Hawai'i for more than a decade. She holds a PhD in Second Language Acquisition from the University of Hawai'i at Manoa and her research interests include second language acquisition, technology for language learning and teaching, and teacher training.

Kira Gor

Kira Gor is Associate Professor of Russian and Second Language Acquisition in the School of Languages, Literatures, and Cultures at the University of Maryland. Her research interests include psycholinguistic mechanisms underlying cross-linguistic and second-language processing of phonology and morphology.

William Grabe

William Grabe is Regents Professor of English at Northern Arizona University, where he teaches in the MATESL and PhD in Applied Linguistics programs. His interests include reading, writing, written discourse analysis, and the disciplinary status of applied linguistics. His most recent book is Reading in a Second Language: Moving from Theory to Practice (2009, Cambridge University Press).

Rick de Graaff

Rick de Graaff is a language teaching consultant/researcher at the IVLOS Institute of Education, Utrecht University, the Netherlands. His main fields of interest include: task effectiveness in language teaching, the role of instruction in L2 pedagogy, the role of peer feedback in collaborative writing, and content and language integrated learning. Most recently he has contributed to the International Journal of Bilingual Education and Bilingualism and ITL - International Journal of Applied Linguistics.

Alex Housen

Alex Housen (MA, UCLA; PhD, University of Brussels) is Senior Lecturer in English, Second Language Acquisition and Bilingualism at the University of Brussels (VUB). His research interests include second/foreign language acquisition, second/foreign language teaching, and bilingualism. His recent publications include Investigations in Instructed Second Language Acquisition (with M. Pierrard, 2005, Mouton de Gruyter) and Bilingualism: Basic Principles and Beyond (with J. M. Dewaele and L. Wei, 2003, Multilingual Matters).

Ken Hyland

Ken Hyland is Professor of Education and director of the Centre for Academic and Professional Literacies at the Institute of Education, University of London. He has published over 130 articles and 13 books on language teaching and academic writing, most recently Academic Discourse (2009, Continuum). He is co-editor of the Journal of English for Academic Purposes.

Eunice Eunhee Jang

Eunice Eunhee Jang is Assistant Professor at Ontario Institute for Studies in Education of the University of Toronto. Her research interests include validity and fairness issues in language testing and cognitive diagnostic assessment. Her research has been published in Journal of Educational Measurement, Language Testing (in press), International Journal of Testing, Journal of Mixed Methods Research, and in the book New Directions in Psychological Measurement with Model-Based Approaches (edited by S. Embretson & J. S. Roberts, American Psychological Association).

Scott Jarvis

Scott Jarvis is an Associate Professor of Linguistics at Ohio University, where he teaches courses on second language acquisition, language testing, and other areas of applied linguistics. His main research interests are cross-linguistic influence (or language transfer) and lexical diversity, and his work has appeared in journals such as Studies in Second Language Acquisition, Language Learning, Applied Linguistics, and Language Testing. He is also co-author with Aneta Pavlenko of Crosslinguistic Influence in Language and Cognition (2008, Routledge), and is the Associate Editor for Language Learning.

Renée Jourdenais

Renée Jourdenais is an associate professor in the MATESOL/MATFL program at the Monterey Institute of International Studies, where she specializes in second language acquisition and in language teacher education. She also has extensive experience in curriculum development and in language assessment. Her recent research work explores the development of teacher knowledge.

Keiko Koda

Keiko Koda is Professor of Second Language Acquisition and Japanese in the Department of Modern Languages at Carnegie Mellon University. Her major research areas include second language reading, biliteracy development, psycho-linguistics, and foreign language pedagogy. Her recent books include Insights into Second Language Reading (2005, Cambridge University Press), Reading and Language Learning (2007, Blackwell), and Learning to Read across Languages (2008, Routledge).

Antony John Kunnan

Antony John Kunnan is Professor of TESOL and Language Education at California State University and the University of Hong Kong respectively. He has published in the Annual Review of Applied Linguistics, Language Testing, and Language

Assessment Quarterly and in many edited volumes and handbooks. He was the President of the International Language Testing Association in 2004 and is the founding editor of Language Assessment Quarterly.

Diane Larsen-Freeman

Diane Larsen-Freeman is Professor of Education, Professor of Linguistics, and Research Scientist at the English Language Institute, University of Michigan. Her most recent book (2008) is Complex Systems and Applied Linguistics, co-authored with Lynne Cameron and published by Oxford University Press.

Michael H. Long

Michael H. Long is Professor of SLA in the School of Languages, Literatures, and Cultures at the University of Maryland, College Park, where he teaches courses and seminars in the PhD program in SLA. Mike is the author of over 100 articles and several books, and has served on the editorial boards of Studies in Second Language Acquisition, TESOL Quarterly, Language Teaching Research, and other journals. His recent publications include The Handbook of Second Language Acquisition, co-edited with Catherine Doughty (2003, Blackwell), Second Language Needs Analysis (2005, Cambridge), and Problems in SLA (2007, Lawrence Erlbaum).

Sandra Lee McKay

Sandra Lee McKay is Professor of English at San Francisco State University, where she teaches courses in sociolinguistics, as well as methods and materials for graduate students in TESOL. Her books include Teaching English as an International Language: Rethinking Goals and Approaches (2002, Oxford University Press, winner of the Ben Warren International Book Award), Sociolinguistics and Language Teaching (edited with Nancy Hornberger, 1996, Cambridge University Press) and Researching Second Language Classrooms (2006, Lawrence Erlbaum). Her newest book, International English in Its Sociolinguistic Contexts: Towards a Socially Sensitive Pedagogy (with Wendy Bokhorst-Heng, 2008, Routledge) is an examination of the social and sociolinguistic context of present-day English teaching and learning.

Rosamond F. Mitchell

Rosamond F. Mitchell is Professor of Education at the University of Southampton. Her research interests are in the area of Second Language Acquisition, especially of French. She is particularly interested in theories of language learning and their empirical implications, and in the interface between linguistic theory and cognitive approaches to the learning of second languages. She is co-editor of Teaching Grammar: Perspectives in Higher Education (1996) and co-author of Second Language Learning Theories (2004).

Silvina Montrul

Silvina Montrul is Associate Professor of Spanish, Linguistics and Second Language Acquisition at the University of Illinois at Urbana-Champaign. She is author of The Acquisition of Spanish (2004, John Benjamins) and Incomplete

Acquisition in Bilingualism. Re-examining the Age Factor (2008, John Benjamins). Her research focuses on linguistic and psycholinguistic approaches to adult second language acquisition and bilingualism, in particular syntax, semantics, and morphology. She is also an expert in language loss and retention in minority-language-speaking bilinguals.

Diane Musumeci

Diane Musumeci is Associate Professor and Head in the Department of Spanish, Italian, and Portuguese at the University of Illinois at Urbana-Champaign. She is the author of Breaking Tradition: An Exploration of the Historical Relationship Between Theory and Practice in Second Language Teaching (1997, McGraw-Hill).

Paul Nation

Paul Nation is professor of Applied Linguistics in the School of Linguistics and Applied Language Studies at Victoria University of Wellington, New Zealand. His specialist interests are language teaching methodology and vocabulary learning. His latest book on vocabulary is Teaching Vocabulary: Strategies and Techniques published by Cengage Learning (2008), and two books, Teaching ESL/EFL Listening and Speaking (with Jonathan Newton) and Teaching ESL/EFL Reading and Writing, have just appeared from Routledge/Taylor and Francis.

John M. Norris

John M. Norris is associate professor in the Department of Second Language Studies at the University of Hawai'i at Manoa. His work focuses on assessment, program evaluation, research methods, and task-based language teaching in foreign and second language education. His recent publications include a single-author book Validity Evaluation in Language Assessment (2008, Peter Lang) and a co-edited volume with Lourdes Ortega Synthesizing Research on Language Learning and Teaching (John Benjamins, 2006).

Lourdes Ortega

Lourdes Ortega is associate professor at the University of Hawai'i, where she teaches graduate courses in second language acquisition and foreign language education. Her most recent book is Understanding Second Language Acquisition (2009, Hodder Arnold).

Robert Phillipson

Robert Phillipson is a Professor Emeritus at Copenhagen Business School, Denmark. His Linguistic Imperialism (1992, Oxford University Press) has also been published in China and India. Recent publications include English-Only Europe? Challenging Language Policy (2003, Routledge) and Linguistic Imperialism Continued (Orient Black-swan). Several articles can be downloaded from www.cbs.dk/staff/phillipson.

Charlene Polio

Charlene Polio is an associate professor at Michigan State University, where she directs the MA TESOL program. She has published research on second language

writing, classroom discourse, and second language acquisition and in journals such as the Journal of Second Language Writing, the Modern Language Journal, and Studies in Second Language Acquisition. She is the incoming editor of the Annual Review of Applied Linguistics and co-editor of Multiple Perspectives on Interaction: Second Language Research in Honor of Susan M. Gass to be published by Routledge.

Hakan Ringbom

Hakan Ringbom is emeritus professor of English at Abo Akademi University, Turku/Abo, Finland. Among his previous publications are The Role of the First Language in Foreign Language Learning (1987) and Cross-Linguistic Similarity in Foreign Language Learning (2007), both with Multilingual Matters.

William P. Rivers

William P. Rivers is Chief Linguist at Integrated Training Solutions, Arlington, VA. His publications include Language and National Security in the 21st Century (with Richard D. Brecht, 2001) and Language and Critical Area Studies after September 11 (with Richard D. Brecht, Ewa Golonka, and Mary E. Hart). His research interests include third language acquisition, computational sociolinguistics, and language policy.

Peter Robinson

Peter Robinson is Professor of Linguistics and SLA in the Department of English, Aoyama Gakuin University, Shibuya, Tokyo, where he teaches and supervises research on second language acquisition, cognitive abilities for language learning, and effects of instruction. Recent publications include Task Complexity, the Cognition Hypothesis and Second Language Instruction, special issue of the International Review of Applied Linguistics (co-edited with Roger Gilabert, 2007), Handbook of Cognitive Linguistics and Second Language Acquisition (co-edited with Nick Ellis, 2008, Routledge), and Second Language Task Complexity: Researching the Cognition Hypothesis of Learning and Performance (in press, John Benjamins).

Carsten Roever

Carsten Roever is a Senior Lecturer in Applied Linguistics in the School of Languages and Linguistics at the University of Melbourne. His research interests include second language acquisition, interlanguage pragmatics, and second language assessment. He has written several book chapters, journal articles, and the book Testing ESL Pragmatics (2005, Peter Lang) and has co-authored Language Testing: The Social Dimension with Tim McNamara (2006, Blackwell).

Steven J. Ross

Steve Ross teaches at the School of Policy Studies, Kwansei Gakuin University. His research has appeared in Language Learning, Applied Linguistics, International Journal of Testing, Language Testing, Journal of Pragmatics, Studies in Second Language Acquisition, Second Language Research, System, International Review of Applied Linguistic, TESOL Quarterly, and in several edited volumes.

Rani Rubdy

Dr Rani Rubdy is Senior Fellow at the National Institute of Education, Nanyang Technological University in Singapore. She is co-editor of two recently published books, English in the World: Global Rules, Global Roles (Continuum, 2006) and Language as Commodity: Global Structures, Local Marketplaces (Continuum, 2008). Her other recent publications include the book chapters, 'Remaking Singapore for the new age: Official ideology and the realities of practice' in Decolonization, Globalization: Language-in-education Policy and Practice (edited by Angel M. Y. Lin & Peter W. Martin, 2005, Multilingual Matters) and 'Language planning ideologies, communicative practices an their consequences' in Springer's Encyclopedia of Language and Education (2008).

Tove Skutnabb-Kangas

Tove Skutnabb-Kangas, emerita (University of Roskilde, Denmark and Abo Akademi University, Finland), bilingual from birth in Finnish and Swedish, has written or edited around 50 monographs and almost 400 articles and book chapters, in 32 languages, about minority education, linguistic human rights, linguistic genocide, subtractive spread of English and the relationship between biodiversity and linguistic diversity. She lives on an ecological farm with husband Robert Phillipson. For publications, see http://akira.ruc.dk/~tovesk/.

Kris Van den Branden

Kris Van den Branden is a professor of linguistics at the Katholieke Universiteit Leuven. He is one of the current directors of the Centre for Language and Education at the same university. His main research interests are in task-based language teaching, the role of interaction in instructed language learning, and the diffusion of innovations in the educational field. He has published in many international journals, and has edited a volume on task-based language teaching in the Cambridge University Press Applied Linguistics Series.

Larry Vandergrift

Larry Vandergrift is Professor at the Official Languages and Bilingualism Institute (OLBI) at the University of Ottawa. His research in the teaching of second/ foreign language listening has been published in Annual Review of Applied Linguistics, Applied Linguistics, Canadian Modern Language Review, Language Learning, Language Teaching, Modern Language Journal, and more. He is currently a co-editor of the Canadian Modern Language Review and director of the research centre at OLBI.

Karen Vatz

Karen Vatz is a graduate student in the Second Language Acquisition PhD program at the University of Maryland. She is currently working on her dissertation on the representation and processing of grammatical gender in advanced L2 learners. Other areas of interest include bilingual lexical representation and critical period effects.

Alan Waters

Alan Waters is a Senior Lecturer in the Department of Linguistics and English Language, Lancaster University, UK. He has taught EFL and trained teachers in the UK and several other parts of the world. He has published a number of books and articles on a range of ELT topics.

Jessica Williams

Jessica Williams is a Professor of Linguistics at the University of Illinois at Chicago, where she also directs the TESOL program. She has published on variety of topics, including second language writing, lexical acquisition, and the effect of focus on form. Her latest publications include an edited volume (with Bill VanPatten, 2006, Routledge), Theories in Second Language Acquisition and the student text, Academic Encounters: American Studies (Cambridge University Press, 2007).

Part I Overview

The Handbook of Language Teaching Editedby MicheelH.LnngandCethermeL.Doughty © 2009 Blackwell PublishDgLtd. IBBN: 978-1-408-15789-5

1 Language Teaching

MICHAEL H. LONG

Hundreds of millions of people voluntarily attempt to learn languages each year. They include adults who seek proficiency in a new language for academic, professional, occupational, vocational training, or religious purposes, or because they have become related through marriage to speakers of languages other than their mother tongue. Then, there are (some would argue, "captive") school-age children who experience their education through the medium of a second language, or for whom one or more foreign languages are obligatory subjects in their regular curriculum. In addition to these easily recognizable groups, language teachers around the world are increasingly faced with non-volunteers. These are the tens of millions of people each year forced to learn new languages and dialects, and sometimes new identities, because they have fled traumatic experiences of one kind or another - war, drought, famine, disease, intolerable economic circumstances, ethnic cleansing, and other forms of social conflict -crossing linguistic borders in the process. Since the horror and frequency of such events show no signs of decreasing, language teaching is likely to remain a critical matter for these groups for the foreseeable future, with the scale of forced mass migrations if anything likely to grow in the twenty-first century, due to the potentially disastrous effects of climate change.

For both groups of learners, volunteers and non-volunteers, language teaching is increasingly recognized as important by international organizations, governments, militaries, intelligence agencies, corporations, NGOs, education systems, health systems, immigration and refugee services, migrant workers, bilingual families, and the students themselves. With the growing recognition come greater responsibility and a need for accountability. LT1 is rarely a matter of life or death, but it often has a significant impact on the educational life chances, economic potential and social wellbeing of individual students and whole societies. Students and entities that sponsor them increasingly want to know not just that the way they are taught works, but that it constitutes optimal use of their time and money.

Demonstrating effectiveness and efficiency is often difficult. Historically, LT has been regarded as an art - or a craft, at least - not a science, with scant regard

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

and little financial support for research. Demand for some languages, notably English and Chinese, has been so great in recent years that, with demand far exceeding supply, few consumers have been in a position to quibble over the quality of their instruction. In the case of some rarely taught languages for which there is a sudden surge in need, e.g., as a by-product of military actions or natural disasters, students and sponsors have no choice but to accept whatever can be found, adequate or not. Even in the case of widely taught languages, like English, Chinese, Arabic, French, German, and Spanish, research that is carried out is sometimes criticized for having been conducted in real classrooms and other "natural" instructional settings, with a resulting lack of control over significant variables that may have influenced the outcomes of interest. Alternatively, when conducted under controlled experimental conditions, studies are sometimes criticized for having produced findings that may not generalize to real classrooms. Series of studies of the same phenomena in both natural and artificial instructional environments, utilizing a variety of research methods, are clearly desirable.

Despite these problems, the situation has gradually improved in recent years, with steady growth in the amount and sophistication of research on LT itself, and in disciplines with much to say about the process LT is designed to facilitate, language learning. Of those feeder disciplines, theory and research in some areas of second language acquisition (SLA) are the most directly relevant, but work in psychology, educational psychology, anthropology, curriculum and instruction, and more, is also valuable. This is not to say that all the answers are known, or even that most of them are, but LT prescriptions and proscriptions that ignore theory and research findings in those fields are gradually and justifiably losing credibility. Where they are kept viable, it is chiefly by commercial interests, which still wield enormous influence, and the continued marketability of whose wares is often best served by ignorance about effectiveness.

The authors of each chapter in this volume were asked wherever possible to draw on research findings when making proposals. This, they have done. Also, while many of them specialize in the teaching of English, on which the greatest number of studies have been carried out, and/or operate in English-speaking countries, they were asked not to focus on the teaching of any one language or any one teaching context - foreign, second, lingua franca, etc. - but to choose examples and synthesize research findings and teaching experience from, and relevant to, a variety of languages and settings. They were asked to provide balanced evaluations of major positions and approaches, but granted scope to advance their own views. This, they have also done.

As is visible in the Table of Contents, in addition to coverage of core foundational issues, The Handbook of Language Teaching contains chapters on a few topics seldom found in comparable anthologies and textbooks. These chapters reflect recent developments and changing emphases in the field, or ones we believe deserve more attention. Examples include chapters on the language-learning brain; on programs designed specifically for heritage learners, about whom there is now an explosion of (sometimes rather uninformed) writing; on advanced learners; study abroad; third language, conversion, and cross-training programs; LCTLs

(less commonly taught languages), which geopolitics are rapidly making a lot more commonly taught; and (not unrelated) on reading new scripts; as well as on radical language teaching and the diffusion of innovation. In another departure from the norm, instead of one chapter on teaching various skills, and a separate one on testing them, we invited one author to cover both in a single chapter. The idea is to avoid overlap and facilitate greater coherence of treatment. We selected individuals whose prior work showed they can handle both at the required level. While certainly not unique to this volume, there is also expert coverage of the increasingly apparent and important politics and social and political context of language teaching.

One author conspicuously missing from the assembled company is the late Craig Chaudron, a widely respected expert on many aspects of LT, and a valued colleague and close personal friend. Craig had agreed to contribute a chapter to the handbook, but as many readers will know, died unexpectedly in 2006. His untimely passing is a tragic loss for all who knew him, and for the field as a whole. This volume is humbly dedicated to his memory.

We are grateful to Danielle Descoteaux, Julia Kirk, and the staff at Wiley-Blackwell for their support at all stages of the development of this volume, and to the reviewers of individual chapters.

1 The following abbreviations are used throughout the volume:

FL - foreign language

L1 - first, or native, language

L2 and SL - second language in the broad sense, including any additional language to the L1

LT - language teaching

SLA - second language acquisition.

Part II Social, Political, and Educational Contexts of Language Teaching

The Handbook of Language Teaching Editedby MichaelH.InngandCathermel.Doughty © 2009 Blackwell PublishDgLtd. IBBN: 978-1-408-15789-5

2 The Social and

Sociolinguistic Contexts of Language Learning and Teaching

SANDRA LEE MCKAY AND RANI RUBDY

We live in an age of linguistic diversity increased greatly by globalization, the movement of people across borders, and the widespread acquisition of additional languages by individuals in their own countries. All of these factors have led to an increase in the number of second-language learners and the kinds of contexts in which they are learning languages.

This chapter is about the social and sociolinguistic context of present-day foreign and second-language learning and teaching. In examining the social context of language learning, we focus on how language teaching contexts are affected by the larger social, political, and educational setting in which the teaching takes place. In examining the sociolinguistic context of language teaching, we focus on how the linguistic features of interactions, both inside and outside of the classroom, are affected by the social context in which the interaction takes place.

Our division is in many ways similar to a traditional distinction made in the field of sociolinguistics where one of the major debates is whether to take social or linguistic factors as primary in investigating the relationship between the social context and language variables. As evidence of this debate, Wardaugh (1992) and others make a distinction between the sociology of language and sociolinguistics. Whereas the sociology of language investigates the manner in which social and political forces influence language use, sociolinguistics takes linguistic factors as primary in its investigations of language and society.

In keeping with this distinction, the first part of the chapter focuses on two areas of investigation typically studied in the sociology of language that influence the social context of language learning: language planning and policy, and societal multilingualism. The second part of the chapter focuses on two areas of investigation typically studied in what Wardhaugh terms sociolinguistics: language contact and variation, and ethnographic sociolinguistics. The final section of

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

the chapter uses case studies of second and foreign language teaching to illustrate how the social and sociolinguistic context can influence language pedagogy.

Whereas this chapter will discuss the learning and teaching context of various languages, a good deal of attention will be given to the learning of English. This is because today English is the most widely studied second and foreign language (Crystal, 1997), causing the study of the social and sociolinguistic context of the learning of this language to be of interest to many learners and practitioners.

The Social Context of Language Learning and Teaching

Language planning and policy

The social context of language learning and teaching is greatly impacted by a nation's political decision to give special status to a particular language or languages. This status can be achieved either by making the language an official language of a country or by giving special priority to the language by requiring its study as a foreign language. Today there are over 75 countries in which English has been or continues to be an official language of the country, with many more nations requiring the study of English in the public schools (Crystal, 1997). This situation provides tremendous incentives for the learning of English.

The political choice of designating an official language is fully discussed in Chapter 3. What is important for our purposes is how this choice affects the social context of language learning and teaching. Three ways in which the designation of an official language has consequences for language learning and teaching are (a) the insight the designation provides into prevalent social attitudes toward particular languages, (b) the effect of the language policy on the stated language-in-education policy, and (c) the setting of linguistic standards.

The designation of an official language can foster a great deal of political tension that polarizes social attitudes toward particular languages. Malaysia's decision, for example, to recognize Bahasa Melayu as the country's sole official language was strongly opposed by the ethnic Chinese and Tamil populations, who preferred giving English equal status. The debate in South Africa over which languages to designate as official was also based on ethnic lines. In both cases the decision of whether or not to give special status to a particular language became a rallying point for social and ethnic groups. Such social attitudes obviously can affect an individual's motivation to learn or not learn a particular language.

A second consequence of a language being designated as one of the official languages of the country is that in most cases the country's official language or languages are used, or at least designated to be used, as the medium of instruction in the schools. The National Educational Policy of South Africa is a case in point. In 1997, the former Minister of Education argued that South Africa's national language-in-education policy was integral to the government's strategy

of redressing the discrimination of the past and building a non-racial nation in South Africa. He contended that being multilingual should be a defining characteristic of being South African (Chick & McKay, 2001).

While providing for choice from a range of language-in-education policy models, the South African National Educational Policy identified additive bilingual-ism/multilingualism as the normative orientation of the language-in-education policy. This policy, however, contradicts the beliefs of many South African parents that the best way to acquire English, the dominant ex-colonial language, is to commence studying the language as early as possible; that maintenance of the first language is unnecessary and perhaps undesirable; and that the best way for speakers of other languages to acquire English is submersion, that is, a subtractive approach. Given the strong desire on the part of many parents for their children to learn English, English-medium education is currently the only option offered by South Africa's most sought-after schools. This situation exemplifies the manner in which the stated language-in-education policy is often undermined by prevalent social beliefs as to the value of particular languages.

National language policies can also influence language learning and teaching by the setting of standards. For example, in Singapore today, there is a segment of the population that speaks a localised dialect of English widely known as Singlish or Singapore Colloquial English (SCE). Like many stigmatized varieties, Singlish has begun to gain immense popularity among young professionals, who increasingly use it in domains of friendship and solidarity. Its negative association with the poorly educated and its accelerated usage among the general populace in recent years alarmed the Singapore authorities sufficiently to warrant the mounting of a Speak Good English Movement, a campaign that overtly promotes the use of standard English, and whose implicit agenda is to stem the spread of Singlish before it becomes an integral part of the cultural life of the present generation of school-goers in Singapore. As in the case of Singapore, government policies can influence not only which language is promoted but also which variety of that language is preferred.

Societal multilingualism

When a country has more than one official language and the majority of the population is bilingual, there are generally particular domains in which each language is used. Ferguson (1959) coined the term diglossia to describe the situation of a community in which most of the population is bilingual and/or bidialectal and the two codes serve different purposes. The term was originally used by Ferguson to describe a context in which two varieties of the same language are used by people of that community for different purposes. Normally one variety, termed the High, or H, variety, is acquired in an educational context and used by the community in more formal domains, such as in churches or universities. The other variety, termed the Low, or L, variety, is acquired in the home and used in informal domains, like the home or social center, to communicate with family and friends.

Later, Fishman (1972) generalized the meaning of diglossia to include the use of two separate languages within one country in which one language is used primarily for formal purposes and the other for more informal purposes. The expansion of the meaning of the term made it applicable to countries in which English is one of the official languages, as in South Africa, Singapore, and India. In these countries, English often assumes the role of what Ferguson calls the High variety, with the other languages of the country, or a different variety of English, being used in informal domains.

The fact that these different languages or varieties serve different purposes has implications for second-language teaching. In many cases the language or dialect that serves the purposes of the Low variety has lower status, so that speakers of this variety are marginalized in society and in the school system. Because of this, speakers of this variety are often given the impression that their home language is inferior; furthermore, their lack of access to the High variety can impede their progress in the educational establishment and, ultimately, in society.

Two additional concepts in the study of societal multilingualism that are important for our purposes are language maintenance and language shift. In the case of language maintenance, members of a language minority group work to promote the maintenance of their first language. This is the case of many language minority groups in the United States who have established after-school first-language maintenance programs, funded print and media programs in their first language, and supported special events in which the first language is used. (See McKay & Wong, 2000.) Language shift, on the other hand, occurs when members of a language minority replace the use of their first language in favor of another one. This is the case for almost all third generation immigrants to the United States.

The concepts of language maintenance and shift are particularly relevant to the topic of linguistic diversity in an era of globalization. Today, many warn of the danger of the spread of English and the threat it poses to the continued existence of indigenous and smaller languages (Nettle & Romaine, 2000; Phillipson, 2003). For such individuals, English is seen as the culprit in the decrease in the number of languages spoken in the world. However, there are others (e.g., Brutt-Griffler, 2002) who maintain that the spread of English is not a step toward a monolingual world of English speakers but rather a step toward a world in which bilingualism is the norm. Indeed the tremendous increase in the number of second-language speakers of English would seem to support this position.

The growth of individuals who are learning another language in their own country in order to partake in regional or global exchanges has important implications for second and foreign language learning and teaching. To begin with, such individuals have another language that serves their informal and intimate needs. Hence, they typically have little need to develop informal registers of the regional or global language. Second, in many instances individuals will acquire the additional language in order to communicate with other non-native speakers of that language. Because of this, much more attention should be given

in language classes to developing strategies that help learners to communicate in exchanges in which neither speaker is fully fluent in the language.

The Sociolinguistic Context of Language Learning and Teaching

Language contact and variation

One common effect of language contact is language change. In such cases, the various languages used within a multilingual context may undergo phonological, lexical, and grammatical changes as bilinguals make use of two or more languages on a regular basis. This situation is occurring in many countries today where English has an official role in the society. In these countries, English is being influenced by the other languages it comes in contact with. In addition, English is often influencing other languages through the borrowing of English terms.

Many studies have been undertaken to determine the types of grammatical changes that are occurring in various multilingual contexts in which English plays a significant role. (See, for example, Kachru, 2005.) Frequently, researchers begin by examining a written corpus of English of a particular multilingual context to determine what kinds of grammatical innovations exist and how acceptable these structures are to both native speakers of English and local speakers of English. In general, when investigations of language change use a written corpus of published English, only very minor grammatical differences are found. (See, for example, Parasher, 1994.)

Often the kinds of grammatical changes that occur tend to be minor differences, such as variation in what is considered to be a countable noun (e.g., the standard use of luggages in English in the Philippines and the use of furnitures in Nigeria) and the creation of new phrasal verbs (e.g., the use of dismissing off in English in India, and discuss about in Nigeria). In contexts in which such features become codified and recognized as standard within that social context, there arises what Kachru (1986) has termed a nativized variety of English.

What is perhaps most puzzling in the development of alternate grammatical standards in the use of English is the fact that whereas lexical innovation is often accepted as part of language change, this tolerance is generally not extended to grammatical innovation. In Widdowson's (1994) view, the reason for this lack of tolerance for grammatical variation is because grammar takes on another value, namely that of expressing a social identity. Hence, when grammatical standards are challenged, they challenge the security of the community and institutions that support those standards.

Investigations of language contact have also focused on the code-switching behavior of bilinguals. One of the most comprehensive theories of codeswitching is that of Myers-Scotton (1993). She explains code-switching in terms of a theory of rights and obligations. She proposes a markedness model of code-switching

which assumes that speakers in a multilingual context have a sense of which code is the one expected to be used in a particular situation. This is termed the unmarked code. However, speakers can also choose to use the marked code, that is, the language or language variety that is not expected in a particular social context. Using data from multilingual African contexts, Myers-Scotton demonstrates how bilingual speakers make code choices to signal a variety of social relationships. Unfortunately, in many language learning and teaching contexts, the rich linguistic repertoire of bilinguals is not recognized, and policies are often implemented to prohibit the use of any code other than the target language.

Studies in language contact have several implications for the teaching and learning of another language. As mentioned above, language contact will inevitably result in language change. Since today many individuals are using English in contact with other languages on a daily basis, their use of English is changing, and they are in the process of establishing their own standards of English grammar and pronunciation. In general, research on these emerging varieties of English indicates that the codified and accepted standard of English that exists in these communities has few differences from other standard varieties of English. Hence, it is important for L2 teachers to recognize the integrity of the varieties of the language they teach, to realize that they are important sources of personal identity and signs of the current mobility of populations, and to avoid promoting negative attitudes toward such varieties.

Studies on code-switching have illustrated the regularity of code-switching behavior and the purposes that code-switching can serve for bilinguals. Given the many contexts today where English is used as one of the additional languages within a country, more research is needed regarding how individuals make use of English in reference to the other languages they speak. Such research will be valuable in establishing classroom objectives that complement the students' use of English within their own speech community. In addition, in classrooms in multilingual contexts where the teacher shares a first language, more research is needed to determine how students' first language can be used to further their competence in a target language.

Ethnographic sociolingistics

A good deal of current work in sociolinguistics falls under what is referred to as an ethnomethodologically oriented approach to the field of sociolinguistics, with linguistic interaction as the focal point. One of the central concepts of ethnographic or interactional sociolinguistics is the term speech community. Hymes (1972) contends that members of a speech community must share the same rules of speaking and be familiar with at least one common linguistic variety. Individuals are typically members of several speech communities and alter their norms of language use to conform to other members of the same speech community. With growing mobility, individuals today can belong to many different speech communities.

Work in linguistic interaction began as a reaction to Chomsky's (1957) focus on the language of an idealized speaker-listener in a homogeneous speech community with complete knowledge of the language. This notion was challenged by Hymes (1974), who insisted that studies on language use should strive to account for the communicative competence of a native speaker of a language. Gumperz (1982) also challenged Chomsky's notion of an idealized speaker in a homogeneous speech community, arguing instead that language use in a speech community is influenced by social and cultural factors. Gumperz's studies on communication between blacks and whites in the United States and between Indians and British in England demonstrated how differences in language use among speech communities can cause misunderstandings leading to racial and ethnic stereotypes and inequalities in power.

The work of Rampton (1995, 1997) has taken the debate about linguistic diversity one step further. He maintains that globalization, as well as late/post-modernity (a term he prefers to postmodernism), warrants a fresh look at the issues important to sociolinguistics and L2 research. Rampton believes that the time has come for sociolinguists to challenge the notion that societies are compact and systematic entities and instead to recognize the heterogeneity and fluidity of modern states. In keeping with much of the discourse of postmodernism, he argues persuasively that sociolinguistics should give more attention to investigating issues related to fragmentation, marginality, and hybridity and recognize that "being marginal is actually a crucial experience of late modernity. Being neither on the inside nor the outside, being affiliated but not fully belonging, is said to be a normal condition" (Rampton, 1997, p. 330).

The ability to signal identity through surface linguistic features has significant ramifications for language learning and teaching. In many contexts around the world, one of the major goals of teaching a second or foreign language is to promote the acquisition of the standard form of the target language. As a result, those who use an alternate form of the target language as a way of signaling their hybridity and affiliation with a particular speech community are often penalized. They are marginalized in the society and often penalized in a school system that uses one standard to determine proficiency in the language.

In order for current sociolinguistic research to be in touch with issues of late modernity, further research is needed that investigates linguistic diversity without preconceived ideas about native speakers and language standards. Such research should examine how particular varieties of language illustrate the fluidity of modern society. This type of research is presently underway in investigations of English as a lingua franca (ELF), which Firth (1996) defines as a contact language "between persons who share neither a common native tongue nor a common (national) culture, and for whom English is the chosen foreign language of communication" (p. 240). Since it is estimated that today about 80 percent of oral exchanges in English do not involve native speakers of English (Seidlhofer, 2004), more research is needed to determine the characteristics of such exchanges. Most of the current research in this area has taken place in Expanding Circle

countries (see below), much of it in Europe. (See Seidlhofer 2004 for a review of ELF research.)

Major Second and Foreign Language Learning and Teaching Contexts

As a way of illustrating how the social and sociolinguistic factors described above have implications for second and foreign language learning and teaching, it is helpful to consider the three major types of learning contexts described by Kachru (1986) in reference to English learning: (a) the Inner Circle, where English is the primary language of the country, such as in Australia, Canada, and the United Kingdom; (b) the Outer Circle, where English serves as a second language in a multilingual country, such as in Singapore, India, and the Philippines; and (c) the Expanding Circle, where English is widely studied as a foreign language, such as in China, Japan, and Korea.

This very broad distinction of learning contexts can be generalized to all second and foreign languages. In doing so, any teaching context in which a learner is learning the major language of the host country would be termed an Inner Circle learning context. This is the type of social context faced by many immigrants today, who in an age of globalization often immigrate to other more prosperous countries, typically for better economic opportunities. This is the case of many Turks in Germany, Poles in Ireland, and Indians in Great Britain. The second major type of second and foreign language learning is an Outer Circle learning context, in which the learner is acquiring one of the major languages spoken in the country. This would be the case of English speakers learning French in Quebec or Zulu speakers learning English in South Africa. Finally, in Expanding Circle learning contexts, learners are learning a language, often a language of regional or global importance, in a country in which the language has no official role. This is the case of American learners studying Mandarin or Japanese learners learning English. The question is, in what way do the social and sociolinguistic constructs described above provide insight into the learning and teaching context? In order to address this, we examine some representative case studies in each context. We begin with Inner Circle learning contexts.

Inner Circle learning contexts

Language maintenance and language shift, as well as language-in-education policy, are particularly relevant to Inner Circle learning contexts. In many instances, government and educational leaders make decisions that can encourage either language maintenance or shift among its minority population. British government policies regarding the acquisition of English among minority language speakers illustrate the effect that government policy can have on language maintenance and shift. During the 1960s and 1970s in the United Kingdom, students with limited English proficiency were placed either in specialist language centers or

withdrawn from mainstream classrooms for special ESL instruction. Then, in 1986, the Commission for Racial Equality decided that language centers constituted a form of racial discrimination in that language minority students, typically members of racial minority groups, were being singled out and that such programs were creating further social and racial barriers between groups.

As Rampton (1995) notes, British government policy in general tolerates the use of minority languages only as a transitional measure during the early years of schooling, and there is no encouragement for the development of bilingual-ism during a child's education. Obviously, such policies are designed to promote language shift, with no encouragement for language maintenance. What are the implications of such policies for language learning and teaching? In general, they promote a view of a language classroom as an English-only environment in which the first-language resources a child has are not recognized. Such policies also do not typically promote favorable attitudes toward bilingualism or code-switching.

These pedagogical policies are in sharp contrast to the linguistic situation many language minority youths encounter outside the classroom. Here, as Rampton (1995) carefully documents in his investigation of British language minority adolescents, code alternation both between Standard English and Stylised Asian English and between English and the students' first language plays an important role in "the organization of interracial adolescent solidarities" (p. 141). Whereas the actual interactions of British adolescents with adults and their peers demonstrate the sociolinguistic sophistication of bilingual youths, this is not recognized as an asset or promoted in classrooms which support an English-only model that recognizes and legitimizes only standard English.

The language-in-education policies in countries like the United Kingdom reflect a huge gap between the sociolinguistic reality of language minority groups and prevalent pedagogical approaches to the teaching of English. If pedagogical models are to be sensitive to the social interactions of the local context, much more attention needs to be given in Inner Circle learning contexts to the value of bilingualism, the maintenance of existing linguistic resources, and the sociolin-guistic knowledge of language learners.

Presently, Australia is making some attempt to encourage the development of bilingualism among its citizens. As an inducement to encourage the learning of languages other than English (LOTE), some Australian universities are offering bonus points to those university applicants who include the study of LOTE in their required university entrance subjects (Smolicz & Secombe, 2003). Even with this incentive, only 10-20 percent of students take LOTE as a university entrance subject. As Smolicz and Secombe point out, this is "particularly striking for students from the majority English-speaking background, many of whom see no obvious benefits from investing the effort required to learn a new language, in view of the availability of what they perceive as 'easier options', as well as the global dominance of English" (2003, p. 16). Attitudes such as this suggest that the English-speaking populations of Inner Circle countries may be the least likely to become bilingual in an increasingly multilingual and multicultural world.

Outer Circle learning contexts

Diglossia, language change, linguistic standards, and code-switching are particularly relevant to many Outer Circle learning contexts. Not least because many of the Outer Circle countries are inherently characterized by a rich and complex multiethnic, multicultural, and multilingual ethos wherein English was transplanted and taught either as an important official language or a language of higher education, or both.

Partly a legacy of colonial history and partly the effect of post-independence policies, English was introduced as a panacea for solving the economic and educational problems of many newly independent states in the Outer Circle, such as India, the Philippines, Singapore, and those of Africa, and soon became a dominant language of the ruling elite. The study of English was therefore initially restricted to a particular socio-economic class, with a focus on literature and culture. Today, its earlier role in these contexts as an "instrument of science and technology" and a language of "development" and "national unity" is increasingly being replaced by economic and pragmatic motivations, due to the meshing of English with globalization. This has meant a strong demand for communication skills and everyday use of English, in place of the study of literature, for creating a competent workforce for the multinational corporations and outsourcing centers that represent this global trend.

Many "new" and distinct indigenized or nativized varieties of English (Kachru, 1986) have evolved, due to the processes of change that English has undergone by acquiring new linguistic and cultural features as a result of contact with indigenous languages and cultures. These contact varieties have become institutionalized and are recognized by their speakers as autonomous varieties of English in their own right, notwithstanding the struggle for legitimacy and equivalence with long-established "native" varieties, particularly British and American Englishes, that continue to be looked upon as powerful reference points in the practice of English language instruction in some of those settings.

A crucial consideration for speakers of English in the Outer Circle has been the choice of a pedagogical model for teaching English as a second/foreign language, that is, whether to adopt an external "native speaker" standard or an internal one as the appropriate model in serving the purposes of their local contexts. Governments and education ministries differ sharply in the positions they have taken. For instance, Indian English (English of educated Indians) is taught and learnt in all educational institutions in India and this is the variety also used in the administration, the judiciary, the military, and the media. This rejection of what Phillipson (2001) terms "a debilitating dependence on native speaker models" (p. 195) is in tune with the sociolinguistic realities of the Indian subcontinent, where most Indians use English as a lingua franca to communicate with other Indians or Asians from language groups not their own, much less frequently with Inner Circle speakers. This also means that Indian English has "multiple identities" that are markers of distinctive regional variation within it.

India has adopted a policy known as the "three language formula," wherein students in the mainstream typically learn their mother tongue or regional language as their first language, and English and Hindi as their second and third languages, respectively. Of the roughly 200 languages that actively contribute to making India a functionally multilingual country and not just a demographically multilingual one, 41 languages, including its 18 official languages, are available for study in the school curriculum (NCERT, 1999, cited in Annamalai, 2004, p. 177). However, because English is the only language taught in all states as either a first (e.g., in privately run English-medium schools), second, or third language and is taught in the largest number of schools overall, from one point of view it has been argued that making English compulsory in Indian schools has rendered hundreds and thousands of children handicapped. Because it marginalizes vernacular medium students and defeats the policy goal of nation-building with equal educational opportunities for all, free India, it is argued, is free only politically and not educationally (Krishnaswamy & Krishnaswamy, 2006).

Whereas English still remains essentially an urban middle-class phenomenon in India, and as the language of the ruling bureaucracy and higher education it is domain-specific and register-based, English in Singapore occupies a unique position among Asian countries. In Singapore, English has first-language status in the educational curriculum at all levels and is the de facto working language of administration, business, and the media. While the more standard variety of English is the one taught in schools, alongside it has developed a variety which has a distinctive phonology, syntax, and lexicon, which shows a high degree of influence from the other local languages such as Hokkien, Cantonese, Malay, and Tamil, a contact variety thought to be pioneered in the classroom and the playground by children (Gupta, 1994). Because English in Singapore and Malaysia was found to display a range, it was initially characterized as a "lectal continuum" within a "post-creole continuum" (Platt & Weber, 1980), extending from the basilect at one end, which showed features of creoles, to the acrolect at the other, approximating Standard (superstrate) English, with the mesolect mediating transitionally between the two. However, since English in these regions was acquired through the formal school system, a description more favored today is one that sees it as a form of diglossia (Gupta, 1994; Foley et al., 1998), with Standard Singapore English (SSE) constituting the High variety, used in public, formal and educational domains and Singapore Colloquial English (SCE), the Low variety used in the home and neighborhood. Members of the Singaporean speech community know well when SCE is or is not appropriate and even where code-mixing between SSE and SCE is commonplace, children do separate the codes at a very early age (Gupta, 1994).

Standards of English have remained a continuing concern at the highest levels of government. Despite the fact that educated Singaporeans have come to enjoy a greater degree of English language proficiency in present-day Singapore, there is currently strong official pressure to promote an exonormative standard - notably British or American English - so as to curb further "decay" of the language through processes of indigenization. Varied measures have been undertaken to bring

about targeted change. The Speak Good English Movement, for instance, was mounted in 2000 as a timely check to contain the popularity of the local vernacular "Singlish," considered a non-standard variety with an alluring potential for symbolizing Singaporean identity and solidarity among young Singaporeans. For some, the movement was viewed as a way to help Singapore plug into the English-dominant global economic network, while others saw it as a way to facilitate linguistic homogenization by devaluing and diminishing the existing linguistic resources of the average Singaporean (Chng, 2003; Kramer-Dahl, 2003; Rubdy, 2001).

The dominance of English in Singapore has been enormously boosted by the institutionalization of English as the medium of instruction and first language in schools since 1979, as a measure to create national unity and forge a national identity and consciousness that transcended ethnic boundaries, as well as by the assignation of the local "mother tongue" languages to second-language status, such that all Singaporeans may be described as "English-knowing bilinguals" (Pakir, 1991). Singapore is, in fact, well on its way toward becoming a largely English-speaking country, certainly one that is English dominant (Lim, 2004). Equally, the Speak Mandarin campaign that was put in place as a measure simultaneously to uphold Asian values and counter the influence of "Western decadence," and that is now paying off with the emergence of China as a powerful trading partner, has had an overwhelming success in increasing the widespread use of Mandarin. However, this has happened at the expense of the many Chinese dialects that it has replaced, leading to extensive language shift and language loss, and threatening intergenerational continuity. The language-in-education policy thus causes the language of school and government to displace the language of home and neighborhood (Tickoo, 1996).

Kamwangamalu (2003) reports on a similar trend in South Africa, where English is "spreading like wildfire" and has infiltrated the family domain, particularly in urban black communities, who "see the language as an open sesame by means of which one can achieve unlimited upward social mobility," and prefer English-medium education over an education in their own native languages, such as Sotho, Zulu, or Xosa. Kamwangamalu maintains that if the current trend toward monolingualism in English continues, the African languages will face attrition and death. He points to the anxiety and agony expressed in South Africa by, on the one hand, some purists "who believe that the language is being mutilated through nativization by its new users (i.e., non-native speakers); and, on the other hand, African language activists and community leaders, who see the spread of English into the family domain as a threat to the maintenance and a prelude to the demise of the indigenous languages" (Kamwangamalu, 2003, pp. 68-9).

The seemingly people-driven spread of English in South Africa reflects a growing but worrying trend observable also in other Outer Circle countries. The "English advantage" that India has as a key to employment in the global market is appreciated by many Indians. So, "English for all" is the new slogan. The demand that English is also for the masses is gaining ground, and it is estimated that with the introduction of a new policy for English at the primary level, about 150 million children at the primary stage will be learning English in India. This

move could bring about a dramatic change in the demographics of the English-speaking population in this country and that of the Anglophone world, as well (Krishnaswamy & Krishnaswamy, 2006).

Expanding Circle learning contexts

Language-in-education policies play an important role in the learning and teaching of foreign languages. Current Chinese language-in-education policies illustrate the manner in which the learning and teaching of English has been strongly influenced by government policies in regard to both the requirements for studying English and the methods promoted in English language classrooms. In 1976, Deng Xiaoping launched a national modernization program in which English education was seen as a key component: "English was recognized as an important tool for engaging in economic, commercial, technological and cultural exchange with the rest of the world and hence for facilitating the modernization process" (Hu, 2005, p. 8).

In 1978, the Ministry of Education issued the first unified primary and secondary curriculum for the era of modernization. This curriculum introduced foreign language learning at Primary 3. The directive also mandated that efforts in promoting English language proficiency were to be aimed at strengthening English language teaching in elite schools, which were expected to produce the English-proficient personnel needed to successfully undertake national modernization. In fact, in 1985, the Ministry of Education exempted poorly resourced schools from providing English instruction. In addition, the Ministry of Education gave several economically developed provinces and municipalities the autonomy to develop their own English curricula, syllabi, and textbooks for primary and secondary education (Hu, 2005). These materials tended to be more innovative, learner-centered, and communicative than earlier classroom texts and materials.

The directives summarized above illustrate the dangers that can arise from state mandated guidelines for language teaching. First, such mandates can determine when foreign language learning begins in the public school system. The Chinese Ministry of Education, like governments in many other Asian countries, is formally promoting the early learning of English, even though the issue of early exposure to foreign language learning is still being debated (see Hyltenstam & Abrahamsson, 2001). Second, state mandates can determine who has access to English language learning. In China, recent policies have tended to support English learning among the Chinese elite, in this way exacerbating educational inequality. Finally, state mandates can determine how a language is taught. In China, as in many other Asian countries, current curriculum developments have tended to promote more learner-centered, communicative methods. The problem, however, has been a lack of teacher education that will ensure the effective implementation of new methods.

Malaysia is a country that historically presents a sharp contrast to China's English teaching policy in that, in its early independence days, Malaysia tried officially to discourage the spread of English. At independence in 1957, the Malays

made up close to half the population, the Chinese a little over a third, and the Indians 10 percent (Gill, 2005). In the previous colonial system, English-medium schools were located in urban areas and were primarily attended by non-Malays and a small number of elite Malays. Many Malay nationalists were frustrated by the fact that those who spoke English were non-Malays and that knowing English gave them a social and economic advantage. The Malays believed that designating Bahasa Melayu as the official language would lead to its development as a language of higher status and thus provide Malays with the linguistic capital previously held by the English-speaking Chinese and Indians (Gill, 2005). To promote the status of Bahasa Melayu, the language was established as the language of education, with all universities required to use Bahasa Melayu as the medium of instruction. The resistance of the Chinese and Indian population to this policy led the government to undertake a rapid implementation of the language-in-education policy, so that the status of Malay language could be established.

What is surprising is that this early attempt in Malysia to control the spread of English among its population is currently being re-examined. As evidence of this change in policy, in 2002 the Prime Minister of Malaysia, Tun Dr. Mahathir Mohamad, announced that starting immediately, science and mathematics were to be taught in English in both the primary and tertiary levels. The question is, what led to this change in policy from one of restraining the spread of English to one of encouraging its development as an additional language.

Gill (2005) contends that the change was brought about because the nationalistic language policy had resulted in a generation of university graduates who were fluent in the national language but not in English. The problem is that in an age of globalization, in order to be competitive, Malays need English to access the tremendous amount of scientific and technological knowledge available in English. Without access to this information, the government believes Malaysia can not be competitive in a global market. In this way, current changes in Malaysia's education policy "were largely influenced by the two domains which are important in the growth and status of any language - the domain of business and the domain of science and technology" (Gill, 2005, p. 256).

The case of Malaysia offers a vivid example of how the larger social and political context can affect language-in-education policies. Malaysia, like many countries today, is struggling with an attempt to balance its nationalistic priorities with the need to stay competitive in a global economy. At the present time, Malaysia appears to be replacing its desire to promote its national language with its felt need to establish an English-knowing population that will make it a competitive society in today's global economy.

In contrast to Malaysia, Sweden is struggling with its national language being dominated by English.

Today Swedish, like many national languages throughout the world, is in an awkward position. It is at the same time a strong national language with the potential to dominate other languages within its borders and a potentially dominated language with respect to English as an international language. (Hult, 2004, p. 182)

According to Hult, the prominence of English in higher domains like education, commerce, and industry threatens Swedish "to the point where there is a risk of a two-tiered society developing, with English used for high status interaction and Swedish for lower status common daily interaction" (Hult, 2004, p. 183). There is also concern within Sweden that the consequence of this situation would be greater social inequality, with those that know English having greater access to high status social positions than those without it. This situation has led the Swedish government to commission the Swedish Language Council, a semi-public Swedish language planning body, to draft a program designed for the promotion and protection of Swedish.

The situation in Sweden is indicative of many European countries today, in which the fear of the growing use of English is in sharp contrast to the prevalent belief that knowledge of English provides access to the global economy. It is this ambivalent attitude that fuels countries to require the study of English while at the same time jealously protecting their own national language.

Conclusion

The social and sociolinguistic context of language learning and teaching has a significant impact on which languages are taught, when they are taught, and how they are taught. This fact has several implications for second and foreign language professionals. First, second and foreign language professionals, no matter which language they are teaching, need to work vigorously to ensure that all individuals be given the opportunity to become multilingual in an increasingly multilingual and multicultural world and to maintain the linguistic resources they have. In order to do this, they need to voice their disapproval of any policies that minimize those opportunities. Second, language professionals need to be sensitive to the local social and sociolinguistic context and to implement language teaching goals and methods that complement the social reality of their language learners. Finally, language professionals need to work to see that all learners have equal opportunities to achieve their language learning goals, so that they can reap the social and economic benefits that come from being a bilingual in the current global culture.

REFERENCES

Annamalai, E. (2004). Medium of power: The question of English in education in India. In J. W. Tollefson & A. Tsui (eds.), Medium of instruction policies: Which agenda? Whose agenda? (pp. 117-94). Mahwah, NJ: Lawrence Erlbaum.

Brutt-Griffler, J. (2002). World English: A study of its development. Clevedon, UK: Multilingual Matters.

Chick, K. & McKay, S. L. (2001). Teaching English in Multiethnic schools in the Durban area: The promotion of multilingualism or monolingualism. South African Linguistics and Applied Language Studies 19, 3 -4, 163 -78.

Chng, H. H. (2003). "You see me no up": Is Singlish a problem? Language Problems and Language Planning 27, 47-62.

Chomsky, N. (1957). Syntactic structure. The Hague: Mouton.

Crystal, D. (1997). English as a global language. Cambridge: Cambridge University Press.

Ferguson, C. (1959). Diglossia. Word 15, 325-40.

Firth, A. (1996). The discursive accomplishments of normality. On "lingual franca" English and conversation analysis. Journal of Pragmatics 26, 237-59.

Fishman, J. (1972). The socoiology of language. Rowley, MA: Newbury House.

Foley, J. A., Kandiah, T., Bao, Z., Gupta, A. E., Alsagoff, L., Ho, C. L., Wee, L., Talib, I. S., & Bokhorst-Heng, W. (eds.) (1998). English in new cultural contexts. Singapore: Oxford University Press.

Gill, S. (2005). Language policy in Malaysia: Reversing direction. Language Policy 4, 24160.

Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press.

Gupta, A. (1994). The step-tongue: Children's English in Singapore. Clevedon, UK: Multilingual Matters.

Hu, G. (2005). English language education in China: Policies, progress and problems. Language Policy 4, 5-24.

Hult, F. M. (2004). Planning for multilingualism and minority language rights in Sweden. Language Policy 3, 181-201.

Hyltenstam, K. & Abrahamsson, N. (2001). Age and L2 learning: The hazards of matching practical "implications" with theoretical "facts." TESOL Quarterly 35, 151-70.

Hymes, D. (1972). On communicative competence. In J. Pride & J. Holmes (eds.), Sociolin-guistics (pp. 269-93). Harmondsworth: Penguin.

Hymes, D. (1974). Foundations of sociolinguistics. Philadelphia: University of Pennsylvania Press.

Kachru, B. B. (1986). The alchemy of English. Oxford: Pergamon Press.

Kachru, B. B. (2005). Asian Englishes: Beyond the canon. Hong Kong: Hong Kong University Press.

Kamwangamalu, N. M. (2003). Globalization of English, and language maintenance and shift in South Africa. International Journal of the Sociology of Language 164, 65-81.

Kramer-Dahl, A. (2003). Reading the "Singlish Debate": Construction of a crisis of language standards and language teaching in Singapore. Journal of Language, Identity, and Education 2, 159-90.

Krishnaswamy, N. & Krishnaswamy, L. (2006). The story of English in India. Delhi: Foundation Books.

Lim, L. (ed.) (2004). Singapore English: A grammatical description. Philadelphia: John Benjamins.

McKay, S. L. & Wong, S. C. (2000). Immigrants in the United States. Cambridge: Cambridge University Press.

Myers-Scotton, C. (1993). Social motivation for codeswitching. Oxford: Claredon Press.

Nettle, D. & Romaine, S. (2000). Vanishing voices: The extinction of the world's languages. Oxford: Oxford University Press.

Pakir, A. (1991). The range and depth of English-knowing bilinguals in Singapore. World Englishes 10, 167-79.

Parasher, S. V. (1994). Indian English: Certain grammatical, lexical and stylistic features. In R. K. Angihotri & A. L. Khanna (eds.), Second language acquisition: Socio-cultural and linguistic aspects of English in India (pp. 145-64). New Dehli: Sage.

Phillipson, R. (2001). English for globalization or the world's people? International Review of Education 47, 185-200.

Phillipson, R. (2003). English-only Europe? Challenging language policy. New York: Routledge. Platt, J. T. & Weber, H. (1980). English in Singapore and Malaysia: Status, features, functions.

Kuala Lumpur: Oxford University Press. Rampton, B. (1995). Crossing: Language and ethnicity among adolescents. New York: Longman. Rampton, B. (1997). Second language research in late modernity. Modern Language Journal 15, 3, 329-33.

Rubdy, R. (2001). Creative destruction: Singapore's Speak Good English Movement. World Englishes 20, 341-55.

Seidlhofer, B. (2004). Research perspectives on teaching English as a lingua franca. Annual

Review of Applied Linguistics 24, 209-39. Smolicz, J. J. & Secombe, M. J. (2003). Assimilation or pluralism? Changing policies for

minority languages education in Australia. Language Policy 2, 3-25. Tickoo, M. L. (1996). Fifty years of English in Singapore: All gains, (a) few losses? In J. A. Fishman, A. W. Conrad, & A. Rubel-Lopez (eds.), Post-imperial English: Status change in former British and American colonies. 1940-1990. (pp. 431-55). Berlin and New York: Mouton de Gruyter.

Wardhaugh, R. (1992). An introduction to sociolinguistics. Cambridge, MA: Blackwell. Widdowson, H. (1994). The ownership of English. TESOL Quarterly 28, 377-88.

3 The Politics and Policies of Language and Language Teaching

ROBERT PHILLIPSON AND TOVE SKUTNABB-KANGAS

As long as we have the language, we have the culture. As long as we have the culture, we can hold on to the land.

Manu Metekingi, Maori from the Whanganui iwi ("tribe")1

Language and culture cannot be separate from each other - if they are, the language only becomes a tool, a thing... Our language and culture are our identity and tell us who we are, where we came from and where we are going.

Task Force on Aboriginal Languages and Cultures, 2005, p. 58

To give millions a knowledge of English is to enslave us. The foundation that Macaulay laid of education has enslaved us.

Gandhi, barrister-at-law, architect of Indian independence, 1908,

quoted in Naik, 2004, p. 255

It was because we were taught in our own language that our minds quickened.. .If the whole mind is not functioning from the beginning its full powers remain undeveloped to the end. While all around was heard the cry for English teaching, my third brother was brave enough to keep us to our Bengali course.

Rabindranath Tagore, 1913 Nobel Prize Laureate for Literature, 1992, pp. 53-4

The English language teaching sector directly earns nearly £1.3 billion for the UK in invisible exports and our other education related exports earn up to £10 billion a year more.

Neil Kinnock, Chair British Council, politician, ex-EU Commission

Vice-President, in Graddol, 2006, p. 4

English in Africa: an imperial language, the language of linguistic Americanization, a language of global capitalism . . . creating and maintaining social divisions serving an economy dominated primarily by foreign economic interests and, secondarily, by a small aspiring African bourgeoisie.

Alamin Mazrui, PhD Stanford, expert on World Bank language policies, 2004, p. 30

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty ©2009 BkckwellPublishDgLtd. ISBN: 978-1-405-15489-5

One feels so handicapped working from here, and also in a language not one's own. The proponents of a worldwide right and duty to learn English will never admit the cost for the non-native speakers on whom the World English is imposed.

Ajit Mohanty, PhD Edmonton, cutting-edge bilingualism researcher,

personal correspondence, June 8, 2004

English is not enough. We are fortunate to speak a global language but, in a smart and competitive world, exclusive reliance on English leaves the UK vulnerable and dependent on the linguistic competence and the goodwill of others... Young people from the UK are at a growing disadvantage in the recruitment market.

Nuffield Languages Enquiry, 2000, www.nuffield.org

Unhappily for those who have sought to devise a "science" of Language Policy and Planning there are no protocols for doing or designing LPP that can be induced from practice, abstracted, tested and refined into procedures and then transferred across contexts and applied in diverse settings . . . the field is too dependent on the descriptive traditions of linguistics from which it derives, and insufficiently in communication with policy analysis sciences, with political science, with sociology and with critical schools of thought.

Joseph Lo Bianco, PhD Melbourne, language policy expert, 2002, pp. 23, 25

Introduction

These glimpses into the world of language policy demonstrate that educational language policy and its analysis are of major significance for the individual, for group vitality, for national identities, and international relations. They also reveal how language policy interlocks with political decision-making. The intrinsic complexity of language policy, in practice and in theory, is increasingly being addressed in books and journals. Language policy requires a multi-disciplinary approach. The concerns of educationalists and language professionals need to draw on insights from political science, international law, and economics. Even within educational language studies, there is a tendency toward excessive specialization. For instance, books on language policy (e.g., Ricento, 2006; Spolsky, 2004) generally ignore translation studies, a productive field of activity that interlocks with language technology, globalization (Cronin, 2003), and conflict (Baker, 2006). Our review will focus on historical contextualization and conceptual clarification, and build on examples of ongoing language policy issues at the global, regional, and national levels.

Committed scholarship should ensure that language policy experience and analysis contribute to the resolution of complex political problems. This requires that academic, educational, and political discourse interact dialectically. It also presupposes recognition of the value of all the world's languages and their use in addressing life's challenges: in the words of the Kenyan Nobel Peace Prize Laureate, Wangari Maathai, referring to the powerful addressing the population

at large: "if you don't speak in their language, you may touch the head, but you may never touch the heart. And that is what a mother tongue does . . . if you lose your language, you lose yourself."2

Historical and Global Contextualization

Foreign languages were traditionally learned by the privileged for the social prestige they represented and facilitated, whether the canonical texts of Arabic, Latin, or Sanskrit, or more recent internationally influential languages. The Russian aristocracy was literate and vocal in French but often had little proficiency in Russian, the language of the masses. It is arguable that in African and Asian former colonies, English may now be functioning rather as did French in nineteenth-century Russia. The main difference is that the class of English users worldwide is now connected to a neoliberal economic system ("the new imperialism," Harvey, 2005) and ideological systems, forms of consciousness, consumption, and growthism that cultural globalization is propagating (Seabrook, 2004). The penetration of contemporary English is thus vastly deeper than that of Latin, French, and other dominant languages in earlier times (see Gandhi, Tagore, and Mohanty above). It is also of considerable economic importance in itself (see Kinnock above), though there are risks if one does not know other languages (see Nuffield Foundation above). English is central to corporate globalization spearheaded by economic, political, and military forces in the "English-speaking" countries (Mazrui above).

The incorporation and co-option of local elites follows the pattern by which the privileged in colonies learned the language of the invader. This was already the case in Roman England (Tacitus, [AD 97] 1948). In Europeanized settler colonies (the Americas, Australasia, South Africa) educational policies were elaborated ad hoc, often by missionary societies, following or preceding military suppression. Although in some parts of the USA in the early nineteenth century there was a bilingual education policy, "throughout the 19th and much of the 20th centuries, federal Indian education policy was one of almost zero tolerance for linguistic and cultural difference" (McCarty, Romero, & Zepeda, 2006, p. 94). In the extraction colonies of the British empire (such as Malaysia or West Africa), the learning of the colonial language was paramount for local elites and intermediaries, as in the French and Portuguese colonies, and has remained a significant legacy in the postcolonial age. In Finland, where 70 percent of the population, mostly Finnish-speakers, were not represented in the parliament during the Russian era (1807-917), the first use of Finnish in the parliament (in 1894, by a representative of the gentry) was called "unbelievably barbaric" and "illegal." The language of parliament was not Russian but Swedish, the language of the former colonial power (1155-1809) and all Finnish-speakers were supposed to know it. Linguistic hierarchization (linguicism) and genocide were the norm in settler colonies, and have also been vigorously pursued by monolingually oriented states elsewhere (Skutnabb-Kangas, 2000).

In the twentieth century, "modern," "living," or "foreign" language learning progressively shifted from a predominantly literary focus to more instrumental purposes. However, departments of English in former colonies generally retain literature as a core concern (on India, see Joshi, 1994; on Sri Lanka, Kandiah, 1999), emulating departments of English as a mother tongue in the major English-using countries. Many "foreign" language departments are similar. Thus the Modern Language Association of the USA (www.mla.org) brings together the study of many languages, but its "divisions" and "discussion groups" are overwhelmingly concerned with literature, and pedagogy is marginal.

The MLA claims leadership in the national education community (i.e., the USA), but also records 30,000 members in 100 countries. In this respect the professional association is comparable to TESOL (a mainly US body that claims to serve "a global group of devoted teachers, educators, researchers, and friends of the profession," www.tesol.org) and its British-based equivalent, IATEFL (whose "mission is to link, develop and support English Language Teaching professionals throughout the world," www.iatefl.org). Even if TESOL and IATEFL generally represent language learning constituencies that are often seen as marginal to mainstream education, their professionalism exercises considerable influence globally on the content of education, teacher training, theoretical paradigms, and methodological innovation, with major cultural consequences. Journals reflect this hegemony: "... most of what gets published in the most influential applied linguistics journals is generally a product of what we could term the Anglo-American centre. It is primarily academics from the academic centre, above all in North America, who are funded and encouraged to do research" (Block, 1997, p. 3). Considerable amounts of US funding are also directed toward scholars worldwide, which can result in reciprocal stimulation, as for instance in a recent book on language policy in China (Zhou and Sun, 2004), but always with the risk of inequality reinforcing Western paradigms, orientalism, and academic imperialism (for criticism of these, see Odora Hoppers, 2002; Smith Tuhiwai, 1999). The handbook in your hands, which aims at global relevance, incurs the same risk.

Capitalism deifies the law of the market, and requires continuous expansion that does not respect national borders. English currently oils the wheels of most global finance and commerce. Educational services are increasingly seen as commodities and market opportunities, rather than as a human right and public good that each country is responsible for. Thus Educational Testing Services of New Jersey, a hugely profitable "non-profit" body, announces its global ambitions proudly (www.ets.org):

As ETS's wholly-owned subsidiary, ETS Global BV is structured to bring ETS's expertise and experience with tests, assessments, and related services to educational and business communities around the world. ETS Global BV now has subsidiaries in Europe and Canada, and it will be expanding into other countries and regions as well . . . Our global mission goes far beyond testing. Our products and services enable opportunity worldwide by measuring knowledge and skills, promoting learning and performance, and supporting education and professional development for all people worldwide.

This linguistic-educational imperial thrust is viscerally connected to corporate-driven globalization. In contrast, human rights, especially economic and social rights, serve, according to human rights lawyer Katarina TomaSevski (1996, p. 104), to act as correctives to the "free" market, instead of giving market forces free range. Tomasevski claims that "The purpose of international human rights law is . . . to overrule the law of supply and demand and remove price-tags from people and from necessities for their survival" (1996, p. 104). The necessities for survival include not only economic and social rights such as basic food and housing, but also basic civil, political, and cultural rights necessary for a dignified life. Basic linguistic human rights in education belong to these. Thus all second and foreign-language teaching that promotes the subtractive spread of "big" languages violates basic human rights (see below).

The distribution of European languages worldwide reflects a global European-ization process. Other major languages, such as Arabic and Japanese, are also promoted globally. The learning of Chinese worldwide is being promoted in Confucius Institutes (following the trail of Cervantes, Dante, and Goethe Institutes) and anticipates an increase from 30 million learners to 100 million within a few years (Graddol, 2006, p. 63). Figures for speakers of a first and especially of a second language are notoriously unreliable (see Skutnabb-Kangas, 2000, pp. 30-46), or fraudulent (a French example, Chaudenson, 2003), not least because of imprecise definitions of language (as well as dialect, sociolect, etc.) and proficiency. Migration, urbanization, and technological developments are making the linguistic mosaic ever more complex, and leading to new forms of multilin-gualism, at the same time as oppressive state language policies continue to erode linguistic diversity.

Politics and/or Policy?

There are no quick fixes in language policy (see Lo Bianco above) or politics. In English there are the two lexical items policy and politics, whereas many languages have a single term for both (Danish politik, French politique, etc.). This tends to blur the distinction between the two concepts. If a user of English as a second language refers to the "language politics" of an institution, what is intended may be the language policy in force. The "politics of language" often indicates political struggles in which linguistic identity and language rights may be contentious, polarizing factors. Language policy can be defined as referring to all the measures, explicit and implicit, which have an impact on the language ecology in a given context, including the rights of speakers of a given language, and the use made of languages for given functions. It is thus a broader term than language planning. The politics of language refers to the political domain and its discourses, what politicians do or say, language promotion or suppression, and to struggles for rights or recognition for a given language. A challenge to scholars is to investigate how their (our?) language policy and politics activities relate

to political power nationally and internationally, and their/our function in upholding a globally oppressive system. A good example of failure to do so occurs when it is claimed that "international" English is a neutral language, disconnected from the power and powers behind it:

English being disembedded from national cultures can never mean that it floats culture-free ( . . . or) is culturally neutral. The point may be simple, but it is often elided; and this elision constitutes a politics of English as a global language which precisely conceals the cultural work which that model of language is in fact performing. (Kayman, 2004, p. 17)

Kayman also makes the intriguing point that the prophets and proponents of English as a global language can be compared to Europeans occupying other continents that were falsely seen as terra nullius. Contemporary linguists who proclaim the neutrality of English treat the language as a cultural terra nullius (Kayman, 2004, p. 18). The forms and functions of "global" English have been extensively analysed (see the review of three books in Phillipson, 2004), including whether its present position is unassailable (Graddol, 2006).

Mainstream political science is archetypically represented by Huntington's The Clash of Civilizations and the Remaking of the World Order (1996), essentially a blueprint for global US dominance. Language, and the cultural universe and ways of thought it embodies, is a key dimension to this global mission. David Rothkopf, director of Kissinger Associates, wrote an article entitled "In praise of cultural imperialism?" in Foreign Policy in 1997 (pp. 45, 48):

It is in the economic and political interest of the United States to ensure that if the world is moving toward a common language, it be English; that if the world is moving toward common telecommunications, safety, and quality standards, they be American; and that if common values are being developed, they be values with which Americans are comfortable. These are not idle aspirations. English is linking the world . . . Americans should not deny the fact that of all the nations in the history of the world, theirs is the most just, the most tolerant, the most willing to constantly reassess and improve itself, and the best model for the future.

It is false to regard globalization as a new phenomenon. There have been blueprints for US dominance of the two American continents since the Monroe Doctrine of 1823 and for global domination for over a century. The need for new markets due to capital over-accumulation was a primary concern of US foreign policy throughout the twentieth century. The role of scholarship in legitimating the thrust for global dominance is explored in American Empire: Roosevelt's Geographer and the Prelude to Globalization (Smith, 2003). Geography is inseparable from economics, politics, and international affairs. The promotion of English globally is part and parcel of this process: it is integral to globalization. In political discourse the dominant economic system of capitalism has been conflated with "democracy" and "freedom."

Linguistic Human Rights, Linguistic Diversity, and Language Maintenance in and through Education: Issues of Language Policy and Politics

In today's world, it is impossible to imagine linguistic diversity consisting of more or less monolingual groups living in isolation from all others. Most probably this has never been the case. Even groups with very high degrees of self-sufficiency, and little risk of ecological disasters such as drought, have traded with each other, and many in the group have known the neighboring language(s) (see, e.g., Corballis, 2002; Diamond, 1998; Nettle, 1999; Wells, 2002). Many, if not most speakers of numerically small languages are necessarily multilingual as adults, and some degree of individual multilingualism is a prerequisite for them for participation in the wider society. Few are today decisively against individual multilingualism, once it has been achieved, while many are trying to prevent children from achieving high levels of bi-or multilingualism, for instance through depriving them of mother tongue medium education. In contrast, many are against societal multilingualism, and, especially, any linguistic human rights for Indigenous peoples and minorities that would enable them to maintain their languages and reproduce themselves as minorities. And many are against the world's linguistic diversity, claiming it is problematic and not cost-effective (see Skutnabb-Kangas, 2007, for an overview of language policy and language rights).

Many researchers plead directly or indirectly for the elimination of both mul-tilingualism and linguistic diversity. They claim it is unnecessary, messy, costly and inefficient (for some examples, see de Swaan, 2001; Glazer, 1997; Kymlicka, 2001; Ladefoged, 1992; Kymlicka & Patten, 2003). To label linguistic diversity as a complication, obstacle, or problem is to deny and lament not suggestions or dreams, but facts. With very few exceptions, the world's countries are multilingual, and, with Debi Pattanayak's subtle Indian understatement, "[o]ne language is an impractical proposition for a multilingual country" (1988, p. 382). The teaching profession often legitimates the normalization of English as The Preferable (and Only) Language of the World, partly by overemphasizing its usefulness and importance, partly by invisibilizing other languages (not mentioning them or belittling them). English can also, paradoxically, be invisibilized when it is presented as the self-evident default norm, as in a Peace Corps advertisement where one FAQ (Frequently Asked Question) to USA applicants was: "Do you need to know a language?" The answer was "No." English was obviously not seen as "a language."

Why is linguistic diversity needed? The following list gives a short summary of reasons that have been presented by various researchers:

• Languages have been called the libraries of the intangible heritage of humankind, in terms of both form (diversity of ways of structuring a language, and of the underlying cognitive categories and processes) and content (most of

humankind's knowledge is encoded in languages; grammars have been called "fossilized experience").

• For many people, languages are cultural core values, central for their identities (see Metekingi, 2003-6; and Task Force, 2005).

• Creativity precedes innovation, which is followed by investment. Creativity can be one of the results of additive teaching and multilingualism. Creativity and new ideas are the main assets (cultural capital) in a knowledge society and a prerequisite for humankind to adapt to change and to find solutions to the catastrophes of our own making. The more linguistically and culturally diverse the world is, the more new ideas and creativity of various kinds are likely to exist. High levels of multilingualism may enhance creativity; mono-lingualism and homogenization kill it.

• English is not enough. "Good" English will fairly soon be like literacy yesterday or computer skills today: employers see it as self-evident and necessary but not sufficient for good jobs. Supply and demand theories predict that when many people possess what earlier was a scarce commodity (near-native English), the price goes down (Grin, 2001). The value of "perfect" English skills as a financial incentive decreases substantially when a high proportion of a country's or a region's or the world's population knows English well.

• Linguistic diversity and biodiversity are correlationally and probably also causally related. Knowledge about how to maintain biodiversity is encoded in small languages because it is their speakers who live in the world's biologically (and linguistically) most diverse areas. Through killing these languages (or letting them die), we kill many of the prerequisites for maintaining biodiversity (see Harmon, 2002; Maffi, 2001; Posey, 1999; Skutnabb-Kangas, Maffi, & Harmon, 2003; Skutnabb-Kangas & Phillipson, 2007, for overviews).

• Finally, Colin Baker sums up the importance of ecological diversity in his review of Skutnabb-Kangas 2000 (Baker, 2001, p. 281):

Ecological diversity is essential for long-term planetary survival. Diversity contains the potential for adaptation. Uniformity can endanger a species by providing inflexibility and unadaptability. As languages and cultures die, the testimony of human intellectual achievement is lessened. In the language of ecology, the strongest ecosystems are those that are the most diverse. Diversity is directly related to stability; variety is important for long-term survival. Our success on this planet has been due to an ability to adapt to different kinds of environment over thousands of years. Such ability is born out of diversity. Thus language and cultural diversity maximises chances of human success and adaptability.

Biocultural diversity (biodiversity + linguistic diversity + cultural diversity) is thus essential for long-term planetary survival because it enhances creativity and adaptability and thus stability. Today we are killing biocultural diversity faster than ever before in human history.

Schools can, even in one generation, make the intergenerational transfer of small Indigenous and minority languages impossible, and thus make the languages

seriously endangered (not learned by the next generation of children). The educational system participates in committing linguistic and cultural genocide, according to two of the five definitions of genocide (II(b) and II(e)) in the United Nations International Convention on the Prevention and Punishment of the Crime of Genocide (E793, 1948). Education offered to most Indigenous and many minority children is intentionally "forcibly transferring children of the group to another group" and "causing serious bodily or mental harm to members of the group" (emphasis added). This is done through assimilationist subtractive submersion education, using a dominant language - instead of the children's own language - as the teaching language (see Skutnabb-Kangas, 2000; Magga et al., 2005, and references in both, for evidence and further discussion). Without binding educational linguistic human rights, especially a right to mainly mother tongue-medium education in state schools, with good teaching of a dominant language as a second language, given by competent bilingual teachers, most Indigenous peoples and minorities have to accept subtractive education through the medium of a dominant/majority language. They learn a dominant language at the cost of the mother tongue(s). These are displaced, and later often replaced by the dominant language. Subtractive teaching subtracts from the child's linguistic repertoire, instead of adding to it. For linguistic genocide to be stopped and for the linguistic diversity of the world to be maintained, basic linguistic human rights (LHRs) are necessary, especially in education.

The most important LHR with these aims is the right to use one's own language as the main teaching language, together with the right to learn a dominant official language and thus to become high-level bilingual or multilingual. There is today no universal right guaranteeing this. Language is seen in most human rights documents as one of the important "characteristics" on the basis of which people are supposed not to be discriminated against, a negative right. But when one moves from the prefaces of the documents to educational rights, language is often not mentioned, and if it is, the provisions are extremely vague or conditional and there are so many modifications and "claw-backs" that the rights are virtually meaningless (see Skutnabb-Kangas, 2000, for thorough exemplification). Two recent European HRs instruments3 offer some possibilities, not so much because of their formulations, which are often weak,4 but mainly because the monitoring committees, which examine the reports that participating states have to write, have tried to prompt states to action. The Organization for Security and Cooperation in Europe and some other organizations (including UNESCO) have issued recommendations and/or guidelines that are steps in the right direction, but these are in no way binding.

Even if they were binding, it is doubtful to what extent they would be implemented. This is an example of how policy in education (a decision to use the mother tongue as the teaching language) interlocks with the politics of the economy (free or for-fee education). Most education of Indigenous and minority children is not even starting to fulfill the requirements posed by Katarina Tomasevski in terms of "the four a's": education has to be available, accessible, acceptable, and

adaptable. TomaSevski, the former United Nations Special Rapporteur on the Right to Education, examines 170 countries in her last report The State of the Right to Education Worldwide. Free or Fee: 2006 Global Report (www.katarinatomasevski. com) to see to what extent education is free or for-fee. Even primary education is fee-paying in more than half of the countries (see her Table 25); thus education is not available to children. The pattern of economic (poverty-based) exclusion from primary school is part of a global strategy for "no poverty reduction." Education is often priced out of the reach of the poor. The trend has been a transition from free-and-compulsory to market-based education where the costs of even primary education have been transferred from governmental to family budgets. But even if children could afford primary education, being forced to accept teaching in a language that is not the (Indigenous or minority) language makes the teaching non-accessible, except in the most trivial sense of children possibly understanding a few contextualized issues and learning some aspects of concrete everyday language. This kind of education often fits well with the conclusions of a study from Zambia: "There is a clear risk that the policy of using English as a vehicular language may contribute to stunting, rather than promoting, academic and cognitive growth" (Williams, 1998). There has to be a major change in the politics of education before we can even start to discuss meaningful language policies in education for Indigenous and (autochthonous or immigrant) minority children, or, for that matter, many other children from subordinated groups/ peoples.

Some changes are on their way. The pessimistic prognoses state that minimally 50 percent, maybe up to 90-95 percent, of today's spoken languages may be extinct or very seriously endangered (not learned by children) around 2100. Language revitalization movements are growing all over the world. The Maori language nests (pre-schools where fluent elders are teaching language and culture to children, parents, and even pre-school teachers) have spread from Aotearoa to Hawai'i, to the Saami areas in Norway and Finland, to Canada, etc. Indigenous immersion programs for revitalization are likewise spreading to many Indigenous peoples, for instance in Canada and the USA, where the languages have been partially lost (on concepts, see Skutnabb-Kangas & McCarty, 2008). There are discussions about redefining mother tongues so that Indigenous individuals whose parent or grandparent generation has experienced linguistic genocide through education could both claim a mother tongue on the basis of identification only, without knowing this mother tongue, and demand compensation for the language lost. In Papua New Guinea, in India, especially Orissa, in Nepal, in Vietnam, in many African countries, and several other places where the languages are still alive, Indigenous/tribal children are starting to have their own languages as teaching/learning languages while at the same time learning official languages. The United Nations Permanent Forum on Indigenous Issues (UNPFII, http://www.un.org/esa/socdev/unpfii/) is making language regenesis, reclamation, revival, revitalization, maintenance, and further development one of the focal program points.

Language Policy, Exemplified by the European Region

Language policies are necessary for the European Union, which is undergoing an intensive process of integrating 27 member states. Policies in education, culture, and language have traditionally been the prerogative of each state. Since 1992 the EU has had a mandate to "supplement" this through a considerable number of programs, such as student and staff mobility, language learning, and awareness raising, as well as funding for minority languages (http://ec.europa.eu/educa-tion/policies/lang/languages_en.html). Currently the EU advocates the learning of two foreign languages by all schoolchildren, building on curriculum development work by the Council of Europe over several decades, such as the Common European Framework of Reference for Languages (www.coe.int/T/DG4/ Linguistic/Default_en.asp). Levels of foreign-language competence vary greatly between the different member states. The supra-statal initiatives show innovation being promoted by bureaucratic elites rather than a profession's internal dynamics. Reform in foreign-language learning is very slow globally, despite some imaginative initiatives, for instance, Content and Language Integrated Learning (CLIL), one of the EU's mantras for achieving more success in foreign-language learning, or the Cultures and Languages Across the Curriculum (CLAC) Movement in the USA (www.clas.pdx.edu/clac/). Another EU language mantra is an early start to foreign-language learning, a notion that is intuitively appealing, hence easy for politicians to latch on to and promote. Experience shows, however, that considerable educational changes need to be in place before the age factor can influence outcomes decisively (see Phillipson, 2003, pp. 95-104). Persuading Europeans to learn two foreign languages represents an attempt to diversify the languages learned.

The EU has commissioned studies of a range of language pedagogy issues, e.g., of foreign-language teacher training in 32 European countries, including case studies of good practice, bilingual education, and a profile of the language teacher of the future (Grenfell, Kelly, & Jones, 2003). Other studies argue persuasively for a paradigm shift in foreign-language education (Dendrinos and Mitsikopoulou, 2004). Unlike the bland, technocratic discourse of many EU policy statements, such authors see it as axiomatic that discourses on language policy and foreign-language education are neither ideologically nor politically neutral.

Among multiple market forces strengthening English is the Bologna process, the integration of higher education and research across 45 countries. The 1999 Bologna objectives declare the intention "within the framework of our institutional competences and taking full respect of the diversity of cultures, languages, national education systems and of University autonomy," to consolidate a European Higher Education Area at the latest by 2010. The ministerial meeting at Bergen, Norway on 19-20 May 2005 (www.bologna-bergen2005.no) focused on the coordination of structural uniformity (a standardized degree structure in 45 countries), quality assurance, the recognition of degrees and study

periods, attractiveness, and competitiveness. What is striking is that not once in the communiqué is there any reference to languages, to bilingual degrees or multilingualism. What emerges unambiguously is that in the Bologna process, "internationalization" means "English-medium higher education" (Phillipson, 2006).

This clearly entails the risk of what political discourse in several European countries refers to as "domain loss." This is a seemingly innocuous but deceptive term. Like the language policy term "language spread," or "language death," it seems to imply a natural, agent-less process (see Skutnabb-Kangas & McCarty, 2008, for conceptual clarification). Clearly there are agents involved in domains being lost or gained, in languages "disappearing" and in any diglossic division of academic labor. The process of domain loss can be seen as linguistic capital accumulation by dispossession. As in the commercial world in its global pursuit of markets and profit (Harvey, 2005), some combination of internal motivation and external pressure contributes to this trend.

Users of English benefit from these processes, just like speakers of all killer languages that are learned at the expense of small disappearing languages. These benefits are exemplified in a study of foreign languages as public policy in education, commissioned by the Haut Conseil de l'évaluation de l'école, Paris (Grin, 2005). The study calculated that the current dominance of English in continental European education systems results in quantifiable privileged market effects, communication savings effects, language learning savings effects, alternative human capital investment effects, and legitimacy and rhetorical effects. Continental European countries are transferring to the UK and Ireland at least €10 billion per annum, and more probably about €16-17 billion. The UK and Ireland benefit since they invest so little in foreign-language learning as compared with their EU partners. If this European study in the economics of language were to be extrapolated globally, it would show a massive global transfer of not only linguistic but economic capital from the rest of the world to the few English-dominant countries. Voices are being raised that demand more equality in sharing the communication burden: some kind of compensation is due to those who go to the expense of learning the languages of more powerful people. Exposing this aspect of the politics of language teaching is an urgent task for applied linguists and economists of language.

Grin has assessed what the costs and benefits are of a language policy which maintains and promotes minority languages, and what the costs (and benefits) are if they are not. Some of his encouraging conclusions, which we endorse, are as follows (Grin, 2003, p. 26):

• diversity seems to be positively, rather than negatively, correlated with welfare;

• the available evidence indicates that the monetary costs of maintaining diversity are remarkably modest;

• devoting resources to the protection and promotion of minority cultures [and this includes languages] may help to stave off political crises whose costs would be considerably higher than that of the policies considered;

• therefore, there are strong grounds to suppose that protecting and promoting regional and minority languages is a sound idea from a welfare standpoint, not even taking into consideration any moral argument.

1 The quote from Manu Metekingi comes from a film shown at the Whanganui Iwi Exhibition, at Te Papa Tongarewa Museum of New Zealand, Wellington, 29 November, 2003 - May, 2006. The Exhibition tells about "our heartland, the Whanganui River, and our place within it." The Whanganui iwi write: "The well-being of our river is intertwined with its people's well-being" (from the brochure describing the exhibition, with the theme: "Ko au te awa, ko te awa ko au. I am the river, the river is me"). Thanks to the staff at Te Papa for identifying the person for us - neither the quote nor his name is in the brochure, only in the film.

2 Talk at University of California, Irvine, 20 March 2006, see http://humanities.uci.edu/ faultline/.

3 The Framework Convention for the Protection of National Minorities (http://conventions. coe.int/treaty/en/Treaties/Html/157.htm) and the European Charter for Regional or Minority Languages (http://conventions.coe.int/treaty/en/Treaties/Html/148.htm).

4 An example is the Framework Convention's Article covering medium of education. It is so heavily qualified that the minority is completely at the mercy of the state (emphases added): "In areas inhabited by persons belonging to national minorities traditionally or in substantial numbers, if there is sufficient demand, the parties shall endeavour to ensure, as far as possible and within the framework of their education systems, that persons belonging to those minorities have adequate opportunities for being taught in the minority language or for receiving instruction in this language."

REFERENCES

Baker, C. (2001). Review of Tove Skutnabb-Kangas Linguistic Genocide in Education - or Worldwide Diversity and Human Rights? Journal of Sociolinguistics 5, 2, 279-83.

Baker, M. (2006). Translation and conflict: A narrative account. London: Routledge.

Block, D. (l997). Publishing patterns and McDonaldization. IATEFL Newsletter (April/ May), 12-15.

Chaudenson, R. (2003). Geolinguistics, geopolitics, geostrategy: The case for French. In J. Maurais and M. A. Morris (eds.), Languages in a globalising world (pp. 291-7). Cambridge: Cambridge University Press.

Corballis, M. C. (2002). From hand to mouth: The origins of language. Princeton/Oxford: Princeton University Press.

Cronin, M. (2003). Translation and globalization. London: Routledge.

Dendrinos, B. & Mitsikopoulou, B. (eds.) (2004). Policies of linguistic pluralism and the teaching of languages in Europe. Athens: Metaixmio, and the National and Kapodistrian University of Athens.

de Swaan, A. (2001). Words of the world: The global language system. Cambridge: Polity.

Diamond, J. (1998). Guns, germs and steel: A short history of everybody for the last 13,000 years. London: Random House.

Glazer, N. (1998). We are all multiculturalists now. Cambridge, MA: Harvard University Press.

Graddol, D. (2006). English next? Why Global English may mean the end of "English as a Foreign Language." London: The British Council. (downloadable at www.britishcouncil.org/ learning-research)

Grenfell, M., Kelly, M., & Jones, D. (2003). The European language teacher: Recent trends and future developments in teacher education. Bern: Peter Lang.

Grin, F. (2001). English as economic value: Facts and fallacies. World Englishes 20, 2, 65 -78.

Grin, F. (2003). Language planning and economics. Current Issues in Language Planning 4, 11, 1-66.

Grin, F. (2005). L'enseignement des Langues Étrangères comme Politique Publique. Paris: Haut Conseil de l'Évaluation de l'École.

Harmon, D. (2002). In light of our differences: How diversity in nature and culture makes us human. Washington, DC: The Smithsonian Institute Press.

Harvey, D. (2005). The new imperialism. Oxford: Oxford University Press. (First edition published 2003.)

Huntington, S. (1996). The clash of civilizations and the remaking of the world order. New York: Simon and Schuster.

Joshi, S. (ed.) (1994). Rethinking English: Essays in literature, language, history. Delhi: Oxford University Press.

Kandiah, T. (1999). Re-visioning, revolution, revisionism: English and the ambiguities of post-colonial practice. The Sri Lanka Journal of the Humanities 24 & 25, 31-64.

Kayman, M. A. (2004). The state of English as a global language: Communicating culture. Textual Practice 18, 1, 1-22.

Kymlicka, W. (2001). Politics in the vernacular: Nationalism, multiculturalism and citizenship. Oxford: Oxford University Press.

Kymlicka, W. & Patten, A. (eds.) (2003). Language rights and political theory. Oxford: Oxford University Press.

Ladefoged, P. (1992). Another view of endangered languages. Language 68, 4, 809-11.

Lo Bianco, J. (2002). Real world language politics and policy. In S. J. Baker (ed.), Language policy: Lessons from global models (pp. 8-27). Monterey, CA: Monterey Institute of International Studies.

Maffi, L. (ed.) (2001). On biocultural diversity: Linking language, knowledge and the environment. Washington, DC: The Smithsonian Institute Press.

Magga, O. H., Nicolaisen, I., Trask, M., Dunbar, R., & Skutnabb-Kangas, T. (2005). Indigenous children's education and indigenous languages. Expert paper written for the United Nations Permanent Forum on Indigenous Issues. New York: United Nations.

Mazrui, A. A. (2004). English in Africa: After the Cold War. Clevedon, UK: Multilingual Matters.

McCarty, T. L., Romero, M. E., & Zepeda, O. (2006). Reimagining multilingual America: Lessons from Native American youth. In O. Garcia, T. Skutnabb-Kangas, & M. E. Torres-Guzman (eds.), Imagining multilingual schools: Languages in education and glo-calization (pp. 91-110). Clevedon, UK: Multilingual Matters.

Metekingi, M. (2003-6). Film shown at the Whanganui Iwi Exhibition, at Te Papa Tongarewa Museum of New Zealand, Wellington, 29 November, 2003 - May, 2006.

Naik, C. (2004). India's language policy in an uncertain century. In P. Dias (ed.), Multiple languages, literacies and technologies: Mapping out concepts, analysing practices and defining positions (pp. 246-63). New Delhi: Books for Change.

Nettle, D. (1999). Linguistic diversity. Oxford: Oxford University Press.

Nuffield Foundation, The (2000). Languages: The next generation. The final report and recommendations of the Nuffield Languages Inquiry. London: The Nuffield Foundation.

Odora Hoppers, C. A. (ed.) (2002). Indigenous knowledge and the integration of knowledge systems: Towards a philosophy of articulation. Claremont: New Africa Books.

Pattanayak, D. P. (1988). Monolingual myopia and the petals of the Indian lotus: Do many languages divide or unite a nation? In T. Skutnabb-Kangas & J. Cummins (eds.), Minority education: From shame to struggle (pp. 379-89). Clevedon, UK: Multilingual Matters.

Patten, A. & Kymlicka, W. (2003). Introduction: Language rights and political theory: Context, issues, and approaches. In W. Kymlicka & A. Patten (eds.), Language rights and political theory (pp. 1-51). Oxford: Oxford University Press.

Phillipson, R. (2003). English-only Europe? Challenging language policy. London: Routledge.

Phillipson, R. (2004). Review article, English in globalization: Three approaches (books by de Swaan, Block and Cameron, and Brutt-Griffler). Journal of Language, Identity, and Education 3, 1, 73-84.

Phillipson, R. (2006). English, a cuckoo in the European higher education nest of languages? European Journal of English Studies 10, 1, 13-32.

Posey, D. A. (ed.) (1999). Cultural and spiritual values of biodiversity: A complementary contribution to the Global Biodiversity Assessment. London: Intermediate Technology Publications, for and on behalf of the/United Nations Environmental Programme.

Ricento, T. (ed.) (2006). An introduction to language policy: Theory and method. Oxford: Blackwell.

Rothkopf, D. (1997). In praise of cultural imperialism?' Foreign Policy 53, 38-53.

Seabrook, J. (2004). Consuming cultures: Globalization and local lives. London: New Internationalist.

Skutnabb-Kangas, T. (2000). Linguistic genocide in education - or worldwide diversity and human rights? Mahwah, NJ: Lawrence Erlbaum.

Skutnabb-Kangas, T. (2007). Language planning and language rights. In M. Hellinger & A. Pauwels (eds.), Linguistic diversity and language change (pp. 365-97), Handbooks of Applied Linguistics 9. Berlin: Mouton de Gruyter.

Skutnabb-Kangas, T., Maffi, L., & Harmon, D. (2003). Sharing a world of difference: The earth's linguistic, cultural, and biological diversity. Paris: UNESCO Publishing. UNESCO, Terralingua, and World Wide Fund for Nature (ISBN UNESCO 92-3-103917-2; also in Catalan, French, Korean and Spanish).

Skutnabb-Kangas, T. & McCarty, T. (2008). Clarification, ideological/epistemological underpinnings and implications of some concepts in bilingual education. In J. Cummins & N. Hornberger (eds.), Bilingual education, vol. 5 of Encyclopedia of Language and Education (2nd edn., pp. 3-17). Kluwer Academic.

Skutnabb-Kangas, T. & Phillipson, R. (2007). Language and ecology. In J.-O. Östman & J. Verschueren (eds.), in collaboration with Eline Versluys, Handbook of pragmatics. Amsterdam: Benjamins.

Smith, N. (2003). American empire: Roosevelt's geographer and the prelude to globalization. Berkeley and Los Angeles, CA: University of California Press.

Smith Tuhiwai, L. (1999). Deconstructing methodologies: Research and Indigenous peoples. Dunedin/New York: University of Otago Press/Zed Books.

Spolsky, B. (2004). Language policy. Cambridge: Cambridge University Press.

Tacitus (1948). Tacitus on Britain and Germany. A new translation of the 'Agricola' and the 'Germania', by H. Mattingly. Harmondsworth: Penguin.

Tagore, R. (1992). My reminiscences. New Delhi: Rupa.

Task Force on Aboriginal Languages and Cultures (2005). Towards a new beginning: A foundational report for a strategy to revitalize First Nation, Inuit and Métis languages and cultures. Report to the Minister of Canadian Heritage, June 2005. Ottawa: Aboriginal Languages Directorate. www.aboriginallanguagestaskforce.ca

Tomasevski, K. (1996). International prospects for the future of the welfare state. In Reconceptualizing the welfare state (pp. 100-17). Copenhagen: The Danish Centre for Human Rights.

Wells, S. (2002). The journey of man: A genetic odyssey. Princeton & Oxford: Princeton University Press.

Williams, E. (1998). Investigating bilingual literacy: Evidence from Malawi and Zambia. Education Research no. 24. London: Department For International Development.

Zhou, M. & Sun, H. (eds.) (2004). Language policy in the People's Republic of China: Theory and practice since 1949. Boston, MA: Kluwer Academic.

FURTHER READING

Dunbar, R. & Skutnabb-Kangas, T. (2008). Forms of education of indigenous children as crimes against humanity? Expert paper written for the United Nations Permanent Forum on Indigenous Issues (PFII). New York: PFII. [In PFII' system: "Presented by Lars-Anders Baer, in collaboration with Robert Dunbar, Tove Skutnabb-Kangas and Ole Henrik Magga".] www.un.org/esa/socdev/unpfii/documents/E_C19_2008_7.pdf Garcia, O., Skutnabb-Kangas, T., & Torres Guzman, M. (eds.) (2006). Imagining multilingual schools: Languages in education and glocalization, Linguistic Diversity and Language Rights series. Clevedon, UK: Multilingual Matters. Phillipson, R. (2009). Linguistic imperialism continued. New Delhi: Orient BlackSwan. Tomasevski, K. (2001). Human rights obligations: Making education available, accessible, acceptable and adaptable. Right to Education Primers 3. Lund/Stockholm: Raoul Wallenberg Institute of Human Rights and Humanitarian Law/Sida (Swedish International Development Cooperation Agency).

Many useful articles can be found in the following journals: Journal of Language, Identity and Education; Language Policy; Language Problems and Language Planning; Critical Inquiry in Language Studies.

4 History of Language Teaching

DIANE MUSUMECI

What is the relevance of the history of language teaching for a volume that contains state-of-the-art perspectives on issues facing the profession? Cutting edge research does not require it. Modern theories need not consider it. Yet none of the topics addressed in this volume is novel. Each has been considered at an earlier - sometimes much earlier - point in the history of language teaching. Most have been part of the disciplinary discourse for centuries. Notwithstanding, the field of applied linguistics devotes scant attention to its history. In his 1983 volume entitled Fundamental Concepts of Language Teaching, H. H. Stern lamented the fact that "language teaching theory has a short memory" and lacks "historical depth" (pp. 76-7). Brumfit and Mitchell (1990), Musumeci (1997), as well as Thomas (2004), also argue for an historical perspective. Mitchell and Myles (2004) address the problem of ahistoricity by grounding their excellent introduction to current second-language learning theories within the "recent history" of the post-WWII period. Aside from the intrinsic merit of historical research, even a passing acquaintance with the people, philosophies, and events that have shaped the history of second-language teaching provides the possibility of contextualizing current trends, practices, and debates.

Given the long and varied history of second-language teaching, a strictly chronological account of that tradition would be necessarily superficial at best in the space of a single essay. Instead, this chapter will outline the teaching of one particular language over the course of several centuries. In doing so, the topics presented in their contemporary context elsewhere in this volume will be treated here within a broader historical perspective. It is hoped that in this manner the reader may be exposed to the complexity that has characterized the history of language teaching and to the recurring issues that form the core of that tradition, as well as encountering relevant, at times perhaps surprising, insights that such a perspective offers.

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 Bkckwell Publishing Ltd. ISBN: 978-1-405-15489-5

The Context of Second-Language Teaching

Whereas systematic research on second-language acquisition is a relatively recent phenomenon, the practice of language teaching enjoys a long tradition, one that is linked to theories of mind and thought, philosophy, culture, economics, and education. One can easily imagine that second-language learning has been going on since peoples with different language systems first encountered one another on the savannahs of Africa, in the Mesopotamian valley, and on the plains of northern Europe. However, despite a history of second-language learning that certainly predates writing, formal accounts of second-language teaching typically begin in the Western tradition with the teaching of Latin, and to a lesser extent, Greek, in the centuries following the decline of the Roman Empire in the fifth century of the Common Era. The reason for this starting point is simple: Latin was the language of wider communication, the language of education, scholarship, and commerce, the lingua franca for almost two millennia. It was very much the equivalent of English in the globalized world today even though, for the vast majority of that time, it was not the first language of anyone who used it.

For centuries, the acquisition of second-language literacy skills; i.e., learning to read and write in Latin, constituted the foundation of all formal education. Erudite people also spoke it. That the historical tradition is based almost exclusively on the model of teaching Latin (and its subsequent influence on the teaching of other Western European languages) reflects a distinct disciplinary and cultural bias: First, because it investigates primarily formal instruction, despite the fact that until late in that history such instruction rarely included anyone but the most elite males of society; and second because the Western written tradition is certainly not the only one, and there is much that one could undoubtedly learn from other, oral, and more ancient traditions (Reagan, 1996).

In addition to a disciplinary bias, concise histories of second-language teaching necessarily have a narrow focus. Although the variety of approaches to language teaching that date from the mid-twentieth century (audiolingualism, the Direct Method, the Silent Way, Suggestopedia, the Natural Approach, communicative language teaching, content-based instruction, and task-based instruction to name the most salient) are well documented, histories typically cite the grammar-translation method used to teach Latin as the predominant instructional approach to the teaching of all languages until that point (Mackey, 1965; Richards & Rogers, 2001; Titone, 1968). It is generally understood that this method consisted of the study of grammatical rules, followed by translation from the second language into the first and back again. In an attempt to provide evidence for a much less hegemonic history of pedagogical practice, Kelly (1969) instead provides a compendium of materials and techniques that have formed a legacy of 25 centuries of language teaching, his eponymous work. More recently, Howatt and Widdowson (2004) present a richly detailed and fascinating account of the history of English language teaching over several centuries in an updated version of Howatt's

classic 1984 text. Other historical accounts that examine the highly contextualized teaching of one language include Kibbee's (1991) examination of the early practice of teaching of French in England.

Taking a decidedly different approach to the historical record, Thomas (2004) offers an impressive scholarly volume that traces over two millennia key philosophical concepts underlying the notion of Universal Grammar, that is, the principles deemed to be part of the innate human capacity to learn language. In a much less exhaustive, but similarly thematic, approach based primarily upon evidence from the fifteenth through the seventeenth centuries - a span which marked the apex of the teaching of Latin as a universal language and the inception of its subsequent demise - Musumeci (1997) explores the disparity between theory and practice in language teaching and argues that language teaching as rule-governed practice may not have been nearly as widely prescribed as the historical summaries suggest. Instead, the rote teaching of forms and rules has consistently coexisted with a concomitant insistence on language as communication and the privileged status of attention to meaning in the acquisition process. The tension between treating the second language as the object of instruction versus a system of communication has persisted throughout the history of language teaching, with the former often a by-product of external, pragmatic forces rather than an instantiation of theoretical stance.

Within either historical approach, broadly diachronic or thematic, two problems immediately arise in any attempt to characterize second-language teaching. First, one cannot presume that actual teaching practice at any given time was monolithic. Even today with the academy's emphasis on experimentation and data-driven research, the contemporary literature contains few detailed and systematic accounts of actual classroom behavior, let alone the numerous samples across a broad range of contexts that would be necessary to provide an informed characterization of current practice. The body of research on instructed second-language acquisition, as Doughty (2003) points out, has been complicated by the fact that "any particular implementation [of a method] by an individual teacher is subject to variation" (p. 263). How much more difficult it is, then, to accurately reconstruct that practice in past centuries when the norms governing scientific discovery, let alone educational practice, were quite different. To further complicate matters, the philosophical treatises, letters to princes and nobles, and sermons which constitute much of the historical record focus largely on reform, with the consequence that, where observations and accounts do exist, they are typically presented for the purpose of demonstrating ineffective practice leading to unsatisfactory outcomes. As an example, the following quote from the seventeenth-century educational reformer and renowned author of language textbooks Johannes Comenius foreshadows the rationale of those who have recently proposed the elimination of foreign language programs from universities in favor of outsourcing to study-abroad providers:

Camp followers and military attendants, engaged in the kitchen and in other menial occupations, learn a tongue that differs from their own, sometimes two or three,

quicker than the children in schools learn Latin only, though children have an abundance of time, and devote all their energies to it. And with what unequal progress! The former gabble their languages after a few months, while the latter, after fifteen or twenty years, can only put a few sentences into Latin with the aid of grammars and of dictionaries, and cannot do even this without mistakes and hesitation. (p. 69)

Not surprisingly, proponents of educational reform select worst-case scenarios to provide evidence for their argument, that is, what needs to change, rather than what is effective or beneficial. In other words, a change of practice is advocated because the current system is reported not to produce the desired results. It is noteworthy that the rhetoric of reform does not require proof that this is indeed the case. Instead it relies on the readers' conviction that the author speaks from authority and experience.

Just as the reformer need only convince the reader of the inadequacy of the current state of affairs, to be effective his or her argument need not necessarily provide evidence that the new system will guarantee success. To that end, a different strategy may be employed: namely, that the advocated change is actually not a new practice at all, rather a "tried and true," albeit little-known, component of the long-standing tradition. As a case in point, the US in the 1960s experienced a sudden interest in the history of language teaching precisely for this reason: the introduction of what was purported to be a "scientific" approach to language learning, the "New Key," as audiolingualism was called. Based on principles of behaviorism in psychology, it was a language teaching method that advocated carefully constructed pattern practice that would result in the formation of good linguistic habits deemed necessary for successful language learning. By arguing that the "new" method, while supported by the latest thinking in psychology and offering a radical departure from grammar-translation, was in fact seeded throughout centuries of second-language teaching, proponents of the approach provided empirical evidence to justify its adoption. Indeed, Kelly's tome, which remains the gold standard as an archive of language teaching techniques, is rife with references to historical exemplars of audiolingual tenets. Mackey and Titone reflect similar interpretive biases.

Spikes in historical interest appear to coincide with the advent of approaches to second-language learning that challenge the explicit teaching of grammar rules as the primary, or even sole, method of instruction. In this way, proponents of the innovation situate the new approach within a broader historical framework and seek validation in past experience. Thomas and Musumeci may well provide further testament to this phenomenon, with regard to Universal Grammar and meaning-based instruction, respectively.

On the other hand, advocates for language teaching methodologies that expound an explicit grammatical focus from the very beginning may not feel compelled to look for historical evidence to support their position. Such lack may reflect an underlying assumption that theirs is the historically supported stance, the norm. Certainly, a prevailing belief that all language teaching prior to the

mid-twentieth century entailed a grammar-translation approach would support such a notion. Blanket references to "traditional" methods of language teaching favor such interpretation. One would be surprised to read of a second-language curriculum characterized as "traditional" that advocated an experiential, multi-sensory, or immersion-type approach. Upon examination, however, language teaching history demonstrates a broad range of approaches, methods, and techniques dependent upon prevailing notions about how language is acquired, the context in which it occurred - which, in turn, affected the perceived purpose in learning the second language - the resources that were available, and the role of the participants. In light of the historical record, the definition of what is "traditional" language teaching becomes impossibly difficult to establish. The next section provides an account of how the first three factors - beliefs about language learning, methodologies, and the historical context - intertwined in the history of second-language teaching. A brief historical perspective on the perceived roles of the teacher and learner concludes this chapter.

Beliefs about Language Learning, Methodologies and Historical Context

The questions that present-day research in second-language acquisition addresses derive from theoretical frameworks that posit a set of underlying beliefs, assumptions, or principles. Doughty and Long (2003) provide an authoritative and comprehensive account of the key issues surrounding current research. Is language an innate ability unique to human beings? Is it a skill that is acquired through exposure and interaction? To what extent does learning a second language resemble first-language acquisition? To what extent is explicit versus implicit attention to linguistic rules a condition for successful acquisition? What is the nature of errors: Are they part of a developmental process that will eventually dissipate with increasing exposure and experience or are they something pernicious to be avoided or immediately corrected before they become a permanent part of the underlying grammar? What is the role of the learners' first language in the learning of the second? These are only a few of the questions surrounding second-language acquisition that have intrigued educators for centuries. The most fundamental question, the extent to which language is the result of an innate ability or learned behavior, forms the basis of early philosophical thought. In turn, the answer to that question intimately informs the construction of language teaching methodology.

At its most extreme, the innatist position holds that language acquisition results from an essential human ability specific to language that resides in the mind. A pedagogical corollary of this position is that language acquisition manifests itself according to a predestined route that instructional intervention serves at best to accelerate. Although the research evidence that substantiates this position dates to the second half of the twentieth century, one can trace the earliest roots for an innatist stance to an ancient philosophical position originating with

the Greek philosopher, Plato (427-347 BCE). Stated in the simplest of terms, Plato proposed that human beings possess knowledge intrinsically. Such knowledge need simply be activated and drawn out. As a result, the teacher's role is to educate, from the Latin educere, which means literally 'to lead forth'. It is from Plato that we learn of the Socratic method, an instructional technique in which the teacher asks a series of carefully constructed questions each based on the student's previous response, leading students to arrive at the answer from what they already know. The basic premise of another ancient Greek philosopher, Aristotle (384-322 BCE), is radically different. Aristotle proposed that a human being comes into the world as a blank slate, tabula rasa, upon which knowledge is inscribed either by experience or by rule. The teacher's role in the Aristotelian tradition was to provide that knowledge, to instruct; from the Latin instruere, meaning 'to build'. Those who believe that language is largely a learned behavior may be more concerned with the environmental features surrounding language acquisition and with explicit instruction in rules. For example, they may fear the consequences of exposing learners to poor examples and leaving errors unchecked, and thus the possibility of acquiring inexpungeable mistakes or bad habits. Quintilian, a Roman educator in the first century CE whose twelve-volume text Institutio oratoria [Education of an orator] is a foundational contribution to Western educational theory, warns that children must be presented only with the best models, such that "Care must be taken that tender minds, which will imbibe deeply whatever has entered them while rude and ignorant of everything, may learn not only what is eloquent, but, still more, what is morally good" (Institutio oratoria I.8.4, in Murphy, 1987, p. 64).

From the very beginning, the Western tradition has struggled with the question of nature versus nurture, the innatist versus environmentalist position, a biological as opposed to a behavioral explanation for language acquisition. Mitchell and Myles (2004) provide a highly readable account of the continuing nature versus nurture debate and its influence on the most recent theories of second-language acquisition. Their discussion reiterates the diversity of frameworks that guide contemporary research on second-language learning.

Clearly, it would be absurd to reduce our current understanding of second-language acquisition to a simple dichotomy. In dealing with a complex knowledge system like language, one would expect that certain theories better explain some aspects of the phenomenon than others. On the one hand, proponents of connectionism address several issues - the contribution of repetition, frequency of stimuli, and automaticity - that find their genesis in an Aristotelean model of learned behaviors. On the other hand, proponents of Universal Grammar have established a rigorous research paradigm in which to investigate the extent to which innate abilities can be reactivated and with what success. What is particularly interesting from an historical perspective is the vantage point that it affords one to examine these basic philosophical questions as they permeate the history of second-language teaching from the earliest records to the present day, receding and reemerging in response to intellectual shifts, as well as external pressures, as that tradition unfolds.

Among the external factors that contribute to underlying beliefs and consequently prescriptions for practice, it is nearly impossible to overestimate the enormous influence of organized religion on education, including the teaching of language. The Aristotelian philosophy as interpreted by the early Church Fathers was the one adopted by the Christian Church, itself the major source and foundation of institutionalized education from the early Middle Ages onward. As a social institution, the Church possessed the infrastructure and the authority necessary for the provision of such education. One must keep in mind, however, that during this time mass public education did not exist. Although purportedly obligatory in the Aztec empire, in Europe the notion of compulsory education of children was unheard of from the fall of the Roman empire until the Protestant Reformation, when the theologian and church reformer Martin Luther (14831547) proposed in 1524 that all children acquire basic literacy in order to read the Bible. Strasbourg passed an edict to that effect in 1598, but it wasn't until the late eighteenth century that similar legislation began to spread slowly across the continent, first arriving in the United States in Massachusetts in 1852 and in England still later in 1870.

Until almost the twelfth century, formal education in Western Europe took place in monasteries that offered instruction in liturgy and prayer. Cathedral schools trained clergy in Canon law, preaching, and disputation, that is, the ability to defend a position through logical argumentation orally. Until the seventeenth century, other than religious schools that trained boys for the clergy, monastery, or choir - in descending order of required proficiency in Latin -education of children was limited to instruction of the elite by private tutors. The foundation of the earliest universities in the late eleventh and twelfth centuries (Bologna in 1088, Paris in 1150, and Oxford in 1167) allowed students the opportunity to build upon that early education. In this way, a student at age 14 or so, having mastered basic literacy in Latin under a tutor's guidance, could study the seven liberal arts - the trivium of grammar, logic, and rhetoric, followed by the quadrivium of arithmetic, geometry, music theory, and astronomy - before proceeding to advanced study in law, medicine, or theology. At the university, Latin was both the initial core subject as well as the medium of instruction. All lectures, readings, and discussion took place in Latin.

From their letters, we learn that university students in the twelfth century shared many of the same concerns of students today: they ask their parents for money and clothes; they complain about the cost of books and supplies; and they request more time to finish their degrees (Haskins, 1958). The Manuale Scholarium, a student guide to the university, provides a lively account of student life in the medieval period. In it we learn of the explicit requirement that Latin be used at all times as the ordinary means of communication in class and outside it, on penalty of a fine should a student be caught using the vernacular. The "wolves" were students appointed to surreptitiously record and report infractions of the rule. That students found the rule difficult to keep is recorded in one of the manual's dialogues:

Cam. May the minions of hell destroy him. If I ever find out his name, he won't get away.

Bar. What ails you?

Cam. Listen to me; I've been up against the wolf twelve times . . . I'll see to it, I'll find him. Later, I'll avenge this injustice.

Bar. It isn't an injustice, but rather the rule. Don't be surprised that you've been reported so often; he could have reported you a hundred times. To tell the truth, I haven't heard a single word from you in Latin for a whole week.

(The Manuale Scholarium, ch. 11, in Seybolt, p. 73)

Clearly, although Latin was the fulcrum upon which the entire curriculum balanced, already in the twelfth century it was neither the students' first nor preferred language for everyday communication even at the university. The need to establish strict rules enforcing its use among the students provides compelling evidence of its contrived status as their lingua franca. However, such rules also underscore a prevailing conviction that oral proficiency was best acquired by using the language for meaningful communication.

Given the dependence on Latin texts to provide the bulk of one's formal education, the rediscovery of original classical texts during the Renaissance fostered an explosion of academic interest. This interest inflamed the desire among certain scholars not only to correct the errors of spelling, grammar, and syntax that had multiplied in medieval manuscripts, but also to recreate, to the greatest extent possible, the best of the classical world, including the widespread diffusion of the Latin language as a means to imitate that experience. To that purpose, they proposed a revision of the medieval curriculum: namely, the studia humanitatis (literally, the study of humanity) designed to make students wise and moral, ideal citizens of the world, and universally proficient in Latin.

The new program of learning is described in great detail in a treatise entitled De ingenuis moribus et liberalibus studiis [On the Conduct and Education of Young People], written around 1400 by the Latin scholar, Vergerius. Taught entirely in Latin, the curriculum was based on early exposure to classical texts and included the study of history, eloquence (persuasive speaking), grammar (writing), rhetoric (public speaking, including oration), poetry, music, arithmetic, geometry, and science. One of the most celebrated humanists, as proponents of the new curriculum were called, was Guarino Guarini (1374-1460), also known as Guarino da Verona, a fifteenth-century scholar of Latin and Greek and tutor to Lionello d'Este, prince of Ferrara, where Guarino eventually opened his own school. Guarino was convinced that the new program of learning provided the essential formative education for all citizens, especially those in positions of power and civil authority, not just the future teacher-scholar. Moreover, he insisted that one of the strengths of the studia humanitatis lay in the fact that it sought to develop high levels of speaking ability in Latin:

To be admired for fluency, to be appreciated for speaking good Latin, is beginning to be true of the period, not just of individuals, so that today one is less likely to be

praised for speaking well than to be criticized for speaking badly; today we are expected to speak good Latin, more so than in the past when we were criticized for speaking a barbarous language. (Guarino Guarini, Letter to Cristoforo Sabbion, Chancellor of Verona, in Garin, 1958, p. 420, translation mine)

The humanists contended that second-language literacy skills alone were necessary but insufficient training for a citizen of the Renaissance world. Instead, the educated person would also be able to speak Latin fluently, confidently, and persuasively. The means to attain such fluency was being immersed in the study of humanities in the second language, Latin.

Like Guarino, other Latin (and Greek) tutors became famous during the Renaissance for the small schools that they operated, often limited to a select group of children who attended lessons along with the patron's children. Vittorino da Feltre (1378-1446) was one of these. His Casa Giocosa (Joyful House) was where he taught the Duke of Gonzaga's children along with others. (Guarino insisted, as a condition of employment, that he be allowed to enroll additional pupils whom he selected based on their intellectual ability rather than their families' social position.) The school is included in histories of education as one of the first boarding schools. It boasted a total immersion experience, with instruction geared to the ability and needs of each child. His curriculum included games and recreation, physical education, and music, in addition to the humanistic subjects of Latin and Greek. Associates of Vittorino attested to his extraordinary success.

Unfortunately, it is impossible to determine from the historical record to what extent the success of Vittorino's school was a result of the teacher's expertise, the methodology (early full immersion to the extreme), specific learner characteristics (aptitude, age, motivation), and the second language itself (highly prestigious, structurally similar to the students' first language). One might suspect that the last factor - shared structural features between Latin and the Italian vernaculars - must certainly have influenced at least the rate of acquisition among Vittorino's pupils. Would the same hold true for structurally more disparate languages? Given students' success in acquiring Greek, as well as Latin, apparently so.

One might argue that the linguistic proficiency expected of a five-year-old is quite different from what one might demand from an educated adult. Nevertheless, it appears that a talented and expert teacher along with a motivated group of pupils given unlimited time and resources to enjoy total immersion in the second language offered the ideal conditions for rapid and successful acquisition. Many of the most learned men of the day were trained in Guarino's and Vittorino Da Feltre's schools.

But despite success stories like those of Guarino and Vittorino Da Feltre, the humanists' campaign to resurrect Latin as the universal language ultimately failed. While their introduction of the studia humanitatis would have lasting effects on the curriculum, the case for spoken Latin could not turn the rising tide of the vernacular languages, which constituted the common system of communication

among the burgeoning and increasingly influential middle class. Even parents of university students began to complain about the waste of time and money spent on learning a language that served no practical purpose. In his Letters to the Majors and Aldermen, in which he makes the case for basic literacy for everyone, Luther warns against an overzealously pragmatic approach that would lead to the abandonment of the content of the humanities along with Latin, a tossing of the baby with the bathwater. He specifies that even in the vernacular, basic literacy is not a sufficient outcome of instruction:

And pay no attention to the contempt which the ordinary devotee of Mammon manifests for culture, so that he says: "Well, if my son can read, write, and cipher, that is enough; for I am going to make a merchant out of him." Without scholars it would not be long till business men in their perplexity would be ready to dig a learned man out of the ground ten yards deep with their fingers; for the merchant will not long remain a merchant, if preaching and the administration of justice cease. (Luther, Letters to the Mayors and Aldermen, in Boyd, 1966, p. 247)

Pragmatic concerns are not so easily set aside, and tensions between Mammon and the Academy reverberate still. The fading utility of Latin for anything but the most scholarly pursuits would eventually be extinguished with the availability of vernacular translations of the Bible made possible by the invention of the printing press in 1440. For the first time, literacy in one's first language could both ensure success in the affairs of this world and guarantee one's salvation in the next. Before that would happen on a wide scale, however, another century would pass. In the meantime, a sophisticated and prestigious program of education that would set the academic standard for foreign language instruction emerged.

The influence of the Catholic Church on education reached its pinnacle with the system of education developed by the Jesuits, a religious order founded in the 1500s by Ignatius of Loyola, and still influential to this day. Known for their intellectual rigor and strict discipline, the Jesuits accepted only the brightest students into their schools, suggesting a suspected link between intelligence and foreign language aptitude, at least in so far as the method by which Latin appears to have been taught. The Jesuits were also instrumental in solidifying the tradition of presenting the whole of the grammar in the first year or even in only half a year. Although a few students, who must have possessed an extraordinary aptitude for language learning, appeared to benefit from this practice and were able to proceed immediately to the study of content in the second year, the majority advanced at a much slower rate. Consequently, the second year consisted of a review of the grammar, with continued review each year until the material had been mastered.

The Jesuits' Ratio studiorum [Plan of Study] offers one of the most detailed extant manuals for school administration and teaching practice, surpassing even that of Quintilian. In it one can find rules covering everything from the responsibilities of the rector to those of the janitor and everyone in between. It also

provides an exquisitely clear description of a language teaching method that later came to be characterized as "grammar-translation."

The general format of the language classes consisted of recitation of memorized passages (from Cicero and a grammar book), a review of the previous lesson, a lecture, and a dictation, followed by the presentation of a new grammar point. Directives for how the lectures in the first-year grammar class were to be conducted are provided in almost excruciating detail:

The prelection [lecture] in Cicero which shall not exceed four lines will be in this form: First, let him [the teacher] read the entire passage continuously, and state its topic very briefly in the vernacular. Second, let him express the sentence in words of the vernacular. Third, starting from the beginning, let him indicate the structure and explain the sentence, telling which words govern which cases; let him go over many things pertaining to the laws of the grammar already explained; let him offer some observations or other of the Latin language, but as simple as possible; let him explain the metaphors by well-known examples, but let him not dictate anything, except perhaps the topic. Fourth, let him again go through the words of the author in the vernacular. (Ratio, Rules to the Professor of Lower Grammar Classes, in Fitzpatrick, 1933, p. 233)

Lectures in the second- and third-year classes followed the same format, changing only the number of lines to be read (from four to seven), with the inclusion of more dictation in the second year and emphasis on word derivations in the third. Outside of lectures, students prepared written translations of sentences (later, passages), to and from the vernacular. Teachers were exhorted to devote scrupulous attention to error correction. Either the teacher himself would correct the translations or a particular type of activity, the concertatio, would serve the same purpose. This activity consisted of students identifying errors in each other's written work and demanding a repetition of the rule that had been broken. For example, if a student had written an incorrect ending on a word, the one who found it would ask for a recitation of "the whole declension or conjugation in order or in broken order, alone or with an adjective or noun or pronoun" (Ratio, in Fitzpatrick, 1933, p. 234). Interestingly, this was one of the rare activities that the Ratio recommended be conducted as an oral exercise entirely in Latin without recourse to the vernacular, limited as it was to the use of fixed expressions and the repetition of grammatical rules.

A comparison with the teaching methods advocated by the humanists of the previous century immediately reveals a glaring difference: namely, the enormous reliance in the Ratio on the use of the students' first language in order to teach the second. The directives of the Ratio indicate clearly that Latin was no longer the medium of instruction, or at least not the sole medium of instruction, in the Jesuit system. Instead, the teacher was explicitly advised to provide learners with a first-language translation of the Latin words and sentences, not just once, but several times over the course of a single lecture, and not only in the first-year course, but in the second and third years as well. The use of the vernacular undoubtedly ensured comprehension on the part of the students. However, it

also obviated the function of Latin as the primary route to access meaning. Whereas the humanists had recommended that, in order for students to achieve both literacy and functional oral proficiency, Latin be the medium of instruction in interesting subject matter, in the grammar-translation method advocated by the Ratio studiorum, it became, instead, the object of study, in what today we might recognize as a "focus on forms" approach (Doughty & Williams, 1998; Long, 1991). Moreover, by separating the study of language and the study of content for a period that could last as long as three years, a divide was created within the curriculum that exists to the present day.

The use of the vernacular languages for instruction of the elite students in the Jesuit schools reflected their rising cachet in society at large. What precipitated their transformation from "barbarous" tongues to such elevated status? The seventeenth century witnessed a combination of social and religious forces that solidified the status and power of the vernacular languages of Europe. Intense religious antagonism, not only between Roman Catholics and Protestants but within the sects themselves, could not allow an international language associated with the Roman Church on the one hand, and classical (pagan) culture on the other. Economic factors, too, favored the vernacular languages of France, Holland, and England as they displaced Catholic Spain as the dominant commercial power.

The nationalistic objectives of seventeenth-century Europe were decidedly at odds with the notion of global communication. The formation of nation states was aided by linguistic, in addition to geographical, boundaries. National languages served to strengthen national identities. During this period, academies were formed to promote, regulate, and preserve the linguistic integrity of the vernacular languages. The Accademia della Crusca in Florence was established in 1582, the Académie Française in 1635, and the Real Academia Española in Madrid in 1713. The Accademia della Crusca published the first dictionary of Italian in 1612, which served in turn as the model for the first dictionaries published in French (1694) and in Spanish (1726-39).

The elevated prestige of the vernacular languages affected more than the status of spoken Latin, already in jeopardy in the Middle Ages; it compromised even the utility of developing reading proficiency in Latin. No longer dependent on ecclesiastic authority and Latin, individuals could interpret Scripture for themselves in their first languages, thanks to the proliferation of translations of the Bible (Luther's translation of the New Testament into German in 1522 and of the Old Testament in 1534; Tyndale's translation into English in 1526, with the authorized King James version in 1611; Czech in 1568; Welsh in 1588; and the London Polyglot Bible in ten languages in 1653). By this time, the printing press had advanced to the point that, not only did students likely have individual copies of texts, even of the same edition, but new versions of books appeared on the market that offered interlinear translations. In this way, students could view synoptically on the same page a line of text in Latin and its vernacular translation. One can imagine how such texts must have facilitated the conduct of the lessons, ensuring comprehension and easing the lives of teachers and students

alike! "Reading between the lines" literally gave instant access to meaning. Such a welcome innovation did not come without cost, however, as it also allowed students to bypass the second language entirely, much like watching a film with subtitles or listening to a simultaneous translation. No longer the necessary vehicle by which to access interesting subject matter, Latin was reduced at best to a linguistic puzzle or a challenging mental exercise, and at worst to a dull catalogue of abstract, nonsensical rules.

Interestingly, the most celebrated educational reformer of the seventeenth century Johannes Amos Comenius was known primarily not for his curricula (which he structured from infancy through university), nor for his interest in the invention of a new universal language to replace Latin, nor for his many philosophical and theological treatises, nor even for his Didactica Magna [The Great Didactic] of 1657, in which he outlined his plan for a universal system of education. Comenius is famous above all for his textbooks: the Janua Linguarum Reserata [The Gate of Tongues, Unlocked] of 1631, the Vestibulum [The Vestibule] of 1633, and perhaps most famous of all, the Orbis Sensualium Pictus [The World of Things in Pictures] of 1658. The acclaim for Comenius' textbooks cannot be overstated. Their popularity was such that they were translated into twelve European languages and several Asian ones, as well; they continued to be published for almost two centuries after the author's death.

Comenius' passion for educational reform stemmed from his frustration with what he perceived to be the appalling inefficiency of the schools, those "slaughterhouses of the mind," a problem that he blames on the instructional methods employed: "Latin grammar was taught us with all the exceptions and irregularities; Greek grammar with all its dialects, and we, poor wretches, were so confused that we scarcely understood what it was all about" (Didactica Magna, ch. 16, p. 122). Schools, he argues, fail for two reasons: First and foremost because they lack appropriate materials and, second, because they do not follow the "natural order," that is, "that the matter come first and the form follow" (Didactica Magna, ch. 16, p. 115). With regard to language teaching, this meant that instruction erroneously begins with the grammar (the form) rather than authors or examples (the matter). Instead, three principles guided Comenius' proposed reform: "that no language be learned from a grammar, but from suitable authors," "that the understanding be first instructed in things, and then taught to express them in language," and finally, "that examples come before rules" (p. 116). In some sense, Comenius' insistence on meaning ("things") before form and a reliance on "authors" to supply that content of that meaning echoes that of the early humanists. However, the intellectual milieu of the seventeenth century - a period of dramatic scientific discovery that challenged long-standing beliefs and established a new science of observation and experimentation - along with the consequences of the Reformation and Comenius' own religious convictions, had undermined the authority of the ancient authors. The suspicion that such authors might not constitute either trustworthy or appropriate material served as an impetus for the design of textbooks whose content and structure were more suitable for children in school.

The first of Comenius' textbooks, the Janua Linguarum Reserata, presents 8,000 of the most common Latin words arranged to form 1,000 sentences in a progression from the shortest, most simple constructions in the early chapters to increasingly more complex constructions in the latter. Congruent with his notion that instruction in content and language proceed simultaneously, each of the 100 chapters deals with one class of phenomena; for example, fire, diseases, trade, arithmetic, angels. Despite the inclusion of a vernacular translation for each Latin sentence, teachers complained that the book was too difficult. In response, Comenius created a preparatory text, the Vestibulum, in which 1,000 of the most common words are arranged into 427 simple sentences. However, subsequent editions of the Vestibulum became simpler still, until they were finally reduced to Latin-vernacular word lists.

Despite an apparent shift in his methodological stance, as evidenced in the later versions of the Vestibulum, as well as related incongruencies in his much less widely known methods book, Didactica Magna, Comenius' acclaim as a textbook author increased. Around 1650 he began to sketch out his third and most significant historical contribution, the Orbis Sensualium Pictus, a true breakthrough in educational practice. Designed as a preschool text, the Orbis was meant to teach the vocabulary of real-world objects and events. Its innovation lies in the use of illustrations, not as decorative elements or as supplements, but as an integral part of the text itself. Typeset as columns of Latin sentences with side-by-side vernacular translations, words in both languages are numerically keyed to the corresponding elements in the accompanying illustration. According to a biographer, the success of the Orbis was "even more extraordinary" than that of the Janua, such that two books, the Orbis and the Bible, formed the essential home library for generations of children (Keatinge, 1910, p. 78).

The creation of the first illustrated "picture book" for children is accomplishment enough to rate permanent inclusion in every history of education. However, despite their popularity, Comenius' language textbooks did not ensure the easy and successful acquisition of Latin any more than the "grammar-translation" approach of the Jesuits. Moreover, another of Comenius' recognized contributions to the history of Western education is the legitimacy that his textbooks gave to the use of the vernacular as the language of instruction (Cole, 1950).

Although his Janua was a bilingual textbook, Comenius championed the notion of primary instruction in the students' first language. His school plan began with the Vernacular School. Only upon its successful completion would those boys whose professional aspirations required a more sophisticated program of study advance to the Latin School. With regard to modern foreign languages, he advised that, once having attained first-language literacy, boys between the ages of 10 and 12 could profitably study foreign languages in the period between the Vernacular and the Latin Schools. In Comenius' opinion, the best means for that study "is to send them to the place where the language that they wish to learn is spoken, and in the new language to make them read, write, and learn the class books of the Vernacular School" (Didactica Magna, p. 273). In other words, he advocated study abroad within an immersion or content-based model. In this

way, Comenius did not disagree with the humanists: he agreed that 10-year-olds could learn second languages without difficulty and that an effective way to ensure successful acquisition was to immerse students in the second-language environment, where the language was also the medium of instruction. The reason that Latin could not be taught like any modern language was that the goal of that instruction had changed. The need for functional proficiency in Latin no longer existed.

No longer the language of wider communication, having been displaced by the vernaculars as the language of commerce, diplomacy, science, and scholarship, Latin's only remaining communicative function had existed solely within the instructional setting, and shared even that with a predomination of the learners' first language, a context so circumscribed and artificial that it probably could not have been sustained by any means on the wide scale demanded by the impending imperative of mass public education. Although it continued to enjoy esteem in the academy, until the early twentieth century the place of Latin in the general curriculum was settled: for the vast majority of students it was a prestigious but arcane academic exercise that served to enrich the learner's lexicon through derivation and word-formation skills, to teach abstract grammatical categories, and perhaps to allow for rudimentary translation of a few lines of Vergil's Aeneid or Julius Caesar's De Bello Gallico. It is difficult to imagine that Latin was once the historical equivalent of World English today, just as it is equally unfathomable that English could be replaced by yet another language, perhaps one that is currently spoken by a relatively small group of people in some remote part of the world. And, yet, this is precisely what happened to a prestigious world language, despite an educational system that was predicated entirely on its preservation and transmission.

The Role and Status of the Language Teacher

Histories of education are commonly organized around the identification of the great philosophers and famous educators (Cole, 1950; Boyd, 1966; Smith & Smith, 1994). Yet another benefit that the student of applied linguistics might derive from the study of the history of second-language teaching is an appreciation of its intellectual legacy. Such illustrious figures to the contrary, however, the historical record suggests that the average language teacher was seldom held in such high esteem. Grammar may have been the cornerstone of the entire curriculum, but the grammar teacher himself was often disparaged. By definition, a grammarian is one versed in the knowledge of grammar, a philologist, and/or a teacher of grammar. It has also come to be used as a term of reproach: a "mere grammarian" or a "dry, plodding grammarian." The pedagogue (from the Greek pedagogos) was originally the slave who accompanied boys to and from school and who supervised their behavior; later the Romans extended the meaning of the term to signify the teacher. It, too, has acquired a decidedly negative connotation. Like the grammarian and the pedagogue, the pedant also refers to a

teacher, schoolmaster or tutor, and consequently to someone who is "excessively concerned with accuracy over trifling details of knowledge or who insists on adherence to formal rules." The Commedia dell'Arte, the Italian improvisational comic theater begun in the mid-1500s and popular through eighteenth century, included as one of the stock characters il Dottore, a pompous, ridiculous figure immediately recognizable by his pedantry. How did the profession arrive at this point?

From the earliest days of language teaching in ancient Rome, educated Greek slaves provided instruction in Greek language and the arts to interested members of the household. Itinerant teachers offered instruction for a fee to whomever they could recruit as students. Some of the greatest educators were tutors, but few were teachers in classrooms with many learners, let alone students of diverse preparation, motivations, and abilities.

That language teaching in the schools was held in lowest regard is evident in the following excerpt from a Jesuit who proposed for himself the severest of penances:

Relieve me of the care of others, take away my preaching and my study, leaving me only my breviary, and bid me come to Rome, begging my way, and there put me to work in the kitchen, or serving table, or in the garden, or at anything else. And when I am no longer good for any of this, put me in the lowest class of grammar and that until death, without any more care for me ... than you have for an old broom. (Letter to Ignatius from Lainez, 1552, in Young, 1959, p. 273, emphasis mine)

As if to further underscore the marginal status of the language teacher, the Jesuits invented a two-tiered system of instruction that forms the basis of second-language programs at many universities to this day: the use of advanced students to instruct the language classes, freeing professors to teach the more prestigious "subject matter" courses. Musumeci (1997) argues that this distinct separation of subject matter (content) and language (skill) further fueled the demise of Latin as a meaningful subject in the curriculum. By reducing language to a mere skill, the esteem in which the teacher was held diminished proportionally.

The introduction of mass public education in northern Europe in the seventeenth century created a demand for teachers that must have outstripped the supply. Rather than raise the status of the teacher according to the usual rules of modern economics, however, it appeared to have had an opposite effect: increasing supply by lowering the acceptable standards. Comenius lamented the fact that good teachers were few and far between, the best having been snatched up by the wealthy to serve as tutors for their children, depriving the state of an important resource.

In Comenius' instructional framework, the teacher is the single source of knowledge that is poured into the students, like water from a fountain or the warmth of the sun (Didactica Magna, pp. 163, 165, 166, 250). Given the student's status as a blank tablet on which the teacher wrote or painted knowledge, Comenius held the teacher entirely responsible for a student's failure to learn: "If the result be not successful, it is more than certain that this is not the fault of the tablet (unless

it have some inherent defect), but arises from ignorance on the part of the writer or painter" (p. 44). Despite the enormous instructional burden that he placed on the teacher's shoulders, Comenius minimized the instructor's relative merits as follows:

An organist can read any piece of music from his notes, though he might not be able to compose it or to sing or play it from memory; and a schoolmaster, in the same way, should be able to teach anything, if he have before his eyes the subject matter and the method by which it should be taught. (Didactica Magna, p. 288)

Comenius' passionate belief in the foremost importance of sound materials and methodology reduces the teacher to a technician who simply puts the plan into action, but has no role in its design. He is a musician who cannot compose or sing or even play from memory. The adage "those who can, do; those who can't, teach" comes painfully to mind.

Finally, despite the ongoing admonition of the importance of supplying learners with good language models from the very beginning of the language learning process, for most of its history the profession had little to say about whether those models need be native speakers. This, too, is an artifact of Latin, a language that for centuries of its instruction was the first language of no one; all teachers were de facto non-native speakers. Howatt and Widdowson (2004) point out that the tradition of the non-native teacher persisted throughout the early days of teaching English as a second language. In fact, the highly proficient non-native speaker who shared the learners' first language and culture and who could predict the difficulties that students would encounter enjoyed a unique advantage over the native-speaker instructor.

Potential effects of the instructor's status as a native or non-native speaker of the second language continue to be investigated in present-day research. Nevertheless, it would be naive to believe that the factors that decide who should teach a language revolve solely around linguistic concerns. When language is understood to be a cultural commodity or a lucrative national product, the native speaker becomes part of the linguistic economy, and the economic reality can be significant. As an example, recent calculations report that English language teaching contributed over £10 billion to the United Kingdom's economy in 2001-2, with the prediction that the market was poised to undergo yet another major expansion (Johnes, 2004). Given such high stakes, one can well understand the motivation to portray native-speaker teachers as highly desirable and valuable resources.

The Role and Status of the Learner

The teacher is only part of the equation; what about the learner? Is the student an active participant in his or her learning or a blank tablet upon which to write? The underlying philosophy in conjunction with the current technology

supply the metaphors for learning. Not surprisingly, Quintilian likens the student to a container, "we are by nature most tenacious of childish impressions, just as the flavour first absorbed by vessels when new persists, and the colour imparted by dyes to the primitive whiteness of wool is indelible. Further it is the worst impressions that are most durable" (Institutio oratoria I, in Ulich, 1954, p. 104). Using almost identical words, one of the early Church doctors, Jerome (347-420), offers advice on the proper education of a daughter when he warns, "Early impressions are hard to eradicate from the mind. When once wool has been dyed purple who can restore it to its previous whiteness? An unused jar long retains the taste and smell of that with which it is first filled" (Letter to Laeta, in Ulich, 1954, pp. 165-6) Such expressions reflect a fear that exposure to poor models and allowing errors to go uncorrected will have lasting detrimental effects for the learner.

Although they reappear with some regularity throughout the history of second-language teaching, these concerns do not reflect the only response to learner error. For example, such metaphors are conspicuously absent in the language of some early humanists. Guarino insisted, as well, on the importance of excellent language models, warning that teachers should refrain from obscene language that students might later imitate. Unlike the tabula rasa models, his metaphors for students, instead, emphasize their active role in a discovery process of learning. He offers the following advice to students who struggle to interpret the meaning of a text:

If, instead, it [the meaning] escapes you and "remains hidden to you" go back, knock so to speak on the door, until even if it takes time, it opens just a crack to your understanding. Here you should imitate your hunting dogs that, if rummaging through bush and shrubs they don't find the bird on the first try, receive the order to repeat the procedure, because that which doesn't emerge at the first attempt might be flushed at the next. (Letter to Lionello d'Este, in Garin, 1958, p. 380, translation mine)

In the expressive skills, as well, Guarino concentrates on the creative rather than the mechanical aspects of the task, comparing the learner to a sculptor or artist who from a heap of raw materials (ideas) creates a polished piece of work.

To support his argument for the importance of early education, Comenius, instead, builds his case by enumerating the classical metaphors for care in the early stages: the initial pliancy of wax, the vessel that retains the essence of what it first held, the sapling that conforms to the shape of its early environment. However, in an interesting twist, he combines these with new metaphors, peculiar to his time; namely, the specialization of labor and mass production. He suggests that just as one goes to the cobbler for shoes and the locksmith for a key, children go to school for instruction. The implication is that the teacher supplies a commodity (knowledge), and the school is the place where scholars are produced, like shoes or keys. Mass education is likened to fish hatcheries or the cultivation of fruit orchards. Moreover, he argues that, given the proper tools and method, instruction can proceed almost effortlessly:

It will be no harder to teach schoolboys, in any number desired, than with the help of the printing press to cover a thousand sheets daily with the neatest writing . .. The whole process, too, will be as free from friction as is the movement of a clock whose motive power is supplied by weights. It will be as pleasant to see education carried out on my plan as to look at an automatic machine of this kind, and the process will be as free from failure as are these mechanical contrivances, when skilfully made. (Didactica Magna, pp. 96-7)

With the appropriate materials and method, instruction will proceed with the efficiency of the printing press and the precision of clockwork. In reaction against the elitist system of the Jesuits, Comenius insisted that all children benefit from education regardless of their natural ability, arguing that "a sieve, if you continually pour water through it, grows cleaner and cleaner, although it cannot retain liquid" (Didactica Magna, p. 67).

At first glance, an examination of the metaphors that were used in the past to describe language teachers and learners may appear antiquated or quaint. However, even a brief overview of the historical panorama reveals the extent to which the existent technology influences the ways in which second-language learning and teaching are construed. It is not surprising that the modern equivalent of the printing press, the computer, serves the same function today. Most of the literature in second-language acquisition research would be impossible to interpret without the metaphors from twentieth-century information technology that allow the profession to refer to language learners as bilingual processors and the pathways and products of that learning in terms of input, intake, output, working memory, and processing capacity.

Conclusion

In historical accounts of second-language teaching, clichés abound: there is nothing new under the sun and the pendulum swings widely from one approach to its apparent counterpart. Indeed, anyone who has been teaching language for more than 20 years or so has experienced such shifts first hand. A familiarity with the history of language teaching reveals a complex constellation of factors that affect those changes. It can also serve to identify some underlying trends with regard to language education reform. When the primary goal of second-language teaching is the development of functional oral skills, the prescribed method employs the second language as the medium of instruction; that is, as a vehicle to convey meaning. The goal of developing second-language literacy requires at least a reliance on the reading of authentic texts and extensive writing practice. On the one hand, it appears that the successful implementation of total immersion requires a combination of features that may be impossibly difficult to control in a large, institutionalized setting, where the expertise of the teacher, the motivations and abilities of the students, the prestige of the second language, and the economic resources are not guaranteed. On the other hand, language teaching

methods that are easier to implement in a large, institutionalized setting with a wide range of teachers and learners, in which the second language is the focus rather than the medium of instruction, and students rely heavily on their first language, result in limited functional skills for the majority of students. The historical record implies that, directives aside, attempts to force students to use a language that is not necessary for the acquisition of subject matter knowledge in the classroom or that serves no practical purpose outside it have small hope of success.

REFERENCES

Boyd, W. (1966). The history of western education. New York: Barnes and Noble.

Brumfit, C. & Mitchell, R. (1990). Research in the language classroom. London: Modern English Publications in Association with The British Council.

Cole, L. (1950). A history of education: Socrates to Montessori. New York: Holt, Rinehart, and Winston.

Comenius, J. A. (1657). Didactica Magna [The Great Didactic]. In M. W. Keatinge, (trans. and ed.), 1907: The Great Didactic of John Amos Comenius, vol. 2. London: Adam and Charles Black.

Doughty, C. J. (2003). Instructed SLA: Constraints, compensation, and enhancement. In C. J. Doughty & M. H. Long (eds.), The handbook of second language research (pp. 256310). Oxford: Blackwell Publishing.

Doughty, C. J. & Long, M. H. (eds.) (2003). The handbook of second language research. Oxford: Blackwell Publishing.

Doughty, C. J. & Williams, J. (eds.) (1998). Focus on form in classroom second language acquisition. New York: Cambridge University Press.

Fitzpatrick, E. A. (ed.) (1933). St Ignatius and the Ratio Studiorum. New York: McGraw-Hill.

Garin, E. (ed.) (1958). Il Pensiero Pedagogico dello Umanesimo. Florence: Giuntine & Sansoni.

Guarini, G. Le Epistole di Guarino da Verona. [The Letters of Guarino da Verona.] In Garin, E. (ed.) (1958), Il Pensiero Pedagogico dello Umanesimo. Florence: Giuntine and Sansoni.

Haskins, C. H. (1958). Studies in mediaeval culture. New York: Frederick Ungar Publishing Company.

Howatt, A. P. R. & Widdowson, H. G. (2004). A history of English language teaching. Oxford: Oxford University Press.

Johnes, G. (2004). The global value of education and training exports to the UK economy. UK Trade and Investment, April. London: The British Council.

Keatinge, M. W. (trans. and ed.) (1907). The Great Didactic of John Amos Comenius, vol. 2. London: Adam and Charles Black.

Keatinge, M. W. (trans. and ed.) (1910). The Great Didactic of John Amos Comenius, vol. 1. London: Adam and Charles Black.

Kelly, L. G. (1969). 25 centuries of language teaching. Rowley, MA: Newbury House.

Kibbee, D. (1991). For to speke Frenche trewely: The French language in England, 1000-1600: Its status, description and instruction. Amsterdam: Johns Benjamin.

Long, M. H. (1991). Focus on form: A design feature in language teaching methodology. In K. de Bot, R. B. Ginsberg, & C. Kramsch (eds.), Foreign language research in cross-cultural perspective (pp. 39-52). Amsterdam: John Benjamins.

Luther, M. (1524). Letter to the mayors and aldermen of all of cities of Germany in behalf of Christian schools. In R. Ulich (ed.) (1954), Three thousand years of educational wisdom (pp. 218-38). Cambridge, MA: Harvard University Press.

Mackey, W. F. (1965). Language teaching analysis. London: Longman.

Mitchell, R. & Myles, F. (2004). Second language learning theories. London: Hodder Arnold.

Murphy, James J. (ed.) (1987). Quintilian on the teaching of speaking and writing. Carbondale: Southern Illinois University Press.

Musumeci, D. (1997). Breaking tradition: An exploration of the historical relationship between theory and practice in second language teaching. New York: McGraw-Hill.

Reagan, T. (1996). Non-western educational traditions. Mahwah, NJ: Lawrence Erlbaum.

Richards, J. C. & Rogers, T. S. (2001). Approaches and methods in language teaching. Cambridge: Cambridge University Press.

Seybolt, T. (trans.) (1921). The Manuale Scholarium: An original account of life in a mediaeval university. Cambridge, MA: Harvard University Press.

Smith, L. G. & Smith, J. K. (eds.) (1994). Lives in education: A narrative of people and ideas. New York: St Martin's Press.

Stern, H. H. (1983). Fundamental concepts of language teaching. Oxford: Oxford University Press.

Thomas, M. (2004). Universal grammar in second-language acquisition: A history. London: Routledge.

Titone, R. (1968). Teaching foreign languages: An historical sketch. Washington, DC: Georgetown University Press.

Ulich, R. (ed.) (1954). Three thousand years of educational wisdom. Cambridge, MA: Harvard University Press.

Vergerius, P. P. (1403). De ingenuis moribus et liberalis studiis. [On the conduct and education of young people]. In W. H. Woodward (trans. and ed.) (1921), Vittorino da Feltre and Other Humanist Educators (pp. 96-118). Cambridge: Cambridge University Press.

Woodward, W. H. (trans. and ed.) (1921). Vittorino da Feltre and other humanist educators. Cambridge: Cambridge University Press.

Young, W. J. (trans. and ed.) (1959). Letters of St Ignatius of Loyola. Chicago: Loyola University Press.

Part III Psycholinguistic

Underpinnings of Language Learning

The Handbook of Language Teaching Editedby MicheelH.LnngandCethermeL.Doughty © 2009 Blackwell PublishDgLtd. BBN: 978-1-408-15789-5

5 The Language-Learning Brain

ALAN BERETTA

In second language learning and teaching, there seem to be two principal ways in which neurolinguistics is of interest: as a source of evidence to support a particular approach to SLA, usually via the critical period issue; and as a source of evidence to promote some supposedly "brain-compatible" classroom teaching technique and to dismiss other techniques that fail on this criterion. The latter use of language-brain research is, to say the least, premature. Even the neuroimaging work on the critical period issue appears to entertain higher expectations than can reasonably be sustained in view of the uncontroversial fact that far less complex problems in language-brain research remain to be solved. This chapter attempts to provide a more realistic perspective.

In an attempt to anchor debate to reality, the question that first needs to be considered is what the enterprise of neurolinguistics could possibly be. After all, there is nothing obvious about what is meant by grafting linguistics onto neuro. Once this question is confronted, the answer that it yields may serve as a basis for judging what claims are warranted and what expectations might reasonably be entertained with respect to brains and second language teaching and learning.

What is Neurolinguistics?

Someone wishing to find out what neurolinguistics is might turn to the Linguistic Society of America's website (www.lsa.org), as it contains sketches of the various sub-disciplines of linguistics. Here is what it says about neurolinguistics:

What is neurolinguistics about? Where in your brain is a word that you've learned? If you know two languages, are they stored in two different parts of your brain? Is the left side of your brain really the language side? If you lose the ability to talk because of a stroke, can you learn to talk again? Do people who read languages written from left to right (like English) think differently from people who read languages written from right to left (like Hebrew and Arabic)? What about if you

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

read a language that is written using some other kind of symbols, like Chinese or Japanese? If you're dyslexic, is your brain different from the brain of someone who has no trouble reading? All of these questions and more are what neurolinguistics is about.

You and I may be interested in some of these questions or not. You and I may find some of them quite bizarre. But you and I would be none the wiser after reading this passage as to what neurolinguistics could possibly be. The problem is not that the passage fails to demarcate neurolinguistics, to specify what counts as neurolinguistic and what does not. Let us be quite clear about that: boundary-setting for domains of interest is never at issue in any kind of scientific inquiry, as is plain from the slightest acquaintance with the history of science. The problem is this, that the term neurolinguistics links together two areas of inquiry, which begs the question of what sort of relationship they could have, given current levels of understanding, and what sort of insight the relationship might ever yield such that anyone could care. A random series of questions, clinical, Whorfian, whatever, does not begin to offer any idea what is at stake, that is, what is at stake theoretically.

Having broached the notion of theory, a caution is in order. Theory certainly matters, but whatever theory is at issue, it cannot be neurolinguistic theory. The term make no sense at all at present. It is true that one constantly hears about political theory, feminist theory, social theory, educational theory, queer theory, critical theory, framing theory, and border theory, but in these sorts of cases, it is difficult to see why common sense and experience of the world could not arrive at the same conclusions as are given by the theories. If the term theory is to be used in the way it is normally understood in the sciences, that is, an idea that yields genuine insight and which agrees with solid and publicly demonstrable evidence, then it is pointless to talk about neurolinguistic theory.

A criticism that the LSA description of neurolinguistics is atheoretical coupled with a statement that the very idea of neurolinguistic theory is at present entirely senseless appears to be paradoxical. Even more so, since it will be maintained in what follows that neurolinguistic inquiry had better be guided by theory or it is worthless. In order to resolve these apparent contradictions, and to give some idea of what combining the neuro and the linguistic might yield, discussion will start from a position of common sense, which is the starting point for many when they contemplate the relationship between language and brain, and which unfortunately remains the apparent guiding principle for many seasoned language-brain researchers. It will be argued that common sense needs to be abandoned, which prepares the ground for a notion of theoretically-guided neurolinguistic inquiry.

The commonsense approach is to wonder how something mental, like language, can interact with a physical brain, and yet know that it must. Thus, starting from common sense, it is possible to approach the language-brain relation as mind-body dualists, no better off than Descartes was more than 300 years ago when he formulated the problem, similarly confronted with a mystery and similarly left

with no way to move forward. Descartes really did have no option but to posit two distinct substances because in his mechanical universe, everything was constrained by the laws of mechanism, everything physical; that is, everything except mind. Researchers today, on the other hand, have the advantage of hindsight, not smarter, as they say, just later. With hindsight, it is known that the idea of a mechanical universe crumbled later in the seventeeth century, when Newton proposed a notion of gravity, but the consequences for the mind-body problem do not seem to have been immediately apparent. However, by the eighteenth century, the chemist Joseph Priestley was able to see that the inevitable consequence was the total rejection of the commonsense mind-body problem and a radical re-conception of the relationship between the mental and the physical. Priestley, too, started out as a mind-body dualist, so what better than to follow his progress from that belief to the standard materialism that is considered normal in cognitive science today.

Priestley initially assumed "the soul to be a substance so entirely distinct from matter as to have no property in common with it" (1777, p. xi). The "soul," or "spirit," or "mind" consisted of something airy and insubstantial, while the "body," or "matter," or the "physical" was solid, inert, and occupied space. He began to wonder whether "either the material or the immaterial part of the universal system was superfluous" but "relapsed into the general hypothesis of two entirely distinct and independent principles in man" (p. xii). Having idly wondered about the issue, Priestley thought it through in earnest in Disquisitions relating to matter and spirit (1777). Mindful of gravitational force and also of more recent insight into electromagnetic forces, he concluded that the commonsense notion of body was dead to science; matter with properties such as inertness, solidity, and so forth, he reasoned, does not exist. Matter instead possesses "powers," forces of attraction and repulsion which lack solidity, do not occupy space, are not inert. "Matter is not the inert substance it has been supposed to be" since "powers of attraction and repulsion are necessary to its very being" (p. xxxviii); without the power of attraction, "there cannot be any such thing as matter . . . for when we suppose bodies to be divested of it, they come to be nothing at all" (p. 5); take away attraction and "solidity itself vanishes" and particles would just fall apart and "be dispersed" (p. 6).

The mind-body problem, in Priestley's argument, was based on a notion of body that science had shown not to exist. Since the mind-body problem was how something lacking solidity and not occupying space could interact with bodies that had solidity and did occupy space, the problem has disappeared because physical properties, in a new conception of matter, also lack solidity and do not occupy space. For Priestley, it made better sense to think of an enlarged conception of matter "especially as we know nothing more of the nature of substance than that it is something which supports properties" (p. 17); "powers of sensation or perception, and thought" (p. 22) could belong to the same substance as supports properties of attraction and repulsion. As he observed, "powers of sensation and thought . . . have never been found except in conjunction with a certain organized system of matter" (p. 26) which comprises both mind and

brain. "Thought is a property of the nervous system, or rather of the brain" (p. 27); whatever matter supports a brain, supports a mind too.

In the same vein as Priestley, it has been argued more recently that the notion of body turns out to be a useless concept which has no more meaning in effect than whatever "finds a place in intelligible explanatory theory" and theory can contain within it anything, "however offensive to common sense" (Chomsky, 1995, p. 5). With the loss of any meaningful notion of body 300 years ago, the mind-body problem no longer had any coherent formulation. The challenge for language-brain research is of a more familiar kind in the sciences: a unification problem, how to unify two bodies of theoretical knowledge (Chomsky, 1995). There is linguistic theory and there is brain theory, but at present neither theory constrains the other. For centuries, physics and chemistry were in the same position and unification only became possible a couple of generations ago (Brock, 1992; Chomsky, 1995). Within physics today, a major question is to how to unify the different forces (gravitational, electromagnetic, strong and weak nuclear). So, when the term neurolinguistics is used, reasonably, all that can be meant is an area of inquiry whose ultimate goal is theory unification, but with a full recognition that unless substantial progress is made in that direction, the term neurolinguistic theory is without content since it presupposes that unification has already occurred.

How can progress toward eventual unification be made? The most obvious answer is for linguists and brain scientists to continue doing what they are already doing, namely, developing ever more explanatory theories of language and of brains, with the hope that with increased understanding, it will one day be possible to perceive how unification is possible. This answer, of course, prescribes no particular role for neurolinguistics at all. Neurolinguistics has a purpose only to the extent that examining language in brains can be informative about either brains or about language. Does examining language in brains tell us anything about language? That is, is it possible to think of a problem in linguistics such that something about brains provides an answer? The question is rhetorical because the only answer, until more is known about brains perhaps, is silence. Let us ask the converse question: Does examining language in brains tell us anything about brains? The answer to this question, one suspects, is not quite so debilitating because it may be reasoned that if something worthwhile is known about language, then at least it is possible to ask the brain something coherent. Of course, no one knows how a brain computes language - that is what it is like to be non-unified - but to the extent that the brain makes such distinctions that agree with the distinctions made in a theory of language, it might, not irrationally, be hoped that something is being learned about how brains care about language. What is learned, however, will have the character of peculiar facts because we do not have a relevant understanding of brains. Again, such is the nature of inquiry when two bodies of knowledge are not unified. The best prospect is to ask the brain a linguistically coherent question and look for an answer that matters to current brain theory. Then the inquiry is coherent in the only ways currently available, and the hope is that it becomes possible to build bridges that will provide eventual insight.

There is an alternative: experiment that treats language as a mere commonsense notion and brains only in terms of location. That description covers much of the vast literature on language and brain, including most of the work that seems closest to the interests of L2 researchers and teachers, the work on differences and similarities between L1 and L2 in brain territory. It is the kind of inquiry that Poeppel and Embick (2005) refer to, without approval, as the standard program in neurolinguistics.

Before discussing this further, a few more comments on common sense and theory. In the early days of modern science, the mechanical philosophy was based on a commonsense apprehension of Nature. Expectations were very high. Think of Bacon's hope that, if a natural and experimental history such as he proposed was provided, "the investigation of nature and of all sciences will be the work of a few years" (1620). From hopes for imminent and comprehensive understanding amenable to common sense, very soon afterwards, with the advent of Newton's gravitational force which defied the prevailing commonsense assumption of mechanism that physical objects could only move through direct contact with other physical objects, expectations changed. From that time on, it would be necessary to resign ourselves to a far more limited access to Nature's secrets, namely, access via the proxy of theory entirely unconstrained by the dictates of common sense (Chomsky, 2000). Centuries later, this has become familiar. The physicist Richard Feynman comments that, "one had to lose one's common sense in order to perceive what was happening at the atomic level" (1985, p. 5). Quantum theory, he notes, is "absurd from the point of view of common sense. And it agrees fully with experiment" (p. 10). Quantum theory tortures common sense, but Feynman observes that physicists have learned not to care if common sense is tortured. What matters is insight-giving theory that agrees with experiment.

Although the shift from common sense to theory must have been hard to take during the early days of modern science, it is rather surprising to find that, so many years later, a great deal of neuroimaging work on language anachronist-ically prefers commonsense notions of language to theoretical principles. In an attempt to draw attention to this atheoretical tendency, Poeppel (1996) criticized several high-profile researchers for their imaging studies of phonological processing which concerned themselves little with what was known about phonological processing or even about phonology. The criticism was not appreciated (Demonet et al., 1996). I recall a discussion a couple of years ago with a well-known scholar who has published many imaging studies of language. I was informed that it was premature in neuroscience for experiment to worry about theoretical issues. What was needed, he explained, was a long period of description, so informal notions of language were appropriate. Venturing dissent, I was asked to explain why so many highly intelligent people were doing descriptive research. I could not answer then and I am unable to answer now. However, imagine physicists, faced with some tough problem, and deciding that what was called for was a period in which theoretical concepts would be put on ice and replaced with informal notions of, say, work or energy or light. Everyone would scoff at such a suggestion,

and yet, when it comes to language, that is precisely the approach that is so often adopted. It amounts to what Chomsky has called methodological dualism, more ruinous than mind-body dualism.

It is not just linguistic theory that is ignored, but theory of brain also seems very thin on the ground in the standard research program in neurolinguistics. Since most neuroscience studies of language use positron emission tomography (PET) or functional magnetic resonance imaging (fMRI), the focus is heavily on location. Although there is by now a vast literature describing locations of this informal notion of language or that, it is far from obvious that neural location does any theoretical work. Function has to occur somewhere, and to some extent we can already navigate a brain in terms of different functions, and certainly these are useful facts to know. But that does not mean that location is doing explanatory work; it may instead be merely epiphenomenal, just as linearity is epiphenomenal in linguistic theory - every sentence has to be heard or produced in a linear string, but that observation has no theoretical status.

Electrophysiological tools, by contrast, record aspects of brain that are theoretically more obviously interesting. Electroencephalography (EEG) and magne-toencephalography (MEG) record brain activity directly, and the signal they pick up is the synchronous firing of a large number of neurons. Synchronous firing has long been considered a central explanation of how neurons form assemblies for particular computational tasks (Abeles, 1991; Gray et al., 1989; Singer, 2000). Also, as Phillips (2005, p. 81) has pointed out, there are substantial hypotheses regarding timing of language function but no detailed hypotheses regarding location. In neural studies of second language processing, however, EEG studies are few and far between and MEG studies, so far as I know, are non-existent.

Some Promising Neurolinguistic Research

The above characterization of neurolinguistics may serve (or not) to keep expectations in perspective, but hopefully not to curb interest. That is far from the intention. So long as we devise experiments that ask the brain theoretically coherent questions, we are doing the only reasonable thing we can do, which at least holds out the possibility of building bridges between language and brain.

For example, some rather intriguing work has been carried out which finds a neural response to phonological categories. This is unusual, as most neurolinguistic studies have focused on acoustic and phonetic features, but have not addressed phenomena relevant to the phonological level of analysis. Phonetic categories, of course, group together a number of similar sounds, and neural responses have been established for within-category discrimination of these sounds. For example, Dehaene-Lambertz et al. (1997) found that within-category contrasts produced a Mismatch Response, a particular neural response to a sequence of standards (identical sounds) interrupted by a deviant (different sound). The deviant elicits the mismatch response at ~150-250 ms in auditory cortex.

But does the auditory cortex register phonological categories? Since phonological categories are abstract symbolic representations that have an all-or-none character, within-category variation is irrelevant; the same rules apply across the board to all members of a category. One body of research was able to use the mismatch response to get at phonological categories. It is a well-known fact that the perceptual cue for voicing in stop consonants (e.g., /1/ and /d/) is voice onset time (VOT) - the amount of time that elapses between the release of a consonant and the onset of voicing. Stop consonants that have shorter VOTs are perceived as voiced, while those with longer VOTs are perceived as unvoiced. The perceptual boundary is quite sharply drawn at ~30 ms. Phillips et al. (2001; see also Aulanko et al., 1993) used a range of different VOT values of /dae/ and /tae/. With VOT varying both within and between /dae/ and /tae/, there was no many-to-one ratio of standards and deviants at acoustic or phonetic levels; all sounds were different in terms of VOT. There could only be a many-to-one ratio if all of the different realizations of each category were treated as the same thing, that is to say, at the phonological level. In the event, different VOT values of one category /tae/ interrupted by a single value of /dae/ elicited the characteristic mismatch response in auditory cortex, indicating that the brain does care about phonological categories.

In addition to the impressive body of work taking us, in Poeppel's phrase, from vibration in the ear to abstraction in the head, thoughtful work has been pursued that investigates lexical access. Embick et al. (2001), in a MEG study of lexical frequency, found that the first peak in the magnetic waveform to vary as a function of frequency occurred at ~350 ms after stimulus presentation. In a lexical decision task, the M350 peak for frequent words was earlier than the peak for non-frequent words. The M350 also occurred earlier for repeated words than for non-repeated words (Pylkkanen et al., 2000). In a further study, Pylkkanen, Stringfellow, & Marantz (2002) observed that the M350 responded to phonotactic probability, but not to density similarity neighborhoods. That is, it is sensitive to the frequencies of sounds and their sequences in words, but not to the extent with which a word sounds like lots of other words. This is interesting because in behavioral studies, it had been shown that phonotactic probability and neighborhood density had opposite effects: phonotactics facilitated sublexical processing, while density had an inhibitory effect, slowing word recognition. What Pylkkanen et al. did was to use stimuli that possessed both properties. The fact that the M350 responded to phonotactics and not density argues that the peak indexes an early stage of (sub)lexical processing rather than a later recognition stage when competition from like-sounding words would obtain.

Further sharpening our grasp of the M350, Fiorentino & Poeppel (2007) showed that it is not an index simply of lexical access but of root access, a rather more precise concept because roots and what is contained in them are important theoretical entities (see, e.g., Halle & Marantz, 1993). Another important theoretical issue is polysemy, the different senses of a word. Syntactic principles operate on identity in sentences such as The newspaper decided to change its format (Chomsky, 1970): newspaper and its have identical reference and yet they have two different

senses (newspaper refers to the publisher and its to the publication). Of interest, then, is whether polysemous senses share the same root, and are thus identical at that level, or whether they have separate roots, as is widely agreed to be the case with homonymy. Both positions have their adherents but evidence one way or the other has been difficult to obtain. Two recent MEG studies have investigated polysemy and homonymy (Beretta, Fiorentino, & Poeppel, 2005; Pylkkanen, Llinas, & Murphy, 2006). Without going into detail, the findings of these studies can be explained if the M350 accesses roots and if polysemous senses, but not homonym-ous meanings, share the same root.

Neurolinguistic Research and L2 Learning and Teaching

One strand of research that is often discussed in the second language literature asks the question: Is L1 like L2 in the brain? There are many reviews of this literature (e.g., Fabbro, 2001; Perani, 2005; Perani & Abutalebi, 2005), so a cursory summary will suffice before it is asked in what ways the body of work is informative.

Starting with PET and fMRI studies, the general picture that seems to be emerging is twofold. First, acquisition of L2 engages very much the same neural regions as L1. Second, these relatively stable common patterns can be mediated by age of acquisition and the level of proficiency attained. Basically, with respect to tasks that are claimed to tap grammatical resources, subjects who acquired their L2 early show the same regional activation, whereas those who acquired their L2 late use the same areas as L1 but also show some additional activation. With respect to tasks that are considered to access lexical-semantic processing, in early or late acquisition of L2, if proficiency is high, the same neural areas are activated.

Why ask a brain if L1 is like L2 with regard to location? For Kim et al. (1997), as for many others, this is self-explanatory: "Here we investigate the fundamental question of how multiple languages are represented in the human brain" (p. 171). No further motivation is offered. Leaving aside, for the moment, the fact that the term fundamental is misplaced given that far simpler questions have barely been addressed (as mentioned earlier), the lack of any clearly stated theoretical focus reflects the descriptive nature of much of the work. The main issue stated in a number of reports concerns critical periods in language acquisition, but unlike much of the behavioral work in this domain of inquiry (such as Johnson & Newport, 1989; White & Genesee, 1996), there is little or no concern for questions that are driven by some aspect of linguistic theory. The result is that the behavioral work is often capable of fairly precise interpretation whereas the neurolinguistic work is not.

As in the behavioral literature, independent variables that are frequently manipulated in the neuroimaging literature are age of first exposure to a language and level of attainment. What distinguishes the two literatures is the content of the dependent variables. Dependent variables in at least some of the behavioral

studies attempt to isolate some linguistically relevant element or other, while the dependent variables in typical PET and fMRI studies bear no obvious relation to what is known about the nature of language. Some studies (e.g., Chee, Tan, & Thiel, 1999) ask subjects to take letter strings like cou and think of a word that begins with those letters, perhaps coulomb or courtesy; or to think of words that end in a letter string like ter, maybe banter. Others ask subjects to think of a word that begins with a single letter that is called out to them (Perani et al., 2003). Still others have subjects listen to stories (Perani et al., 1996), or silently describe events that happened during the previous day (Kim et al., 1997), or listen to (unspecified) sentences in different languages (Pallier et al., 2003), or translate and read words (Price, Green, & von Studnitz, 1999), listen to dialogues (Nakai et al., 1999), and so forth. Language, in some sense, is plainly involved in all of these tasks, but in none of the reports is there any indication as to which theoretical element is being tapped into, and it is hard to imagine candidates. Perani et al. (1998) start out by discussing parameter setting, something behavioral studies have pursued, but in the event they end up using a story-telling task, a task that does not pick out any particular parameter at all.

Part of the problem of relating L2 to L1 in brain space is that we know precious little about L1 in brain space. Even the rare experiment that contrasts sentences that differ only in respects that have relevance in the linguistic literature, such as subject vs. object movement in relative clauses in L1, has found it difficult to nail down anything solid and reliable. For example, Caplan and his colleagues have carried out a series of experiments looking only at relative clauses. In the first study, Stromswold et al.'s (1996) PET study of relative clauses, syntactic processing showed increased activation, not only in Broca's area, but in a specific portion of that area, the pars opercularis. A follow-up study (Caplan, Alpert, & Waters, 1998) with the same materials, but with female instead of male subjects, found increased activation in one part of the pars opercularis, higher and more anterior than for males. This difference was not predicted, but after the fact, it was speculated that this might indicate a male-female difference. A separate experiment in the same report looked at relative clauses with two propositions vs. relative clauses with one proposition. No increased activation was found in the traditional language areas, but in visual cortex. Clearly, this was not predicted, but it was considered that it was "most likely due to differences in some aspect of post-interpretive processing . . . visual mental imagery processes" (p. 548). A later study found increased regional cerebral blood flow for one kind of relative clause in a part of Broca's area not implicated in the earlier studies, the pars triangularis (Caplan et al., 2000). Why should that be? Apparently, "both these areas can be activated by syntactic processing" (p. 70). Undeterred, Caplan et al. (2001) proceed to examine relative clauses with and without padding. They find no difference for the critical part of the sentence (immediately following the padding). Perhaps this was because relative clauses involved movement out of subject or object position, perhaps different task demands, possibilities that "can only be answered by further detailed studies" (2001, p. 37). A further detailed study (Caplan, Waters, & Alpert, 2003) used the same materials and tasks as

Stromswold et al. (1996) and Caplan et al. (1998), but this time with subjects aged 70-79. Unlike the earlier studies, increased blood flow was found in the left inferior parietal, which must be "related to sentence processing ... and syntactic comprehension" (p. 118). There was also increased activation in left superior frontal cortex, near the midline, an area which, it is surmised, "may be involved in syntactic processing" (p. 118). Why might older subjects apparently use totally different parts of the brain from the younger subjects? One possibility mooted is that they had received two years less education. This was explored in a second experiment in Caplan et al. (2003). Younger subjects with less education were tested. Now, increased activation was observed in superior frontal, but a part of it quite distinct from any previously seen. Increases were also reported in the motor planning region, which might be due to different eye movements, or "could reflect greater deployment of attention" (p. 122). An increase in the precuneal parietal, not the parietal area observed in the older subjects, showed increased activation too, and it was considered that this could be due to subtle aspects of language processing. But, perhaps the older subjects' different neural activation was not due to their lower education, but to their less accurate task performance, a possibility also examined by Caplan et al. (2003). Older subjects whose performance matched the younger subjects were tested. Increased regional blood flow was found in one more area of prefrontal cortex not implicated in the earlier studies, an area normally considered to be involved in executive functioning. Caplan et al. note that it would appear also to be involved in syntactic processing.

This series of studies is of interest because of the persistent attention to a quite well-circumscribed range of sentence types. In the end, we know that if you place subjects in a PET or fMRI scanner, then something is going to happen, but it is very hard to say in advance what that will be, and after the fact, very difficult to narrow candidate explanations. This is not just a problem for Caplan and his colleagues, but a problem for everyone at a time when theories of language and brain seem far away from each other. If one asks subjects to listen to stories in L1 or L2, for example, then surely prediction and explanation are rendered all the more indeterminate.

Turning to the EEG studies, the picture that emerges from the relatively fewer studies that have been carried out is, roughly speaking, that as proficiency increases ERPs (event-related potentials) for L2 processing become more like those in L1. Research by Friederici and her colleagues has suggested that semantic violations modulate the N400, while phrase structure violations have an effect on early anterior negativity and afterwards on late positivity (the P600). An early study by Weber-Fox and Neville (1996) found qualitatively different ERPs between L2 and L1 for syntactic processes, but quantitative differences for semantic processes (such as a delayed N400 for those who acquired the L2 at a later age). Isel, Hahne, and Friederici (2000) and Hahne and Friederici (2001), using similar stimuli but for different L2s, found N400 modulation for semantic violations, along with a lack of the expected early anterior negativity modulation for phrase structure violations and no P600. Hahne (2001), again using similar stimuli, found an N400 effect for both L1 and L2 groups, while for phrase structure violations

there was no modulation of early anterior negativity for the L2 group, in contrast to the L1 group, suggesting that automaticity of processing was reduced. The P600 was similar for both L1 and L2. Reflecting on these findings across different studies, Hahne (2001) notes that the subjects in her study were more proficient in L2 than the subjects in Isel et al. (2000) and Hahne and Friederici (2001), and while the main difference between her study and the other two is that the more proficient subjects showed a P600 effect, all three studies revealed N400 modulation, as does an experiment by McLoughlin, Osterhout, and Kim (2004). She infers from these facts that semantic processes are the first to become L1-like, and syntactic processes become so with greater proficiency. This does not address the difference in anterior negativities, but a recent study (Ojima, Nakata, & Kakigi, 2005) finds that high-proficiency L2 subjects do show anterior negativities for syntactic processing whereas low-proficiency L2 subjects do not. In addition, a study by Friederici, Steinhauer, and Pfeifer (2002) finds that training on an artificial language called BRICANTO that reportedly has UG-like constraints yields the same early automatic negativities and P600 effects as seen in normal L1 processing. Age does not seem to be a contributory factor in these experiments, but level of proficiency does. These considerations are used to argue against a very strong form of the critical period hypothesis.

The EEG studies of critical period issue are conducted against a background of increasingly well-understood electrophysiological responses in the time course of linguistic processing, namely the N400, LAN (left anterior negativity), and the P600. As such, the program is seeking to construct language-brain relations in a way that has some neural justification (timing matters in proposals regarding neuronal population coding), and also in a way that is not incoherent linguistically, though rather blunt. Blunt, because the contrasts are relatively coarse and there is no attention to the very fine computations that linguistic theory cares about.

Some Confusions

The above treatment of the L1/L2 neuroimaging studies is perhaps slightly less sanguine than is to be found in some reviews, but if reservations have been expressed about research that is plainly serious, it is hard to know what to say about some of the views that have been advanced in sections of the second language literature regarding what claims neuroscience entitles us to make about second language acquisition and teaching.

It transpires there is such a literature whose concern is for brain-compatible language teaching (Bimonte, 1998; Dhority & Jensen, 1998; Genesee, 2000; Lombardi, 2004). Apparently, the brain is complex and adaptive, and varies across individuals, so language teaching should be approached with a variety of techniques, such as group work, and so forth. The brain is social, so teaching should involve cooperative learning strategies. The brain is emotional, so it is critical to lower the "affective filter." Left brain and right brain are respectively associated with rational and artistic orientations, so one can deploy art or music to teach

math, physics, or language; though some of the authors regard teaching based on left-brain/right-brain considerations to be outmoded. The brain links language to other parts of the brain, so good teaching practice should, for example, not present phonics independently of meaning. To do so would be misguided. Brain connections are shaped by the environment, so teaching makes a difference and one should not give up on older learners. And so on.

The blandest comment that presents itself is that the conclusions do not follow from the premises. However, even this is tantamount to saying that the authors have abandoned the most rudimentary notions of rationality. So, lacking any understanding of what counts as argument in this literature, I will not, indeed cannot, criticize it any further.

With regard to SLA, the picture is not much clearer. A typical account might, for example, seek to promote a language socialization paradigm and jettison cognitive science models, and claim that brain research provides the impetus (e.g., Watson-Gegeo, 2004). The appeal to neuroscience is littered with misconceptions, such as that it is a problem for cognitive science approaches that "research has discovered no structure in the brain that corresponds to a Language Acquisition Device as argued by Chomsky" (p. 333). Chomsky argued no such thing, of course, but profound confusions of this sort will not be pursued here (however, see Gregg, 2006 for a thorough critique).

Conclusion

Neurolinguistics is an area of inquiry whose ultimate goal is theory unification. To engage in neurolinguistic work is to attempt to integrate what is known about language, which I have been referring to as '"linguistic theory," with what is known about brains, which I have been informally calling "brain theory." Given that at present neither theory constrains the other, it is obvious that neurolinguistics is at a very preliminary stage of development. The challenge facing neurolinguistics is awe-inspiring, and there are those who have concluded (prematurely, since we simply do not know) that the relationship of mind to brain will forever remain a mystery (e.g., McGinn, 1993, 1999; Uttal, 2005). It is unfortunate, then, that much of language-brain research is conducted without reference to either linguistic or brain theory, because it suggests that there is some other way of finding out about nature than via theory. Indeed, as noted earlier, the standard research program in neurolinguistics largely proceeds as if what is called for is description. The prospects for theory unification are remote indeed if most current work is atheoretical.

Commenting on an earlier research program that favored theory-free description, the biologist Ernst Mayr (cited in Hull, 1988) offered a cautionary tale. In the nineteenth century, in an attempt to solve the mysteries of inheritance, one scientist, Carl Friedrich von Gärtner, judged that the best approach was theory-free Stakhanovite data collection, relentless unbiased description. He carried out literally 10,000 cross-breeding experiments. Unfortunately no one was ever able

to make use of his work; they were always forced to conduct their own experiments because, even in as many as 10,000 experiments, the contrasts demanded by one hypothesis or another were never found. Not surprising when we consider that theories contain within them exotic elements that torture common sense. The relevant contrasts would not occur to anyone absent a theory. Von Gartner's data collections now lie "buried in the stacks of a few libraries around the world, as sterile as the day that they were conceived" (Hull, 1988, p. 489). It is not unreasonable to wonder if today's mass accumulation of descriptive neuroimaging data on language and the brain can have much more hope of chancing upon the right contrasts that might permit insightful language-brain connections than von Gartner's work contributed to an understanding of inheritance.

Given the preliminary nature of neurolinguistic inquiry and the descriptive bent of much of the work currently conducted, the upshot is that brain research can at present provide evidence that is little more than suggestive to a model of L2 acquisition and less than that to the practice of language teaching. The neuroimaging studies on the critical period question can be seen, at best, as to be taken under advisement. It may seem compelling to use brain data to make a point, but there is little justification for privileging brain data over other kinds of data. The same goes for teaching research. If there was some reason to support group work or dispense with it before, there is no better reason to be found in the neurolinguistic literature. Practical applications, such as to teaching, are remote, beyond remote.

REFERENCES

Abeles, M. (1991). Corticonics: Neural circuits of the cerebral cortex. Cambridge: Cambridge University Press.

Aulanko, R., Hari, R., Lounasmaa, O. V., Naatanen, R., & Sams, M. (1993). Phonetic invariance in the human auditory cortex. NeuroReport 4, 1356-8.

Bacon, F. (1620). Preparative toward a natural and experimental history. http://etext.library. adelaide.edu.au/b/bacon/francis/preparative/complete.html.

Beretta, A., Fiorentino, R., & Poeppel, D. (2005). The effects of homonymy and polysemy on lexical access: An MEG study. Cognitive Brain Research 24, 57-65.

Bimonte, R. (1998). Mysteries of the brain: Students thrive in a brain-compatible learning environment. Momentum 16-18.

Brock, W. H. (1992). The Norton history of Chemistry. New York: Norton.

Caplan, D., Alpert, N., & Waters, G. (1998). Effects of syntactic structure and propositional number on patterns of regional cerebral blood flow. Journal of Cognitive Neuroscience 10, 541-52.

Caplan, D., Alpert, N., Waters, G., & Olivieri, A. (2000). Activation of Broca's area by syntactic processing under conditions of concurrent articulation. Human Brain Mapping 9, 65-71.

Caplan, D., Vijayan, S., Kuperberg, G., et al. (2001). Vascular responses to syntactic processing: Event-related fMRI study of relative clauses. Human Brain Mapping 15, 26-38.

Caplan, D., Waters, G., & Alpert, N. (2003). Effects of age and speed of processing on rCBF correlates of syntactic processing in sentence comprehension. Human Brain Mapping 19, 112-31.

Chee, M. L. W., Tan, E. W. L., & Thiel, T. (1999). Mandarin and English single word processing studied with functional magnetic resonance imaging. The Journal of Neuroscience 19, 3050-6.

Chomsky, N. (1970). Remarks on nominalism. In R. A. Jacobs & P. S. Rosenbaum (eds.), Readings in English transformational grammar (pp. 184-221). Waltham, MA: Ginn & Co.

Chomsky, N. (1995). Language and nature. Mind 104, 1-61.

Chomsky, N. (2000). Linguistics and brain science. In A. Marantz, Y. Miyashitu, & W. O'Neil (eds.), Image, Language, Brain (pp. 13-23). Cambridge, MA: MIT Press.

Dehaene-Lambertz, S., Dupoux, E., Mehler, J., et al. (1997). Anatomical variability in the cortical representation of first and second languages. NeuroReport 8, 3809-15.

Demonet, J. F., Fiez, J. A., Paulesu, E., et al. (1996). PET studies of phonological processing: Critical reply to Poeppel. Brain and Language 55, 352-79.

Dhority, L. & Jensen, E. (1998). Joyful fluency: Brain-compatible second language acquisition. San Diego: Brain Store.

Embick, D., Hackl, M., Schaeffer, J., et al. (2001). A magnetoencephalographic component whose latency reflects lexical frequency. Cognitive Brain Research 10, 345-8.

Fabbro, F. (2001). The bilingual brain: Cerebral representation of languages. Brain and Language 79, 211-22.

Feynman, R. P. (1985). QED: The strange theory of light and matter. Princeton, NJ: Princeton University Press.

Fiorentino, R. & Poeppel, D. (2007). Compound words and structure in the lexicon. Language and Cognitive Processes 12, 953-1000.

Friederici, A., Steinhauer, K., & Pfeifer, E. (2002). Brain signatures of artificial language processing: Evidence challenging the critical period hypothesis. Proceedings of the National Academy of Sciences USA 99, 529-34.

Genesee, F. (2000). Brain research: Implications for second language learning. Center for Research on Education, Diversity & Excellence, Occasional Report, ERIC_00_12_brain. ERIC Clearinghouse on Language and Linguistics: Washington, DC.

Gray, C. M., König, P., Engel, A. K., & Singer, W. (1989). Oscillatory responses in cat visual cortex exhibit inter-columnular synchronization which reflects global stimulus properties. Nature 338, 334-7.

Gregg, K. (2006). Taking a social turn for the worse: The language socialization paradigm for second language acquisition. Second Language Research 22, 1-30.

Hahne, A. (2001). What's different in second-language processing? Evidence from event-related brain potentials. Journal of Psycholinguistic Research 30, 251-66.

Hahne, A. & Friederici, A. (2001). Processing a second language: Late learners' comprehension mechanisms as revealed by event-related potentials. Bilingualism: Language and Cognition 4, 123-41.

Halle, M. & Marantz, A. (1993). Distributed morphology. In K. Hale & S. J. Keyser (eds.), The view from Building 20 (pp. 111-76). Cambridge, MA: MIT Press.

Hull, D. (1988). Science as a process: An evolutionary account of the social and conceptual development of science. Chicago: University of Chicago Press.

Isel, F., Hahne, A., & Friederici, A. (2000). Auditory processing of German sentences by French late bilinguals. Poster presented at the Conference of Architectures and Mechanisms of Language Processing (Amlap). Leiden, The Netherlands, September.

Johnson, J. S. & Newport, E. L. (1989). Critical period effects in second language learning: The influence of maturational state on the acquisition of English as a second language. Cognitive Psychology 21, 60-99.

Kim, K. H., Relkin, N. R., Lee, K. M., & Hirsch, J. (1997). Distinct cortical areas associated with native and second languages. Nature 388, 171-4.

Lombardi, J. (2004). Practical ways brain-based research applies to ESL learners. Internet TESL Journal 10, 8.

McGinn, C. (1993). Problems in philosophy: The limits of inquiry. Oxford: Blackwell.

McGinn, C. (1999). The mysterious flame: Conscious minds in a material world. New York: Basic Books.

McLoughlin, J., Osterhout, L., & Kim, A. (2004). Neural correlates of second-language word learning: Minimal instruction produces rapid change. Nature Neuroscience 7, 703-4.

Nakai, T., Matsuo, K., Kato, C., et al. (1999). A functional magnetic resonance imaging study of listening comprehension of language in human at 3-tesla comprehension level and activation of the language areas. Neuroscience Letters 263, 33-6.

Ojima, S., Nakata, H., & Kakigi, R. (2005). An ERP study of second language learning after childhood: Effects of proficiency. Journal of Cognitive Neuroscience 17, 1212-28.

Pallier, C., Dehaene, S., Poline, J.-B., et al. (2003). Brain imaging of language plasticity in adopted adults: Can a second language replace a first? Cerebral Cortex 13, 155-61.

Perani, D. (2005). The neural basis of language talent in bilinguals. Trends in Cognitive Sciences 9, 211-13.

Perani, D. & Abutalebi, J. (2005). The neural basis of first and second language processing. Current Opinion in Neurobiology 15, 202-6.

Perani, D., Abutalebi, J., Paulesu, E., et al. (2003). The role of age of acquisition and language usage in early, high-proficient bilinguals: An fMRI study during verbal fluency. Human Brain Mapping 19, 170-82.

Perani, D., Dehaene, S., Grassi, F., et al. (1996). Brain processing of native and foreign languages. NeuroReport 7, 2439-44.

Perani, D., Paulesu, E., Sebastian-Galles, N., et al. (1998). The bilingual brain: Proficiency and age of acquisition of the second language. Brain 121, 1841-52.

Phillips, C. (2005). Electrophysiology in the study of developmental language impairments: Prospects and challenges for a top-down approach. Applied Psycholinguistics 26, 79 -96.

Phillips, C., Pellathy, T., Marantz, A., et al. (2001). Auditory cortex accesses phonological categories: An MEG mismatch study. Journal of Cognitive Neuroscience 12, 1038-55.

Poeppel, D. (1996). A critical review of PET studies of phonological processing. Brain and Language 55, 317-51.

Poeppel, D. & Embick, D. (2005). The relation between linguistics and neuroscience. In A. Cutler (ed.), Twenty-first century psycholinguistics: Four cornerstones (pp. 1-16). Mahwah, NJ: Lawrence Erlbaum.

Price, C. J., Green, D. W., & von Studnitz, R. (1999). A functional imaging study of translation and language switching. Brain 122, 2221-35.

Priestley, J. (1777). Disquisitions relating to matter and spirit. Kila, MT: Kessinger.

Pylkkanen, L., Llinas, R., & Murphy, G. (2006). Representation of polysemy: MEG evidence. Journal of Cognitive Neuroscience 18, 1-13.

Pylkkanen, L., Stringfellow, A., Flagg, E., & Marantz, A. (2000). A neural response sensitive to repetition and phonotactic probability: MEG investigations of lexical access.

Proceedings of the Twelfth International Conference on Biomagnetism (pp. 363-7). Helsinki University of Technology.

Pylkkanen, L., Stringfellow, A., & Marantz, A. (2002). Neuromagnetic evidence for the timing of lexical activation: An MEG component sensitive to phonotactic probability but not to neighborhood density. Brain and Language 81, 666-78.

Singer, W. (2000). Response synchronization: A universal coding strategy for the definition of relations. In M. Gazzaniga (ed.), The new cognitive neurosciences, (2nd edn., pp. 325-38). Cambridge, MA: MIT Press.

Stromswold, K., Caplan, D., Alpert, N., & Rausch, S. (1996). Localization of syntactic comprehension by positron emission tomography. Brain and Language 52, 452-73.

Uttal, W. R. (2005). Neural theories of mind: Why the mind-brain problem may never be solved. Mahwah, NJ: Lawrence Erlbaum.

Watson-Gegeo, K. A. (2004). Mind, language, and epistemology: Toward a language socialization paradigm for SLA. Modern Language Journal 88, 331-50.

Weber-Fox, C. M. & Neville, H. J. (1996). Maturational constraints on functional specializations for language processing. Journal of Cognitive Neuroscience 8, 231-56.

White, L. & Genesee, F. (1996). How native is near-native? The issue of ultimate attainment in second language acquisition. Second Language Research 12, 233-65.

6 Sequences and Processes in Language Learning

LOURDES ORTEGA

In this chapter I review major findings concerning the sequences and processes associated with the development of a second language. I begin by characterizing the nature of interlanguage. The bulk of the chapter is then devoted to an overview of sequences, or systematic patterns uncovered in the development of various areas of the L2 grammar. Next, I examine some central processes by which learners attend selectively to aspects of the L2 input, reorganize their mental representations in light of new evidence as this becomes available to them, and then move (or not) along the sequences. I conclude the chapter with five generalizations that summarize the implications of knowledge about sequences and processes for second and foreign language teaching.

Learner Language or Interlanguage

During the 1950s and 1960s, researchers interested in understanding language teaching and learning set out to compare external differences between a given first language (L1) and a given target language (L2). The hope was that such a Contrastive Analysis approach would help them uncover areas of difficulties for L2 learners (e.g., Stockwell, Bowen, & Martin, 1965). However, beginning in the 1970s, and influenced by seminal findings that had begun to accumulate about child L1 acquisition, researchers turned to a different strategy and began analyzing the actual language samples that learners produced when they attempted to use their L2 in speaking. (Writing and signing data have also been examined, although less extensively; see, e.g., Bardovi-Harlig, 1995, for L2 writing; and Mayberry, in press, for L2 signing.) A central concept that emerged with the study of actual learner language was interlanguage (Selinker, 1972), or the language system that each learner constructs at any given point in development. Interlanguage reflects an interim competence that contains elements from both the L1 and the L2 grammar, but also elements that go beyond both. The new associated methodologies of Error Analysis and later Performance Analysis (Long & Sato, 1984) led to an

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

unprecedented wealth of findings about interlanguages. More recent developments in the last 20 years have added detail and explanatory power to the older findings.

Interlanguages are natural languages

Interlanguages are believed to be systematic, natural languages in their own right. In other words, learners are constrained in their development of a second language by the same natural principles that constrain the development of any human language. An unresolved question among second language acquisition researchers is whether these constraints and natural limits are provided by a genetic endowment specific to the human language faculty, as proponents of Universal Grammar posit (e.g., White, 2003), or by the same general cognitive learning mechanisms that help humans process and learn any other kind of information, as language psychologists argue (e.g., DeKeyser, this volume; Ellis, this volume). In this chapter, I adopt the latter position. I assume that general cognition principles can help us understand how L2 learners develop grammar knowledge and the ability to use it in order to communicate, at least for the great majority of interlanguage phenomena that are of relevance and interest for language teaching.

Interlanguages are more than the sum of input and first language

The nature of interlanguage is delimited by three well-known but nevertheless striking facts. First, learners build mental representations that are rather different from what the input in their surrounding environment looks like. For example, consider the following sentence from Oshita (2000, p. 313), who identified it in an L2 English essay written by an L1 Spanish speaker:

(1) It [a wall] was falled down in order to get a bigger greenhouse.

We can see that the English morpheme -ed has been added to a verb whose past and past participle forms are irregular in English (fall, fell, fallen) and furthermore that a verb that is intransitive in English (fall) has been made into a causative verb with a transitive meaning ('make to fall' or 'tear down'). When L2 English learners produce such interlanguage solutions, they have most certainly not picked them up from their surrounding input. These are not forms or meanings that can be learned from, say, English-speaking friends or textbooks.

Second, although knowledge of the first language definitely influences interlanguages in various ways (see Ringbom & Jarvis, this volume), many so-called errors cannot be explained by recourse to the L1 alone, including example (1) above. Spanish has regular and irregular verbal morphology and uses it to mark past tense and past participles, and it also has a verb for the intransitive meaning 'fall' (caer) and another for the transitive meaning 'make to fall' (derribar). Moreover, even when the L1 appears logically to explain an interlanguage

solution, other explanations may in fact be needed. To illustrate, if we hear How I do this? from an L1 Spanish learner and an L1 Punjabi learner, whose languages do not have inversion, we may conclude it is their respective L1s that are inducing this choice. However, if we sampled learners from a wide enough range of L1 backgrounds, including languages where inversion does exist (e.g., Dutch and German), we would find that they, too, use un-inverted questions in their English interlanguage at an early stage of L2 development. In the presence of this additional evidence, we must then conclude that the L1 cannot be the correct explanation for lack of inversion. In fact, we will later see that the utterance How I do this? results from a universally attested interlanguage solution to the problem of question formation in English, called fronting (cf. stage 3 in Table 6.5).

A third, striking fact is that many interlanguage solutions are also attested in the production of children acquiring their first language (who, therefore, do not have any other L1 knowledge on which to rely, for better or worse). Thus, the solutions in (1) have been also amply documented in data from children learning L1 English. How can we explain interlanguage solutions that are neither directly attributable to the input nor to the L1, and that are shared by first and second language acquirers?

The unavoidable conclusion is that these forms are interim systematic innovations that learners independently create when they are trying to figure out the workings of the new language system they are learning. As will become clear, these interlanguage inventions are motivated by the complex interaction of multiple forces. Those include the evidence available or absent in the input (White, 2003) and knowledge of the L1 and other known languages (Ringbom & Jarvis, this volume). But additional forces stem from the interaction between the universal shape of languages and the conceptual apparatus of the human mind. They include syntactic, semantic-discoursal (Anderson & Shirai, 1994; Hyltenstam, 1987), and statistical, as well as conceptual and sensorimotor, processing influences (N. Ellis, 2006; this volume), on the one hand, and communicative pressures and social incentives learners experience as they use the language to make meaning (Klein & Perdue, 1997; von Stutterheim & Klein, 1987), on the other.

Sequences in Language Learning

Interlanguage development is systematic, not haphazard. For a substantial number of language areas, learners are seen to traverse several stages, each consisting of predictable solutions, on their way to developing the various full-fledged subsystems of the target language. Patterned development through stages should not be equated with linear progression from inaccurate to accurate use of the L2 in a form-by-form or function-by-function, piecemeal fashion. For all the phenomena described in this section, learners undergo non-linear and unevenly paced increases and decreases in accuracy. For all phenomena, as well, some learners may never progress all the way to a full targetlike system represented in their mental grammars. In order to illustrate L2 sequences, we will examine findings for five interlanguage domains.

Systematicity in the development of accuracy: Morpheme orders

A robust case of systematicity in interlanguage, well known since the 1970s, pertains to a set of English inflectional morphemes found to be mastered by L2 English users in a certain order, shown in Table 6.1. This order represents the point at which learners across studies reached a conventional level of 80 or 90 percent accurate suppliance for each of the forms (see Krashen, 1977).

In Table 6.1, I have grouped morphemes into four sets to acknowledge the fact that, on occasion, some studies have reported slightly different ranks for structures within a given grouping. For example, Jia and Fuse (2007) reported that ten Mandarin Chinese-speaking children, who had arrived in the United States between the ages of 5 and 16 and whose L2 development was followed by the researchers for five years, found regular past tense -ed more challenging than third-person -s (possessive -'s was not investigated). Three of them were seen to master third-person -s to 80 percent accuracy after a year and a half or later in the United States, but none had yet mastered -ed at the end of the five-year study period. However, the rank of structures is never violated across the groups in Table 6.1 (see Goldschneider & DeKeyser, 2001, Appendix A). The accuracy order has been shown to be similar for both young and adult L2 learners, for both naturalistic and instructed learners, regardless of L1 background or whether the data are oral or written.

Table 6.1 Morpheme accuracy order, from earliest to latest mastery

Morpheme Illustration

-ing the girl is watching shop window with the food

Plural -s Chaplin give away a lot of cigars and chocolate to the kids

Be copula she is the one

Be auxiliary the girl is watching shop window with the food

a/the she steals a bread.. .he took the bread

Irregular past the police misunderstood

Regular past -ed she crashed with a man

third-person -s she steals a bread

Possessive -'s the bread shop's owner

Note. Illustrations are from L2 oral narratives produced by college-level learners of English in Japan after watching Alone and Hungry, a short video clip from Charlie Chaplin's Modern Times; unpublished author data © Ortega, Iwashita, Rabie, & Norris.

Larsen-Freeman (1976) identified the frequency and salience of these forms in the input as possible powerful determinants of the order, but it was not until 25 years later that Goldschneider and DeKeyser (2001) were finally able to garner concrete empirical evidence about this possibility. They established frequency benchmarks and operationalized salience via a scoring of four formal and functional dimensions. They then pooled the data for twelve morpheme studies and entered them into a regression analysis to see whether frequency plus the four dimensions of salience would shed light on the previously speculative explanations for the accuracy order shown in Table 6.1. They found that, indeed, each of the five factors on its own correlated reasonably well with the pooled morpheme data (in the range of r = 0.40 to 0.60), and that, when combined, they explained 71 percent of the variance in the data. Thus, we can conclude that the systematicity found in interlanguage development, at least for many aspects of morphology, reflects properties in the language input related to statistical distribution (frequency) and perceptual and functional cues (salience) as they interact with human cognitive capacities to learn from the linguistic evidence (N. Ellis, 2006, this volume).

Focus on complex form-function mappings: Sequenced development of temporal expression

Researchers have also uncovered systematicity in how learners come to express a given function in the L2 through the range of forms available. The expression of temporality has been well studied from this functional perspective, leading to the robust conclusion that learners of a second language (as well as children learning their first language) undergo three phases, each characterized by reliance on a different set of resources that help them express temporality: pragmatic, lexical, and grammatical.

In the initial phase, learners can mark tense only by means of pragmatic devices. For example, they stick to the chronological order in which things occur and use the sequential order of discourse to convey temporal relations among events, as illustrated in the following two contrasting examples from von Stutterheim and Klein (1987, p. 198):

(2) Schule fertig, Deutschland komm 'school finish, Germany come'

(3) Deutschland komm, Schule fertig 'Germany come, school finish'

Most interlocutors would naturally interpret (2) as meaning 'after finishing school I came to Germany' and (3) 'after I came to Germany I finished school.'

At a second stage, L2 learners begin expanding their repertoire and are able to recruit lexical devices in order to mark temporal relations more explicitly, as shown in example (4) from Schumann (1987, p. 25):

(4) and me come in '47 'and I came in 1947'

Lexical resources at this stage can help express a variety of temporal and aspectual meanings, such as calendric reference (in '47), anaphoric time (after, before), duration (two hours), frequency (always), and even two-point temporality (again) (Klein & Perdue, 1997).

It is only after this lexical stage that morphological forms may emerge in interlanguage to express grammatical notions related to time and temporality, such as grammatical tense and aspect. Not all forms emerge at once, and only one function or meaning is expressed by a given form initially (something to which we will return when we examine processes). Some grammatical meanings appear to be more basic or earlier acquired than others, for example, aspect before tense (Andersen, 1991), progressive before past (cf. Table 6.1), and perfective before imperfective (Andersen, 1991; cf. Table 6.2).

The patterned development of tense and aspect during this third phase, moreover, is guided by the inherent aspect or lexical semantics of each verb to which morphology is attached, reflecting an initial trajectory that relies on prototypical pairings of verb morphology with verb semantics. This is the prediction, in a nutshell, of the Aspect Hypothesis (see Andersen & Shirai, 1994). For example, the English imperfective marking -ing, which carries a prototypical durative meaning, emerges in interlanguage in combination with verbs depicting situations that imply duration, such as run, walk, or sing (such situations are called "activities" in linguistic theories of tense and aspect) and then spreads to other verbs with less prototypically durative meaning. Conversely, the English simple past marker -ed emerges first in combination with verbs that imply an action with duration and also a beginning and end, such as meet someone, catch something, see someone/ something (called "achievements"), and only later appears in combination with other verb semantics.

Table 6.2 summarizes the stages of development of perfective and imperfective aspect in L2 Spanish (expressed by the morphology of pretérito and imperfecto). As can be seen, development unfolds in a gradual form-function mapping process guided by prototypical pairings of verbal morphology with verb semantics. Similar stages of development have been found for other L2s that have the perfective/imperfective distinction, as well (e.g., French passé composé and imparfait; Italian passato prossimo and imperfetto; see review in Bardovi-Harlig, 2000). Although the tense/aspect systems of other languages may be rather different, similar prototype influences have been shown to shape the development of grammatical expression of temporality in those different cases too, such as L2 Japanese (Sugaya & Shirai, 2007) and L2 Korean (Lee & Kim, 2007).

More systematicity: Developmental sequences of negation

Not only morphology, but also syntax unfolds in predictable sequences in the production of L2 learners. One case is negation, which has been studied in L2

Table 6.2 Stages in the development of perfective (pretérito) and imperfective (imperfecto) aspect in L2 Spanish

Verb semantics (inherent lexical aspect)

Form- function development Stages Achievements +Punctuality +Telicity +Dynamicity Accomplishments -Punctuality +Telicity +Dynamicity Activities -Punctuality -Telicity +Dynamicity States -Punctuality -Telicity -Dynamicity

Emergence of one form in one context 1 PRETERIT

Preterit in achievements: por fin los dos líderes de la parroquia cambiaron su actitud hacia mí 'finally, the two leaders of the parish changed their attitudes towards me')

2 Preterit IMPERFECT

Imperfect in states: cuando era. pequeña ('when I was young')

Spread to additional contexts 3 Preterit PRETERIT IMPERFECT Imperfect

Imperfect in activities: me dolía la cabeza mucho por la altitud ('my head hurt a lot because of the altitude') Preterit in accomplishments: en las navidades pasadas vení a casa de mis padres ('last Christmas I comed to my parents' house')

4 Preterit Preterit IMPERFECT Imperfect Imperfect

Imperfect in accomplishments: cada navidad venia a casa de mis padres ('every Christmas I would come to my parents' house')

5 Preterit Preterit Imperfect Imperfect PRETERIT Imperfect

Preterit in activities: no sé por qué, pero ayer me dolió la cabeza toda la tarde ('I don't know why, but my head hurt all afternoon yesterday')

Full form- function mapping 6 Preterit IMPERFECT Preterit Imperfect Imperfect Preterit Imperfect PRETERIT

Preterit in states: aquel día .. .fuefatal ('that day . . . was terrible') Imperfect in achievements: se fue . . . porque no encontraba trabajo aquí en Dinamarca ('he left.. . because he couldn't find a job here in Denmark')

Note. Capitalized labels indicate first emergence of a form with a given semantic verb type. Illustrations show cutting edge of interlanguage (i.e., new attested form-function pairings) at each successive stage. All illustrations are from Cadierno (2000) and were produced in essays by ten advanced college-level learners of Spanish in Denmark, except for illustrations for stages 4 and 5, which are invented examples that have been added here.

English (Cancino, Rosansky, & Schumann, 1978; Stauble, 1978), as well as several other target languages. The four stages found for L2 English negation are summarized in Table 6.3.

These negation stages reflect internal grammar representations that learners build and gradually revise as they are better able to approximate the target system. Learners "outgrow" each stage as they develop. However, it is important to emphasize that each new stage represents a more advanced solution to the problem of how to negate in English, even though only the last stage results in a solution that converges with the target system. It is also important to appreciate that pre-verbal negation is the first stage not only for L1 Spanish learners whose L1 is consistent with that solution (no + verb) but also for other L2 learners whose L1, just like English, only allows post-verbal negation. Hyltenstam (1987) suggests that this first stage may be related to the fact that, across languages of the world, pre-verbal negation is a more common grammar configuration than postverbal negation. This is not to say, however, that the L1 does not play any role in sequenced development. Speakers of languages (e.g., Italian, Greek, Russian, and Spanish) where pre-verbal negation is the grammatical norm will remain in the first pre-verbal negation stage in English longer than, for example, L1 Norwegian or L1 Japanese speakers, whose L1s, just like the L2 in this case, require postverbal negation (Schumann, 1979; Zobl, 1982). In other words, L2 development may be slowed down by the influence of the L1, but not altered.

Table 6.3 Developmental stages for negation in L2 English

Stage Pattern Illustration

1 Pre-verbal negation with no/not No/Not + verb No saw him. 'I didn't see him.'

2 Pre-verbal negation with don't Don't + verb I don't saw him. 'I didn't see him.'

3 Postverbal negation in restricted contexts AUX + not/don't I will don't see you tomorrow. 'I will not see you tomorrow.'

4 Postverbal negation in all contexts They didn't see nobody. 'They didn't see anybody.'

Note. Illustrations are all reported in Stauble (1978), collected over a 10-month period in interviews with the same L2 speaker, 12-year-old Jorge from Colombia (who was one of the informants in Cancino et al., 1978). The verb which the negating functor modifies is underlined in each example.

Development of word order and questions

In the late 1970s, Meisel, Clahsen, and Pienemann (1981) uncovered another pattern in L2 development, this time for word order in L2 German, based on data from 45 migrant workers from Romance language backgrounds who were living in Germany. The stages are summarized in Table 6.4. Unlike the negation sequence in Table 6.3, which represents successive stages that learners gradually outgrow, the word order stages are cumulative. This means that each stage adds an important piece to the increasingly more complete repertoire of syntactic options, until the interlanguage system matches the full complexity of the repertoire available in the target grammar. Only stage 2 happens to lead to a solution that is inconsistent with the target input, i.e., that results in an ungrammatical word order, due to lack of inversion. This is because, while the context that calls for inversion emerges in stage 2, the actual syntactic operation of inversion will not be acquired (at stage 4) until learners also acquire the ability to handle verb separation in production, at stage 3. Learners who develop high levels of accuracy may eventually be able to shed the ungrammatical solution represented by stage 2, but not all learners will.

Table 6.4 The emergence of word order in L2 German according to Meisel, Clahsen, & Pienemann (1981)

Stage Description Illustration

1 Canonical word order Subject-Verb-X die Kinder spielen mim Ball 'the children play with the ball'

2 Adverb preposing X-Subject-Verb da Kinder spielen 'there children play'

3 Particle separation Verb . . . AUX/COMP or Particle alle Kinder muß die Pause machen 'all children must the break have'

4 Inversion X-Verb-Subject dann hat sie wieder die Knoch gebringt 'then has she again the bone bringed'

5 Verb-end Final position for verbs in subordinate clauses er sagte, daß er nach House kommt 'he said that he home comes'

Note. All illustrations are from Pienemann (1985) and were produced by elementary school children in Germany from L1 Italian background (except for the example in stage 5, which is invented by Pienemann). Close English translations, also taken from Pienemann, reflect German word order.

It should be remembered, moreover, that developmental sequences are about emergence and not accuracy. Thus, for example, a learner could apply the inversion rule to only one relevant case and miss its application to another 10 cases, and we would still consider him or her to be at stage 4, not 2. Indeed, learners may make use of the full repertoire of German word order once they reach stage 5, but they will do so with varying levels of accuracy in their production.

Initially, the developmental sequence for German word order was explained by Meisel et al. via two processing strategies hypothesized to be relevant: A canonical word order strategy (COS) and an initialization/finalization strategy (IFS). The COS strategy posited that producing canonical subject-verb-object word order (stage 1) is easier than producing more varied word orders. The IFS strategy was predicated on the assumption that initial and final components of strings are perceptually more salient to learners. Therefore, moving material to the initial (stage 2) or final (stage 3) position of a syntactic string is easier than moving material to (stage 4) or from (stage 5) positions inside the same string. The same COS and IFS rationale was later applied by Pienemann and colleagues in Australia to explain the developmental order of emergence of English question formation (Pienemann, Johntson, & Brindley, 1988). The stages are summarized in Table 6.5.

In the mid-1990s, Pienemann reconceptualized the explanations for the COS and IFS as part of a more precise and complex Processability Theory (Pienemann, 1998), which combined the linguistic framework of Lexical-Functional Grammar with an assumed corresponding psycholinguistic processing architecture in the mind of the learner. In a nutshell, the newly theorized explanation was that at beginning proficiencies, L2 learners are limited in their capacity for what syntactic information they can hold in memory during processing (hence the term "processability" in the name of the theory). They need gradually to develop the psycholinguistic capacity to match grammatical information contained within and across units in the linguistic material they encounter, and they are capable of doing so gradually with more distant elements in linguistic units. The Processability rationale has been successfully applied to a range of sequences across many target languages besides German and English (Pienemann, 2005).

Implicational, hierarchical acquisition of relative clauses

The final area of interlanguage systematicity we will examine is relativization. The L2 patterns uncovered to date in this aspect of grammar are related to wider patterns found across human languages affecting not only how different types of relative clauses emerge in L1 acquisition (Diessel & Tomasello, 2005), but also how frequent the various types are in L1 natural use, and the ease with which they are processed and comprehended by adult L1 users in experimental studies (e.g., Fox and Thompson, 2007; Reali & Christiansen, 2007). The explanations for these universal patterns are likely to include a number of syntactic, semantic-discoursal, cognitive, and statistical influences.

Table 6.5 The emergence of questions in L2 English according to Pienemann, Johnston, & Brindley (1988)

Stage Description

1 Words and fragments with rising intonation

2 Canonical word order with rising intonation

3 Fronting of a questioning element (wh-word, do, something else)

4 Inversion in two restricted contexts: (1) in w^-questions with copula, (2) in yes/no questions with auxiliaries other than do

6 Negative questions Question tags

Questions in embedded clauses

Illustration

One astronaut outside the space ship? A ball or a shoe?

He have two house in the front? Two children ride a bicycle? The boy threw a shoes? The dog don't have a spot?

Where the little children are?

What the boy is throwing?

What the boy with the black short throw?

Do the boy is beside the bus?

Do you have a shoes on your picture?

Is the boy is beside the bus?

Where is the sun?

Where is the space ship?

The ball is it in the grass or in the sky?

Is there a dog on the house?

Garbage, is it full?

Doesn't your wife speak English?

You live here, don't you?

Can you tell me where the station is?

Inversion expands to the full How many astronauts do you have? range of targetlike contexts What is the boy throwing?

Note. All illustrations are from Spada and Lightbown (1993 and 1999); questions for stages 1 through 5 were produced by francophone 10- to 12-year-olds in intensive English programs in Canadian schools during task-based oral interactions with a researcher. Questions for stage 6 are examples invented by the researchers.

In the L2, the most widely investigated influence stems from the typological concept of markedness (see Batistella, 1996, for a broad discussion), which was applied to relative clauses by Keenan and Comrie (1977) and resulted in their proposal of a Noun Phrase Accessibility Hierarchy that describes relativization options across all human languages. The hierarchy is illustrated with L2 data in Table 6.6. At the highest position of the hierarchy (that is to say, the most frequent and least marked) are subject relative clauses; at the lowest (i.e., the least frequent and most marked) are object of comparison relative clauses.

Table 6.6 Relative clauses in L2 German following Keenan & Comrie's (1977) Noun Phrase Accessibility Hierarchy

Clause type

L2 illustration

English equivalent

Subject Direct object

Indirect object

Object of preposition

Genitive

Object of comparison

Ich bin auf einen Baum geklettert, der neben das Ufer stand.

Du bist der Jungen, der ich in dem Sturm sah. [targetlike relativizer: den]

Ein funktionierender Rechtsstaat, offene Gesellschaft und die Achtung der menschenrechte sind die Anforderungen, den die Beitrittskandidaten entsprechen müssen. [targetlike relativizer: denen]

Es war ein fremdes Land, in dem die Leute eine fremde Sprache sprachen.

Und die Bäume hatten nur ein bisschen Blätte, deren Schatten am Boden spielten.

I climbed up a tree that stood next to the bank.

You are the boy that I saw in the storm.

A functioning constitutional state, an open society, and attention to human rights are the requirements that the candidate countries for accession to the European Union must meet. [German meet = entsprechen + dative]

It was a foreign land, in which people spoke a foreign language.

And the trees had only a few leaves, whose shadows played on the ground.

The job went to the driver who Peter is more qualified than.

Note. All L2 illustrations are from the longitudinal corpus of L2 writing produced by college-level German learners in the United States investigated by Byrnes & Sinicrope (2008); they appear here by courtesy of the authors. The English example for object of comparison is invented. na = not attested in the L2 data and not allowed in German L1

Cross-linguistically, these six possible types of relative clauses are in a marked-ness relationship which is hierarchical and implicational. That is, each lower (more marked) type is possible in a given language only if all other preceding (less marked) types are also possible. Some languages have only the simplest type of subject relativization at the top of the hierarchy (e.g., Maori and Tagalog) and, at the other extreme, some languages allow the full range of six types (e.g., Classical Arabic and English). Many languages have five of the six types (e.g., Indo-European languages, such as French, German, Italian, and Spanish, but also

genetically unrelated languages, such as Korean), missing the extremely rare object of a comparison type. However few or many possibilities for relativization a given language allows, it will do so following the pattern from highest to lowest in the hierarchy, without gaps.

The same implicationally related markedness hierarchy has been observed within learner grammars in studies that have examined not only L2 English (e.g., Pavesi, 1986), but target languages as varied as Chinese (Hu & Liu, 2007), German (Byrnes & Sinicrope, 2008), Italian (Croteau, 1995), Swedish (Hyltenstam, 1984), and several other L2s (Shirai & Ozeki, 2007; Tarallo & Myhill, 1983). The evidence is particularly robust for subject, object, and object of preposition types (evidence on the other three types is scarcer and more difficult to interpret, as noted by Tarallo & Myhill, 1983). Specifically, the ability to relativize in an L2 appears to unfold in the same order proposed in Keenan and Comrie's hierarchy and depicted in Table 6.6, based on the fact that when a given learner is able to produce more marked types (particularly the, in English, conspicuous object of a preposition), he or she will also likely be able to produce the highest (least marked) positions of subject and direct object, and not the other way around. As in all empirical research, of course, observations are never perfect, and most studies report a few learners whose production showed some gaps, particularly around the overall infrequent types of indirect object, genitive, and object of comparison. These gaps typically affect as low as 9-13 percent of the samples obtained with experimental elicitation procedures (e.g., Hyltenstam, 1984; Pavesi, 1986) or as high as 20 or 30 percent of learners in corpus data made up of extended discourse (Byrnes & Sinicrope, 2008). Not only the ability to produce a given type, but also the frequency and accuracy with which a given type is used, conform to observations in the L1 studies. Thus, types in the higher positions of the hierarchy are more frequent in L2 production, and also tend to contain fewer errors, whereas lower, more marked, types tend to induce more solutions that may be non-targetlike in the L2, such as resumptive pronouns (The teacher who you introduced me to her works for me now) (see Hu & Liu, 2007; and Hyltenstam, 1984, for further discussion).

Processes

Processes are the manifestation of putative mechanisms by which learners develop (or fail to develop) their internal grammars. They help characterize changes in accuracy levels and refinements in scope and degree of systematicity, as representations of the L2 grammar are assembled, expanded, revised, and elaborated. These processes attest to the importance of variability (both systematic and random) as a property of all natural languages that helps explain cycles of stability and instability in the dynamic self-organization of the language system as development proceeds (de Bot, Lowie, & Verspoor, 2007). For reasons of space, only a few central processes will be covered here: simplification, overgeneralization, restructuring, U-shaped behavior, and fossilization.

Simplification reflects a strategy that is called upon when messages, however simple or complex, must be conveyed with little language. But as Corder (1981) noted, "you cannot simplify what you do not possess" (p. 110). Thus, simplification may be a misnomer, as it results from incipient or incomplete learning of items (language forms or low-scope rules), or isolated instances that have not yet been fully represented and integrated into a broader system by the internal grammar.

Massive simplification (sometimes compared to pidginization, see Seigel, 2006) is seen during very early stages of L2 development, particularly among naturalistic learners who begin using vocabulary and discourse with little syntax and no morphology, drawing on only the most basic meaning-making principles, or what Klein and Perdue (1997) have called the Basic Variety. Later on, when (and if) complex syntax and some morphology emerge, simplification is also seen across all subsystems. Thus, even though a full range of formal choices is available in the morphology of the target language, a base (invariant) form tends to be chosen by learners at first; and even though multiple form-meaning mappings exist in the target language, a one-meaning-one-form mapping is initially represented in the learner grammar (Andersen, 1984a). For example, Anthony, a 12-year old L1 English learner of Spanish investigated by Andersen (1984b), used two invariant forms of the Spanish article, one devoted to mark definiteness (la, 'the' in feminine singular form) and one to mark indefiniteness (un, 'a' in masculine singular form), despite the full choice of eight forms that are available in the Spanish input. Which one meaning or form will be chosen out of those available may depend on the same input properties that explain other L2 phenomena: frequency and (perceptual and functional) salience. As an illustration, Sugaya and Shirai (2007) showed that even though the Japanese marker te i-ru can have a progressive (Ken-ga utat-te i-ru, "Ken is singing") and a resultative meaning (e.g., Booru-ga oti-te i-ru, "The ball has fallen"), L2 Japanese learners at first use it for the progressive meaning only. They argued that the progressive meaning is preferred because te i-ru is the only form that can convey action in progress in Japanese, whereas the resultative meaning (which in fact is more frequent in the input) can be also conveyed by several additional forms, such as ta and te-a-(ru). In this area of L2 Japanese, functional salience rather than frequency initially affects learners' developing grammars.

Overgeneralization involves the application of a form or rule not only to contexts where it applies in the target language, but also to others where it does not apply. Random or unsystematic overgeneralization, also called oversuppliance, does not appear to correspond to the systematic application of any pattern or logic. For example, naturalistic as well as instructed learners begin using -ing from very early on, but they also overgeneralize it to many nontargelike contexts, sometimes for substantial periods, and even when simultaneously they may fail to supply it in other contexts. This is shown in the following two contrasting cases attested by Pica (1985, p. 143) with L1 Spanish learners of English as a foreign language:

(5) I like to studying English.

(6) I was study languages last year.

However, even in cases of apparently random oversuppliance, an underlying systematicity may eventually be discovered if the analyst has rich enough data and searches deep enough. Certainly, it is the case that oversuppliance tends to be congruent with developmental constraints. For example, Camps (2002) found that second-semester college-level learners of L2 Spanish overgeneralized both the preterit and the imperfect. However, 39 percent of the time they overused the more stable preterit in contexts where the less consolidated, later emerging imperfect was required, whereas cases involving overgeneralization of the more recently acquired imperfect to preterit contexts were rare (only 4.5 percent).

Overgeneralization can also be systematic. An important case of systematic overgeneralization in morphology involves overregularization, or the attempt to make irregular forms fit regular patterns. It typically emerges after a certain level of development has been reached, in that it presupposes that learners have at least partially figured out some form or rule. The overuse of -ed with irregular verbs (as in falled, comed, goed, or even wented) is a well-known case of over-regularization (cf. example (1) at the beginning of this chapter, and Table 1.1; and see discussion in Clahsen, 2006, and Marslen-Wilson & Tyler, 1998, for L1; and in Leung, 2006, for L2). After systematically overgeneralizing, the learning task is to retreat from the overgeneralization and to adjust the application of the form or rule to increasingly more relevant contexts. This task calls upon the related processes of restructuring and U-shaped behavior, to which we turn.

Restructuring is the process of self-reorganization of grammar knowledge representations. During periods when restructuring of internal representations is happening, learners may seem to "backslide" and produce "errors" they did not seem to produce earlier, producing a pattern known as U-shaped behavior. Sharwood Smith and Kellerman (1989) define it as "the appearance of correct, or nativelike, forms at an early stage of development which then undergo a process of attrition, only to be reestablished at a later stage" (p. 220).

An oft-cited L2 illustration of successive restructuring accompanied by U-shaped behavior can be found in Huebner's (1983) study of Ge, a naturalistic learner of English as an L3 (he spoke Hmong as L1 and Lao as L2). Ge's development of the definite English article underwent several restructuring phases. Initially, he mostly marked nouns either with no articles or with da (his rendition of the) in what appears to have been a one-form-one meaning purpose for it: to encode the meaning of "assumed known to hearer" (which only partially overlaps with the notion of definiteness). Given that Ge's other languages did not have an article system, the internal grammar representation at this point may have simply been a unique rule that could be expressed as "in English, nouns must be marked as -/+ 'assumed known to hearer' with -/+ da." A month and a half into the study, this representation was destabilized, and Ge began using da to mark between

80 and 90 percent of all noun contexts he produced. Huebner called this extremely pervasive overgeneralization "flooding." This interlanguage solution may have been random, or it may have been motivated by a restructured rule like "nouns must always be marked by an article in English." In fact, this new "rule" can be considered a better (albeit overly general) approximation to the target rule, which says "all nouns must be marked by a three-way choice: zero article/ the/a." Be that as it may, the flooding of da to a majority of contexts naturally resulted in much higher levels of non-targetlike use of the definite article, leading to an appearance of regression or backsliding to Ge's English (the U-shaped behavior). A little over five months into the study, oversuppliance of da retreated from first-mention contexts (which are - 'assumed known to hearer' but specific; as in a woman is walking down the street), and shortly before reaching the seventh month, da began to retreat from even more non-targetlike indefinite contexts, giving way to a restructured rule that yielded stable targetlike suppliance for da at 80-90 percent levels for the rest of the observation period. Ge's development of the article da (the) shows that an item sometimes has to get worse before it can get better in language development. Or, as Kellerman (1985) put it, learners must go "beyond success" (an expression he borrowed from Annette Karmiloff-Smith) and restructure in what may appear to be a non-targetlike direction, before they can refine their representations of the L2.

Simplification, overgeneralization, restructuring, and other fundamental processes help learners move along the sequences. But there is no guarantee that the outcomes of these processes will keep propelling all learners toward convergence with the target system. Despite apparently favorable conditions for learning, many L2 users may stop anywhere along a given sequence of development, perhaps permanently. The term fossilization was coined by Selinker (1972) to refer to such cases of "premature cessation of development in defiance of optimal learning conditions" (Han, 2004, p. 23).

In a study that spanned seven years and focused on grammaticality judgments, as well as free writing, Han (2000, 2006) investigated whether the ability to use passives in English had fossilized in the grammar of Geng and Fong, two advanced users of English from Chinese L1 backgrounds. Over the seven-year period, both Geng and Fong consistently failed to supply passive in some cases where English requires it, as illustrated in (7), and sometimes oversupplied it in other contexts where the active voice would be pragmatically and discoursally preferred in English but where the passive voice was not ungrammatical, as shown in (8):

(7) I do not know whether these problems have solved in the newest release.

(written by Fong in 1996; Han, 2000, p. 89)

(8) What I can do for you is to give you a list of professors . . . The list will be

sent to you later.

(written by Fong in 1996; Han, 2000, p. 94)

In addition, both L2 users showed indeterminacy in their knowledge of English unaccusativity, on occasion ungrammatically overgeneralizing the passive voice to unaccusative verbs, as illustrated in (9), and other times using unaccusatives grammatically, as shown in (10):

(9) Thanks to John's blocking the event were stopped after 3/7/03. (written by Geng in 2003; Han, 2006, p. 69)

(10) The action already stopped on 1/6 probably after receiving our mail. (written by Geng in 2003; Han, 2006, p. 69)

The persistence of both the nativelike and non-nativelike solutions over the seven years is indeed suggestive of permanent cessation of learning in this one area of the L2. Surprisingly, Geng and Fong were otherwise extremely advanced learners under optimal learning circumstances. They had had formal English instruction for six years in their home country, China, and had scored over 600 on the TOEFL before moving to the United Kingdom to obtain their doctoral degrees. Upon receipt of their degrees each continued living in English-speaking environments, and both actively published in English in international journals in their fields. Han suggested Gong and Feng's apparent case of fossilization was caused by the subtle influence of the L1, which unlike English lacks passive morphology and is a topic-prominent language. On the other hand, reviewing this and other studies, Long (2003) proposed that sensitivity to the input (or lack thereof) may be a better explanation for fossilization in general.

In the end, the notion of fossilization, despite its immense popularity, has proved to be extremely problematic. Long (2003) critically examined the reasons for this, only some of which can be mentioned here. First, complete and permanent cessation of learning is difficult to demonstrate empirically, unless learners are followed longitudinally and over their life time, or at least over a very long period. Yet, most fossilization studies have only documented "lack of" learning for short periods of time, rarely more than two or three years. Notable exceptions are multi-year longitudinal studies of Patty by Lardiere (2007), Ayako by Long (2003), and the one just discussed of Fong and Geng by Han (2006). Second, many so-called fossilized learners investigated in the literature may not have enjoyed optimal learning conditions, missing out in one or more of three areas: sufficient exposure and practice, positive attitudes toward the target language and society, and the aid of (high quality) instruction. Third, there is conceptual lack of clarity with regard to at least two questions: (1) Is fossilization meant to denote a process or mechanism that causes cessation of learning, or is it simply being used to denote lack of success in ultimate attainment, that is, the product or consequence of some other mechanism? (2) Where should the evidence for cessation of learning be sought? In the absence of any change (measuring stability versus systematic change), or in what Long calls volatility, or a stable mixture of systematic plus random variability that never restructures into any systematic new representation?

Another issue that muddies the waters of fossilization can be added here. Namely, some researchers discuss fossilization as a universal process (or product) that allegedly characterizes all L2 learners by definition, be it as part of the discussion of the availability of Universal Grammar (Lardiere, 2007) or as related to the entrenchment of L1 knowledge (Han, 2006). From either research perspective, all learners will inevitably fossilize sooner or later, and fossilization may ultimately mean that so-called nativeness is unattainable in an L2 (a tautology of doubtful value for researchers who have begun to examine L2 learning through a bilingual prism; see Ortega, 2007). Other times, however, fossilization is discussed as a process or product that helps explain individual differences in ultimate attainment, as a unique tendency for some particularly unsuccessful learners to "stop" learning much sooner than others, as was the oft-discussed case of Alberto, reported by Schumann (1978). This is perhaps the notion of fossilization that concerns most language teachers, but it seems to be the one gradually being dispreferred by most contemporary research on fossilization. In fact, we may want to avoid speaking of "fossilized learners" or even fossilized grammars. Until our understanding of cessation of learning phenomena is clearer, it may be better to speak of stabilization and to consider this process as affecting only local areas of grammar (Han, 2004; Long, 2003).

Sequences, Processes, and Instruction: Five Generalizations and a Coda

I conclude this chapter with five generalizations about the relationship between interlanguage development and instruction that summarize the implications of the developmental sequences and processes for second and foreign language teaching.

1 Instruction cannot affect the route of L2 development in any fundamental way. Several studies have shown that students in classrooms proceed along the developmental sequences regardless of the order in which they are taught. For example, German textbooks typically introduce the inversion rule before they feature the verb separation rule (cf. Table 6.4), perhaps on the basis of a data-free intuition of most textbook writers that verb separation is logically more "complex" than subject-verb inversion. Yet, R. Ellis (1989) found that the order in which 39 students of German in the United Kingdom learned the word order fitted the developmental sequence in Table 6.4 and was unaffected by the textbook order followed by their teachers. Likewise, Bardovi-Harlig (1995) reported no relationship between when the past perfect (John entered college in 1980. He had graduated from high school five years earlier) was taught to 16 college learners of English by their teachers and when it emerged in these students' L2 writing. Rather than instruction, two developmental prerequisites were the predictors of emergence of past perfect: (1) learners first had to reach stable accuracy of about 80 percent in the suppliance of -ed; and (2) they then had to begin creating contexts

involving reverse-order reports (cf. the example above) that semantically call for the use of the past perfect. Over a period of 15 months, 10 of the 16 learners met these two prerequisites and, regardless of instruction, began using the new form.

2 Instruction can have some effect on processes, fostering some and inhibiting others. It has been shown that overgeneralization is fostered and, conversely, simplification inhibited, by instruction. Among the earliest studies to document this are Lightbown (1983) and Pica (1985). Particularly in classrooms where teachers teach grammar explicitly and have students practice the language in mechanical drills, oversuppliance of morphology is to be expected. For example, Camps (2002) reported that the 15 second-semester Spanish students he investigated suddenly oversupplied subjunctive in about 20-30 percent of contexts that required preterit and imperfect, coinciding with the point toward the end of the semester at which teachers taught this form to their students. Temporary overgeneralization, however, does not need to be interpreted negatively. As Long (1988) suggested, high levels of overgeneralization coupled with low levels of simplification may be indicative of "healthier error profiles" (p. 120), in that overgeneralization (particularly if aided by development-sensitive instruction) may give way to restructuring and, ultimately, to more development.

3 Instruction can be ineffective and even counterproductive when it ignores developmental readiness. In two different studies involving ten 7- to 9-year-old children and three adult classroom learners, respectively, Pienemann (1984, 1989) discovered that the teaching of stage 4 (inversion) benefited only those learners who were developmentally ready, that is, at stage 3 of verb separation. For some of the learners who were not ready, nothing was gained and, moreover, an undesirable effect of instruction was that they began avoiding the use of adverb preposing (stage 2), presumably to avoid making errors. Other findings by Mackey (1999) and Spada and Lightbown (1999) have suggested that some developmentally unready learners may accrue benefits from instruction and advance to the next stage prior to the one taught, but that they will not skip stages. Thus, teachers cannot teach everything they want and instead should carefully consider what their students are developmentally ready to learn. Premature instruction may be counterproductive if it discourages learners from taking risks, which may delay development.

4 Not all sequences present equal challenges for instruction. On the other hand, it is important to recognize that the systematicity of interlanguage development is of different kinds for different areas of language. More specifically, not all developmental sequences are equally open or impervious to instruction in similar ways. For areas of language where accuracy improvements have been shown to pose differential difficulty for learners (such as the morpheme order in Table 6.1), instruction may be possible at any level. For areas where development through stages means that an increasingly more complex option in a repertoire is cumulatively added (such as the sequences in Tables 6.2, 6.4, and 6.5), learners may not skip stages and teachers will do well to consider developmental readiness. Finally, for areas of the grammar that exhibit cross-linguistic markedness relations (such as the hierarchy in Table 6.6), instruction of a more marked case

can help learners get the less marked cases for free, as it were. This has been shown in a good number of studies that successfully targeted the teaching of object-of-preposition relative clauses to students who were already able to handle relativization, but only of the subject type, and who after instruction showed gains not only in the taught type, but also all the intervening types in the hierarchy (e.g., Doughty, 1991).

5 Instruction has large positive effects on rate of development and level of ultimate attainment. It would be mistaken to conclude that instruction does not matter, just because it cannot override development. Instructed learners progress at a faster rate, they are likely to progress further along the sequences, and they typically become more accurate overall than uninstructed learners. For example, after daily use of English, but without specific instruction, many naturalistic learners do not produce -ed, or produce it with extremely low levels of accuracy (cf. Table 6.1). This has been shown even with learners who had the advantage of a relatively early start when they were first surrounded by the L2, as was the case for the two 10- and 12-year-old boys studied by Sato (1990) for ten months or the ten young children and adolescents studied by Jia and Fuse (2007) for five years. By comparison, Bardovi-Harlig (1995) found that 135 instructed English learners in a college-level intensive English program exhibited levels of accuracy in the use of -ed that averaged about 70 percent for the lowest curricular level and 90 percent for the highest. Likewise, many naturalistic L2 German learners may not reach the particle-separation stage (cf. stage 3 in Table 6.4) even after several years of living in the L2 environment (Meisel et al., 1981). By contrast, in the foreign language classroom, findings by R. Ellis (1989) and Jansen (2008) suggest that most students will have reached that stage (and some may even have traversed the entire developmental sequence!) by the end of the second semester. For relativization, too, Pavesi (1986) found that only about a fourth of 38 naturalistic learners with six years on average of living in the L2 environment had reached object of preposition (cf. Table 6.6), whereas the same stage had been reached by about 40 percent of 48 high school students in Italy with an average of four years of instruction in English as a foreign language. Byrnes and Sinicrope (2008) found evidence of the object of preposition stage at just the end of the second year of study for about a fourth of 23 college students of German in the United States.

In light of these five generalizations, language teachers may ask themselves whether they should teach to the sequences and processes. It would be unadvisable to develop instructional curricula around the known sequences and processes of L2 development, for several reasons. First, we do not have sufficient descriptions of all aspects of the grammar of any target language to do so. Furthermore, the holistic question of how different sequences relate to each other in the grammar of individual learners has rarely been examined, giving textbook writers and curriculum developers little guidance as to how to sequence grammatical targets according to developmental learner readiness principles. More importantly, language learning amounts to much more than the learning of syntax and morphology. It also involves the learning of vocabulary, pragmatics, phonology,

and so on. Although much is also known about how these areas are learned by L2 users, it would be difficult to exploit all this knowledge in a syllabus or curricular plan. The most fundamental objection, however, is that learning an L2 calls for a much more encompassing approach than a focus on bits and pieces of language could possibly afford us. Rather than trying to organize instruction around grammar in what Wilkins (1976) called a synthetic syllabus, we have a wide range of options more attuned to what we know about psycholinguistic, cognitive, and socioeducational principles for good language teaching. Many of these options are discussed in sections IV through VI in this handbook.

Nevertheless, knowledge about the sequences and processes of interlanguage development can inform good teaching by helping teachers (and their students) cultivate a different attitude toward "errors," and more enlightened expectations for "progress." It can help them recognize that many so-called errors are a healthy sign of learning, that timing is hugely important in language teaching, and that not all that can be logically taught can be learned if learners are not developmentally ready. Knowledge about sequences and processes can also help counter the deficit view that interlanguages are defective surrogates of the target language by making it clear that interlanguages are shaped by the same systematicity and variability that shape all other forms of human language.

This chapter draws on material from Ortega (2009).

REFERENCES

Andersen, R. W. (1984a). The one-to-one principle of interlanguage construction. Language Learning 34, 77-95.

Andersen, R. W. (1984b). Whats gender good for, anyway? In R. W. Andersen (ed.), Second languages: A cross-linguistic perspective (pp. 77-100). Rowley, MA: Newbury House.

Andersen, R. W. (1991). Developmental sequences: The emergence of aspect marking in second language acquisition. In T. Huebner & C. A. Ferguson (eds.), Crosscurrents in language acquisition research and linguistic theories (pp. 305-23). Amsterdam: John Bejamins.

Andersen, R. W. & Shirai, Y. (1994). Discourse motivations for some cognitive acquisition principles. Studies in Second Language Acquisition 16, 133-56.

Bardovi-Harlig, K. (1995). The interaction of pedagogy and natural sequences in the acquisition of tense and aspect. In F. Eckman, D. Highland, P. Lee, J. Mileham, & R. Weber (eds.), Second language acquisition theory and pedagogy (pp. 151-68). Mahwah, NJ: Lawrence Erlbaum.

Bardovi-Harlig, K. (2000). Tense and aspect in second language acquisition: Form, meaning, and use. Malden, MA: Blackwell.

Batistella, E. L. (1996). The logic of markedness. New York: Oxford University Press.

Byrnes, H. & Sinicrope, C. (2008). Advancedness and the development of relativization in L2 German: A curriculum-based longitudinal study. In L. Ortega & H. Byrnes (eds.), The longitudinal study of advanced L2 capacities (pp. 109-38). New York: Lawrence Erlbaum Taylor & Francis.

Cadierno, T. (2000). The acquisition of Spanish grammatical aspect by Danish advanced language learners. Spanish Applied Linguistics 4, 1-53.

Camps, J. (2002). Aspectual distinctions in Spanish as a foreign language: The early stages of oral production. International Review of Applied Linguistics 40, 179-210.

Cancino, H., Rosansky, E., & Schumann, J. (1978). The acquisition of English negatives and interrogatives by native Spanish speakers. In E. Hatch (ed.), Second language acquisition: A book of readings (pp. 207-30). Rowley, MA: Newbury House.

Clahsen, H. (2006). Dual-mechanism morphology. In Keith Brown (ed.), Encyclopedia of language and linguistics, vol. 4. Oxford: Elsevier.

Corder, P. (1981). Error analysis and interlanguage. Oxford: Oxford University Press.

Croteau, K. C. (1995). Second language acquisition of relative clause structures by learners of Italian. In F. R. Eckman, D. Highland, P. W. Lee, J. Mileham, & R. R. Weber (eds.), Second language acquisition theory and pedagogy (pp. 115-28). Mahwah, NJ: Erlbaum.

de Bot, K., Lowie, W., & Verspoor, M. (2007). A Dynamic Systems Theory approach to second language acquisition. Bilingualism: Language and Cognition 10, 7-21.

Diessel, H. & Tomasello, M. (2005). A new look at the acquisition of relative clauses. Language 81, 882-906.

Doughty, C. (1991). Second language instruction does make a difference: Evidence from an empirical study of ESL relativization. Studies in Second Language Acquisition 13, 431-69.

Ellis, N. C. (2006). Language acquisition as rational contingency learning. Applied Linguistics 27, 1-24.

Ellis, R. (1989). Are classroom and naturalistic acquisition the same? A study of the classroom acquisition of German word order rules. Studies in Second Language Acquisition 11, 305-28.

Fox, B. A. & Thompson, S. A. (2007). Relative clauses in English conversation: Relativizers, frequency, and the notion of construction. Studies in Language 31, 293-326.

Goldschneider, J. & DeKeyser, R. M. (2001). Explaining the "natural order of L2 morpheme acquisition" in English: A meta-analysis of multiple determinants. Language Learning 51, 1-50.

Han, Z. (2000). Persistence of the implicit influence of NL: The case of the pseudo-passive. Applied Linguistics 21, 78-105.

Han, Z. (2004). Fossilization in adult second language acquisition. Clevedon, UK: Multilingual Matters.

Han, Z. (2006). Fossilization: Can grammaticality judgment be a reliable source of evidence? In Z. Han & T. Odlin (eds.), Studies of fossilization in second language acquisition (pp. 56-82). Clevedon, UK: Multilingual Matters.

Hu, X. & Liu, C. (2007). Restrictive relative clauses in English and Korean learners' second language Chinese. Second Language Research 23, 263-87.

Huebner, T. (1983). A longitudinal analysis of the acquisition of English. Ann Arbor, MI: Karoma.

Hyltenstam, K. (1984). The use of typological markedness conditions as predictors in second language acquisition: The case of pronominal copies in relative clauses. In R. W. Andersen (ed.), Second languages: A cross-linguistic perspective (pp. 39-58). Rowley, MA: Newbury House.

Hyltenstam, K. (1987). Markedness, language universals, language typology, and second language acquisition. In C. Pfaff (ed.), First and second language acquisition processes (pp. 55-78). Cambridge, MA: Newbury House.

Jansen, L. (2008). Acquisition of German word order in tutored learners: A cross-sectional study in a wider theoretical context. Language Learning 58, 185-231.

Jia, G. & Fuse, A. (2007). Acquisition of English grammatical morphology by native Mandarin-speaking Children and adolescents: Age-related differences. Journal of Speech, Language & Hearing Research 50, 1280-99.

Keenan, E. & Comrie, B. (1977). Noun phrase acessibility and universal grammar. Linguistic Inquiry 8, 63-99.

Kellerman, E. (1985). If at first you do succeed. In S. M. Gass & C. Madden (eds.), Input in second language acquisition (pp. 345-53). Rowley, MA: Newbury House.

Klein, W. & Perdue, C. (1997). The basic variety (or: Couldn't natural languages be much simpler?). Second Language Research 14, 301-47.

Krashen, D. (1977). Some issues relating to the Monitor Model. In H. D. Brown, C. A. Yorio, & R. H. Crymes (eds.), On TESOL '77. Teaching and learning English as a second language: Trends in research and practice (pp. 144-58). Washington, DC: TESOL.

Lardiere, D. (2007). Ultimate attainment in second language acquisition: A case study. Mahwah, NJ: Lawrence Erlbaum.

Larsen-Freeman, D. (1976). An explanation for the morpheme acquisition order of second language learners. Language Learning 26, 125-34.

Lee, E. & Kim, H.-Y. (2007). On crosslinguistic variations in imperfective aspect: The case of L2 Korean. Language Learning 57, 651-85.

Leung, Y.-K. I. (2006). Verb morphology in L2A vs. L3A: The representation of regular and irregular past participles in English-Spanish and Chinese-English-Spanish inter-languages. In S. Foster-Cohen, M. M. Krajnovic, & J. M. Djigunovic (eds.), EUROSLA Yearbook 6 (pp. 27-56). Amsterdam: John Benjamins.

Lightbown, P. M. (1983). Exploring the relationships between developmental and instructional sequences in L2 acquisition. In H. W. Seliger & M. H. Long (eds.), Classroom oriented research in second language acquisition (pp. 217-45). Rowley, MA: Newbury House.

Long, M. H. (1988). Instructed interlanguage development. In L. Beebe (ed.), Issues in second language acquisition: Multiple perspectives (pp. 115-41). Rowley, MA: Newbury House.

Long, M. H. (2003). Stabilization and fossilization in interlanguage development. In C. J. Doughty & M. H. Long (eds.), Handbook of second language acquisition (pp. 487535). Malden, MA: Blackwell.

Long, M. H. & Sato, C. J. (1984). Methodological issues in interlanguage studies: An interactionist perspective. In A. Davies, C. Criper, & A. Howatt (eds.), Interlanguage (pp. 253-79). Edinburgh: Edinburgh University Press.

Mackey, A. (1999). Input, interaction, and second language development: An empirical study of question formation in ESL. Studies in Second Language Acquisition 21, 557-87.

Marslen-Wilson, W. & Tyler, L. K. (1998). Rules, representations, and the English past tense. TRENDS in Cognitive Sciences 2, 428-35.

Mayberry, R. I. (in press). Second language learning of sign languages. In Keith Brown (ed.), Encyclopedia of language and linguistics, 2nd edn., vol. 10. Oxford: Elsevier.

Meisel, J., Clahsen, H., & Pienemann, M. (1981). On determining developmental stages in natural second language acquisition. Studies in Second Language Acquisition 3, 109-35.

Ortega, L. (2007). Second language learning explained? SLA across nine contemporary theories. In B. VanPatten & J. Williams (eds.), Theories in second language acquisition: An introduction (pp. 221-46). Mahwah, NJ: Lawrence Erlbaum.

Ortega, L. (2009). Understanding second language acquisition. London: Hodder Arnold.

Oshita, H. (2000). What is happened may not be what appears to be happening: A corpus study of "passive" unaccusatives in L2 English. Second Language Research 16, 293-324.

Pavesi, M. (1986). Markedness, discoursal modes, and relative clause formation in a formal and an informal context. Studies in Second Language Acquisition 8, 38-55.

Pica, T. (1985). Linguistic simplicity and learnability: Implications for language syllabus design. In K. Hyltenstam & M. Pienemann (eds.), Modelling and assessing second language acquisition (pp. 137-51). Philadelphia: Multilingual Matters.

Pienemann, M. (1984). Psychological constraints on the teachability of languages. Studies in Second Language Acquisition 6, 186-214.

Pienemann, M. (1985). Learnability and syllabus construction. In K. Hyltenstam & M. Pienemann (eds.), Modelling and assessing second language acquisition (pp. 23-76). Philadelphia: Multilingual Matters.

Pienemann, M. (1989). Is language teachable? Psycholinguistic experiments and hypotheses. Applied Linguistics 10, 52-79.

Pienemann, M. (1998). Language processing and second language development: Processability Theory. Amsterdam/Philadelphia: John Benjamins.

Pienemann, M. (ed.) (2005). Cross-linguistic aspects of Processability Theory. Amsterdam: John Benjamins.

Pienemann, M., Johnston, M., & Brindley, G. (1988). Constructing an acquisition-based procedure for second language assessment. Studies in Second Language Acquisition 10, 217-43.

Reali, F. & Christiansen, M. H. (2007). Processing of relative clauses is made easier by frequency of occurrence. Journal of Memory and Language 57, 1-23.

Sato, C. (1990). The syntax of conversation in interlanguage development. Tübingen: Gunter Narr.

Schachter, J. (1974). An error in error analysis. Language Learning 24, 205-14.

Schumann, J. (1978). Second language acquisition: The pidginization hypothesis. In E. Hatch (ed.), Second language acquisition: A book of readings (pp. 256-71). Rowley, MA: Newbury House.

Schumann, J. (1979). The acquisition of English negation by speakers of Spanish: A review of the literature. In R. W. Andersen (ed.), The acquisition and use of Spanish and English as first and second languages (pp. 3-32). Washington, DC: TESOL.

Schumann, J. (1987). The expression of temporality in basilang speech. Studies in Second Language Acquisition 9, 21-41.

Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics 10, 219-31.

Sharwood Smith, M. & Kellerman, E. (1989). The interpretation of second language output. In H. W. Dechert & M. Raupach (eds.), Transfer in language production (pp. 217-36). Norwood, NJ: Ablex.

Shirai, Y. & Ozeki, H. (2007). The L2 acquisition of relative clauses in East Asian languages. Special issue of Studies in Second Language Acquisition 29, 155-374.

Siegel, J. (2006). Links between SLA and Creole studies: Past and present. In C. Lefebvre, L. White, & C. Jourdan (eds.), L2 acquisition and creole genesis (pp. 15-46). Amsterdam: John Benjamins.

Spada, N. & Lightbown, P. M. (1993). Instruction and the development of questions in the L2 classroom. Studies in Second Language Acquisition 15, 205-21.

Spada, N. & Lightbown, P. M. (1999). Instruction, first language influence, and developmental readiness in second language acquisition. The Modern Language Journal 83, 1-22.

Stauble, A. E. (1978). The process of decreolization: A model for second language development. Language Learning 28, 29-54.

Stockwell, R., Bowen, J., & Martin, J. (1965). The grammatical structures of English and Spanish. Chicago: University of Chicago Press.

Sugaya, N. & Shirai, Y. (2007). The acquisition of progressive and resultative meanings of the imperfective aspect marker by L2 learners of Japanese: Trasfer, universals, or multiple factors. Studies in Second Language Acquisition 29, 1-38.

Tarallo, F. & Myhill, J. (1983). Interference and natural language processing in second language acquisition. Language Learning 33, 55-73.

von Stutterheim, C. & Klein, W. (1987). A concept-oriented approach to second language studies. In C. W. Pfaff (ed.), First and second language acquisition (pp. 191-205). Rowley, MA: Newbury House.

White, L. (2003). Second language acquisition and Universal Grammar. New York: Cambridge University Press.

Wilkins, D. A. (1976). Notional syllabuses. Oxford: Oxford University Press.

Zobl, H. (1982). A direction for contrastive analysis: The comparative study of developmental sequences. TESOL Quarterly 16, 169-83.

7 The Importance of Cross-Linguistic Similarity in Foreign Language Learning

hAkan ringbom and

SCOTT JARVIS

Second language research has tended to concentrate on differences, as they are manifested in linguistic variation of numerous kinds, rather than on similarities. To the learner, however, similarities have a much more direct effect on language learning and performance than differences do. Learners are constantly trying to establish links between the TL (target language) and whatever prior linguistic knowledge they have. Instead of seeking out differences, they tend to look for similarities wherever they can find them. They make use of intra-lingual similarities, which are perceived from what they have already learnt of the target language. However, at early stages of learning, when their TL knowledge is limited, the L1 is generally the main source for perceiving linguistic similarities, though other known languages, especially if they are related to the target language, and if they have been acquired to a high level of proficiency, may also have an important part to play. Perceiving and making use of cross-linguistic similarities to prior knowledge is important in the learner's striving to facilitate the learning task, and these are processes central to transfer. Some of the most influential work in this area includes, in chronological order, Selinker (1969), Kellerman (1977), Gass and Selinker (1983), Kellerman and Sharwood Smith (1986), Ringbom (1987), Dechert and Raupach (1989), Odlin (1989), Kellerman (1995), Jarvis (2000), and recently Odlin (2003). Jarvis and Pavlenko (2008) is a survey of cross-linguistic influence, especially of the work that has been done in the last 20 years.

Actual, Perceived, and Assumed Similarities

Perceiving cross-linguistic similarities is of course a subjective process that often results in an inaccurate or incomplete awareness of the actual similarities that exist across languages. Accordingly, it is important to distinguish between actual similarities and assumed similarities. These two types of similarity, in fact, relate to

The Handbook of Language Teaching Editedby Michael H.Longand Catherin eJ.Doughty © 2009 BkcWvell Publishin gLtd. ISBN: 978-1-405-15489-5

different levels. Actual similarities or differences belong to the domain of linguistics and can accordingly be analyzed linguistically, whereas what the learner does with or assumes about the TL relates to the processes taking place in the learner's mind. Actual and assumed similarities can hypothetically be fully congruous if the learner accurately perceives the objective similarities between two languages, but this appears to be relatively rare. Indeed, in the foreign-language learning setting, the disparity between actual and assumed similarities can be great, and this results from (1) learners' failure to notice a number of the actual similarities that exist across languages, (2) learners' misperception of the nature of many of the similarities that they do notice, and (3) learners' assumptions that there exist certain similarities between the languages that actually do not exist and which the learners have correspondingly never previously encountered (see Jarvis & Pavlenko, 2008; Ringbom, 2006; 2007).

Attempts have been made to establish criteria for actual similarities. These attempts include the work of Ellegard (1978) and Ard and Homburg (1992). There is also an ongoing project in the Research Institute at the University of Groningen that involves the use of computational methods for determining cross-linguistic similarity (see www.rug.nl/let/onderzoek/). So far, however, there is still little agreement in the field concerning how to define or measure the actual, objective similarities existing across languages.

Even if there were complete agreement concerning how to measure the degree of actual similarity between languages, the literature on foreign and second language learning, as well as the literature on cross-linguistic influence, make it clear that assumed similarities have a greater and more direct effect on language learning and performance than actual similarities do (Kellerman, 1978; Odlin, 1989, p. 142; 2006, pp. 23-4). Although actual similarities do seem to account quite well for learners' rate of acquisition and the amount of time they need to achieve certain levels of proficiency in the target language (Ard & Homburg, 1992; Odlin, 1989; Ringbom, 1987, p. 66), there are a number of constraints on the effects that actual similarities have on foreign-language learning, and these constraints appear to be related primarily to learners' perceptions and assumptions. For example, Kellerman (1978) has shown that learners rely on only certain types of actual similarities and not on others, depending on how language-specific they consider the language features in question to be. Ringbom (2007) has additionally explained that actual similarities are symmetrical across languages (applying equally from Language A to Language B and vice versa), whereas assumed similarities may have a stronger effect in one direction than in the other. Finally, actual similarities are constant over time, whereas the perception of similarities changes as the learner's TL experience and proficiency increase (Kellerman, 1979).

Although assumed similarities have the strongest and most direct impact on language learning and performance, there are different types of assumed similarities, and they work differently in comprehension and production. In comprehension, especially of a related language, learners directly perceive similarities, by which we mean that learners encounter and become aware of features of the

TL that they recognize as bearing resemblances to a language they already know. These similarities are normally formal similarities - i.e., similarities pertaining to the spelling, pronunciation, and/or morphological make-up of words or multiword structures. When producing the new language, on the other hand, learners are not engaged in perceiving similarities but rather in encoding their ideas into language structures they have previously learned or into language structures that they create in the absence of learned knowledge, which tend to be based heavily on the similarities they assume to exist between the two languages. These assumed similarities may be formed on the basis of previously perceived similarities in comprehension, but if such similarities have not been perceived, as is usually the case across wholly unrelated languages, where few formal similarities can be established, learners only assume that the new language works in the same way as their L1, and this tends to lead to errors. In cases where learners are learning a third or fourth language, their assumption tends to be that the TL is semantically and pragmatically similar to the L1 but formally similar to the language (the L1, L2, L3, etc.) that they perceive as being typologically closest to the TL (Ringbom, 1987; 2007). (Note: Several studies have noted a foreign-language effect in L3 acquisition, whereby under certain circumstances the L2 may exert a greater influence on L3 production than the L1 does. This is especially true if the L2 is typologically similar to the L3, and factors such as proficiency and order of acquisition also appear to have an effect. See De Angelis, 2005; Dewaele, 1998; Singleton, 2006, pp. 136-8; Williams & Hammarberg, 1998.)

Learners have a strong tendency to assume semantic and pragmatic similarities between the L1 and TL without having ever perceived those similarities, and this is true regardless of how typologically distant the two languages are (e.g., Biskup, 1992; Eisenstein & Bodman, 1993; Jarvis, 1998; Ringbom, 1987). Concerning formal similarities, on the other hand, learners tend not to assume that the formal properties of words and multi-word structures are similar until they have actually perceived those similarities, although, crucially, this depends largely on the typological distance between the two languages and also on the learners' levels of proficiency. In typologically similar languages, there are more similarities to perceive, and when learners have crossed a certain threshold of perceived similarities, they often assume additional formal similarities that they have never perceived.

Learners of English in Finland have been found to cross this threshold in relation to the similarities between Swedish and English, related languages, whereas they do not cross this threshold in relation to the similarities between Finnish and English, which are typologically unrelated languages. The actual formal similarities between Swedish and English are numerous and cut across all word classes (e.g., Sw. arm = 'arm'; Sw. fot = 'foot'; Sw. tanka = 'think'; Sw. ata = 'eat', Sw. rod = 'red'; Sw. bla = 'blue'; Sw. vad = 'what'; Sw. fran = 'from'; Sw. vi = 'we') and they also cut across numerous inflectional and syntactic patterns (Sw. den bla bilen = 'the blue car'; Sw. Henriks hus ar litet = 'Henrik's house is little'). Concerning the similarities between Finnish and English, a number of actual formal similarities can be found here, as well, but most of these involve only

medium- to low-frequency loanword nouns that represent entities and notions that Finnish has adopted from other languages, primarily from Swedish and English (e.g., Fi. auto = 'automobile'; Fi. radio = 'radio'; Fi. televisio = 'television'; Fi. filmi = 'film'; Fi. presidentti = 'president'). Aside from words of this type, Finnish shares very few grammaticalization and lexicalization patterns with English. (See, however, Seppanen, 1998, who lists seven grammatical correspondences between Finnish and English, many of which also occur between Finnish and other Germanic languages. Cf. Ringbom, 2007.)

As mentioned, learners' perceptions of the similarities between Swedish and English cross a crucial similarity threshold (cf. Eckman, 2004; Wode, 1976), whereas their perceptions of the similarities between Finnish and English do not. The consequences of this threshold are clearest in the English production of learners who know both Swedish and Finnish. Learners whose L1 is Finnish and L2 is Swedish, and who are at a low level of English proficiency, have been found to be especially prone to over-assuming the similarities between Swedish and English, producing errors such as the following (errors in the Finnish National Matriculation Exam): A teacher is a forebild for pupils (Sw. forebild = 'model', 'good example'); He is good at mathematics but he success at other amnys, too (Sw. amne = 'subject'). They rarely produce corresponding errors based on an over-assumption of formal similarities between Finnish and English. (Note: Anglophone learners of Greek have likewise been found to disregard the similarities between English and Greek due to the perceived distance between the languages. See Singleton, 2006, p. 139.) Like the Finns, Swedes also sometimes over-rely on assumed similarities between Swedish and English, even on chance similarities: A crucial problem is the drinking of alcohol before sitting behind the rat (Sw. ratt = 'steering wheel'). Mostly, however, Swedish learners at the intermediate and advanced levels of English proficiency tend to over-assume formal similarity in low-frequency words, which may often result in the use of an existing Swedish word instead of a formally similar English one: I am not an eremit (Sw. eremit = 'hermit'; A new house made of marmor (Sw. marmor = 'marble'). (For statistics and examples, see Ringbom, 1987.)

Types of Cross-Linguistic Similarity Relationships

In the continuum of cross-linguistic similarity relationships, three distinct types can be discerned: similarity, contrast, and zero relations. The similarity relation means that an item or pattern in the TL is perceived as formally and/or functionally similar to a form in the L1 (or other previously learned language). Full-scale cross-linguistic similarity of both form and function is rare, except in closely related languages that are mutually comprehensible, such as Norwegian and Swedish.

In a contrast relation, the learner perceives a TL pattern as in important ways differing from an L1 form or pattern, though there is also an underlying similarity between them. An English-speaking learner of German, who is used to

a specific third-person suffix used with the present tense of verbs, will notice that German has a host of other personal endings for the verb as well. This means that there will be problems for the learner in producing the correct verb forms, but the learner is basically aware of the existence of a system and does not have to expend great effort to learn to understand the types of functions expressed by verbal suffixes. An example of a contrast relation is native speakers of English learning a Germanic or Romance target language: They will encounter similarities and differences in varying proportions, and the foundation of similarities will allow them to contrast what is different. As James (1980) says, "it is only against the background of sameness that differences are significant" (p. 169).

The zero relation does not mean that the learner finds nothing at all that is relevant to his L1 as the learning progresses. There are, after all, some linguistic universals common to all languages. But the level of abstraction in these is so high that an average learner cannot easily notice features that a totally different TL has in common with the L1. A zero relation simply means that items and patterns in the TL at early stages of learning are seen to have little or no perceptible relation to the L1 or any other language the learner knows. A learner who knows only Indo-European languages and starts learning Chinese will find it difficult to relate anything in the TL to his previous linguistic knowledge. The zero, or near-zero relation between Chinese and English poses great difficulties at early stages of learning, since the learner has to spend considerable time figuring out how the new language really works. As the learner's proficiency develops, he or she may become aware of an increasing number of points of contrast between the two languages. Thus, what represents a zero or near-zero relation for a beginning learner may in many respects eventually come to be seen as a contrast relation by that learner.

Item Transfer and System Transfer in Comprehension, Learning, and Production

Learners' reliance on perceived and assumed cross-linguistic similarities - i.e., transfer - can be manifested at three different levels: (1) item transfer, (2) system transfer or procedural transfer, and (3) overall transfer. These different levels of transfer can best be understood against the backdrop of the distinction between item learning and system learning, where the term item is defined as an individual form (e.g., sound, letter, morpheme, word, phrase, syntactic unit), and system is defined as a set of principles for organizing forms paradigmatically (e.g., assigning different functions to different forms of a word, as in go, goes, going, gone, went) and syntagmatically (e.g., rules for forming compound words, word order rules), and for mapping meanings onto those forms. Probably the first scholar to refer to the distinction between item learning and system learning as it pertains to language acquisition (L1 in this case) was Cruttenden (1981), who said that item learning "involves a form which is uniquely bonded with some other form or with a unique referent, whereas system-learning involves the

possibility of the commutation of forms or referents while some (other) form is held constant" (p. 79).

The distinction between item learning and system learning sheds light on one of the basic questions in transfer research: What is actually transferred? Item transfer means that a one-to-one relationship is established in the learner's mind between an item in the TL and either an item or a concept in the L1. Especially in the early stages of acquiring a new language, the unique bondage characteristic of item learning is predominantly cross-linguistic, since the (adult) language learner already has a full system of linguistic and conceptual representations, even if the structure and substance of those representations are not identical across languages. Initially, learning takes place on an item-by-item basis in all areas of language: phonological, morphological, syntactic, lexical, and phraseological.

The cross-linguistic similarities that underlie item transfer are a concretely perceived similarity of form usually combined with an associated assumed similarity of function or meaning. Being grounded in perceived (not just assumed) formal similarities between languages, item learning has a predominantly positive effect on learning, notably on learning for comprehension. When they are able to perceive similarities between L1 and L2 items, learners make use of an oversimplified L2 = L1 equivalence hypothesis, mapping the functions or meanings of L2 items directly onto existing L1 items during comprehension, and mapping the functions or meanings of L1 items onto L2 items during production. This is the case particularly at early stages of learning, when the learner's linguistic resources in the L2 are insufficient to make much use of intra-lingual similarities. The learner establishes oversimplified cross-linguistic one-to-one relationships to reduce the workload. At these early stages of learning, learners tend to focus on form rather than meaning or function, as meaning and function are more abstract and less accessible to direct observation and analysis. Perceived formal similarities aid learners in establishing cross-linguistic relations (also known as interlingual identifications, Odlin, 1989) in long-term memory.

Perceived similarity of form combined with an assumed similarity of meaning/ function provide the basis for establishing simplified one-to-one relations between L2 items and L1 items, which come to be represented as "primary counterparts" (a term used by Arabski, 1979) in the learner's mind. In system transfer - which might be better termed procedural transfer - abstract principles of organizing information are transferred. Positive procedural transfer lies behind the easy comprehension of a related TL, although it is not easy to point to concrete evidence for it. In this type of transfer, the learner assumes that there is cross-linguistic functional equivalence, but does not necessarily assume any formal item similarity between languages. Procedural transfer in the foreign-language learning context is predominantly transfer from L1, or possibly from another language the learner knows very well; by contrast, a language the learner knows only superficially generally serves as the source of only item transfer but not procedural transfer. Apparently, grammatical rules and semantic properties must be well internalized, or even fully automatized, in order to be transferred. Since functional and semantic systems in two languages are hardly ever fully congruent,

procedural transfer tends to lead to error, a negative effect, though the difficulty of recognizing positive effects of transfer must be acknowledged. What has been called interference or negative transfer in L2 production, might, in fact, be better described as the absence of relevant concrete (positive) transfer, leading to subsequent wrong assumptions about cross-linguistic similarities between L1 and L2. Positive transfer could then be described as "the application of at least partially correct perceptions or assumptions of cross-linguistic similarity."

Inevitable oversimplifications that occur during item learning will be modified as learning progresses. The learning of systems is always preceded by the learning of items, and, in fact, system learning involves a gradual modification of the oversimplified one-to-one relations that have been formed during item learning. The procedural transfer that occurs in learners' TL production is of three kinds: intrusive, inhibitive, and facilitative (see Hammerly, 1991). Intrusive transfer leads to the inappropriate use of L1-based items and structures. Inhibitive transfer prevents or inhibits the learner from learning how to use new words and structures appropriately. For example, TL words and structures that have no parallel in the L1 provide the learner with no concrete basis for positive item transfer and are therefore often avoided (cf. Sjoholm, 1995). Facilitative (positive) procedural transfer boosts learners' ability to access, process, and organize TL information due to similarities between the L1 and TL systems. Facilitative procedural transfer is often difficult to recognize, but it clearly constitutes an important factor affecting the learner's ease of comprehending a language perceived to be close. In language production, facilitative procedural transfer is especially hard to pin down because it coincides with successful acquisition of the TL. Nevertheless, the fact that learners of a related language regularly outperform learners of an unrelated language in their TL production attests to its presence (see, e.g., Ringbom, 1987).

Overall transfer is an umbrella term referring to the learner's reliance on conceptions of both formal similarities across individual items and functional equivalences between the underlying systems. The amount of overall transfer that will be present in a learner's language performance (both comprehension and production) depends on how much cross-linguistic similarity the learner can generally perceive between items and systems in the two languages, beginning from a common alphabet, phonemes in common and similar phonotactics, and extending to similarities in grammatical categories (case, gender, word classes), as well as to the number of cognates and other lexical similarities that exist across both languages. Quick and effective item learning for comprehension is what above all distinguishes the learning of a related TL from the learning of an unrelated language. As for procedural transfer, the matter is more complex. Learners normally assume that L1 procedures work also for L2 comprehension. The extent to which these assumptions actually work determines whether the effect is positive or negative. If inappropriate L1 procedures are applied to L2 comprehension, misinterpretations are likely to arise. Syntactic congruence plays a key role: if the L2 has categories that are absent in the L1, the functions of these have to be learnt for full comprehension to occur. Across closely related languages this does not pose much of a problem: L1 procedures are fairly appropriate and tend

to work well for comprehension. When learners have reached a stage of knowing a sufficient number of items (words), approximate comprehension can easily occur.

The two distinctions between item learning and system learning and between comprehension and production can be joined together to provide a simplistic model of the development of foreign-language learning. In SLA research, the concept of learning has been used in two different ways. It has referred to the learner's gradual achievement of an ability to use the target language productively, and in doing so has neglected the prior ability of being able to understand the language. In SLA research, the term learning has also been applied to both receptive and productive aspects of language proficiency, but then the very different mechanisms of comprehension and production are ignored.

We should, in fact, distinguish four different types of learning: (1) item learning for comprehension; (2a) item learning for production; (2b) system learning for comprehension; and (3) system learning for production. Cross-linguistic similarity is relevant to all of these stages but influences them somewhat differently. The starting point for the learner is item learning for comprehension. Across related languages, a fair receptive knowledge can be attained quite quickly. When learners can make use of positive item transfer to facilitate their acquisition of the ability to understand a new language, they can free many cognitive resources for other aspects of receptive learning. If, on the other hand, there is a zero or near-zero similarity relation between the languages, the normal gap between receptive and productive vocabulary will be reduced, in that the acquisition of receptive competence takes a long time for such learners. This can be seen quite clearly in a study by Takala (1984), who found similar-sized "active" and "passive" vocabularies in Finnish-speaking learners of English in a formal foreign-language learning situation.

Item learning for production and system learning for comprehension are stages following the initial stage of item learning for comprehension. They usually develop in parallel, but the learner can, of course, focus more on either of these, depending on the aims, the learning situation, individual learner characteristics, and so forth. In vocabulary production, formal similarity between words such as cognates may help the learner locate the intended L2 word in his or her mental lexicon for the TL (De Groot, 1993). But errors are more likely to occur in production than in comprehension, partly because of the absence of a context in production that would rule out the wrong interpretations that formal similarities can induce.

System learning for comprehension means that oversimplified one-to-one relations are gradually modified. The learner learns, for example, that one L2 word can have several meanings, that the past tense of verbs can be expressed by different morphemes, and, in general, that pragmatic aspects of the target language are also relevant for learning. Gradually the learner develops an improved understanding of how different L2 units correspond to L1 units and how they relate to the underlying concepts. Cross-linguistic relations are not as easy to establish in system learning as they are in item learning since the functional and semantic systems in two languages are reasonably congruent only in very closely

related languages. In system learning, the relevant cross-linguistic similarities are functional or semantic, and formal similarity plays a subordinate role, if any. This corresponds to what several studies of the organization of the learner's mental lexicon have found: that as learning progresses, learners rely less on phonological similarity and more and more on semantic similarity (Singleton, 1994, p. 54; 1999, p. 189; Kroll & De Groot, 1997, p. 174).

The message of L2 texts can be at least approximately comprehended if inflectional affixes and function words are left unnoticed or unanalyzed. For production, on the other hand, accuracy is more vital, and if grammatical morphemes are omitted, the result is errors causing misunderstanding or at least irritation in the listener. The learner has to recognize and appropriately make use of the many linguistic means of expressing the same meaning and also become able to use the same linguistic form for several purposes. This applies to both grammar and vocabulary. As system learning progresses, learners gradually learn to expand their lexical network in many dimensions. In this process, the importance of cross-linguistic similarity decreases at the same time as intra-lingual similarities within the TL become more and more important.

Implications for Teaching

Finally, concerning the relationship between learning and teaching, it is necessary to point out that much more knowledge is needed about the mechanisms through which language learning proceeds before the field is justified in pronouncing definitive statements about how languages can be taught most effectively. Nevertheless, we can point to some implications that are relatively clear. First, given the important role that cross-linguistic similarities play in language learning, a natural question to follow is whether and to what extent they could be put to effective use in teaching. In general terms, a good strategy would be to make use of, and even overuse, actual similarities at early stages of learning.

A project that has a didactic aim along these lines is EuroCom, where textbooks and other materials are being produced to make use of the facilitative potential for reading comprehension inherent in speakers of languages that are related, but not as close as to be mutually comprehensible. The project's main focus has been on the Romance languages, and the initiative has come from speakers of German. So far, German has been the main L1 used in the project. For an account in English of the EuroComRom project, see McCann, Klein, and Stegmann (2003). The basic idea of EuroComRom is simple. If you know one Romance language, you already have a lot of relevant knowledge that can be used for understanding all others. The learner is helped on the way by descriptions of what are called "The Seven Sieves," where systematic correspondences between all Romance languages are set out. The seven sieves are (1) international vocabulary, (2) pan-Romance vocabulary, (3) sound correspondences, (4) spelling and pronunciation, (5) syntactic structures, (6) morphosyntax, and (7) affixes. A project led by Britta Hufeisen applying the same principles on Germanic

languages has also begun (see Duke, Hufeisen, & Lutjeharms, 2004; Hufeisen, 2005).

Another clear implication is that learners of a closely related language have a far smaller learning burden than learners of a distant language. This means that there is less that they need to learn, that what they do need to learn is likely to be incorporated more easily into their existing knowledge, and that it will take them less time to arrive at a criterion level of language proficiency (see Ingram, 1975, p. 272; Odlin, 1989). Depending on the relative absence of cross-linguistic similarities, learners of a distant language will take considerably longer to reach that criterion level of proficiency in the TL (e.g., Ard & Homburg, 1992), they may need to pass through more stages of acquisition (e.g., Master, 1997), and they will tend to produce a higher proportion of errors (e.g., Schumann, 1986). Nevertheless, even though they have a larger learning burden, learners of a distant language can be ultimately as successful as learners of a closely related language when learning conditions are favorable (cf. Palmer, 1917/1968, p. 33; Sweet, 1899/ 1964, p. 54).

As in any concrete teaching situation, making use of cross-linguistic similarities requires that a number of both contextual and learner variables be considered. At least the following are particularly relevant:

The actual relation between the LI and L2. If the two languages are closely related, the teacher certainly needs to briefly outline the systematic recurring correspondences. But the closer the relation, the more teaching can concentrate on the actual differences that exist. Thus the teaching of Scandinavians learning another Scandinavian language would naturally focus on differences in pronunciation and the most common false friends. Where the languages are very distant, such as Japanese and English, the existence of mid- and high-frequency loanwords may facilitate learning. It is true that the loanwords from English into Japanese have been considerably modified, and the similarities have been obscured particularly by the differences in scripts and phonological systems. Even so, according to recent views (Daulton, 2007; Uchida & Scholfield, 2000), this built-in lexicon, gairaigo, provides a powerful tool for more effective learning. Developing strategies for maximally efficient use of high-frequency English loanwords provides a challenge for teachers and researchers in Japan.

Comprehension vs. production. If the learner is already able to comprehend the TL, at least approximately, teaching spoken and written production of that language may be necessary only for learners having a strong need to integrate with the TL community. This is the case when their job situation requires writing proficiency in the TL. For most purposes in the inter-Scandinavian context, however, Swedes, Norwegians, and Danes can count on generally being understood if they speak or write their own standard L1. Teaching a language not immediately comprehensible should acknowledge that cross-linguistic similarities are particularly important for comprehension.

Language proficiency. When a non-native language is closer to the TL than to the L1, comprehension can be facilitated by focusing on the similarities between the L2 and L3. For the productive aspects of L3 learning, however, it seems very

likely that only a high level of proficiency in a similar L2 is really useful. The inherent risks of confusion between related languages are more easily actualized if the learner has not successfully internalized grammatical rules and semantic properties in the L2.

Individual learner characteristics. While some learners appear to have blinders on, in that they do not take notice of even obvious cross-linguistic similarities, others are too prone to assuming similarities where they do not exist. Teaching needs to strike a balance between encouraging learners to make use of actual similarities and preventing exaggerated reliance on merely assumed similarities (cf. Haastrup, 1991, p. 341).

REFERENCES

Arabski, J. (1979). Errors as indications of the development of interlanguage. Katowice, Poland: Uniwersytet Slasky.

Ard, J. & Homburg, T. (1992). Verification of language transfer. In S. Gass & L. Selinker (eds.), Language transfer in language learning (pp. 47-70). Amsterdam: Benjamins.

Biskup, D. (1992). L1 influence on learners' rendering of English collocations: A Polish/ German empirical study. In P. Arnaud & H. Béjoint (eds.), Vocabulary and applied linguistics (pp. 85-93). London: Macmillan.

Cruttenden, A. (1981). Item-learning and system-learning. Journal of Psycholinguistic Research 10, 79-88.

Daulton, F. E. (2007). Japan's built-in lexicon of English-based loanwords. Clevedon, UK: Multilingual Matters.

De Angelis, G. (2005). Interlanguage transfer of function words. Language Learning 55, 379-414.

Dechert, H. & Raupach, M. (eds.) (1989). Transfer in language production. Norwood, NJ: Ablex.

De Groot, A. M. B. (1993). Word-type effects in bilingual processing tasks: Support for a mixed-representational system. In R. Schreuder & B. Weltens (eds.), The bilingual lexicon (pp. 27-51). Amsterdam: Benjamins.

Dewaele, J.-M. (1998). Lexical inventions: French interlanguage as L2 versus L3. Applied Linguistics 19, 471-90.

Duke, J., Hufeisen, B., & Lutjeharms, M. (2004). Die sieben Siebe des EuroCom für den multilingualen Einstieg in die Welt der germanischen Sprachen. In H. G. Klein & D. Rutke (eds.), Neuere Forschungen zur Europäischen Interkomprehension (pp. 109-34). Aachen: Shaker.

Eckman, F. R. (2004). From phonemic differences to constraint rankings. Studies in Second Language Acquisition 26, 513-49.

Eisenstein, M. & Bodman, J. (1993). Expressing gratitude in American English. In G. Kasper & S. Blum-Kulka (eds.), Interlanguage pragmatics (pp. 64-81). Oxford: Oxford University Press.

Ellegârd, A. (1978). On measuring language similarity. In J. Weinstock (ed.), Nordic languages and linguistics (pp. 195-216). Austin, TX: University of Texas.

Gass, S. & Selinker, L. (eds.) (1983). Language transfer in language learning. Rowley, MA: Newbury House.

Haastrup, K. (1991). Lexical inferencing procedures or talking about words: A book about receptive procedures in foreign language learning with special reference to English. Tübingen: Gunter Narr.

Hammerly, H. (1991). Fluency and accuracy. Clevedon, UK: Multilingual Matters.

Hufeisen, B. (2005). Mehrsprachigkeit: Fit für Babel. Gehirn & Geist: Das Magazin für Psychologie und Hirnforschung 6, 28-33.

Ingram, E. (1975). Psychology and language learning. In J. B. P. Allen & S. P. Corder (eds.), Papers in applied linguistics: The Edinburgh Course in Applied Linguistics, vol. 2, Language Learning Series (pp. 218-90). London: Oxford University Press.

James, C. (1980). Contrastive analysis. London: Longman.

Jarvis, S. (1998). Conceptual transfer in the interlingual lexicon. Bloomington, IN: Indiana University Linguistics Club Publications.

Jarvis, S. (2000). Methodological rigor in the study of transfer: Identifying L1 influence in the interlanguage lexicon. Language Learning 50, 245-309.

Jarvis, S. & Pavlenko, A. (2008). Crosslinguistic influence in language and cognition. Mahwah, NJ: Lawrence Erlbaum.

Kellerman, E. (1977). Towards a characterization of the strategy of transfer in second language learning. Interlanguage Studies Bulletin 2, 58-145.

Kellerman, E. (1978). Giving learners a break: Native language intuitions as a source of predictions about transferability. Working Papers on Bilingualism 15, 59-92.

Kellerman, E. (1979). Transfer and non-transfer: Where are we now? Studies in Second Language Acquisition 2, 37-57.

Kellerman, E. (1995). Crosslinguistic influence: Transfer to nowhere? Annual Review of Applied Linguistics 15, 125-50.

Kellerman, E. & Sharwood Smith, M. (eds.) (1986). Crosslinguistic influence in Second Language Acquisition. Oxford: Pergamon.

Kroll, J. & De Groot, A. M. B. (1997). Lexical and conceptual memory in the bilingual: Mapping form to meaning in two languages. In A. M. B. De Groot & J. Kroll (eds.), Tutorials in bilingualism: Psycholinguistic perspectives (pp. 169-99). Mahwah, NJ: Lawrence Erlbaum.

Master, P. (1997). The English article system: Acquisition, function, and pedagogy. System 25, 215-32.

McCann, W. J., Klein, H. G., & Stegmann, T. D. (2003). EuroComRom - the seven sieves: How to read all the Romance languages right away, 2nd revised edn. Aachen: Shaker.

Odlin, T. (1989). Language transfer: Cross-linguistic influence in language learning. Cambridge: Cambridge University Press.

Odlin, T. (2003). Crosslinguistic influence. In C. Doughty & M. Long (eds.), The handbook of second language acquisition (pp. 436-86). Oxford: Blackwell.

Odlin, T. (2006). Could a contrastive analysis ever be complete. In J. Arabski (ed.), Cross-linguistic influences in the second language lexicon (pp. 22-35). Clevedon, UK: Multilingual Matters.

Palmer, H. (1917/1968). The scientific study and teaching of language, Language and Language Learning Series. London: Oxford University Press.

Ringbom, H. (1987). The role of the first language in foreign language learning. Clevedon, UK: Multilingual Matters.

Ringbom, H. (2006). The importance of different types of similarity in transfer studies. In J. Arabski (ed.), Cross-linguistic influences in the second language lexicon (pp. 36-45). Clevedon, UK: Multilingual Matters.

Ringbom, H. (2007). Cross-linguistic similarity in foreign language learning. Clevedon, UK: Multilingual Matters.

Schumann, J. (1986). Locative and directional expressions in basilang speech. Language Learning 36, 277-94.

Selinker, L. (1969). Language transfer. General Linguistics 9, 67-92.

Seppänen, A. (1998). Finnish and English from a comparative perspective. In R. W. Cooper (ed.), Compare or contrast: Current issues in cross-language research. Tampere English Studies 6 (pp. 15-51). Tampere, Finland: Tampere University.

Singleton, D. (1994). Learning L2 lexis: A matter of form? In G. Bartelt (ed.), The dynamics of language processes: Essays in honor of Hans W. Dechert (pp. 45-57). Tübingen: Gunter Narr.

Singleton, D. (1999). Exploring the second language mental lexicon. Cambridge: Cambridge University Press.

Singleton, D. (2006). Lexical transfer: Interlexical or intralexical? In J. Arabski (ed.), Cross-linguistic influences in the second language lexicon (pp. 130-43). Clevedon, UK: Multilingual Matters.

Sjöholm, K. (1995). The influence of crosslinguistic, semantic, and input factors on the acquisition of English phrasal verbs: A comparison between Finnish and Swedish learners at an intermediate and advanced level. Äbo, Finland: Äbo Akademi University Press.

Sweet, H. (1899/1964). The practical study of languages, Language and Language Learning Series. London: Oxford University Press.

Takala, S. (1984). Evaluation of the students' knowledge of English vocabulary in the Finnish comprehensive school. Reports from the Institute for Educational Research 350. Jyväskylä, Finland: University of Jyväskylä.

Uchida, E. & Scholfield, P. (2000). Why words differ in difficulty: The learnability of English-Japanese cognates for Japanese learners of English. Last accessed: July 18, 2006. http:// privatewww.essex.ac.uk/~scholp/emimart2.htm.

Williams, S. & Hammarberg, B. (1998). Language switches in L3 production: Implications for a polyglot speaking model. Applied Linguistics 19 (pp. 295-333).

Wode, H. (1976). Developmental sequences in naturalistic second language acquisition. Working Papers on Bilingualism 11, 1-13.

FURTHER READING

Cenoz, J., Hufeisen, B., & Jessner, U. (eds.) ( 2001). Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives. Clevedon, UK: Multilingual Matters.

Cook, V. (ed.) (2002). Portraits of the L2 user. Clevedon, UK: Multilingual Matters.

Ellis, R. (2008). The study of second language acquisition, 2nd edn. Oxford: Oxford University Press.

Jarvis, S. & Pavlenko, A. (2008). Crosslinguistic influence in language and cognition. New York/ London: Routledge.

Nation, I. S. P. (1990). Teaching and learning vocabulary. New York: Newbury House.

Odlin, T. (ed.) (1994). Perspectives on pedagogical grammar. Cambridge: Cambridge University Press.

Ringbom, H. (2007). Cross-linguistic similarity in foreign language learning. Clevedon, UK: Multilingual Matters.

Schmitt, N. & McCarthy, M. (eds.) (1997). Vocabulary: Description, acquisition and pedagogy. Cambridge: Cambridge University Press.

8 Cognitive-Psychological Processes in Second Language Learning

ROBERT M. DEKEYSER

This chapter discusses core concepts in the cognitive psychology of second language learning: what are the various components of L2 knowledge? how are these components used? and most importantly, how are they learned or acquired, monitored, practiced, and consolidated?

The Components of Second Language Knowledge

The competence-performance distinction

While this is the oldest and most widely known of the distinctions we are discussing in this chapter, having been formulated by Chomsky (1965, p. 4), and having been the subject of countless books and articles, both in the Ll and L2 domains (e.g., Chipere, 2003; Hymes, 1972; McNamara, 1995; Sorace, 2003; Tarone, 1983; Taylor, 1988), it is probably the least useful for our purposes. It leaves out the whole area of processing, explicitly so (see, e.g., Chomsky, 1965, p. 9), and gives the impression that whatever is not part of competence is not systematic and not of linguistic interest. At the same time, of course, it presupposes that the rules of grammar are indeed rules of the mind (cf. e.g., Pinker, 1999; Pinker & Prince, 1988; Pinker & Ullman, 2002a, 2002b; Ullman, 2004), and not just a convenient summary of probabilistic behaviors, as the connectionists would have it (cf. e.g., Elman et al., 1996; McClelland & Patterson, 2002a, 2002b; Rumelhart & McClelland, 1986), and that competence is not merely an emergent property of performance (as Hall, Cheng, & Carlson, 2006, propose).

On the other hand, the terms are often used in a much broader sense, coinciding more or less with the knowledge-use distinction discussed below. This terminological vagueness is a source of great confusion and tends to sweep a variety of phenomena under the carpet, whether they be entirely cognitive in nature, such as syntactic processing mechanisms (e.g., Boland, 2005; Gibson, 1998; Pritchett, 1992), or more socially determined, such as accommodation of the interlocutor (e.g., Bell, 1984), or both (e.g., Tarone, 1983). It could even be argued that neither

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

competence nor performance is of much interest to the applied linguist, the former being a highly abstract, almost philosophical concept, and the latter an uninteresting collection of anomalies and restrictions, while the whole area in between, stretching from representation in a wider sense to processing and to socially variable use, is at the core of second language acquisition research and of applied linguistics more broadly.

The representation-processing distinction

A more useful distinction for the second language acquisition researcher than competence versus performance is representation versus processing. A large body of psycholinguistic literature has accumulated that deals with phonological, morphological, syntactic, and lexical processing, both in L1 and L2. It aims to find out what goes through the mind, in real time, of somebody who listens to, speaks, reads, or writes in a native or near-native language: what are the processes that bridge the gap between the form in which knowledge is stored in the mind and the form in which it is audible or visible in perception and production? This literature draws both on symbolist (e.g., Boland, 1995; Gibson, 1998; Pritchett, 1992) and connectionist (e.g., Elman et al., 1996) accounts of the representations that underlie processing, with symbolist accounts predominating. Clearly, the nature of processing depends on the nature of representation. It could be argued, therefore, that the study of processing should wait for more agreement on what representation looks like; others would argue, however, that the study of realtime processes gives unique insights into what representations processing draws on. As both the nature of representation and the nature of processing are not fully understood, it can even be hard to determine whether certain behavioral phenomena are due to the nature of representation or the nature of processing. Well-known foci of this debate in L1 are the discussion of whether subjacency restrictions are an issue of representation or of processing (see e.g., Huang, 1982; Pesetsky, 1987; Rizzi, 1982) and the research on the extent to which prepositional phrase attachment is influenced by discourse context or an issue of autonomous syntax (see e.g., Altmann, 1998; Frazier, 1987).

In L2 research, the representation-processing dichotomy has become prominent in the area of syntax recently through the work of Juffs (1998, 2004; Juffs & Harrington, 1995) on relative clauses, w^-questions, and various types of garden path effects, that of Lieberman, Aoshima, and Phillips (2006) and Williams, Möbius, and Kim (2001) on w^-questions, and of Jiang (2004, 2007) on plural -s, and of Trenkic (2007) on third-person -s. There is also recent work in the areas of L2 phonological processing (e.g., Escudero & Boersma, 2004; Sebastian-Galles & Soto-Faraco, 1999) and L2 lexical processing (e.g., Hernandez & Meschyan, 2006; Sunderman & Kroll, 2006).

The declarative-procedural distinction

This distinction is well known in cognitive psychology - even though it came out of computer science - and is increasingly used by SLA researchers. Declarative

knowledge is knowledge THAT something is, and can further be divided into semantic memory (knowledge of concepts, words, facts) and episodic memory (knowledge of events experienced). Procedural knowledge is knowledge HOW to do something, whether this involves psychomotor skills (knowing how to swim, ride a bicycle, or play tennis) or cognitive skills (knowing how to solve an equation, write a computer program, conjugate a verb, or read a text). The declarative-procedural distinction is central to models of skill acquisition, such as ACT-R (Anderson & Lebiere, 1998), has neurological support (e.g., Eichenbaum,

2000, 2002; Squire, 1992; Squire & Kandel, 1999; Squire & Knowlton, 2000; Squire & Zola, 1996), and has recently been brought to bear on issues of L1 processing (e.g., Pinker & Ullman, 2002a, 2002b) and of L1-L2 differences (e.g., Ullman,

2001, 2004, 2005).

This distinction is often equated with the explicit-implicit dichotomy, and the two pairs of concepts do overlap greatly, and can often be equated in certain contexts, but explicit is not exactly the same as declarative, and implicit not exactly the same as procedural. Declarative knowledge is not necessarily explicit, because it is not necessarily accessible to awareness (linguistic competence in the Chomskyan sense being a good example). On the other hand, procedural knowledge is not necessarily implicit, because it can be the result of proceduralization (and partial) automatization of declarative knowledge, and still allow or even require a certain degree of conscious access when being used. Nor is implicit knowledge necessarily procedural: knowledge of category prototypes, even including the knowledge of chunk strength involved in artificial grammar learning, may be implicit, but this implicit knowledge is neither declarative nor procedural (Squire & Knowlton, 2000); similarly, priming in amnesic patients seems to draw on knowledge that is implicit (no conscious access) but declarative (Kihlstrom, Dorfman, & Park, 2007).

In the second language domain, the declarative-procedural distinction has been used both in behavioral experiments (e.g., DeKeyser, 1997) and in neurophysio-logical studies (ERP, e.g., Morgan-Short, Steinhauer, Sanz, & Ullman, 2007; Mueller, Hahne, Fujii, & Friederici, 2005; Tokowicz & MacWhinney, 2005). Both lines of inquiry suggest a shift from reliance on declarative to reliance on procedural knowledge during the learning process within an individual. Such a shift has also been demonstrated, of course, in neurological studies of other forms of skill learning (e.g., Poldrack et al., 2001). Furthermore, the much more obvious difficulty of adult learners with grammar than with vocabulary suggests that the grammatical/procedural system is less available to them than lexical/declarative memory (Ullman, 2001, 2005).

The explicit-implicit distinction

Explicit knowledge is knowledge that one is aware of, that one has conscious access to. As a result in can be verbalized, at least in principle; not everybody has the cognitive and linguistic wherewithal to articulate that knowledge clearly and completely. Implicit knowledge is outside awareness, and therefore cannot be verbalized, only inferred indirectly from behavior, whether it be, for example,

semantic priming showing implicit declarative knowledge (faster than usual reaction to the word cat because one has recently seen a related word such as dog, even though one may not be aware of having seen this word), or apparently rule-governed behaviors, such as complex mental calculation or the production of a complex sentence both showing implicit procedural knowledge, i.e., knowledge of the rules underlying these behaviors, without being able to conceptualize those rules, let alone verbalize them. Most of us have long forgotten the steps involved in various forms of mental arithmetic that we were taught in grade school, and most of us have never even known, in the explicit sense, the basic rules of grammar of our native language (just ask the average American undergraduate when one puts an -s on the verb in English!).

In the second language domain, the extent to which L2 speakers are aware of the grammar knowledge they have - whatever the extent of this knowledge may be - depends, of course, on how they have learned the language. Furthermore, given the problems involved in operationalizing concepts such as awareness and verbalization (see, e.g., R. Ellis, 2004, 2005), it is virtually impossible to design "pure" measures of implicit or explicit knowledge of L2. Perhaps such "pure" measures should not even be a goal anyway, given that at least some psychologists see the implicit-explicit distinction as a matter of degree, as a continuum whose extremes may be rarely represented (e.g., Dienes & Perner, 1999).

The distinction is very important, however, for determining how L2 learners acquire and use L2 knowledge (see below), and therefore researchers have used a variety of imperfect measures (such as metalinguistic tests for explicit knowledge and timed grammaticality judgments for implicit knowledge). R. Ellis has taken the issue a step further by carrying out a factor analysis in order to determine to what extent two factors, implicit and explicit knowledge, can be identified as underlying a variety of L2 tests. While the methodological specifics are very tricky (Isemonger, 2007), this approach does seem promising (depending on the results of the factor analysis, one could put different weights on different tests for subsequent investigations of implicit and explicit knowledge), and further refinement seems to be leading to increasingly clear-cut results (Ellis & Loewen, 2007).

A special note is in order here about the term "metalinguistic (awareness)." This term is used with at least three different meanings in our field, which further obscures the debate. One can distinguish three layers of "metalinguistic knowledge": the first one is simply the awareness of what is right or wrong in a given sentence, without necessarily knowing why. This is the layer that Gombert (1992, p. 10) has called epilinguistic and that Karmiloff-Smith (1992, p. 52) refers to as "the E1 representational format," which "involves a redescription of information into a format that is accessible to certain tasks outside normal input-output relations but not yet to metalinguistic explanation." Clearly, this kind of knowledge can be entirely implicit, and therefore the term "metalinguistic" in this context is very confusing, because many equate "metalinguistic" with explicit. A second layer is "metalinguistic" in the sense of metacognition about language.

Just as one can have metacognitive knowledge about how one types, drives, or does mental arithmetic (and perhaps even of how one's typical behavior deviates from what one was taught!), one can have the same metacognitive knowledge about one's linguistic behavior. This is the sense of "metalinguistic" that corresponds to "explicit." The third layer is "metalinguistic" in the sense of language about language; this is the ability to verbalize, for which explicit knowledge is a necessary but not a sufficient condition.

The item-rule distinction

The use of rules can be laborious, even when they are proceduralized and even partially automatized. Therefore, under conditions of time pressure, it is very useful to have a shortcut available, and for those situations where a certain input is frequently processed by a rule and therefore frequently yields the same output, an associative learning process for that input-output pair may be more efficient. To take an example from math again, we all learned early on that instead of laboriously calculating 7 x 5 as being 7 + 7 + 7 + 7 + 7, going from 7 to 14 to 21 to 28 to 35, it is easier to just memorize the item from the table of multiplication that tells us 7 x 5 = 35 (note that the same distinction can be made at the lower level of addition in the sense that 7 + 7 = 14 is itself a shortcut for 7 + 1 + 1 + 1 + 1 + 1 + 1 + 1 = 14).

In the same way, for verb forms we often produce, it becomes more economical to store them as an "item" rather than having to put them together on the basis of a rule; for seldom used combinations, storage of all the items would be too cumbersome, and rule use is more economical. It is not surprising, therefore, that irregular forms are almost always high-frequency forms: as they are high-frequency, they tend to be retrieved as items, hence there is no analogical pressure, i.e., no pressure to make them conform to a rule; for low-frequency items, people tend to fall back on the rule, even if the prescriptive grammar requires an irregular, which over time leads to regularization. According to some, this need to combine rapid reaction with flexible knowledge that can be used under changing circumstances constitutes the evolutionary pressure that led to the declarative-procedural distinction (see e.g., Poldrack et al., 2001). (Note that the item-rule distinction is yet another one that often overlaps with the declarative-procedural distinction (cf., e.g., Pinker & Ullman, 2001), but cannot be equated with it: rules can be stored in declarative form before being proceduralized and automatized, as is traditionally held by skill acquisition theory).

Not only inflectional morphology can be produced this way; entire strings of words that are often used together may be stored and retrieved as "chunks," saving even more mental resources. Hence the well-known phenomenon of formulaic knowledge, both in L1 and L2 (see, e.g., Myles, 2004; Myles, Hooper, & Mitchell, 1998; Schmitt, 2004; Wray, 2001): L2 speakers, in particular, can often produce whole strings very rapidly and without errors, even if in other contexts they would have to slow down in order to produce some of the same morphological forms, and L2 learners can often produce such strings long before they

can use the constitutive elements productively. Even what appear to be subtle grammar errors may in fact be due to the use of certain lexemes with the grammatical morphemes they are most frequently used with, regardless of the appropriateness of those morphemes to express the specific meaning intended; this is especially likely to happen for structures where the form-meaning mapping is already obscure, as is the case for the Spanish "reflexive" pronoun se (Zyzik, 2006).

Neurological data confirm differences in storage for items and rules, and a shift from one to the other over time (e.g., Doeller et al., 2006; Fletcher et al., 1999; Opitz & Friederici, 2003, 2004).

Conclusion: Multiple sources of knowledge

As should be clear from the above, all L2 speakers have L2 knowledge of many kinds. On top of universal grammar (assuming it exists and is still accessible in L2 learning) and of elements transferred from L1, all learners have both declarative and procedural knowledge, implicit and explicit knowledge, knowledge of items and rules, and knowledge of how to process language in real time for production or comprehension. The good language learner will acquire as much knowledge of as many kinds as possible, and perhaps more importantly, will be able to draw on these various kinds of knowledge effectively and efficiently during language use, as we will see in the two sections that follow.

How the Components of Second Language Knowledge Are Used

The knowledge-use distinction

As mentioned above, this distinction is not to be confused with the narrower competence-performance dichotomy. Knowledge, as the previous section illustrates, includes much more than competence in the Chomskyan sense, the latter being necessarily implicit. Often students will have considerable amounts of explicit knowledge about parts of the L2, but little or no competence, i.e., implicit, intuitive knowledge, of the same elements in the same L2. Nor is use the same as performance. Many systematic differences in use that depend on both social and cognitive constraints are outside the scope of grammatical competence, while being more than moment-to-moment, more or less random, deviations from competence. Grammatical competence is only part of communicative competence, which is only part of knowledge, and performance is only part of use (cf. Canale & Swain, 1980; McNamara, 1995; Taylor, 1988). The age-old problem, then, of students who know all the rules but can't speak is not one of a gap between competence and performance but one of a gap between insufficiently proceduralized/automatized explicit knowledge, on the one hand, and very limited implicit and/or automatized knowledge on the other hand (see DeKeyser, 2007b).

Using implicit and explicit knowledge

A number of authors tried to document the extent to which students draw on implicit and explicit knowledge in language use (e.g., Bialystok, 1979; R. Ellis, 2004, 2005, 2006; Ellis & Loewen, 2007; Golonka, 2006; Green & Hecht, 1992; Hernandez & Meschyan, 2006; Hu, 2002; Hulstijn & Hulstijn, 1984; Jiang, 2004, 2007; Macrory & Stone, 2000; Renou, 2000; Schulz, 2002; Trenkic, 2007). While these authors define use in different ways, and while methodological problems abound in this area of research (see especially DeKeyser, 2003; and Isemonger, 2007), three points are fairly generalizable:

1 different tasks draw on the two kinds of knowledge to different extents;

2 just about all tasks draw on both explicit and implicit or automatized knowledge to some extent;

3 the gap between explicit and implicit or fully automatized knowledge is particularly wide for some elements of grammar.

R. Ellis (2006), in particular, shows that among the 17 ESL/EFL structures he examined, the rank order was quite different depending on whether a test was used that drew predominantly on implicit or explicit knowledge. No structures elicited significantly better scores for implicit than for explicit tests, but some were descriptively better. Several structures, however, yielded significantly higher scores on the explicit than on the implicit test; the biggest difference was found for the indefinite article, question tags, plural -s, and regular past -ed. Ellis also shows how, for some structures, explicit knowledge was a predictor of overall proficiency scores on tests of listening, speaking, reading and writing (especially for relative clauses and indefinite articles), whereas for other structures, it was implicit knowledge rather than explicit knowledge that predicted proficiency (especially for comparatives, unreal conditionals, since/for). These two kinds of data clearly suggest that different kinds of knowledge are accessed depending on the structure and the test, assuming, of course, that implicit and explicit are the best labels for the different kinds of tests involved; the factor analyses in R. Ellis (2005, 2006) and Ellis and Loewen (2007) can only show that two different components of knowledge are involved, not that the implicit-explicit distinction is the best label, rather than some correlated dichotomy.

As a case in point, Hu (2002) argues strongly, on the basis of ESL learners' differential performance on prototypical versus peripheral items, and their differential sensitivity to different factors conducive to monitoring (focus on form, sufficient time, knowledge of rule), that the phenomenon of differential knowledge access is not so much an issue of retrieval of explicit versus implicit knowledge, but of differentially automatized structures, i.e., various levels of explicit knowledge. Similarly, Jiang (2004, 2007) and Trenkic (2007) argue convincingly, the former for plural -s, and the latter for third-person -s, that, while the difficulty in using these structures spontaneously may be very sensitive to processing constraints, the processing constraints involved are a matter of

general cognition in the absence of implicit knowledge (and hence are evidence of a representational rather than a processing deficit in the psycholinguistic sense described above).

Using declarative, procedural, and automatized knowledge

While the implicit-explicit distinction may be the most useful for assessment at a given point in time, the declarative-procedural-automatized distinction is the most useful from a developmental perspective. To what extent is the learner's L2 knowledge represented in these three forms at various stages of the learning process, and to what extent can one form of knowledge be instrumental in developing a different form of knowledge? This question has been the subject of a decades-long debate in applied linguistics and has often been referred to as the interface issue, the acquisition-learning distinction or even the implicit-explicit distinction. Again, while in many cases these terms can be used interchangeably, it is most useful to use the distinction from skill acquisition theory to investigate to what extent second language learning (in adults) can develop along this path. Krashen has stated that "learning cannot turn into acquisition" (1985, pp. 42-3) and that automatization of explicitly learned rules only appears to be the case, because explicit knowledge of a given rule often far precedes its automatized form, but this temporal ordering does not imply causation (1985, p. 41), only that acquisition takes longer than learning (in other words, the two processes happen in parallel; one is not conducive to the other).

More recently Hulstijn (2002, p. 211) has argued that explicit knowledge of second language grammar cannot turn into implicit knowledge through automatization, and that, strictly speaking, there is no "automatization of rules" (emphasis in the original), because automatization implies a qualitative change, not mere speeding-up. This may appear to be another clear "no" answer to DeKeyser's question as to "whether the declarative knowledge that results from explicit learning processes can be turned into a form of procedural knowledge that is accessible in the same way as implicitly acquired knowledge" (2003, p. 328). It is important, however, not to misinterpret the word "turn into" in this context. The interface hypothesis, of which DeKeyser is a proponent, does not imply that one kind of knowledge turns into another in the sense that the more there is of one kind (procedural, automatized, or implicit), the less there is of the other (declarative, explicit) (pace Segalowitz & Hulstijn, 2005, p. 378). What it means is that the presence of one is conducive to, or plays a causal role in the development of the other, and on that point the literature is quite clear, in the sense that, while explicit learning certainly does not necessarily lead to eventual automatized, let alone implicit, knowledge, the likelihood of learners achieving a fairly high degree of automaticity in their use of a structure keeps evolving in parallel with their declarative knowledge about that structure, even in the latter years of foreign language instruction. (This position is a variant of the strong, not the weak interface position as these two are described, e.g., in R. Ellis (1993, 1994),

because the weak interface position implies that explicit knowledge is merely helpful in triggering or speeding up implicit learning processes, whereas the strong interface hypothesis implies a more direct causality in the sense that explicit declarative knowledge, given practice of sufficient quantity and quality, leads to proceduralization, (at least partial) automatization, and only in some cases, eventually, after proceduralization and automatization have taken place, to implicit representation.)

Schulz (2002), for instance, showed that declarative knowledge (even metalinguistic in the sense of verbalization) among learners of German L2 in college was a better predictor of their grammatical proficiency on a written test (r = 0.52) than experience abroad was. Other studies that demonstrate a link between students' declarative knowledge of a structure and their ability to use it, even fairly rapidly and correctly, which implies proceduralization and a certain degree of automatization as a function of the initial declarative knowledge, are de Jong (2005) and DeKeyser (1997). Both of these studies show that a certain degree of automaticity can be reached in learners who have solid declarative knowledge (whereas DeKeyser (1995) showed that no procedural, let alone automatized or implicit, knowledge developed in the absence of declarative knowledge, even after thousands of exposures, for the rather straightforward morphosyntactic form-meaning mappings investigated there). Both studies also demonstrate the skill specificity of the procedural knowledge documented (in the sense of little transfer between receptive and productive skills). As stated above, procedural knowledge is highly specific and therefore very limited in its transfer potential; only the more abstract declarative knowledge transfers relatively easily to new tasks.

Even for vocabulary, and even in the rather advanced learner, automatization may not be complete, and a certain degree of attention shifting (Segalowitz & Frenkiel-Fishman, 2005) or executive control (Hernandez & Meschyan, 2006) may be necessary; this does not prevent the learner (or indeed the bilingual) from being highly fluent.

In conclusion, then, while it may be the case that declarative knowledge does not turn into procedural or automatized knowledge in the sense that the more there is of the latter, the less there is of the former, and while it may also be the case that highly automatized knowledge is still not necessarily qualitatively the same as implicit knowledge, there is no reason to doubt the causal effect of declarative knowledge on the development of automatized knowledge.

How the Components of Second Language Knowledge Are Learned, Acquired, Practiced, Monitored, and Consolidated

The previous section describes how learners draw on their various sources of L2 knowledge for (communicative) use. This ability to use their knowledge is, of course, the result of an often very long process of learning, practicing, monitoring,

receiving feedback, modifying hypotheses, and practicing again. This section goes into some detail about each of these processes.

Learning and acquisition

The distinction between explicit and implicit learning, usually referred to in the SLA literature as learning versus acquisition, is a huge topic in cognitive psychology, where it is the subject of hundreds of empirical studies. The part of that literature that is most directly relevant to language learning, especially the research on artificial grammar learning, was covered in quite some detail in DeKeyser (2003), who came to the conclusion that "a thorough reading of the literature on implicit learning . . . must leave one very skeptical about the possibility of implicit learning of abstract structure, at least by adults" (p. 321). This is not only a vast, but highly controversial literature, leading some to draw more tentative conclusions at that time: "while there can be no doubt that both spontaneous and more deliberate L2 performance exist, what type of knowledge underlies each, and whether there is any connection between the two during SLA and L2 use, are contentious issues that are from settled in SLA, let alone any other domain of human cognition" (Doughty, 2003, p. 258).

The debate in experimental psychology has gone on since then, through empirical laboratory studies, literature reviews in journals and edited volumes, and conference panels. Space does not permit going into much detail here, so only some key references will be discussed. On the one hand, Pothos (2007), on the basis of his literature review and conceptual analysis of implicit and explicit knowledge of artificial grammars, comes to the conclusion that both forms of knowledge exist in parallel, or at least that different task demands lead to implicit versus explicit retrieval of the same kind of knowledge. He stresses, however, that there "has been no evidence that such rules [a rule-based representation of the artificial grammar, corresponding 'to the view of rules postulated in language (Chomsky, 1965)'] are developed in AGL [artificial grammar learning], nor a proposal for a corresponding learning mechanism" (2007, p. 239).

On the other hand, Sallas, Mathews, Lane, and Sun (2007) conclude from their experiments on implicit and explicit learning, also with artificial grammars, that it is important to draw learners' attention to structure, at exactly the right stage of the learning process; participants "achieved both fast and accurate string generation only if they were aided by structural (diagram) information during training and relevant information was highlighted just when it was needed" (p. 2130). These results contrast with those of Domangue et al. (2004), who managed to elicit either fast or accurate performance, but not both, depending on whether the instruction had focused on structural information (explicit knowledge) or memorization of lower-level information (implicit learning of structure). Crucial to the Sallas et al. findings were the facts that they did not quantify accuracy as the number of correct chunks in a string, but rather the number of perfect strings; that the learning of the model was facilitated through computer animation, enhancing both salience and attention; and that ample practice was provided to

allow proceduralization and automatization. Similar results were found by Bitan and Karni (2004) for letter decoding and word recognition in an artificial script.

Thus, explicit learning followed by practice for proceduralization and automatization can be superior to implicit learning, even in this artificial grammar domain, which has been a favorite paradigm for demonstrating the importance of implicit learning. Sallas et al. (2007) suggest that their results have practical implications for learning in a variety of domains taught in school: "Although our artificial grammar may seem different from topics normally taught in the classroom, such findings appear consistent with research suggesting that even topics whose underlying structure is relatively explicit and salient are typically learned better with guided instruction than relatively unguided discovery learning" (p. 2132). Here, of course, the attention is shifting to the distinction between inductive versus deductive explicit learning, and the educational psychology literature on this point is increasingly in favor of deductive presentation, in spite of many attempts over the decades, with a variety of conceptual variants and different terminologies, to show advantages of inductive learning. For a particularly interesting overview on this point, see Kirschner, Sweller, and Clark (2006).

The psychological literature on artificial grammar learning and other forms of implicit learning remains hard to interpret, however, in part because of the difficulty of distinguishing a number of overlapping dichotomies used as more or less equivalent in different studies. For a very good discussion of the relationship between these dichotomies as they apply to that literature, however, see Kihlstrom et al. (2007).

Meanwhile, in the domain of second language learning, the debate about implicit and explicit learning continues as well, and there too, a consensus is gradually developing that a synergy may be at play. Contrary to the development in cognitive psychology, however, where the synergy is seen as operating on the same specific element, synergy in SLA is often conceived of as implicit learning for some elements and explicit learning for others. In John Williams' work, for instance, there are indications that form-meaning mappings are learned explicitly, while form-form mappings may be learned implicitly (see especially Williams & Lovatt, 2003); other studies by the same author, however, suggest that both have to be learned explicitly (Williams, 2003) or that even abstract form-meaning mappings can be learned implicitly (Williams, 2005). The particularly rich discussion in Williams (2005) makes it clear that much more work is needed to try to isolate the role of various potential confounding variables, such as implicit-explicit knowledge of agreement patterns (in particular gender) in other languages and the degree of "abstractness" of the meaning involved (e.g., animacy vs. proximity). Furthermore, the 2005 study used an artificial pattern learned in the context of English sentences, which may also have influenced the nature or the degree of attention that subjects paid to the form-meaning mappings in question (articles agreeing with the noun for animacy and proximity).

Williams' research, even though conducted with a "natural language" in the sense of form-meaning mappings as in any existing human language, is still based on a miniature artificial language in the sense that learners are only

exposed to a very small system, with words and rules created for the experiment. In an experiment with Samoan as L2, Robinson (2005) too found evidence for different learning processes depending on the structure involved, in the sense that the relationship between the percentage correct for old items, new grammatical items, and new ungrammatical items was very different, depending on the structure involved (ergative, locative, incorporation).

Like Williams' research, Robinson's was carried out in a laboratory context. Most research on the development of implicit and explicit knowledge with classroom learners in recent years has involved different kinds of feedback, rather than the presentation aspect of instruction. Studies such as Egi (2007), R. Ellis, Loewen, and Erlam (2006), Lyster (1998a, 1988b, 2004), Mackey (2006), Mackey, Gass, and McDonough (2000), Nassaji (2007), and Sheen (2006) all point to the importance of the explicitness of feedback, along with other aspects of feedback that make it easier to notice and learn explicitly. For more information on learning from L2 feedback, see especially Leeman (2007), Mackey and Goo (2007), Nicholas, Lightbown, and Spada (2001), and Russell and Spada (2006).

Given the difficulty of defining, and especially of operationalizing implicit and explicit knowledge (see above), the issue of the relationship between implicit and explicit learning is likely to remain controversial for some time to come (R. Ellis, 2006), in the field of SLA as in cognitive psychology (see above). In the meantime, however, the preponderance of the evidence justifies N. Ellis' statement: "Many aspects of a second language are unlearnable - or at best are acquired very slowly - from implicit processes alone" (2005, p. 307).

The debate about the interface of, or overlap between, implicit and explicit knowledge extends into discussions of practice and monitoring, two areas to which we turn next.

Practice

If practice is understood as "specific activities in the second language, engaged in systematically, deliberately, with the goal of developing knowledge of and skills in the second language" (DeKeyser, 2007a, p. 8), then it is immediately clear that no discussion of practice can avoid the distinctions between implicit and explicit learning and between declarative and procedural knowledge. While there can be no doubt that the ultimate goal is to have (highly automatized) procedural L2 knowledge, this does not mean that declarative knowledge is not useful: it is a handy crutch to lean on whenever our procedural knowledge is insufficient, as is the case for early stages of learning, as well as for advanced learners who need to extend their performance into new domains or registers, or who are having occasional trouble retrieving their procedural knowledge.

Nor does the importance of procedural knowledge mean that initial learning activities should always be aimed at procedural knowledge or at implicit learning processes. It does mean, however, that extensive practice is necessary, and that this practice has to bridge the gap between the initial presentation of the L2 knowledge (in traditional deductive learning from the teacher's presentation) or

the initial hypotheses formed on the basis of the input (in more inductive learning, be it implicit or explicit) and the desirable end stage of fully proceduralized grammar. It follows that this practice will take different forms, from repeated use of the same formulae in cases where little declarative knowledge is available in analyzed form (cf. e.g., Gatbonton & Segalowitz, 1988), to processing instruction (see esp. VanPatten, 2004; cf. also DeKeyser, 2007c, 2007d) where explicit rule knowledge is available but not proceduralized (especially not in comprehension), to communicative drills in cases where this knowledge needs to be proceduralized in production, and eventually to large amounts of real-life or real-life-like two-way communicative practice for automatization and integration of this knowledge. "Slot-and-frame patterns, drills, mnemonics, and declarative statements of pedagogical grammar . . . all contribute to the conscious creation of utterances that then partake in subsequent implicit learning and proceduralization" (N. Ellis, 2005, p. 308). It is important that learners receive practice both in comprehension and production, as several laboratory studies (e.g., DeKeyser, 1997; de Jong, 2005) have shown that fluent and accurate performance in both modes for a given structure requires practice of that very structure in both modes. Besides the laboratory research by DeKeyser and de Jong, classroom research (e.g., Toth, 2006, in press) also shows the importance of production practice (and the teacher's role in bringing about procedural knowledge), and research on study abroad shows the long-term consequences of the lack of such production practice: failure to proceduralize in the classroom, which prevents automatization from the ample practice received during study abroad (DeKeyser, 2007b). (For a broader account of the importance of deliberate practice in learning advanced skills in a variety of domains, see especially Ericsson, 2006.)

Monitoring

Monitoring one's speech production is comparing one's output (or imminent output) to one's knowledge. This can be done on the basis of either explicit or implicit knowledge; Krashen referred to the former as "Monitoring" and the latter as "monitoring" or "monitoring by feel."

If monitoring is a resource-consuming process in L1 (Oomen & Postma, 2001, 2002), it is even more so in L2, especially at early stages, when much knowledge is only available in declarative form - or not at all. As long as no solid implicit knowledge has been acquired, monitoring on the basis of explicit knowledge is the only realistic possibility for the second language learner. This process takes not only knowledge, but also time and motivation, as Krashen has pointed out many times (e.g., 1982, 1985). Kormos (1999, 2000, 2006) acknowledges this, but also points out that learners are better able to monitor both form and meaning of their L2 production at the same time than much of the literature would suggest, especially learners who have received substantial explicit teaching.

Moreover, monitoring plays an essential role both in the detection of gaps in one's competence or knowledge and in the process of proceduralization and automatization, and hence, somewhat paradoxically, plays an important role in

bringing about practically useful forms of knowledge, which may eventually become implicit (DeKeyser, 2007b; Izumi, 2003; Kormos, 1999; Swain, 1985, 2005; Swain & Lapkin, 1995). The importance of monitoring for the development of L2 knowledge and skill is, given its reliance on explicit knowledge, yet another argument for not minimizing the importance of explicit knowledge in the adult learner.

Knowledge consolidation: Toward robust L2 knowledge

Monitoring, whether self-induced or prompted by corrective feedback, and the ensuing proceduralization and automatization, lead to a form of knowledge where the declarative and the procedural, the implicit and the explicit, rules and items, are all interlinked in a way that is maximally efficient: highly specific procedural knowledge is drawn on for the most familiar uses of language in the most familiar circumstances; more abstract and more flexible declarative knowledge is drawn on, often seamlessly, to fill in the rest. Such integrated knowledge is not only maximally efficient at a given point in time; it is also robust over time because a certain degree of redundancy between the various components of knowledge guarantees smooth retrieval and use of the knowledge required for fluent and accurate performance, in one way or another, even if some components or subcomponents deteriorate over time because of forgetting or interference. Such robust knowledge is only possible, however, if the various components exist and have been used often enough in varied enough circumstances for this process of integrative use to have been solidly established.

REFERENCES

Altmann, G. T. M. (1998). Ambiguity in sentence processing. Trends in Cognitive Sciences 2, 4, 146-52.

Anderson, J. R. & Lebiere, C. (1998). The atomic components of thought. Mahwah, NJ: Lawrence Erlbaum.

Bell, A. (1984). Language style as audience design. Language in Society 13, 145-204.

Bialystok, E. (1979). Explicit and implicit judgements of L2 grammaticality. Language Learning 29, 81-103.

Bitan, T. & Karni, A. (2004). Procedural and declarative knowledge of word recognition and letter decoding in reading an artificial script. Cognitive Brain Research 19, 229-43.

Boland, J. (2005). Cognitive mechanisms and syntactic theory. In A. Cutler (ed.), Twenty-first century psycholinguistics: Four cornerstones (pp. 23-42). Mahwah, NJ: Lawrence Erlbaum.

Canale, M. & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics 1, 1, 8-24.

Chipere, N. (2003). Understanding complex sentences: Native speaker variation in syntactic competence. New York: Palgrave.

Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.

de Jong, N. (2005). Can second language grammar be learned through listening? An experimental study. Studies in Second Language Acquisition 27, 2, 205-34.

DeKeyser, R. M. (1995). Learning second language grammar rules: An experiment with a miniature linguistic system. Studies in Second Language Acquisition 17(3), 379 -410.

DeKeyser, R. M. (1997). Beyond explicit rule learning: Automatizing second language morphosyntax. Studies in Second Language Acquisition 19, 2, 195-221.

DeKeyser, R. M. (2003). Implicit and explicit learning. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 313-48). Oxford: Blackwell.

DeKeyser, R. M. (2007a). Situating the concept of practice. In R. M. DeKeyser (ed.), Practice in a second language: Perspectives from applied linguistics and cognitive psychology (pp. 118). New York: Cambridge University Press.

DeKeyser, R. M. (2007b). Study abroad as foreign language practice. In R. M. DeKeyser (ed.), Practice in a second language: Perspectives from applied linguistics and cognitive psychology (pp. 208-26). New York: Cambridge University Press.

DeKeyser, R. M. (2007c). The future of practice. In R. M. DeKeyser (ed.), Practice in a second language: Perspectives from applied linguistics and cognitive psychology (pp. 287-304). New York, Cambridge University Press.

DeKeyser, R. M. (2007d). Skill acquisition theory. In B. VanPatten & J. Williams (eds.), Theories in second language acquisition: An introduction (pp. 97-112). Mahwah, NJ: Lawrence Erlbaum.

Dienes, Z. & Perner, J. (1999). A theory of implicit and explicit knowledge. Behavioral and Brain Sciences 22, 5, 735-55.

Doeller, C. F., Opitz, B., Krick, C., Mecklinger, A., & Reith, W. (2006). Differential hippocampal and prefrontal-striatal contributions to instance-based and rule-based learning. NeuroImage 31, 1802-16.

Dornyei, Z. (2005). The psychology of the language learner: Individual differences in second language acquisition. Mahwah, NJ: Lawrence Erlbaum.

Domangue, T. J., Mathews, R. C., Sun, R., Roussel, L. G., & Guidry, C. E. (2004). Effects of model-based and memory-based processing on speed and accuracy of grammar string generation. Journal of Experimental Psychology: Learning, Memory, and Cognition 30, 1002-11.

Doughty, C. J. (2003). Instructed SLA: Constraints, compensation, and enhancement. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 256-310). Oxford: Blackwell.

Egi, T. (2007). Interpreting recasts as linguistic evidence: The roles of linguistic target, length, and degree of change. Studies in Second Language Acquisition 29, 4, 511-37.

Eichenbaum, H. (2000). A cortico-hippocampal system for declarative memory. Nature Reviews Neuroscience 1, 41-50.

Eichenbaum, H. (2002). The cognitive neuroscience of memory: An introduction. New York: Oxford University Press.

Ellis, N. (2005). At the interface: Dynamic interactions of explicit and implicit language knowledge. Studies in Second Language Acquisition 27, 2, 305-52.

Ellis, R. (1993). The structural syllabus and second language acquisition. TESOL Quarterly 27, 1, 91-113.

Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford University Press.

Ellis, R. (2004). The definition and measurement of L2 explicit knowledge. Language Learning 54, 2, 227-75.

Ellis, R. (2005). Measuring implicit and explicit knowledge of a second language: A psychometric study. Studies in second language acquisition 27, 2, 141-72.

Ellis, R. (2006). Modelling learning difficulty and second language proficiency: The differential contributions of implicit and explicit knowledge. Applied Linguistics 27, 3, 431-63.

Ellis, R. & Loewen, S. (2007). Confirming the operational definitions of explicit and implicit knowledge in Ellis (2005): Responding to Isemonger. Studies in Second Language Acquisition 29, 1, 119-26.

Ellis, R., Loewen, S., & Erlam, R. (2006). Implicit and explicit corrective feedback and the acquisition of L2 grammar. Studies in Second Language Acquisition 28, 2, 339-68.

Elman, J. L., Bates, E. A., Johnson, M., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1996). Rethinking innateness: A connectionist perspective on development. Cambridge, MA: MIT Press.

Ericsson, K. A. (2006). The influence of experience and deliberate practice on the development of superior expert performance. In K. A. Ericsson et al. (eds.), The Cambridge handbook of expertise and expert performance (pp. 683-703). New York: Cambridge University Press.

Escudero, P. & Boersma, P. (2004). Bridging the gap between L2 speech perception research and phonological theory. Studies in Second Language Acquisition 26, 4, 551-85.

Fletcher, P., Büchel, C., Josephs, O., Friston, K., & Dolan, R. J. (1999). Learning-related neuronal responses in preferontal cortex studies with functional neuroimaging. Cerebral Cortex 9, 168-78.

Frazier, L. (1987). Sentence processing: A tutorial review. In M. Coltheart (ed.), Attention and performance 12: The psychology of reading (pp. 559-86). Mahwah, NJ: Lawrence Erlbaum.

Gatbonton, E. & Segalowitz, N. (1988). Creative automatization: Principles for promoting fluency within a communicative framework. TESOL Quarterly 22, 3, 473-92.

Gibson, E. (1998). Linguistic complexity: Locality of syntactic dependencies. Cognition 68, 1-76.

Golonka, E. M. (2006). Predictors revisited: Linguistic knowledge and metalinguistic awareness in second language gain in Russian. The Modern Language Journal 90, 4, 496-505.

Gombert, J.-E. (1992). Metalinguistic development. Chicago, IL, Chicago University Press.

Green, P. & Hecht, K. (1992). Implicit and explicit grammar: An empirical study. Applied Linguistics 13, 168-84.

Hall, J. K., Cheng, A., & Carlson, M. (2006). Reconceptualizing multicompetence as a theory of language knowledge. Applied Linguistics 27, 2, 220-40.

Hernandez, A. E. & Meschyan, G. (2006). Executive function is necessary to enhance lexical processing in a less proficient L2: Evidence from fMRI during picture naming. Bilingualism: Language and cognition 9, 2, 177-88.

Hu, G. (2002). Psychological constraints on the utility of metalinguistic knowledge in second language production. Studies in Second Language Acquisition 24, 3, 347-86.

Huang, C.-T. (1982). Logical relations in Chinese and the theory of grammar. Cambridge, MA: MIT Press.

Hulstijn, J. (2002). Towards a unified account of the representation, acquisition, and automatization of second-language knowledge. Second Language Research 18, 3, 193-223.

Hulstijn, J. & Hulstijn, W. (1984). Grammatical errors as a function of processing constraints and explicit knowledge. Language Learning 34, 23-43.

Hymes, D. (1972). On communicative competence. In J. B. Pride & J. Holmes (eds.), Socio-linguistics (pp. 269-93). Harmondsworth, UK: Penguin.

Isemonger, I. M. (2007). Operational definitions of explicit and implicit knowledge: Response to R. Ellis (2005) and some recommendations for future research in this area. Studies in Second Language Acquisition 29, 1, 101-18.

Izumi, S. (2003). Comprehension and production processes in second language learning: In search of the psycholinguistic rationale of the output hypothesis. Applied Linguistics 24, 2, 168-96.

Jiang, N. (2004). Morphological insensitivity in second language processing. Applied Psycholinguistics 25, 4, 603-34.

Jiang, N. (2007). Selective integration of linguistic knowledge in adult second language learning. Language Learning 57, 1, 1-33.

Juffs, A. (1998). The acquisition of semantics-syntax correspondences and verb frequencies in EL materials. Language Teaching Research 2, 2, 93-123.

Juffs, A. (2004). Representation, processing and working memory in a second language. Transactions of the Philological Society 102, 2, 199-225.

Juffs, A. & Harrington, M. (1995). Parsing effects in second language sentence processing: Subject and object asymmetries in wh-extraction. Studies in Second Language Acquisition 17, 4, 483-516.

Karmiloff-Smith, A. (1992). Beyond modularity: A developmental perspective on cognitive science. Cambridge, MA: MIT Press.

Kihlstrom, J. F., Dorfman, J., & Park, L. (2007). Implicit and explicit memory and learning. In M. Velmans & S. Schneider (eds.), The Blackwell companion to consciousness (pp. 52539). Oxford: Blackwell.

Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist 41, 2, 75-86.

Kormos, J. (1999). Monitoring and self-repair in L2. Language Learning 49, 2, 303-42.

Kormos, J. (2000). The role of attention in monitoring second language speech production. Language Learning 50, 2, 343-84.

Kormos, J. (2006). Speech production and second language acquisition. Mahwah, N: Lawrence Erlbaum.

Krashen, S. D. (1982). Principles and practice in second language acquisition. Englewood Cliffs, NJ: Prentice-Hall.

Krashen, S. D. (1985). The input hypothesis. London/New York: Longman.

Leeman, J. (2007). Feedback in L2 learning: Responding to errors during practice. In R. DeKeyser (ed.), Practice in a second language: Perspectives from applied linguistics and cognitive psychology (pp. 111-37). New York: Cambridge University Press.

Lieberman, M., Aoshima, S., & Phillips, C. (2006). Nativelike biases in generation of wh-questions by nonnative speakers of Japanese. Studies in Second Language Acquisition 28, 3, 423-48.

Lyster, R. (1998a). Negotiation of form, recasts, and explicit correction in relation to error types and learner repair in immersion classrooms. Language Learning 48, 2, 183-218.

Lyster, R. (1998b). Recasts, repetition, and ambiguity in L2 classroom discourse. Studies in Second Language Acquisition 20, 1, 51-81.

Lyster, R. (2004). Differential effects of prompts and recasts in form-focused instruction. Studies in Second Language Acquisition 26, 3, 399-432.

Mackey, A. (2006). Feedback, noticing, and instructed second language learning. Applied Linguistics 27, 3, 405-30.

Mackey, A., Gass, S., & McDonough, K. (2000). How do learners perceive interactional feedback? Studies in Second Language Acquisition 22, 4, 471-99.

Mackey, A. & Goo, J. (2007). Interaction research in SLA: A meta-analysis and research synthesis. In A. Mackey (ed.), Conversational interaction in second language acquisition (pp. 407-52). New York: Oxford University Press.

Macrory, G. & Stone, V. (2000). Pupil progress in the acquisition of the perfect tense in French: The relationship between knowledge and use. Language Teaching Research 4, 1, 55-82.

McClelland, J. L. & Patterson, K. (2002a). "Words OR rules" cannot exploit the regularity in exceptions. Trends in Cognitive Sciences 6, 11, 464-5.

McClelland, J. L. & Patterson, K. (2002b). Rules or connections in past-tense inflections: What does the evidence rule out? Trends in Cognitive Sciences 6, 11, 465-72.

McNamara, T. F. (1995). Modelling performance: Opening Pandora's box. Applied Linguistics 16, 2, 159-79.

Morgan-Short, K., Steinhauer, K., Sanz, C., & Ullman, M. T. (2007). An ERP investigation of second language processing: Effects of proficiency and explicit and implicit training. Journal of Cognitive Neuroscience (Supplement), 164.

Mueller, J. L., Hahne, A., Fujii, Y., & Friederici, A. G. (2005). Native and nonnative speakers' processing of a miniature version of Japanese as revealed by ERPs. Journal of Cognitive Neuroscience 17, 8, 1229-44.

Myles, F. (2004). From data to theory: The over-representation of linguistic knowledge in SLA. Proceedings of the National Academy of Sciences, USA 102, 2, 139-68.

Myles, F., Hooper, J., & Mitchell, R. (1998). Rote or rule? Exploring the role of formulaic language in classroom foreign language learning. Language Learning 48, 3, 323-63.

Nassaji, H. (2007). Elicitation and reformulation and their relationship with learner repair in student-teacher dyadic interaction. Language Learning 57, 4, 511-48.

Nicholas, H., Lightbown, P. M., & Spada, N. (2001). Recasts as feedback to language learners. Language Learning 51, 4, 719-58.

Oomen, C. C. E. & Postma, A. (2001). Effects of time pressure on mechanisms of speech production and self-monitoring. Journal of Psycholinguistic Research 30, 2, 163-84.

Oomen, C. C. E. & Postma, A. (2002). Limitations in processing resources and speech monitoring. Language and Cognitive Processes 17, 2, 163-84.

Opitz, B. & Friederici, A. D. (2003). Interactions of the hippocampal system and the prefrontal cortex in learning language-like rules. NeuroImage 19, 1730 -7.

Opitz, B. & Friederici, A. D. (2004). Brain correlates of language learning: The neural dissociation of rule-based vs. similarity-based learning. Journal of Neuroscience 24, 843640.

Pesetsky, D. (1987). Wh-in-situ: Movement and unselective binding. In E. Reuland & A. G. B. ter Meulen (eds.), The representation of indefiniteness (pp. 98-129). Cambridge, MA: MIT Press.

Pinker, S. (1999). Words and rules: The ingredients of language. New York: Basic Books.

Pinker, S. & Prince, A. (1988). On language and connectionism: Analysis of a parallel distributed processing model of language acquisition. In S. Pinker & J. Mehler (eds.), Connections and symbols (pp. 73-193). Cambridge, MA: MIT Press.

Pinker, S. & Ullman, M. T. (2002a). The past and future of the past tense. Trends in Cognitive Sciences 6, 11, 456-63.

Pinker, S. & Ullman, M. T. (2002b). Combination and structure, not gradedness, is the issue. Trends in Cognitive Sciences 6, 11, 472-4.

Poldrack, R. A., Clark, J., Pare-Blagoev, E. J., et al. (2001). Interactive memory systems in the human brain. Nature 414, 546-9.

Poldrack, R. A. & Gabrieli, J. D. E. (2001). Characterizing the neural mechanisms of skill learning and repetition priming: Evidence from mirror reading. Brain 124, 67-82.

Pothos, E. M. (2007). Theories of artificial grammar learning. Psychological Bulletin 133, 2, 227-44.

Pritchett, B. L. (1992). Grammatical competence and parsing performance. Chicago, IL: University of Chicago Press.

Renou, J. M. (2000). Learner accuracy and learner performance: The quest for a link. Foreign Language Annals 33, 168-80.

Rizzi, L. (1982). Issues in Italian syntax. Dordrecht: Foris.

Robinson, P. (2005). Cognitive abilities, chunk-strength and frequency effects in implicit artificial grammar and incidental L2 learning: replications of Reber, Walkenfield and Hernstadt (1991) and Knowlton and Squire (1996) and their Relevance for SLA. Studies in Second Language Acquisition 27, 2, 235-68.

Rumelhart, D. & McClelland, J. (1986). On learning the past tenses of English verbs. In J. McClelland, D. Rumelhart, & PDP Research Group (eds.), Parallel distributed processing, vol. 2: Psychological and biological models (pp. 217-71). Cambridge, MA: MIT Press.

Russell, J. & Spada, N. (2006). The effectiveness of corrective feedback for the acquisition of L2 grammar: A meta-analysis of the research. In J. Norris & L. Ortega (eds.), Synthesizing research on language learning and teaching (pp. 133-64). Philadelphia/Amsterdam: John Benjamins.

Sallas, B., Mathews, R. C., Lane, S., & Sun, R. (2007). Developing rich and quickly accessed knowledge of an artificial grammar. Memory and Cognition 35, 8, 2118-33.

Schmitt, N. (ed.) (2004). Formulaic sequences. Philadelphia. Benjamins.

Schulz, R. A. (2002). Hilft es die Regel zu wissen um sie anzuwenden? Das Verhältnis von metalinguistischem Bewusstsein und grammatischer Kompetenz in DaF. Die Unterrichtspraxis 35, 1, 15-24.

Sebastiän-Galles, N. & Soto-Faraco, S. (1999). On-line processing of native and non-native contrasts in early bilinguals. Cognition 72, 111-23.

Segalowitz, N. & Frenkiel-Fishman, S. (2005). Attention control and ability level in a complex cognitive skill: Attention shifting and second-language proficiency. Memory and cognition 33, 644-53.

Segalowitz, N. & Hulstijn, J. (2005). Automaticity in bilingualism and second language learning. In J. F. Kroll & A. M. B. de Groot (eds.), Handbook of bilingualism: Psycholinguistic perspectives (pp. 371-88). Oxford: Oxford University Press.

Sheen, Y. (2006). Exploring the relationship between characteristics of recasts and learner uptake. Language Teaching Research 10, 4, 361-92.

Sorace, A. (2003). Near-nativeness. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 130-51). Oxford: Blackwell.

Squire, L. (1992). Memory and the hippocampus: A synthesis from findings with rats, monkeys, and humans. Psychological Review 99, 2, 195-231.

Squire, L. R. & Kandel, E. R. (1999). Memory: From mind to molecules. New York: Scientific American.

Squire, L. R. & Knowlton, B. J. (2000). The medial temporal lobe, the hippocampus, and the memory systems of the brain. In M. S. Gazzaniga (ed.), The new cognitive neurosciences (pp. 765-79). Cambridge, MA: MIT Press.

Squire, L. R. & Zola, S. M. (1996). Structure and function of declarative and nondeclarative memory systems. Proceedings of the National Academy of Sciences, USA 93, 13515-22.

Sunderman, G. & Kroll, J. F. (2006). First language activation during second language lexical processing: An investigation of lexical form, meaning, and grammatical class. Studies in Second Language Acquisition 28, 3, 387-422.

Swain, M. (1985). Communicative competence: some roles of comprehensible input and comprehensible output in its development. In S. M. Gass & C. G. Madden (eds.), Input in second language acquisition (pp. 235-53). Rowley, MA: Newbury House.

Swain, M. (2005). The output hypothesis: Theory and research. In E. Hinkel (ed.), Handbook of research in second language teaching and learning (pp. 471-83). Mahwah, NJ: Lawrence Erlbaum.

Swain, M. & Lapkin, S. (1995). Problems in output and the cognitive processes they generate: A step towards second language learning. Applied Linguistics 16, 3, 371-91.

Tarone, E. (1983). On the variability of interlanguage systems. Applied Linguistics 4, 2, 14363.

Taylor, D. S. (1988). The meaning and use of the term "competence" in linguistics and applied linguistics. Applied Linguistics 9, 2, 148-68.

Tokowicz, N. & MacWhinney, B. (2005). Implicit and explicit measures of sensitivity to violations in second language grammar: An event-related potential investigation. Studies in Second Language Acquisition 27, 2, 173-204.

Toth, P. (2006). Processing instruction and a role for output in second language acquisition. Language Learning 56, 2, 319-85.

Toth, P. (in press). Teacher- and learner-led discourse in task-based grammar instruction: Providing procedural assistance for L2 morphosyntactic development. Language Learning.

Trenkic, D. (2007). Variability in second language article production: Beyond the representational deficit vs. processing constraints debate. Second Language Research 23, 3, 289-327.

Ullman, M. T. (2001). A neurocognitive perspective on language: The declarative/procedural model. Nature Reviews: Neuroscience 2, 717-26.

Ullman, M. T. (2004). Contributions of memory circuits to language: The declarative/ procedural model. Cognition 92, 231-70.

Ullman, M. (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 (pp. 141-78). Washington, DC: Georgetown University Press.

VanPatten, B. (ed.) (2004). Processing instruction: Theory, research, and commentary. Mahwah, NJ: Lawrence Earlbaum.

Williams, J. N. (2003). Inducing abstract linguistic representations: Human and connection-ist learning of noun classes. In R. van Hout, A. Hulk, F. Kuiken, & R. Towell (eds.), The interface between syntax and the lexicon in second language acquisition (pp. 151-74). Amsterdam: Benjamins.

Williams, J. N. (2005). Learning without awareness. Studies in Second Language Acquisition 27, 2, 269-304.

Williams, J. N. & Lovatt, P. (2003). Phonological memory and rule learning. Language Learning 53, 1, 67-121.

Williams, J. N., Möbius, P., & Kim, C. (2001). Native and non-native processing of English wh-questions: Parsing strategies and plausibility constraints. Applied Psycholinguistics 22, 4, 509-40.

Wray, A. (2001). Formulaic language and the lexicon. New York: Cambridge University Press.

Zyzik, E. (2006). Transitivity alternations and sequence learning: Insights from L2 Spanish production data. Studies in Second Language Acquisition 28, 3, 449-85.

9 Optimizing the Input: Frequency and Sampling in Usage-Based and Form-Focused Learning

NICK C. ELLIS

Estimating How Language Works: From Tokens to Types to System

Learners' understanding of language and of how it works is based upon their experience of language. They have to estimate the system from a sample. This chapter considers the effects of input sample, construction frequency, and processing orientation on learning. It draws out implications for usage-based acquisition and form-focused instruction for second (L2) and foreign (FL) language learners.

A language is not a fixed system. It varies in usage over speakers, places, and time. Yet despite the fact that no two speakers own an identical language, communication is possible to the degree that they share constructions (form-meaning correspondences) relevant to their discourse.1 Language learners have to acquire these constructions from usage, and beginners don't have much to go on in building the foundations for basic interpersonal communication. They have to induce the types of construction from experience of a limited number of tokens. Their very limited exposure poses them the task of estimating how linguistic constructions work from an input sample that is incomplete, uncertain, and noisy. How do they achieve this, and what types of experience can best support the process?

Nativelike fluency, idiomaticity, and selection are another level of difficulty again. For a good fit, every utterance has to be chosen, from a wide range of possible expressions, to be appropriate for that idea, for that speaker, for that place, and for that time. And again, learners can only estimate this from their finite experience. What are the best usage histories to support these abilities?

Language, a moving target, can neither be described nor experienced comprehensively, and so, in essence, language learning is estimation from sample. Like other estimation problems, successful determination of the population characteristics is

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

a matter of statistical sampling, description, and inference. For language learning the estimations include: What is the range of constructions in the language? What are their major types? Which are the really useful ones? What is their relative frequency distribution? How do they map function and form, and how reliably? How can this information best be organized to allow its appropriate and fluent access in recognition and production? Are there variable ways of expressing similar meanings? How are they distributed across different contexts? And so on. Et cetera. And so forth. Like.

Frequency of usage, in various guises, determines acquisition (Ellis, 2002a, 2002b). There are three fundamental aspects of this conception of language learning as statistical sampling and estimation:

• The first and foremost concerns sample size: As in all surveys, the bigger the sample, the more accurate the estimates, but also the greater the costs. Native speakers estimate their language over a lifespan of usage. L2 and FL learners just don't have that much time or resource. Thus, both of these additional language (AL) learner groups are faced with a task of optimizing their estimates of language from a limited sample of exposure. Broadly, power analysis dictates that attaining nativelike fluency and idiomaticity requires much larger usage samples than does basic interpersonal communicative competence in predictable contexts. But for the particulars, what sort of sample is needed adequately to assess the workings of constructions of, respectively, high, medium, and low base occurrence rates, of more categorical versus more fuzzy patterns, of regular versus irregular systems, of simple versus complex "rules," of dense versus sparse neighbourhoods, et cetera?

• The second concerns sample selection: Principles of survey design dictate that a sample must properly represent the strata of the population of greatest concern. Thus, Needs Analysis (Brown, this volume) is relevant to all AL learners. Thus, too, the truism that FL learners, who have much more limited access to the authentic natural source language than L2 learners, are going to have greater problems of adequate description. But what about learning particular constructions? What is the best sample of experience to support this? How many examples do we need? In what proportion of types and tokens? Are there better sequences of experience to optimize estimation? What learning increment comes from each experience? Is this a constant or does it diminish over time as dictated by the power law of practice? And so forth.

• A final implication of language acquisition as estimation concerns sampling history: How does knowledge of some cues and constructions affect estimation of the function of others? What is the best sequence of language to promote learning new constructions? And what is the best processing orientation to make this sample of language the appropriate sample of usage? Like.

This chapter first describes the units of language acquisition - linguistic constructions - and then considers how sample size and sample selection affect the development of constructions (their consolidation, generalization, and probabilistic

tuning) from naturalistic input. There are established effects of input token frequency, type frequency, Zipfian frequency distribution2 of the construction-family, and neighborhood homogeneity.

Next, it describes how sample size and sample selection affect usage-based language acquisition across the board - native and AL both. It reviews how learners' models of language broadly reflect the constructions in their sample of experience and how they unconsciously tally and collate a rich knowledge of the relative frequencies of these constructions in their input history. Because language learning is less an issue of the collection of linguistic constructions than of their cataloguing, organization, and marshalling for efficient appropriate use, this implicit knowledge is essential to fluent processing. In order for the estimation procedures rationally to produce a model of the language that optimizes the probabilistic knowledge of constructions and their mappings, learners must be exposed to a representative sample of authentic input that is appropriate to their needs. The chapter also considers the implications of modularity and transfer-appropriate processing for tuning the full range of necessary representative modalities and functions of usage.

Finally it nods at analyses of transfer in AL acquisition, how prior estimation of L1 biases the usage-based estimation of an AL, and why form-focused instruction may be necessary to reset some counters to tally the L2 more appropriately.

The Units of Language Acquisition

Construction Grammar (Goldberg, 1995, 2003, 2006; Tomasello, 2003) and other Cognitive Linguistic theories of first (Croft & Cruise, 2004; Langacker, 1987; Taylor, 2002; Tomasello, 1998) and second language (Robinson & Ellis, 2008b) acquisition hold that the basic units of language representation are constructions. These are form-meaning mappings, conventionalized in the speech community, and entrenched as language knowledge in the learner's mind. Constructions vary in specificity and in complexity, including morphemes (anti-, -ing, N-s), words (aard-vark, and), complex words (antediluvian, multimorphemic), idioms (hit the jackpot), semi-productive patterns (Good <time of day>), and syntactic patterns [Subj [V Obj 1 Obj2]]; [Subj be- Tns V -en by Obl]. Hence morphology, lexicon, and syntax are uniformly represented in Construction Grammar. Constructions are symbolic, in that their defining properties of morphological, lexical, and syntactic form are associated with particular semantic, pragmatic, and discourse functions. Constructions form a structured inventory of a speaker's knowledge of the conventions of their language, where schematic constructions can be abstracted over the less schematic ones, which are inferred inductively by the speaker in acquisition. A construction may provide a partial specification of the structure of an utterance; hence, an utterance's structure is specified by a number of distinct constructions. Constructions are independently represented units in a speaker's mind. Any construction with unique, idiosyncratic formal or functional properties must be represented independently in order to capture a speaker's knowledge of their

language. However, absence of any unique property of a construction does not entail that it is not represented independently and simply derived from other, more general or schematic constructions. Frequency of occurrence may lead to independent representation of even "regular" constructional patterns.

Acquiring Constructions

Usage-based theories of naturalistic language acquisition hold that we learn language through using language. Creative linguistic competence emerges from learners' piecemeal acquisition of the many thousands of constructions experienced in communication, and from their frequency-biased abstraction of the regularities in this history of usage. Competence and performance both emerge from the conspiracy of memorized exemplars of construction usage, with competence being the integrated sum of prior usage and performance its dynamic contextualized activation (Ellis, 1998, 2003, 2006a, 2007; Ellis & Larsen Freeman, 2006).

Many of the constructions we know are quite specific, formulaic utterances based on particular lexical items, ranging, for example, from a simple "Wonderful!" to increasingly complex phrases like "One, two, three," "Once upon a time," or "Won the battle, lost the war." These sequential patterns of sound, like words, are acquired as a result of chunking from repeated usage (Ellis, 1996; Pawley & Syder, 1983; Wray, 2002). In building up these sequences, learners bind together the chunks that they already know, with high-frequency sequences being more strongly bound than lower-frequency ones (Ellis, 2002a). In analyzing these sequences, the highest-frequency chunks stand out as the most likely constituents of the parse. The constructions already acquired by the learner constitute the sample of evidence from which they implicitly and explicitly identify regularities, so generalizing their knowledge by inducing unconscious schemata and prototypes that map meaning and form, and by abducing conscious metalinguistic hypotheses about language, too. These are the foundations, then, of new expressions and new understandings.

Constructionist approaches to language acquisition (Bybee & Hopper, 2001; Goldberg, 2003; Robinson & Ellis, 2008b; Tomasello, 1998, 2003) thus emphasize piecemeal learning from concrete exemplars. A high proportion of children's early multi-word speech is produced from a developing set of slot-and-frame patterns. These patterns are often based around chunks of one or two words or phrases, and they have "slots" into which the child can place a variety of words, for instance subgroups of nouns or verbs (e.g., I can't + VERB; where's + NOUN + gone?). Children are very productive with these patterns, and both the number of patterns and their structure develop over time. But initially, they are lexically specific. For example, if a child has two patterns, I can't + X and I don't + X, the verbs used in these two X slots typically show little or no overlap, suggesting (1) that the patterns are not yet related through an underlying grammar (the child doesn't "know" that can't and don't are both auxiliaries or that the words that

appear in the patterns all belong to a category of Verb), and (2) that learners are picking up frequent patterns from what they hear around them and only slowly making more abstract generalizations as the database of related utterances grows (Pine & Lieven, 1993; Pine, Lieven, & Rowland, 1998; Tomasello, 1992). Tomasello's (1992) Verb Island hypothesis holds that it is verbs and relational terms that are the individual islands of organization in young children's otherwise unorganized grammatical system: the child initially learns about arguments and syntactic markings on a verb-by-verb basis, and ordering patterns and morphological markers learned for one verb do not immediately generalize to other verbs. Positional analysis of each verb island requires memories of the verb's usage, the exemplars of its collocations, and the constructions it commonly inhabits. Over experience, syntagmatic categories emerge from the regularities in this data set, the learner's sample of language.

The chapters in Robinson and Ellis (2008b) extend these cognitive linguistic/ construction grammar theories of child language acquisition to the naturalistic acquisition of ALs in adulthood, so developing a usage-based approach to SLA. Some of the key features are as follows.

Frequency and the Roles of Input

AL Learners' knowledge of a linguistic construction depends, too, on their experience of its use, the sample of its manifestations of usage. Different frequencies of exemplification, and different types of repetition of a linguistic pattern, have different effects upon acquisition - the consolidation, generalization, and productive use of constructions. A key separation is between type and token frequency.

Type and token frequency

The token frequency of a construction is how often in the input that particular word or specific phrase appears; we can count in a sample corpus the token frequency of any specific form (e.g., the syllable [ka], the trigram aze, the word frog, the phrase on the whole, the sentence I love you). Type frequency, on the other hand, is the calculation of how many different lexical items a certain pattern, paradigm, or construction applies to, i.e., the number of distinct lexical items that can be substituted in a given slot in a construction, whether it is a word-level construction for inflection or a syntactic construction specifying the relation among words. For example, the "regular" English past tense -ed has a very high type frequency because it applies to thousands of different types of verbs, whereas the vowel change exemplified in swam and rang has much lower type frequency. Similarly the prepositional transfer construction [Subj [V ObjDir to ObjInd]] has a high type frequency (give, read, pass, donate, display, explain ...) because many different verbs can be used in this way, whereas the ditransitive alternative [Subj [V ObjInd ObjDir]] is only used with a small set of verbs like give, read, and pass and not others (*donate, *display, *explain).

Consolidating a particular formulaic construction: The role of token frequency

Like other concrete constructions, a word can be sketchily learned from a single exposure, as a fast mapping (Carey & Bartlett, 1978), a relation between an approximation of its sound and its likely meaning, forged as an explicit episodic memory relating its form and the perception of its likely referent (Ellis, 2005). The hippocampus and limbic structures in the brain allow us such unitary bindings from single experiences, rapid explicit memory, one-off learning, the establishment of new conjunctions of arbitrarily different elements (Ellis, 2002b; Squire, 1992), the learning of separate discrete episodes - what you saw across the field as your friend said gavagai or the particular color of tray that accompanied hearing chromium for the first time. There is benefit in being able to keep such episodic records distinct. But fast mappings are rough, ready, fragile, and, without reiteration, often transient. Repetition strengthens memories (Ebbinghaus, 1885), and there are clearly defined effects of frequency, spacing, and distribution of practice in the consolidation, elaboration, and explicit learning of foreign-language vocabulary, both naturalistically and from flash-cards, CALL programs, and the like (Ellis, 1995).

Repeated processing of a particular construction facilitates its fluency of subsequent processing, too, and these effects occur whether the learner is conscious of this processing or not. Your reading of the various occurrences of the word chunk in this chapter so far has primed the subsequent reading of this word and contributed to your lifetime usage practice of it, despite the fact that you cannot remember where in the text these occurrences fell. Although you are conscious of words in your visual focus, you definitely did not just now consciously label the word focus as a noun. On reading it, you were surely unaware of its nine alternative meanings, though in a different sentence you would instantly have brought a different meaning to mind. What happens to the other meanings? Psycholinguistic evidence demonstrates that some of them exist unconsciously for a few tenths of a second before your brain decides on the right one. Most words (over 80 percent in English) have multiple meanings, but only one of these can become conscious at a time. So your reading of focus has primed subsequent reading of that letter string (whatever its interpretation), and your interpretation of focus as a noun has primed that particular subsequent interpretation of it. In this way, particular constructions (e.g., [ba], ave, kept, man, dead boring, on the whole, I love you, [wAn] = 'one') with high token frequency are remembered better, recognized faster, produced more readily and otherwise processed with greater facility than low token frequency constructions (e.g., [za], aze, leapt, artichoke, sublimely boring, on the organelle, I venerate you, [wAn] = won) (see Ellis, 2002a for review). Each token of use thus strengthens the memory traces of a construction, priming its subsequent use and accessibility following the power law of practice relationship, whereby the increase in strength afforded by early increments of experience is greater than that from later additional practice. In these ways,

language learning involves considerable unconscious "tallying" (Ellis, 2002a) of construction frequencies, and language use requires exploitation of this implicit statistical knowledge (Bod, Hay, & Jannedy, 2003; Bybee & Hopper, 2001; Chater & Manning, 2006).

High token repetition is said to entrench constructions (Langacker, 1987), protecting them from change. Thus it is that it is the high frequency past tenses in English that are irregular (went, was, kept), their ready accessibility holding off the forces of regularization from the default paradigm (*goed, *beed, *keeped), whereas neighbors of lower frequency eventually succumb (with leaped starting to rival leapt in usage). Bybee (2008) calls this the conserving function of high token frequency. High token frequency also leads to autonomy, whereby creative constructions learned by rote may never be analyzed into their constituent units, e.g., learners may never have considered that gimme consists of give + me, nor the literal roots of a dicey situation. Finally, considerable practice with a particular token also results in automaticity of production and processes of reduction, assimilation, and lenition involving loss and overlap of gestures. A maxim of Bybee (2003, p. 112), on a variant of Hebb's "Cells that fire together wire together," is that "Items that are used together fuse together." The phenomenon is entirely graded - the degree of reduction is a continuous function of the frequency of the target word and the conditional probability of the target given the previous word and that of the target given the next word (Bybee & Hopper, 2001; Ellis, 2002a; Jurafsky et al., 2001). Such changes underpin grammaticalization in language change (Bybee, 2000; Croft, 2000).

In sum, although a particular construction can be roughly learned from a single exposure, multiple repetitions of that same token in different contexts are needed to enmesh and elaborate it into the meaning system - to turn it from a fast-mapped, tentative working hypothesis to a more complete, rich representation of the full connotations of a word (Carey & Bartlett, 1978). For example, it has been estimated that between 8 and 12 encounters are needed of a novel word in text before its meaning will be adequately comprehended from inference and its form and meaning retained (Horst, Cobb, & Meara, 1998; Saragi, Nation, & Meister, 1978). Multiple repetitions are also necessary for entrenched representation, ready accessibility, automatized processing, idiomatic autonomy, and fast, fluent, and phonetically reduced production.

Generalizing a construction from formula to limited scope pattern to productive abstract schema: The role of type frequency

The productivity of phonological, morphological, and syntactic patterns is a function of their type rather than token frequency (Bybee, 1995; Bybee & Hopper, 2001). Type frequency determines productivity because: (1) The more lexical items that are heard in a certain position in a construction, the less likely it is that the construction is associated with a particular lexical item and the more likely it is

that a general category is formed over the items that occur in that position. As novel exemplars are added in memory, they affect the category too, their features resonate with the whole population, adding their weight to the prototype, and stretching the bounds slightly in their direction. (2) The more items the category must cover, the more general are its criterial features and the more likely it is to extend to new items. (3) High type frequency ensures that a construction is used frequently, thus strengthening its representational schema and making it more accessible for further use with new items (Bybee & Thompson, 2000).

When a construction is variously experienced with different items occupying a position, it allows the parsing of its schematic structure. Having an initial formulaic exemplar of the Caused-Motion construction [Subj V Obj Prep Oblpath/loc], perhaps she pushed it down the road, subsequent experience of she pushed it ((up) the hill), she pushed it ((to) the service station), she pushed it ((to) the gas pump) allows identification of the common components, their structural commonalities, and their regularities of reference. Common items (pronouns like she, he, I, rather than complex noun phrases Mrs Struthers, the miraculous moose, the distressed driver, etc.; high frequency prepositions like to, up, down, etc., rather than complex locatives Alabama-way, paralleling the path of flight, etc.) repeat more in these slots and thus help to bring out the commonalities of the adjacent slot-fillers. Braine (1987) showed in experiments involving the learning of artificial languages that it was relatively easy to learn "categories" and rules for combining them, providing the "words" exemplifying these categories were either preceded or followed by a fixed item. Otherwise, the categories were difficult or impossible to learn. In natural language, it is the grammatical words that often serve as anchors like this. It is the closed class "little words," the grammatical functors, that have both the highest frequency in the language and the highest connectivity or degree. When the sequential co-occurrences of words in discourse are described in terms of graphs of word connections, mapping the interactions like social networks, the world wide web, or other complex systems, these graphs show so-called small-world properties of being highly clustered and richly interconnected (Ferrer i Cancho & Solé, 2001; Ferrer i Cancho, Solé, & Köhler, 2004). Despite having many thousands of nodes (the > 450,000 words populating a language), the average number of jumps in the path needed to get from any word to any other in this graph is remarkably small, at less than three. A small number of highly connected words allows these properties. And it is the function words, the prepositions, pronouns, determiners, etc., that do this, having both high token frequency and high degree of connectivity.3

So, these highest frequency components and chunks are the recurrent constituents of the construction that anchor its parse: as sub-unit constructions with high token frequency, they are recognized faster, produced more readily and otherwise processed with greater facility than low token frequency constructions, and, thus, they outline and bracket the schematic structure of the construction more readily. In 11-month-old infants, it is these frequently occurring functor forms that serve as a framework against which potential candidates for vocabulary membership may be identified and extracted from the speech stream (Shi et al.,

2006). In these ways, although verb islands predominate in seeding generalizations, patterns based on other high frequency lexical types, such as bound morphemes, auxiliary verbs and case-marking pronouns ("pronoun islands"), are also important in the parsing and identification of the schematic structure of constructions (Childers & Tomasello, 2001; McClure, Lieven, & Pine, in press; Pine, Lieven, & Rowland, 1998; Wilson, 2003).4 In growth, too, these are the high-degree nodes of the kernel lexicon of the language network, to which new sub-unit constructions are preferentially attached, allowing scale-free growth distribution according to the so-called Barbarási-Albert model (Barbarási & Albert, 1999; Ferrer i Cancho & Solé, 2001).

Chunking is a ubiquitous feature of human learning and memory. Chunking affords the ability to build up structures recursively, with the embedding of small chunks within larger ones leading to a hierarchical organization in nature (Simon, 1962, see particularly his parable of the two watchmakers, Hora and Tempus), in memory (Newell, 1990), and in the hierarchies and tree structures of grammar (Bybee, 2003; Ellis, 1996, 2003). In these ways, constituent structure is emergent, with constructions as grammatical schemata at all levels of specificity (from very specific (my chapter), through limited scope (my + NOUN), more general (POSSESSIVE + NOUN), to fully general (DETERMINER + NOUN)) emerging from the conspiracy of component constructions whose commonalities, in turn, are defined by their inclusion in the networks of other constructions (Bybee, 2003, 2008).

Functional motivations

Constructions are useful because of the symbolic functions that they serve. It is their communicative functions, semantic, pragmatic, or discursive, that motivate their learning. Goldberg (1995) claims that verb-centered constructions are more likely to be salient in the input because they relate to certain fundamental perceptual primitives, and, thus, that this construction of grammar involves in parallel the distributional analysis of the language stream and the analysis of contingent perceptual activity. It has been argued that basic level categories (e.g., hammer, dog) are acquired earlier and are more frequently used than superordinate (tools, canines) or subordinate (ball pein hammer, weimaraner) terms because, besides their frequency of use, this is the level at which the world is optimally split for function, the level where objects within the class share the same broad visual shape and motoric function, and, thus, where the categories of language most directly map onto perceptual form and motoric function (Lakoff, 1987; Rosch et al., 1976; Rosch, Varela, & Thompson, 1991). Goldberg extends this notion to argument structure more generally:

Constructions which correspond to basic sentence types encode as their central senses event types that are basic to human experience . . . that of someone causing something, something moving, something being in a state, someone possessing something, something causing a change of state or location, something undergoing a change of state or location, and something having an effect on someone. (Goldberg, 1995, p. 39)

Ninio (1999) and Goldberg, Casenhiser, and Sethuraman (2004) show for child language acquisition that individual "pathbreaking" semantically prototypic verbs form the seeds of verb-centered argument-structure patterns, with generalizations of the verb-centered instances emerging gradually as the verb-centered categories themselves are analyzed into more abstract argument structure constructions. The verb is a better predictor of sentence meaning than any other word in the sentence and plays the central role in determining the syntactic structure of a sentence. Since the same functional concerns motivate AL and L1 both, we should expect the same pattern for L2 and FL acquisition.

Learning categories and prototypes: From tokens to types

Because constructions are linguistic categories, we need to consider the psychology of concept and category learning (Ashby & Maddox, 2005; Cohen & Lefebvre, 2005): Humans can readily induce a category from experience of exemplars. Categories have graded structures (Rosch & Mervis, 1975). Rather than all instances of a category being "equal," certain instances are better exemplars than others. The prototype is the best example among the members of a category and serves as the benchmark against which the surrounding "poorer," more borderline instances are categorized; it combines the most representative attributes of that category in the conspiracy of its memorized exemplars. People have memory for the tokens they have seen before - previously experienced patterns are better judged than novel ones of equal distortion from the prototype. Although we don't go around consciously counting types and tokens, we nevertheless have very accurate implicit knowledge of the underlying distributions and their most usual settings. Similarity and frequency are, thus, important determinants of learning and generalization:

The more similar an instance is to the other members of its category and the less similar it is to members of contrast categories, the easier it is to classify (e.g., we better classify sparrows (or other average-sized, average-colored, average-beaked, average-featured specimens) as birds than we do birds with less common features or feature combinations, like geese or albatrosses) (Tversky, 1977). The greater the token frequency of an exemplar, the more it contributes to defining the category, and the greater the likelihood it will be considered the prototype of the category (e.g., sparrows are rated as highly typical birds because they are frequently experienced examples of the category birds). The unmarked forms of linguistic oppositions are more frequent than their marked forms (Greenberg, 1966). Token frequency is particularly important in this way in early and intermediate levels of learning, less so as learning approaches asymptote (Homa, Dunbar, & Nohre, 1991; Nosofsky, 1988).

There are important effects of presentation order in the implicit tallying that underlies category formation. In learning, the greater the variability of exemplars, the lower the rate of acquisition but the more robust the categorization/the less variability of distortion, the faster the category is learned (Posner & Keele, 1968, 1970). But it looks like there's an optimal balance to be had here. When

people try to teach a category to someone else explicitly, there is high agreement on the teaching sequences that are naturally adopted: The typical sequence starts with several ideal positive cases, followed by an ideal negative case and then borderline cases (Avrahami et al., 1997). Avrahami et al. tested to see whether this is indeed an optimal instruction sequence by comparing it with other orders that emphasized the full breadth of category from the outset. Exemplifying category breadth from the outset, borderline cases and central cases all, produced slower and less accurate explicit learning. For implicit learning of categories from exemplars, so, too, acquisition is optimized by the introduction of an initial, low-variance sample centered upon prototypical exemplars (Elio & Anderson, 1981, 1984). This low variance sample allows learners to get a "fix" on what will account for most of the category members. Then the bounds of the category can later be defined by experience of the full breadth of exemplars.

Form, function, and frequency: Zipfian family profiles

Goldberg, Casenhiser & Sethuraman (2004) tested the applicability of these generalizations to the particular case of children acquiring constructions. Phrasal form-meaning correspondences (e.g., X causes Y to move Zpath/loc [Subj V Obj Oblpath/loc]) do exist independently of particular verbs, but there is a close relationship between the types of verb that appear therein (in this case put, get, take, push, etc.). Furthermore, in natural language, the frequency profile of the verbs in the family follows a Zipfian profile (Zipf, 1935) whereby the highest frequency words accounted for the most linguistic tokens. Goldberg et al. demonstrated that in samples of child language acquisition, for a variety of constructions, there is a strong tendency for one single verb to occur with very high frequency in comparison to other verbs used (e.g., the [Subj V Obj Oblpath/loc] construction is exemplified in the children's speech by put 31% of the time, get 16%, take 10%, and do/pick 6%). This profile closely mirrored that of the mothers' speech to these children (with, e.g., put appearing 38% of the time in this construction that was otherwise exemplified by 43 different verbs). Ellis and Ferreira-Junior (Ellis & Ferreira-Junior, in press a, b) have replicated the Zipfian family profiles of these same constructions for the speech of naturalistic adult learners of English as a second language in the ESF project (Perdue, 1993).

The same can be seen in the constructions for compliments. Manes and Wolfson (1989) examined a corpus of 700 examples of compliments uttered in day-to-day interactions. Just three constructions accounted for 85% of these: [NP <is/looks> (really) ADJ] (53%), [I (really) <like/love> NP] (16%), and [PRO is (really) (a) ADJ NP] (15%). Eighty percent of these depended on an adjective to carry the positive semantic load. While the number of positive adjectives that could be used is virtually unlimited, in fact two-thirds of all adjectival compliments in the corpus used only five adjectives: nice (23%), good (20%), pretty (9%), beautiful (9%), and great (6%). Non-adjectival compliments were focused on a handful of semantic-ally positive verbs, with like and love accounting for 86%.

Thus, it appears that in natural language, at least for the constructions considered in this way so far, tokens of one particular verb account for the lion's share of instances of argument frames, and that the pathbreaking verb for each is the one with the prototypical meaning from which that construction is derived. How about that? As Morales and Taylor (2007) put it: "Language is exquisitely adaptive to the learning capabilities of its users." The natural structure of natural language seems to provide exactly the familial type:token frequency distribution to ensure optimized acquisition of linguistic constructions as categories.

Optimizing instruction samples for construction learning

What are the implications for instruction using curriculum-driven input samples? What we know about category formation suggests that these type:token frequency considerations should apply here too. Optimal acquisition should occur when the central members of the category are presented early and often.

For syntactic constructions, Goldberg, Casenhiser, and Sethuraman (2004) tested whether, when training novel patterns (a construction of the form [Subj Obj V-o] signaling the appearance of the subject in a particular location, for example, the king the ball moopo-ed) exemplified by five different novel verbs, it is better to train with relatively balanced token frequencies (4-4-4-2-2) or with a family frequency profile where one exemplar had a particularly high token frequency (8-2-2-2-2). Undergraduate native speakers of English learned this novel construction from three minutes of training using videos. They were then tested for the generalization of the semantics of this construction to novel verbs and new scenes. Learners in the high token frequency condition showed significantly better learning than those in the balanced condition, a finding Goldberg (Goldberg, 2006; 2008) has now observed in studies of child acquisition too.

For morphological constructions, Bybee (2008) analyzed the ways that natural frequency skewing affects the acquisition of verbal inflexions. The most frequent forms of a paradigm (third person/first person singular) either have no affix or a short affix, and the other forms of the paradigm can typically be derived from them. Thus, she argues, the high token frequency forms of the paradigm are the anchoring points of the other forms. Lower frequency forms are analyzed and learned in terms of these more robust forms creating a relationship of dependency.

Frequency variation is ubiquitous across natural languages. Morales and Taylor (2007) present connectionist simulations evidencing how learning can be enhanced through frequency variation: training samples where there were variable numbers of tokens per type produced more accurate and more economical learning than did training with more uniform frequency profiles.

There is clearly a need to extend these initial studies to explore more thoroughly the sampling of exemplars of a wide range of second language constructions for optimal acquisition, but in the interim, the best informed practice is to introduce a new construction using an initial, low-variance sample centered upon prototypical exemplars to allow learners to get a "fix" on the central tendency that will

account for most of the category members. Tokens that are more frequent have stronger representations in memory and serve as the analogical basis for forming novel instances of the category.

Corpus and cognitive linguistic analyses are essential to the determination of which constructions of differing degrees of schematicity are worthy of instruction, their relative frequency, and their best (= central and most frequent) examples for instruction and assessment (Biber, Conrad, & Reppen, 1998; Biber et al., 1999). Gries (2008) describes how the three basic methods of corpus linguistics (frequency lists, concordances, and collocations) inform the instruction of second language constructions. Achard (2008), Tyler (2008), Robinson and Ellis (2008a) and other readings in Robinson and Ellis (2008b) show how an understanding of the item-based nature of construction learning inspires the creation and evaluation of instructional tasks, materials, and syllabi, and how cognitive linguistic analyses can be used to inform learners how constructions are conventionalized ways of matching certain expressions to specific situations and to guide instructors in precisely isolating and clearly presenting the various conditions that motivate speaker choice.

Tuning the System: Frequency and the Attainment of Nativelike Fluency and Selection

Language is fundamentally probabilistic: every piece is ambiguous. Each of these example formulas ("One, two, three," "Once upon a time," "Wonderful!," "Won the battle, lost the war") begins with the sound "wAn". At this point, what should the appropriate interpretation be? A general property of human perception is that when a sensation is associated with more than one reality, unconscious processes weigh the odds, and we perceive the most probable thing. Psycholinguistic analyses demonstrate that fluent language users are sensitive to the relative probabilities of occurrence of different constructions in the speech stream (Bod, Hay, & Jannedy, 2003; Bybee & Hopper, 2001; Chater & Manning, 2006; Ellis, 2002a, 2002b; Jurafsky & Martin, 2000). Since learners have experienced many more tokens of "one" than they have "won," in the absence of any further information, they typically favor the unitary interpretation over that involving gain or advantage. But they need to be able to suppress this interpretation in a context of "Alice in wAn . . ." Learners have to figure language out: their task is, in essence, to learn the probability distribution P(interpretation\cue, context), the probability of an interpretation given a formal cue, a mapping from form to meaning conditioned by context. This figuring is achieved, and communication optimized, by implicit tallying of the frequency, recency, and context of constructions.

This incidental learning from usage allows language users to be rational in the sense that their mental models of the way language works are optimal given their linguistic experience to date (Ellis, 2006b). The words that they are likely to hear next, the most likely senses of these words, the linguistic constructions

they are most likely to utter next, the syllables they are likely to hear next, the graphemes they are likely to read next, the interpretations that are most relevant, and the rest of what's coming next across all levels of language representation, are made more readily available to fluent speakers by their language processing systems. Their unconscious language representations are adaptively probability-tuned to predict the linguistic constructions that are most likely to be relevant in the ongoing discourse context, optimally preparing them for comprehension and production. With practice comes modularization too, the development of autonomous specialist systems for different aspects of language processing. These "zombie agents" are independent - experience of reading a word facilitates subsequent reading of that word, experience of speaking a word facilitates subsequent speaking of that word, but cross-modal priming effects are null or slight in fluent speakers. So reading practice tallies the reading system, speaking practice tunes the speaking system, etc. Fluency in each separate module requires its own usage practice (see Gatbonton & Segalowitz, 2005 for communicative approaches designed to engender this). This specificity of practice gain from different forms of processing underlies many failures of learning and generalization as summarized in the Transfer-Appropriate Processing (TAP) framework (Morris, Bransford, & Franks, 1977). Lightbown (2007) reviews the implications of TAP for L2 instruction, how there is a need to increase the number of settings and processing types in which learners encounter the material they need to learn.

Just as extensive sampling is required for nativelike fluency, so it is, too, for nativelike selection. Many of the forms required for idiomatic use are, nevertheless, of relatively low frequency, and the learner thus needs a large input sample just to encounter them. More usage still is required to allow the tunings underpinning nativelike use of collocation - something which even advanced learners have particular difficulty with. Hence the emphasis on the representative samples necessary for English for Academic and Specific Purposes (EAP/ESP) (e.g., Swales, 1990). Linguists interested in the description of language (e.g., British National Corpus, 2006) have come to realize that really large corpora are necessary to describe it adequately - 100 million words is just a start, and each genre, dialect, and type requires its own properly targeted sampling. Child language researchers have also begun the relevant power analyses to explore the relations between construction frequency and sample size for accurate description, reaching the conclusion that for many constructions of interest, dense corpora are an absolute necessity (Tomasello & Stahl, 2004). So, too, in learners' attainment of fluent language processing, whether in L1 or AL, there is no substitute for usage, lots of appropriate usage.

Becoming fluent requires a sufficient sample of needs-relevant authentic input for the necessary implicit tunings to take place. The "two puzzles for linguistic theory," nativelike selection and nativelike fluency (Pawley & Syder, 1983), are less perplexing when considered in these terms of frequency and probability. There's a lot of tallying to be done here. The necessary sample is certainly to be counted in terms of thousands of hours on task.

The Language Calculator Has No "Clear" Button

A final implication of language acquisition as estimation relates again to sampling history, this time in terms of the difference between first langauge (L1) and adult language (AL) acquisition. AL learners are distinguished from infant L1 acquirers by the fact that they have previously devoted considerable resources to the estimation of the characteristics of another language - the native tongue in which they have considerable fluency (and any others subsequently acquired). Since they are using the same apparatus to survey their additional language too, their computations and induction are often affected by transfer, with L1-tuned expectations and selective attention (Ellis, 2006c) blinding the computational system to aspects of the AL sample, thus rendering biased estimates from naturalistic usage and the limited endstate typical of L2A. These effects have been explored within the traditions of contrastive analysis (James, 1980), language transfer (Odlin, 1989), and more recently within cognitive linguistics (Robinson & Ellis, 2008b). From our L1 we learn how language frames the world and how to use it to describe action therein, focusing our listeners' attention appropriately. Cognitive linguistics is the analysis of these mechanisms and processes that underpin what Slobin (1996) called "thinking for speaking." But learning an AL requires "rethinking for speaking" (Robinson & Ellis, 2008a). In order to counteract the L1 biases to allow estimation procedures to optimize induction, all of the AL input needs to be made to count (as it does in L1A), not just the restricted sample typical of the biased intake of L2A. Certain types of form-focused instruction can help to achieve this by recruiting learners' explicit, conscious processing to allow them to consolidate unitized form-function bindings of novel AL constructions (Ellis, 2005). Once a construction has been represented in this way, so its use in subsequent processing can update the statistical tallying of its frequency of usage and probabilities of form-function mapping.

Language is its dynamic usage. It ever changes. For learners and linguists alike, its sum can only ever be estimated from limited samples of experience. Understanding the units and the processes of their estimation helps guide theory and application, learning and instruction.

I thank Patsy Lightbown for constructive comments on a previous draft of this chapter.

1 Depending as well, of course, upon degree of shared context, embodiment, attention, cultural understandings, communicative intent, etc.

2 Whereby the frequency of the tokens of verbs seeding a construction type decays as a power function of their rank (Zipf, 1935).

3 The high token frequency of these items, though, means that in the course of language use, they have become phonetically eroded. These items lack perceptual salience and are consequently difficult to perceive from bottom-up, data-driven sources alone, a

factor which makes their second language acquisition difficult (Ellis, 2006c, 2008). They are also semantically light, abstract, and often homonymous, factors also making them difficult to acquire (Ellis, 2008). So it is the semantically rich and basic verbs which seed the constructions, these other grammatical functors making their contribution by marking the commonalities of the parse pattern.

4 Again, emphasizing the proviso concerning their low salience, low contingency, and abstractness.

REFERENCES

Achard, M. (2008). Cognitive pedagogical grammar. In P. Robinson & N. C. Ellis (eds.), Handbook of cognitive linguistics and second language acquisition (pp. 432-55). London: Routledge.

Ashby, E. G. & Maddox, W. T. (2005). Human category learning. Annual Review of Psychology 56, 149-78.

Avrahami, J., Kareev, Y., Bogot, Y., Caspi, R., Dunaevsky, S., & Lerner, S. (1997). Teaching by examples: Implications for the process of category acquisition. The Quarterly Journal of Experimental Psychology: Section A 50, 3, 586-606.

Barbarasi, A.-L. & Albert, R. (1999). Emergence of scaling in random networks. Science 286, 509-11.

Biber, D., Conrad, S., & Reppen, R. (1998). Corpus linguistics: Investigating language structure and use. New York: Cambridge University Press.

Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spoken and written English. Harlow, UK: Pearson Education.

Bod, R., Hay, J., & Jannedy, S. (eds.) (2003). Probabilistic linguistics. Cambridge, MA: MIT Press.

Braine, M. D. (1987). What is learned in acquiring word classes - a step towards acquisition theory. In B. MacWhinney (ed.), Mechanisms of language acquisition (pp. 65-87). Hillsdale, NJ: Lawrence Erlbaum Associates.

British National Corpus (2006). http://www.natcorp.ox.ac.uk/

Bybee, J. (1995). Regular morphology and the lexicon. Language and Cognitive Processes 10, 425-55.

Bybee, J. (2000). Mechanisms of change in grammaticalization: The role of frequency. Unpublished manuscript.

Bybee, J. (2003). Sequentiality as the basis of constituent structure. In T. Givon & B. F. Malle (eds.), The evolution of language out of pre-language (pp. 109-32). Amsterdam: John Benjamins.

Bybee, J. (2008). Usage-based grammar and second language acquisition. In P. Robinson & N. C. Ellis (eds.), Handbook of cognitive linguistics and second language acquisition (216-36). London: Routledge.

Bybee, J. & Hopper, P. (eds.) (2001). Frequency and the emergence of linguistic structure. Amsterdam: Benjamins.

Bybee, J. & Thompson, S. (2000). Three frequency effects in syntax. Berkeley Linguistic Society 23, 65-85.

Carey, S. & Bartlett, E. (1978). Acquiring a single new word. Proceedings of the Stanford Child Language Conference / Papers and Reports on Child Language Development 15, 1729.

Chater, N. & Manning, C. (2006). Probabilistic models of language processing and acquisition. Trends in Cognitive Science 10, 335-44.

Childers, J. B. & Tomasello, M. (2001). The role of pronouns in young children's acquisition of the English transitive construction. Developmental Psychology 37, 739-48.

Cohen, H. & Lefebvre, C. (eds.) (2005). Handbook of categorization in cognitive science. Amsterdam: Elsevier.

Croft, W. (2000). Explaining language change: An evolutionary approach. London: Longman.

Croft, W. & Cruise, A. (2004). Cognitive linguistics. Cambridge: Cambridge University Press.

Ebbinghaus, H. (1885/1913). Memory: A contribution to experimental psychology, trans. H. A. Ruger & C. E. Bussenius. New York: Teachers College, Columbia.

Elio, R. & Anderson, J. R. (1981). The effects of category generalizations and instance similarity on schema abstraction. Journal of Experimental Psychology: Human Learning & Memory 7, 6, 397-417.

Elio, R. & Anderson, J. R. (1984). The effects of information order and learning mode on schema abstraction. Memory & Cognition 12, 1, 20-30.

Ellis, N. C. (1995). The psychology of foreign language vocabulary acquisition: Implications for CALL. Computer Assisted Language Learning 8, 2-3.

Ellis, N. C. (1996). Sequencing in SLA: Phonological memory, chunking, and points of order. Studies in Second Language Acquisition 18, 1, 91-126.

Ellis, N. C. (1998). Emergentism, connectionism and language learning. Language Learning 48, 4, 631-64.

Ellis, N. C. (2002a). Frequency effects in language processing: A review with implications for theories of implicit and explicit language acquisition. Studies in Second Language Acquisition 24, 2, 143-88.

Ellis, N. C. (2002b). Reflections on frequency effects in language processing. Studies in Second Language Acquisition 24, 2, 297-339.

Ellis, N. C. (2003). Constructions, chunking, and connectionism: The emergence of second language structure. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 33-68). Oxford: Blackwell.

Ellis, N. C. (2005). At the interface: Dynamic interactions of explicit and implicit language knowledge. Studies in Second Language Acquisition 27, 305-52.

Ellis, N. C. (2006a). Cognitive perspectives on SLA: The Associative Cognitive CREED. AILA Review 19, 100-21.

Ellis, N. C. (2006b). Language acquisition as rational contingency learning. Applied Linguistics 27, 1, 1-24.

Ellis, N. C. (2006c). Selective attention and transfer phenomena in SLA: Contingency, cue competition, salience, interference, overshadowing, blocking, and perceptual learning. Applied Linguistics 27, 2, 1-31.

Ellis, N. C. (2007). Dynamic Systems and SLA: The wood and the trees. Bilingualism: Language & Cognition 10, 23-5.

Ellis, N. C. (2008). The dynamics of second language emergence: Cycles of language use, language change, and language acquisition. Modern Language Journal 92, 2, 232-49.

Ellis, N. C. & Ferreira-Junior, F. (in press, a). Construction learning as a function of frequency, frequency distribution, and function. Modern Language Journal.

Ellis, N. C. & Ferreira-Junior, F. (in press, b). Constructions and their acquisition: Islands and the distinctiveness of their occupancy. Annual Review of Cognitive Linguistics.

Ellis, N. C. & Larsen Freeman, D. (2006). Language emergence: Implications for Applied Linguistics (Introduction to the Special Issue). Applied Linguistics 27, 4, 558-89.

Ferrer i Cancho, R. & Solé, R. V. (2001). The small world of human language. Proceedings of the Royal Society of London, B 268, 2261-5.

Ferrer i Cancho, R., Solé, R. V., & Köhler, R. (2004). Patterns in syntactic dependency networks. Physical Review E69, 0519151-8.

Gatbonton, E. & Segalowitz, N. (2005). Rethinking communicative language teaching: A focus on ACCESS to fluency. Canadian Modern Language Review 61, 325-53.

Goldberg, A. E. (1995). Constructions: A construction grammar approach to argument structure. Chicago: University of Chicago Press.

Goldberg, A. E. (2003). Constructions: a new theoretical approach to language. Trends in Cognitive Science 7, 219-24.

Goldberg, A. E. (2006). Constructions at work: The nature of generalization in language. Oxford: Oxford University Press.

Goldberg, A. E. (2008). The language of constructions. In P. Robinson & N. C. Ellis (eds.), A handbook of cognitive linguistics and SLA (pp. 197-215). Mahwah, NJ: Lawrence Erlbaum.

Goldberg, A. E., Casenhiser, D. M., & Sethuraman, N. (2004). Learning argument structure generalizations. Cognitive Linguistics 15, 289-316.

Greenberg, J. (1966). Language universals. The Hague: Mouton.

Gries, S. T. (2008). Corpus-based methods in analyses of SLA data. In P. Robinson & N. C. Ellis (eds.), A handbook of cognitive linguistics and SLA (pp. 406-31). Mahwah, NJ: Erlbaum.

Homa, D., Dunbar, S., & Nohre, L. (1991). Instance frequency, categorization, and the modulating effect of experience. Journal of Experimental Psychology: Learning, Memory, and Cognition 17, 444-58.

Horst, M., Cobb, T., & Meara, P. (1998). Beyond A Clockwork Orange: Acquiring second-language vocabulary through reading. Reading in a Foreign Language 11, 207-23.

James, C. (1980). Contrastive analysis. London: Longman.

Jurafsky, D., Bell, A., Gregory, M., & Raymond, W. D. (2001). Probabilistic relations between words: Evidence from reduction in lexical production. In J. Bybee & P. Hopper (eds.), Frequency and the emergence of linguistic structure (pp. 229-54). Amsterdam: Benjamins.

Jurafsky, D. & Martin, J. H. (2000). Speech and language processing: An introduction to natural language processing, speech recognition, and computational linguistics. Englewood Cliffs, NJ: Prentice-Hall.

Lakoff, G. (1987). Women, fire, and dangerous things: What categories reveal about the mind. Chicago: University of Chicago Press.

Langacker, R. W. (1987). Foundations of cognitive grammar, vol. 1: Theoretical prerequisites. Stanford, CA: Stanford University Press.

Lightbown, P. M. (2007). Transfer appropriate processing in classroom second language acquisition. In Z.-H. Han & E. S. Park (eds.), Understanding second language process (pp. 27-44). Clevedon, UK: Multilingual Matters.

Manes, J. & Wolfson, N. (1989). The compliment formula. In F. Coulmas (ed.), Conversational routine (pp. 116-32). The Hague: Mouton.

McClure, C., Lieven, E., & Pine, J. M. (in press). Investigating the abstractness of children's early knowledge of argument structure. Journal of Child Language.

Morales, F. & Taylor, J. R. (2007). Learning from relative frequency. Available as LAUD (Linguistic Agency, University of Duisburg) preprint, paper no. 690.

Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning & Verbal Behavior 16, 519-33.

Newell, A. (1990). Unified theories of cognition. Cambridge, MA: Harvard University Press.

Ninio, A. (1999). Pathbreaking verbs in syntactic development and the question of prototypical transitivity. Journal of Child Language 26, 619-53.

Nosofsky, R. M. (1988). Similarity, frequency, and category representations. Journal of Experimental Psychology: Learning, Memory, and Cognition 14, 54-65.

Odlin, T. (1989). Language transfer. New York: Cambridge University Press.

Pawley, A. & Syder, F. H. (1983). Two puzzles for linguistic theory: Nativelike selection and nativelike fluency. In J. C. Richards & R. W. Schmidt (eds.), Language and communication (pp. 191-225). London: Longman.

Perdue, C. (ed.) (1993). Adult language acquisition: Crosslinguistic perspectives. Cambridge: Cambridge University Press.

Pine, J. M. & Lieven, E. V. (1993). Reanalyzing rote-learned phrases: Individual differences in the transition to multi-word speech. Journal of Child Language 20, 551-71.

Pine, J. M., Lieven, E. V., & Rowland, C. F. (1998). Comparing different models of the development of the English verb category. Linguistics 36, 807-30.

Posner, M. I. & Keele, S. W. (1968). On the genesis of abstract ideas. Journal of Experimental Psychology 77, 353-63.

Posner, M. I. & Keele, S. W. (1970). Retention of abstract ideas. Journal of Experimental Psychology 83, 304-8.

Robinson, P. & Ellis, N. C. (2008a). Conclusion: Cognitive linguistics, second language acquisition and L2 instruction: Issues for research. In P. Robinson & N. C. Ellis (eds.), Handbook of cognitive linguistics and second language acquisition (pp. 489-545). London: Routledge.

Robinson, P. & Ellis, N. C. (eds.) (2008b). A handbook of cognitive linguistics and second language acquisition. London: Routledge.

Rosch, E. & Mervis, C. B. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General 104, 192-233.

Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M., & Boyes-Braem, P. (1976). Basic objects in natural categories. Cognitive Psychology 8, 382-439.

Rosch, E., Varela, F., & Thompson, E. (1991). The embodied mind. Boston, MA: MIT Press.

Saragi, T., Nation, I. S. P., & Meister, G. F. (1978). Vocabulary learning and reading. System 6, 72-128.

Shi, R., Cutler, A., Werker, J. E., & Cruikshank, M. (2006). Frequency and form as determinants of functor sensitivity in English-acquiring infants. Journal of the Acoustic Society of America 119, EL61-EL67.

Simon, H. A. (1962). The sciences of the artificial. Cambridge, MA: MIT Press.

Slobin, D. I. (1996). From "thought and language" to "thinking for speaking." In J. J. Gumperz & S. C. Levinson (eds.), Rethinking linguistic relativity (pp. 70-96). Cambridge: Cambridge University Press.

Squire, L. R. (1992). Memory and the hippocampus: A synthesis from findings with rats, monkeys, and humans. Psychological Review 99, 195-231.

Swales, J. M. (1990). Genre analysis: English in academic and research settings. Cambridge: Cambridge University Press.

Taylor, J. R. (2002). Cognitive grammar. Oxford: Oxford University Press.

Tomasello, M. (ed.) (1998). The new psychology of language: Cognitive and functional approaches to language structure. Mahwah, NJ: Lawrence Erlbaum.

Tomasello, M. (1992). First verbs: A case study of early grammatical development. New York: Cambridge University Press.

Tomasello, M. (2003). Constructing a language. Boston, MA: Harvard University Press.

Tomasello, M. & Stahl, D. (2004). Sampling children's spontaneous speech: How much is enough? Journal of Child Language 31, 101-21.

Tversky, A. (1977). Features of similarity. Psychological Review 84, 327-52.

Tyler, A. (2008). Cognitive linguistics and second language instruction. In P. Robinson & N. C. Ellis (eds.), Handbook of cognitive linguistics and second language acquisition (pp. 456-88). London: Routledge.

Wilson, S. (2003). Lexically specific constructions in the acquisition of inflection in English. Journal of Child Language 30, 75-115.

Wray, A. (2002). Formulaic language and the lexicon. Cambridge: Cambridge University Press.

Zipf, G. K. (1935). The psycho-biology of language: An introduction to dynamic philology. Cambridge, MA: MIT Press.

Part IV Program Design

The Handbook of Language Teaching Editedby MicheelH.LnngandCethermeL.Doughty © 2009 Blackwell PublishDgLtd. BBN: 978-1-408-15789-5

10 Bilingual and Immersion Programs

JIM CUMMINS

Introduction

The term bilingual education refers to an organized and planned program that uses two (or more) languages of instruction. The central defining feature of bilingual programs is that the languages are used to teach subject matter content rather than just the languages themselves. Bilingual instruction can be implemented at any grade or age level, ranging from pre-school through university or college. Bilingual education can be traced back to Greek and Roman times and currently a large majority of countries throughout the world offer some form of bilingual education either in public or private school settings (Cummins & Hornberger, 2008).

The goals of bilingual programs vary widely across contexts. Some programs aim to develop proficiency in two languages; others do not. For example, the most common form of bilingual education for linguistic minority students in the United States during the past 40 years, transitional bilingual education, aims only to promote students' proficiency in English. When it is assumed that students have attained sufficient proficiency in the school language to follow instruction in that language, home language instruction is discontinued and students are transitioned into mainstream classes taught exclusively in English.

The term "immersion" is used in two very different ways in educational discourse. In the first sense, immersion programs are organized and planned forms of bilingual education in which students are "immersed" in a second-language instructional environment with the goal of developing proficiency in two languages. First-language instruction is typically introduced within a year or two of the start of the program and forms an integral part of the overall plan. In its second sense, the term "immersion" refers to the immersion of immigrant or minority language children in a classroom environment where instruction is conducted exclusively through their second (or third) language (frequently the dominant language of the society or a global language of wider communication). The intent is to develop proficiency in the language of instruction. Such programs vary in the amount of support they provide to enable students to acquire

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

proficiency in the language of instruction - in some cases extensive support is provided by specialist language teachers, but in other cases students are left to "sink or swim." This second sense of the term "immersion" reflects popular usage but, as described below, is diametrically opposed to the conceptualization of immersion education within the educational research community. In the remainder of this paper, "immersion education" will be used to describe the first sense of the term - a planned program aimed at bilingual development - while "immersion" or "submersion" will be used to refer to the exclusive use of students' second language (L2) as a medium of instruction with the goal of developing proficiency only in the language of instruction.

The term "immersion education" came to prominence in Canada during the 1960s to describe innovative programs in which the French language was used as an initial medium of instruction for elementary school students whose home language was English. Immersion programs explicitly aim to promote fluency and literacy in students' first and second languages (L1 and L2). These programs were originally implemented at the Kindergarten level (age 5 - termed early immersion) but were later also implemented in Grades 4 or 5 (termed middle immersion) and Grades 7 or 8 (termed late immersion). About 300,000 Canadian students currently participate in immersion programs. This represents about 6 percent of the national school population. In early immersion programs, students whose L1 is English are initially "immersed" in a French language school environment for two to three years prior to the introduction of formal teaching of English. Instruction through French is designed specifically to enable students to gain access to academic content despite their initially low levels of French proficiency. English language arts are typically introduced in Grade 2 and English is used as a medium for teaching other subject matter (e.g., science, math, social studies) by Grades 3 or 4. Generally, by Grade 4, 50 percent of the instructional time is spent through each language.

Johnson and Swain (1997) point out that there is nothing new in the phenomenon of teaching students through the medium of a second language. In fact, throughout the history of formal education, the use of an L2 as a medium of instruction has been the rule rather than the exception. The Canadian French immersion programs, however, were the first to articulate a set of pedagogical principles underlying immersion education (Lambert & Tucker, 1972). They were also the first to be subjected to intensive long-term research evaluation, although some large-scale research had been undertaken in other contexts prior to the Canadian experience (e.g., Macnamara, 1966 in Ireland, and Malherbe, 1946 in South Africa).

Johnson and Swain (1997) summarize eight core features of immersion programs:

• The L2 is a medium of instruction.

• The immersion curriculum parallels the local L1 curriculum.

• Overt support exists for the L1.

• The program aims for additive bilingualism where students "add" L2 proficiency while continuing to develop their L1.

• Exposure to the L2 is largely confined to the classroom.

• Students enter with similar (and limited) levels of L2 proficiency.

• The teachers are bilingual.

• The classroom culture is that of the local L1 community.

It is clear that immersion education represents a carefully planned program that goes far beyond simply instructing students through a second language. In practice, however, when applied to immigrant and minority language students, the term "immersion" is frequently used to refer to programs that fall far short of the conditions specified by Johnson and Swain.

The Sociopolitical Context of Bilingual Education

There are an estimated 5,000 languages spoken in the world's 200 or so sovereign states. Thus, the majority of states encompass multiple languages within their boundaries. About two-thirds of all children in the world grow up in a bilingual or multilingual environment. To illustrate, 90 million of China's more than one billion population belong to a national minority and most of these minority groups speak languages other than Mandarin, the official language of the country. Linguistic diversity also exists among the Han majority group as a result of multiple "dialects" that represent mutually unintelligible spoken languages, even though all share the same writing system. Singapore, Switzerland, India, and most African countries are just a few other examples of countries that recognize multiple national languages and which regulate the status and use of these languages in education, government, and other social arenas.

In the current era of globalization, with unprecedented human mobility and social interchange across cultural and linguistic boundaries, processes of language learning (and language loss) are apparent in societies around the world. Government policies attempt to influence these processes by supporting the teaching of certain languages in schools and, in some cases, by actively discouraging the maintenance of other languages, usually the languages of subordinated groups within the society. Bilingual programs have emerged in recent years as a viable option for governments and communities interested in promoting more effective learning of socially valued languages and/or maintaining languages that are endangered, such as many indigenous languages in North America.

Despite their utility as a tool for language planning, bilingual programs have also aroused considerable controversy in some countries. Opposition to bilingual education tends to be highly selective. It focuses only on the provision of L1 instruction to students from minority or socially subordinated groups (e.g., Spanish-speakers in the United States, Turkish-speakers in Germany, etc.). There is virtually no controversy about the provision of bilingual programs or second-language immersion programs to children of the dominant group(s) in society. For example, French immersion programs for anglophone students in Canada have been minimally controversial during the past 40 years because they serve

the interests of the dominant group. Similarly in Europe and the United States, when the target students are from the dominant group, instruction through the medium of a second language is seen as educational enrichment - a more efficient way of teaching additional languages and adding to the cultural capital of the student.

Thus, opposition to bilingual education is fueled primarily by ideological concerns relating to diversity and power. Use of a language as a medium of instruction confers recognition, status, and often economic benefits (e.g., teaching positions) on speakers of that language. Consequently, bilingual education is not simply a politically-neutral instructional innovation. It is also a sociopolitical phenomenon that is implicated in the ongoing competition between social groups for material and symbolic resources.

Types, Goals, and Participants

Typologies of bilingual education focus on characteristics of students in the program, the goals of the program, and organizational structures. The more important distinctions are outlined below:

• Majority/minority languages or students. These terms refer to whether a language is the language of the numerically dominant group in a society or that of a numerically non-dominant group.

• Dominant/subordinated students or groups. These terms are often used interchangeably with majority/minority but they refer explicitly to power and status relations between societal groups rather than to the numerical size of the groups. Minoritized is sometimes used interchangeably with subordinated (Skutnabb-Kangas & McCarty, 2008).

• Enrichment/remedial programs. The term enrichment bilingual education refers to programs that aim to enrich students' educational experience by strongly promoting bilingualism and biliteracy. French immersion programs in Canada and dual language programs involving both majority and minority language students in the United States are examples of enrichment programs. Dual language programs are also termed two-way immersion programs. Remedial programs, by contrast, aim to remediate or compensate for presumed linguistic deficits that bilingual children bring to school.

• Maintenance/transitional programs. Maintenance programs aim to help language minority students maintain and develop their proficiency in their home language while transitional programs are designed as a temporary bridge to instruction exclusively through the dominant language of the school and society.

• Late-exit/Early-exit programs. Transitional bilingual programs are often distinguished according to the grade level at which students transition from the bilingual program into mainstream monolingual classes. Early-exit programs are typically motivated by the assumption that students will benefit by transitioning from the bilingual program into the mainstream program as

rapidly as possible. The transition usually occurs by Grade 2 or 3. By contrast, late-exit programs, also known as developmental programs in the United States, transition students close to the end of elementary school (Grade 5 or 6). The assumption is that academic outcomes in both the majority language and students' L1 will benefit from strong promotion of both languages. • Immersion/submersion programs. Immersion programs, as conceptualized within the educational research community, are a form of bilingual education that immerse students in a second-language instructional environment for between 50 and 100 percent of instructional time with the goal of developing fluency and literacy in both languages. Students may be either from the dominant linguistic group or members of an ethnocultural or indigenous community whose heritage language is one of the languages of instruction. In this latter case, the goal is usually to maintain or revitalize an endangered language. Submersion programs, by contrast, provide 100 percent of instruction through the dominant language (students' L2); teachers typically do not understand students' L1, and few instructional supports are available to help students understand instruction or express themselves through either L1 or L2. These programs are also termed sink-or-swim programs. The term "structured immersion" has been used in the United States (e.g., by Rossell & Baker, 1996) to refer to English instructional programs that provide comprehension supports (including the possibility of some very limited use of students' L1) to enable English language learners to understand instruction. These programs are dismissed by advocates of bilingual education as simply another form of submersion (e.g., Skutnabb-Kangas & McCarty, 2008).

Bilingual programs can also be categorized according to who participates in the program. Four broad overlapping categories can be distinguished. The first category involves programs intended for indigenous students (e.g., Maori students in New Zealand) and those from nationally recognized minority groups (e.g., students of Breton heritage in France or of Basque heritage in the Basque Autonomous Community in Spain). Typically these programs are intended to either maintain or revitalize the minority language.

The second category involves students from the dominant or majority group. The goal is to develop bilingual and biliteracy skills among these students. Examples are the Canadian French immersion programs and dual language programs in the United States that enroll both majority and minority language students.

The third category involves students who come from immigrant communities. Most of these programs are transitional and remedial in nature with the primary goal of supporting students' academic development in the majority language.

The final category of bilingual education programs involves children who are deaf or hard-of-hearing. These programs use a natural sign language, such as American Sign Language (ASL), as a medium of instruction together with the dominant language of the society, frequently with a focus on the written form of this language. Bilingual-bicultural programs are common and well accepted in Scandinavian countries such as Sweden and Denmark (Mahshie, 1995) but are

still struggling to gain acceptance in North America and many other parts of the world (Small & Mason, 2008).

General Outcomes of Bilingual Education Programs

Formal academic research has been conducted on bilingualism and bilingual education since the 1920s and a voluminous literature has accumulated on these topics (e.g., August & Shanahan, 2006; Cummins, 2001; Garcia & Baker, 2007; Genesee et al., 2006; May, 2008). At this point, considerable confidence can be placed in some general conclusions about the outcomes of bilingual education; specifically, the research evidence is clear that for both minority and majority language students, well-implemented bilingual programs are an effective way of promoting proficiency in two languages (e.g., August & Shanahan, 2006).

A finding common to all forms of bilingual education is that spending instructional time through two languages entails no long-term adverse effects on students' academic development in the majority language. This pattern emerges among both majority and minority language students, across widely varying socio-linguistic and sociopolitical contexts, and in programs with very different organizational structures. Three additional outcomes of bilingual programs can be highlighted.

1 Significant positive relationships exist between the development of academic skills in first and second languages

In order to account for these findings and the fact that instruction through a minority language entailed no adverse consequences for students' academic development in the majority language, Cummins (1979, 1981) proposed the "interdependence hypothesis." This hypothesis was formally expressed in the following way:

To the extent that instruction in Lx is effective in promoting proficiency in Lx, transfer of this proficiency to Ly will occur provided there is adequate exposure to Ly (either in school or environment) and adequate motivation to learn Ly. (1981, p. 29)

In concrete terms, what this hypothesis means is that in, for example, a Basque-Spanish bilingual program in the Basque Country in Spain, Basque instruction that develops Basque reading and writing skills is not just developing Basque skills, it is also developing a deeper conceptual and linguistic proficiency that is strongly related to the development of literacy in the majority language (Spanish). In other words, although the surface aspects (e.g., pronunciation, fluency, etc.) of different languages are clearly separate, there is an underlying conceptual proficiency, or knowledge base, that is common across languages. This common

underlying proficiency (or what Genesee et al. (2006) call a cross-linguistic reservoir of abilities) makes possible the transfer of concepts, literacy skills, and learning strategies from one language to another. This is true even for languages that are dissimilar (e.g., American Sign Language and English, Spanish and Basque; Dutch and Turkish). The transfer of skills, strategies, and knowledge explains why spending instructional time through a minority language entails no adverse consequences for the development of the majority language.

There is extensive empirical research that supports the interdependence hypothesis (see reviews by Dressler & Kamil, 2006; Baker, 2001; Cummins, 2001; Genesee et al., 2006). The most comprehensive review was conducted by Dressler and Kamil as part of the Report of the National Literacy Panel on Language-Minority Children and Youth (August & Shanahan, 2006). They conclude:

In summary, all these studies provide evidence for the cross-language transfer of reading comprehension ability in bilinguals. This relationship holds (a) across typologically different languages . . . (b) for children in elementary, middle, and high school; (c) for learners of English as a foreign language and English as a second language; (d) over time; (e) from both first to second language and second to first language; (p. 222)

Cummins (2008) has suggested that, depending on the sociolinguistic situation, five types of cross-linguistic transfer are possible:

• transfer of conceptual elements (e.g., understanding the concept of photosynthesis);

• transfer of metacognitive and metalinguistic strategies (e.g., strategies of visualizing, use of visuals or graphic organizers, mnemonic devices, vocabulary acquisition strategies, etc.);

• transfer of pragmatic aspects of language use (willingness to take risks in communication through L2, ability to use paralinguistic features such as gestures to aid communication, etc.);

• transfer of specific linguistic elements (e.g., knowledge of the meaning of photo in photosynthesis);

• transfer of phonological awareness - the knowledge that words are composed of distinct sounds.

The documentation of multiple forms of cross-linguistic transfer (e.g., Dressler & Kamil, 2006) raises the pedagogical issue (to be considered in more detail in a later section) of whether teachers should actively aim to promote transfer across languages among bilingual or emergent bilingual students. A number of researchers have argued for the adoption of bilingual instructional strategies (e.g., Cummins, 2008; Jessner, 2006), but this orientation contravenes the long-term assumption that bilingualism is best developed within bilingual programs through the implementation of monolingual instructional strategies (e.g., Lambert, 1984).

2 The most successful bilingual programs are those that aim to develop bilingualism and biliteracy

Short-term transitional programs are less successful in developing both L2 and L1 literacy than programs such as dual language or maintenance programs that continue to promote both L1 and L2 literacy throughout elementary school. Lindholm-Leary and Borsato (2006) express this pattern of findings as follows:

There is strong convergent evidence that the educational success of ELLs [English language learners] is positively related to sustained instruction through the student's first language . . . most long-term studies report that the longer the students stayed in the program, the more positive were the outcomes. (p. 201)

This pattern of results refutes the assumption underlying many transitional bilingual programs that students should be transferred out of the bilingual program as rapidly as possible.

3 Bilingual education for minority students is, in many situations, more effective in developing L2 literacy skills than monolingual education in the dominant language, but it is not, by itself, a panacea for underachievement

The National Literacy Panel on Language-Minority Children and Youth (August & Shanahan, 2006) concluded that bilingual instruction exerts a moderate but significant effect on minority students' English academic achievement.

In summary, there is no indication that bilingual instruction impedes academic achievement in either the native language or English, whether for language-minority students, students receiving heritage language instruction, or those enrolled in French immersion programs. Where differences were observed, on average they favored the students in a bilingual program. The meta-analytic results clearly suggest a positive effect for bilingual instruction that is moderate in size. This conclusion held up across the entire collection of studies and within the subset of studies that used random assignment of students to conditions. (Francis, Lesaux, & August, 2006, p. 397)

This finding concurs with the results of other recent comprehensive reviews (e.g., Genesee et al., 2006; Rolstad, Mahoney, & Glass, 2005). However, it is important to emphasize that underachievement among subordinated group students derives from many sources (e.g., socioeconomic status, inferior schools, low teacher expectations, etc.) and simply providing some L1 instruction will not, by itself, transform students' educational experience nor reverse the effects of social discrimination and poverty.

Dissenting Perspectives

As noted above, opposition to bilingual education for linguistic minority students derives primarily from ideological concerns related to immigration and national identity in societies that are increasingly diverse. However, two groups of researchers in the United States and Germany respectively have disputed the general pattern of findings presented above regarding the outcomes of bilingual education (Esser, 2006; Rossell & Baker, 1996; Rossell & Kuder, 2005). Rossell and Baker carried out a literature review of studies, which (they claimed) compared bilingual education with "structured immersion" in the dominant language of the school. In a detailed review, Cummins (1999) argued that the Rossell and Baker review is "characterized by inaccurate and arbitrary labeling of programs, inconsistent application of criteria for 'methodological acceptability', and highly inaccurate interpretation of the results of early French immersion programs" (p. 30). The credibility of their review can be gauged from the fact that 90 percent of the studies they claimed as support for "structured immersion" (English-medium programs) are interpreted by the authors of these studies as supporting the effectiveness of bilingual and even trilingual education. Similar problems characterize the more recent review written by Rossell and Kuder (2005).

Esser's (2006) arguments against bilingual education for immigrant and minority students in the German context are based on an uncritical acceptance of the claims made by Rossell and her colleagues (Rossell & Baker, 1996; Rossell & Kuder, 2005) together with inferences drawn from analysis of large-scale international studies such as the Programme for International Student Assessment (PISA) conducted by the Organisation for Economic Cooperation and Development (OECD) (Stanat & Christensen, 2006). His general argument against bilingual education is based on the claim that lack of proficiency in the school language is a major cause of academic difficulties among immigrant students and, consequently, language assimilation through immersion in the school language is a necessary condition for both academic success and social integration. Esser's analysis of the PISA data suggests that knowledge and use of the school language in the home is strongly related to academic success whereas knowledge of the home language either makes no contribution or is negatively related to school success (depending on whether L1 knowledge is accompanied by strong L2 knowledge). He finds no evidence that bilingual education promotes academic development for minority students and suggests that "retention of the first language usually takes place at the cost of second language acquisition (and vice-versa)" (pp. 97-8).

Esser's (2006) argument is unconvincing because he interprets correlational data as causal and fails to take account of the fact that the relationship within PISA between home language and achievement disappeared for a large majority (10 out of 14) of OECD-member countries when socioeconomic status and other background variables were controlled (Stanat & Christensen, 2006, table 3.5, pp. 200-2). The disappearance of the relationship in a large majority of countries

suggests that language spoken at home does not exert any independent effect on achievement but is rather a proxy for variables such as socioeconomic status and length of residence in the host country. Furthermore, any relationship between home language use and achievement is tangential to the issue of whether bilingual education is a legitimate and potentially useful policy option for teaching immigrant and linguistic minority students. The research data (summarized above) overwhelmingly demonstrate the legitimacy of bilingual education and neither Esser nor Rossell and her colleagues provide any credible evidence to the contrary.

Outcomes of Immersion Programs

The outcomes of second-language immersion programs are consistent with the more general findings from bilingual education. The immersion data derive primarily from the Canadian French immersion programs, which have been researched extensively, but also from studies in countries such as Spain (Huguet, Lasagabaster, & Vila, 2008), Japan (Bostwick, 1999), Ireland (Harris, 2007), Singapore (Pakir, 2008), South America (de Mejia, 2008), Sweden (Buss & Lauren, 1995), and the United States (Genesee & Lindholm-Leary, 2008). Note that "immersion" in these contexts is a form of bilingual education that aims to develop fluency and literacy in two languages. The Canadian findings are summarized below as illustrative of the more general trends.

In early immersion programs, students gain fluency and literacy in French at no apparent cost to their English academic skills. Within a year of the introduction of formal English language arts, students catch up in most aspects of English standardized test performance. Usually students require additional time to catch up in English spelling, but by Grade 5 there are normally no differences in English test performance between immersion students and comparison groups whose instruction has been totally through English. One potential limitation of these findings is that standardized tests do not assess all aspects of English academic skills; in particular, writing development is usually not assessed in such tests. However, the few studies that have examined English writing development specifically show no evidence of problems among immersion students in this regard (e.g., Swain, 1975). There is also no evidence of any long-term lag in mastery of subject matter taught through French in early, middle, or late immersion programs.

With respect to French skills, students' receptive skills in French are better developed (in relation to native speaker norms) than are their expressive skills. By the end of elementary school (Grade 6) students are close to the level of native speakers in understanding and reading of French but there are significant gaps between them and native speakers in spoken and written French. The gap is particularly evident with respect to accuracy of grammar and range of vocabulary knowledge and use. These gaps are clearly related to the restricted input that students receive in French. There is typically minimal contact or interaction with French speakers outside the school context and very few students read for

pleasure in French. After the initial grades, reading in French tends to be primarily textbook reading, which is typically not particularly engaging for students. Thus, there are few opportunities for students to extend their exposure to French and expand their knowledge of vocabulary and grammar.

Writing also tends to be carried out only within the school context and applied to academic tasks that are often not highly engaging for students. Students seldom write for authentic purposes where they are encouraged to invest their identities in creative writing projects. As discussed later, a change in pedagogical approach that would emphasize extensive reading and writing across a wide range of genres might significantly improve students' range of vocabulary and grammatical accuracy in their expressive French.

The overall outcomes of French immersion programs can be summarized as follows:

• Students acquire good receptive skills (listening and reading) in French but their productive skills (speaking and writing) are limited with respect to grammatical accuracy and range of vocabulary.

• Teaching through L2 entails no adverse effects on L1 literacy development.

• In early immersion programs, students are able to develop decoding skills in French despite the fact that their French proficiency in the early grades is very limited.

• A large majority of students spontaneously develop English decoding skills in Grades 1 and 2 with no formal instruction in English reading.

• Immersion appears appropriate for a wide variety of students - not just an academic elite. Students with special needs, as well as those who speak a language other than English or French at home, can succeed in immersion programs.

In short, while immersion programs by themselves typically do not result in nativelike French proficiency, they do provide an excellent foundation for students to later "re-immerse" themselves in a genuine French language context, if they so desire, and develop their L2 skills closer to native speaker norms.

In the next section, I briefly sketch bilingual and immersion programs in different parts of the world in order to illustrate the range of sociolinguistic and sociopolitical contexts within which these programs have been implemented.

Illustrative Sketches of Bilingual and Immersion Programs

Malawi

Williams (1996) examined the impact of language of instruction on reading ability in L1 and L2 in Malawi and Zambia. In Malawi, Chichewa is the language of instruction for years 1-4 of primary school, with English taught as a subject. In

Zambia, English is the medium of instruction, with one of seven local languages taught as a subject. Williams administered an English reading test and a local language reading test (Chichewa in Malawi and the almost identical Nyanja in Zambia) to year 5 learners in six schools in each country. He reported no significant difference in English reading ability between students in each country, despite the huge difference in amount of English instruction. However, there were large differences in favor of Malawi in local language reading ability. Williams concluded that these results "are consistent with research on minority groups suggesting that instruction in L1 reading leads to improved results in L1 with no retardation in L2 reading" (p. 183).

Singapore

Pakir (2008) points out that the complexity of the language situation in Singapore does not fit neatly into dichotomous majority/minority language categorizations. English is one of the four official languages of Singapore together with Mandarin, Malay, and Tamil. English, the language of the former colonial power, was initially seen as a "neutral" language and was adopted as the major medium of instruction in school and the "first school language." The other languages were labeled "ethnic mother tongues" and given status as "second school languages." Thus, the bilingual education policy privileges English but also places strong emphasis on the Asian languages of the population. These languages are taught as subjects within the English-medium system. The Singapore educational system appears to be working effectively, as judged by international comparisons. Students from the major language backgrounds have performed well in international comparisons, not only in mathematics and science but also on measures of English literacy, where their scores are at similar levels to several countries where English is the first language of students (e.g., New Zealand, Scotland).

Mexico

Hamel (2008) notes that in 2005, approximately 55,000 indigenous teachers instructed over 1.2 million primary school students who were speakers of one of the 62 indigenous languages still spoken in Mexico. About half of the total indigenous primary school population are now taught by indigenous teachers. Unfortunately, however, the predominant focus in schools serving indigenous students has been on assimilation. Hamel points out that reading primers in indigenous languages funded and produced by the Mexican state are not extensively or effectively used. Reading is typically taught in Spanish from Grade 1. According to Hamel, "the attempt to teach literacy in a second language without sufficient acquisition of the necessary oral skills leads the teachers to under-exploit the communicative potential of the primers, and to return to traditional practices of synthetic methods and structural pattern drill" (p. 317).

However, in recent years, new experimental projects have been implemented based on a pluralist conception of the state and full respect for indigenous peoples

and their ethnic rights. These projects aim to maintain or revitalize indigenous cultures and languages. As one example, Hamel described how, in 1995, the P'urhepecha (Tarascan) teachers from two bilingual elementary schools in Michoacán, in the central Highlands of Mexico, changed the curriculum so that all subject matter including literacy and mathematics was taught in P'urhepecha, the children's L1. Teachers had to create their own materials and develop a writing system. Comparative research several years later reported that students who had acquired literacy in their L1 achieved significantly higher scores in both languages than those who were taught reading and writing in Spanish.

Pedagogical Issues within Bilingual and Immersion Programs

A number of pedagogical and organizational issues have been debated in the context of bilingual and immersion programs. One of these concerns the allocation of languages with respect to both instructional time and academic content to be taught through each language. A related issue concerns the appropriate language for initial reading instruction - should students be introduced to reading in their L1, the L2, or both languages more or less simultaneously? A third issue concerns the extent to which the two languages within a bilingual or immersion program should be kept separate or, alternatively, brought into contact, with the goal of encouraging transfer across languages and developing awareness of language.

Language allocation

It is generally accepted that within bilingual and immersion programs strong emphasis should be put on development of conversational and academic skills in the minority language. For dominant group students (e.g., in a second-language immersion program), exposure to the minority language is usually minimal outside of the school context; therefore, the development of proficiency in that language depends almost exclusively on input within the school. Students from language minority groups, on the other hand, are typically exposed to the minority language within the home. However, the status of this language is often low in comparison to the status of the dominant language. Students frequently internalize the status differential between the languages and, in the absence of L1 instruction, adopt the majority language as their language of choice with consequent loss of their L1 proficiency. Thus, within bilingual programs for minority students, strong emphasis on the minority language is intended to counteract the status imbalance between the languages and enable students to feel proud of their bilingual skills and develop literacy in both languages.

These considerations have led some policy makers and researchers to recommend maximizing instructional time through the minority language, particularly in the early stages of bilingual and immersion programs. However, reinforcement

of the minority language is not just a matter of quantity of instruction. Some of the most successful bilingual and dual language programs in the United States have divided instructional time equally between Spanish and English (e.g., Freeman, 1998). Programs that have initially emphasized the minority language over the majority language (e.g., 90 percent Spanish, 10 percent English in the early grades) have also demonstrated a high level of success (e.g., Lindholm-Leary, 2001). Thus, a variety of options are possible and the research does not point to the superiority of any particular model of language allocation. There is consensus, however, that at least 50 percent of the instructional time should be spent through the minority language for as long as possible throughout the elementary school years.

In the context of language revitalization efforts, immersion programs often maximize instruction through the minority language as a means of extending the domains in which the language is used and the functions served by the language. Most Maori immersion programs in Aoteroa/New Zealand, for example, use Maori exclusively from pre-school through Grade 4 (and sometimes longer), with English introduced only at Grade 5. Typically, English is taught in a classroom separate from the rest of the school so that the school functions essentially as an "English-free zone." The rationale for this policy is that the school is one of the very few places where Maori is normalized as a legitimate language of communication, and academic skills developed through Maori will transfer to English, which is the home language of most of the students. Although debate continues in the Aoteroa/New Zealand context about when and how English should be introduced (e.g., May, Hill, & Tiakiwai, 2003), the decision is essentially a local one, since the broader research suggests that a range of options are feasible and consistent with successful bilingual development.

Similar considerations apply to the issue of which subjects should be taught through each language. This is essentially a decision that should be taken at the local level, taking account of issues such as parent preferences, textbook availability, teacher expertise in particular subject matter, and assessment regime in the wider educational context. For example, in the United States, high-stakes tests are typically administered in Grade 3 through the majority language (English). This reality may lead some policy makers to adopt a 50/50 rather than a 90/10 model and ensure that subjects that will be tested (e.g., reading and mathematics) are taught through English for a sufficient period of time to ensure that students will be successful on the tests.

Language of initial reading instruction

Most immersion programs provide initial reading instruction through students' L2 (e.g., French immersion programs). However, this practice is not based on any research suggesting that introducing reading in L2 is superior to teaching children to read in their L1. It is simply consistent with the philosophy of immersion and the fact that countless evaluations have demonstrated that students can acquire decoding skills through a language that is still inadequately developed.

Immersion and dual language programs that teach reading through students' L1 have also demonstrated success. Similarly, teaching literacy in both languages simultaneously or in quick succession appears to be quite feasible (e.g., Freeman, 1998). As in the case of language allocation, the decision regarding initial language of reading instruction is best viewed as a local option.

There is considerable consensus among researchers, however, that for minority students in bilingual programs, reading should normally be introduced in L1. In some cases, the home language has a more regular sound-symbol relationship than is the case with the dominant language (e.g., Spanish and English in the United States). There is also the consideration that many minority students from low-income backgrounds may come to school with relatively little exposure to literacy in the home; under these circumstances, it makes sense to introduce reading through the language the student already knows. Literacy instruction through minority students' L1 also facilitates the involvement of parents in their children's literacy development and reinforces the status of students' L1.

However, there is also extensive research that demonstrates that many language minority students acquire L2 decoding skills under conditions of initial L2 literacy instruction (Geva, 2006). Thus, the issue of initial literacy instruction remains a local option even though most bilingual programs serving language minority students introduce reading to students through their L1 for the reasons outlined above.

Monolingual or bilingual instructional strategies?

Lambert (1984) clearly expressed the monolingual instructional philosophy underlying French immersion programs:

No bilingual skills are required of the teacher, who plays the role of a monolingual in the target language . . . and who never switches languages, reviews materials in the other language, or otherwise uses the child's native language in teacher-pupil interactions. In immersion programs, therefore, bilingualism is developed through two separate monolingual instructional routes. (p. 13)

Adoption of monolingual instructional strategies within immersion programs reflects what Howatt (1984), in his history of English language teaching, referred to as the "monolingual principle." This principle emphasizes instructional use of the target language (TL) to the exclusion of students' L1, with the goal of enabling students to think in the TL with minimal interference from L1. This principle initially gained widespread acceptance more than 100 years ago in the context of the direct method and has continued to exert a strong influence on various language teaching approaches since that time (Howatt, 1984). According to Yu (2000), "the direct method imitated the way that children learn their first language, emphasizing the avoidance of translation and the direct use of the foreign language as the medium of instruction in all situations" (p. 176). Consistent with direct method principles, translation across languages is seen as unacceptable within immersion (and many bilingual) programs.

There is certainly a rationale for creating largely separate spaces for each language within a bilingual or immersion program. However, there are also compelling arguments to be made for teaching for transfer across languages. The reality is that students are making cross-linguistic connections throughout the course of their learning in a bilingual or immersion program (Jessner, 2006), so why not nurture this learning strategy and help students to apply it more efficiently?

Teaching for cross-linguistic transfer is consistent with both the interdependence hypothesis and the extensive research supporting the crucial role that prior knowledge plays in all learning (e.g., Bransford, Brown, & Cocking, 2000). The interdependence hypothesis has drawn attention to the reality of cross-linguistic transfer in virtually all second-language learning situations. It is reasonable to argue that learning efficiencies can be achieved if teachers explicitly draw students' attention to similarities and differences between their languages and reinforce effective learning strategies in a coordinated way across languages. For example, if the teacher is explaining the meaning of the term predict in science (taught in English) within a French immersion program, it makes sense to explain the meaning of the root (from the Latin dicere meaning 'to say') and the prefix (meaning 'before'), as well as drawing students' attention to the fact that the root and prefix operate in exactly the same way in the French word prédire.

Similarly, the centrality of prior knowledge in the learning process implies that instruction should explicitly attempt to activate students' prior knowledge and build relevant background knowledge as necessary. This holds true regardless of whether students are being instructed through L1 or L2. However, monolingual instructional approaches appear at variance with this fundamental principle of learning because they regard students' L1 (and, by implication, the knowledge encoded therein) as potentially an impediment to the learning of L2. As a result, these approaches are unlikely to focus on activation of students' prior knowledge. In cases where monolingual approaches do acknowledge the role of prior knowledge, they are likely to limit its expression to what students can articulate through their L2.

Among the bilingual instructional strategies that can be employed to promote literacy engagement in both L1 and L2 are the following (Cummins, 2008):

• focus on cognates in contexts where the languages share common linguistic origins;

• creation and web-publication of dual language multimedia books and projects (see, for example, www.multiliteracies.ca and http://thornwood.peelschools. org/Dual/); the creation of dual language books clearly involves translation across languages, a practice that has hitherto been viewed as pedagogically unacceptable in immersion and bilingual programs;

• sister class exchanges in which students use the Internet to connect with other bilingual students and use both L1 and L2 to create literature and art and/or to explore issues of social relevance to them and their communities.

Immersion researchers are beginning to acknowledge that students' use of their L1 serves some legitimate and useful learning functions within the L2-medium classroom (Swain & Lapkin, 2000, 2005). According to Swain and Lapkin (2005), students' use of the L1 enables them to develop strategies to carry out tasks in the target language and to work through complex problems more efficiently than they might be able to do through their L2. They also point to the changing demographic realities of immersion education in Canada and in other contexts -an increasing number of students from language backgrounds other then English and French are now in immersion programs. They argue that it is important to support the home language development of these students within the immersion program, in addition to the teaching of French and English.

In short, although most bilingual and immersion programs continue to rely almost exclusively on monolingual instructional strategies, there is emerging recognition that students' L1 can function as a cognitive and linguistic resource to scaffold more accomplished performance in the L2.

Conclusion

Research during the past 40 years has clearly established bilingual and immersion programs as a legitimate educational option for both majority and minority language students. For majority language students, bilingual/immersion education provides an effective means of developing proficiency in a target language at no cost to students' fluency or literacy in their L1. For minority students, bilingual education similarly promotes development of fluency and literacy in two languages; furthermore, in the case of minority students who are at risk of school failure, bilingual education has demonstrated its potential to support students' overall academic development more effectively than programs conducted exclusively through the majority language.

In the current era of unprecedented population mobility, the economic and personal utility of bilingual and multilingual skills has become increasingly obvious, and this phenomenon has propelled awareness of, and interest in, bilingual and immersion education. Population mobility also increases the number of children from linguistically diverse groups in countries around the world. Although it is clearly not feasible to implement bilingual programs on a large scale in school situations that are highly multilingual, there is increasing recognition among educators in many contexts that minority students' home languages represent

(1) a significant intellectual and personal resource for the students themselves,

(2) an important communicative tool within families, and, (3) in an interdependent world, an economic and diplomatic resource for the nation as a whole. In Ontario, Canada, for example, Ministry of Education documents now highlight the importance of students' home languages and provide concrete strategies to enable educators to support students' languages within the mainstream (English-medium) classroom (Ontario Ministry of Education, 2006).

Although there is no longer serious debate about the scientific legitimacy of bilingual education for linguistic minorities, the ideological debate is likely to continue for the foreseeable future, partly because it has very little directly to do with education. The issues concern the extent to which societies should adopt a pluralist approach that encourages children and communities to maintain and develop their languages and culture, in addition to acquiring the majority language, or alternatively, whether schools should promote the assimilation of immigrants and encourage minority students to abandon their home languages and cultures. In contexts where this debate is raging, bilingual programs are frequently seen as valuable and worthy of public funding when they are directed toward the acquisition of additional languages by dominant group students, but highly problematic when the beneficiary of bilingual education is a minority or subordinated group.

REFERENCES

August, D. & Shanahan, T. (eds.) (2006). Developing literacy in second-language learners. Mahwah, NJ: Lawrence Erlbaum.

Baker, C. (2001). Foundations of bilingual education and bilingualism, 3rd edn. Clevedon, UK: Multilingual Matters.

Bostwick, R. M. (1999). A study of an elementary English language immersion school in Japan. Unpublished Doctoral dissertation, Temple University, Philadelphia.

Bransford, J. D., Brown, A. L., & Cocking, R. R. (2000). How people learn: Brain, mind, experience, and school. Washington, DC: National Academy Press.

Buss, M. & Laurén, C. (1995). Language immersion: Teaching and second language acquisition: From Canada to Europe. Proceedings of the University of Vaasa Research Papers. Tutkimuksia No. 192. Vaasa: University of Vaasa.

Cummins, J. (1979). Linguistic interdependence and the educational development of bilingual children. Review of Educational Research 49, 222-51.

Cummins, J. (1981). The role of primary language development in promoting educational success for language minority students. In California State Department of Education (ed.), Schooling and language minority students: A theoretical framework. Evaluation, Dissemination and Assessment Center, California State University, Los Angeles.

Cummins, J. (1999). Alternative paradigms in bilingual education research: Does theory have a place? Educational Researcher 28, 26-41.

Cummins, J. (2001). Negotiating identities: Education for empowerment in a diverse society, 2nd edn. Los Angeles: California Association for Bilingual Education.

Cummins, J. (2008). Teaching for transfer: Challenging the two solitudes assumption in bilingual education. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 65-75). New York: Springer Science/Business Media LLC.

Cummins, J. & Hornberger, N. H. (eds.) (2008). Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education. New York: Springer Science/Business Media LLC.

de Mejia, A. M. (2008). Enrichment bilingual education in South America. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn.,

vol. 5: Bilingual education (pp. 323-31). New York: Springer Science/Business Media LLC.

Dressler, C. & Kamil, M. (2006). First- and second-language literacy. In D. August & T. Shanahan (eds.), Developing literacy in second-language learners: Report of the National Literacy Panel on Language-Minority Children and Youth (pp. 197-238). Mahwah, NJ: Lawrence Erlbaum.

Esser, H. (2006). Migration, language, and integration. AKI Research Review 4. Berlin: Programme on Intercultural Conflicts and Societal Integration (AKI), Social Science Research Center. Retrieved 21 December 2007 from http://www.wzb.eu/zkd/aki/ files/aki_research_review_4.pdf.

Francis, D., Lesaux, N., & August, D. (2006). Language of instruction. In D. August & T. Shanahan (eds.), Developing literacy in second-language learners: Report of the National Literacy Panel on Language-Minority Children and Youth (pp. 365-413). Mahwah, NJ: Lawrence Erlbaum.

Freeman, R. D. (1998). Bilingual education and social change. Clevedon, UK: Multilingual Matters.

García, O. & Baker, C. (2007). Bilingual education: An introductory reader. Clevedon, UK: Multilingual Matters.

Genesee, F. & Lindholm-Leary, K. (2008). Dual language education in Canada and the USA. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 253-63). New York: Springer Science/Business Media LLC.

Genesee, F., Lindholm-Leary, K., Saunders, W. M., & Christian, D. (eds.) (2006). Educating English language learners: A synthesis of research evidence. New York: Cambridge University Press.

Geva, E. (2006). Second-language oral proficiency and second-language literacy. In D. August & T. Shanahan (eds.), Developing literacy in second-language learners: Report of the National Literacy Panel on Language-Minority Children and Youth (pp. 123-39). Mahwah, NJ: Lawrence Erlbaum.

Hamel, R. E. (2008). Bilingual education for indigenous communities in Mexico. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 311-21). New York: Springer Science/Business Media LLC.

Harris, J. (2007). Bilingual education and bilingualism in Ireland North and South. International Journal of Bilingual Education and Bilingualism 10, 359-68.

Howatt, A. (1984). A history of English language teaching. Oxford: Oxford University Press.

Huguet, A., Lasagabaster, D., & Vila, I. (2008). Bilingual education in Spain: Present realities and future challenges. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 225-37). New York: Springer Science/Business Media LLC.

Jessner, U. (2006). Linguistic awareness in multilinguals: English as a third language. Edinburgh: Edinburgh University Press.

Johnson, R. K. & Swain, M. (1997). Immersion education: International perspectives. Cambridge: Cambridge University Press.

Lambert, W. E. (1984). An overview of issues in immersion education. In California State Department of Education (ed.), Studies on immersion education: A collection for United States educators (pp. 8-30). Sacramento: California State Department of Education.

Lambert, W. E. & Tucker, G. R. (1972). Bilingual education of children: The St. Lambert experiment. Rowley, MA: Newbury House.

Lindholm-Leary, K. (2001). Dual language education. Clevedon, UK: Multilingual Matters.

Lindholm-Leary, K. J. & Borsato, G. (2006). Academic achievement. In F. Genesee, K. Lindholm-Leary, W. Saunders, & D. Christian (eds.), Educating English language learners (pp. 176-222). New York: Cambridge University Press.

Macnamara, J. (1966). Bilingualism and primary education. Edinburgh: Edinburgh University Press.

Mahshie, S. N. (1995). Educating Deaf children bilingually: With insights and applications from Sweden and Denmark. Washington, DC: Gallaudat University.

Malherbe, E. G. (1946). The bilingual school. Johannesburg: The Bilingual School Association.

May, S. (2008). Bilingual/immersion education: What the research tells us. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 19-33). New York: Springer Science/Business Media LLC.

May, S., Hill, R., and Tiakiwai, S. (2003). Bilingual/immersion education: Indicators of good practice. Final Report for New Zealand Ministry of Education. Hamilton: Wilf Malcolm Institute of Educational Research.

Ontario Ministry of Education (2006). Many roots, many voices. Toronto: Ministry of Education. Available at: http://www.edu.gov.on.ca/eng/document/manyroots/.

Pakir, A. (2008). Bilingual education in Singapore. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 191-203). New York: Springer Science/Business Media LLC.

Rolstad, K., Mahoney, K., & Glass, G. V. (2005). The big picture: A meta-analysis of program effectiveness research on English language learners. Education Policy 10, 57294.

Rossell, C. H. & Baker, K. (1996). The effectiveness of bilingual education. Research in the Teaching of English 30, 7-74.

Rossell, C. H. & Kuder, J. (2005). Meta-murky: A rebuttal to recent meta-analyses of bilingual education. In J. Söhn (ed.), The effectiveness of bilingual school programs for immigrant children, AKI Research Review 2 (pp. 43-76). Berlin: Programme on Intercultural Conflicts and Societal Integration (AKI), Social Science Research Center. Retrieved 21 December 2007 from http://www.wzb.eu/zkd/aki/files/aki_bilingual_ school_programs.pdf.

Skutnabb-Kangas, T. (2000). Linguistic genocide in education - or worldwide diversity and human rights. Mawah, NJ: Lawrence Erlbaum.

Skutnabb-Kangas, T. & McCarty, T. L. (2008). Key concepts in bilingual education: Ideological, historical, epistemological and empirical foundations. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 3-17). New York: Springer Science/Business Media LLC.

Small, A. & Mason, D. (2008). American Sign Language (ASL) bilingual bicultural education. In J. Cummins & N. H. Hornberger (eds.), Encyclopedia of language and education, 2nd edn., vol. 5: Bilingual education (pp. 133-5). New York: Springer Science/Business Media LLC.

Stanat, P. & Christensen, G. (2006). Where immigrant students succeed: A comparative review of performance and engagement in PISA 2003. Paris: Organisation for Economic Cooperation and Development.

Swain, M. (1975). Writing skills of grade three French immersion pupils. Working Papers on Bilingualism 7, 1-38.

Swain, M. & Lapkin, S. (2000). Task-based second language learning: The uses of the first language. Language Teaching Research 4, 253-76.

Swain, M. & Lapkin, S. (2005). The evolving sociopolitical context of immersion education in Canada: Some implications for program development. International Journal of Applied Linguistics 15, 169-86.

Thomas, W. P. & Collier, V. P. (2002). A national study of school effectiveness for language minority students' long-term academic achievement. Santa Cruz, CA: Center for Research on Education, Diversity and Excellence, University of California-Santa Cruz. Available at: http//www.crede.ucsc.edu.

Williams, E. (1996). Reading in two languages at Year 5 in African primary schools. Applied Linguistics 17, 2, 183-209.

Yu, W. (2000). Direct method. In M. Byram (ed.), Routledge encyclopedia of language teaching and learning (pp. 176-8). New York: Routledge.

11 Heritage Language Programs

SILVINA MONTRUL

Introduction

The last decade has seen an important change in second language classrooms as a result of recent immigration patterns all over the world, especially in Western Europe, Australia, and North America. Partly due to these demographic changes and to the impact of globalization on international relations, today more than ever before, an increasing number of bilingual speakers of minority languages wish to maintain and/or (re)learn their family language. As a consequence, post-secondary foreign/second language classes typically geared to students who have little to no background in the target language (L2 learners) have had to accommodate speakers who were exposed to the language at home early in childhood (henceforth, heritage language learners or HL learners) and whose levels of oral proficiency in the language range from minimal to advanced.1 While many institutions place HL learners in foreign/second language classrooms that follow a traditional L2 curriculum, in the 1970s other institutions started to create special courses to address the specific linguistic and cultural needs of HL learners. These are heritage language programs.

With this increasing trend, there is growing recognition that heritage language programs should be informed by a theory of HL acquisition and teaching (Kondo-Brown, 2003; Lynch, 2003; Valdés, 1997, 2005). To date, there is very little systematic research on HL learners and HL acquisition, although the situation is rapidly changing.

My purpose in this chapter is to show how a combination of the linguistic and cognitive views of first and second language acquisition can be extended to make predictions about the heritage language learning process that teaching is designed to bring about. I will then discuss what most recent research has so far uncovered about heritage speakers of different languages and their language learning process. Finally, I will evaluate how existing programs and models for heritage language teaching are consistent with the predictions of the theoretical position I advance here.

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 Bkckwell Publishing Ltd. ISBN: 978-1-405-15489-5

Heritage Languages and Speakers

Although originally coined in Canada, the term heritage language began to be used in the United States in the 1990s to refer to minority languages (Cummins, 2005). Commonly used terms in different parts of the world are international, community, immigrant, ethnic, indigenous, minority, ancestral, third, and non-official language. In the United States, a heritage language is a language other than English spoken by immigrants and their children (Valdes, 2001).2 One problem with this definition is that immigrant languages may have dual or multiple statuses. For example, in the United States, Spanish is the most widely studied foreign language in elementary and high schools. It is also a second language, since the United States ranks fifth in the world for the number of Spanish speakers. And because Spanish is spoken by Latino immigrants, it is also a heritage language. Polish, on the other hand, is a heritage language whose distribution is highly restricted.

Like the term heritage language, heritage speaker is not easy to define. From an ethnolinguistic perspective, what defines a heritage speaker is cultural and linguistic identity (Carreira, 2004; Wiley, 2001). A heritage speaker is a member of an ethnolinguistically minority culture, that is, an individual who may or may not have knowledge of the language. From this perspective, individuals who have one grandparent or great-grandparent from the minority language culture are considered heritage speakers, even if the person was not exposed to the language at home and does not speak the language. Similarly, African Americans wishing to learn Swahili to understand their origins may be considered heritage speakers, as may third- or fourth-generation Italians, with no knowledge of the language, living in Argentina or the United States. Two main reasons motivating these individuals to learn the heritage language are familiarity with the culture and/or a desire to reconnect with their ancestral roots.

While linguistic and cultural identity is certainly a feature that characterizes heritage speakers, researchers approaching the issue from linguistics and education take linguistic proficiency in the heritage language as a defining factor. In the US context, Valdes (2001) defines a heritage speaker as a bilingual individual "raised in a home where a non-English language is spoken, who speaks or merely understands the heritage language, and who is to some degree bilingual in English and the heritage language" (p. 1). The term bilingual, like the term speaker in heritage speaker, implies that the individual must have some oral command of the heritage language. But what makes matters even more complicated is that the profiles of these speakers vary greatly. Sociolinguists often relate generation of immigration and degree of bilingualism as a result of acculturation patterns (Silva-Corvalan, 1994). Table 11.1 illustrates some characteristics of heritage speakers based on generation and degree of bilingualism.

First generation immigrants immigrate to the host country as adults. They are typically monolingual speakers of the heritage language, and most of them learn the majority language late in life (and imperfectly). Command of the heritage

Table 11.1 Linguistic characteristics of heritage speakers

Generation

Possible language characteristics

First generation (parents)

Second generation (children)

Third generation (grandchildren)

Monolingual in the heritage language

Dominant in the heritage language

Dominant in the majority language

Incipient L2 learner of the majority language

Dominant in the majority language

Monolingual in the majority language

language is strong in this immigrant group, although there can be some attrition after more than 10 years of intense exposure to the majority language (Kopke and Schmid, 2003).

The children of the first generation are the second generation immigrants. This group may include the children of first generation immigrants born in the host country to at least one first generation parent. It also includes immigrant children who come to the host country before the age of 5. In terms of types of bilingual profile, this group may include: (1) simultaneous bilinguals, those exposed to the heritage and the majority language before the age of 5; (2) sequential bilinguals or child L2 learners, those exposed to the heritage language at home until age 4-5, and to the majority language once they start pre-school; and (3) late child L2 learners, children monolingual in the heritage language, who received some elementary schooling in their home country and immigrated around age 7-8.

It is in the second generation when language shift in the home typically occurs, due to the fact that children are schooled in the majority language and have a strong desire to fit in with the new society. With language shift, there are concomitant changes in the bilingual balance of second generation children until adulthood. As the majority language begins to be used more than the home language, the heritage language may be either incompletely acquired or undergo attrition.3 In early childhood, many of these children are either monolingual or dominant in the heritage language. As bilingualism progresses during the elementary school period, the children can be balanced in the two languages (typically at age 10-11, according to Kohnert, Bates, & Hernández, 1999) and eventually become dominant in the majority language. When they reach adolescence, they are already dominant in the majority language, and by the time they are adults the majority language is both stronger and dominant in proficiency and domains of use. Due to this rapid shift, which is very common in the United States as a result of its covert English-only language policies, by the third generation (the grandchildren of the first generation immigrants) heritage speakers are native speakers of the majority language. Some may have partial knowledge of the heritage language (if a grandparent or other relative lives in the home), while

most do not. By the fourth and subsequent generations, the language is no longer used in the family. This pattern of declining bilingualism shows how heritage languages do not survive intergenerational transmission in many countries even when they are supported by continuous immigration.

The Acquisition of Heritage Languages

From an acquisition perspective, heritage speakers are bilinguals who were exposed to the heritage language early in childhood. Thus, age of onset of bilin-gualism and type of linguistic support throughout the lifetime are key factors in understanding different heritage speaker profiles. Adult heritage speakers may have failed to develop full linguistic competence as they began using the majority language intensively at different ages in childhood and did not receive schooling in the heritage language. Heritage speakers are a case of incomplete acquisition (Montrul, 2002; Polinsky, 1997). As a result, many of these individuals want to acquire, reacquire, or expand their knowledge of the heritage language in a classroom setting in early adulthood. Currently, there are two main models to instruct such HL learners: L2 programs or HL programs. Ignoring for the moment practical issues affecting the implementation of one model over the other, important theoretical assumptions underlie the two models, even if these assumptions are not spelled out explicitly.

The L2 program model assumes that HL and L2 learners with little or no previous knowledge of the language are linguistically alike, while the HL model assumes that there are important cultural and linguistic differences between heritage and non-heritage learners. Two questions remain unanswered: first, how are HL learners similar to or different from L1 and L2 learners? And second, what theoretical approach(es) can account for the type of knowledge HL learners have or lack? These questions are crucial for understanding the linguistic needs of these learners and identifying the best way to address their needs in the classroom.

L1 and L2 acquisition

Early L1 acquisition happens through the aural medium and takes place in a naturalistic setting by means of interaction with caregivers and with limited access to correction and feedback on grammatical form. L1 acquisition is universal, uniform, and complete. It is universal because by 3-4 years of age all normally developing monolingual children master the basic structure of their native language, including its phonology, morphosyntax, semantics, and some aspects of pragmatics and sociolinguistic conventions. It is uniform because children exposed to the same language or dialect reach the same level of linguistic development (and competence) despite variations in input. In effect, children converge on the grammar of other members of the speech community. It is complete because the outcome of the acquisition process is successful, although this does not mean

it is entirely error free. While basic acquisition of the structure and conventions of the language is relatively rapid, children make errors along the way. In all languages there are well-documented developmental stages in different areas of linguistic knowledge, such that some structures and sounds are controlled earlier than others. However, the phenomenon of fossilization (Lardiere, 2006; Long, 2003; Selinker, 1972), or arrested development, so typical of adult L2 acquisition, does not occur.

With the linguistic foundations of the language and the essentials of native-speaker competence in place by age 3-4, language acquisition continues beyond this early period. Around age 4, children's metalinguistic ability develops through emergent literacy and continues at school, where children learn to read and write, expand their vocabulary, and acquire more complex structures. Exposure to rich oral and written input allows children to learn to communicate in different registers and styles, both orally and in writing. At the end of the process, children become educated adult native speakers capable of functioning in many social and professional contexts.

Postpuberty L2 acquisition, on the other hand, typically occurs in a classroom setting, with heavy emphasis on reading and writing, and grammatical explanations, practice, feedback, and assessment of the developing L2 skills. Unlike L1 acquisition by children, adult L2 acquisition is not universal - not everybody learns a second language. L2 acquisition is variable rather than uniform - not all L2 learners attain the same level of linguistic competence in the second language. In terms of outcome, L2 acquisition is typically incomplete - most learners never reach the competence of a native speaker. Throughout the process of acquisition, L2 learners make both developmental errors, like L1 learners, and transfer errors due to influence from their L1, especially at early stages. A key difference between L1 and L2 acquisition, however, is that while child L1 learners overcome developmental errors without need for instruction, L2 learners continue to make many errors even after receiving instruction, practice, and correction. Although some researchers argue that attainment of full linguistic competence in the L2 is in principle possible, it is by no means guaranteed. Fossilization can occur at any point in L2 development.

Due to these characteristics, it has been hypothesized that L1 and L2 acquisition utilize very different learning mechanisms, as spelled out in Bley-Vroman's (1989) Fundamental Difference Hypothesis (FDH). According to Bley-Vroman (and other generative linguists), child L1 acquisition happens so rapidly and efficiently because the process is guided largely by innate mechanisms, which are assumed to be part of Universal Grammar (Chomsky, 1981). That is, at the outset of language acquisition, children are guided by the inventory of principles and constraints subsumed under Universal Grammar. To explain the apparent differences between L1 and L2 acquisition in terms of outcome, the main claim of the FDH is that access to Universal Grammar is subject to a critical period, such that when learning a second language, postpuberty L2 learners can only rely on their L1 knowledge (a particular instantiation of Universal Grammar, but not the full spectrum of linguistic options) and the principles and parameters active in their

language. Unable to utilize domain-specific (i.e., purely linguistic) mechanisms as L1 learners, L2 learners resort to domain-general problem-solving skills, like analogy or pattern matching, instead.4

This particular position within generative linguistics is echoed within cognitive and neurolinguistic perspectives on L2 acquisition, which do not necessarily view language and language learning as innate, but take into account the distinction between procedural and declarative knowledge and implicit and explicit language learning (DeKeyser, 2003; Paradis, 2004). Implicit knowledge refers to that learned without awareness of what is being learned, and is learned incidentally or not (depending on the researcher). Implicit knowledge is stored in procedural memory, and when this knowledge is accessed or recalled, it is executed automatically and quickly. By contrast, explicit knowledge is acquired with awareness of what is being learned, and with conscious effort. Because explicit knowledge is learned explicitly, individuals can verbalize this knowledge on demand. It is stored in declarative or episodic memory, where our world knowledge is stored.

Adult educated native speakers have both systems of learning available and use them as needed. According to Paradis (2004), when young children speak or comprehend language, they use implicit competence (or knowledge) only. This is also true of adults who are illiterate. By contrast, incipient L2 learners use explicit knowledge of the L2 when producing or understanding the L2, and steadily and in tandem develop implicit competence of it. In agreement with Bley-Vroman's position, De Keyser (2000, 2003) also contends that adult L2 learners use a different cognitive system to learn an L2 because maturational constraints apply to implicit linguistic competence acquired early in childhood. The decline of procedural memory and loss of implicit cognitive mechanisms for language somewhere in childhood - what Bley-Vroman takes to be Universal Grammar and domain-specific mechanisms - force late L2 learners to rely on explicit learning. Where does HL acquisition fit in this model?

Heritage language acquisition

Descriptively, HL acquisition has characteristics of both L1 and L2 acquisition. It is incomplete L1 acquisition that takes place in a bilingual environment rather than in a monolingual one. As such, HL acquisition shares the development path (or lack thereof) and characteristics of L2 grammars as well. Table 11.2 summarizes the characteristics of these three types of acquisition: L1, L2, and HL. Bold type represents the intersecting subset between L1 and L2 acquisition which characterizes HLA.

As can be seen, HL and L1 learners are both exposed to the language in early childhood through the aural medium, and before the emergence of literacy. Recall that this is the period during which the essence of native-speaker competence develops. HL learners may command basic structures of the language if they received abundant input, or only a subset of those structures, if the input was less abundant. Because HL acquisition takes place in a bilingual environment,

Table 11.2 Characteristics of L1, L2, and HL acquisition

L1 Acquisition

Early exposure to the language Naturalistic setting (aural input)

Abundant input

Control of features of language acquired very early in life (phonology, some vocabulary, some linguistic structures)

Developmental errors

Outcome is successful and complete

Fossilization does not occur

No clear role for motivation and affective factors to develop linguistic competence

More complex structures and vocabulary developed at school after age 5, when metalinguistic skills develop.

Bold type = HL acquisition

L2 Acquisition

Late exposure to the language

Instructed and/or naturalistic setting (aural and written input)

Varying amount of input

Grammar may be incomplete (no chance to develop other structures and vocabulary)

Developmental and transfer errors

Outcome is variable proficiency. It is typically incomplete

Fossilization is typical

Motivation and affective factors play a role in language development

Experience with literacy and formal instruction

as a HL learner develops command of the majority language, he or she also makes transfer errors. The outcome of HL acquisition is also variable and often incomplete, as in L2 acquisition, due to reduction of input and use of language in restricted contexts. Fossilization, so typical of L2 acquisition but unheard of in L1 acquisition, is also frequent in incomplete HL acquisition. Notice that the last row of Table 11.2 is not bold. This is because HL learners do not typically receive schooling in their HL like monolingual children after age 4-5, unless they attend a bilingual or dual immersion program that teaches the HL. Because they do not experience literacy training or formal instruction in their HL, whatever skills they possess are probably transferred from their metalinguistic and literacy development in the majority language.

If some cases of HL acquisition are incomplete or interrupted L1 acquisition in a bilingual environment, then one theoretical prediction is that HL learner's knowledge of the language (prior to instruction) has been acquired implicitly

and through access to Universal Grammar in childhood, before the closure of the critical period. That is, HL learners should have implicit knowledge of aspects of phonology and morphosyntax (basic word order, Pro-drop parameter, binding principles), which emerge very early in childhood and that are not overly dependent on a heavy amount of continuous input. In turn, aspects of language that are context-dependent, acquired after age 5, and reinforced through reading and formal instruction at school - such as specific vocabulary, forms of address and honorifics, complex structures like relative clauses, and semantically and pragmatically conditioned uses of the subjunctive in Spanish and Russian, for example - should either be missing or remain imperfectly acquired, depending on the amount of input received. Table 11.3 lists six predictions this cognitive-linguistic perspective makes for HL acquisition.

In summary, if HL learners are interrupted L1 learners who received some crucial input during the critical period, instruction should be able to turn incomplete native speakers into complete educated native speakers, given optimal amounts of input and time to develop the underdeveloped skills. Timing and amount of input will vary as a function of the HL learner's competence in the language. In principle, it should be faster for HL learners to reach certain linguistic milestones than for L2 learners. By contrast, regardless of amount of input and experience with the target language, L2 learners are not necessarily guaranteed to attain native-speaker competence in the L2. Thus, the theory of language learning just discussed predicts that HL learners have the potential to reach nativelike competence in the HL. With this background, the next section presents recent research findings that bear on this theory.

Research Findings

Comparing L2 and HL learners

The first prediction listed in Table 11.3 - that heritage speakers should be better than L2 learners in basic aspects of phonology and morphosyntax - is confirmed by a series of recent findings. Au et al. (2002) and Oh et al. (2003) compared accent ratings and the production of stop consonants in Spanish (voiceless [p, t, k], voiced [b, d, g]) and Korean (aspirated [p, t, k], lax [p, t, k] and tense [p, t, k]) by very low-proficiency HL and L2 learners. Both studies found that HL learners outperformed L2 learners in both accent ratings and VOT measures. While Au et al. also claimed there was no advantage for HL learners in morphosyntax, this conclusion is premature. Montrul (2005) looked at the syntactic distribution of intransitive verbs in Spanish, and Montrul et al. (2006) focused on the acquisition of object clitic placement and word order in Spanish. Both studies showed that low-proficiency HL learners were more accurate than proficiency-matched L2 learners in a grammaticality judgment task and an online processing task. These advantages for HL learners over L2 learners in basic grammatical areas was also supported by Montrul's (2006) study of subject pronouns, agreement, and word

Table 11.3 Some predictions of a cognitive and linguistic approach to HL acquisition

1 HL learners should be better than L2 learners with aspects of grammar that are acquired early in childhood (phonology; basic aspects of syntax, like word order; basic aspects of morphosyntax, like nominal and verbal agreement), even when their proficiency in the language is low to intermediate.

2 Transfer errors from the stronger language (L2 in HL learners and L1 in L2 learners) should occur in grammatical areas where the two languages differ parametrically.

3 HL learners should be more accurate and faster than L2 learners in oral than in written production and comprehension tasks.

4 HL learners should be more accurate and faster than L2 learners in written tasks that require less metalinguistic awareness.

5 If adults are left with their explicit learning skills for learning language beyond childhood, explicitly learned knowledge may mask implicit linguistic competence in L2 learners, especially if the tasks used to tap that knowledge are metalinguistic.

6 For many researchers, it is not clear that L2 learners eventually develop implicit competence of their L2, although this remains controversial. If L2 learners have lost access to implicit linguistic competence, then they should not be able to perform like native speakers, even after several years of instruction and immersion. By contrast, if HL learners have some implicit knowledge of the HL acquired in childhood, they should eventually catch up with the missing explicit and metalinguistic knowledge that they did not get at school through reading and writing instruction.

order in Spanish. Results showed that while intermediate HL learners had the pro-drop parameter in place, the L2 learners had not yet reset the parameter to the Spanish value, since they did not produce sentences with post-verbal subjects like the HL learners.

Although HL learners may have an advantage over L2 learners in many grammatical domains, both groups make similar types of transfer errors, as prediction 2 states. Montrul (2004) documented striking similarities between L2 and HL learners, where the two groups displayed statistically comparable patterns in their errors with aspectual interpretations in Spanish. Similarly, Kim, Montrul, and Yoon (in press) studied anaphor interpretations in Korean and found that, like L2 learners, Korean HL learners tended to reject long-distance binding interpretations of the anaphor caki, for instance, when this anaphor appeared in structural configurations not allowed in English.

Prediction 3 states that, due to gaps in their acquisition of literacy skills in their heritage language, HL learners should be better and faster than L2 learners in tests of oral production (this is also due to the mode of acquisition). Matsunaga

(2003) offers confirming evidence for this prediction. Matsunaga tested Japanese heritage speakers' oral and reading skills against those of L2 learners (some of whom were Chinese and knew kanji, while others did not know any kanji, like the Japanese heritage speakers). Results revealed that the Japanese HL learners had significantly higher oral proficiency than the Japanese L2 learners, but their reading proficiency was about the same. In support of prediction 4, Montrul, Foote, & Perpinan (2008) tested knowledge of gender agreement in nouns in Spanish heritage speakers and L2 learners using two written tasks and an oral task. Results showed that the L2 learners were statistically more accurate on gender marking than the HL learners in the two written tasks, while the HL learners were significantly better and faster than the L2 learners in the spontaneous oral picture-naming task.

A large part of prediction 5, that some near-natives can pass as natives on metalinguistic tasks, has been addressed by several critical period studies over the years. Some of these studies have documented that even when advanced L2 learners can pass for native speakers impressionistically, they turn out to be non-native when examined in a controlled experimental situation (Coppieters, 1987; Sorace, 1993).

Prediction 6 refers to outcome differences between L2 and HL learners. What is needed to address this prediction are studies of very advanced heritage speakers who are also extremely fluent in the majority language. The question is whether advanced HL learners will behave like L2 learners or like native speakers. The findings from Guillelmon and Grosjean's (2001) study of English-French early and late bilinguals may be relevant here, although it is not entirely clear whether the early bilinguals in this study are heritage speakers (i.e., can English be considered an immigrant language in France?). Results of a psycholinguistic task showed that the early bilinguals performed like monolingual controls in the processing of grammatical and ungrammatical sentences testing gender agreement in French. By contrast, the late learners were insensitive to gender errors.

Finally, prediction 6 also refers to the positive effects of instruction on advanced L2 and HL learners. There is virtually no research that directly addresses these issues. One exception is the study of Song et al. (1997) on the effects of instruction on child HL learners of Korean attending a Korean community school. They tested knowledge of case markers, word order, and reflexive pronouns. The children who were initially found not to control these aspects of Korean grammar improved significantly on a post-test after receiving explicit instruction on these topics. Although these results with children are encouraging, similar experiments with young adults are sorely needed to assess the effectiveness of this type of instruction for HL development at the postsecondary level.

Differences among HL learners

Empirical evidence showing differences and similarities between HL and L2 learners is important for understanding whether these two groups need similar or different types of language instruction. However, the pedagogical challenges

for HL educators do not stop there. The other reality is that it is also difficult, if not impossible, to generalize about the HL learner population, since there are considerable individual differences in linguistic skills in the target HL. As discussed above, most of the variation occurs in second generation speakers, but is not restricted to those immigrants (see Carreira, 2004, for some other complex profiles). Those with only one parent who is a full speaker of the HL will probably speak less fluently than those with both parents who speak the HL. If only the mother speaks the HL, the child is more likely to be proficient in the HL than if only the father speaks the HL. Furthermore, age of onset of bilingualism is a crucial factor in determining proficiency. Montrul's (2002) study of tense-aspect distinctions in Spanish included three groups of heritage speakers: simultaneous bilinguals, sequential bilinguals or child L2 learners of English, and late child L2 learners. Results showed that those bilinguals who received the most exposure at home during childhood before they started learning English after age 5 (the child L2 learners or sequential bilinguals) had more knowledge of Spanish tense-aspect than the simultaneous bilinguals who were exposed to English from birth. The late child L2 learners, who received some schooling in the HL in their countries of origin before moving to the United States, were the closest to the adult native-speaker control group.

Further research including HL learners with different linguistic profiles is Kondo-Brown (2005), a study of Japanese HL learners in Hawai'i. Kondo-Brown looked at the degree of connection to the heritage language (from a merely ethnic connection to having actually been exposed to the language). Kondo-Brown used a locally developed proficiency test and self-assessment measures to investigate differences in listening, speaking, reading, and writing in L2 learners with no previous exposure to Japanese and in HL learners. HL learners were defined in a very broad sense and included individuals with some ethnic connection to Japan (some relative or grandparent), but no active knowledge of the language, as well as HL learners born to Japanese-speaking parents, who were exposed to and used Japanese at home from an early age. Results of tests showed that HL learners with no previous knowledge of Japanese patterned with the L2 learners, whereas HL learners who were bilingual to some degree were significantly different and superior on all measures. This study clearly shows that exposure to the language early in childhood brings advantages to heritage speakers (consistent with prediction 1 in Table 11.3), and that HL learners with just an ethnic connection to the language are like L2 learners.

Also focusing on Japanese learners, Kanno et al. (2008) examined how the linguistic profiles of 15 ACTFL-defined advanced HL and L2 learners correlated with several measures of linguistic proficiency. (See Lee et al., 2008 for a replication of this study with Korean heritage and L2 learners.) All learners completed a guided oral narrative, a written test, and a free conversation. L2 learners were divided into two groups based on learning environment (naturalistic or instructed), while the HL learners were divided into three different bilingual subgroups based on degree of experience and exposure to Japanese beyond the home (those who attended Saturday school, those who received exposure in other contexts, and

those whose exposure was limited to the home environment). Results indicated that the oral test provided a good measure of the students' advanced proficiency, but the written test allowed the researchers to identify structural areas that are already controlled by this level of learner, together with systematic gaps in their linguistic knowledge. Overall, the HL learners who were exposed to Japanese from birth and attended Japanese Saturday school outperformed all other L2 and HL learner groups in grammar and vocabulary. In conclusion, research so far supports both important similarities and fundamental differences between HL and L2 learners. Let us now examine the two most current teaching models in light of these findings.

(Re)Learning the Heritage Language in a Formal Setting

Why do heritage speakers come to the classroom to learn the heritage language in a formal setting? At the postsecondary level, two main reasons why heritage speakers turn to language classes are to fulfill linguistic and identity needs (in addition to, perhaps, a foreign language requirement). Languages that are well represented in the community also offer career incentives. Despite these general goals, particular challenges arise in the teaching of heritage languages, challenges that have to do with the number of HL and L2 learners interested in learning the target language (i.e., enrollments), the size of the heritage speaker population in the community, and whether or not the target language has a dual status as a foreign, a second, or a heritage language. Two common models currently in place for teaching HL learners are to place both HL and L2 learners in foreign language classes, or to develop dual tracks for L2 learners and heritage speakers if the enrollments and institutional resources warrant it. For example, in areas where Chinese and Japanese have a significant number of community members, they may be taught as dual tracks. French, German, and Italian are typically taught as foreign languages, and where the heritage speaking community is small, HL learners are placed in the same classes. Russian, which has been undergoing declining enrollments as a foreign language, today mostly serves HL learners, especially in urban areas with a high concentration of Russian speakers. South Asian languages and Arabic classes may also combine heritage and non-heritage learners. An additional challenge that arises with these two less commonly taught languages is that many HL learners may be speakers of a different dialect or a cognate language, such as Gujurati speakers learning Hindi.

The FL/L2 model

Typical foreign language classes in North America serve college students whose primary (and very often only) language is English and who may or may not have received some instruction in the target language in high school. Some of the goals of a foreign language program at the postsecondary level are to introduce

students to the language and culture, to help them develop an intermediate to high-intermediate level of communicative competence in the language, to promote critical thinking about their own culture and the culture of the language they are learning, and to develop skills to allow them to continue advanced study in the language, literature, or culture. Some L2 programs also offer language for specific purposes (business, health professions, law, etc.) to promote use of the target language in professional contexts. Informed by research on the nature of language acquisition and the effectiveness of formal instruction, many foreign language programs today adopt a communicative, content-based, and task-based methodology, where students work on listening, speaking, reading, and writing by means of individual, pair, and small-group activities whose main focus is negotiating meaning. Students are introduced to the culture of the language as outsiders, and while the topics covered in class are appropriate for college students, the vocabulary and structures presented are very elementary at first, increasing in level of difficulty in a two- to four-year period. Although the teaching and practice of grammar does not take center stage, it is not completely ignored either, since the linguistic objectives of the course must be met as well.

Until now, HL learners have individually enrolled or have been formally placed in the available FL classes, especially when the number of HL learners is very small, or the institution does not offer another option. But how suitable are FL programs for HL learners? This question only becomes significant if student demand is high and if the educational institution has the means to make changes accordingly. Assuming the number of HL learners is substantial, the answer depends on the particular linguistic profile of the HL learners and their level of proficiency in the language. Unlike L2 learners, who typically have similar proficiency in the four skills, HL learners tend to have uneven proficiency in different areas: oral vs. written skills, vocabulary, idiomatic expressions, and some aspects of grammar. Those speakers who have some command of the language are usually bored in classes that start teaching the language from zero. Furthermore, mixing the two types of students raises other affective issues, and this has been investigated by Potowski (2002). L2 learners who have poor oral skills feel intimidated by the advanced oral and pronunciation skills of HL learners. Similarly, HL learners, who may feel that they have a good command of the language, are surprised when they do poorly on written tests requiring them to know the labels for grammatical structures that they can use without reflection. Therefore, HL learners who have a certain command of the language might not be well served by the same approach and curriculum designed for L2 learners. Furthermore, there is the issue of available materials and resources. Textbooks for FL/L2 learners may be cognitively and age-appropriate for HL learners in terms of content, but are not linguistically challenging when they focus on basic pronunciation, grammar, and vocabulary.

This does not mean that a FL/L2 curriculum is always unsuitable. HL learners with no command, or close to minimal command, of the language might be better served by an FL program, since, as Kondo-Brown (2005) showed, HL learners who are not speakers of the language pattern with L2 learners on a number of

proficiency measures. Because these particular HL learners will not have any linguistic advantage or disadvantage as compared to L2 learners, the affective issues that arise in mixed classrooms due to disparate proficiency levels may not be relevant in this case. The underlying theoretical assumption behind this practice is that L2 and HL learners are very similar in the types of linguistic knowledge they lack and, being adults, in the way they learn the target language. Therefore, they can benefit from the same instructional approach. Nevertheless, while this model may fulfill the linguistic proficiency goals of students with minimal previous knowledge of the language, it may do little to fulfill the identity goal.

Dual tracks for L2 learners and HL learners

HL practitioners who recognize that HL learners are very different from typical L2 learners advocate different classes and curricula for the two types of learners (Carreira, 2004; Kagan & Dillon, 2003; Potowski, 2005; Potowski & Carreira, 2004; Valdes, 1997, 2001), and a considerable number of postsecondary institutions offer separate classes, or dual tracks, for L2 and HL learners.5 This tendency is growing. Specialized classes for heritage speakers are typically offered at the elementary and intermediate levels of instruction for languages like Spanish, Chinese, and Japanese. An underlying assumption of this model is that bilingual HL learners are different from L2 learners and more similar to L1 learners (due to their linguistic and cultural past). As such, they are able to learn the target HL at a faster pace than L2 learners, even if their initial proficiency in the language is quite low. This assumption has not yet been empirically verified, however.

There appears to be agreement that the communicative and content-based approach currently in vogue for L2 programs is also ideal for HL classes (Krashen, 2000; Lynch, 2003). A key difference between FL and HL classes is the cultural content, with HL classes focusing on the cultural issues, interests, and social problems affecting the HL community (poverty, discrimination, immigration, separation, diaspora, social, political and educational consequences of bilingualism, clash of cultures, etc.). Moreover, the communicative activities implemented in the classroom (through literature, music, videos) aim to reaffirm the learners' cultural identity. As for fulfilling the linguistic content, Potowski (2005) argues that specialized classes for HL learners should more closely follow the language arts curriculum taught to monolingual elementary school children. That is, strong emphasis should be placed on vocabulary expansion and enrichment, reading, writing, and spelling, in addition to basic grammatical concepts. Furthermore, the sociolinguistic dimension of language, including a discussion of regional dialects and registers and the linguistic diversity in the target language-speaking world, is crucial for promoting linguistic awareness in both HL teachers and learners.

While this model offers specialized instruction for HL learners and seeks to meet both identity and linguistic needs, it faces practical challenges, especially with regard to instructional materials. Language arts textbooks designed for elementary school children learning their first language may be appropriate from the point of view of linguistic coverage, but are unsuitable because the content

is not age-appropriate or challenging for college level students. Although some textbooks already exist for Spanish as a heritage language, little is available for other heritage languages.

Another problem is that the dual track program is currently implemented only at the basic level of language instruction (elementary and intermediate). In many language departments that offer a major in the language, HL learners who elect to continue advanced study are once again placed in the same classes as L2 learner majors. As the research discussed above showed, advanced HL learners have gaps in their linguistic knowledge, especially when it comes to knowledge of registers and dialectal variation. HL learners who possess advanced proficiency in informal, non-standard language varieties (what they acquired at home) exhibit low proficiency in formal, standard, and academic language varieties in oral and written discourse (Valdes & Geoffrion-Vinci, 1998). This is especially manifested in languages that use a complex system of honorifics and forms of address ( Japanese, Korean, the formal and informal second-person difference in Spanish). HL learners of these languages either do not know these forms or use them randomly. Thus, ideally, more linguistic support for HL learners at the advanced level should be provided.

A third possibility discussed by Carreira (2004), and which heritage languages with fewer L2 and HL learners than others may already be implementing, is to create a hybrid model. This hybrid model takes a HL approach for all courses that enroll HL learners, without making the distinction between L2 and HL classes. All language classes ranging from beginner to advanced would incorporate cultural topics of interest to the two groups of learners, together with more specialized instruction on problem areas for the two groups (e.g., pronunciation for L2 learners and spelling for HL learners). This solution might be pedagogically and administratively ideal for serving the two types of learners, but it would also necessitate reconceptualizing teaching materials and resources.

Placement and proficiency tests

Once a program offers distinct courses and curricula for students with different linguistic and cultural profiles, questions arise over the specific criteria that should be used to place students in those classes. It is common for L2 learners to take a locally developed placement test (Kondo-Brown, 2005), or an ACTFL interview, specifically designed to place L2 learners. Unfortunately, placement procedures for identifying different levels of heritage speaker ability are anything but straightforward. In extending the ACTFL proficiency guidelines to heritage students, Valdes (1997) was able to identify 14 possible proficiency outcomes, clearly showing that heritage speakers do not fit neatly in the guidelines developed for L2 learners. The least proficient HL learners (in a written task) have the intermediate or high intermediate skills of a foreign language learner. And while virtually all HL learners pass the criteria of comprehensibility (their phonology is typically very good), they may lack proficiency in vocabulary and in a range of grammatical areas that go undetected by this measure. Kagan and Dillon (2003) advocate

using the linguistic assessment offered by the Foreign Language Standards, instead, which takes into account different communicative modes (or registers): interpersonal, interpretive, and presentational. HL learners may demonstrate command of the interpersonal mode, but lack experience with the interpretive and presentational domains, since they have not had experience using the language in an academic setting.

Finally, HL practitioners and researchers have found it useful to develop extensive language background questionnaires to understand the linguistic past of heritage speakers and their family circumstances. This information also becomes important for assessing the degree of linguistic ability in the HL acquired in the home. Combined with a proficiency test, this information helps identify both linguistic gaps and professional goals of HL learners, and helps HL practitioners decide on the best placement for each learner.

Conclusion

I have discussed two main models for the teaching of heritage languages, and the underlying theoretical assumptions behind them. Due to the infancy of the field of heritage language acquisition and teaching, there is ample room for other teaching models. While the current models are consistent with what research findings have so far uncovered about the nature of heritage language competence and its potential for acquisition and development, more theoretically informed psycholinguistic studies as well as classroom-oriented research on HL learners are needed to test their true pedagogical effectiveness and to learn more about the HL (re)acquisition process in general.

I thank Melissa Bowles, Mike Long, Diane Musumeci, Kim Potowski, Marc Thompson,

and Gabriela Zapata for invaluable feedback. All remaining errors are my own.

1 For the purposes of this chapter, a HL learner is a heritage speaker relearning the HL in a language class.

2 The term is also used to refer to indigenous languages in the US, even though aboriginals and immigrants vary significantly in their historical, social, linguistic, and demographic characteristics (Fishman, 2001; Wiley, 2001). In this chapter, I will only focus on the acquisition and learning of immigrant languages. For issues related to learning and revitalization of indigenous languages, see Hornberger (2005).

3 Attrition is the process by which aspects of language that were fully acquired at an earlier age are now eroded and eventually lost.

4 Other researchers consider that L2 learners have full access to Universal Grammar, like L1 learners, and deny the implication of maturational constraints (see White, 2003). An elaboration of this position is not relevant for this chapter.

5 This is almost 20 percent of all US universities for Spanish (Ingold et al., 2002).

REFERENCES

Au, T., Knightly, L., Jun, S., & Oh, J. (2002). Overhearing a language during childhood. Psychological Science 13, 238-43.

Bley-Vroman, R. (1989). The logical problem of second language learning. In S. Gass & J. Schachter (eds.), Linguistic perspectives on second language acquisition (pp. 41-68). Cambridge: Cambridge University Press.

Carreira, M. (2004). Seeking explanatory adequacy: A dual approach to understanding the term "Heritage Language Learner." Heritage Language Journal 2, 1. www. heritagelanguages.org.

Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris.

Coppieters, R. (1987). Competence differences between native and near-native speakers. Language 63, 544-73.

Cummins, J. (2005). A proposal for action: Strategies for recognizing language competence as a learning resource within the mainstream classroom. Modern Language Journal 89, 585-91.

DeKeyser, R. (2000). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition 22, 499-534.

DeKeyser, R. (2003). Implicit and explicit learning. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 313-48). Malden, MA: Blackwell.

Fishman, J. (2001). 300-plus years of heritage language education in the United States. In J. Peyton, D. Ranard, & S. McGinnis (eds.), Heritage languages in America (pp. 81-99). McHenry, IL: Center for Applied Linguistics.

Guillelmon, D. & Grosjean, F. (2001). The gender marking effect in spoken word recognition: The case of bilinguals. Memory and Cognition 29, 3, 503-11.

Hornberger, N. (ed.) (2005). Heritage/community language education: US and Australian perspectives. Special issue of International Journal of Bilingual Education and Bilingualism 8, 2 & 3.

Ingold, C., Rivers, W., Tesser, C., & Ashby, E. (2002). Report on the NFLC/AATP Survey of Spanish Language Programs for Native Speakers. Hispania 85, 2, 324-9.

Kagan, O. & Dillon, K. (2003). A new perspective on teaching Russian: Focus on the heritage learner. Heritage Language Journal 1, 1. www.heritagelanguages.org.

Kanno, K., Hasegawa, T., Ikeda, K., Ito, Y., & Long, M. H. (2008). Relationships between prior language-learning experience and variation in the linguistic profiles of advanced English-speaking learners of Japanese. In D. Brinton, O. Kagan, & S. Bauckus (eds.), Heritage language education: A new field emerging. New York: Routledge.

Kim, J.-H., Montrul, S., & Yoon, J. (in press). Dominant language influence in acquisition and attrition of binding: Interpretation of the Korean reflexive caki. Bilingualism: Language and Cognition.

Kohnert, J., Bates, E., & Hernández, A. (1999). Balancing bilinguals: Lexical-semantic production and cognitive processing in children learning Spanish and English. Journal of Speech, Language and Hearing Research 42, 1400-13.

Kondo-Brown, K. (2003). Heritage language instruction for post-secondary students from immigrant backgrounds. Heritage Language Journal 1, 1. www.heritagelanguages.org.

Kondo-Brown, K. (2005). Differences in language skills: Heritage language learner subgroups and foreign language learners. Modern Language Journal 89, 563-81.

Kopke, B. & Schmid, M. (2003). Language attrition. The next phase. In M. Schmid, B. Kopke, M. Keijzer, & L. Weilemar (eds.), First language attrition (pp. 1-43). Amsterdam: John Benjamins.

Krashen, S. (2000). Bilingual education, the acquisition of English and the retention and loss of Spanish. In A. Roca (ed.), Research on Spanish in the United States: Linguistic issues and challenges (pp. 432-44). Sommerville, MA: Cascadilla Press.

Lardiere, D. (2006). Ultimate attainment in second language acquisition: A case study. Mahwah, NJ: Lawrence Earlbaum.

Lee, Y.-G., Kim, H.-S., Kong, H., Hong, J.-M., & Long, M. (2008). Variation in the linguistic profiles of advanced English-speaking learners of Korean. In D. Brinton, O. Kagan, & S. Bauckus (eds.), Heritage language education: A new field emerging (pp. 165-80). New York: Routledge.

Long, M. H. (2003). Stabilization and fossilization in interlanguage development. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 487536). Malden, MA: Blackwell.

Lynch, A. (2003). The relationship between second and heritage language acquisition: Notes on research and theory building. Heritage Language Journal 1, 1. www.heritagelanguages.org.

Matsunaga, S. (2003). Instructional needs of college-level learners of Japanese as a heritage language: Performance based analysis. Heritage Language Journal 1, 1. www. heritagelanguages.org.

Montrul, S. (2002). Incomplete acquisition and attrition of Spanish tense/aspect distinctions in adult bilinguals. Bilingualism: Language and Cognition 5, 1, 39-68.

Montrul, S. (2004). Convergent outcomes in second language acquisition and first language loss. In M. Schmid, B. Kopke, M. Keijzer, & L. Weilemar (eds.), First language attrition (pp. 259-80). Amsterdam: John Benjamins.

Montrul, S. (2005). Second language acquisition and first language loss in adult early bilinguals: Exploring some differences and similarities. Second Language Research 21, 3, 199-249.

Montrul, S. (2006). Incomplete acquisition as a feature of L2 and bilingual grammars. In R. Slabakova, S. Montrul, & P. Prévost (eds.), Inquiries in language development (pp. 335-59). Amsterdam: John Benjamins.

Montrul, S., Foote, R., & Perpinân (2008). Gender agreement in adult second language learners and Spanish heritage speakers: The effects of age and context of acquisition. Language Learning 58, 3, 503-53.

Montrul, S., Foote, R., Perpinân, S., Thornhill, D., & Vidal, S. (2006). Full access and age effects in adult bilingualism: An investigation of accusative clitics and word order. In N. Sagarra & J. Toribio (eds.), Selected papers from the Hispanic Linguistics Symposium (pp. 217-28). Sommerville, MA: Cascadilla Press.

Oh, J., Jun, S., Knightly, L., & Au, T. (2003). Holding on to childhood language memory. Cognition 86, B53-B64.

Paradis, M. (2004). A neurolinguistic theory of bilingualism. Amsterdam: John Benjamins.

Polinsky, M. (1997). American Russian: Language loss meets language acquisition. In W. Browne, E. Dornisch, N. Kondrashova, & D. Zec (eds.), Proceedings of the Annual Workshop on Formal Approaches to Slavic Linguistics (pp. 370-406). Ann Arbor: Michigan Slavic Publications.

Potowski, K. (2002). Experiences of Spanish heritage speakers in university foreign language courses and implications for teacher training. ADFL Bulletin 33, 3, 35-42.

Potowski, K. (2005). Fundamentos de la enseñanza del español a hispanohablantes en los EE. UU. Madrid: Arco Libros.

Potowski, K. & Carreira, M. (2004). Towards teacher development and national standards for Spanish as a heritage language. Foreign Language Annals 37, 3, 427-37.

Selinker, L. (1972). Interlanguage. IRAL 10, 3, 209-31.

Silva-Corvalán, C. (1994). Language contact and change: Spanish in Los Angeles. Oxford: Oxford University Press.

Song, M., O'Grady, W., Cho, S., & Lee, M. (1997). The learning and teaching of Korean in Community Schools. Korean Language in America 111-27. American Association of Teachers of Korean.

Sorace, A. (1993). Incomplete vs. divergent representations of unaccusativity in native and non-native Italian grammars of Italian. Second Language Research 9, 22-47.

Valdés, G. (1997). The teaching of Spanish to bilingual Spanish-speaking students: Outstanding issues and unanswered questions. In M. C. Colombi & F. Alarcón (eds.), La enseñanza del español a hispanoahablantes (pp. 8-44). Boston, MA: Houghton Mifflin.

Valdés, G. (2001). Introduction. Spanish for native speakers, vol. 1. AATSP Professional Development Series Handbook for teachers K-16. New York: Harcourt College.

Valdés, G. (2005). Bilingualism, heritage language learners, and SLA research: Opportunities lost or seized? Modern Language Journal 89, 410-26.

Valdés, G. & Geoffrion-Vinci, M. (1998). Chicano Spanish: The problem of the "underdeveloped" code in bilingual repertoires. Modern Language Journal 82, 4, 473-501.

White, L. (2003). Second language acquisition and universal grammar. Cambridge: Cambridge University Press.

Wiley, T. (2001). On defining heritage languages and their speakers. In J. Peyton, D. Ranard, & S. McGinnis (eds.), Heritage languages in America (pp. 29-36). McHenry, IL: Center for Applied Linguistics.

FURTHER READING

Brinton, D., Kagan, O., & Bauckus, S. (eds.) (2008). Heritage language education: A new field emerging. New York: Routledge.

Gambhir, V. (ed.) (1995). The teaching and acquisition of South Asian languages. Philadelphia: University of Pennsylvania Press.

Kondo-Brown, K. (ed.) (2006). Heritage language development: Focus on East Asian immigrants. Amsterdam: John Benjamins.

Krashen, S., Tse, L., & McQuillan, J. (eds.) (1998). Heritage language development. Culver City, CA: Language Education Associates.

Montrul, S. (2008). Incomplete acquisition in bilingualism: Re-examining the age factor. Amsterdam: John Benjamins.

Peyton, J., Ranard, D., & McGinnis, S. (eds.) (2001). Heritage languages in America. McHenry, IL: Center for Applied Linguistics.

Valdés, G., Fishman, J., Chavez, R., & Pérez, W. (2006). Developing minority language resources: The case of Spanish in California. Clevedon, UK: Multilingual Matters.

13 Study Abroad Research: Findings, Implications, and Future Directions

JOSEPH COLLENTINE

One of the most important variables that affects the nature and the extent to which learners acquire a second language (L2) is the context of learning, that is, whether the learning takes place within the society in which the L2 is productive or where the first language (L1) is productive. Second language acquisition (SLA) research takes place in three primary contexts, which differ in sociological and functional terms. The foreign-language (FL) classroom exists in the domestic setting of the L1, and learners tend to use the L2 within the classroom and as it relates to academic purposes. The intensive domestic immersion setting is different from the FL context in that students dedicate the majority (if not all) of their academic term to studying the L2, and this context often entails an increase in the functional purposes of the L2 when learners sign a "contract" not to use the L1. The study abroad (SA) context takes place in countries where the L2 enjoys an important sociological and functional status, entailing a combination of planned curriculum and a host family. This chapter provides an overview of the state of the art of SA research. This review uncovers a pattern not clearly articulated to date: specifically, the most salient domains of interest (e.g., cognitive, pragmatic, sociolinguistic factors) that arise out of the SA literature are quite different from the salient domains in other contexts of learning. Informed by this literature, I conclude with recommendations as to the ideal design of SA programs in a modern FL curriculum.

It is extremely challenging for researchers to isolate the effects of the learning context on acquisition because one must be aware that, within any of these situations, learners acquire the L2 in two sub-contexts: communicative contexts as well as learning contexts (Batstone, 2002). Communicative contexts require that the learner use the L2 to exchange information and engage in essential social and interpersonal functions. Learning contexts manage input and output so that learners will attend to form and take intentional steps toward improving their linguistic expertise. The influence of these covariates complicates the assessment of a learning context since the FL classroom heavily favors learning contexts, intensive domestic immersion settings attempt to provide both communicative and learning contexts, and SA presumably provides more opportunities for

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 Bkckwell Publishing Ltd. ISBN: 978-1-405-15489-5

processing and using the L2 in communicative contexts. SA learners must determine the relationship between the L1, the L2, and their identities as social individuals and language learners. For researchers, the complication then becomes which theoretical frameworks can capture the cognitive and social developments through which learners pass in the SA setting. As Collentine and Freed (2004) note, research in a SA context provides an important contextualization for understanding the interaction between cognitive, sociolinguistic, and socio-cultural factors in the construction of a comprehensive theory of SLA. This chapter attempts to delineate the contribution of SA research to SLA theory by exploring its (brief) history as a sub-discipline. The chapter also details the types of populations that researchers have studied, the efficacy of this learning context, how researchers operationalize "efficacy" (and the effects on the interpretation of results), the roles of input and interaction, the cognitive changes (e.g., phonological memory) that occur abroad, important issues identified in the research relating to learner identity, and the role of pedagogy in SA contexts on acquisition.

A Brief History of Study Abroad Research

SA research can be seen as having two periods. The first attempted to understand the overall efficacy of SA programs. These studies concentrated on measuring the gains learners make abroad largely from broad measurement instruments. This period extends from the 1960s to Barbara Freed's publication of her seminal volume Second language acquisition in a study abroad context in 1995 (Freed, 1995b). Freed succeeds in framing SA research within the SLA theory-building enterprise, challenging researchers to view SA research as a means of studying the effects of "learning context" on acquisition.

The first period examined gains (or simply post-treatment abilities) with instruments that sought to assess learners' overall L2 abilities.1 Carroll's (1967) widely-cited study looked at the language skills of 2,782 college seniors on tests that measured linguistic skills in the L2 (i.e., their metalinguistic knowledge), finding that even a short duration abroad (touring or summer) predicted higher levels of proficiency. Willis et al. (1977) summarized a series of studies on British students, concluding that these studies lacked an overall systematic assessment of learners gains. Willis et al.'s (1977) own study on British students in France and Germany showed general support for residency abroad. The strongest gains were in listening and speaking and less in reading abilities. Dyson (1988) conducted another macro study on 229 British students in France, Germany, and Spain, showing that the learning context improved listening and speaking skills. Opper, Teichler, and Carlson (1990) conducted a large-scale study on the efficacy of SA on students in more than 80 programs in Britain, France, Germany, and Sweden. The study is limited in its validity because it relies on self-reported assessments of general language skills (listening, reading, speaking, and writing) whereby the learners generally reported important gains. It provided nonetheless an important clue to a key factor that has been identified in the SA research: SA appeared to be particularly powerful for learners with lower levels of proficiency.

Perhaps the key observation to be gleaned from this study is the notion that there are threshold levels of development at which SA will be optimally beneficial (discussed further below). Another key study of this period that laid the foundation for current SA research was Mohle and Raupach (1983), who found SA not to have an important effect on improving morphosyntactic abilities but to have a positive effect on fluency, a factor in SLA research that has become an important focus of late since it helps us to understand the interaction between online processing mechanisms and linguistic competence such as working and phonological memory (see Segalowitz, 2003). The final noteworthy studies of this period are Brecht and Davidson (1991) and Brecht, Davidson, and Ginsburg (1995), who examined 668 American learners' acquisition of Russian in a SA context. The data reinforced a growing hunch that individual differences (e.g., reading aptitude) are exceptionally evident in this learning context and that preprogram explicit grammar instruction predicts gains abroad.

Freed (1995a) is the first effort to synthesize SA research. She identifies a number of issues that researchers in her volume and others have addressed since its publication. She notes a growing suspicion that the linguistic benefits of the SA context were not the same as those of the traditional classroom (Regan, 1995), and there were surprisingly few empirical studies that actually compared these two learning contexts. She also recognizes the hypothesis that there might be a proficiency threshold at which learners most benefit from SA (Brecht & Davidson, 1991; Brecht, Davidson, & Ginsburg, 1995; Regan, 1995). Additionally, there was almost no research that examined SA gains within current theoretical frameworks that dealt with how new linguistic information becomes internalized in the learner's competence (e.g., interactionist, input-oriented, and socio-cultural models). Freed (1995b) presents studies addressing two issues in her volume, providing important SA-AH (at home) studies comparing orders of acquisition abroad to orders documented in FL contexts (Guntermann, 1995; Lafford, 1995). Freed (1995a) also provides studies on the acquisition of pragmatic competence within a sociolinguistic framework (Regan, 1995; Siegal, 1995).

Since Freed (1995b), noteworthy collections of SA experiments have been published. And, while these collections have concentrated on Americans going abroad (Collentine & Freed, 2004; DuFon & Churchill, 2006; Gore, 2005; Pellegrino-Aveni, 2005), Murphy-Lejeune (2002) examines SA in the European context. What follows delineates the key topics that SA research is currently addressing. The research addresses a variety of issues that relate to the internal cognitive mechanisms affecting acquisition and the external sociolinguistic mechanisms, as well as the socio-cultural issues of SLA in a SA context of learning.

Populations of Study and the Threshold Hypothesis

Kinginger (2007), as well as Coleman (1997), characterizes the existing body of SA research as falling into two categories, with each focusing on different populations and distinct levels of development. Research on American, university-level learners tends to examine acquisition at the beginning stages of

development, with study participants sampled from first- and second-year language programs. While there is a good amount of research on learners at the beginning levels of development, Lafford and Collentine (2006) report that SA research on American students of Spanish has sampled participants whose preprogram proficiency ranged from the novice to the advanced high levels. It is true that there exists little information about American learners' success in advanced-level, direct-enrollment programs (i.e., where students sit in classes with otherwise proficient/native speakers of the L2). American, university-level L2 programs - and so SA curricula - concentrate on fostering acquisition at the financially lucrative early stages of acquisition (where FL enrollments in general are highest). Kinginger (2007) asserts that SA research such as Murphy-Lejeune (2002) has focused on European learners at more advanced levels, stemming from the European Commission's inter-university programs, such as the ERASMUS program, which allows students to complete part of their university studies in another EU country and university (41 percent of such students study language or philology; http://www.erasmus.ac.uk/statistics/contents_04_05.htm).

There is a growing interest in SA SLA issues along the Pacific Rim, and these learners - Japanese, Chinese, Korean, and Taiwanese university-level adults slightly beyond the initial stages of development - tend to target English as their L2, either in North America or the South Pacific (Churchill, 2006; Tanaka & Ellis, 2003). If current estimates that 80 percent of foreign students in the world are from Asian countries are correct (Altbach & Bassett, 2004), these populations will become an increasingly important source of data.

There are important pockets of SA contexts and learner profiles that have yet to be studied. Students are attending so-called language camps with increasing frequency. Korean.net, for instance, reports that nearly 40,000 Korean students enrolled in domestic EFL immersion programs in 2005.2 The New York Times recently reported that nearly 75 percent of American summer camps have foreign nationals attending their activities, and many camps provide some EFL instruction (Bick, 2007). To this researcher's knowledge, there is no existing literature either documenting these experiences or providing data on their value, although the usual unsubstantiated and anecdotal claims of efficacy are not hard to uncover.

Learners of all levels of development (who are largely self-selecting) have been studied in the SA literature. However, Lafford and Collentine (2006) discuss the growing consensus amongst researchers that there is a threshold which learners must reach to benefit fully from the SA context of learning. And, while the general notion of a threshold level is important for program designers to keep in mind, from a linguistic competence perspective it is probably too broad in scope. There are most likely specific domains that require a particular developmental threshold for overall gains to occur. Golonka (2006) - refining the Brecht, Davidson, and Ginsberg (1995) analysis - presents evidence suggesting that preprogram linguistic (grammar, vocabulary, accuracy) and metalinguistic (self-corrected errors and sentence repair) levels predicted which SA learners of Russian would attain the Advanced level on the ACTFL proficiency scale. Segalowitz and Freed (2004) found that, amongst Spanish L2 learners, an initial threshold level of basic word recognition and lexical access processing abilities may be necessary for oral

proficiency and fluency to develop abroad significantly. Lafford (2004) surmises that advanced learners bring more formulaic expressions to SA communicative contexts and can therefore spend more attention resources on form, whereas novices must attend primarily to meaning. Finally, Segalowitz and Freed (2004) report that the most important gains that SA learners of Spanish make abroad are in the domain of fluency (as measured by temporal/hesitation phenomena), and O'Brien et al. (2007) report that these fluency gains abroad are a function of students' potential for phonological memory storage, which varies from adult to adult. While this last factor is not controllable from a preprogram perspective, it argues that a student's cognitive and linguistic abilities will mediate developmental gains in the program.

The Issue of Study Abroad Efficacy

Many researchers and educators have surmised that the SA context is the sine qua non for achieving global L2 competence (Rivers, 1998). Freed (1995a) acknowledged that, while some empirical studies conducted up until 1995 provided evidence that SA facilitates acquisition (e.g., DeKeyser, 1991; Teichler & Steube, 1991), a sizable amount of evidence challenged researchers to consider whether the study abroad context might impede acquisition at the beginning stages (e.g., Freed, 1990; Spada, 1985, 1986), which, again, points to a threshold effect. Cohen and Shively (2007) summarize the research to date, asserting: "An intriguing finding in the study abroad research literature that prompted the current study was that study-abroad students do not necessarily achieve greater language gains than their peers who stay home and study the target language" (p. 189). Clearly, overall efficacy is difficult to assess in the absence of data stemming from large-scale studies such as those reported by Brecht and Davidson (1991) and Segalowitz et al. (2005). Yet, SA studies tend to be longitudinal in nature, as opposed to much SLA research that examines short-term effects. These studies support Carroll's (1967) initial assertion about the advantages of SA, but they require a great deal of qualification: whereas SA affects gains in certain language-specific domains, it does not affect development in all aspects of a learner's competence. Interestingly, linguistic aspects that do indeed seem to benefit from SA, such as fluency and discursive abilities, are often not those in which AH FL program directors hope to see improvements, such as those grammatical aspects around which the AH, focus-on-forms syllabus is designed. The following two sections summarize what we know to date about the important cognitive and linguistic aspects of L2 development with which SA interacts.

Important Cognitive Constructs in the Study Abroad Literature

One consistent theme in the SA literature is individual differences and individual variation (variation is addressed in the next section). An area of SLA research

that is making important contributions to our understanding of individual differences - or, between-subject variation - examines the interaction of short-term memory stores (e.g., working memory, phonological memory), speed of information access, attention control, and acquisition (see DeKeyser, 2001).

Segalowitz and Freed (2004) are interested in the cognitive processing abilities that underlie "expert abilities." They present data indicating that cognitive abilities interact with development abroad in complex ways. They found lexical access to be related to overall proficiency gains amongst SA learners of Spanish, such that learners who access lexical items faster show greater gains. They also present data suggesting that learners exhibiting greater attention control at the completion of their SA program spoke less fluently. Segalowitz and Freed (2004) speculate that the increased attention control in the SA context may reflect increased monitoring of output. O'Brien oversees two studies examining the interaction of phonological memory and SA gains in learners of Spanish, both of which indicate that phonological memory assists not only children (in vocabulary gains) but also adults in an L2 context. O'Brien et al. (2006) showed that phonological memory abilities have a positive effect on one's abilities to produce multi-propositional utterances at the early stages of Spanish learners' development (i.e., narrative abilities) in addition to a positive effect on the acquisition of particular grammatical functors at later stages. O'Brien et al. (2007) add to our understanding of the cognitive mechanisms underlying the general fluency gains associated with SA experiences, showing that phonological memory also predicts fluency gains. Finally, Tokowicz, Michael, and Kroll (2004) examine working memory capacity and single-word translation errors, presenting data that suggest that SA learners engage in more approximate than precise translations but that this tendency is limited to those with higher working memory capacities.

This literature suggests that learners' predisposed cognitive abilities determine how much one can produce (and how fast) as a result of SA, and may impact the ways that learners approach attention-demanding processes (such as translations and story-telling). It also indicates that researchers ought to consider carefully how they interpret observable behaviors, such as fluency. The Segalowitz and Freed (2004) study provides some evidence that where there is a lack of fluency, there may be more monitoring occurring (due to greater attention control), which is not an unreasonable conclusion, since Golonka (2006) shows a strong relationship between metalinguistic abilities and overall proficiency gains.

Important Linguistic Constructs in the Study Abroad Literature

As mentioned above, there has been a growing concern about the overall lack of efficacy of SA on acquisition (cf. Cohen & Shively, 2007). In the following I show that, while it is not accurate to claim a superiority for SA on fostering acquisition, it is likewise erroneous to conclude that important, positive changes

(toward native-speaker norms) do not occur as a result of the SA context of learning. The developmental growth one finds in this learning context appears not to be what one finds in the AH setting (although the field admittedly has a dearth of comparative AH-SA studies) nor what learners and their AH teachers expect (e.g., dramatically improved grammatical and phonological accuracy). The linguistic-specific growth that occurs is in the domain of global discursive abilities, the expansion of knowledge of verb paradigms, and in the domain of pragmatics. A review of this research reveals much about the communicative needs and demands placed on learners in the SA context.

One of the consistent conclusions that researchers have drawn from the SA literature has been that the development of grammatical abilities does not seem to outpace that of the AH context. DeKeyser (1990, 1991) found that residence abroad had little impact on the development of overall grammatical abilities and that SA learners were equal to or inferior to their AH counterparts in their use of grammar. Collentine (2004) gauged SA learners' acquisition of a variety of morphosyntactic features, showing that they do not make as much progress as AH learners on precisely those grammatical aspects that many FL teachers emphasize, namely, verbs and subordinate conjunctions.

These studies, in sum, indicate that the appreciable development of morpho-syntax and general grammatical abilities is not to be expected, at least within the timeframe of a semester to a year abroad. Indeed, two of these studies (Collentine, 2004; DeKeyser, 1990) suggest that the AH experience affords certain advantages as regards overall grammatical development for intermediate learners. A notable exception is Isabelli and Nishida (2005), who reveal that SA has an advantage with respect to subjunctive development when learners are at more advanced stages, thus supporting the threshold hypothesis.

A further complication to this scenario arises out a comparison of Golonka (2006), who found preprogram metalinguistic knowledge to predict proficiency gains, and Izumi and Iwasaki (2004). The latter examined the effects of amount of SA experience on a grammaticality judgment test by Japanese-speaking English FL learners, where the participants were asked to give reasons for their evaluations. The purely classroom EFL learners used intuitive, analyzed, and metalinguistic knowledge more or less equally, while the learners who experienced living abroad for several years used intuitive knowledge and very little metalinguistic knowledge. It may be that preprogram metalinguistic knowledge is an important prerequisite, but that not all learners depend on that knowledge in communicative situations to process grammar. It may also be that, as the Segalowitz and Freed (2004) data suggest, some processes become more automatized abroad, so that metalinguistic awareness is converted into implicit grammatical knowledge. Clearly, more research is needed in this area.

Research is starting to suggest that the organizing principles around which SA learners develop their grammatical abilities stem from the discursive demands they face in communicative contexts. This ought not to be surprising, given what we have known for a while about uninstructed SLA. The Second Language Acquisition by Adult Immigrants study was conducted from 1981 to 1988 in five

European countries by the Max Planck Institut für Psycholinguistik, mostly on L2 learners of German (Perdue & Klein, 1992), showing that L2 grammaticaliza-tion arises out of functional pressures to achieve discursive coherence. Collentine (2004) used corpus-based techniques (see Biber, 1988) to compare the development of narrative abilities, as well as growth in the semantic density, of AH and SA learners of Spanish, revealing that the SA context afforded a significant advantage. Similarly, Cheng and Mojica-Diaz (2006) compared SA learners of Spanish on their subjunctive abilities (using a native-speaker baseline), finding no significant improvement after a two-month period, although they report some learners started producing more tightly structured argumentation over time.

Fine-grained analyses of SA learners' grammatical performance over time suggest another plausible explanation for why researchers in the past have not found an appreciable quantifiable advantage for SA context in grammatical development: the SA context fosters grammatical variation. Regan (1995) examined SA learners of French and their use of the ne negation particle, showing that AH learners tended to adopt a single, standardized construct for negation whereas their SA counterparts varied between the inclusion and omission of the ne particle as a function of sociolinguistic factors. Howard (2002, 2006), who has examined SA Irish learners of French, has presented convincing evidence that SA learners spend much time varying within and accommodating the individual elements of a paradigm (e.g., present-tense French inflections). His data also suggest that L2 phonological development interacts with the increased inflectional variability (or perhaps inflectional confusion) in the SA context, which is not surprising given that inflectional morphology tends to comprise short phonemic segments (and suffers neutralization in modern French; cf., Howard, Lemee, & Regan, 2006).

Some researchers suspect that one of the reasons for a weak SA grammar effect is that learners have much less access to the L2 than one might suspect, especially as it relates to its pragmatic features. Native speakers find it difficult to abandon certain mentor-apprentice modes of interaction with learners (Pellegrino-Aveni, 2005; Wilkinson, 2002). Barron (2003), examining Irish SA learners of German, argues that learners form social networks with other speakers of their L1, limiting their access to native speakers. Others suggest that learners are not adequately aware of the rules of pragmatics within the target culture (Wilkinson, 2002). Cohen and Shively (2007) present data from a controlled experiment suggesting that preprogram efforts to raise French and Spanish learners' awareness of speech acts had only a marginal effect on their ability to mitigate the intensity of requests and no significant effect on their ability to recognize the appropriateness of other acts, such as apologies.

Some recent research on SA pragmatics reinforces O'Brien et al.'s (2006) position that SA has differential effects on learners depending on their preprogram level of development.3 Shardakova (2005) reports that SA learners of Russian with low preprogram proficiency adopt culturally appropriate apologies, while more advanced candidates develop their own strategies, which are not consistent

with cultural prescriptions, reflecting the extent to which linguistic behaviors in a SA environment can be a function of identity as much as input and types of interaction (see below for more on the notion of identity). This pattern of tension between pragmatic appropriateness and rejection of target cultural (linguistic) practices in terms of routines and speech acts is a recurring theme in the SA literature (DuFon & Churchill, 2006). This may account for the general consensus amongst researchers of SA pragmatics that, while SA learners outpace their AH counterparts in pragmatic development, they may spend years attaining native-speaker behaviors, if they so choose. All told, pragmatic competence seems to develop quite slowly in the SA context (Hoffman-Hicks, 1999; Rodriguez, 2001). Protracted pragmatic development is not surprising since learning new scripts/ discourse grammar (i.e., in the case of speech acts) and illocutive meanings alongside locutive ones (i.e., double meanings for certain phrases and constructs, such interrogatives that represent imperatives) represents a considerable task.

This review indicates that grammatical and pragmatic development abroad becomes complicated by the sociocognitive and socio-cultural pressures that learners face in the SA context, a situation that sends many more messages to learners than does the AH context as to the complete repertoire of skills and behaviors one needs to be communicatively functional. Communicative demands at the discursive, as opposed to the sentential, level may well force the learner to make adjustments to his or her preprogram, personal or internal syllabus (Lafford, 2004). If linguistic variation is more prevalent in the SA context than the AH context, SA learners may need to incorporate non-standard forms into their development competence. This greater variability in the SA input may force the SA learner to accommodate the L2 forms with which AH learners may only need to (re)familiarize themselves at test time. Finally, pragmatic development seems to be slow in the SA context, yet important gains are made in the SA context. However, issues of (self-) identify may well interact with the extent to which pragmatic norms become adopted by the learner.

The (Assumed) Roles of Input and Interaction

One of the most undisputed assumptions about the SA context is that learners receive vast amounts of input and have numerous opportunities for communicative interaction. There is a tendency in the field to attribute gains that learners make to the enormous amount of available input. However, there has been no attempt independently to document in a fully quantified manner the types of input and interaction that learners have abroad. The assumptions that exist in the literature may be too strong. Pellegrino-Aveni (2005) documents from a qualitative perspective the types of interactions SA learners of Russian have, arguing that self-preservation (e.g., face-saving) needs effectively impede learners' contact with native speakers. There have been attempts to document the types of input and interaction learners have in a SA context, such as the Language Contact Profile (LCP) employed by Segalowitz and Freed (2004). This sort of

assessment tool is helpful, in that it has built-in redundancies, so as to triangulate the reported contact for any given subject. However, it is a self-report, and its validity is limited in the same way all self-reports are (such as in clinical research), where data sets represent self-perceptions and require objective, third-party validation, such as by host families and professors abroad.

The lack of primary data on the amount of input learners receive abroad is, in fact, symptomatic of a larger problem in SA research and in the enterprise as a whole. Large-scale studies of how much input and interaction learners have in AH contexts really do not exist either. Yet, the field of SLA knows a good deal about the effects of input and interaction on acquisition because we have almost three decades of controlled treatment data (Mackey & Goo, 2007; Norris & Ortega, 2001). No concerted research agenda exists either in theory or practice to control the amounts and types of input and interaction learners receive in the SA context. Taking this observation a step further, we are confronted with the conclusion that we do not know how input differs in the SA context, comprehensible or otherwise. Barron (2003), for instance, notes that learners often misinterpret important aural cues and draw erroneous conclusions about the L2, since negative evidence is not available. Additionally, Magnan and Back (2007), using a modified version of the LCP, present data suggesting that the living situation and access to authentic aural media do not predict oral proficiency gains abroad. The field of L1 acquisition has developed a number of tools for documenting what learners are exposed to in naturalistic settings, and SA researchers could use these methods and instruments as a starting point.

McGeeking (2006), studying SA students of Japanese, is one of the few studies to document how learners negotiate for meaning in the home-stay environment, reporting that numerous opportunities exist (see also Dings & Jobe, 2003, as well as Smartt & Scudder, 2004). Yet the opportunities to negotiate appear to be mitigated in the SA context by personal and interpersonal factors. Wilkinson (2002) reports that home-stay families find it difficult to use naturalistic language with their SA learners of French, preferring to use teacher-talk, denying the learners opportunities for authentic input via interaction. Churchill (2006) documents that SA learners of English from Japan vary in the amount of authentic interactions they have depending on length of stay, which in turn is a function of the extent to which the learners are fully integrated into the target learning community, which is difficult in five-week programs.

What is unclear is whether there are more reports that have made it to press in the SA literature than in the AH literature about the interpersonal mitigating factors that impede opportunities for negotiating for meaning. Much research purporting to comment on negotiation opportunities is conducted within a socio-cultural framework (cf. Kinginger, 2007; DuFon & Churchill, 2006), and so the epistemology is not the same as that from which SLA has traditionally examined interaction (i.e., from a largely cognitive perspective; McGeeking (2006) is a notable exception). Until these two epistemologies can achieve a common terminological interface, our understanding of negotiation in SA contexts relative to the existing SLA literature will remain weak.

Identity

Researchers approaching the SA context from a socio-cultural perspective have focused on the individual histories of students and the tension that exists between maintaining individuality, issues of self esteem, worldviews (e.g., social hierarchies), and the need to advance their own development through native-speaker interactions (Wilkinson, 1998). Churchill (2006) documents how the manner in which learners are received (or not) affects the amount of interaction they have. Kinginger (in press) provides an exhaustive summary of the socio-cultural literature relating to SA. She concludes that the most important theme is that, in those moments when learners' personal sense of identity with the target culture (or with the representatives that they know, such as the home-stay family) is distant, or when learners sense that they have too many obstacles to attaining higher levels of proficiency (either through negative interactions or from a sense of linguistic inadequacy), SA learners abandon their role as "language learners," thus impeding the development process.

Programmatic Considerations

There are a number of programmatic implications to be gleaned from the research presented to date. From an administrative perspective, it is difficult for a home institution to have a strong effect on a host institution's syllabi and methodological approaches. SA instructors are often employees of other institutions and adjunct faculty who have less of a stake in the long-term needs of the learner or program. Programs such as the University of Delaware's, where there is a tight integration between the home and abroad curricula and a faculty development program, can serve as models for ensuring greater quality control over pedagogy (cf. Chieffo & Zipser, 2001). The last ten years have seen a noticeable interest in, and research on, the role of tasks and authentic interactions within an L2 curriculum. For instance, Doughty and Long (2003) suggest that Task-Based Language Teaching provides one such curricular framework that, based on SLA research, details the types of interactions in which learners can engage according to developmental level. An exploration of the plausibility and the outcomes of this sort of language program within a SA curriculum might at least provide a principled starting point from which to study the "missing SA methodology" (Lafford & Collentine, 2006).

It is unclear what the ideal duration of a SA program might be, although a consideration of this question raises intriguing questions for SLA researchers. Programs tend to range from about five weeks to a year in length. Expectations about (1) how much development occurs and, more importantly, (2) which aspects of a learner's competence develop must be measured against the observations made above that exponential growth will not occur within such a time frame. In the absence of a solid base of SA-AH comparative studies, it is possible

(naively) to look at SA research solely as a set of longitudinal SLA investigations. This would lead SLA researchers to the conclusion that L2 acquisition is a protracted process. If the field were to imagine (again, naively) that AH research represents the possible effects of pedagogical interventions in a laboratory setting, SA research indicates that sociopragmatic variables and the inherent linguistic variation existing outside this laboratory will essentially increase the "content" that learners must acquire. The SA learner is confronted with functional demands that invite us to view the AH student as one who learns the language for so-called special purposes. As Freed (1995b) notes, it is no longer tenable to consider a student who enrolls in a SA program as a FL learner, since the SA converts him/her into the learner of a L2. Thus, the end goal of the SA student (whether or not the learner completely abandons his/her role as a language learner; cf. Kinginger, in press) is often different from the AH learner's, and the amount of time needed to complete the L2 agenda becomes longer than that needed for the FL agenda.

If there is a consensus, it is that a student may begin a SA program too early in his/her development. The research considered above indicates that a certain level of metalinguistic knowledge (we do not know yet how much) is a prerequisite for developing the L2 abroad (Brecht & Davidson, 1991; Magnan & Back, 2007). At the very least, SA outcomes are sensitive to learners' preprogram competence levels (Golonka, 2006; O'Brien et al., 2006). At this point, it is possible to suggest: (1) the amount and type of preprogram preparation should be studied in greater depth; (2) these preprogram considerations should inform the types of L2 knowledge and levels of development in programs.

The research does not control for the effects of the homestay with host families. A review of the literature indicates that most students stay with a host family. Lazar (2004), however, reports that the actual amount of time that learners spend with their host families varies both in quantity and quality, and these interactions have an appreciable effect on acquisition in general. Lafford (2004) found a significant negative correlation between the amount of time spent talking with host families and the use of communication strategies to bridge communication gaps. The literature reported here indicates that issues of (1) identity and (2) the host family's perceived role as a mentor may determine how much one learns (Wilkinson, 2002).

Concluding Remarks

Most research conducted to date on SA has concentrated on the American university experience, as seen by the above literature review. As Kinginger (2007) notes, other parts of the world such as Europe have a different relationship with the SA experience, depending on their proximity to other languages and the value they place on bilingualism. Perhaps because the field of SLA is still in the early stages of building a theory of acquisition, it may not be surprising that very little research exists on the effects of SA on learners such as those in the ERASMUS

programs that explicitly purport to promote learners' proficiency. We need to learn much more about the effects of SA on advanced learners, especially given that we suspect there is a developmental threshold at which it starts to be generally effective.

Researchers also need to consider the effects of particular teaching strategies and syllabus design on learner development. The knowledge base reviewed above indicates that this area of SA research is still in its descriptive phase. There are good models of curricular design whose impact we should start to investigate.

1 See Freed (1995a) for an extensive, historical overview of the research on SA up to 1995.

2 www.kois.go.kr/news/News/newsprint.asp?serial_no=20050725004.

3 See DuFon and Churchill (2006) for an extensive overview of the research to date on pragmatic research in the SA context.

REFERENCES

Altbach, P. & Bassett, R. (2004). The brain trade. Foreign Policy 44, 30-1.

Barron, A. (2003). Acquisition in interlanguage pragmatics: Learning how to do things with words in a study abroad context. Amsterdam: John Benjamins.

Batstone, R. (2002). Contexts of engagement: A discourse perspective on "intake" and "pushed output." System 30, 1-14.

Biber, D. (1988). Variation across speech and writing. Cambridge: Cambridge University Press.

Bick, J. (2007). Around the campfire, global ghost stories. New York Times, June 21.

Brecht, R. & Davidson, D. (1991). Language acquisition gains in study abroad: Program assessment and modification. Paper presented at the NFLC Conference on Language Testing, Washington DC.

Brecht, R., Davidson, D., & Ginsberg, R. (1995). Predictors of foreign language gain during study abroad. In B. Freed (ed.), Second language acquisition in a study abroad context (pp. 37-66). Philadelphia: John Benjamins.

Carroll, J. (1967). Foreign language proficiency levels attained by language majors near graduation from college. Foreign Language Annals 1, 131-51.

Cheng, A. & Mojica-Díaz, C. (2006). The effects of formal instruction and study abroad on improving proficiency: The case of the Spanish subjunctive. Applied Language Learning 16, 17-36.

Chieffo, L. & Zipser, R. (2001). Integrating study abroad into the foreign language curriculum. ADFL Bulletin 32, 79-85.

Churchill, E. (2006). Variability in the study abroad classroom and learner competence. In M. DuFon & E. Churchill (eds.), Language learners in study abroad contexts (pp. 203-27). Clevedon, UK: Multilingual Matters.

Cohen, A. & Shively, R. (2007). Acquisition of requests and apologies in Spanish and French: Impact of study abroad and strategy-building intervention. Modern Language Journal 91, 189-21.

Coleman, J. (1997). Residence abroad within language study. Language Teaching 30, 1-20.

Collentine, J. (2004). The effects of learning contexts on morphosyntactic and lexical development. Studies in Second Language Acquisition 26, 227-48.

Collentine, J. & Freed, B. (2004). Introduction: Learning context and its effects on second language acquisition. Studies in Second Language Acquisition 26, 153-72.

DeKeyser, R. (1990). From learning to acquisition? Monitoring in the classroom and abroad. Hispania 73, 238-47.

DeKeyser, R. (1991). Foreign language development during a semester abroad. In B. Freed (ed.), Foreign language acquisition: Research and the classroom (pp. 104-19). Lexington, MA: D. C. Heath.

DeKeyser, R. M. (2001). Automaticity and automatization. In P. Robinson (ed.), Cognition and second language instruction (pp. 125-51). Cambridge, MA: Cambridge University Press.

Dings, A. & Jobe, T. (2003). Negotiating obstacles: Extended repair sequences in native/ non native interaction. In P. Kempchinsky & C. Pineros (eds.), Theory, practice, and acquisition: Papers from the sixth Hispanic Linguistics Symposium and the fifth Conference on the Acquisition of Spanish and Portuguese (pp. 371-81). Somerville, MA: Cascadilla.

Doughty, C. & Long, M. (2003). Optimal psycholinguistic environments for distance foreign language learning. Language Learning and Technology 7, 50-80.

DuFon, M. & Churchill, E. (2006). Language learners in study abroad contexts. Clevedon, UK: Multilingual Matters.

Dyson, P. (1988). The year abroad: Report for the Central Bureau for Educational Visits and Exchanges. Oxford: Oxford University Language Teaching Centre.

Freed, B. (1990). Language learning in a study abroad context: The effects of interactive and non interactive out-of-class contact on grammatical achievement and oral proficiency. In J. Alatis (ed.), Linguistics, language teaching and language acquisition: The interdependence of theory, practice, and research (pp. 459-77), Georgetown University round table on languages and linguistics. Washington, DC: Georgetown University Press.

Freed, B. (1995a). Introduction. In B. Freed (ed.), Second language acquisition in a study abroad context (pp. 3-34). John Benjamins: Philadelphia.

Freed, B. (1995b). Second language acquisition in a study abroad context. John Benjamins: Philadelphia.

Golonka, E. (2006). Predictors revised: Linguistic knowledge and metalinguistic awareness in second language gain in Russian. Modern Language Journal 90, 496-505.

Gore, J. (2005). Dominant beliefs and alternative voices: Discourse, belief, and gender in american study abroad. New York: Routledge.

Guntermann, G. (1995). The peace corps experience: Language learning in training and in the field. In B. Freed (ed.), Second language acquisition in a study abroad context (pp. 14970). John Benjamins: Philadelphia.

Hoffman-Hicks, S. (1999). The longitudinal development of French foreign language pragmatic competence: Evidence from study abroad participants. Unpublished doctoral dissertation, Indiana University.

Howard, M. (2002). Prototypical and non-prototypical marking in the advanced learner's aspectuo-temporal system. EuroSLA Yearbook 2, 87-114.

Howard, M. (2006). Variation in advanced French interlanguage: A comparison of three (socio)linguistic variables. Canadian Modern Language Review 62, 379-400.

Howard, M., Lemee, I., & Regan, V. (2006). The L2 acquisition of a socio-phonetic variable: The case of /l/ deletion in French. Journal of French Language Studies 16, 1-24.

Isabelli, C. & Nishida, C. (2005). Development of the Spanish subjunctive in a nine-month study-abroad setting. In D. Eddington (ed.), Selected proceedings of the sixth Conference on the Acquisition of Spanish and Portuguese As First and Second Languages (pp. 78-91). Sommerville, MA: Cascadilla Press.

Izumi, S. & Iwasaki, M. (2004). Development of explicit and implicit knowledge of different grammatical forms by learners with different ESL learning backgrounds. Sophia Linguistica 52, 1-33.

Kinginger, C. (2007). Research on language education through programmes of study and residence abroad: Recent publications. International Journal of Bilingual Education and Bilingualism 10, 104-12.

Kinginger, C. (in press). Language learning in study abroad: Case studies of Americans in France. Modern Language Journal.

Lafford, B. (1995). Getting into, through, and out of a survival situation: A comparison of communicative strategies used by students studying Spanish-abroad and "at home." In B. Freed (ed.), Second language acquisition in a study abroad context (pp. 97-122). John Benjamins: Philadelphia.

Lafford, B. (2004). The effect of context of learning on the use of communication strategies by learners of Spanish as a second language. Studies in Second Language Acquisition 26, 201-26.

Lafford, B. & Collentine, J. (2006). The effects of study abroad and classroom contexts on the acquisition of Spanish as a second language: From research to application. In B. Lafford & R. Salaberry (eds.), Spanish second language acquisition: From research findings to teaching applications (pp. 103-26). Washington, DC: Georgetown University Press.

Lazar, N. (2004). A short survey on causal inference, with implications for context of learning studies of second language acquisition. Studies in Second Language Acquisition 26, 329-48.

Mackey, A. & Goo, J. (2007). Interaction research in SLA: A meta-analysis and research synthesis. In A. Mackey (ed.), Conversational interaction in second language acquisition (pp. 407-52). Oxford: Oxford University Press.

Magnan, S. & Back, M. (2007). Social interaction and linguistic gain during study abroad. Foreign Language Annals 40, 43-61.

McGeeking, A. (2006). Negotiation in a Japanese study abroad setting. In M. DuFon & E. Churchill (eds.), Language learners in study abroad contexts (pp. 177-202). Clevedon, UK: Multilingual Matters.

Möhle, D. & Raupach, M. (1983). Planen in der Fremdsprach. Frankfurt: Peter Lang.

Murphy-Lejeune, E. (2002). Student mobility and narrative in Europe: The new strangers. New York: Routledge.

Norris, J. & Ortega, L. (2001). Does type of instruction make a difference? Substantive findings from a meta-analytic review. Language Learning 51, 157-213.

O'Brien, I., Segalowitz, N., Collentine, J., & Freed, B. (2006). Phonological memory and lexical, narrative, and grammatical skills in second-language oral production by adult learners. Applied Psycholinguistics 27, 377-402.

O'Brien, I., Segalowitz, N., Freed, B., & Collentine, J. (2007). Phonological memory predicts second language oral fluency gains in adults. Studies in Second Language Acquisition 29, 557-81.

Opper, S., Teichler, U., & Carlson, J. (eds.) (1990). Impact of study abroad programmes on students and graduates. London: Jessica Kingsley.

Pellegrino-Aveni, V. (2005). Study abroad and second language use: Constructing the self. Cambridge: Cambridge University Press.

Perdue, C. & Klein, W. (1992). Why does the production of some learners not gram-maticalize? Studies in Second Language Acquisition 14, 259-72.

Regan, V. (1995). The acquisition of sociolinguistic native speech norms: Effects of a year abroad on second language learners of French. In B. Freed (ed.), Second language acquisition in a study abroad context (pp. 245-68). John Benjamins: Philadelphia.

Rivers, W. (1998). Is being there enough? The effects of home stay placements on language gain during study abroad. Foreign Language Annals 31, 492-500.

Rodriguez, S. (2001). The perception of requests in Spanish by instructed learners of Spanish in second- and foreign-language contexts: A longitudinal study of acquisition patterns. Unpublished doctoral dissertation, Indiana University.

Segalowitz, N. (2003). Automaticity and second languages. In C. J. Doughty and M. H. Long (eds.), Handbook of second language acquisition (pp. 382-408). Malden, MA: Blackwell.

Segalowitz, N. & Freed, B. (2004). Context, contact, and cognition in oral fluency acquisition. Studies in Second Language Acquisition 26, 173-99.

Segalowitz, N., Freed, B., Collentine, J., Lafford, B., Lazar, N., & Diaz-Campos, M. (2005). A comparison of acquisition of Spanish as a second language in two different contexts of learning: Study Abroad and the versus regular academic classroom. Frontiers 10, 1-18.

Shardakova, M. (2005). Intercultural pragmatics in the speech of American L2 learners of Russian: Apologies offered by Americans in Russian. Intercultural Pragmatics 2, 42351.

Siegal, M. (1995). Individual differences and study abroad: Women learning Japanese in Japan. In B. Freed (ed.), Second language acquisition in a study abroad context (pp. 22544). John Benjamins: Philadelphia.

Smartt, J. & Scudder, R. (2004). Immersion study abroad in Mexico: Using repair behaviors to assess proficiency change. Foreign Language Annals 37, 592-607.

Spada, N. (1985). Effects of informal contact on classroom learners' proficiency: A review of five studies. TESL Canada Journal 2, 51-62.

Spada, N. (1986). The interaction between types of contact and types of instruction: Some effects on the second language proficiency of adult learners. Studies in Second Language Acquisition 8, 181-99.

Tanaka, K. & Ellis, R. (2003). Study-abroad, language proficiency, and learner beliefs about language learning. JALT Journal 25, 63-85.

Tiechler, U. & Steube, W. (1991). The logics of study abroad programs and the impacts. Higher Education 21, 325-49.

Tokowicz, N., Michael, E., & Kroll, J. (2004). The roles of study-abroad experience and working-memory capacity in the types of errors made during translation. Bilin-gualism: Language and Cognition 7, 255-72.

Wilkinson, S. (1998). On the nature of immersion during study abroad: Some participants' perspectives. Frontiers 4, 121-38.

Wilkinson, S. (2002). The omnipresent classroom during summer study abroad: American students in conversation with their French hosts. Modern Language Journal 86, 157-73.

Willis, F., Doble, G., Sankarayya, U., & Smithers, A. (1977). Residence abroad and the student of modern languages: A preliminary survey. Bradford, UK: Modern Languages Centre, University of Bradford.

14 Less Commonly Taught Languages: Issues in Learning and Teaching

KIRA GOR AND KAREN VATZ

Introduction

A "Less Commonly Taught Language" (LCTL), a definition stemming out of educational policy, is a language considered important by the government, but unsustainable by the market. Political priorities drive a nation's linguistic need, although developing the educational infrastructure to support this need presents serious challenges. While "LCTL" is predominantly a US term referring to languages other than French, German, and Spanish (Brecht & Walton, 1997), the concept exists globally and it is a nation's current educational policy and political situation that determine what languages are classified as less commonly taught (LCT). For instance, in the US, languages such as Persian and Japanese have only recently become thought of as LCT, not because of the number of speakers of those languages, which has not changed significantly, but because those countries have come to play an increasingly important role in the global political arena and economy, whereas a language such as Dutch, which is equally, if not more, uncommon in foreign language curricula, is not classified as LCT. Furthermore, LCTLs are not universal, that is, a language classified as LCT in one country may be the predominant foreign language in another country. For example, Chinese is a LCTL in the US, but in many Asian countries it is offered, and often required, as part of the standard foreign language curriculum.

Brecht and Walton (1997) outline political, economic, social, and communication factors as playing a crucial role in determining a nation's foreign language needs. Specifically, language is a political tool for people to assert their identity in territorial and cultural conflicts. For example, Kazakhstan has recently replaced Russian as a lingua franca with its own language, making Kazakh important not only in Russia, but also in countries with a political or economic stake in that part of the world. International trade drives economic relationships, which also has consequences as to what languages are considered important to a nation. Social issues such as humanitarian aid and environmental protection require international cooperation that relies on communication among more people from

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 Bkckwell Publishing Ltd. ISBN: 978-1-405-15489-5

different language backgrounds. And finally, ease of international travel and advances in communication technology facilitate direct interaction between people of different countries, necessitating extensive language skills.

However, identifying a language as critical does not automatically result in that language achieving popular status in a country's foreign language curriculum, either from a financial or learner perspective. Often, the reason for teaching a language originates from historical circumstances, as is the case for colonial languages (e.g., French in North Africa) and autochthonous languages (e.g., Gaelic in Ireland). A language may also be considered "classical," (e.g., Latin in the US) and remain prevalent despite its lack of political value. Therefore, because a country's current language needs are not always addressed, government intervention is frequently necessary to support LCTLs and incorporate them into existing educational policy, which is met by challenges. First, developing the infrastructure necessary to support the teaching of LCTLs is challenged by the shortage of programs, materials, trained teachers, and access to immersion opportunities. Within the US, LCTLs are typically only offered at the undergraduate and graduate levels at major universities and elite liberal arts colleges, and even within these institutions they are largely supported by federal funding from Title VI of the Higher Education Act (Brecht & Walton, 2000), and more recently, the National Security Language Initiative. Second, developing programs for all the potential LCTLs is in many cases unfeasible. For instance, some 50 languages have been identified as relevant to worldwide security interests (The National Language Conference, 2004). It is not practical for any one country to invest in developing language programs to support this breadth of language capability. Finally, LCTLs are often genetically, typologically, and culturally distant from the native language of the learner, decreasing the language's learnability and making it difficult for learners to achieve functional proficiency without a considerable time investment and often an extended immersion experience (Brecht & Walton, 2000). While this challenge attracts highly motivated language learners, it also results in low enrollments, making it difficult to justify maintaining the program financially. This chapter will look at LCTLs through the prism of research and theories in second language acquisition (SLA) and identify cognitive and linguistic difficulties inherently associated with learning a LCTL. It will then adopt a teaching perspective and address general problems and possible solutions to building a curriculum for teaching LCTLs.

Are LCTLs Difficult?

The reasons for languages being "less commonly taught" are many, as we note above. However, there is a prevalent attitude among educators in the United States that a major factor discouraging students from studying certain languages is their inherent difficulty. While this statement is generally correct, not everything about learning a given LCTL is problematic, and insights from SLA and several other disciplines are needed to develop a coherent framework to account

for the level of difficulty of various linguistic and cultural phenomena. Identification of these areas of difficulty is a first step in elaborating an efficient pedagogical approach.

On the one hand, the term Less Commonly Taught Language does not imply in its definition that a language is inherently more difficult. It is, instead, a cultural and political statement referring to the status of a given language's position within a country's foreign language curriculum. Therefore, the term LCTL, which encompasses the status of Norwegian, Japanese, and Avestan, refers to a politically motivated definition and makes no reference to the language's difficulty. On the other hand, there is a general belief that were Norwegian, Japanese, and Avestan taught to a group of American students, those studying Japanese and Avestan would surely show a slower rate of acquisition than those studying Norwegian, all other factors being equal (i.e., teachers and resource availability). Indeed, the notion that some languages are more difficult than others has been supported by a National Foreign Language Center (NFLC) study assessing the impact of in-country study on language proficiency (Frank, 2000). Proficiency gains were measured for National Security Education Program (NSEP) study abroad fellows in China, Russia, and Spain. Entering the program at the intermediate-low to intermediate-mid proficiency levels on the oral proficiency scale adopted by the American Council of Teachers of Foreign Languages (ACTFL), learners of Spanish were most likely to achieve advanced proficiency, followed by learners of Mandarin Chinese, and finally learners of Russian. While the differences were not significant, due to a limited sample size, it appears that despite equivalent pre-program proficiency scores and comparable program duration, rates of advanced level attainment are higher in Spanish than in Mandarin Chinese and Russian.

The notion of language difficulty has been addressed by US government language training institutions in terms of "the inherent difficulty posed to native speakers of American English in learning the target language" (Frank, 2000). In this approach, languages are categorized into three levels of difficulty according to the number of hours of study typically required to achieve ILR (Interagency Language Roundtable) level 3 proficiency, corresponding to Superior on the ACTFL scale. For example, Spanish, requiring relatively few hours of study (approximately 575-600) to achieve ILR level 3, is a Category 1 language, whereas Korean, requiring approximately 2,200 class hours, is classified as a Category 3 language (Languages of the World, 2007). Therefore, it appears that some LCTLs may be inherently more difficult for American learners. The question, then, is, why are certain languages more difficult? More specifically, what cognitive factors contribute to one language being more difficult to acquire than another?

A major factor that is likely to contribute to the difficulty of a LCTL is its non-cognate status to the learner's native language. Traditionally, cognate languages are defined as related in origin, having descended from the same ancestral root. Accordingly, the term non-cognate language refers to those that are not genetically related and implies typological distance, primarily relating to the languages' differing structures and lexicon. However, the term non-cognate may also be

applied to different writing systems, and, perhaps more significantly, cultural distance. While structural distance between languages is an attractive candidate for explaining language-learning difficulty, research shows that, first, typological distance does not translate directly into the level of difficulty, and second, there are other powerful factors at work in the acquisition of LCTLs, such as the lexicon, phonological system, writing system, and cultural distance. Thus, the level of difficulty of a particular LCTL is defined by a combination of factors, and while there is a practice of assigning levels of difficulty to particular LCTLs for American learners, which is driven by the needs of curricular planning, more research is needed to provide a sound theoretical and empirical foundation to this global, holistic scale. At the same time, SLA research provides insights into the operation of individual factors, or components, contributing to the overall level of difficulty. The following sections will, first, address the general approaches to language difficulty arising from theoretical perspectives on SLA, and then focus on individual aspects of language acquisition, which constitute sources of difficulty.

Theoretical and Empirical Approaches to LCTLs in SLA: Sources of Difficulty

Language typology, influence of the native language, and universal constraints

Second language acquisition research identifies two major factors shaping inter-language (IL, language of L2 learners): the influence of the native language, or transfer from the native language to the target language (L1 transfer); and universal constraints on language perception and production. In addition to these major forces, whose interplay shapes the course of SLA, processing constraints and properties of the input to the learner, in particular, input frequencies and the effects of practice, interact with working and long-term memory as the learner gradually achieves automaticity in speech perception and production. This section will briefly review the main claims and findings from these fields of inquiry, and draw implications for LCTLs.

L1 transfer implies that certain linguistic structures, patterns, or rules from L1 are transposed and applied to L2, which may lead to two possible outcomes: a facilitative effect when both L1 and L2 indeed share the structure (positive transfer, see Ringbom & Jarvis, this volume); and a negative effect, resulting in errors in L2, when L1 and L2 do not share the structure (negative transfer). Whether L1 transfer of either kind takes place depends on a set of conditions. First, there should be certain proximity between the two phenomena in L1 and L2 in order for this transfer to be possible. For example, when acquiring the Russian case system with numerous inflectional markers, native speakers of English have nothing to transfer from their native language, or in other words, there are no conditions for L1 transfer. Native speakers of Czech, another Slavic language with a similar inflectional system, however, have been observed to

substitute Czech case inflections for Russian ones (Duskova, 1984, cited by Gass & Selinker, 2001, p. 74). This constraint on L1 transfer has received the name of "crucial similarity measure" (Wode, 1983), and is especially operative in L2 phonological acquisition (see below). Second, "perceived similarity measure," a subjective assessment of similarity between L1 and L2 phenomena (Kellerman, 1978), appears to determine whether L1 transfer will be used as a strategy when L2 processing involves meaning and reaches the conscious level, as in the use of lexical items or idioms. And finally, there are factors external to the linguistic properties of L1 and L2 per se, which pertain to task and learner characteristics. For example, a study of L2 acquisition of Japanese progressive and resultative meanings of the imperfective aspect marker by L1 speakers of English, German, and Slavic languages demonstrated that both the task type, or rather, modality, production, and perception, as well as proficiency level, determined whether L1 transfer would take place (Sugaya & Shirai, 2007). In particular, lower proficiency was more conducive to L1 transfer, while higher proficiency relied more on universals in creating form-meaning associations. Considering what is known about L1 transfer, what role is it expected to play in the acquisition of LCTLs? Given that LCTLS are often typologically distant from learners' native language, or are non-cognate languages, the role of L1 transfer should be rather limited, both on the level of linguistic structure and vocabulary. This applies to both positive and negative transfer. At the same time, lower-proficiency learners are more likely to apply L1 transfer as a strategy, especially, under communicative pressure, as this may be the only available resource for them in an attempt to repair a communication breakdown. The most "extreme" case of L1 transfer in such a case would be code-switching.

Universal constraints are defined in different terms depending on the theoretical framework, and they can either pertain to linguistic structures or to cognitive processing and memory. Several assumptions and empirical findings in cross-linguistic and SLA research underlie the notion of universal constraints. First of all, Universal Grammar (UG) (Chomsky, 1986), research on typological universals (Greenberg, 1976), and markedness theory (Jakobson, 1936/1972), while differing in scope and theoretical underpinnings, all point to the fact that languages share some general universal properties. At the same time, UG emphasizes innateness and specificity of language, while typological studies focus on how certain linguistic properties are distributed across all the languages, which ones are common, and which are rare. Typological universals may take the shape of implicational universals, when the presence of feature A in a certain language also implies the presence of feature B, but the presence of feature B does not imply the presence of feature A. For example, the presence of voiced consonants /b, d, g/, either in the individual language in general, or in a particular position in the word, word-initial, medial or final, implies the presence of voiceless consonants /p, t, k/, but the presence of voiceless consonants does not imply the presence of voiced ones.

Markedness theory assigns the marked status to one member of a binary opposition, as in voiced and voiceless consonants mentioned above, or organizes a set

of structures in a certain hierarchical order, as in the most researched implicational universal concerning relative clause formation, Noun Phrase Accessibility Hierarchy (Keenan & Comrie, 1977). Markedness is usually associated with a cluster of properties, thus a marked item requires more effort, is more difficult, more complex structurally, less frequent, and less productive than its unmarked counterpart. It is not obvious what kind of causal relationship applies to this cluster of properties, but it appears that difficulty and complexity are at the core of the factors underlying markedness. If this is indeed the case, then it is to be expected that more complex or difficult linguistic structures occur less frequently across languages, but also that less frequent structures are inherently more complex and difficult.

Second, while typologically distant languages are likely to exhibit more differences than typologically proximate languages, some linguistic features are "scattered" across languages and do not follow a predictable pattern. For example, agglutination, a typological feature whereby words are formed of long strings of mostly unmodified affixes, occurs in genetically unrelated languages, such as Hungarian, Turkish, Korean, and Bantu languages. Third, universal constraints, whether they come in the form of UG, observations from typological studies, or markedness, imply that not every imaginable structure or combination of structures is possible in human languages. The question arises whether interlanguages are constrained in the same way, and researchers coming from different theoretical backgrounds espouse this view (Eckman, 2004; O'Grady, 2003; White, 2004). Eckman proposed the Structural Conformity Hypothesis (SCH), according to which "All universals that are true for primary languages are also true for ILs" (1996, p. 204). If this is the case, then markedness and universal constraints determine different patterns in SLA, as they do in native languages. Under the SCH, it is not the learner's native language that affects the acquisition of certain L2 structures, but rather the universal markedness of the L2 structures. In other words, a learner may have difficulty learning a marked structure in the L2, even if that same structure exists in the learner's L1. Conversely, a learner may not have difficulty in learning an unmarked L2 structure, even if that structure does not exist in the learner's L1. Since LCTLs, at least in the American context, represent typologically distant and highly diverse languages, many features found in them are typologically marked, and likely to be problematic for L2 learners. Therefore, acquisition of many aspects of LCTLs can be predicted to pose difficulties for L2 learners. In psycholinguistic terms, acquisition of typologically distant or non-cognate languages requires control of numerous new concepts, structures, etc. This novelty effect leads to depletion of attentional resources and overload on working memory, which ultimately delays automatization of L2 processing.

Four likely candidates that contribute to the difficulty in acquiring a non-cognate language, in addition to linguistic structure, are the lexicon, the phonological system, the writing system and script, and cultural distance. The lexicon, if no part of it is shared with the L2, will create an additional burden on processing and make attention to and processing of new aspects of the language all the more difficult. The phonological system, if it contains novel sounds, features, or

contrasts, may involve difficulties at the level of production and/or perception. The writing system and script may also add to difficulty in acquisition if they are different from that of the native language. For one thing, they will limit the opportunity for input, and in addition, they may contain ambiguities that lead to unpredictability in word recognition. Finally, a language of a non-cognate culture will likely consist of complex pragmatic and cultural references that are closely linked to linguistic performance. The following sections will address these four aspects of learning LCTLs.

The lexicon

The role of the lexicon in SLA has been thoroughly examined in the SLA literature, and specifically in the context of positive transfer between proximate and distant languages. Nation (2001) proposes that different words will have different "learning burdens" according to how much effort is required to learn the word. Specifically, Nation proposes that patterns and knowledge of an L2 word may be available to the learner from the learner's L1, previous knowledge of the L2, or even other L2s. For example, similarities between the L1 and L2 in sounds and spelling patterns, the pairing of certain words together (e.g., collocations), and grammatical patterns will reduce the learning burden and thus facilitate acquisition of a particular word(s). The learning burden is likely to be less for vocabulary of a closely related, or cognate, language, than for a non-cognate language.

Ringbom (1992; Ringbom & Jarvis, this volume) examined this phenomenon in Finnish and Swedish learners of English pointing out that Swedish learners achieve higher scores, even with fewer years of study, than Finnish learners. Ringbom attributes the Swedes' advantage to positive lexical transfer, which facilitates reading and listening comprehension (although more so for reading than listening). Ringbom argues that a learner of a related language can easily convert the L1-based lexical knowledge to the L2, "because the procedures for comprehending and using identical or very similar L1 words in L2 are already automatized" (1992, p. 102). For a learner of an unrelated language, however, the L1 knowledge is not accessible for automatized use in the L2, thus resulting in a heavier learning burden.

Furthermore, learning a non-cognate lexicon will impede learning other aspects of the language, in that a heavier learning burden requires cognitive resources that would otherwise be allocated elsewhere. Odlin (2003) summarizes the advantage of learning a linguistically proximate language: "the advantage that cognate vocabulary confers can allow learners to take advantage of positive transfer to increase their comprehension of the target language with far greater ease, thereby freeing many cognitive resources for other language learning tasks" (p. 441). Therefore, learners of an L2 that is lexically related to their L1, such as Swedish learners of English, will have more resources available, facilitating, and even speeding up, lexical, and subsequently, general L2 acquisition. Acquisition of LCTLs will show the opposite inhibitory tendency.

Phonological system

Phonology is always a major issue for adult learners, and few, if any, are able to overcome a foreign accent. Phonemic discrimination studies have shown that infants lose their ability to discriminate between non-native contrasts as early as 10-12 months (Werker & Tees, 1984), and SLA researchers have placed the close of the critical period for acquiring a nativelike accent around age 6 (Long, 1993).

The mechanisms underlying production and perception are asymmetrical, with some sounds being more difficult to hear, and others to pronounce. It is true, however, that perceptual difficulties usually constitute the core with articulatory difficulties superimposed on them, hence the focus on perception in the major models of phonological acquisition (see below). A rare exception is the /r/-/l/ distinction in English, which is difficult for Japanese learners to hear, yet Sheldon and Strange (1982) found that learners were able to accurately produce both sounds despite an inability to perceive the difference. LCTLs often have difficult sounds and sound contrasts (e.g., murmured consonants in Hindi/Urdu; clicks in Bantu languages; plain, tense, and aspirated consonants in Korean, etc.), making accurate perception and production challenging for learners of non-cognate languages.

Kuhl (1993), Flege (1995), and Best (1994) propose three separate models of language perception that attempt to explain non-native phonemic representation within the context of a learner's native speech system. Kuhl's Native Language Magnet model, focusing mainly on vowel perception, posits that native speakers develop acoustic prototypes for phonemic categories, which act as perceptual magnets. A non-native phoneme that is perceptually close to a native prototype will be drawn to and represented by that prototype, whereas a non-native phoneme that is perceptually distant from native prototypes will not be influenced by the magnet and will exist independently in its own space. Therefore, a non-native phonemic contrast that is drawn to a native prototype will be more difficult for learners to discriminate than a non-native contrast that is represented independently of native prototypes.

Flege's (1995) Speech Learning Model also proposes that the distance between native and non-native sounds determines ease of acquisition. He argues that non-native categories will develop if there is no native equivalent, which will facilitate perception and production. Non-native sounds that are similar to native sounds, however, will be incorporated into existing native categories, making perception and production inaccurate.

Finally, Best's (1994; Best, McRoberts, & Goodell, 2001) Perceptual Assimilation Model claims that non-native phonemes may be assimilated into the native system as a categorized exemplar of a native phoneme, an uncategorized phoneme, or a nonassimilable nonspeech sound. How a non-native phoneme is assimilated determines its perceptual difficulty. Sounds that are assimilated into a native exemplar may pose no problem for the learner if there is a good fit between the non-native sound and the native exemplar; however, a poor fit will result in perceptual difficulty. In the case of the uncategorized phoneme, the

native system will have less of an impact on perception, and difficulty will depend on perceived proximity to nearby phonemes. Finally, nonassimilable nonspeech sounds are not processed linguistically, and will, therefore, not pose a problem.

All three models have one point in common: non-native sounds that are easily assimilated into the native language phonological system are problematic (compare this conclusion to the concept of the "crucial similarity measure" discussed above). How do these theories apply to acquisition of LCTLs, which often contain sounds and sound contrasts not found in most common Indo-European languages, and do they account for the level of difficulty they present? First, phonological difficulties in SLA almost never involve only the perceptual level, which is the focus of all three models. Typically, phonological difficulties arise either both in perception and production or in production only. Second, while it is uncontroversial that similar sounds easily pass through the "phonological sieve" and are perceived as their native counterparts (often leading to a perceptual error), there are types of sounds that are difficult to perceive and/or produce, in the typological markedness sense. And third, the most challenging task in phonological acquisition is differentiation of sound contrasts when neither member of the contrast is found in L1, and they belong to a typologically marked domain. For example, tones found in Mandarin, Cantonese, Thai, and Vietnamese are difficult, as the learner needs to learn to perceive and produce four or more phonological tones accurately and be aware of the distinctions. To illustrate, consider the different facets of phonological difficulties involved in the acquisition of Arabic consonants. There are pharyngeal and laryngeal consonants in Arabic, and these sounds are difficult to differentiate both in perception and production, since pharyngeal and laryngeal articulations, typologically rare and marked, are notoriously problematic. Furthermore, Arabic has a phonological distinction between plain and emphatic consonants with additional pharyngeal articulation, the latter being difficult to perceive, but especially to pronounce. To summarize, the phonology of LCTLs is likely to present problems for L2 learners at the level of perception, production, or both. These problems will mostly be due to universal constraints, as learners are dealing with new and marked phonological units, features, and contrasts, and to a lesser degree to L1 transfer resulting from perceptual assimilation of L2 sounds (see Nguyen & Macken, 2008).

Writing system and script

The L2 writing system and script can also provide strategic support if similar to the L1 and, conversely, can create an obstacle if different from the L1 (see Koda, this volume). The writing system refers to the most general distinction between alphabetic, syllabic, and logographic languages, while script refers to the actual graphic symbols (e.g., Roman and Cyrillic scripts). Orthography, or the rules of graphic-phonological conversion, will have less impact when the writing system and script already differ, as is often the case with LCTLs. MacWhinney (2006) discusses recoding, a compensatory strategy strongly affected by the L1, that L2 learners may use to enhance language learning. Recoding, according to

MacWhinney, "involves the construction of alternative images of new words and phrases ... the easiest way to do this is to represent the new word orthographically" (p. 151). He argues that orthographic learning provides learners with a "solid recoding of transient auditory input" (p. 151) and allows them access to input from additional sources, such as books, signs, and product labels. Recoding will be relatively easy for a learner whose writing system and script are the same in the L1 and L2 (for example, French-English), however, it will be quite difficult for a learner who has to learn a new script, and especially, a writing system. An English speaker learning Russian must map the Roman script onto the Cyrillic script. Because both of these scripts are alphabetic and rely on a grapheme-phoneme correspondence, this task can be accomplished with some difficulty. However, learning a logographic writing system (e.g., Chinese Hanzi characters) will create a major obstacle for older learners and may prevent them from taking advantage of the recoding strategy, making the language-learning process all the more difficult for learners of a language with a different writing system.

An additional complexity is present in languages with wide gaps between the written and spoken varieties, such as the numerous dialects of Arabic compared to written Modern Standard Arabic. A student learning Moroccan Arabic must not only learn the spoken variety, but also the written variety, which is significantly different structurally, as well as phonologically. In this case, access to written input requires substantially more work than if the student were learning Spanish, where the spoken language is closely represented by the writing system. Therefore, a student learning Arabic not only has to learn an L2 script that is different from the L1, but also an L2 writing system that is different from the spoken L2 system.

However, the importance of the writing system as a language acquisition tool may be debated in light of recent findings on the acquisition of Chinese. NSEP studies on the acquisition of Mandarin Chinese by American learners (Frank, 2000) have suggested that Chinese is a Category 2 language in speaking, despite its difficult writing system. That is, it seems that speaking can be acquired successfully independent of the writing system in an immersion-type learning environment with heavy emphasis on oral communication. Nonetheless, access to written media can be a rich and important source of both linguistic and socio-cultural input for language learners. An L2 with a different writing system may prevent a learner from taking advantage of written input, especially in the early stages of acquisition, and thereby slow the rate of acquisition.

Pragmatics and cross-cultural communication

Languages of non-cognate cultures are often characterized by very different pragmatic rules than those of the native language. While there is limited research on L2 pragmatics, these often subtle and complex linguistic aspects of basic communication, such as the speech acts of giving and receiving compliments, accepting and declining an offer or invitation, and apologizing or adhering to an honorifics

hierarchy, contribute to the difficulty of acquiring a typologically distant language. For example, the pragmatic strategies involved in Japanese honorifics are closely linked to linguistic performance; while the linguistic forms themselves are not necessarily difficult, knowing when and to whom to apply the proper address form is culturally foreign to native English speakers and, therefore, difficult to acquire.

Another component of pragmatics that is necessary for successful communication is what Kramsch (1991) identifies as sociolinguistic knowledge: "Background knowledge and shared assumptions have been shown to be a crucial element in understanding oral and written forms of discourse" (p. 217). For example, the ability to understand and employ literary references in languages such as Arabic and Chinese, whose literary culture dating back 1,500 years is incorporated into daily speech, presents a challenge for L2 learners. In other languages, such as Persian, there is an equivalent depth in the oral tradition that is still present in the language today. Again, these skills are not necessarily linguistic in nature, but rather require cognitive skills and cultural sensitivity that develop from cultural experience and education. Because they are manifested in language comprehension and production, they contribute to the practical difficulty of the language. It would be difficult to acquire this type of knowledge from studying LCTL in a foreign language classroom outside its broader context; a cultural immersion is most likely the only route to acquiring what Kramsch (2006) has termed symbolic competence.

To conclude, structural (typological) differences between non-cognate languages, as well as L2 lexicon, phonological system, writing system and script, and cultural distance are major inhibiting factors in the acquisition of certain LCTLs.

Teaching Less Commonly Taught Languages

Teaching LCTLs in the US presents numerous challenges, and while some of them are related to the inherent linguistic difficulty of many of the languages discussed above, others transcend the realm of linguistic debates and are grounded in the social and educational context. Despite fundamental differences between the principles underlying the grouping of LCTLs in the US and lesser-used languages in the European Union, some of these challenges are shared by both educational communities. This section will address the issues in teaching LCTLs, and for lack of space will be restricted to adult L2 learners, which will leave two major educational issues outside the scope of the present discussion: child L2 acquisition and K-12 instruction in LCTLs (see Brecht, 2007), and the special instructional needs of heritage speakers (see Montrul, this volume).

It is true that non-cognate status and typological distance of LCTLs from English create an additional level of difficulty in the acquisition of their linguistic structure, sound systems, writing systems and scripts, vocabulary, including idiomatic use, and socio-cultural and linguo-pragmatic aspects. What are the implications of these additional learning problems for teaching LCTLs? They

increase the need for appropriate pedagogical approaches and instructional techniques aimed at developing metalinguistic awareness and structural knowledge in L2 learners. Two aspects of instruction play a key role in addressing these problems: the type and amount of input (Gor & Long, forthcoming) and practice (DeKeyser, 2007). And indeed, developing metalinguistic awareness when dealing with novel linguistic concepts may require explicit input, while high quantities of structured input will ensure internalization of high-frequency items and structures. Koda (this volume) claims that both quality and quantity of L2 print are to a large extent responsible for the development of the metalinguistic awareness necessary for reading new writing systems and scripts (as well as orthography). Intensive practice in LCTLs will promote control and automaticity in the use of linguistic structures, and decrease overload on working memory, thus releasing L2 processing resources.

Automaticity refers to the way psychological mechanisms operate, with automatic processing often characterized as being fast, unstoppable, load independent, effortless, and unconscious (Segalowitz, 2003, p. 384). In other words, automatic processing is not affected by the load of information to be processed, nor does it require effort, leaving attentional resources free for other tasks. Research to date on automaticity in the field of SLA has focused mainly on word recognition and grammatical structure acquisition, where speed of processing is linked to fluency. According to Nation (1993), a certain degree of automaticity in basic vocabulary must be achieved before new vocabulary can be acquired. Considering LCTLs' non-cognate status, developing automaticity for novel vocabulary and typologic-ally distant grammatical structures will take time, delaying fluency and reducing the attentional resources, such as working memory, available for processing and subsequently acquiring additional aspects of the language.

What kind of teaching method would be most appropriate for LCTLs? In the last two decades, foreign language teaching in many countries has been dominated by the communicative approach, stemming from the construct of communicative competence (Hymes, 1974), which came to replace grammar-translation and audio-lingual methods. Communicative language teaching (CLT), closely associated with the influential proficiency movement, emphasizes the development of learners' ability to communicate, express themselves, get their meaning across, and engage in social interactions. Pedagogical practices developed to accomplish this mission produce relatively quick results, empowering the student with a sense of gains made, and not just effort invested. This agenda, a welcome change from the teaching methods it replaced, is heavily geared toward speaking, even at the early stages of L2 learning. In recent years, communicative language teaching has become the target of criticism summarized below by Magnan:

Following most textbooks used in the United States, CLT encourages personalized activities through which students talk about themselves with their classmates (Magnan, 2006). This practice introduces three problems for language learning: (a) talk about self generally does not elicit the analytical language that collegiate language departments consider pertinent to their intellectual missions; (b) too much

talk about self perpetuates self-referential notions of language and culture, preparing students to present an egocentric view when abroad; (c) talk with U.S. classmates fosters a U.S. frame of reference and discourse, although the words to express them are foreign. (2007, p. 250)

In a discussion devoted to the future of CLT in The Modern Language Journal (volume 90(2), 2006), Kramsch advocated an alternative approach promoting the development of symbolic competence (Kramsch, 2006). The proposed shift of focus to the study of texts and culturally bound meanings (Larson, 2006) highlights the challenge of reading authentic texts in LCTLs, which provides a window into the target culture, but often stumbles at different writing systems and scripts, while understanding a new culture through language increases in importance with the greater distance between the native and target culture. The arguments raised against CLT mostly target lower-level instruction and as such are not in conflict with the core principles underlying this approach. Teaching and learning LCTLs need to address several goals: to provide explicit and abundant input and opportunities for focused practice required for learning language structure, to pay attention both to oral and written communicative competence, and to complement functional language ability with the development of cultural sensitivity.

Since the distance between L1 and LCTLs, linguistic and cultural, is often greater than for commonly taught languages, their teaching (as learning) requires additional effort and resources. Ironically, educational policy and practices, which were shaped by an interaction of academic and socio-political circumstances, led to the opposite outcome, with the lack of available resources exacerbating inherent difficulties in developing efficient curricula for teaching LCTLs. Indeed, academic programs offering instruction in LCTLs are plagued by the lack of a trained cadre of instructors, adequate and rich pedagogical materials for different learner levels, institutional support, and national and international infrastructure (see Al-Batal, 2007, about the teaching of Arabic). As mentioned above, based on the estimates of the number of contact hours needed to reach ILR level 3 proficiency by the State Department, all of the languages falling into Category 3 (exceptionally difficult, most time required), Arabic, Cantonese, Mandarin, Japanese, and Korean, are LCTLs critical for US national interests. Consequently, in order to achieve the same level of proficiency in one of these LCTLs as in a language belonging to Categories 1 or 2, smaller classes and/or more coursework will be necessary. However, requests for limiting the number of seats per class may meet resistance from the administration of academic institutions, as small classes are often perceived as a sign that a certain language has low enrollments due to low demand or, conversely, they will be in conflict with the need to accommodate very high enrollments in some of these languages. One of the inevitable consequences of the fact that LCTLs typically require more time and effort to learn is decreasing learners' motivation when immediate results and rapid gains in proficiency are lacking. Thus, to build successful academic programs, LCTL curricula need to stimulate student motivation. The ability to communicate from

the very start fostered within the paradigm of CLT is one of the means to boost learner motivation, engaging students in a new fascinating cultural experience being another one.

The critical role a number of LCTLs play in US national security has generated an enormous momentum, with an increasing demand for highly proficient speakers of these languages and public interest in learning them. Several organizations and research centers in the US focus on providing resources and teaching support for LCTLs: The National Council of Less Commonly Taught Languages (NCOLCTL), the hub of the web-based network CouncilNet, The National Foreign Language Center (NFLC), whose LangNet offers support for several dozen LCTLs, and The Center for Applied Linguistics (CAL), as well as university-affiliated research centers, such as The Center for Advanced Study of Language (CASL) at the University of Maryland and The Center for Advanced Research on Language Acquisition (CARLA) at the University of Minnesota. The European Bureau for Lesser-Used Languages (EBLUL) performs a similar mission in the European Union. Only joint efforts of SLA researchers, instructors, and experts in language pedagogy and cross-cultural communication, administrators, and language policy makers will win a brighter future for LCTLs.

REFERENCES

Al-Batal, M. (2007). Arabic and national language educational policy. The Modern Language Journal 91, 2, 268-71.

Best, C. T. (1994). The emergence of native-language phonological influences in infants: A perceptual assimilation model. In J. C. Goodman & H. C. Nusbaum (eds.), The development of speech perception: The transition from speech sounds to spoken words (pp. 167-224). Cambridge, MA: MIT Press.

Best, C. T., McRoberts, G. W., & Goodell, E. (2001). Discrimination of non-native consonant contrasts varying in perceptual assimilation to the listener's native phonological system. Journal of the Acoustical Society of America 109, 2, 775-94.

Brecht, R. D. (2007). National language educational policy in the nation's interest: Why? How? Who is responsible for what? The Modern Language Journal 91, 2, 264-5.

Brecht, R. D. & Walton, A. R. (1997). National language needs and capacities: A recommendation for action. In J. Hawkins et al. (eds.), International education in the new global era: Proceeding of a national policy conference on the Higher Education Act, Title VI, and Fulbright-Hays programs (pp. 92-102). Los Angeles: UCLA International Studies and Overseas Programs.

Brecht, R. D. & Walton, A. R. (2000). System III: The future of language learning in the United States. In R. Lambert & E. Shohamy (eds.), Language policy and pedagogy: Essays in honor of A. Ronald Walton (pp. 111-27). Amsterdam: John Benjamins.

Chomsky, N. (1986). Knowledge of language: Its nature, origin and use. New York: Praeger.

DeKeyser, R. (2000). The robustness of critical period effects in second language acquisition. Studies in Second Language Acquisition 22, 499-533.

DeKeyser, R. (ed.) (2007). Practice in a second language: Perspectives from applied linguistics and cognitive psychology. Cambridge: Cambridge University Press.

Duskova, L. (1984). Similarity: An aid or hindrance in foreign language learning? Folia Lingüistica 18, 103-15.

Eckman, F. R. (1996). A functional-typological approach to second language acquisition theory. In W. C. Ritchie & T. K. Bhatia (eds.), Handbook of second language acquisition (pp. 195-211). San Diego: Academic Press.

Eckman, F. R. (2004). Universals, innateness, and explanation in second language acquisition. Studies in Language 28, 3, 682-703.

Flege, J. E. (1995). Second language speech learning: Theory, findings and problems. In W. Strange (ed.), Speech perception and linguistic experience: Theoretical and methodological issues in cross-language speech research (pp. 233 -72). Timonium, MD: York.

Frank, V. (2000). Impact of in-country study on language ability: National Security Education Program undergraduate scholarship and graduate fellowship recipients. Technical Report, The National Foreign Language Center.

Gass, S. M. & Selinker, L. (2001). Second language acquisition: An introductory course. London: Lawrence Erlbaum.

Gor, K. & Long, M. H. (forthcoming). Input and second language processing. In W. C. Ritchie & T. K. Bhatia (eds.), Handbook of second language acquisition. New York: Academic Press.

Greenberg, J. (1976). Language universals. The Hague: Mouton.

Hymes, D. (1974). Foundations in sociolinguistics: An ethnographic approach. Philadelphia: University of Pennsylvania Press.

Jakobson, R. (1936/1972). Child language, aphasia, and phonological universals, 2nd edn. The Hague: Mouton.

Keenan, E. & Comrie, B. (1977). Noun phrase accessibility and Universal Grammar. Linguistic Inquiry 8, 63-99.

Kellerman, E. (1978). Giving learners a break: Native language intuitions as a source of predictions about transferability. Working Papers on Bilingualism 15, 59-92.

Kramsch, C. (1991). Culture in language learning: A view from the United States. In K. De Bot, R. Ginsberg, & C. Kramsch (eds.), Foreign language research in cross-cultural perspective (pp. 217-40). Philadelphia: John Benjamins.

Kramsch, C. (2006). From communicative competence to symbolic competence. The Modern Language Journal 90, 2, 249-52.

Kuhl, P. K. (1993). Innate predispositions and the effects of experience in speech perception: The native language magnet theory. In B. de Boysson-Bardies, S. de Schonen, P. Jusczyk, P. McNeilage, & J. Morton (eds.), Developmental neurocognition: Speech and face processing in the first year of life (pp. 259-74). Dordrecht: Kluwer Academic.

Languages of the World (2007). Retrieved February 24, 2008, from www.nvtc.gov/lotw/ index.html.

Larson, P. (2006). The return of the text: A welcome challenge for less commonly taught languages. The Modern Language Journal 90, 2, 255-8.

Long, M. H. (1993). Second language acquisition as a function of age: Research findings and methodological issues. In K. Hyltenstam & A. Viberg (eds.), Progression and regression in language (pp. 196-221). Cambridge: Cambridge University Press.

MacWhinney, B. (2006). Emergent fossilization. In Z. Han & T. Odlin (eds.), Studies of fossilization in second language acquisition (pp. 134-56). Clevedon, UK: Multilingual Matters.

Magnan, S. (2006). Enjeux et défis de l'enseignement du français langue étrangère en France et aux Etats-Unis. French Review 80, 332-52.

Magnan, S. (2007). Reconsidering communicative language teaching for national goals. The Modern Language Journal 91, 2, 249-52.

Nation, I. S. P. (1993). Using dictionaries to estimate vocabulary size: Essential, but rarely followed, procedures. Language Testing 10, 27-40.

Nation, I. S. P. (2001). Learning vocabulary in another language. Cambridge: Cambridge University Press.

The National Language Conference: A call for action (2004, June 22). Retrieved February 28, 2008, from http://www.nlconference.org.

Nguyen, N. & Macken, M. A. (2008). Factors affecting the production of Vietnamese tones. Studies in Second Language Acquisition 30, 49-77.

Odlin, T. (2003). Cross-linguistic influence. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 436-86). Oxford: Blackwell.

O'Grady, W. (2003). The radical middle: Nativism without Universal Grammar. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 19-42). Oxford: Blackwell.

Ringbom, H. (1992). On L1 transfer in L2 comprehension and L2 production. Language Learning 42, 1, 85-112.

Segalowitz, N. (2003). Automaticity and second languages. In C. J. Doughty & M. H. Long (eds.), The handbook of second language acquisition (pp. 382-408). Oxford: Blackwell.

Sheldon, A. & Strange, W. (1982). The acquisition of /r/ and /l/ by Japanese learners of English: Evidence that speech production can precede speech perception. Applied Psycholinguistics 3, 243-61.

Sugaya, N. & Shirai, Y. (2007). The acquisition of progressive and resultative meanings of the imperfective aspect marker by L2 learners of Japanese: Transfer, universals, or multiple factors? Studies in Second Language Acquisition 29, 1-38.

Werker, J. F. & Tees, R. C. (1984). Cross-language speech perception: Evidence for perceptual reorganization during the first year of life. Infant Behaviour and Development 7, 49-63.

White, L. (2004). "Internal" versus "external" universals. Studies in Language 28, 3, 704-7.

Wode, H. (1983). Contrastive analysis and language learning. In H. Wode (ed.), Papers on language acquisition, language learning and language teaching (pp. 202-12). Heidelberg: Groos.

15 Third Language Acquisition Theory and Practice

WILLIAM P. RIVERS AND EWA M. GOLONKA

Introduction

The popular belief that bilinguals or multilinguals learn a subsequent language more easily than monolinguals has received substantial attention from researchers in recent years, as minority language communities in the European Union achieve greater educational autonomy and as United States government (USG) language training programs seek to respond to rapidly changing requirements for global expertise. In the European context, researchers have paid significant attention to the acquisition of a third language (L3) by childhood bilinguals, such as Basque-Spanish or Catalan-Spanish bilingual secondary school students learning English in Spain, or Swedish-Finnish bilingual children learning English in Finland. These acquisition paradigms have sparked recent research on transfer in L3 acquisition (Cenoz, 2001; Ringbom, 2001, 2007), metalinguistic awareness in L3 acquisition (Cenoz & Valencia, 1994), and parameter setting within the Universal Grammar (UG) paradigm (Klein, 1995; Zobl, 1992).

In this chapter, we present the current state of the theory and practice of adult L3 acquisition in instructional settings. We label this deliberately, in that we are concerned with the theoretical issues arising when adults who know more than one language learn another, as well as the practical issues arising in organized, formal programs for such adult learners. We will show from the literature and from our work that this type of language acquisition differs substantially from adult L2 acquisition.

We begin with a discussion of the state of the research literature in L3 acquisition, and then turn to the praxis, based on our own qualitative research on directed adult L3 acquisition programs in the United States. Throughout, we seek to address the following questions: Why should L3 acquisition differ from L2 acquisition? What are the theoretical implications of the evidence?

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 Bkckwell Publishing Ltd. ISBN: 978-1-405-15489-5

Review of the Literature

In examining the research basis for L3 acquisition, we draw upon a number of areas of inquiry in SLA and psycholinguistics. Some of these areas, such as transfer, do at times address L3 acquisition specifically. Others, such as metacognition in adult language learners, may find that L3 acquisition is a factor in one or another result, at times an unexpected factor. We examine, in turn, findings from work on linguistic transfer, typology and "linguistic distance"; UG and the Interlanguage Structural Conformity Hypothesis (ISCH), which can be seen as outgrowths of the parameter-setting school of the late 1980s and early 1990s; metacognition and metalinguistic awareness; learner autonomy; and last, the small body of work on directed L3 acquisition for adults.

Transfer

From a psycholinguistic perspective, the main focus of research in L3 acquisition is the concept of cross-linguistic influence (see Ringbom & Jarvis, this volume), which includes some aspects of phonetic, morphological, lexical, syntactic, and pragmatic transfer, as well as interference and language attrition related to L3 acquisition. Earlier studies of transfer in L3 sought to identify error as an effect of prior linguistic experience; Ahukanna, Lund, and Gentile (1981), for example, identified semantic errors in learners of French in Nigeria for whom Igbo was L1 and English L2. They found that more advanced students made fewer errors and were thus presumably less susceptible to interference effects, that L2 caused more interference than L1, and that semantic errors were more frequent than other types. The majority of studies on L3 acquisition that have focused on cross-linguistic influence have revealed that it is generally productive and that it facilitates L3 learning. Two related issues arise in these studies: typological distance/ phylogenetic relatedness and transfer. Typological distance has been shown to play an important role in transfer (Cenoz, 2001; Fouser, 1995; Kellerman, 1983; Singleton, 1987). The results of Cenoz's (2001) study indicated that Spanish-Basque bilinguals learning English demonstrated a stronger influence from Spanish, typo-logically a closer language to English, than from Basque, a non-Indo-European language.1 In his recent book, Ringbom (2007) argued that prior cross-linguistic knowledge is very important to the language learner and that the extent to which it may influence the learning of a new language depends on language proximity, i.e., similarities, not language distance, i.e., differences. When discussing language proximity, researchers often refer to the concept of psychotypology (Kellerman, 1983), that is, the way the learner perceives differences between the languages based on previously acquired knowledge, noting that perception of language distance might not correspond to the actual distance between the languages.

Few extant measures are designed to quantify language proximity and thereby determine the distance between languages. Studies of the measurement of

language distance usually address specific pairs of languages and specific aspects of languages, and are not framed in terms of the relative learnability of one or another language. An exception is Chiswick and Miller's (2004) study, in which the researchers discussed and tested the Foreign Service Institute's (FSI) quantitative measure of the distance between English and a number of other languages. Their study, based on US Census data, examined the English proficiency of immigrants in the 1990 US and Canadian Censuses, and is worth some consideration, as it is the first (and thus far, only) attempt to validate a scale of language difficulty in the general population. They impute the titular language of the country of origin as L1, and use the Census question on "how well is English spoken at home" as the outcome variable, assigning null if "not at all" or "not well" and unity if "well" or "very well." From the perspective of SLA research, both the imputation of L1 and the use of self-assessed global proficiency ratings are cause for caution in interpreting the results, but they do find a linear correlation between the FSI difficulty scale and learnability of English for speakers of some 43 languages. Moreover, they test the scale in a way that FSI did not intend when it was developed, in that the FSI scale measures the learnability of a range of foreign languages for English speakers (and more precisely, English-speaking students at the Foreign Service Institute), whereas Chiswick and Miller test the scale against a range of foreign language speakers learning English. In other words, they assume that the scale is reciprocal.2

Recent computational work has led to promising text-based approaches for estimating the relative distance among languages by applying computational measures of entropy between samples in parallel corpora (Juola, 2003), or by estimating genetic drift between related languages, again using parallel texts (Nakhleh, Ringe, & Warnow, 2005). Juola applies the Kullback-Liebler distance, which measures the degree of similarity between two samples of text, providing an "entropy number" which indicates the information loss between two samples, to parallel texts of the Bodleian Oath in more than 30 languages. He thus derives a scalar measure of the degree to which the transliterated sound symbols in each translated version of the oath differ from the English version. Nakhleh et al. apply methods from genetic analysis to estimate the degree of drift among related languages, using primarily Bible corpora from a mother language (such as Latin) and the same texts in daughter languages to estimate the degree of drift. Again, they obtain scalar measures of relatedness among the languages. However, neither method has been used to test learnability, and as with scales such as the FSI and DLI learnability scales, these two methods depend on fixing one language (e.g., English) as the starting point against which difficulty or distance is measured. This brings up two other issues: first, even assuming English as L1, there are a large number of L1-L2-L3 triads possible - tens of thousands. Thus, developing relatedness measures among L1-L2-L3 triads would be impractical. Second, any measure of difficulty must be relative - there are no abstract models to which one might appeal to create an absolute difficulty or learnability measure. Clearly, the linguistic distance between languages cannot easily be measured or quantified.

Besides the distance between the languages, the researchers also have cited the activation of source languages and the order in which they were acquired as interacting with transfer. In the reviewed studies, the source of transfer can be either the learner's L1 (Ringbom, 2001), or a second language or other non-native language (De Angelis & Selinker, 2001; Hammarberg, 2001; Sikogukira, 1993; Leung, 1998). Several factors determine conditions in which transfer occurs and conditions for choosing the source language. Ringbom (2001), for example, argued that transfer can be either L1-based or L2-based, and that the choice of the donor language can depend on typological distance or degree of L2 proficiency. However, what is most important is that the source of transfer depends on the type of transfer investigated, whether lexical, morphological, or syntactic. Ringbom concluded that although lexical transfer is usually L2-based, grammatical and semantic transfers are nearly always L1-based. De Angelis and Selinker (2001) described interlanguage transfer as the influence of one non-native language on another non-native language. They analyzed the language production of two adult multilinguals and concluded that the learners' interlanguages, not the learners' previously learned languages, needed to be taken into consideration while analyzing lexical and morphological transfer. De Angelis and Selinker found evidence for lexical and morphological transfer and noted that this evidence is restricted to the transfer of form only. They argued that transfer of meaning, as well as multilingual language processing, are in need of further research. Leung (1998) also investigated transfer between interlanguages on three grammatical tasks given to Chinese (L1) speakers with a background in English (L2) who were learning French (L3). Leung concluded that two of these tasks revealed L2 transfer to L3 and one reflected more universal factors than transfer from L2.

Another factor discussed in the literature as influencing transfer is L2 status (Hammarberg, 2001). Hammarberg described the different roles that L2 (German) played in speech production of L3 (Swedish) in the speech of his polyglot subject, Sarah Williams. Hammarberg observed that Sarah Williams' lexicon was most often influenced by her L2 German, and that German, not her native English, provided the most supply for her L3 Swedish. After ruling out all other possibilities, he concluded that the reason for L2 being a donor language was L2 status, in other words, the fact that German and Swedish were "foreign languages" for his subject. Hammarberg also referred to the notion of "recency" of the language; that is, the use of the most recently acquired language by the learner as the source of transfer.

In general, researchers agree that transfer from L2 to L3 is most salient in the lexicon, and that lexical transfer has the most facilitative effects on L3 learning. Lexical transfer can help in successful comprehension of texts because learners can rely on a large number of cognates in related languages. Clearly, the existence of false cognates between L2 and L3 may sidetrack learners, but usually the number of true cognates is much larger than the number of faux amis (Ringbom, 2001). Semantic transfer or meaning-based transfer is very often L1-based, even if the L2 and L3 are typologically closely related (Ringbom, 2001). De Angelis and Selinker (2001) analyzed transfer between interlanguages and concluded that the

evidence of lexical and morphological transfer in their study is restricted only to transfer of form. They argued that transfer of meaning, as well as multilingual language processing, was in need of further research.

UG and parameter setting

Zobl (1992) and Klein (1995) examined the effect of prior linguistic experience on UG and parameter setting and concluded that certain qualities, such as less conservative TL grammars, metalinguistic skills, and enhanced lexical knowledge, play a role in facilitating the setting of UG parameters. Zobl (1992), for example, on the basis of trends found in his empirical study, claimed that multilinguals create less conservative interlanguage grammars because of over-generalizations of grammars that enable them to learn TL at a faster rate.

Klein (1995) presented empirical evidence to support the argument that L2 acquisition and L3 acquisition are indeed different. She tested for differences between groups of monolinguals and multilinguals in the acquisition of lexical items and certain syntactic constructions. Specifically, two grammaticality judgment tests were administered to both groups of participants, one on the acquisition of preposition stranding, that is, syntactic constructions, and the other on the acquisition of specific verbs and their prepositional compliments, that is, lexical items. Klein hypothesized that, according to the parameter-setting model of language acquisition, there should be no difference between L2 and L3 learners in terms of parameter setting alone.3 In other words, L3 learners should not have advantages over L2 learners because the parameters had been set in L1, and it should not matter whether it is the learner's L2, L3, or L4. However, the results of Klein's study showed differences between the two groups, with multilinguals outperforming monolinguals on the two types of tests (grammaticality judgments of preposition stranding and verb subcategorization); however, the difference between the groups on a third test (grammaticality judgments of null-prep) was not statistically significant.4 Klein concluded that heightened metalinguistic skills, enhanced lexicon, and less conservative learning helped multilinguals in re-setting of UG parameters.

Metacognition and metalinguistic awareness

The facilitative role of previous linguistic experience in L3 learning can be found in the area of metacognition, in that the experience of learning another language may later employ the learner's metacognitive self-assessment and metacognitive self-management, in addition to providing the learner with explicit, declarative metalinguistic knowledge about how languages work. Ramsay (1980) discovered that multilinguals tended to perform far better than monolinguals on an achievement test. Successful learners in Ramsay's study tended to use more cognitive and metacognitive strategies, including substantially more verbalization and vocal practice, use of mnemonic devices, a more positive attitude toward the learning process (an affective strategy), use of positive affect reinforcement, use

of more sources of information, and more risk taking and less fear of errors. Ramsay noted that metacognitive strategies can be distinguished as the primary difference between novice and expert learners across a broad set of abstract systems of knowledge, specifically in language, at both a very discrete level (the processing of specific constituent units) and a discourse level. Wenden (1999) reviewed literature on metacognition in language learning and drew a similar conclusion: Good language learners as well as self-directed language learners exhibit metacognitive behaviors. Mohle (1989) examined learning strategies in German multilingual university students taking a variety of Indo-European languages (French, Spanish, and English), hypothesizing that the narration of a film with no overt linguistic information would be influenced by cognitive processing. She found evidence of controlled lexical transfer, again a metacognitive strategy.

Metalinguistic awareness is believed to play an important role in L3 acquisition and in the investigation of the differences between L2 and L3 acquisition (Bartlet, 1989; Jessner, 1999; Klein, 1995; Thomas, 1988; Zobl, 1992). Bartlet (1989) examined the influence of L3 on procedural (metalinguistic) knowledge in a case study on interference in L3 (English) from L2 (Spanish) among multilingual Yaqui Indians in Arizona. Bartlet found evidence of broad use of metacognitive and communicative strategies in oral discourse and oral narrative. Jessner (1999) investigated the role of metalinguistic awareness in multilinguals within the framework of the Dynamic Model of Multilingualism. She analyzed qualitative data collected from trilingual adults via think-aloud protocol sessions to illustrate different strategies learners use in searching for a word in L3. All of the cited examples relate to metalinguistic thinking that involves three languages, all of which are typologically related. For instance, Jessner provided examples of code-switching involving either two or three languages, along with learners' attempts to look for equivalents or cognates in the three languages. Basing her conclusions on the relevant literature and this study alone, Jessner argued that metalinguistic awareness is desirable and can be increased through teaching similarities between languages and through activating students' prior linguistic knowledge. Two other studies, Zobl (1992) and Klein (1995), attributed better performance of multi-linguals than monolinguals on grammaticality judgment tests to metalinguistic awareness.

Learner autonomy

Learner autonomy has also surfaced as an area of interest in the study of L3 acquisition. We define learner autonomy here as the active, independent management of learning by the learner (rather than independent study outside the classroom), where the learner sets or attempts to control the goals, curriculum, pedagogical method, or content of the learning program. A qualitative method was applied by Rivers (1996, 2001), who used a variety of data-collection techniques, including questionnaires, focus groups, classroom observation, and interviews, to investigate the characteristics and behaviors of college students learning a third language. Rivers (1996) compared proficiency outcomes of L3

learners from three programs: a program in languages of the former Soviet Union at the University of Maryland, a cross-training program at the Defense Language Institute (DLI), and a DLI study of immersion training with learners enrolled in L2 courses. He found that L3 learners were more successful than L2 learners of the same target language; that is, they learned more of the language and in a shorter time than L2 learners, based on end-of-course proficiency tests administered by DLI and other USG agencies. In another study, Rivers (2001) analyzed self-directed language learning behaviors of 11 adult learners of Georgian and Kazakh as a third language. All subjects' L1 was English and all had advanced (Interagency Language Roundtable (ILR) 3/3/3 or better) proficiency in L2 Russian. Rivers found that all learners accurately assess their progress, learning styles, and learning strategy preferences, as well as conflicts with teaching styles and behaviors of other learners within the class. Next, based upon these self-assessments of learner styles and preferences, all learners revealed a high tendency toward controlling their learning process, the tendency toward learner autonomy being demonstrated by their requesting and demanding changes to course content and structure. Finally, the majority of learners made attempts to modify the learning environment by using self-directed language learning strategies that referred to different aspects of course structure, for example, type and mode of input, workload, and classroom activities.

Praxis: Research on Adult L3 Instruction

In an effort to promote Portuguese language courses in American colleges and universities, Holton (1954) provided an overview of similarities between Spanish and Portuguese, claiming that learners with a good command of Spanish could acquire a reading facility in Portuguese in a very short time and with minimum effort. As he put it, "It would seem to be a valuable piece of intellectual merchandise obtained at a wonderful bargain price" (p. 447). Jensen (1989) and Jordan (1991) suggested that the high degree of mutual intelligibility between Portuguese and Spanish could be used in teaching Portuguese to students who had Spanish as a second language. Jensen (1989) administered a series of listening proficiency tests in Spanish to Portuguese speakers, and tests in Portuguese to Spanish speakers, and found that Portuguese was 60 percent intelligible to Spanish speakers and that Spanish was 50 percent intelligible to Portuguese speakers. Jordan (1991) argued for the use of contrastive analysis techniques in teaching Portuguese to speakers of Spanish, and listed the pedagogical benefits and risks of using this technique for teaching closely related languages. Based on the principle that there is a high degree of mutual intelligibility between closely related languages, such as Spanish and Portuguese, Gribble (1987) created a Bulgarian course for Russian speakers and Townsend (1995) a Czech course for Russian speakers. The efficacy of these courses, which are highly contrastive in nature, has not been measured; we note them here as they do, at least implicitly, draw

upon the assumption of mutual intelligibility and lexical and linguistic transfer between closely related languages.

A notably different view was offered by Teixeira-Leal Tarquinio (1977), who recommended that, for English-speaking students at American universities, it is unsound to take both Spanish and Portuguese concurrently, pointing to the interference of Spanish, especially in beginning Portuguese classes. She warned that this practice could lead to a hybrid product of "espangues." Teixeira-Leal Tarquinio provided a list of specific phonological, morphological, and syntactical items whose transfer may cause difficulties for these students, claiming that one of the two languages must be mastered before beginning to learn the other. It needs to be noted that the author based her view on observations of students in American universities learning Portuguese and Spanish concurrently. This view seems to support claims about the importance of L2 proficiency in L3 acquisition. What needs to be determined in future research is the effect of the proximity of Spanish and Portuguese, or any other closely related languages, on proficiency level and order of acquisition.

In the US, the primary locus of deliberate L3 instruction has been the language training institutes of the USG. Over the past 15 years, rapidly changing and emerging government language requirements, coupled with the assumption that significant time savings could be achieved in L3 training courses, have led to courses in Serbian/Croatian for Russian, Polish, and Czech second language learners; Tausug for Tagalog speakers; Malaysian for Bahasa Indonesian speakers; Portuguese for Spanish speakers; Kazakh, Kyrgyz, and Georgian for Russian speakers; Haitian Creole for French Speakers; and several courses in one or another Arabic vernacular for speakers of Modern Standard Arabic.

Several courses at the Defense Language Institute Foreign Language Center (DLI) have retrained speakers of one language in another closely related language: Czech L2 speakers in Serbian/Croatian (Corin, 1994), French speakers in Haitian Creole, and Russian, Polish, and Czech speakers in Serbian/Croatian. This type of instruction, in which the target language and the learners' previously known languages are closely related, is called conversion. Corin (1994) reported on a Serbian/Croatian conversion course at the DLI that retrained 40 Czech linguists in Serbian/Croatian in a 3-month period. Based on the outcomes of the three-month course reported on the ILR scale (median oral proficiency score = Level 2; mode = Level 1+), Corin concluded that conversion works, that L2 proficiency may influence L3 gains, and that learner style interacted with the teaching materials and methods (global learners performed better). There were no traditional grammars; the learners had to derive rules for the target language from their L2s. Kulman and Tetrault (1993) reviewed USG L3 courses and proposed Rapid Survey courses for closely related languages, for example, a Ukrainian course for Russian speakers. According to Kulman and Tetrault, such courses would make use of the phonologies, morphologies, and syntaxes of the L2 and TL, as well as contrastive analysis, to enable L3 learners to predict parallel and divergent structures in languages.

Current best practices in USG conversion training were summarized in Brecht et al. (1998). The authors claimed that best practices in conversion training concentrate on attempts to capitalize upon the students' prior knowledge and prior language learning experience. Brecht et al. list three advantages that individuals with previous language learning background possess over persons without such experience: (1) prior knowledge of the donor language (L2), (2) greater metalinguistic awareness, and (3) general language-learning skills. The authors also explain two models of USG cross-training and conversion programs: the Quick Response Short-Term Demand Model and the Long-Term Programmed Career-Based Model. In the former, also called the "crisis model," cross-training/ conversion is a method for responding to changing requirements as a result of rapidly developing geopolitical events that can unfold much faster than the time normally required to train a linguist. However, such courses are often organized on short notice, with little time to develop curricula, methods, and materials, select students, and to train teachers. The latter, also called the "programmed career" model, is based on the idea of expanding the language background of USG personnel with language qualifications as part of their career development. Brecht et al. argued that very often the "crisis model" is applied rather than the more desirable "programmed career" model.

For closely related languages, the potential programmatic benefits of this training are immediately clear for the rate of acquisition. This advantage is vital: faster rate of acquisition leads to shorter courses; shorter courses require less investment and are more responsive to rapidly emerging language requirements, such as those that arose during the US interventions in Bosnia and Kosovo. In fact, Rivers (2001) reported that cross-training courses in Kazakh and Kyrgyz (with Russian as L2) were one third the length of basic courses in these Category 3 languages. Corin (1994) and Brecht, et al. (1998) reported that Serbian/Croatian conversion courses were between one half and two thirds the length of the basic course.

Regarding the putative benefit of L3 acquisition, and the use of cross-training and conversion courses by USG training schools, we return to the hypothesis initially stated: For the learner, the acquisition of a third language is aided by knowledge of a second language. The literature is in broad agreement that L3 learners are at an advantage, be they adults who, having acquired a second language as adults, turn to a third, as is the case in most USG cross-training and conversion programs, as well as much of higher education when L3 learning occurs, or children in multilingual societies learning a foreign language in addition to their native language (typically a minority language) and the titular language. The advantages such learners enjoy in comparison to ab initio adult second language learners include:

• the use of more metacognitive behaviors, particularly metacognitive self-management, in the learning process used by the learner to direct the learner's in-class learning behaviors, as well as language use behaviors outside of the formal instructed environment;

• the use of a wider variety of cognitive learning strategies;

• more demonstration of autonomous learning;

• better affective behaviors and, in particular, a more positive attitude toward the learning process;

• higher proficiency outcomes for a given course length.

However, these advantages do not accrue equally for all adult L3 learners. Rivers (1996, 2001) noted distinct differences in affective behaviors with respect to adjusting to varying learning styles among the students and between the students and instructors. Some learners were able to negotiate these differences with little adverse effect on themselves or the classroom environment, while others were not. Rivers attributed this to poor metacognitive self-assessment. Other studies on linguistic transfer have noted that the advantages often afforded by transfer can be mitigated by inaccurate transfer: false cognates (lexical), overgeneralization of morphological and syntactic constructions, and so forth. However, overall, the second language learner with reasonably high L2 proficiency is a prime candidate for L3 training. The findings from studies on linguistic transfer also point to the typological distance between languages as an important factor in positive transfer - the more closely related the languages, the more linguistic transfer is possible. There is still not enough research conducted on typologically unrelated languages to provide evidence on the role of linguistic knowledge during L3 learning.

Summary of current research and praxis

Our rather cursory review reveals several areas where one might well posit that L3 acquisition is qualitatively and quantitatively different from SLA. We would argue that L3 acquisition then requires either a new theoretical framework unique to L3 acquisition (such as proposed by Flynn, Foley, & Vinnitskaya, 2004; Herdina & Jessner, 2002; and Hufeisen, 2000) or an extended SLA framework (such as argued by Marx & Hufeisen, 2003) to account for L3 acquisition within a broader framework of adult language acquisition. Such an extended framework would likely focus on the psycholinguistic and neurolinguistic changes wrought by acquiring a second language to some significant level of competence. One might further argue that such a framework would necessarily focus on metacognitive and metalinguistic traits to a significant degree, in that these areas are rather less well investigated than typology or relatedness, for example, and would tend to be independent of the particular L1-L2-L3 triads in question for a given learner.

Insofar as the applications and results of L3 acquisition differ from SLA - we know from established practice in L3 acquisition that the programs and outcomes can differ significantly, and that learner processes may differ considerably at the psycholinguistic, sociolinguistic, neurolinguistic, cognitive, and metacognitive levels - one might well argue that programming in L3 acquisition ought to capitalize on these factors. We turn next to a case study of the state of the art for directed L3 acquisition programs for adults in the US.

A Case Study of USG L3 Programs

The authors recently conducted a study of practices in USG L3 training with the goal of enhancing instruction in third language courses. We collected retrospective survey data from program managers and instructors of current and past L3 courses in:

• Tagalog to Tausug

• Bahasa Indonesian to Malaysian

• Spanish to Portuguese

• Russian to Serbian/Croatian

• Russian to Ukrainian

• Ukrainian to Russian

• Multiple languages to Uzbek

• Mandarin to Cantonese

• Modern Standard Arabic (MSA) to Egyptian Arabic

• MSA to Levantine Arabic

• Hindi to Urdu

Most courses met for 16 weeks, 6 hours per day, 5 days per week, with the exception of Russian to Serbian/Croatian courses that met for 12 or 16 weeks and the Bahasa Indonesian to Malaysian course that met for 34 weeks. A total of 39 people were interviewed: 18 language instructors, 5 program managers, and 16 students. Students in these courses were adult learners of the target language and may or may not have been native speakers of the language from which they were retrained. In some courses proficiency in L2 was measured, and in some, attainment of a certain proficiency level was required.

Analytical methodology

The authors took a qualitative, grounded theory approach to analyze a corpus of ethnographic data collected from classroom observations and learner, instructor, and course manager focus groups at six language schools. The focus groups were semi-structured face-to-face interviews with small groups of informants. The team also observed two basic MSA classes (six and eight students in each) for comparison purposes. All observed classes were one hour long. Two or three researchers were present in each of the classes; they sat in the back and took notes. The notes included: a running list of classroom activities, notes on classroom behaviors, direct quotes of the learners and teachers.

Results and discussion

The courses we examined differed on external factors related to learning environment, such as course goals, methods and materials used, and teacher

characteristics. For the most part, they were intended as L3 courses, with the exception of the Uzbek course, which was initially designed as a basic course. However, because of the presence of students with backgrounds in typologically close languages, such as Azeri and Turkish, as well as languages that have donated substantial lexis to Uzbek, such as Arabic and Persian, the Uzbek course evolved into an L3 course. L3 goals were not always explicitly set and communicated to students. In Serbian/Croatian courses, for example, L3 goals were not explicitly included in the course description, but students were told that they were taking an L3 course.

Beyond confirming that the primary perceived benefit for L3 courses in the language training institutes of the USG is the more rapid acquisition of the target language, the signal result of this preliminary case study of L3 acquisition course structure and goals is that the courses themselves differed greatly in terms of which putative advantages of L3 acquisition were exploited. The courses we examined differed significantly in relation to factors external to the learners themselves, particularly with respect to course goals (whether the course was explicitly an L3 course and whether that goal was expressed in course design, pedagogical methodology, or course materials); transfer (whether transfer was explicit in course design and materials and whether contrastive analysis was used); instructor background (some instructors had a background in the learners' second language, while others did not); and materials used (materials were typically adopted from basic language courses with little or no modification for the L3 course, although there were notable exceptions).

On the other hand, in our interviews, instructors, program managers, and students indicated striking similarities in their reports of learner behavior. We summarize these perceptions in Figure 15.1. Learners accessed transfer phenomena, typically but not exclusively doing so in an overt fashion, comparing their L2 to L3, regardless of course design or instructor input. Learners clearly noticed and took advantage of differences and similarities between the two languages, but whether the phenomenon was noticing (without awareness) or metalinguistic awareness, remains unanswered. L3 learners also demonstrated a high degree of autonomy in language learning processes and usage of metacognitive learning strategies, such as organization, planning, and evaluation. All in all, they were taking advantage of their previous learning experience. Most courses proceed from an assumption that the greatest benefit is derived from retraining highly proficient learners into a closely related language. This may confound some of the putative benefits. That is, by screening for such advanced, experienced learners, USG L3 programs tend to select learners who might enjoy these and other advantages due to aptitude, verbal skills, maturity, or general intelligence, rather than their experience with a second language.

Directions for further research

This study, and the larger project to which it belongs, raises interesting questions about L3 learning in general. First, the best evidence we have thus far is that L3

Learning resources

Moderating factors

Instructional issues

Best practices

Figure 15.1 Maximizing prior language learning and experience in conversion training

courses in the USG tend to succeed - that is, the goal of reaching a specified proficiency in one half to two thirds of the time required for attainment of comparable proficiency via an L2 course is achieved - in a remarkably uniform and consistent way, regardless of the specifics of course design, instructional methods, or materials. We intend over the next several years to test this assumption by a careful ethnographic examination of diverse L3 training courses, coupled with a correlational study of learner factors predictive of success in L3 training courses. It may be that the chief influence on the rate of acquisition is some construct of typological and genetic distance between the learners' second and third languages. The metalinguistic and metacognitive effects noted in the research literature and echoed in our interviews hold some significant promise for SLA, in particular in advanced and independent learning situations, such as extended immersion abroad or daily professional use of the language.

We would advance a brief set of research questions for L3 acquisition, in hopes of sparking a discussion of the kinds of research reviewed in our chapter, as well as some of the initial results from studies of L3 acquisition. In other words, we would advance a set of questions unified around L3 acquisition as a theme, given the interest in L3 acquisition in the US for adults and the prevalence of L3 acquisition in other countries with complex multilingual societies.

• Does bilingualism - of any type - bring about changes in language acquisition in general?

• Does bilingualism - of any type - bring about changes in general cognition and/or the cognitive processes associated with adult language acquisition?

• Does bilingualism - of any type - bring about changes in general metacognition and/or the metacognitive processes associated with adult language acquisition?

• Does bilingualism - of any type - bring about changes in neurophysiology? How do such changes affect adult SLA?

We believe that these approximate a research agenda for the L3 acquisition field, one that would greatly enrich our understanding of the processes underlying all adult language acquisition.

1 By linguistic distance, Cenoz means phylogenetic relationship. See Nakhleh, Ringe, and Warnow (2005) for a detailed assay at defining and measuring the relative distance among Indo-European languages and Juola (2003) for an attempt to create scalar measures for relative distances among unrelated languages. Distance has also been defined in terms of learnability; the difficulty scales employed by the Foreign Service Institute (FSI) and the Defense Language Institute (DLI) are examples.

2 One should note that other attempts to assay the learnability of specific features in L1-L2 dyads, such as the Interlanguage Structural Conformity Hypothesis (see, for example, Carlisle, 1998), do not assume reciprocity for learnability of specific features.

3 In broad terms, the Principles and Parameters model holds that linguistic structures -syntactic, lexical, phonological - are governed by a set of parameters, the setting of which allows or disallows phenomena such as trace movement in question formation or obligatory devoicing of final stops. These settings are fixed during first language acquisition, and if they should vary in the second (or third) language, then the corresponding parameter must also be "reset" for that language, to allow proper production of the particular structure.

4 Preposition stranding may occur in question formation when the trace noun is found in a prepositional phrase: "The students worried about the test," can become "About which test did the students worry?" (so-called Pied-Piping) or "Which test did the students worry about?" (preposition stranding) (Klein, 1995, p. 432). The incorrect response "Which tests did the student worry" occurs among learners of English whose L1 does not have prepositions which govern and mark case. In Klein's study, verb sub-categorization refers to the complements taken by the allowable prepositions governed by the verb; and null-prep refers to incorrect stimuli without the required preposition.

REFERENCES

Ahukanna, J., Lund, N., & Gentile, J. (1981). Acquisition of problem solving skills. In J. Anderson (ed.), Cognitive skills and their acquisition (pp. 191-230). Hillsdale, NJ: Lawrence Erlbaum.

Bartlet, G. (1989). The interaction of multilingual constraints. In H. Dechert & M. Raupach (eds.), Interlingual process (pp. 5-16). Tübingen: Gunter Narr.

Brecht, R., Frank, V., Rivers, W., & Walton, A. (1998). A guide for foreign language cross-training. Research Report No. 98-01. Monterrey, CA: Defense Language Institute Foreign Language Center.

Carlisle, R. (1998). The acquisition of onsets in a markedness relationship: An empirical study. Studies in Second Language Acquisition 20, 245-60.

Cenoz, J. (2001). The effect of linguistic distance, L2 status and age on cross-linguistic influence in third language acquisition. In J. Cenoz, B. Hufeisen, & U. Jessner (eds.), Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives (pp. 820). Clevedon, UK/Buffalo: Multilingual Matters.

Cenoz, J. & Valencia, J. F. (1994). Additive trilingualism: Evidence from the Basque Country. Applied Psycholinguistics 15, 195-207.

Chiswick, B. & Miller, P. (2004). Linguistic distance: A quantitative measure of the distance between English and other languages. Journal of Multilingual and Multicultural Development 26, 1, 1-18.

Corin, A. (1994). Teaching for proficiency: The conversion principle. A Czech to Serbo-Croatian conversion course at the Defense Language Institute. ACTR Letter: Newsletter of the American Council of Teachers of Russian 20, 1, 1-5.

De Angelis, G. & Selinker, L. (2001). Interlanguage transfer and competing linguistic systems in the multilingual mind. In J. Cenoz, B. Hufeisen, & U. Jessner (eds.), Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives (pp. 42-58). Clevedon, UK/Buffalo: Multilingual Matters.

Flynn, S., Foley, C., & Vinnitskaya, I. (2004). The Cumulative-Enhancement Model for language acquisition: Comparing adults' and children's patterns of development in

first, second and third language acquisition of relative clauses. International Journal of Multilingualism 1, 1, 3-16.

Fouser, R. (1995). Problems and prospects in third language acquisition research. Language Research 31, 2, 387-415.

Gribble, C. (1987). Reading Bulgarian through Russian. Columbus, OH: Slavica.

Hammarberg, B. (2001). Roles in L1 and L2 in L3 production and acquisition. In J. Cenoz, B. Hufeisen, & U. Jessner (eds.), Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives (pp. 21-41). Clevedon, UK/Buffalo: Multilingual Matters.

Herdina, P. & Jessner, U. (2002). A dynamic model of multilingualism: Perspectives of change in psycholinguistics. Clevedon, UK: Multilingual Matters.

Higgs, T. & Clifford, R. (1982). The push toward communication. In T. Higgs (ed.), Curriculum, competence, and the foreign language teacher. Skokie, IL: NTC & ACTFL.

Holton, J. (1954). Portuguese for Spanish Speakers. Hispania 37, 4, 446-52.

Hufeisen, B. (2000). A European perspective: Tertiary languages with a focus on German as L3. In J. Rosenthal (ed.), Handbook of undergraduate second language education: English as a second language, bilingual, and foreign language instruction for a multilingual world (pp. 209-29). Mahwah, NJ: Lawrence Erlbaum.

Jensen, J. (1989). On the mutual intelligibility of Spanish and Portuguese. Hispania 72, 84852.

Jessner, U. (1999). Metalinguistic awareness in multilinguals: Cognitive aspects of third language learning. Language Awareness 8, 201-9.

Jordan, I. (1991). Portuguese for Spanish speakers: A case for contrastive analysis. Hispania 74, 788-92.

Juola, P. (2003). The time course of language change. Computers and the Humanities 37, 1, 77-96.

Kellerman, E. (1983). Now you see it, now you don't. In S. Gass & L. Selinker (eds.), Language transfer and language learning (pp. 112-34). Rowley, MA: Newbury House.

Klein, E. (1995). Second versus third language acquisition: Is there a difference? Language Learning 45, 419-65.

Kulman, A. & Tetrault, E. (1993). Language cross training: Alternatives and permutations.

Cryptologic Quarterly.

Leung, Y.-K. I. (1998). Transfer between interlanguages. In A. Greenhill, M. Hughes, H. Littlefield, & H. Walsh (eds.), Twenty-Second Boston University Conference on Language Development (pp. 477-87). Medford, MA: Cascadilla Press.

Marx, N. & Hufeisen, B. (2003). Multilingualism: Theory, research methods and didactics. In G. Brauer & K. Sanders (eds.), New vision in foreign and second language education (pp. 178-203). San Diego: LARC Press.

Möhle, D. (1989). Multilingual interaction in foreign language production. In H. Dechert & M. Raupach (eds.), Interlingual process (pp. 179-94). Tübingen: Gunter Narr.

Nakhleh, L., Ringe, D., & Warnow, T. (2005). Perfect phylogenetic networks: A new methodology for reconstructing the evolutionary history of natural languages. Language 81, 2, 381-420.

Ramsay, R. (1980). Language-learning approach styles of adult multilinguals and successful language learners. In V. Teller & S. White (eds.), Studies in child language and multilingualism (pp. 73-96). New York: New York Academy of Sciences.

Ringbom, H. (2001). Lexical transfer in L3 production. In J. Cenoz, B. Hufeisen, & U. Jessner (eds.), Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives (pp. 59-68). Clevedon, UK/Buffalo: Multilingual Matters.

Ringbom, H. (2007). Cross-linguistic similarity in foreign language learning. Clevedon, UK: Multilingual Matters.

Rivers, W. (1996). Self-directed language learning and third-language learners. ERIC Document ED411679.

Rivers, W. (2001). Autonomy at all costs: An ethnography of metacognitive self-assessment and self-management among experienced language learners. Modern Language Journal 85, 279-90.

Sikogukira, M. (1993). Influence of languages other than the L1 on a foreign language: A case of transfer from L2 to L3. Edinburgh Working Papers in Applied Linguistics 4, 11032.

Singleton, D. (1987). Mother and other tongue influence on learner French. Studies in Second Language Acquisition 9, 327-46.

Teixeira-Leal Tarquinio, L. (1977). The interference of Spanish in beginning Portuguese class. Hispania 60, 1, 82-7.

Thomas, J. (1988). The role played by metalinguistic awareness in second and third language learning. Journal of Multilingual and Multicultural Development 9, 235-46.

Townsend, C. (1995). Teaching the Czech language through Russian: Преподавание ческого языка посредством русского (Prepodavanije ceskogo jazyka posredstvom russkogo). Columbus, OH: Slavica.

Wenden, A. (1999). Metacognitive knowledge and language learning. Applied Linguistics 19, 515-37.

Zobl, H. (1992). Prior linguistic knowledge and the conservatism of the learning procedure: Grammaticality judgments of unilingual and multilingual learners. In S. Gass & L. Selinker (eds.), Language transfer in language learning (pp. 176-96). Amsterdam/ Philadelphia: John Benjamins.

Part V Course Design and Materials Writing

The Handbook of Language Teaching Editedby MicheelH.LnngandCethermeL.Doughty © 2009 Blackwell PublishDgLtd. ISBN: 978-1-405-15489-5

16 Foreign and Second

Language Needs Analysis

JAMES DEAN BROWN

Introduction

To start this chapter, I will address the three most basic questions about needs analysis: What is needs analysis? What literature is available on language needs analyses? And, what are the steps in needs analyses? The rest of the chapter will then provide more details about each step.

What is needs analysis?

The concept of needs analysis (also known as needs assessment) commonly refers to the processes involved in gathering information about the needs of a particular client group in industry or education. Naturally, in educational programs, needs analyses focus on the learning needs of students, and then, once they are identified, needs are translated into learning objectives, which in turn serve as the basis for further development of teaching materials, learning activities, tests, program evaluation strategies, etc. Thus needs analysis is the first step in curriculum development.

In Brown (1995, p. 36), I provided a more formal definition which combined a number of other definitions I had found in the literature. Unfortunately, the resulting definition was a rather ponderous one: "the systematic collection and analysis of all subjective and objective information necessary to define and validate defensible curriculum purposes that satisfy the language learning requirements of students within the context of particular institutions that influence the learning and teaching situation."

I will simplify that definition here by breaking it into several pieces. Needs analysis (NA) is the systematic collection and analysis of all information necessary for defining a defensible curriculum. A defensible curriculum is one that satisfies the language learning and teaching requirements of the students and teachers within the context of particular institution(s) involved. (I will necessarily return to this issue below.) Naturally, the information necessary to achieve this

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. IBBN:978-l-404-15449-5

Table 16.1 Steps in the NA process

Schntz and Deriving Jordan (1997, p. 23) Graves (2000, p. 100): Stages (steps combined)

(1981, p. 35)

1 Define the purpose 1 Define the purpose of 1 Decide what A Get ready to do NA

2 Delimit the target analysis information to 1 Define the purpose of the NA

population 2 Delimit student gather and why 2 Delimit the student population

3 Delimit the population 2 Decide the best 3 Decide upon approaches) and

parameters 3 Decide upon way to gather it: syllabus(es)

of investigation approach(es) when, how and 4 Recognize constraints

4 Select the 4 Acknowledge from whom 5 Select data collection procedures

informationgathering constraints /limitations 3 Gather the B Do the NA research

instrument 5 Select methods of information 6 Collect data

5 Collect the data collecting data 4 Interpret the 7 Analyze data

6 Analyze the results 6 Collect data information 8 Interpret results

7 Interpret the 7 Analyze and interpret 5 Act on the C Use the NA results

results results information 9 Determine objectives

8 Critique 8 Determine objectives 6 Evaluate the effect [Implement decisions (assessments,

the project 9 Implement decisions and effectiveness materials, teaching strategies, etc.)]

(i.e., decide upon of the action 10 Evaluate and report on the NA

syllabus, content, 7 Decide on further project

materials, methods, etc.) or new information [Decide on further information to gather

10 Evaluate procedures to gather (for ongoing curriculum evaluation)]

and results

defensible curriculum includes all subjective and objective information, and any other types of information that turn out to be appropriate in the particular NA.

What literature is available on language NA?

Over the years, a few books have been devoted to the specific notion of language NA, including Richterich and Chancerel (1977, 1987), Munby (1978), Trim (1980), Buckingham (1981), Richterich (1983), and most recently Long (2005b). In addition, a number of papers have been published that review the literature on NA, most recently those by West (1994, 1997) and Long (2005a). Generic language curriculum books have also included increasingly large sections on NA over the years: Dubin and Olshtain (1986, pp. 13-14); Nunan (1988, pp. 43-53; 1991, pp. 13-24, 75-84), Clark (1987, pp. 35-7), Yalden (1987, pp. 131-8), White (1988, pp. 83-93), Brown (1995, pp. 37-70), Graves (2000, pp. 97-121), and Richards (2001, pp. 51-111).

What are the steps in a NA?

Schutz and Derwing (1981) advocated using eight steps in a NA, "which would seem to constitute an absolute minimum for any needs assessment effort worthy of the name" (p. 35); Jordan (1997, p. 23) argued for ten steps, while Graves (2000, p. 100) listed seven steps. Notice that all three sets of steps have been combined in the last column of Table 16.1 into what I consider three general stages of NA (Get ready to do NA; Do the NA research; and Use the NA results), with ten secondary steps listed under those stages.1 The rest of this chapter will be organized around the ten steps shown in the last column of Table 16.1 followed by a conclusion section.

Get Ready to Do NA

Define the purpose of the NA

A number of perspectives have been proposed for thinking about the purposes of NA. These perspectives generally take the form of frameworks within which NAs can be conducted or types of information that can serve as the basis of a NA.

Frameworks

Stufflebeam et al. (1985, cited in Brown, 1995, pp. 38-9) identified the following four philosophies that can underlie NA (with language-related definitions supplied by me):

1 Discrepancy philosophy - needs are any differences between future desired student language performances and what they can currently do.

2 Democratic philosophy - needs are any learning goals that are preferred by a majority of the stakeholders involved.

3 Analytic philosophy - needs are whatever the students would naturally learn next based on what is known about them and the learning processes involved.

4 Diagnostic philosophy - needs are any language elements or skills that would be harmful if missing.

Similarly, three different types of needs were discussed by Hutchinson and Waters (1987):

1 Necessities are "objective needs," or what learners need to know to successfully function in the target L2.

2 Lacks are differences between target L2 proficiency and what learners currently know (see the discrepancy philosophy above).

3 Wants are "subjective needs," or what and how the learners would like to learn.

Brindley (1984, p. 28) pointed to even more types of needs including wants, desires, demands, expectations, motivations, lacks, constraints, and requirements.

West (1994, pp. 8-12; 1997, pp. 71-4) and Jordan (1997, pp. 23-8) provided lists of different types of NA that overlap considerably with each other. A combination of the three lists would include the following nine types of language NA:

1 Target-situation analysis seeks information on the language requirements learners face in learning a specific type of language.

2 Deficiency analysis accounts for learners' current wants and needs and their target situation deficiencies or lacks.

3 Present-situation analysis focuses on the students' proficiencies at the outset of instruction.

4 Learning-oriented analysis takes the view that needs (in terms of syllabus, content, teaching methods, materials, etc.) should be negotiated between students and other stakeholders.

5 Strategy analysis focuses on learners' preferences in terms of learning strategies, error correction, group sizes, amount of homework, etc.

6 Means analysis focuses on the learning situation, with as few preconceptions as possible in terms of practicality, logistics, cultural appropriateness, etc.

7 Language audits take a large-scale view of NA in terms of strategic language policies for companies, professional sectors, governmental departments, countries, etc.

8 Set menu analysis sets out to create a menu of main courses from which the sponsors or learners can select.

9 Computer-based analysis is done by computer to match perceived needs to a database of materials ". . . from which the course content can be negotiated between students and teacher ..." (West, 1997, p. 74).

Types of information

Berwick (1989, pp. 49-51) listed six planning orientations for the types of information in educational systems:

1 Organized body of knowledge

2 Specific competencies

3 Social activities and problems

4 Cognitive or learning processes

5 Feelings and attitudes

6 Needs and interests of the learner

A variety of other different types of information can serve as the basis of a NA, for example:

1 Goal-oriented vs. process-oriented language needs (Widdowson, 1981, p. 2)

2 Language content vs. learning content (Brindley, 1984, pp. 31-2)

3 Content vs. methodology parameters (Nunan, 1985)

4 Linguistic content vs. learning process (Brown, 1995, p. 41)

5 Language needs vs. situation needs (Brown, 1995, p. 40; Richards, 2001, pp. 90-1)

Some combination of all these alternative frameworks and types of information will probably serve best in most present day NAs.

Delimit the student population

There are at least two ways in which the student population should be delimited. One is in terms of the scale of the project and the other is in terms of the focus of the NA.

Scale of the NA

The scale of a NA has to do with how broadly it is aimed. Some NAs have been large-scale indeed. For example, the international, multi-language NA efforts of the Council of Europe were designed to identify the needs of adult foreign language learners (e.g., Conseil de la Coopération Culturelle, 2000; Council of Europe, 2001; Richterich, 1983, 1985; Richterich & Chancerel, 1977, 1987; Trim, 1973, 1980). Somewhat more modestly scaled NAs have been described for the societal level in the United States (Brecht & Rivers, 2005), for the US military (Lett, 2005), and for large organizations like universities (Coleman, 1988).

In theory, a NA could be conducted on international, national, state or provincial, county or school district, multi-program, program, or classroom levels, all of which represent quite different scales. However, most commonly, NAs are conducted at a local level. Indeed, Purpura et al. (2003, p. 9) go so far as to argue that "All needs assessments are situation-specific" (italics in original). The questions of how specific the situation should be and, by extension, how specific a NA should be are crucial ones that must be addressed in the early stages of any such efforts (for more on this issue, see Hyland, 2002; Dovey, in press).

Focus of the NA

Seedhouse (1995) discusses NA for the general-English classroom. However, by and large, NAs have been conducted for much more specific purposes. As a result, in teacher-training texts on English for specific purposes (ESP), NA usually figures prominently. For example, there are 11 pages on NA in Hutchinson and Waters (1987, pp. 53-64), 22 pages in Jordan (1997, pp. 20-42), and 12 pages in Dudley-Evans and St John (1998, pp. 25-6, 57-9, 121-7). Typically, ESP is further divided into two categories, one for occupational purposes (EOP) and the other for academic purposes (EAP).

EOP includes many subcategories in which NAs have been conducted: business and workplace (e.g., Chew, 2005; Crosling & Ward, 2002; Edwards, 2000; Holliday, 1995; Holmes, 2005; So-mui & Mead, 2000; Tanaka, 2001; van Hest & Oud-De Glas, 1990; Vandermeeren, 2005), healthcare (e.g., Bosher &Smalkoski, 2002; Cameron, 1998; Chia et al., 1999; Lepetit & Cichocki, 2002; Uvin, 1996), hotel workers (Jasso-Aguilar, 1999, 2005), journalists (Gilabert, 2005), footballers (Kellerman, Koonen, & van der Haagen, 2005), and coffee shop workers (Downey Bartlett, 2005).

EAP has also been the focus of many NAs including general EAP needs (e.g., Braine, 2001; Chan, 2001; Jordan, 1997; Kim et al., 2003; Waters, 1996), writing needs (Casanave & Hubbard, 1992; Hale et al., 1996; Leki, 1995), oral/aural needs (Ferris, 1998; Ferris & Tagg, 1996a, 1996b; Kim, in press; Teng, 1999), and heritage learner needs (see discussion and references in Kondo-Brown, 2007).

Outside the traditional EOP/EAP distinction, other NAs have focused on survival English for immigrants (e.g., Nunan, 1990; Winn, 2005) and computer assisted language learning (e.g., Decamps & Bauvois, 2001; Gonzalez-Lloret, 2003).

NA for languages other than English

So far the discussion of NA appears to be all about English, in ESL settings. However, a number of NAs have been conducted in non-English-speaking countries for English as a foreign language (EFL) (e.g., Bacha, 2003; Kormos, Kontra, & Csolle, 2002; and Moreno, 2003 to cite three among many). Other NAs have been conducted for languages other than English. Some of these have been the multi-country efforts in Europe discussed above. Others have included a large number of languages by addressing foreign languages (e.g., Oukada, 2001; Porcher, 1983; and Purpura et al., 2003), or focused a bit more narrowly on three languages like Spanish, French, and German (Alalou, 2001; Alalou & Chamberlain, 1999; and Horwitz, 1988). NAs have also been conducted for specific languages like French (Dalgalian, 1983; Harlow, Smith, & Garfinkel, 1980; Licari, Londei, & Mandolini, 1983; Sapin-Lignieres, 1983), German (von der Handt, 1983), Japanese (Iwai et al., 1999), and Korean (Chaudron et al., 2005; Kim et al., 2003).

Decide upon approach(es) and syllabus(es)

Approaches

Brown (1995, p. 5) described approaches as "ways of defining what and how the students need to learn" and listed the following approaches:

1 Classical

2 Grammar-translation

3 Direct

4 Audiolingual

5 Communicative

Naturally, other belief systems that could be called approaches may exist (the cognitive approach comes immediately to mind).

Syllabuses

Curriculum design has often been discussed in terms of types of syllabuses. Wilkins (1976, pp. 2-14) distinguished between synthetic syllabuses ("organized in terms of tasks derived from the description of the language," p. 3) and analytic syllabuses ("organized in terms of the purposes for which people are learning language and the kinds of language performance that are necessary to meet those purposes," p. 13). White (1988, pp. 44-61) discussed two types of syllabuses from a somewhat different perspective: Type A syllabuses ("What is to be learnt?", p. 44) are distinguished from Type B syllabuses ("How is it to be learnt?", p. 44).

Brown (1995, pp. 6-14) defined syllabuses as "ways of organizing the course and materials" and listed the following seven:

1 Structural

2 Situational

3 Topical

4 Functional

5 Notional

6 Skills-based

7 Task-based

In recent years, other syllabuses have also gained prominence (lexical and problemsolving syllabuses come immediately to mind).

Given the above ways that teachers, students, and administrators may vary in their belief systems about what and how students should learn languages (approaches), it is not surprising that those same stakeholders may often disagree about how language courses and materials should be organized (syllabuses).

Recognize constraints

The constraints that are faced in NAs tend to be situation related rather than language related. For instance, in his article focusing on ESP communication constraints, Singh (1983) listed constraints, all of which focus on the situation and could impinge on the conduct of a NA: adequacy of syllabuses, administrative attitudes toward second/foreign language teaching and learning, availability of audio-video aids, class size, government language policies, availability of

time for NA, departmental organization, relevance of foreign/second language learning to other subjects, scheduling, society's general attitudes toward second/ foreign language learning, status of teachers, as well as teaching/learning traditions.

Singh's list provides a good starting point, but there are many other possible constraints, which will differ widely from situation to situation. Some may focus on the teachers in terms of their: training, qualifications, language proficiency, skills, expertise, experience, language, morale, motivation, styles, beliefs, and principles (list of factors from Richards, 2001, p. 99). More crucially, the teachers' willingness to cooperate in a NA project can determine whether it lives or dies, and their readiness to relinquish sovereignty over their classrooms could also be a factor, as could how satisfied they are with their working conditions, or how well they get along with their administrators, etc.

Consider also the overall effects of factors like money or religion. There are rare settings where money to support language education is plentiful as in Saudi Arabia - during my time working there (1982-5), I found the quality of the buildings, amount of classroom space, quality and amount of equipment, numbers of teachers, numbers and adequacy of textbooks were all excellent. Earlier, when I was teaching in China in (1980-2), all of the same factors were in short supply because money was in short supply. In constrast, the reverse was true with regard to religion. In Saudi Arabia, the students had the right to leave the classroom any time they wanted to pray. That is a constraint that was unthinkable where I was teaching in China.

Many other constraints may unexpectedly rear their heads in a particular NA. For example, the overall amount of time available for studying a language and the intensity of that time may be important constraints. Other constraints may have to do with the degree to which classrooms are homogeneous or heterogeneous with regard to their language background, age, academic status, proficiency level, etc. Such differences could clearly impose constraints on what a NA can find or accomplish.

Some scholars make a distinction labeled means analysis. According to Purpura and Graziano-King (2004, pp. 4-5), "a means analysis examines those factors that impede or facilitate curriculum implementation or change. A means analysis is not so much concerned with the language or the learner per se, but with the contextual variables of the learning/teaching environment (Jordan, 1997; Richterich, 1983)." Indeed, some researchers might consider the constraints listed above more relevant to means analysis than to needs analysis. However, as Dudley-Evans and St John (1998, p. 124) point out, "means analysis is suggested (Holliday and Cooke, 1982, p. 133) as an adjunct to needs analysis." I would go even further and say that needs analysts ignore such constraints at their peril.

In addition, given the large number of potential constraints, needs analysts may be tempted to just give up. However, most constraints can be overcome by simply setting realistic goals for the NA with an eye to attaining a "balance between 'what is needed' and 'what is possible'" (Singh, 1983, p. 156).

Select data collection procedures

According to Purpura et al. (2003, pp. 9-11), a NA should be situation specific, learner centered, pragmatic, and systematic. These characteristics are particularly important for planning the data collection procedures that will be used in a NA because those procedures must be appropriate for the specific situation, should be learner centered (though all other stakeholders must also be involved), must be practical (within the constraints imposed by the situation), and must be systematic in the sense that they will function well together and lead to defensible results.

Therefore, at some relatively early stage, needs analysts must carefully plan what procedures they will use for gathering their information. Stufflebeam et al. (1985, pp. 90-1) suggest five factors that affect choices of information-gathering procedures:

1 Characteristics of the information source

2 Situational characteristics

3 Type of information needed

4 Technical measurement criteria

5 Level of accuracy desired

Buckingham (1981) noted early on that "a great variety of assessment instruments and processes are available, and the use of more than one means of assessment is desirable" (p. 15). Hutchinson and Waters (1987) provide a few suggestions for the types of NA data gathering procedures: questionnaires, interviews, observations, data collection (e.g., gathering texts), informal consultations with sponsors, learners, etc. (p. 58). Jordan (1997) provided a bit longer list of data gathering procedures: "documentation, tests, questionnaires, forms/checklists, interviews, record-keeping and observation," with special emphasis on advance documentation, language test at home, language test on entry, self-assessment, observation and monitoring, class progress tests, surveys, structured interview, learner diaries, case studies, final tests, evaluation/feedback, follow-up investigations (pp. 30-8).

Perhaps the most complete lists of NA procedures are provided by Brown (1995, 2001) and Long (2005a), as shown in Table 16.2. In Brown (1995), I listed 26 different procedures and described each in turn (pp. 45-55). Long (2005a) provided what turns out to be a list that is in almost complementary distribution to my list. However, Long provided more information about each by describing each procedure, but also by citing examples of studies that used each procedure along with references for additional information about each. All the procedures in Table 16.2 taken together give a broad perspective on the many options available for gathering information in NA (including well over 40 different procedures). Clearly, trying to use all of them in any given NA would be impractical. So selecting among them will prove important. Naturally, needs analysts should select those procedures that best fit the purpose, scale, focus, approaches, syllabuses, and constraints of the particular NA.

I should make one last point about Table 16.2. Notice that, by far, the majority of the procedures listed in the table have an asterisk after them. These are the ones that lend themselves to qualitative research methods rather than quantitative methods. I will expand on this below.

Do the NA Research

Collect data

What questions should be addressed? In NA, we are in the business of gathering information, and gathering information inevitably leads to asking questions. I have found at the beginning of a NA that I often have no idea what the appropriate questions are, much less which of those questions are the most important. So the first task a needs analyst faces is often to figure out what types of questions are relevant.

I find it useful to refer to the question categories suggested by Patton (1987) and Rossett (1982) in trying to think of useful questions. These question types are all shown in Table 16.3 along with some ideas for what each question type might ask about.

The profile of communicative needs provided by Munby (1978) can also serve as nine aspects of communication about which needs analysts might want to ask questions:

Table 16.2 Procedures for NA

Listed in Brown (1995, 2001) Listed in Long (2005a)

Existing information

Intuitions

Records analysis* Systems analysis* Literature review* Email, letter writing, phone calls*

Aptitude

Proficiency

Placement

Diagnostic

Achievement

Task-based CR performance tests

Non-expert intuitions* Expert practitioner intuitions*

Table 16.2 (cont'd)

Listed in Brown (1995, 2001) Listed in Long (2005a)

Observations

Interviews

Meetings

Questionnaires

Case studies* Diary studies* Behavior observation* Interactional analysis Inventories*

Target language

Individual (in person, telephone, Internet)* Group*

Delphi technique* Advisory* Focus group* Interest group* Review*

Biodata surveys Opinion surveys (closed-response)

Opinion surveys (open-response)*

Self-ratings (closed-response) Self-ratings (open-response)* Judgmental ratings (closed-response)

Judgmental ratings (open-response)*

Q sort (closed-response)

Text analysis* Discourse analysis*

Diaries, journals, logs*

Participant observations* Non-participant observations* Classroom observations*

Unstructured interviews* Structured interviews* Interview schedules*

Surveys and questionnaires

Discourse analysis* Role plays, simulations* Content analysis* Register/rhetorical analysis* Computer-aided corpus analysis* Genre analysis*

Table 16.3 Ideas for types of questions to address in NA

Author Question type

Ask about

Patton Behaviors/experiences (1987)

Opinions/values Feelings Knowledge Sensory

Demographic/background

Rossett Problems (1982)

Priorities

Abilities

Attitudes

Solutions

Encounters or experiences in certain

language learning settings, what they do when those things happen, how they behave or act in those settings, etc. Thoughts, impressions, attitudes, values, opinions, etc. on various aspects of their language or language learning processes Sentiments and emotional reactions about particular topics, issues, and components of the learning and teaching processes Facts, information, and knowledge about the language learning and teaching processes in a particular context Visual, auditory, tactile, and/or olfactory aspects of the language learning and teaching processes Biographical, descriptive, or historical information that has bearing on the language learning teaching processes

Difficulties and problems participants

perceive in a particular language learning and teaching context Topics, functions, skills, activities, grammar points, etc. that participants feel are most important, second most important, third most important, etc. Language aptitudes, proficiencies of participants, especially with regard to reading, writing, speaking, listening, pragmatics, etc. Wants, wishes, and attitudes (e.g., toward the language being studied, toward native speakers of the language and their culture, toward course objectives, etc.) Answers or solutions to whatever problems or quandaries are uncovered in the above question types

1 Personal (important background information)

2 Purpose (reason for communicating)

3 Setting (place where communication happens)

4 Interaction variables (power, gender, psychological, etc. relationships)

5 Medium, mode, and channel (means of communicating)

6 Dialects (variations that will be encountered)

7 Target level (level of proficiency needed)

8 Anticipated communicative events (different levels of functions or tasks that will be required)

9 Key (particular way communication is accomplished)

How should data collection proceed?

This is a tricky part of any NA because needs analysts inevitably find themselves encroaching on the territory of other people. The first goal should therefore be to get all pertinent stakeholders to cooperate in the data-gathering process. Winn (2005, pp. 293-4) listed some useful strategies to use, if possible, when doing NA data collection:

1 Use inside connections in trying to gain entrée.

2 Be sure to go through channels in seeking permission in early stages.

3 Gradually get to know the participants.

4 Situate yourself as a learner, not an expert.

5 Observe a relatively large number of participants.

6 Use pre-interview conversations and follow-up interviews.

7 Use a variety of information sources to allow for comparisons of the data.

8 Be creative and flexible.

9 Volunteer and work in organizations where you are gathering data.

Analyze data

Quantitative or qualitative?

Research can vary from quantitative to qualitative on a number of dimensions: the quantative research approach tends to use quantitative (data) and be experimental, statistical, highly intervening, highly selective, variable operationalizing, hypothesis testing, deductive, controlled, cross-sectional, large sample, and etic, while the qualitative research approach tends to use qualitative (data) and be non-experimental, interpretive, non-intervening, non-selective, variable defining, hypothesis forming, inductive, natural, longitudinal, small sample, and emic (see Brown, 2004a for definitions of all this terminology). Clearly, these are at least two very different approaches that can be used in NA research.

Concerns with consistency, fidelity, verifiability, and meaningfulness Long (2005a) argued that "... all approaches to NA, new or old, could benefit from some serious work on issues of reliability and validity" (p. 22). I agree, but I would argue that we must first be clear about how we think and talk about

Table 16.4 Standards for judging of quantitative and qualitative research reports (summarized from Brown, 2004a)

Quantitative RESEARCH STANDARDS Qualitative

Reliability CONSISTENCY Dependability

Validity FIDELITY Credibility

Replicability VERIFIABILITY Confirmability

Generalizability MEANINGFULNESS Transferability

issues like reliability and validity. Table 16.4 (based on Brown, 2004a) shows the standards that are used for judging the worth of research, especially in quantitative and qualitative research reports. Note that all researchers must concern themselves with research standards related to consistency, fidelity, verifiability, and meaningfulness (shown down the middle of the table). However, quantitative researchers will do so by focusing on the concepts of reliability, validity, replicability, and generalizability, while qualitative researchers will zero in on dependability, credibility, confirmability, and transferability.

In discussing Table 16.2, I pointed out that the vast majority of the potential NA procedures lend themselves to qualitative rather than quantitative research methods. While much has been written about quantitative research methods in applied linguistics (see Brown, 2004b for a review of the available books on this topic), fewer resources are available specifically focused on doing qualitative research in second language settings. Four books offer at least a chapter-length treatment of qualitative language research methods (Brown, 2001; Freeman, 1998; Johnson, 1992; and Nunan, 1992), and a number of articles have appeared in the TESOL Quarterly on qualitative language research methods (e.g., Davis, 1992, 1995; Lazaraton, 1995).

In addition, some NAs have openly applied qualitative approaches (e.g., Holliday, 1995; Sawyer, 2001). Many other NAs have applied qualitative methods without realizing it. Indeed, most of the NAs cited earlier in this chapter took what I would call at least a predominantly qualitative research approach. Certainly, questionnaires with Likert scales were often involved, but careful scrutiny will reveal that most of the interpretations were actually qualitative in nature.

What I am suggesting here is that we should probably admit that NAs are usually at least predominantly qualitative in methodology, and that, therefore, we should frame our NAs, not in quantitative terms, but rather in qualitative terms. Notice that, while the qualitative terminology defined in Table 16.5 can be said to be analogous to the quantitative research concepts of reliability, validity, replicability, and generalizability, each is defined quite differently (see second column) from the quantitative analogs. Note also that the strategies used to enhance dependability, confirmability, credibility, and transferability (listed in the last column) are quite different from the strategies used to enhance the analogous concepts in quantitative research. All of this is discussed in much more depth elsewhere (e.g., Brown, 2001, 2004a; Davis, 1992, 1995).

Table 16.5 Key concepts in qualitative research methods (summarized from Brown, 2001)

Key Concepts in Qualitative Research Definition Enhanced by

Dependability Consistency of observations, effects of changing conditions in the objects of study, etc. to help better understand the context being studied Overlapping methods Stepwise replications Inquiry audits

Credibility Fidelity of identifying and describing the object(s) of study especially as judged by the various parties being studied Prolonged engagement Persistent observation Triangulation Peer debriefing Negative case analysis Referential analysis Member checking

Confirmability Verifiability of the data upon which all interpretations in a study are based. Audit trails Data records

Transferability Meaningfulness of the results of a study and their applicability in other settings Thick description

Triangulation

One concept that has been surfacing with increasing frequency in the NA literature is that of triangulation. Triangulation is not really a specific procedure, but rather is a research strategy that has been applied in a number of NAs (e.g., Bosher & Smalkoski, 2002; Gilabert, 2005; Jasso-Aiguilar, 1999, 2005; Kikuchi, 2004; Long, 2005a). As Long (2005a, p. 28) puts it, "Triangulation is a procedure used by researchers . . . to increase the credibility of their data and thereby, eventually, to increase the credibility of their interpretations of those data." He adds that it "... involves the researchers comparing different sets and sources of data with one another..." In Brown (2001, p. 229), I took a slightly larger view of triangulation when I identified seven potential types of triangulation: data triangulation (using multiple types of procdures), investigator triangulation (using multiple needs analysts), theory triangulation (using multiple conceptual frameworks), methodological triangulation (using multiple data gathering procedures), interdisciplinary triangulation (using the perspectives of multiple disciplines), time triangulation (using multiple data gathering occasions), and location triangulation (using multiple locations).

However, simply using multiple measures and triangulation does not guarantee that a qualitative NA will be dependable and credible. As Fielding and Fielding (1986, p. 31) put it, "using several different methods can actually increase the chance of error. We should recognize that the multi-operational approach implies a good deal more than merely piling on of instruments." Instead, the combinations of different perspectives, sources, data types, etc. must be carefully thought through and planned so that they cross-validate each other. As Huberman and Miles (1994, p. 438) suggested, "A general prescription has been to pick triangulation sources that have different biases, different strengths, so they can complement one another."

In addition, if appropriate, the combinations should be sequenced in such a way that each builds on what was learned from previous procedures. As Long (2005a, p. 33) put it, "In particular, carefully sequenced use of two or more procedures can be expected to produce better quality information."

For an example of how multiple procedures can be combined and sequenced effectively, see the "Participatory appraisal" NA reported in Holme and Chalauisaeng (in press), which used a number of procedures and combined them in such a way that the total information gathered was much greater in quality than the sum of the information collected with each individual procedure.

Interpret results

Given that NAs are generally qualitative in nature, what can needs analysts do to make sure their interpretations are sound? First, it is crucial that the NA data be gathered and analyzed in such a way that the interpretions will be seen as dependable, confirmable, credible, and transferable. Second, the interpretations should be done very carefully following the three suggestions I made in Brown (2001, p. 230) for minimizing researcher bias: (1) carefully arrange your triangulation so it minimizes the biases of different data sources and maximizes their strong points, (2) carefully examine how your preconceptions may be affecting your data choices, and (3) study how you may be attracted to unusual or salient data. Third, the results should be examined not only for how various data sources cross-validate each other, but also for any differences that occur among data sources. For instance, if there is a mismatch between the views of teachers and students on preferred classroom activities (as found in Spratt, 1999), that may be important information.

Use the NA Results

Determine objectives

One natural outcome for any language NA will be the specification of objectives for specific language courses. Put another way, objectives are the link that connects the NA to the rest of the curriculum (i.e., to the materials, testing, teaching, and program evaluation). Indeed, specifying objectives is a way of fitting what was learned in the NA to the actual instruction that will be delivered.

In addition, the act of fleshing out objectives based on the NA will help needs analysts begin to understand everything that is involved in meeting the students' perceived needs. Aspects of the students' needs that were overlooked during the NA will tend to surface at this point. Thus, specifying objectives is a natural and important next step in a NA.

A number of benefits can be gained from using objectives (for a list of ten such benefits, see Brown, 1995, p. 96), but there are also some potential pitfalls involved. To avoid these pitfalls, Brown (1995, pp. 96-7) provides six guidelines to remember:

1 Objectives can range in type and level of specificity.

2 Objectives are not permanent. They must remain flexible enough to respond to changes in perceptions of students' needs and changes in the types of students served.

3 Objectives must be developed by consensus among all of the teachers involved.

4 Objectives must not be prescriptive in terms of restricting what the teacher does in the classroom to enable students to perform well by the end of the course.

5 Because of all of the above, objectives will necessarily be specific to a particular program.

6 Above all else, the objectives must be designed to help the teachers, not hinder their already considerable efforts.

Evaluate and report on the NA project

What a report might contain

Overall, an effective NA report will explain the results with an eye to evaluating the effectiveness of the NA itself. Such a report might contain a description of the theoretical and practical context of the project. It should also include a clear description of the research methodology employed, including descriptions of the participants (administrators, teachers, students, future employers, etc.), materials (observation forms, interview schedules, questionnaires, tests, etc.), procedures (step-by-step, how was the NA information gathered?), and analyses (including explanation of the ways credibility, confirmability, and transferability were enhanced). Naturally, such a report would also usefully supply a discussion (of what the results of the NA mean in terms of the needs of the various participants in the program) and a set of conclusions drawn from the results and discussion. It is also useful to end such a NA report by supplying all the objectives that were derived from the NA.

Audiences and dissemination

While the report is being written, it is also vital to consider who the audiences will be. Will the audience for the report be just the teachers in the program? Or would it also be useful and politically wise to create shorter and simpler reports for students, administrators, higher authorities, politicians, or even for the general public?

Thinking about dissemination is also important. After all, if you have gone to all the trouble of doing a NA, you probably want key groups of people to know about it. Some options for disseminating NA results include oral presentations at faculty meetings, public meetings, and professional conferences, or written reports distributed through institutional curriculum documents, newsletters, journals, and books.

Conclusion

In this chapter, I have attempted to show the stages and steps that can be included in a sound NA. I hope that this discussion of the many alternatives available to needs analysts will help them do a better job in the future. However, in looking back over the chapter, I cannot help feeling that there is still something crucial missing.

I think the missing component has something to do with our overemphasis on language needs. The many articles about NA and the dozens of actual NA reports cited in this chapter tended to focus almost entirely on the language needs of the students. The problem with this approach is that it is not going far enough. It ignores the fact that language teachers, students, administrators, parents, etc. often have quite different views about what constitutes good language teaching and learning.

Once we accept that the various participants in a program may have quite differing views, it should be clear that we need to know what the various groups of stakeholders are thinking. NA is the perfect means for finding out what people are thinking and for exploring how the views of various groups are similar and different. That does not mean that everybody is right, nor does it mean that everybody should get what they think they want. It does mean that we need to know what everybody is thinking so we can deal politically with the similarities and differences in their views. Naturally, we must also use the best available knowledge and theory about language learning and teaching. If we can then combine all of these elements into a sound NA, we stand a good chance of creating a defensible curriculum.

In short, an obsessive focus on language needs that seeks the "truth" is probably destined to fail. A NA designed to investigate what bits of language the students "really need" to learn is doomed to collapse from the weight of its own single-minded focus. In contrast, a needs analysis that balances information about what options are available (approaches and syllabuses) with information about what various groups of stakeholders think about those options necessarily recognizes that there is no single truth, that NA is a political process, and therefore, that the goal of a NA is a defensible curriculum.

To actually do such a NA, I think we should indeed be gathering data about the students' language needs, but we should also be assembling information on the views of different groups of stakeholders and use that information to:

1 discover what the options (in perceived student needs with regard to approaches, syllabuses, objectives, etc.) are and what people think about those options;

2 decide which options are most likely to serve as a defensible basis for curriculum (i.e., options that might lead to a sort of average or consensus "truth");

3 marshall information and formulate arguments (sometimes alternative arguments) for the most viable options in perceived student needs with regard to approaches, syllabuses, objectives, etc.;

4 work to get all stakeholders to come to agreement, to form a consensus, or to at least compromise on those perceived needs;

5 work to accommodate the views of those who disagree with perceived needs, if possible;

6 try to change the views of those stakeholders who disagree with the perceived needs, when necessary;

7 show respect for all participants by listening and taking their views seriously (even if we ultimately have no intention of doing what they want).

Above all else, it is crucial to involve the teachers in all aspects of the NA processes. They are the people who will have to deliver the curriculum and live with it long after the current students (and perhaps the needs analysts) have moved on. More than any other group of stakeholders, the teachers will need to feel respected. To ignore the teachers in a NA is to doom the resulting curriculum to failure. To include the teachers is crucial because, in one way or another, any curriculum project will require them to make changes in their working habits, to do extra work, and, more importantly, to relinquish some portion of their classroom sovereignty. Needs analyses rightly focus on the needs of students, but we must never forget that teachers have needs, too.

1 Note that two of the steps in the last column of Table 16.1 are shown in brackets. I bracketed them and exclude them as NA steps because I view assessment, materials, teaching, and program evaluation as separate curriculum elements (see Brown, 1995).

REFERENCES

Alalou, A. (2001). Reevaluating curricular objectives using students' perceived needs: The

case of three language programs. Foreign Language Annals 34, 453-69. Alalou, A. & Chamberlain, E. (1999). Using student expectations and perceived needs to rethink pedagogy and curriculum: A case study. Foreign Language Annals 32, 2744.

Bacha, N. (2003). English across academic and professional communities: A study of EFL learners' needs at the Lebanese American University. AAICU Journal 2, 29 pp. Retrieved August 3, 2006 from: www.anatolia.edu.gr/act/aaicu2/pdf/nahla_bacha.pdf.

Berwick, R. (1989). Needs assessment in language programming: From theory to practice. In R. K. Johnson (ed.), The second language curriculum (pp. 48-62). New York: Cambridge University Press.

Bosher, S. & Smalkoski, K. (2002). From needs analysis to curriculum development: Designing a course in health-care communication for immigrant students in the USA. English for Specific Purposes 21, 1, 59-79.

Braine, G. (2001). Twenty years of needs analysis: Reflections on a personal journey. In J. Flowerdew & M. Peacock (eds.), Research perspectives on English for Academic Purposes (pp. 195-207). Cambridge: Cambridge University Press.

Brecht, R. D. & Rivers, W. P. (2005). Language needs analysis at the societal level. In M. H. Long (ed.), Second language needs analysis (pp. 79-104). Cambridge: Cambridge University Press.

Brindley, G. (1984). Needs analysis and objective setting in the Adult Migrant Education Program. Sydney: Adult Migrant Education Service.

Brown, J. D. (1995). The elements of language curriculum: A systematic approach to program development. Boston: Heinle and Heinle.

Brown, J. D. (2001). Using surveys in language programs. Cambridge: Cambridge University Press.

Brown, J. D. (2004a). Research methods for Applied Linguistics: Scope, characteristics, and standards. In A. Davies & C. Elder (eds.), The handbook of applied linguistics (pp. 476500). Oxford: Blackwell.

Brown, J. D. (2004b). Resources on quantitative/statistical research for applied linguistics. Second Language Research 20, 4, 408-29.

Buckingham, T. (1981). Needs assessment in ESL. Washington, DC: Center for Applied Linguistics.

Cameron, R. (1998). A language-focused needs analysis for ESL-speaking nursing students in class and clinic. Foreign Language Annals 31, 2, 203-18.

Casanave, C. & Hubbard, P. (1992). The writing assignments and writing problems of doctoral students: Faculty perceptions, pedagogical issues, and needed research. English for Specific Purposes 11, 33-49.

Chambers, F. (1980). A re-evaluation of needs analysis in ESP. ESP Journal 1, 1, 25-33.

Chan, V. (2001). Determining students' language needs in a tertiary setting. English Teaching Forum 39, 3, 16-27.

Chaudron, C., Doughty, C. J., Kim, Y., et al. (2005). A task-based needs analysis of a tertiary Korean as a foreign language program. In M. H. Long (ed.), Second language needs analysis (pp. 225-61). Cambridge: Cambridge University Press.

Chew, K.-S. (2005). An investigation of the English language skills used by new entrants in banks in Hong Kong. English for Specific Purposes 24, 4, 423-35.

Chia, H. U., Johnson, R., Chia, H. L., & Olive, R. (1999). English for college students in Taiwan: A study of perceptions of English needs in a medical context. English for Specific Purposes 18, 2, 107-19.

Clark, J. L. (1987). Curriculum renewal in school foreign language learning. Oxford: Oxford University Press.

Coleman, H. (1988). Analyzing language needs in large organizations. English for Specific Purposes 7, 155-69.

Conseil de la Coopération Culturelle (2000). Un cadre Européen commun de reference pour les langues: Apprendre, enseigner, évaluer. Strasbourg, France: Conseil de la Coopération

Culturelle, Comité de l'Éducation, Division des Langues Vivantes. Also viewed August 23, 2006 at www.coe.int/t/dg4/linguistic/Source/Framework_FR.pdf.

Council of Europe (2001). Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge: Cambridge University Press. Also viewed August 23, 2006 at: www.coe.int/t/dg4/linguistic/Source/Framework_EN.pdf.

Crosling, G. & Ward, I. (2002). Oral communication: The workplace needs and uses of business graduate employees. English for Specific Purposes 21, 1, 41-57.

Dalgalian, G. (1983). Identifying needs in a production-training project. In R. Richterich (ed.), Case studies in identifying language needs (pp. 117-26). Oxford: Pergamon.

Davis, K. A. (1992). Validity and reliability in qualitative research on second language acquisition and teaching: Another researcher comments. . . . TESOL Quarterly 26, 6058.

Davis, K. A. (1995). Qualitative theory and methods in applied linguistics research. TESOL Quarterly 29, 427-53.

Decamps, S. & Bauvois, C. (2001). A method of computer-assisted language learning: The elaboration of a tool designed for an "un-schoolable" public. Computer Assisted Language Learning 14, 1, 69-96.

Dovey, T. (in press). What purposes, specifically? Re-thinking purposes and specificity in the context of the "new vocationalism." English for Specific Purposes.

Downey Bartlett, N. J. (2005). A double shot 2% mocha latte, please, with whip: Service encounters in two coffee shops and at a coffee cart. In M. H. Long (ed.), Second language needs analysis (pp. 305-43). Cambridge: Cambridge University Press.

Dubin, F. & Olshtain, E. (1986). Course design. Cambridge: Cambridge University Press.

Dudley-Evans, T. & St John, M. J. (1998). Developments in English for Specific Purposes: A multidisciplinary approach. Cambridge: Cambridge University Press.

Edwards, N. (2000). Language for business: Effective needs assessment, syllabus design and materials preparation in a practical ESP case study. English for Specific Purposes 19, 291-6.

Ferris, D. (1998). Students' views of academic aural/oral skills: A comparative needs analysis. TESOL Quarterly 32, 289-318.

Ferris, D. & Tagg, T. (1996a). Academic listening/speaking tasks for ESL students: Problems, suggestions, and implications. TESOL Quarterly 30, 2, 297-320.

Ferris, D. & Tagg, T. (1996b). Academic oral communication needs of EAP learners: What subject-matter instructors actually require. TESOL Quarterly 30, 1, 31-58.

Fielding, N. G. & Fielding, J. L. (1986). Linking data. Beverly Hills, CA: Sage.

Freeman, D. (1998). Doing teacher research: From inquiry to understanding. Boston, MA: Heinle & Heinle.

Gilabert, R. (2005). Evaluating the use of multiple sources and methods in needs analysis: A case study of journalists in the Autonomous Community of Catalonia (Spain). In M. H. Long (ed.), Second language needs analysis (pp. 182-99). Cambridge: Cambridge University Press.

Gonzâlez-Lloret, M. (2003). Designing task-based CALL to promote interaction: En busca de esperaldas. Language Learning and Technology 7, 1, 86-104.

Graves, K. (2000). Designing language courses: A guide for teachers. Boston, MA: Newbury House.

Hale, G., Taylor, C., Bridgeman, B., Carson, J., Kroll, B., & Kantor, R. (1996). A study of writing tasks assigned in academic degree programs. Princeton, NJ: Educational Testing Service.

Harlow, L. L., Smith, W. F., & Garfinkel, A. (1980). Student-perceived communication needs: Infrastructure of the functional/notional syllabus. Foreign Language Annals 13, 11-22.

Holliday, A. (1995). Assessing Language needs within an institutional context: An ethnographic approach. English for Specific Purposes 14, 2, 115-26.

Holliday, A. & Cooke, T. (1982). An ecological approach to ESP. Lancaster Practical Papers in English Language Education (Issues in ESP) 5, 123-43.

Holme, R. & Chalauisaeng, B. (in press). The learner as needs analyst: The use of participatory appraisal in the EAP reading classroom. English for Specific Purposes.

Holmes, J. (2005). When small talk is a big deal: Sociolinguistics challenges in the workplace. In M. H. Long (ed.), Second language needs analysis (pp. 344-71). Cambridge: Cambridge University Press.

Horwitz, E. K. (1988). The beliefs about language learning of beginning university foreign language students. The Modern Language Journal 72, 283-94.

Huberman, A. M. & Miles, M. B. (1994). Data management and analysis methods. In N. K. Denzin & Y. S. Lincoln (eds.), Handbook of qualitative research (pp. 428-44). Thousand Oaks, CA: Sage.

Hutchinson, T. & Waters, A. (1987). English for Specific Purposes. Cambridge: Cambridge University Press.

Hyland, K. (2002). Specificity revisited: How far should we go now? English for Specific Purposes 21, 385-95.

Iwai, T., Kondo, K., Lim, D. S. J., Ray, G., Shimizu, H., & Brown, J. D. (1999). Japanese language needs assessment 1998-1999 (NFLRC NetWork #13) [HTML document]. Honolulu: University of Hawai'i, Second Language Teaching and Curriculum Center. Retrieved from: www.lll.hawaii.edu/nflrc/NetWorks/NW13.

Jasso-Aguilar, R. (1999). Sources, methods, and triangulation in needs analysis: A critical perspective in a case study of Waikiki hotel maids. English for Specific Purposes 18, 1, 27-46.

Jasso-Aguilar, R. (2005). Sources, methods, and triangulation in needs analysis: A critical perspective in a case study of Waikiki hotel maids. In M. H. Long (ed.), Second language needs analysis (pp. 127-58). Cambridge: Cambridge University Press.

Johnson, D. M. (1992). Approaches to research in second language learning. New York: Longman.

Jordan, R. R. (1997). English for Academic Purposes. Cambridge: Cambridge University Press.

Kellerman, E., Koonen, H., & van der Haagen, M. (2005). "Feet speak louder than the tongue": A preliminary analysis of language provisions for foreign professional footballers in the Netherlands. In M. H. Long (ed.), Second language needs analysis (pp. 20022). Cambridge: Cambridge University Press.

Kikuchi, K. (2004). Triangulating perceptions of learners' needs: An alternate way of conducting needs analysis. 2004 JALT Pan-SIG Proceedings. Tokyo: JALT.

Kim, S. (in press). Academic oral communication needs of East Asian international graduate students in non-science and non-engineering fields. English for Specific Purposes.

Kim, Y., Kong, D.-K., Lee, Y.-G., Silva, A., & Urano, K. (2003). A task-based needs analysis for the English Language Institute at the University of Hawaii at Manoa. Korean Journal of Applied Linguistics 19, 2, 93-114.

Kondo-Brown, K. (2007). Issues and future agendas for teaching Chinese, Japanese, and Korean heritage students. In K. Kondo-Brown & J. D. Brown (eds.), Teaching Chinese, Japanese and Korean heritage language students: Curriculum needs, materials, and assessment (pp. 17-43). Mahwah, NJ: Lawrence Erlbaum.

Kormos, J., Kontra, E. H., & Csolle, A. ( 2002). Language wants of English majors in a non-native context. System 30, 517-42.

Lazaraton, A. (1995). Qualitative research in applied linguistics: A progress report. TESOL Quarterly 29, 3, 455-72.

Leki, I. (1995). Coping strategies of ESL students in writing tasks across the curriculum. TESOL Quarterly 29, 2, 235-60.

Lepetit, D. & Cichocki, W. (2002). Teaching languages to future health professionals: A needs assessment study. Modern Language Journal 86, 3, 384-96.

Lett, J. A. (2005). Foreign language needs assessment in the US military. In M. H. Long (ed.), Second language needs analysis (pp. 105-24). Cambridge: Cambridge University Press.

Licari, C., Londei, D., & Mandolini, M. (1983). Notes on a questionnaire for identifying the language needs of French in Bologna. In R. Richterich (ed.), Case studies in identifying language needs (pp. 88-96). Oxford: Pergamon.

Long, M. H. (2005a). Methodological issues in learner needs analysis. In M. H. Long (ed.), Second language needs analysis (pp. 19-76). Cambridge: Cambridge University Press.

Long, M. H. (2005b). Second language needs analysis. Cambridge: Cambridge University Press.

Moreno, A. I. (2003). Análisis de necesidades para el aula de lengua inglesa en filología inglesa: Un estudio de caso. Barcelona English Language and Literature Studies 12 (no page numbers). Accessed August 25, 2006 from: www.publicacions.ub.es/revistes/ bells12/PDF/art10.pdf.

Munby, J. (1978). Communicative syllabus design: A sociolinguistic model for defining the content of purpose-specific language programs. Cambridge: Cambridge University Press.

Nunan, D. (1985). Language teaching course design: Trends and issues. Adelaide: National Curriculum Resource Centre.

Nunan, D. (1988). The learner-centered curriculum. Cambridge: Cambridge University Press.

Nunan, D. (1990). Using learner data in curriculum development. English for Specific Purposes 9, 17-32.

Nunan, D. (1992). Research methods in language learning. Cambridge: Cambridge University Press.

Oukada, L. (2001). Toward responsive beginning language curricula. Foreign Language Annals 34, 107-17.

Patton, M. Q. (1987). How to use qualitative methods in evaluation. Newbury Park, CA: Sage.

Porcher, L. (1983). Reflections on language needs in the school. In R. Richterich (ed.), Case studies in identifying language needs (pp. 127-49). Oxford: Pergamon.

Purpura, J. E. & Graziano-King, J. (2004). Investigating the foreign language needs of professional school students in international affairs: A case study. Teachers College, Columbia University Working Papers in TESOL and Applied Linguistics 4, 1, 1-33.

Purpura, J. E., Graziano-King, J., Chang, J., et al. (2003). An analysis of the foreign language needs of SIPA students at Columbia University. Unpublished Technical Report submitted to the Mellon Foundation through the Arts and Sciences at Columbia University, New York.

Richards, J. C. (2001). Curriculum development in language teaching. Cambridge: Cambridge University Press.

Richterich, R. (ed.) (1983). Case studies in identifying language needs. Oxford: Pergamon.

Richterich, R. (1985). Besoins langagiers et objectifs d'apprentissage. Paris: Hachette.

Richterich, R. & Chancerel, J.-L. (1977). L'Identication des besoins des adultes apprenant une langue étrangère. Strasbourg: Conseil de l'Europe.

Richterich, R. & Chancerel, J.-L. (1987). Idenifying the needs of adults learning a foreign language. Englewood Cliffs, NJ: Prentice-Hall International.

Rossett, A. (1982). A typology for generating needs assessments. Journal of Instructional Development 6, 1, 28-33.

Sapin-Lignieres, B. (1983). A method of research into needs applied to teachers of French in northern Greece. In R. Richterich (ed.), Case studies in identifying language needs (pp. 97-105). Oxford: Pergamon.

Sawyer, R. (2001). An ethnographic approach to needs analysis for international graduate students of science in Japan. In E. F. Churchill & J. W. McLaughlin (eds.), Temple University Japan Working Papers in Applied Linguistics: Qualitative research in applied linguistics: Japanese learners and contexts (pp. 102-17). Tokyo: Temple University Japan.

Schutz, N. W. & Derwing, B. L. (1981). The problem of needs assessment in English for specific purposes: Some theoretical and practical considerations. In R. Mackay & J. D. Palmer (eds.), Languages for specific purposes: Program design and evaluation (pp. 40-5). Rowley, MA: Newbury House.

Seedhouse, P. (1995). Needs analysis and the general English classroom. ELTJournal 49, 1, 59-65.

Singh, R. K. (1983). ESP: Communication constraints. System 11, 2, 155-8.

So-mui, F. L. & Mead, K. (2000). An analysis of English in the workplace: The communication needs of textile and clothing merchandisers. English for Specific Purposes 19, 35167.

Spratt, M. (1999). How good are we at knowing what learners like? System 27, 141-55.

Stufflebeam, D. L., McCormick, C. H., Brinkerhoff, R. O., & Nelson, C. O. (1985). Conducting educational needs assessments. Hingham, MA: Kluwer-Nijhoff.

Tanaka, H. (2001). English communication needs of an international advertising company: An ethnographic appraisal. In E. F. Churchill & J. W. McLaughlin (eds.), Temple University Japan Working Papers in Applied Linguistics: Qualitative research in applied linguistics: Japanese learners and contexts (pp. 143-63). Tokyo: Temple University Japan.

Teng, H. (1999). Needs analysis of EFL listening by Taiwanese college students. KOTESOL Proceedings of PAC2 (pp. 169-77). The Second Pan Asian Conference, Seoul, Korea.

Trim, J. L. M. (1973). Système d'apprentissage des langues vivantes par les adultes: Un système Européen d'unités capitalisables. Strasbourg: Conseil de la Coopération Culturelle du Conseil Européen.

Trim, J. L. M. (1980). The place of needs analysis in the Council of Europe Modern Language Project: Foreign language teaching: Meeting individual needs. Oxford: Pergamon.

Uvin, J. (1996). Designing workplace ESOL courses for Chinese health-care workers at a Boston nursing home. In K. Graves (ed.), Teachers as course developers (pp. 39-62). Cambridge: Cambridge University Press.

van Hest, E. & Oud-De Glas, M. (1990). A survey of techniques used in the diagnosis and analysis of foreign language needs in industry. Brussels: Lingua.

Vandermeeren, S. (2005). Foreign language needs of business firms. In M. H. Long (ed.), Second language needs analysis (pp. 159-81). Cambridge: Cambridge University Press.

von der Handt, G. (1983). Needs identification and curricula with particular reference to German for migrant workers. In R. Richterich (ed.), Case studies in identifying language needs (pp. 24-38). Oxford: Pergamon.

Waters, A. (1996). A review of research into needs in English for Academic Purposes of relevance to the North American higher education context. TOEFL monograph series #6. Princeton, NJ: Educational Testing Service.

West, R. (1994). Needs analysis in language teaching. Language Teaching 27, 1-19.

West, R. (1997). Needs analysis: State of the art. In R. Howard & G. Brown (eds.), Teacher education for LSP (pp. 68-79). Clevedon, UK: Multilingual Matters.

White, R. V. (1988). The ELT curriculum: Design, innovation and management. Oxford: Blackwell.

Widdowson, H. G. (1981). English for specific purposes: Criteria for course design. In M. Selinker, E. Tarone, & V. Hanzeli (eds.), English for Academic Purposes: Studies in Honor of Louis Trimble (pp. 1-11). Rowley, MA: Newbury House. Wilkins, D. (1976). Notional syllabuses. Oxford: Oxford University Press. Winn, M. (2005). Collecting target discourse: The case of the US naturalization interview. In M. H. Long (ed.), Second language needs analysis (pp. 265-304). Cambridge: Cambridge University Pres. Yalden, J. (1987). Principles of course design for language teaching. Cambridge: Cambridge University Press.

17 Syllabus Design

PETER ROBINSON

Introduction: Two Basic Distinctions

Units and sequence

Syllabus design is based essentially on a decision about the "units" of classroom activity, and the "sequence" in which they are to be performed. The syllabus thus formalizes the content to be learned in a domain of knowledge or behavior, and "arranges this content in a succession of interim objectives" (Widdowson, 1990, p. 127). As in other areas of instruction (see Reigeluth, 1999) there are options in, and differing theoretical rationales for, the units to be adopted in specifying and sequencing pedagogic content for second language (L2) learners, and a number of these will be described in this chapter. Theoretical rationales, of course, should be concerned with issues of how the L2 is internalized and learned, and also accessed and acted upon, since it is the cognitive processes leading to learning and successful performance, as they take place in specific pedagogic contexts, that the syllabus is intended to promote. Individual differences between learners in the cognitive and other abilities contributing to their "aptitudes" for learning and performing in the L2 will also modulate, and contribute to variance in, the effectiveness of specific pedagogic contexts and sequencing decisions at the group or program level (Robinson, 2002, 2005a, 2007a). These are theoretical and empirical issues for research into instructed second language acquisition (SLA) to address, in order to establish an optimally effective, learner-sensitive approach to syllabus design. Experimental and classroom-based research into a number of psycholinguistic issues in instructed SLA has begun with this prospect in view.

Perhaps the most fundamental issue for syllabus design addressed by this research so far is the following: Is the L2 best learned explicitly, by understanding and practicing a series of formal units of language, however characterized, or is it best learned incidentally from exposure to the L2 during communicative activities and tasks (see, e.g., Doughty, 2001; Doughty & Williams, 1998; N. Ellis, 2005, this volume; Ortega, this volume; Robinson, 1996a, 1997, 2001)? Commitment to

The Handbook of Language Teaching Editedby MichaelH.Longand CatherineJ.Doughty © 2009 Bkckwell Publishing Ltd. ISBN: 978-1-405-15489-5

one or the other of these broad pedagogic orientations and psycholinguistic positions underlies a number of proposals that have been made for units of syllabus design. Units have been based on an analysis of the language to be learned, in terms of a series of grammatical structures, graded in difficulty, as in Ellis (1993, 1997), or of lexical items and collocations, graded in frequency, as in Willis (1990). Units have also been based on an analysis of the components of simple versus complex skilled behavior in the second language, e.g., the reading microskills described by Richards (1990) or the communicative skills forming part of Munby's (1978) communicative needs profiler, and Johnson's (1996) work. Units have also been based on observed real-world performative acts involving the L2, or "target tasks" for a population of learners identified during a needs analysis (see Brown, this volume). Target tasks involve varied real-world activities, such as greeting passengers and serving meals on an airplane (Long, 1985, 2005) or finding a journal article in a library with the aid of library technology and then using it to find needed answers to questions (Robinson & Ross, 1996). Target task L2 performance is gradually approximated during classroom performance of simpler pedagogic versions of these tasks. Examples of these, and other units, that have been proposed will be given below.

Along with choices in the units to be adopted, there are choices in the "sequence" in which they can be presented. Some sequencing criteria are specific to a particular unit, or units, as will be described below - such as more to less "frequent" in the case of lexical items. However, the relative merits of broader sequencing options, and educational philosophies supporting them, have been discussed in the literature on L2 syllabus design. A syllabus can consist of a prospective and fixed decision about what to teach, and in what order. In this case the syllabus will be a definition of the contents of classroom activity. This is largely the approach to syllabus design discussed in this chapter. However, sequencing decision can also be made online, during classroom activity as in the "process" syllabus (Breen, 1984) or the "negotiated" syllabus (Clarke, 1991). In this case the initial syllabus will only guide, but not constrain, the classroom activities. Finally, Candlin (1984) has proposed that a syllabus can be retrospective, in which case no syllabus will emerge until after the course of instruction. In this case, the syllabus functions only as a record of what was done, imposing no controlling constraint on the classroom negotiation of content.

The role of the learner in approaches to syllabus design

Another distinction which is useful in conceptualizing options in syllabus design was made initially by Wilkins (1976) and refers to the learner's role in assimilating the content provided during group instruction and applying it individually to real-world language performance and interlanguage development. Synthetic syllabi involve a focus on specific elements of the language system (such as grammatical structures, or language functions), often serially and in a linear sequence. The easiest, most learnable, most frequent, or most communicatively important (sequencing decisions can be based on each of these often non-complementary

criteria, and on others) are presented before their harder, later learned, less frequent, and more communicatively redundant counterparts. These syllabi assume the learner will be able to put together, or synthesize in real-world performance, the parts of the language system they have been exposed to separately. In contrast, analytic syllabi do not divide up the language to be presented in classrooms, but involve holistic use of language to perform communicative activities. The learner's role in these syllabi is to analyze or attend to aspects of language use and structure as the communicative activities require them to, in line with: (1) their developing interlanguage systems; (2) their preferred learning style and aptitude profile; and (3) to the extent that they are motivated to develop to an accuracy level which may not be required by the communicative demands of the task. For these reasons analytic approaches to syllabus design have been argued to be more sensitive to SLA processes and learner variables than their synthetic counterparts. This distinction will also be related to the following description of syllabi that have been proposed.

Traditional Approaches to Syllabus Design

Grammatical syllabi

Since the 1920s, and the work of Harold Palmer and others in the Reform Movement (see Howatt & Widdowson, 2004) who emphasized the controlled presentation of grammatical structures and oral practice following classroom presentation, grammatical syllabi have featured prominently in programs for second and foreign language learners. Intuitive criteria of relative usefulness, and simplicity were used as a basis of selection and grading: "The grammatical material must be graded. Certain moods and tenses are more useful than others; let us therefore concentrate on the useful ones first . . . we will not learn off the whole set of prepositions, their uses and requirements, but we will select them in accordance with their degree of importance" (Palmer, 1922, p. 68). These intuitive criteria, reflected in the decisions about selection and grading in basic structure lists for language teaching (e.g., Hornby, 1959) continue to be influential (see the discussion of R. Ellis, 1993, 1997 below). While SLA research into learnability, teachability, and developmental sequences (e.g., Pienemann, 1989) has had little influence on grading, more recent developments in corpus analysis (see Flowerdew, this volume; Gries, 2008; Sinclair, 2004) which identify central patterns of use in specific discourse domains attempt to put at least the "useful" criterion on an empirical footing. SLA research has shown that the additive "accumulation" of increasingly complex and accurate grammatical structures in a linear sequence is not what happens during second language development, but this is what a structural syllabus would seem to predicate as evidence of learning in classrooms that employ it. Nonetheless, Wilkins comments; "The use of a grammatical syllabus can be regarded as the conventional approach to language teaching since the majority of syllabi and published courses have as their core an

ordered list of grammatical structures" (1976, p. 7). A concern to develop syllabi which attempted to at least supplement structural criteria for grading teaching sequences by emphasizing the "uses" to which structures could be put during communicative activities led to alternative proposals in the 1970s that also continue to be influential.

Notional-functional syllabi

In the 1970s, the Council of Europe initiated a project (see Richterich, 1972; Van Ek, 1976; Wilkins, 1976) which aimed to specify a common framework for teaching and assessing "communicative competence" in foreign language education. The aim was to specify syllabi in terms of three categories of meaning common across languages: semantico-grammatical meaning, or notions, such "time" and "quantity"; modal meanings, such as degree of certainty and scale of commitment; and communicative functions, such as agreeing, requesting, complimenting. These provided a way of going from specified types of meaning, or universal communicative and conceptual categories, to their realization in specific languages. This is an unordered inventory: what provides sequencing constraints on these notional categories, and their realizations in language, is the idea of a common core of notions/functions, useful for all communicative goals and purposes, which must be mastered before those particular to specific communicative purposes. Brockett (2000) is a contemporary example of the use of this approach to syllabus design which aims to sets standards for Japanese instruction across various state and private sector institutions in the United States. Core communicative functions are grouped into superordinate categories, which include, socializing, getting things done, communicating factual information. These are themselves subcategorized and the functions matched to essential patterns and phrases. For example, "leave taking" within the superordinate "socializing" category, is specified in terms of formal parting (ja, shitsurei shimasu); at night (oyasumi nasai); and inviting to come again (mata kite kudasai). Concepts or notions are also listed in this way, so the notional category of "existential concepts" is divided into subcategories, such as "possibility and impossibility," and matched with an essential pattern such as . . . koto ga dekiru. At a level beyond mastery of the core functions and notions, lesson content or topics appropriate to learners with specific occupational needs (such as nurses) also provide a constraint on which further functions/notions, and their linguistic exponents are to be taught, and these too can be sequenced using criteria such as going from concrete to more abstract (e.g., from "giving a patient an injection," to " taking care of the elderly," to "illness"), or more to less common, or more to less useful, however defined. The Notional-Functional Syllabus, then, is little more than an inventory of notions and functions to be covered at different levels of a language teaching program. As many have noted (e.g., White, 1988, p. 76), functions can be realized with any number of forms or structures (How about going to see/Would you like to see/ Have you seen the new Woody Allen film?). Consequently, Brockett comments, "It is probable that curricula can be most successfully organized around the

principle of introducing students to structural patterns in the context of their communicative functions, and concepts within the specific topic areas" (2000, p. 19). Grammatical criteria for grading and sequencing the formal exponents of functions are therefore as important to this approach to syllabus design as they are in the structural syllabus described above.

A further problem with the functional syllabus is that, in almost all cases, the functional value of an utterance in extended discourse is a binary relation between two contributions (Widdowson, 1978). Simply listing grammatical exponents of functions misrepresents the fact that B's contribution in the following has a different speech act value depending on A's contribution:

A: Where are you?/John is on the phone.

B: I'm busy on the computer.

Crombie (1985a) is an exhaustive attempt to list all the possible inter- and intra-propositional semantic relations, and exemplify how they can be linguistically coded and signaled in English. The general semantic relations described in the "Relational Approach" to syllabus design (Crombie, 1985b) include temporal relations (e.g., chronological sequence); matching relations (e.g., simple comparison/ contrast); cause-effect relations (reason-result/means-result), and so on. Once again, however, while such inter- and intra-propositional relations can be called "units," it is not clear on what principled basis they are to be sequenced for presentation. Further, the taxonomy given is not definitive, but rather "one which I hope will prove useful in the design of language teaching programs" (Crombie, 1985a, p. 17).

Contemporary Rationales for Syllabus Design

The structural syllabus

Like other contemporary rationales for syllabus design, Ellis (1993, 1997) draws extensively on SLA research and theory to motivate his arguments for a role for a structural syllabus. Ellis' argument rests on two distinctions: between explicit conscious knowledge and implicit tacit knowledge, and between declarative knowledge of facts and procedural knowledge of how to do things (see Anderson, 1992; DeKeyser, 2001; Robinson, 1996a). He argues that explicit, declarative knowledge of L2 grammar can influence the development of implicit declarative knowledge, and through communicative activity, implicit declarative knowledge can be proceduralized and used in spontaneous skilled performance. This is a "weak interface" model, which allows explicit knowledge, under some conditions, to influence the development of tacit representations or competence. The main condition is that the learner must be developmentally ready to incorporate the explicit grammar instruction into their interlanguage. Ellis cites research by himself (1989), Pienemann (1989) and others showing that learners pass through

stages of development in the acquisition of, amongst other things, word order rules, question forms, and negation. Unless grammatical instruction is timed to the learner's point of development it will not influence the developing implicit knowledge base. Since stages of development are learner internal and hidden from the teacher, timing is difficult to manage. However, Ellis argues explicit grammatical knowledge serves a number of other functions: it can be used to monitor production; it can help learners notice features in the input; and it can help learners compare their own production with a target model, and in some cases notice the gap between between them. Knowing about grammar, Ellis argues, is therefore useful. Tasks promote consciousness-raising, and noticing of target grammar rules. Tasks are therefore pedagogic devices for teaching units of grammar (examples are described in Ellis, 2003; Ellis & Noboyushi, 1993; Fotos & Ellis, 1991), and are used to implement a prospective synthetic structural syllabus.

The lexical syllabus

Drawing on a different type of empirical evidence - large-scale corpora of spoken and written language use - Willis also argues for a synthetic syllabus, where word and collocation are the units of analysis. Willis nowhere draws on SLA research to the extent Ellis does to motivate his proposal, but does conclude that SLA research findings show "input does not equal intake" and that "the assumption that language can be broken down into a series of patterns which can then be presented to learners and assimilated by them in a predictable sequence" is wrong (Willis, 1990, p. iii). Arguing against "a methodology which presents learners with a series of patterns" in a presentation, practice, production sequence, Willis proposes taking "meaningful exposure as a starting point" (1990, p. iv). Exposure should be organized in three ways: (1) language is graded in difficulty; (2) language exemplifying the commonest patterns is selected; and (3) the language syllabus is itemized to highlight important features. Exposure is thus tightly controlled. Rather than linguistically grading the content of the syllabus, Willis argues for lexically grading it, using corpora of language use to identify word frequency at the 700-word, the 1,500-word, and the 2,500-word levels. Words in the corpora are itemized as collocations exemplifying each word's typical patterns of use. In effect, though, lexical grading leads to linguistic grading since, as Willis notes, by identifying the commonest words, "inevitably it focuses on the commonest patterns too . . . the lexical syllabus not only subsumes a structural syllabus, it also indicates how the structures which make up the syllabus should be identified" (1990, p. vi). In the lexical syllabus these three corpora are the bases of exposure at three levels of learner development. Willis claims that exposure is not sequenced or controlled within these levels, and the lexical syllabus "does not dictate what will be learned and in what order"; rather "it offers the learner experience of a tiny but balanced corpus from which it is possible to make generalizations about the language as a whole" (1990, p. vii). In other words, the learner corpus which forms the basis of exposure at each level is carefully itemized, but these items are not presented individually and serially.

Willis describes the development of the COBUILD Course (an exemplar of the lexical syllabus) as a process of first intuitively deciding on interesting topics, then developing tasks and choosing texts to complement them, and then highlighting lexical items within, e.g., the first 700-word level, as they occurred in the texts. This series of highlighted items constitutes syllabus content, but items are sequenced according to no criteria that are discussed, apart from teacher intuition (see Willis, 1990, pp. 74-90). The methodology accompanying the syllabus (described in Willis, 1990, and in more detail by Jane Willis, 1996) involves a pre-task introduction to a topic, and exposure to texts; a task cycle where a task is planned, drafted and rehearsed; and a final language focus where learners consciously focus on forms used during the task. Course planning and content, hence the syllabus, is thus largely determined by the choices of texts and tasks -topics about which the lexical syllabus says nothing. This is, then, a language-focused synthetic syllabus, but with some control given to the learner about which forms to attend to and focus on, since the itemized corpora at each level function as a guide, rather than as a prospective plan, allowing more online negotiation of content than Ellis allows.

The skill syllabus

Drawing on the work of Anderson (1992) and the declarative, procedural distinction referred to by Ellis (1997), Johnson argues that SLA and general skill learning draw on the same general cognitive mechanisms. Traditionally, skill acquisition has been viewed as a speed-up in the use of initially attention-demanding declarative knowledge. With practice, attentional demands diminish and declarative knowledge is proceduralized. Johnson argues that many aspects of L2 learning can be viewed as the reverse process, from initially fast, unattended and un-analyzed use, drawing on procedural knowledge alone, to declarative knowledge. This occurs when formulaic language is used fluently at first, without any knowledge of its internal structure. As this becomes attended to and analyzed, declarative knowledge emerges. Declarative knowledge is valuable because it allows greater generalizability of language use, and is not context-dependent, in contrast to procedural knowledge. Johnson concludes that his proposals support a skills syllabus, similar to, but going beyond, the earlier attempts of Munby (1978) and Wilkins (1976) to specify the units of communicative syllabus design. In essence, Johnson proposes a four-tier model of syllabus design. Occupying the first tier are what Johnson, following Munby, calls language-specific skills, such as "identifying the present perfect," or correctly contrasting /i/ and /i:/: "In our attempts to break language behavior down into subskills, the general areas of phonetics/phonology and syntax would, then, follow traditional lines and would not pose any new difficulties for syllabus designers" (Johnson, 1996, p. 164). But the old difficulties are surely difficulties enough. Are separate subskills to be identified for each phonetic contrast, for example? And how does "learning difficulty" affect decisions about selecting and sequencing subskills? Another tier would contain semantic categories, such as notions and functions, "but only

those about which pedagogically accessible generalizations can be made" (1996, p. 165), that is, notions and functions which can be generalized to many contexts. An example given is inviting versus being polite. Johnson claims inviting need not be taught, and so need not be part of the syllabus, since it is largely phrasal and situation-specific, whereas in being polite, "useful generalizations . . . can be made about such things as 'being circumspect and indirect in approach'" (1996, pp. 165-6). A third tier would involve skills often referred to in "process" approaches to teaching writing skills, such as generating new ideas, drafting essays, structuring and evaluating them. It seems then that skill is being used as a term to cover three different types of unit: language item, semantic category, and writing strategy. This is because Johnson is concerned with the transition from knowledge states - procedural to declarative and vice versa - that learning all these units has in common. The fourth and final tier of Johnson's skills syllabus concerns processing demands; the level of complexity of the classroom task should also be specified and enter into sequencing decisions. In summary, Johnson also favors a synthetic syllabus, prospectively organized, based on subskills at a number of levels, linguistic, semantic and pragmatic, and strategic. The role of the syllabus designer is to draw up an inventory of the subskills at each of these levels (as Munby, 1978 attempted to do), then sequence them, and weave them together in a principled way.

The task syllabus

While in practice still not a common choice of unit, tasks have been increasingly researched and theorized as a basis for syllabus design in recent years (see Bygate, Skehan, & Swain, 2001; R. Ellis, 2003, 2005; Long, 1985, 2007; Garcia-Mayo, 2007; Nunan, 2004; Robinson, 1995, 1996b, 2001, 2005b, 2007b; Robinson, Ting, & Urwin, 1995; Skehan & Foster, 2001). Target tasks (see Long & Crookes, 1993) are units of real-world activity involving language use identified on the basis of a needs analysis (see Long, 2005), subsequently broken down into simpler versions, which are presented in order of increasing complexity, so as eventually to approximate the full complexity of the target task demands. In this view, the features of tasks contributing to their relative complexity are the basis of sequencing decisions. However, in many discussions of tasks, and examples of what claim to be task-based materials, tasks are used to force attention to, or to practice a particular structure, function, or subskill. Skehan (1998) refers to these as "structure-trapping" tasks. These include the tasks advocated by Ellis (1997, 2003), and Loschky and Bley-Vroman (1993), where the use of tasks to direct attention to grammatical form is theoretically motivated and an explicit part of the rationale for their use, as well as those in commercially available task-based courses, such as Richards, Gordon, and Harper (1995) and Nunan (1996). In these latter cases, what were typically called exercises or activities in older course books are now called tasks, but there is no difference between them. The organizing principle of these course books, apparent from the syllabus descriptions at the front, are grammatical structures, listening microskills, functions, topics, and often more.

In contrast to structure-trapping tasks, and in contrast to course books using task as a synonym for language exercise, Skehan and Long view tasks as purely meaningful activities. Tasks do not implement a covert grammatical or lexical syllabus; tasks alone are the units of syllabus design.

Long (2000, 2007; Long & Crookes, 1993; Long & Robinson, 1998) and Skehan (1996, 1998; Skehan & Foster, 2001) are in broad agreement about the SLA motivation for analytic syllabi, and task-based syllabi in particular, citing research showing: (1) little resemblance between acquisitional sequences and instructional sequences based on linguistic forms (e.g., Ellis, 1989; Lightbown, 1983); (2) evidence that learning is non-linear and cumulative, rather than linear and additive, as synthetic language syllabi imply (see Selinker & Lakshmanan, 1992 on backsliding, and Kellerman, 1985, on U-shaped behavior); and (3) research showing the influence of learnability on the order in which items can be learned (e.g., Mackey, 1999; Pienemann, 1989). Even if a structural syllabus could be sequenced based on what is known of learnability and language development, it would be impossible to time and target instruction at the stage learners are ready to progress to accurately, since there is variation in rate of acquisition, meaning groups of learners do not progress in lockstep, homogeneously through acquisition sequences. Additionally, as Long (2000) points out, linguistic grading, as required by many synthetic structural approaches, at least in the early stages, results in classroom language and texts which are artificial, and functionally and linguistically impoverished, prohibiting exposure to language that learners may be ready to learn. Given their broad agreement over the motivation for choice of task-based syllabi, there are some differences of scope and focus in their proposals.

Long (1985, 2000; Long & Crookes, 1993) describes a number of steps to be taken in implementing task-based language teaching. First, conduct a needs analysis to identify the target, real-world tasks learners need to perform in the L2, then classify the target tasks into types or superordinate categories such as "making/ changing reservations." From the target tasks, derive pedagogic tasks: "Adjusted to such factors as learners' age and proficiency level, these are a series of initially simple, progressively more complex approximations to the target task" (Long, 2000, p. 185). These tasks are then sequenced to form a syllabus, and the program is implemented with appropriate methodology and pedagogy. One methodological principle Long advocates is "focus on form." That is, where individuals or groups of learners are heard repeatedly producing non-target like forms, teacher intervention to provide corrective feedback is recommended. This can take several forms, such as implicit negative feedback, or recasts of learner forms, brief written illustration of the correct form, brief rule explanations, input enhancement of forms in aural and written texts used on task, and a variety of other techniques (see Doughty & Williams, 1998 for an extended summary). Like Long, Skehan rejects linguistic grading as a criterion for task and syllabus design, defining a task as an activity in which, "Meaning is primary; There is a goal which needs to be worked on; the activity is outcome-evaluated; There is a real world relationship" (Skehan, 1998). Skehan concludes that this definition rules out "an activity that focuses on language itself" such as a transformation drill, or

the consciousness-raising tasks described by Ellis (1997), and many of the tasks in Nunan (1996, 2004).

Grading and sequencing tasks

Researched proposals for grading and sequencing tasks in terms of their information-processing and interactional demands have begun to be developed in recent years. One position, taken by Skehan (1998; Skehan & Foster, 2001), is that more demanding tasks "consume more attentional resources . . . with the result that less attention is available for focus on form" (1998, p. 97), therefore sequencing tasks from less cognitively demanding to more demanding optimizes opportunities for attention allocation to language form. Task design is also seen as a means to promote "balanced language development" in the areas of accuracy, fluency, and complexity of production. This can be done because certain task characteristics "predispose learners to channel their attention in predictable ways, such as clear macrostructure towards accuracy, the need to impose order on ideas towards complexity, and so on" (1998, p. 112). However, due to scarcity of attentional resources, tasks can lead either to increased complexity, or accuracy of production, but not to both. Tasks should therefore be sequenced by choosing those with characteristics that lead to each, at an appropriate level of difficulty, as determined by three factors. (1) Code complexity is described in "fairly traditional ways," as in descriptions of structural syllabi, or developmental sequences (1998, p. 99). (2) Cognitive complexity is the result of the familiarity of the task, topic or genre, and the processing requirements, information type, clarity and organization, and amount of computation required. (3) Communicative stress involves six characteristics, including time pressure, number of participants, and opportunities to control interaction. Unlike Johnson, Willis, or Ellis, Skehan does not argue that tasks should be used to deliver and practice a linguistic syllabus. Tasks are sequenced from less to more difficult to minimize what he argues are the negative effects, given limited attentional capacity, of increased cognitive and attentional demands on linguistic performance. The goals of task-based instruction are to promote language development in the areas of accuracy, fluency, and complexity of speech, as well as comprehension, and task selection and classification are not constrained by the need to articulate pedagogic tasks with target tasks identified in a needs analysis.

In contrast, Robinson (2005b, 2007b, 2007c) assumes behavior descriptions of target tasks for populations of learners are the starting point for pedagogic task design. Based on behavior descriptions, task conditions (i.e., the interactional demands of target tasks) are classified using task characteristics, distinguishing them in terms of participation and participant variables (see Table 17.1). Participation variables include: (1) whether the solution to the task is optional (open) or fixed (closed); (2) whether information exchange goes from A to B (one-way), or is reciprocal (two-way); (3) whether agreement is required (convergent) or the opposite (divergent), etc. Participant variables concern interlocutors' relative status, familiarity with each other, and the extent of shared cultural background, etc.

Table 17.1 Characteristics for pedagogic task design and sequencing decisions

Task complexity Task condition

(Cognitive factors) (Interactive factors)

(Classification criteria: cognitive demands)

(Classification procedure: information-theoretic analyses)

Subcategories:

(a) cognitive variables making cognitive/conceptual demands

+/- here and now +/- few elements -/+ spatial reasoning -/+ causal reasoning -/+ intentional reasoning -/+ perspective-taking

(b) cognitive variables making performative/procedural demands

+/- planning time

+/- prior knowledge

+/- single task

+/- task structure

+/- few steps

+/- independency of steps

(Classification criteria: interactional demands)

(Classification procedure: behavior descriptive analyses)

Subcategories:

(a) participation variables making interactional demands

+/- open solution

+/- one-way flow

+/- convergent solution

+/- few participants

+/- few contributions needed

+/- negotiation not needed

(b) participant variables making interactant demands

+/- same proficiency +/- same gender +/ - familiar

+/- shared content knowledge +/- equal status and role +/- shared cultural knowledge

Similarly, following behavior descriptions of target tasks, task complexity (i.e., the cognitive demands of target tasks) is classified using task characteristics, distinguishing them in terms of cognitive/conceptual, and performative/procedural demands. The cognitive/conceptual demands include: (1) whether the task requires reference to events happening now, in a mutually shared context (Here-and-Now) versus events that occurred in the past, elsewhere (There-and-Then); (2) reference to few, easily distinguished, versus many similar elements; (3) reference to spatial location, where easily identifiable and mutually known landmarks can be used, versus reference to location without this support, etc. (see Table 17.1). The Cognition Hypothesis (Robinson, 2001, 2003a, 2003b, 2005b; Robinson & Gilabert, 2007) argues that sequencing pedagogic tasks from simple to complex in terms of cognitive/conceptual demands leads to interlanguage development and L2 learning. This is because expending the mental

effort needed to make more demanding cognitive/conceptual distinctions in language should prime learners - and direct their attentional and memory resources - to aspects of the L2 system required to understand and convey them accurately, thereby facilitating "noticing" of these, and so speeding up L2 grammaticization in conceptual domains, as well as promoting the use of more complex syntax. In contrast, performative/procedural demands of tasks (see Table 17.1) concern variables such as planning time, or whether the task requires one versus multiple simultaneous outcomes to be accomplished. Increasing complexity along these dimensions of cognitive demand (e.g., from tasks with planning time to tasks without) requires increasingly skilled access to and control over current inter-language L2 resources.

In this proposal, the criteria for grading and sequencing tasks, using these characteristics, are explicit. Interactional demands of pedagogic tasks are not graded and sequenced. The task conditions, e.g., +/- one-way flow of information, +/-equal status and role, are replicated each time pedagogic task versions are performed. A rationale for this, offered only briefly here, is that holding task conditions constant is important to ensuring transfer of training to real-world contexts. The more task conditions are practiced in pedagogic versions, the more elaborate and consolidated the scripts become for real-world performance (Schank & Abelson, 1977), on which successful transfer will draw, outside the classroom (Broad, 1997). Cognitive demands of pedagogic tasks, however, are graded and sequenced. Simpler versions with respect to all relevant cognitive demand characteristics are performed first, and then task complexity (i.e., cognitive demands) is gradually increased on subsequent versions to target task levels. Task complexity is, therefore, the sole basis for pedagogic task sequencing.

There are two stages in which task complexity is increased, and which are decision points for task and syllabus design. In each sequence of pedagogic tasks, relevant performative/procedural variables are first increased in complexity (so if the target task requires dual task performance, without planning time, then planning time is first provided, and the dual task characteristics are first performed separately). The rationale for this is to promote access to and consolidate the learner's current L2 interlanguage system during pedagogic task performance. Subsequently increasing performative and procedural demands to target task levels thereby promotes increased automatic access to, and learner "control" over, the current system in responding to pedagogic task demands.

In the second stage, once the performative/procedural demands have reached targetlike levels, cognitive/conceptual demands are gradually increased to targetlike levels. As described above, these can direct learners' attentional and memory resources to aspects of the L2 system needed to encode increasingly complex concepts, and to meet increasingly complex functional demands requiring their expression in language. This promotes analysis and development of the current interlanguage system. Increasing these demands should lead to more accurate and complex learner production, more noticing of task relevant input, and heightened memory for it, and so lead to more uptake of forms made salient in the input through various focus on form interventions.

Conclusion

The proposal for grading and sequencing tasks made above, and for task characteristics that can be manipulated by materials and syllabus designers for this purpose, has been, in large part, motivated by theories of cognitive processing in cognitive psychology and by findings from SLA research. Only in the later stages of the twentieth century did it become possible to motivate pedagogic design and decision-making in this way. Very similar cognitive-processing approaches to task analysis, grading, and sequencing are now currently widespread in many other areas of instructional design (see e.g., Hollnagel, 2003; Schraagen, Chipman, & Shalin, 2000). Very similar philosophies and broad principles of instructed learning, too, also underlie the proposals that have been made for learn-by-doing simulations, and the use of tasks, in other areas of education: ". . . children are learning in a decontextualised fashion. Lessons are taught in a way in which use of knowledge or skills is divorced from how they would be used in real life . . . When students learn how, they inevitably learn content knowledge in the service of accomplishing their task" (Schank, Berman, & MacPherson, 1999, pp. 165-6). The shift from synthetic to analytic approaches to syllabus design, reflected in the sequencing of sections in this chapter, is one that can be expected to continue. Future research, theory, and practice in L2 syllabus design would do well to look to these other areas of instructional theory and curriculum development for the insights they can offer, while continuing to integrate them with what is known of the processes constraining and promoting L2 acquisition.

REFERENCES

Anderson, J. R. (1992). Automaticity and the ACT* theory. American Journal of Psychology 105, 165-80.

Breen, M. (1984). Process syllabuses for the language classroom. In C. J. Brumfit (ed.), General English syllabus design (pp. 47-60). Oxford: Pergamon.

Broad, M. (1997). Overview of transfer of training: From learning to performance. Performance Improvement Quarterly 10, 2, 7-21.

Brockett, C. (2000). A communicative framework for introductory Japanese language curricula. Washington State Japanese Language Curriculum Guidelines Committee: Second Language Teaching and Curriculum Center, Technical Report # 20. Honolulu: University of Hawai'i Press.

Bygate, M., Skehan, P., & Swain, M. (eds.) (2001). Researching pedagogic tasks: Second language learning, teaching and testing. London: Longman.

Candlin, C. (1984). Syllabus design as a critical process. In C. J. Brumfit (ed.), General English syllabus design (pp. 29-46). Oxford: Pergamon.

Clarke, D. (1991). The negotiated syllabus: What is it and is it likely to work? Applied Linguistics 12, 13-28.

Crombie, W. (1985a). Process and relation in discourse and language learning. Oxford: Oxford University Press.

Crombie, W. (1985b). Discourse and language learning: A relational approach to syllabus design. Oxford: Oxford Uniiversity Press.

DeKeyser, R. M. (2001). Automaticity and automatization. In P. Robinson (ed.), Cognition and second language instruction (pp. 125-51). Cambridge: Cambridge University Press.

Doughty, C. J. (2001). Cognitive underpinnings of focus on form. In P. Robinson (ed.), Cognition and second language instruction (pp. 206-57). Cambridge: Cambridge University Press.

Doughty, C. J. & Williams, J. (1998). Pedagogical choices in focus on form. In C. J. Doughty & J. Williams (eds.), Focus on form in classroom second language acquisition (pp. 197-262). New York: Cambridge University Press.

Ellis, N. (2005). At the interface: Dynamic interactions of implicit and explicit knowledge. Studies in Second Language Acquisition 27, 305-52.

Ellis, R. (1989). Are classroom and naturalistic acquisition the same? A study of the classroom acquisition of German word order rules. Studies in Second Language Acquisition 11, 305-28.

Ellis, R. (1993). The structural syllabus and second language acquisition. TESOL Quarterly 27, 91-113.

Ellis, R. (1997). SLA research and language teaching. Oxford: Oxford University Press.

Ellis, R. (2003). Task-based language learning and teaching. Oxford: Oxford University Press.

Ellis, R. (ed.) (2005). Planning and task performance in a second language. Amsterdam: John Benjamins.

Ellis, R. & Noboyushi, J. (1993). Focused communication tasks. ELT Journal 47, 203-10.

Fotos, S. & Ellis, R. (1991). Communicating about grammar: A task-based approach. TESOL Quarterly 25, 87-112.

Garcia-Mayo, M. P. (ed.) (2007). Investigating tasks in formal language learning. Clevedon, UK: Multilingual Matters.

Gries, S. Th. (2008). Corpus analysis and second language acquisition data. In P. Robinson & N. C. Ellis (eds.), Handbook of cognitive linguistics and second language acquisition (pp. 406-31) London: Routledge.

Hollnagel, E. (ed.) (2003). Handbook of cognitive task design. Mahwah, NJ: Lawrence Erlbaum.

Hornby, A. S. (1959). The teaching of structural words and sentence patterns. Oxford: Oxford University Press.

Howatt, A. P. & Widdowson, H. G. (2004). A history of English language teachng, 2nd edn. Oxford: Oxford University Press.

Johnson, K. (1996). Language teaching and skill learning. Oxford: Blackwell.

Kellerman, E. (1985). Input and second language acquisition theory. In S. Gass & C. Madden (eds.), Input in second language acquisition (pp. 345-53). Rowley, MA: Newbury House.

Lightbown, P. (1983). Exploring relationships between developmental and instructional sequences. In H. G. Seliger & M. H. Long (eds.), Classroom oriented research in second language acquisition (pp. 217-43). Rowley, MA: Newbury House.

Long, M. H. (1985). A role for instruction in second language acquisition. In K. Hyltenstam & M. Pienemann (eds.), Modeling and assessing second language acquisition (pp. 77-99). Clevedon, UK: Multilingual Matters.

Long, M. H. (2000). Focus on form in task-based language teaching. In R. Lambert & E. Shohamy (eds.), Language policy and pedagogy: Essays in honor of Ronald Walton (pp. 181-96). Amsterdam: John Benjamins.

Long, M. H. (ed.) (2005). Second language needs analysis. New York: Cambridge University Press.

Long, M. H. (2007). Problems in SLA. Mahwah, NJ: Lwrence Erlbaum.

Long, M. H. & Crookes, G. (1993). Units of analysis in syllabus design: The case for task. In G. Crookes & S. Gass (eds.), Tasks in a pedagogical context (pp. 9-54). Clevedon, UK: Multilingual Matters.

Long, M. H. & Robinson, P. (1998). Focus on form: Theory, research, and practice. In C. J. Doughty & J. Williams (eds.), Focus on form in classroom SLA (pp. 15-41). New York: Cambridge University Press.

Loschky, L. & Bley-Vroman, R. (1993). Grammar and task-based methodology. In G. Crookes & S. Gass (eds.), Tasks and language learning: Integrating theory and practice (pp. 12367). Clevedon, UK: Mulitilingual Matters.

Mackey, A. (1999). Input, interaction and second language development. Studies in Second Language Acquisition 21, 557-87.

Munby, J. (1978). Communicative syllabus design. Cambridge: Cambridge University Press.

Nunan, D. (1996). Atlas. Boston, MA: Heinle and Heinle.

Nunan, D. (2004). Task-based language teaching. Cambridge: Cambridge University Press.

Palmer, H. (1922). The scientific study and teaching of languages. London: Harrap.

Pienemann, M. (1989). Is language teachable? Applied Linguistics 10, 52-79.

Reigeluth, C. M. (ed.) (1999). Instructional-design theories and models: A new paradigm of instructional theory, vol. 2. Mahwah, NJ: Lawrence Erlbaum.

Richards, J. (1990). The language teaching matrix. Cambridge: Cambridge University Press.

Richards, J., Gordon, D., & Harper, A. (1995). Listen for it: A task-based listening course. Oxford: Oxford University Press.

Richterich, R. (1972). A model for the definition of language needs of adults learning a modern language. Strasbourg: Council of Europe.

Robinson, P. (1995). Task complexity and second language narrative discourse. Language Learning 45, 99-140.

Robinson, P. (1996a). Consciousness, rules and instructed second language acquisition. New York: Peter Lang.

Robinson, P. (ed.) (1996b). Task complexity and second language syllabus design: Data-based studies and speculations. Special issue of University of Queensland Working papers in Language and Linguistics. Brisbane: CLTR.

Robinson, P. (1997). Individual differences and the fundamental similarity of implicit and explicit adult second language learning. Language Learning 47, 45-99.

Robinson, P. (2001). Task complexity, task difficulty and task production: Exploring interactions in a componential framework. Applied Linguistics 22, 27-58.

Robinson, P. (2002). Learning conditions, aptitude complexes and SLA: A framework for research and pedagogy. In P. Robinson (ed.), Individual differences and instructed language learning (pp. 110-33). Amsterdam: John Benjamins.

Robinson, P. (2003a). Attention and memory during SLA. In C. J. Doughty & M. H. Long (eds.), Handbok of second language acquisition (pp. 630-78). Oxford: Blackwell.

Robinson, P. (2003b). The Cognition Hypothesis of adult, task-based language learning. Second Language Studies 21, 45-107. www.hawaii.edu/sls/uhwpesl/21(2)/Robinson.pdf.

Robinson, P. (2005a). Aptitude and second language acquisition. Annual Review of Applied Linguistics 25, 46-73.

Robinson, P. (2005b). Cognitive complexity and task sequencing: A review of studies in a componential framework for second language task design. International Review of Applied Linguistics 43, 1-32.

Robinson, P. (2007a). Aptitude, abilities, contexts and practice. In R. M. DeKeyser (ed.), Practice in second language learning: Perspectives from applied linguistics and cognitive psychology (pp. 256-86). Cambridge: Cambridge University Press.

Robinson, P. (2007b). Criteria for classifying and sequencing pedagogic tasks. In M. P. Garcia-Mayo (ed.), Investigating tasks in formal language learning (pp. 7-27). Clevedon, UK: Multilingual Matters.

Robinson, P. (2007c). Task complexity, theory of mind and intentional reasoning: Effects on L2 production, interaction, uptake and perceptions of task difficulty. In P. Robinson & R. Gilabert (eds.), Task complexity, the Cognition Hypothesis and second language instruction. Special issue of International Review of Applied Linguistics 45, 195-215.

Robinson, P. & Gilabert, R. (2007). Task complexity,the Cognition Hypothesis and second language learning and performance. In P. Robinson & R. Gilabert (eds.), Task complexity, the Cognition Hypothesis and second language. Special issue of International Review of Applied Linguistics 45, 161-76.

Robinson, P. & Ross, S. (1996). The development of task-based testing in EAP programs. Applied Linguistics 17, 455-76.

Robinson, P., Ting, S., & Urwin, J. (1995). Investigating second language task complexity. RELC Journal 25, 62-79.

Schank, R. & Abelson, R. (1977). Scripts, plans, goals and understanding. Hillsdale, NJ: Lawrence Erlbaum.

Schank, R., Berman, T., & Macpherson, K. (1999). Learning by doing. In C. M. Reigeluth (ed.), Instructional-design theories and models: A new paradigm of instructional theory, vol. 2. (pp. 161-82). Mahwah, NJ: Lawrence Erlbaum.

Schraagen, J. M., Chipman, S. F., & Shalin, V. L. (eds.) (2000). Cognitive task analysis. Mahwah, NJ: Lawrence Erlbaum.

Selinker, L. & Lakshmanan, U. (1992). Language transfer and fossilization. In S. Gass & L. Selinker (eds.), Language transfer in language learning (pp. 196-215). Amsterdam: John Benjamins.

Sinclair, J. McH. (2004). Trust the text. London: Collins.

Skehan, P. (1996). A framework for task-based approaches to instruction. Applied Linguistics 17, 34-59.

Skehan, P. (1998). A cognitive approach to language learning. Oxford: Oxford University Press.

Skehan, P. & Foster, P. (2001). Cognition and tasks. In P. Robinson (ed.), Cognition and second language instruction (pp. 183-205). Cambridge: Cambridge University Press.

Van Ek, J. (1976). The threshold level for modern language learning in schools. London: Longman.

White, R. (1988). The ELT curriculum: Design, management, innovation. Oxford: Blackwell.

Widdowson, H. G. (1978). Teaching language as communication. Oxford: Oxford University Press.

Widdowson, H. G. (1990). Aspects of language teaching. Oxford: Oxford University Press.

Wilkins, D. (1976). Notional syllabuses. Oxford: Oxford University Press.

Willis, D. (1990). The lexical syllabus: A new approach to language teaching. London: Collins.

Willis, J. (1996). A framework for task-based learning. Oxford: Longman.

FURTHER READING

Council of Europe (2001). Common European Framework of Reference for Languages: Learning,

teaching, assessment. Cambridge: Cambridge University Press. Long, M. H. & Crookes, G. (1992). Three approaches to task-based syllabus design. TESOL

Quarterly 26, 27-56. Nunan, D. (1988). Syllabus design. Oxford: Oxford University Press.

Prabhu, N. S. (1987). Second language pedagogy. Oxford: Oxford University Press.

Richards, J. (2001). The second language curriculum. Cambridge: Cambridge University Press.

Robinson, P. (2001). Task complexity, cognitive resources, and second language syllabus design: A triadic framework for examining task influences on SLA In P. Robinson (ed.), Cognition and second language instruction (pp. 285-317). Cambridge: Cambridge University Press.

Robinson, P. & Gilabert, R. (eds.) (2007). Task complexity, the Cognition Hypothesis and second language instruction. Special issue of International Review of Applied Linguistics 45, 161289.

18 Advances in Materials Design

ALAN WATERS

To derive the design of an object from its natural functions and conditions.

Walter Gropius

Introduction

Any discussion of the topic of this chapter must begin with a clarification of the central terms involved. "Materials" can, of course, mean any or all of the very wide range of resources capable of aiding language learning. Here, however, it refers to major international language teaching textbooks, such as the Headway (e.g., Soars & Soars, 2003), Interchange (e.g., Richards, Hull, & Proctor, 2004) and other series. This is partly for reasons of space - the field is vast, and only a small part of it can therefore be covered here (and even then, not comprehensively) -but primarily in order to connect what follows with previous and ongoing discussions on the topic of developments in materials design, many of which have focused on the same type of resource. Also, because of the relative ubiquity of materials of this kind, such a focus makes it more likely that readers will already be familiar with and/or able to refer personally to the examples cited.

This leaves much out of account, particularly as the majority of language teaching materials used around the world, especially in state educational systems, are probably locally produced. Although unavoidable in this context, this is obviously an important omission, and one of a number of aspects of materials design (as will be seen) where a good deal more research might be undertaken. It should also be mentioned that all the samples of materials referred to are concerned with the teaching of English as a foreign language. Nevertheless, the underlying principles and issues involved should be of relevance to all areas of foreign language teaching.

It is also important to attempt to define the sense in which the term "advances" is being used. At least two major perspectives exist regarding the matter. The first of these is an applied linguistics point of view, as expressed in, e.g., Littlejohn (1992), Tomlinson (1998, 2003), and others. In a nutshell, this perspective argues that the design of teaching materials should, as far as possible, reflect advances in

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

academic theorizing and research concerning language, language learning, and education. The second view is what might be termed an "audience-based" one -that is, a perspective about materials design which makes primary reference to perceptions of the needs of end users of teaching materials. Though present in the literature on materials design to a limited extent (see, e.g., Bell & Gower, 1998; Mares, 2003; Richards, 2001), this view has been expressed mainly in the form of concrete developments in various aspects of materials design, as prompted by authors' and publishers' "readings" of audience feedback, i.e., their understanding of what design features are seen to work best in order to facilitate language learning and teaching in the situations where the materials will be used.

In practice, these two trends have often intermingled, with many sets of teaching materials attempting to reflect, to varying degrees, insights from applied linguistics, as well as those based on perceptions of audience need. Nevertheless, in academic discussions of the topic, it is the applied linguistics perspective which has tended to predominate. "Advances" in materials design have therefore usually been interpreted rather one-sidedly, as a reflection of the extent to which textbooks have succeeded in incorporating features which have been seen as desirable from an academic point of view.

It has been recognized for some time, however, that the general relationship between applied linguistics and language teaching should be a dialectical one, thereby granting due cognizance to issues of practice as the starting point (though not the only end) for "applied" intellectual enquiry (see, e.g., Widdowson, 2000). In what follows, therefore, the term "advances" has been conceived of in the first instance as a function of those developments in materials design that can be seen as likely to contribute to making classroom language learning, in the type of situations the materials are intended for, a more positive experience than would otherwise be the case. In other words, the overall criterion being used is one of "fitness for purpose," regardless of "theoretical 'correctness'." As a corollary, a realigned and expanded research agenda in this area is delineated, with a view to creating the means for the two main materials design perspectives identified to develop a more productive symbiosis than has occurred hitherto.

The remainder of this chapter consists, first of all, of a brief reprise of the findings of two "benchmark" reviews of developments in materials design from the late 1980s. This serves as a "baseline" for the subsequent sections, which are concerned with investigating how a number of the main features of materials analyzed in the two reviews, as well as other aspects, have developed subsequently, via an examination of a range of contemporary textbooks. The final section attempts to draw the main threads together and to consider the implications for further development in this area.

Departure

The two earlier textbook surveys in question are Rossner (1988) and Clarke (1989). In Rossner, given the general acceptance at the time, as now, of the view that the

goal of language learning is communication, the point of departure is "to examine what materials-writers have understood to be the role of their work in the communicative classroom" (Rossner, 1988, p. 140). On the basis of an analysis of a representative cross-section of coursebooks (e.g., Swan & Walter, 1984), supplementary materials (e.g., Frank, Beres, & Rinvolucri, 1989) and "resource books" (e.g., Sion, 1985) of the time, Rossner concludes that:

materials . .. have not suddenly become "communicative" .. .; rather, materials have become more and more varied as the drive for more and more interesting, and less and less constraining, ways of carrying out language "practice" in the classroom has gathered pace. (p. 142)

In other words, a more "traditional" focus is perceived to have remained intact, despite the addition of a communicative "overlay."

The following main criticisms of the materials are also made. First, they are seen to suffer from an "embarras de richesses." Getting to know, selecting appropriately, and using them judiciously are therefore viewed as more complex and time-consuming tasks than was the case in the pre-communicative era. Second, most of their communication activities are not regarded as providing "true communication with a real purpose" (Rossner, 1988, p. 160), because of their frequent artificiality and lack of relevance to learners' lives. Third, the predominance of UK and US publishers is seen as "dangerous," since, in general, they are "unable to avoid projecting through their topics and their approaches to them, through the language they select, and through the very ethos of the activities they craft, values and educational attitudes which are intrinsically Western and mainly 'Anglo-Saxon'" (p. 160). The review as a whole concludes by saying that "In the communicative era, more than ever before, materials should not seek to mold teachers and learners but should be available for molding by them" (p. 161). As a whole, thus, the materials in question are seen as insufficiently suited to the needs of their intended audiences because of their logistical complexity, lack of personalization, and the cultural bias of their content and methodology. How subsequent sets of materials have fared with respect to all these features will also be explored in due course.

Like Rossner, the purpose of Clarke (1989) was to "indicate some of the ways in which the now established principles of communicative language learning have been translated into actual teaching materials" (p. 73). In doing so, the first part of the paper (i.e., sections 2-8) focuses mainly on the "authenticity debate," in order to prepare the way, in the second part, for an analysis of the extent to which various concepts concerned with the notion of "authenticity" - of text, text use, and learner response - can be seen to have influenced the design of a range of then contemporary textbooks (e.g., Abbs & Freebairn, 1980; Soars & Soars, 1986; Swan & Walter, 1985).

With respect to authenticity of text, many of the materials are seen as having gone "to considerable lengths to stimulate [sic] real materials and to reflect 'real life' in order to create an aura of authenticity" (Rossner, 1988, p. 79). However, at

the same time, some of these features, such as photographs, are seen as having "very little or no direct pedagogical value," and there is a "widespread phenomenon of 'simulated realia'," i.e., graphics, which "do not involve reproduction of the actual text but seek to suggest an identity with some authentic original by devices such as drawing book shapes round lists of appointments to suggest diaries" (p. 79). The latter is also seen to be a feature of many of the listening texts used in the materials, most of which are "scripted or semi-scripted," despite claims to authenticity (p. 80).

In terms of the authenticity of tasks to texts, traditional wh- comprehension questions continue to dominate: "the authentic input data provide little more than pretexts for traditional 'reading comprehension' activities based, for the most part, on irrelevant details being excavated for no particular reason, with no further use proposed for this information" (Clarke, 1989, p. 81). "Authentic materials (which are by definition user-specific)" are thereby seen as being used "for the development of 'general' reading and 'comprehension' skills" (p. 81). Similarly, authentic texts are also reported as being frequently used as vehicles for traditional language focus exercises, such as substitution tables, and the texts themselves are often modified, e.g., by having gaps inserted in them. As a result "the focus of those materials tends to remain on the forms or functions of the language rather than the use to which the language can be put" (p. 82).

Authenticity of context (i.e., the building up of a realistic context of use around the language being focused on) is also seen as underdeveloped in most of the materials in question, with the language practice factor remaining dominant. Similarly, the development in the materials of authenticity of the task to the learner is seen as circumscribed by the difficulties of creating sufficient individuation and personalization at the same time as attempting to cater to a mass market: "Commercial requirements to sell as widely as possible necessarily vitiate the authenticity of much of this material insofar as the discourse types, situations and roles proposed can by no means be guaranteed to evoke learner authentication" (p. 83).

In overall terms, thus, the most obvious feature of Clarke's analysis, like Rossner's, is the identification of a number of fundamental ways - principally to do with issues of "realism" - in which the majority of the teaching materials reviewed were seen to have failed to live up to the theoretical ideals of the communicative approach. However, Clarke concludes by pointing out that the theoretical basis for advocating authenticity in teaching materials is characterized by contradictory stances, even though there is a tendency for this to be overlooked, and for the pro-authenticity view to predominate regardless. He feels it is therefore important to note that "the use of authentic materials does not inevitably result in performance-based activity, while such activity can be generated without the use of authentic materials" (1989, p. 84, original emphasis). He characterizes the teaching materials of the time in overall terms as follows:

While most modern textbooks work hard at achieving at least the aura of authenticity, it should be noted that much of their content still focuses on knowledge of the

language rather than its use . . . it is quite apparent that there is no escaping from the production of pedagogical materials and no need to conceal the fact that there will always be a need for transitional materials, which, while not in themselves authentic, can be authenticated by the learner. The extent to which modern materials tend not to exemplify the communicative principles they purport to embody seems to support this assertion. ( p. 84)

The implication here - that materials have to address needs that extend beyond or in contradistinction to the incorporation of "authentic data" - echoes the flavor of the discussion in the second part of the introduction to this chapter. Clarke appears to argue that the reason why the materials in question have the characteristics identified is not so much because of shortcomings on the part of the authors and publishers, but because the goal of authenticity in its conventional formulation is inappropriate, since what counts as "authentic" from the learner's perspective may well be a different matter.

In other words, a fundamental dichotomy of tendencies is perceived to exist. On the one hand, the main thrust of applied linguistics was (and has continued to be) toward accounting for the factors involved in "natural" language use, and advocating that these should form the basis of teaching materials. On the other hand, however, we have teaching materials which, while decked out in the trappings of target situation authenticity, remain, at root, fundamentally language learning oriented, i.e., based on the view that the kind of authenticity most required for foreign language learning should relate first and foremost to the learner as a current interim acquirer rather than as a potential future user of the language. As will be seen, a similar tension underlies most subsequent developments in language teaching textbook design.

Analysis

Authenticity

Authenticity of text

In the New Headway Intermediate Teacher's Book (Soars, Soars, & Sayer, 2003), the authors state that the Student's Book reading and listening texts "come from authentic sources with the necessary adaptations to suit the level" (p. 4). This is confirmed by an examination of the texts in the Student's Book itself, which have many of the attributes of authentic texts in terms of layout, subject-matter, cohesion, and so on, but, in most cases, the language, while natural-sounding, lacks the idiomaticity and complexity of normal native-speaker discourse. They are thus "simple accounts" (Widdowson, 1979, p. 184). Much the same appears to hold true for a wide range of other recent coursebooks, e.g., New Hotline Elementary (Hutchinson, 1998), Cutting Edge Intermediate (Cunningham & Moor, 1999), Language to go Intermediate (Crace & Wileman, 2002), In English Elementary (Viney & Viney, 2004a), Interchange Student's Book 1 (Richards, Hull, & Proctor, 2005),

and so on. In other words, rather than a move toward greater use of genuinely authentic texts, there seems to have been a consolidation of the status ante quo, i.e., the use of pseudo-authentic or specially-constructed texts has become something of a norm in modern textbook design.

Authenticity of text use

The way in which such texts are used in most modern textbooks tends to be mainly for comprehension or language work purposes, although there are also often pre- and/or post-comprehension or language work activities which involve the learners in relating the information content of the texts to their own lives. Thus, in Language to go Intermediate (Crace & Wileman, 2002, p. 69), the "Listening" section begins by getting the students to look at pictures of a number of possible future electronic communication inventions and describe them. They then listen to four short texts about the inventions and match each of them to the relevant picture. After this, they listen again in order to match a list of the inventions with the year when it is predicted they will be available. This is followed by a "Grammar focus" section containing a number of exercises concerned with the main language point (will and will have done) in the texts. The page ends with a "Get talking" section in which students make predictions about their futures.

Although there is variety in terms of whether the main focus is more or completely on the comprehension or language development side, and as to whether there are also activities which relate the language and/or ideas to the students' own lives, many other recent textbooks follow the same basic pattern in terms of listening and reading text exploitation (see, e.g., Cutting Edge Intermediate (Cunningham & Moor, 1999, pp. 104-5); Interchange Student's Book 1 (Richards et al., 2005, p. 111); New Hotline Elementary (Hutchinson, 1998, pp. 58-9); In English Elementary (Viney & Viney, 2004a, pp. 142-8); New Headway Intermediate Student's Book (Soars & Soars, 2003, pp. 84-5), and so on). Thus, in overall terms, as with the type of text used, most contemporary textbooks, while allowing for a measure of authentic text use via activities which encourage the learners to relate the content to their own lives, tend to focus primarily on exploiting them for comprehension and language development work.

Authenticity of task to learner

As with trends in terms of text type and text use, communication tasks in many modern textbooks appear to have become more closely based on the likely knowledge and interests of the typical learner rather than involving communication situations taken more directly from real life. They are therefore mostly of a kind that might also occur outside the classroom, but appear to have been constructed primarily in order to provide an opportunity to use the language being studied in conjunction with the students' existing world. Thus, for example, in Cutting Edge Intermediate, Module 7, the main task (Cunningham & Moor, 1999, pp. 74-5) is concerned with getting the learners to make a list of "dos and don'ts" about everyday behavior for visitors to their countries; in New Headway Intermediate,

Unit 11, one of the main tasks (Soars & Soars, 2003, pp. 90-1) involves the students in making a poster concerned with asking and answering a question about an area of world knowledge; in Language to go Intermediate, the main task in Unit 29 (Crace & Wileman, 2002, pp. 60-1) consists of having the learners discuss what they will or might do if a variety of everyday situations were to occur; in In English Elementary, Unit 27 culminates in a task (Viney & Viney, 2004a, p. 141) in which the learners give their views about the likelihood of a variety of predictions about the future of the world; in New Hotline Elementary, Unit 7 ends by getting the learners to stage a class fashion show (Hutchinson, 1998, p. 62); and so on. The primary concern of such tasks appears to be one of authenticity to the learning situation, i.e., the provision of tasks that have the potential to enable the learners to put the language knowledge in question into practice in a lifelike way, by being geared sufficiently closely to their level, interests, and so on - a matter of attempting to simulate rather than replicate real-life use.

In overall terms, thus, many major textbooks have evolved to reinforce and develop further the tendencies noted by Rossner and Clarke at the end of the 1980s, namely, on the one hand, to deploy pseudo-authentic texts and exploit them in inauthentic ways, and, on the other, to move away from (more authentic) "target" tasks and closer toward (less authentic) "pedagogical" ones (Nunan, 2004, pp. 1-4). This has occurred despite a continuing growth of interest in and belief in the pedagogical value of "real language data" and real-world tasks in applied linguistics (see, e.g., Carter, Hughes, & McCarthy, 1998; Skehan, 1998).

This trend appears to have occurred because of the important pragmatic advantages which can thereby accrue to textbook writers in their attempts to make their materials fit for their primary purposes. In reflecting on their experiences in writing a major international textbook series, Bell and Gower (1998) say, "our original intention to draw target language out of authentic texts failed at the intermediate level, partly because of the difficulty of finding texts which contained clear examples of the focus language together with interesting content" (p. 127). They also go on to add that, as the writing process progressed:

it was clear there were going to be problems with unadapted authentic texts. Finding texts with a generative topic of the right length and the right level of compre-hensibility for the level . . . as well as an accessible degree of cultural reference and humour was not easy. So we compromised on this ambition and wonder now whether we should have compromised more and simply gone for texts which were interesting. ( p. 128)

In other words, texts for learning purposes clearly need to satisfy a number of needs, primarily related to fitness for the learning purpose, and "authentic" texts may often be inappropriate in this respect. It therefore seems plausible to regard the long-term trend away from "authenticity" and toward "artificiality" in textbooks as concerned with increasing their potential to cater more adequately to the full range of students' learning needs. Artificial but life-like texts make it possible to provide a much greater variety of texts that are likely to be accessible

to most learners than would otherwise be the case; their use as vehicles for comprehension and/or language work is not inauthentic to the purposes for which they were constructed; and the use of "pedagogic" tasks can increase relevance to the learners' world, while still providing a meaningful simulation of real-life language use.

Language practice

Despite the addition of various elements giving the textbooks they reviewed a more communicative "gloss," both Rossner (1988) and Clarke (1989), as has already been pointed out, saw them as nevertheless retaining an overall focus on "language practice." Since those reviews were written, academic ideas in this area have, in general, continued to move further and further away from viewing a direct focus on the conscious study of language forms as being a desirable pedagogical strategy, with various more "indirect" alternatives being favored instead, such as "consciousness-raising" (Ellis, 1993). However, while many textbooks have for some time incorporated activities of this kind, a language practice element has also remained prominent and, over the years, rather than declined, appears to have actually increased.

Thus, for example, in Unit 7 of the original, 1986 version of Headway Intermediate (Soars & Soars, 1986), the language work element comprises two Presentation and Practice sections, the first consisting of two consciousness-raising (C-R) exercises concerning the meaning of the main language structure in focus (the present perfect) and two related "practice" exercises, the second a further C-R activity and related practice exercise, plus a "language review" section (i.e., overall explanation of the rules), a translation exercise, and an "analysis" exercise. In the subsequent Skills Development section, there is a further analysis activity and related practice exercise, and a focused role-play. So all in all, the unit contains three C-R and five language practice activities. In the equivalent unit in the latest edition of the same textbook (Soars & Soars, 2003), the language work concerning the main language focus (once again, the present perfect) is as follows. In the first two pages (pp. 54-5) there is a practice exercise (Test your grammar) in which the students show their existing ability to use the form under focus, as well as to analyze it. This is followed by two fill-in-the-gap practice exercises and a C-R exercise. In the next two pages (pp. 56-7) there is one C-R activity, three analysis ones, and seven language practice exercises. The Listening and Speaking section on p. 61 contains two further practice exercises. There are, thus, altogether, two C-R and twelve language practice exercises concerned with the main language focus of the unit in this edition.1 Similarly, in Unit 6 of the original edition of Hotline Elementary (Hutchinson, 1991), which focuses on the past continuous, there were three C-R and seven language practice exercises. In the later edition, New Hotline Elementary (Hutchinson, 1998), the equivalent unit contains three C-R and thirteen language practice exercises. Thus, while the number of CR exercises in the two textbooks has remained reasonably similar, the language practice ones have increased substantially.

Many other modern textbooks also contain a substantial proportion of exercises of both kinds. On the basis of their survey of nine sets of materials published between 1991 and 2000 (with the majority having been issued in the latter half of this period), Nitta and Gardner (2005) show that approximately equal numbers of C-R and language practice exercises were used. As they conclude, "While recent SLA research continues to provide arguments against the efficacy of practising tasks, the evidence from our analysis suggests that they still occupy an essential place in general ELT materials ... moreover, many suggest using workbooks for further practice" (p. 9). However, as they also go on to say:

Notwithstanding this continued emphasis on practice, our findings have revealed that contemporary coursebooks usually juxtapose C-R tasks with practising tasks. Rather than exclusively selecting one approach, material writers tactfully design grammar syllabuses building on both C-R and practice. Accordingly, though researchers insist on the effectiveness of C-R rather than practice in theory - and rationally their arguments are convincing - ELT practitioners may not be prepared to abandon the familiar, tried and true "practice" exercises. (p. 10)

In other words, it appears that the pragmatic experience of classroom teaching -as it feeds into materials writing via data obtained from practitioners by publishers' agents (Donovan, 1998) - has added a dimension to textbook design in this area which is generally lacking in the academic paradigm, based, as much of it is, mainly on a model of "natural" language use not redolent of the circumstances pertaining in the average school-level or adult language learning situation. As Widdowson (2003) points out (cf. Prabhu, 1992):

a moment's reflection makes it clear that what is taught in classrooms in certain crucial respects cannot be in accordance with actual language use. Actual language use occurs naturally within the continuities of social life, appropriately activated by context, and motivated by the needs of communication and the expression of communal and individual identity. The language subject does not occur naturally at all: it appears, like other subjects, discontinuously on the timetable, fitted into a schedule as suited to administrative convenience. Usually, there is no natural communal or individual impetus to use the language: contexts have to be contrived and motivation created. And this is done within restricted units of time called periods and units of activity called lessons, which are organized into such things as exercises, tasks, tests, group work, and so on. Furthermore, these events are, for the most part, controlled and orchestrated by teacher authority, and directed at an eventual measurable outcome. On the face of it, it is hard to see any resemblance to the natural conditions of actual language use at all. (p. 112, original emphasis)

Thus, rather than perceiving the continued or increased provision for so much practice work in textbooks as regrettable (as in, e.g., Tomlinson et al., 2001), it can be viewed more positively as evidence of a need for re-thinking research and theorizing in this area. In other words, as Swan (2006) argues, the policy adopted by textbooks in this matter can be regarded as helping to avoid the perils of a "subtractive" approach:

Changes in theoretical or pedagogic fashion often come about because of disillusionment: our teaching doesn't seem to be getting very good results, and the temptation is to drop what we are doing and look for alternatives. But this may not bring about any net gain. If we are doing too much formal input and not enough communicative output, the solution is to balance things up, not to move to a position where we are doing too much communicative output and not enough formal input. .. Such approaches [i.e., the latter] are nearly always subtractive as well as additive, putting a great deal of emphasis on one or other ingredient of language teaching while neglecting others. (pp. 53-4)

The implications of this policy for the further development of a materials design research agenda within applied linguistics will be considered in the final part of this chapter.

Syllabus/Unit structure

An additive approach is also evident in the way that contemporary textbooks tend to be structured in terms of their "horizontal" (syllabus) and "vertical" (unit) dimensions. In terms of the former, the "multi-syllabus" (Swan & Walter, 1990) is nowadays the norm. Thus, the Contents of each unit in In English Elementary include "Language Focus," "Communication Skills/Functions/Formulas," "Topic/Vocabulary," and "Extension" (additional language work, etc.) (Viney & Viney, 2004b, pp. 3-7); in New Headway Intermediate (Soars & Soars, 2003), in addition to the Unit Topic (e.g., "It's a wonderful world!"), the categories are "Grammar," "Vocabulary," "Everyday English," "Reading," "Speaking," "Listening," and "Writing"; in Cutting Edge Intermediate (Cunningham & Moor, 1999), in addition to the overall Module topic, (e.g., "About you"), the main contents headings are "Language focus," "Vocabulary," "Speaking," "Reading/Listening," "Task," and "After the task." Other books also include additional elements, such as a "learning to learn" syllabus (see, e.g., New Hotline Elementary (Hutchinson, 1998)). Such multi-stranded forms of organization typify nearly all modern international textbooks.

Such a trend can be interpreted as a "cover all bases" approach, aimed at ensuring that no "market need" is overlooked, or, more positively, as an implicit recognition of the complex, multi-layered nature of the language learning "task." It can be argued that the earlier type of textbook syllabus, in which only a single aspect tended to predominate (e.g., structures, functions, situations, etc.), probably lessened the extent to which other elements were properly developed. The more explicit acknowledgment of a wider range of syllabus components in modern materials has the potential to reduce this problem, and is more in keeping with (some) contemporary views about language and language learning.

It might be argued, however, that a more appropriate textbook design response to a growing recognition of the complex nature of language learning would have been to reduce, rather than increase, the degree of explicit segmentation and specification of the ingredients in the language learning "cake." As Skehan (1996, p. 19) states:

SLA research . . . has established that teaching does not and cannot determine the way the learner's language will develop. The processes by which the learner operates are "natural" processes. Teachers and learners cannot simply "choose" what is to be learned. To a large extent the syllabus is "built in" to the learner.

The development of the multi-syllabus as a mainstay of the modern textbook is therefore noteworthy in terms of the way that, in this respect also, theorizing and research appear to have gone in one direction, and textbook design in another.

In addition to "syntagmatic" textbook structuring of this kind, something of a consensus also seems to have emerged over the last two decades regarding textbook unit structure. The mold appears to have been established in this respect by the appearance of the first Headway series (Soars & Soars, 1986) in the 1980s, and boils down to an initial section in which the main area(s) of language being focused on are presented, analyzed, and practiced, and the main subject-matter theme of the unit is introduced; this is followed by a series of skills-based sections in which the same and, often, additional, related language points are cycled through a series of reading, listening, speaking, and, sometimes, writing texts and related activities, most or all of them linked thematically to the subject matter introduced in the first section (see, e.g., Cunningham & Moor, 1999, pp. 48-57; Hutchinson, 2000, pp. 14-21; Soars & Soars, 1998, pp. 38-46). There are, of course, variations on this theme. Some textbooks contain extra sections on matters such as everyday expressions, (see, e.g., Soars & Soars, 2003, p. 93), "learner training" (see, e.g., Hutchinson, 1998, pp. 43 & 52), or include a more integrated, macro task section (see, e.g., Cunningham & Moor, 1999, pp. 96-9), and so on, but it appears that most modern textbooks use a unit structure along the lines indicated.

It can be argued that such a framework is, at root, a development and refinement of the traditional "PPP" (Presentation-Practice-Production) paradigm (Tomlinson, 2001, p. 69), the main differences being a greater focus on communicative practice and production work and the more consistent development of "carrier" content throughout the unit. Thus, the widespread use of a structure of this kind shows that, once again, in this respect as well, the evolution of the textbook since the late 1980s has continued to be not so much a process of wholesale "communicativization," but, rather, the grafting of a communicative "veneer" on to what has remained a basically language-focused stock. It seems likely that this has also occurred for reasons similar to those already discussed, i.e., primarily as a response to the wishes of end users, for whom an overall focus on language work, with a communicative gloss, appears to have remained the priority, despite the widespread criticisms of all forms of PPP in the professional literature in recent years.

However, it is also important to note the existence of exceptions to some aspects of a unit structure of this kind. For example, Crace & Wileman (2002) has a unit length which is much shorter than is typical. Likewise, the Interchange series (Richards et al., 2004), although also typically centering around a structural (or, in later parts, functional) area of language, and also possessing a common overall content theme, has a unit structure consisting of 10-12 "mini"-sections

(vs. the 4-5 longer sections typical of other textbooks), each of which usually occupies only half a page or less (vs. the whole page or more typical of many other textbooks). Each unit contains two "cycles," each comprising four core sections (three concerned with language presentation and practice, and one with language use), plus additional sections on, e.g., pronunciation, reading, speaking, etc. As a result, there is a proportionally greater amount of language focus than skills work in each unit.

In addition, each of the sections is relatively self-contained, in contrast to the way interconnections will often occur across sections in other textbooks, e.g., in the cases of Activity 4 on p. 56 and Language Focus 2 on p. 62 in Cunningham and Moor (1999), pp. 46-8 and pp. 80-1 in Soars & Soars (2003), and so on. The former also consist of only one, two, or sometimes three activities, none of which is subdivided, once again in contrast to the layout of many other textbooks, in which there are usually 4-5 activities, often subdivided, in each section. Furthermore, texts are relatively short, even at higher levels, as are activity rubrics, giving only the main instructions (once again, in contradistinction to many other coursebooks).

The primary difference between the Interchange unit structure and the more typical one described earlier is that the former has a much less "all-embracing" feel to it. Because sections are relatively shorter and less complex, and more self-contained, it seems reasonable to assume that the learning focus is likely to be clearer and the teaching-learning process easier to manage, with both learners and teachers having greater potential freedom to use their preferred modus operandi, since they do not have to engage in so much detail and at such length with texts and activities of the kind used in many other textbooks. The very comprehensiveness of the more typical textbook unit model - with its extensive program of interlinked texts and activities - can thus also be seen as its main potential inherent weakness. Such a structure commits teachers and learners to adopting a particular type of learning-teaching methodology - one that appears to be less widespread and less well suited than other models, among non-native speaker teachers, in particular (see, e.g., Medgyes, 1994: ch. 3) - or requires them to modify or reject it and attempt to use the book in a way to which it may not readily lend itself.

There are also other grounds for asking whether the current predominant textbook unit structure serves the best interests of teachers and learners as well as it might. An important concern in both earlier, as well as more recent, critiques of textbooks (e.g., Atlan, 1995; Tomlinson et al., 2001) is the extent to which, on the one hand, the content of texts and the setting of activities tend to promote an Anglo-Saxon worldview, and on the other, restrict the opportunities for learners to learn by using the context of their own lives. Shorter texts with fewer activities directly related to their content, in the manner of books such as Interchange, can, on the face of it, lessen the amount of "investment" learners (and teachers) are required to make before reaching a point where the language can be used meaningfully for their own purposes. This would seem especially important to take into account in view of the increasing recognition in recent years that English is

nowadays most often learned for the purpose of communication with other non-native speakers, rather than with native speakers (see, e.g., McKay, 2002).

In other words, just as the communicative approach has gradually come to be seen as a culture-specific rather than context-free methodology, so, perhaps, it is time for some of the design principles underpinning many modern international textbooks to be similarly reappraised, so that they become more attuned to building on the potential that exists within the main styles of language learning and teaching that exist around the world, and cater better to the communication needs of an English as an International Language situation, rather than reflecting so strongly so much in terms of learning methods and content that can be regarded as specific to Western/Anglo-Saxon culture. Rossner's stricture in this respect (see above) can, therefore, still be seen as applying to the majority of cases.

Conclusion

The earlier parts of this chapter have attempted to document a number of the main developments in language teaching materials design over recent years. An historical perspective was adopted, in order to clarify the extent to which the current situation can be seen as a reflection of previous trends. The findings indicate that, first, by and large, in terms of the aspects analyzed, earlier tendencies have been reinforced, with something of an orthodoxy in textbook design having now emerged. Nevertheless, as has also been adduced, other well-established and widely used publications exist which serve to provide something of an alternative to, and potential critique of, the more prevalent model. In addition, in both cases, the analysis indicated that the ontogenesis of the majority of the developments appeared to be related to perceptions of audience need, and were largely in contradistinction to concurrent trends in academic conceptualization.

In terms of the prevailing academic viewpoint, thus, the extent to which the current situation can be seen as representing a set of "advances" in materials design is, of course, problematic. As was noted at the outset, there has been a tendency in academic circles to view advances in textbook design as a function of the extent to which materials reflect succeeding developments in applied linguistics. However, in addition to its "applied science" orientation, such a stance assumes that writers and publishers, in taking the opposite tack, would willingly pursue a course that is against their best interests. It also implies that teachers' and learners' views, when they differ from academic perspectives, are to be discounted (cf. Widdowson, 2003, pp. 130-1).

A sounder stance would, therefore, appear to be one which views academic ideas more as means of illuminating and critiquing textbook design features, rather than being regarded as their prime determiner. Equally, a more humane attitude to the perceptions of textbook users and the efforts of writers and publishers seems called for. As accounts such as Bell and Gower (1998), Richards (2001, ch. 7), and Mares (2003) indicate, arriving at what might constitute a satisfactory textbook design is a difficult, complex, and highly skilled process,

involving, in particular, the notion of a compromise between what might be theoretically desirable and what is practicable and appropriate in audience terms.

Alternative perspectives of this kind might involve, for example, viewing the notion of authenticity less as a wholesale prescription and more as a safeguard against too much artificiality in texts and activities. Similarly, the less linear, more process-oriented focus on form approach to the teaching of language structures, rather than being seen as a replacement for the PPP paradigm used in many current textbooks, can instead be viewed more productively as a way of raising awareness of the limitations of a language teaching approach which is too segmentary and deterministic; and so on.

Such a realignment of attitudes might also desirably underpin a far larger program of empirical enquiry in applied linguistics into textbook design than has occurred hitherto. For example, very little academic research appears to have been done on attempting to establish what learners and teachers in various situations actually think about competing textbook designs, in terms of the various features which have formed the main focus of the analysis in earlier parts of this chapter. Equally, a larger number of studies of the kind conducted by Hutchinson (1996), into actual textbook use, would be a very useful source of further data for informing optimization of their design. And, of course, third-person accounts of the textbook design process itself would be likely to increase understanding in the area. In such ways, it is to be hoped, both a better basis for determining what might constitute advances in materials design, as well as enhanced means of enabling them to occur, might be achieved.

1 At the same time, the later C-R exercises adopt an inductive approach, a trend which has become widespread in other modern textbooks as well.

REFERENCES

Abbs, B. & Freebairn, I. (1980). Developing strategies: Student's book. London: Longman.

Atlan, M. (1995). Culture in EFL contexts: Classroom and coursebooks. Modern English Teacher 4, 2, 58-60.

Bell, J. & Gower, R. (1998). Writing course materials for the world: A great compromise. In B. Tomlinson (ed.), Materials development in language teaching (pp. 116-29). Cambridge: Cambridge University Press.

Carter, R., Hughes, R., & McCarthy, M. (1998). Telling tails: Grammar, the spoken language and materials development. In B. Tomlinson (ed.), Materials development in language teaching (pp. 67-86). Cambridge: Cambridge University Press.

Clarke, D. (1989). Communicative theory and its influence on materials production: State-of-the-art article. Language Teaching 22, 2, 73-86.

Crace, A. & Wileman, R. (2002). Language to go intermediate student's book. Harlow, UK: Pearson Education.

Cunningham, S. & Moor, P. (1999). Cutting edge intermediate student's book. Harlow, UK: Pearson Education.

Donovan, P. (1998). Piloting: A publisher's view. In B. Tomlinson (ed.), Materials development in language teaching (pp. 149-89). Cambridge: Cambridge University Press.

Ellis, R. (1993). Talking shop: Second language acquisition research: How does it help teachers? ELT Journal 47, 1, 3-11.

Frank, C., Beres, M., & Rinvolucri, M. (1989). Challenge to think. Oxford: Oxford University Press.

Hutchinson, E. G. (1996). What do teachers and learners actually do with textbooks? Teacher and learner use of a fisheries-based ELT textbook in the Philippines. Unpublished PhD thesis, Lancaster University.

Hutchinson, T. (1991). Hotline starter student's book. Oxford: Oxford University Press.

Hutchinson, T. (1998). New hotline elementary student's book. Oxford: Oxford University Press.

Hutchinson, T. (2000). Project 3 student's book. Oxford: Oxford University Press.

Littlejohn, A. (1992). Why are English language teaching materials the way they are? Unpublished PhD thesis, Lancaster University.

Mares, C. (2003). Writing a coursebook. In B. Tomlinson (ed.), Developing materials for language teaching (pp. 130-40). London: Continuum.

McKay, S. L. (2002). Teaching English as an International Language: Rethinking goals and approaches. Oxford: Oxford University Press.

Medgyes, P. (1994). The non-native teacher. London: Macmillan.

Nitta, R. & Gardner, S. (2005). Consciousness-raising and practice in ELT coursebooks. ELT Journal 59, 1, 3-13.

Nunan, D. (2004). Task-based language teaching. Cambridge: Cambridge University Press.

Prabhu, N. S. (1992). The dynamics of the language lesson. TESOL Quarterly 26, 2, 16176.

Richards, J. C. (2001). Curriculum development in language teaching. New York: Cambridge University Press.

Richards, J. C., Hull, J., & Proctor, S. (2004). Interchange: Student book 2. Cambridge: Cambridge University Press.

Richards, J. C., Hull, J., & Proctor, S. (2005). Interchange: Student book 1. Cambridge: Cambridge University Press.

Rossner, R. (1988). Materials for communicative language teaching and learning. Annual Review of Applied Linguistics 8, 140-63.

Sion, C. (ed.) (1985). Recipes for tired teachers. Reading, MA: Addison-Wesley.

Skehan, P. (1996). Second language acquisition research and task-based instruction. In D. Willis & J. Willis (eds.), Challenge and change in language teaching (pp. 17-30). Oxford: Macmillan Heinemann.

Skehan, P. (1998). A cognitive approach to language Learning. Oxford: Oxford University Press.

Soars, J. & Soars, L. (1986). Headway intermediate student's book. Oxford: Oxford University Press.

Soars, J. & Soars, L. (1998). New headway English course upper-intermediate. Oxford: Oxford University Press.

Soars, L. & Soars, J. (2003). New headway intermediate student's book. Oxford: Oxford University Press.

Soars, L., Soars, J., & Sayer, M. (2003). New headway intermediate: Teacher's book. Oxford: Oxford University Press.

Swan, M. (2006). Two out of three ain't enough: The essential ingredients of a language course. In B. Beaven (ed.), IATEFL Selections 2006 (pp. 45-54). Canterbury: IATEFL.

Swan, M. & Walter, C. (1984). The Cambridge English course 1. Cambridge: Cambridge University Press.

Swan, M. & Walter, C. (1985). The Cambridge English course 2. Cambridge: Cambridge University Press.

Swan, M. & Walter, C. (1990). The new Cambridge English course 2: Teacher's book. Cambridge: Cambridge University Press.

Tomlinson, B. (ed.) (1998). Materials development in language teaching. Cambridge: Cambridge University Press.

Tomlinson, B. (2001). Materials development. In R. Carter & D. Nunan (eds.), Teaching English to speakers of other languages (pp. 66-71). Cambridge: Cambridge University Press.

Tomlinson, B. (ed.) (2003). Developing materials for language teaching. London: Continuum.

Tomlinson, B., Dat, B., Masuhara, H., & Rubdy, R. (2001). Survey review: EFL courses for adults. ELT Journal 55, 1, 80-101.

Viney, P. & Viney, K. (2004a). In English elementary student's book. Oxford: Oxford University Press.

Viney, P. & Viney, K. (2004b). In English elementary teacher's book. Oxford: Oxford University Press.

Widdowson, H. G. (1979). Explorations in applied linguistics. Oxford: Oxford University Press.

Widdowson, H. G. (2000). On the limitations of linguistics applied. Applied Linguistics 21, 1, 3-25.

Widdowson, H. G. (2003). Defining issues in English language teaching. Oxford: Oxford University Press.

FURTHER READING

Byrd, P. (ed.) (1995). Material writers guide. Boston, MA, Heinle and Heinle.

Cunningsworth, A. (1995). Choosing your coursebook. Oxford: Heinemann.

Hall, D., Hidalgo, A. C., & Jacobs, G. M. (eds.) (1995). Getting started: Materials writers on

materials writing. Singapore: SEAMEO Regional Language Centre. Mishan, F. (2005). Designing authenticity into language learning materials. Bristol: Intellect. Renandya, W. A. (ed.) (2003). Methodology and materials design in language teaching, Anthology Series 44. Singapore: SEAMEO RELC.

19 Corpora in Language Teaching

JOHN FLOWERDEW

Introduction

Applications of corpus linguistics to language teaching began in the late eighties and early nineties. Examples of early work are Higgins and Johns (1984), Higgins (1988), Johns (1988, 1991), Tribble and Jones (1990), Stevens (1991), and J. Flowerdew (1993a). Most work in this area, as in other areas of applied linguistics and language teaching, has focused on English. However, some examples on other languages are Wichmann (1995), Ahmed and Davies (1997), Dodd (1997), King (1997), Kennedy and Miceli (2001), Rule et al. (2003), Belz (2004), Rule (2004), and Bolly (2005). Previous overviews of the field are Leech (1997), Aston (2001), Biber and Conrad (2001), Bernardini (2004), and Stubbs (2004).

Interest in corpus-based approaches to language teaching has developed quite rapidly in recent years, so that now there is a wealth of literature and, although less, still considerable application in this area. Application at the level of primary and secondary schools, however, has not kept pace with the considerable developments that are now going on at the tertiary level, especially in languages for specific purposes (LSP). Gavioli (2005), for example, is only able to cite one project at primary level (Sealey & Thompson, 2004).

One of the reasons for the relatively slow rate of classroom application has been the limitations of the technology. However, as Leech stated already in 1997, "computers have grown smaller, cheaper, and massively more powerful" (1997, p. 2). Since that statement, this trend has continued. In addition, more and more corpora have become available and it is easier to create personalized corpora. Furthermore, people are becoming increasingly computer literate and are therefore more easily introduced to the new approach. But if the start has been rather slow, this has a positive side, in that corpus applications to pedagogy have avoided, to quote Leech (1997, p. 4) again, the "bandwagon" effect. In developing more slowly, the risk of corpus-based approaches to language teaching following the path of the language laboratory, for example, with its its meteoric rise and ultimate demise, may be avoided.

The Handbook of Language Teaching Editedby MichaelH.Longand Catherin eJ.Doughty © 2009 BlackwellPublishDgLtd. ISBN:978-l-405-15489-5

What are the Principles in Corpus Linguistics that Can Be Applied to Language Teaching?

A corpus is a large database of language. Although the first corpora were relatively small - the Brown corpus (developed at Brown University, USA in the early 1960s) consisted of one million words - there now exist corpora consisting of hundreds of millions of words (e.g., the British National Corpus (BNC), 100 million words; and the Bank of English (COBUILD at Birmingham University, UK), over 500 million words). At the same time, however, much smaller corpora with as few as 100,000 words or less are being created all the time for specialist applications. It should be borne in mind, however, that, as pointed out by Gavioli and Aston (2001, p. 238), even the very large corpora contain less language than the average user will have experienced in their daily life.1 In addition, the linguistic content of corpora is different from what is experienced by individuals in real life, many of them consisting largely of written language. Furthermore, while each text is given equal weighting in a corpus, in real life some texts will hold more value and be experienced more times than others (poetry and religious texts, for example, might be highly valued and heard or read many times). While some corpora are kept in a "raw" state (e.g., Bank of English), many are "tagged" (i.e., coded, according to parts of speech) and "parsed" (i.e., analyzed for grammatical structure) (e.g., BNC).

The potential of corpus techniques for investigating patterns of lexis, grammar, semantics, pragmatics, and textual features is well established (e.g., Sinclair, 1991; McEnery & Wilson, 1996; Biber et al., 1998; Kennedy, 1998; Hunston, 2002; Stubbs, 2004). Most work in corpus linguistics to date has relied on word frequency lists, which provide criteria upon which to base a search, and keyword in context (KWIC) concordances, the presentation of every instance of a selected word, phrase, or particle in the corpus down the middle of the page, with a limited amount of cotext on either side. Figure 19.1 provides part of a concordance for the word meaning.

The power of the corpus approach lies in the combination of frequency data regarding all the words in a corpus and the verbal environment in which these words occur. This combination permits the detailed investigation of typical patterns of use of lexis and grammar - information which can be obtained at the click of a mouse. A concordance output may appear to be a reified object, but this is not the case, because it may be ordered in various ways from left and right of the keyword or phrase and these changes reveal different collocational and grammatical patterns. Some critics have complained that concordance lines provide no information about situational context. This is also indeed accepted by proponents (e.g., Sinclair, 1991, p. 34; McEnery & Wilson, 1996, p. 79). However, it may be pointed out that situational information can be built into or accompany a corpus and that there is no reason why corpus evidence may not be supported by ethnographic investigation (L. Flowerdew, 2005, 2008). On the other hand, as Stubbs (2004, p. 108) notes, practice has demonstrated that the meaning of a

dictionary like we all do and the first meaning in the dictionary is all to do with # a proper Italian pronuncation for that meaning leaf or sheet and so really in its basi heet and so really in its basic basic # meaning portfolio means a case for carrying loo ussed before she'd also assessed orally meaning they want to check if her teeth are oka ather different it has a very different meaning and that is that you can't override tes following a step input oscillates meaning it it goes backwards and forwards and that o push out the residual demand curve meaning it would comm not only command a e static properties of the equilibrium meaning we shifted one of the curves and look kets i'm going to use the word bundles meaning a collection of or a combination of tw y're falling apart i think what you're meaning is that you've got an i # an interfacia and it fits the syntax and it fits the meaning that's it i'll # i'll go for that hy defined concepts which we all know the meaning of but don't need or not even able to d pecies and i take you understand the meaning of those numbers here minus-two m index and a little label at the top N meaning the time index so that's the times th en out # and that's the same same meaning the reason the word's choosed in this holism bec # for our intentionality meaning our human concepts our human understan e that it has managed to bleed all the meaning out of maternity we don't have matria but you under what you understand the meaning of what i'm saying here no i don't but in economics it has a very specific meaning input-output models were invented by er goods yeah but this is the literal meaning of this if you increase final demand e in one country and have very little meaning in another but it has # been a key i gn so we must talk about the government meaning the legislature and the executive the u're part of what this can be taken as meaning is that you can consent to be under the solution but it's a compromise solution meaning that each party will have to give up q ed of snakes that's F-L-plus S-N-minus meaning this model monkey was afraid of flow nitive case possessive case of a word meaning kind a kind of something right a nou about three-thousand years ago # word meaning exactly the same thing is like this

Figure 19.1 Concordance of the word meaning from a corpus of academic lectures

search word or phrase is often identifiable within a short span of cotext, enough to fit into one line on the computer screen. Furthermore, most concordancers allow the user to inspect the wider cotext of a selected instance, so the analyst is not limited to the single corpus line.

As already mentioned, corpus techniques have created new knowledge about the behaviour of lexis, grammar, semantics, pragmatics, and textual features. Because corpus linguistics is based on the theory that language varies according to context - across space and time - the potential for finding out new facts about language is infinite. If this theoretical insight is applied to pedagogy, then the case for the use of corpora in teaching becomes very powerful. Because no dictionary or grammar is able to fully describe the language, the educationist, whether materials designer or classroom practitioner - or indeed learners themselves -may play an important role in identifying regularities in the language which are not to be found in such texts.

Another proven benefit of corpus-based approaches to the study of language is that analysis is based on empirical, as opposed to introspected or elicited, data, "real" language as many proponents refer to it.2 As Aston (2001, pp. 7-8) has pointed out (see also J. Flowerdew, 1996), native-speaker intuition about language is often wrong - on the one hand, many uses included in traditional descriptions

do not occur with any frequency in large general corpora and, on the other hand, many uses which occur in corpus data are not recognized in traditional descriptions. This means that teachers and learners have been being given inaccurate and incomplete descriptions of the language.

What Information Can the Corpus Provide?

A corpus can yield various types of information which can be of potential use in language pedagogy. It can provide information about the behavior of words, multi-word phrases, grammatical patterns, semantic and pragmatic features, and textual properties. Knowledge of these features and their relative frequencies can be helpful to language practitioners in deciding what items to teach and when to teach them, as well as, importantly, providing input for reference materials. In this section various concepts regarding different aspects of language behavior will be presented, each followed by an indication of how the concept might be applied in language pedagogy.

Word frequency

Concept

At the most basic level, the corpus can provide word lists organized either according to frequency or alphabetically. Used in conjunction with the concordancer, frequency is not limited to the word forms, but may extend to the different meanings of a given word or phrase; the editing function of the concordancer can be used to group the items according to the different meanings. Frequency data can also be obtained for recurrent sequences (variously referred to as n-grams, pre-fabs, and lexical bundles) e.g., I don't know, all of a sudden, all over the place, don't have a clue. Furthermore, relative frequencies between two or more corpora can be calculated, those words occurring significantly more frequently in one corpus than another being referred to as keywords (Scott & Tribble, 2006).

Application

Frequency information is immensely useful in helping to prioritze what to teach. Aston (2001, p. 8) quite rightly mentions other relevant criteria: range, availability, coverage, learnability, and prototypicality (see also Widdowson, 2004, p. 87; Kaltenbock & Mehlmauer-Larcher, 2005, p. 77), but, other things being equal, frequency of occurrence is an important criterion for syllabus design and teaching. A considerable time ago Nattinger and de Carrico (1992) recommended the application of lexical phrases to language teaching. If one takes the view that all language teaching is LSP teaching, insofar as learners need to acquire a range of registers and genres, then comparative data, as provided by keyword analysis, will provide information regarding what to teach and when to teach in relation to specific genres and registers (see Scott & Tribble, 2006).

Collocation

Concept

Collocation is concerned with how words typically occur (or do not occur) together. Recurrent patterns highlighted by the concordancer will indicate typical collocations, although programs can provide lists of collocates. Hunston (2002, p. 12) gives the example of shed, which collocates with light, tears, garden, jobs, blood, cents, image, pounds, staff, skin, and clothes. Typically different collocates will affect the precise meaning of the word, e.g., shed blood means to suffer, shed pounds means to lose weight, and shed image means a deliberate changing of how one is perceived (Hunston, 2002, p. 12).

Application

How does one correct the learner who says "I will open the air-conditioner," where the collocates are not appropriate? One way, of course is to explain that in standard English one says "switch on" rather than "open" when referring to an air-conditioner or other electrical appliance. However, this lesson is likely to be more powerful and therefore more effective if, instead, the learner can look at concordances of the word open + noun phrase and see that while open collocates with other concrete nouns such as gate, door, and window, there are no instances of open + air-conditioner. On the other hand, a concordance of air-conditioner will probably yield numerous examples with switch on and switch off. In addition to this use with students, the concordancer can also give confidence to teachers, especially to less proficient non-native speakers, where they are unsure of their intuitions (J. Flowerdew, 1996).

Colligation

Concept

A distinction can be made between collocation, which is the combination of individual words, and colligation, which refers to how lexical words are associated with particular grammatical words or categories. Hunston (2002, p. 13), again, gives the example of the word head which has the following colligations: of, over, on, back, and off. Again the colligations affect the meaning of the word, thus Hunston gives examples such as head of department, hit someone over the head, throw one's head back.

Application

Kaltenbock and Mehlmauer-Larcher (2005, pp. 73-4) give some examples of pedagogical activities designed to develop colligational awareness. For example, in one simple task students are given a set of sentences where deleted prepositions have to be inserted after searching a corpus:

The building is adjacent . . . the train station. It is usually a good idea to abide . . . the law. You should give clear indication .. . your intentions. He was aghast . . . the violence he witnessed.

In another example task (with the International Corpus of English (ICE) GB corpus, which is tagged for parts of speech), students consider verb complementation with the gerund and the infinitive:

TASK: What can the corpus tell us about the difference in meaning/use between remember doing something and remember to do something?) (Try [searching for]: "remember to" and "remember" <V> = remember followed by a verb.

Kaltenbock and Mehlmauer-Larcher (2005) make the important point that corpus queries such as those required for these tasks require less "expert knowledge" than would be needed if a reference grammar were used, with the knowledge of grammatical metalanguage that would be implied for the latter task.

Semantic preference

Concept

Here we are concerned with how a word or phrase relates to a group of collocating words that (1) share an element of meaning, (2) are related to particular genres or registers, or (3) belong to lexical sets in terms of synonymy, meronymy, antonymy, etc. Semantic preference is arrived at by sorting collocates into groups based on semantic relations such as those just mentioned. The specific semantic preference is labelled by a gloss, such as "words or phrases relating to measurement," "words or phrases belonging to the register of production engineering," or "words or phrases relating to history."

Application

Semantic field theory, which can be seen as an introspective precursor of semantic preference, has been applied (mostly intuitively) in language teaching for a very long time (Corder, 1973, p. 316). Indeed it can be seen as closely related to situational ("at the post office," "at the airport," "in the supermarket," "in the office," etc.) and topical ("travel," "shopping," "family," etc.) syllabuses. It is also implicitly applied in notional syllabuses (Wilkins, 1972). The assumption here is that certain lexical (and grammatical) items belonging typically in given fields are likely to co-occur and can be learned together in semantic sets. However, a corpus approach takes us beyond introspection to identify empirically established relationships. The choice of corpus here is crucial, larger corpora being more reliable, because smaller corpora will not be likely to provide enough data to determine general preferences. On the other hand, specialist corpora consisting of specific genres or registers have great potential for application to LSP.

Semantic prosody

Concept

If semantic preference can tell us about the semantics of a word or phrase, "semantic prosody" (Sinclair, 1991; Louw, 1993), or for Stubbs (2001) "discourse prosody," can tell us about typical pragmatic values - the attitude or evaluation a speaker or writer attaches to what they are saying. Semantic prosody is similar to connotation. However, it does not just apply to a single word, but to the word and its association with its collocates. Thus, to take an example from Stubbs (1996), the word cause typically collocates with negatively loaded words - e.g., accident, concern, damage, death, trouble - and thereby takes on a negative semantic prosody; provide, on the other hand, is typically used with positive collocates -e.g., aid, care, food, opportunities, relief, support - and thus takes on a positive semantic prosody. Most studies of semantic prosody describe examples in simple terms of positive or negative evaluation. However, it seems that finer grained analysis is possible. Thus Hunston (2002, p. 141) gives the example of sit through, which is often used with lexical items which indicate boring or lengthy things.

Application

The analysis of the semantic prosodies associated with the lexical items in a corpus is a way to acquire context knowledge which is important for writers trying to master tasks within a specific genre (Tribble, 2000). This sort of information is now starting to be incorporated into dictionaries, but a learning activity in the form of analyzing words in context and identifying their semantic prosodies might be a more effective learning strategy, insofar as learners are more likely to remember what they themselves have discovered.3

Register and genre

Concept

Research in corpus linguistics has done much to show how patterns may vary across various registers or genres. As Biber and Conrad (2001, p. 332) put it, "strong patterns of use in one register often represent only weak patterns in other registers." To illustrate this, Biber and Conrad show, for example, how the 12 most frequent lexical verbs (say, get, go, know, think, see, make, come, take, want, give, and mean) in a corpus of 20 million words drawn from four registers (conversation, fiction, newspaper language, and academic prose) are very unequally distributed across the four registers. These verbs, for example, represent 45 percent of all verbs in conversation versus only 11 percent for academic prose.

Application

Biber and Conrad (2001) argue that the verbs referred to above should be given priority in pedagogy. In practice, however, they note that low-level ESL grammar books tend not to use these verbs, preferring activity verbs such as eat, play, work,

run, travel, and study, which, as they concede in a footnote, are easier to learn. Nevertheless, Biber and Conrad argue that just because they are more difficult to learn does not mean that the high-frequency verbs should be neglected, "as these are the ones students will most often hear in their day-to-day interactions with native speakers" (p. 333).4

At a more micro level, working with small corpora composed of specific text types, Gavioli (2001) has shown how particular recurrent patterns tend to occur within such corpora. Comparing two corpora, one composed of lonely hearts ads and the other of letters to a newspaper agony aunt, for example, she (or, in fact, her students, because in her paper Gavioli is showing how corpus analysis can be done by learners (see below)) shows certain similarities and differences. Taking the adjectives pretty, attractive, and beautiful, for example, she shows that pretty and attractive always refer to people's physical appearance in both corpora. Beautiful, in the letters, however, also refers to music and the home. In addition, neither pretty nor attractive occurs in a series of adjectives. However, beautiful occurs in a co-ordinate pattern with another positive adjective, in phrases such as mature and beautiful; beautiful and well-behaved; beautiful and wonderful; sweet and beautiful. This is not the sort of information that can be found in reference grammars or dictionaries and it provides a strong argument for a corpus-based approach to the development of genre awareness (J. Flowerdew, 1993b).

What Corpora?

One of the problems with applying corpora to language teaching is deciding which the most appropriate corpora are. As Leech (1997, p. 18) has pointed out, "the corpora which are easiest to compile are not necessarily those which are most useful for language learning purposes." Not all corpora will be suitable for all learners.

Until recently, the most pressing problem in this area was the dearth of spoken corpora, most corpora being wholly or primarily made up of written language. The reason for this is simple. It is difficult and expensive to collect spoken language, which then has to be recorded and transcribed. It is true that spoken corpora are starting to be created - for example the CANCODE corpus of spoken English developed jointly by the University of Nottingham and Cambridge University Press - but there is still an emphasis on the written word (not to mention problems of accessibility). The BNC, for example, has 90 million written words compared to 10 million of speech). Given the emphasis in modern-day language pedagogy on the spoken word, this is a serious problem.

Another problem is that most corpora are based on native-speaker models. In a climate where there is much discussion of the role of world Englishes in language pedagogy, the use of native-speaker models may be questioned (Hunston, 2002). This does not just concern lexico-grammar. As Carter (1998b, p. 49) has demonstrated, colloquial speech is deeply embedded in cultural understandings. The simplest of phrases may require knowledge of the culture for understanding.

Among many examples, Carter provides the following service encounter from CANCODE:

[In a fish and chip shop]

A: Can I have chips, beans, and a sausage?

B: Chips, beans, and a sausage.

A: Yeah.

B: Wrapped up?

A: Open please.

Carter points out how in this extract the word "open" in this particular context is used as an antonym of "wrapped up" and "carries a specific cultural meaning of food being served in paper so that it can be eaten immediately, even perhaps while walking home" (p. 48). Carter asks to what extent such cultural allusions should be removed. Furthermore, he asks how relevant it is to be able to make observations such as that fish and chip shops in Britain serve not only fish and chips, but also other food, such as sausages, burgers, and curry.

The foregoing suggests that corpora made up of different language varieties might be needed. Hong Kong learners or Filipino learners, it might be argued, should have as their target educated Hong Kong or educated Filipino English, not British English. Similarly, it would seem sensible that learners of French in Canada might want a standard and hence a corpus based on Canadian French rather than the metropolitan variety. The problem is being addressed to a degree, for English, with the ICE corpora, referred to earlier, a suite of corpora of 15 different national/regional varieties, such as Australian English, British English, East African English, Filipino English, Indian English, etc. Given the complexity of coordinating and collecting such a range of corpora, however, it is perhaps understandable that these corpora are relatively small, at one million words each. Of course, the question of what standard to adopt is itself controversial. To take the example of Canadian French, many learners want to acquire the metropolitan standard, even though they will be using their French in Canada. To take another example, at a recent corpus conference an Indian member of the audience was asked if Indians would want to learn Indian English. His answer was that they would definitely not want to be associated with such a variety which they did not even acknowledge as such, preferring so-called "standard" English. This raises the question of language rights, in this case the rights of the learners (or their parents) to have the target variety that they want. In addition, where regional or local varieties have developed, in a globalizing world, with all that goes with it - mass migration, mass tourism, international business travel, the internationalization of (especially tertiary) education, use of the Internet and other electronic communication devices, and the internationalization of popular culture and mass media - learners may need not only the local variety, but also some standard for international intelligibility.

An alternative solution in terms of appropriate models might lie in lingua franca corpora, corpora composed of language produced by proficient non-

native speakers who are interacting with each other or with native speakers. In English, it is said that more English is spoken in the world among non-natives than natives, so lingua franca corpora might seem a logical way to go. However, research to date has not come up with systematic descriptions of such language and it is questionable whether - certainly at the level of phraseology and the grammatical code - such systematic patterns are discernible. Interestingly Anna Mauranen (2006; personal communication June 2006) has identified in her ELFA (English as a Lingua Franca for Academic Purposes) corpus certain pragmatic regularities, such as greater use of grammatical rephrasing and a greater tolerance for ambiguity. But she has not identified any new lexico-grammatical or phraseological regularities.

Further confusing the picture as regards suitable corpora, there are other learner differences that need to be taken into account. For example, a model for young learners might be child language, teenagers may want a teenage model, women and men might want models of the speech of either gender; then again, learners may want specific academic or professional language (see below on corpora and LSP). It is true that there are different types of corpora or sub-corpora (for example, the CHILDES corpus of children's speech (MacWhinney & Snow, 1991) and the British National Corpus has a section on young people's spoken language, referred to separately as the COLT corpus (Stenstrom et al., 2002)). However, these corpora or sub-corpora have not been designed with language teaching specifically in mind and their suitability, certainly in terms of their size and representativeness, might be questioned.5

Finally, the authenticity of corpus data may mean that it is difficult for less advanced learners to process. Perhaps corpora of simplified language might be needed for such learners, or some sort of filter which removes concordances which contain vocabulary items which do not occur in a pre-established list (Kuo et al., 2001).

Applications

Applications of corpus linguistics to language teaching may be direct or indirect (Stubbs, 2004). A direct application would be advanced users of academic English using a corpus of the language of their speciality to assist them in writing academic papers (see Lee and Swales, 2006 for an account of such a procedure). Indirect applications would be the application of corpus findings to the creation or refinement of dictionaries, reference grammars, and pedagogic materials.

Indirect applications

Use in developing reference material

One of the first applications in this area was Collins COBUILD English Language Dictionary (1987) edited by John Sinclair; Other dictionaries have made use of

corpora to a greater or lesser extent, e.g., Longman Dictionary of Contemporary English, Oxford Advanced Learner's Dictionary, Macmillan English Dictionary for Advanced Learners. As Leech (1997, p. 14) points out, some of the advantages of corpus-based lexicography are that corpus data:

• can be searched quickly and exhaustively,

• can provide frequency data,

• can be easily processed to produce updated lists of words,

• can provide authentic examples for citation,

• can readily be used by lexicographical teams for updating and verifying other levels of descriptions such as dictionary definitions.

A precursor of grammars totally based on corpus data was A Comphrehensive Grammar of the English Language (Quirk et al., 1985), which relied on manually collected examples of use that were stored in a giant database. This might be called a corpus-informed grammar. The first grammar to be fully based on a corpus, what might be called a corpus-driven grammar, was Collins COBUILD English Grammar (Sinclair, 1990). This has been followed by the Longman Grammar of Spoken and Written English (Biber et al., 1999), and, more recently, by the Cambridge Grammar of English (Carter & McCarthy, 2006). At this point, it might be noted that, while the Longman Grammar is no doubt a great achievement, especially in the great advance achieved in the incorporation of frequency data according to four different domains of use, the corpus-driven grammars, especially COBUILD and Cambridge tend to be inconsistent in their coverage of the basic features of the language. In terms of comprehensiveness, Quirk et al. (1985) cannot be beaten. No doubt with further work the comprehensiveness of corpus-driven grammars will improve.

Pedagogic materials

Again, Collins (now HarperCollins), with John Sinclair as editor in chief, were the first here, with an extensive series called Collins COBUILD English Guides. Titles focused on such linguistic features as Determiners (Berry, 1996)